A Novel Interaction between hScrib and PP1γ Downregulates ERK Signaling and Suppresses Oncogene-Induced Cell Transformation

Previous studies have shown that the cell polarity regulator hScrib interacts with, and consequently controls, the ERK signaling pathway. This interaction occurs through two well-conserved Kinase Interacting Motifs, which allow hScrib to bind ERK1 directly, resulting in a reduction in the levels of phospho-ERK. This suggests that hScrib might recruit a phosphatase to regulate this signaling pathway. Using a proteomic approach we now show that Protein Phosphatase 1γ (PP1γ) is a major interacting partner of hScrib. This interaction is direct and occurs through a conserved PP1γ interaction motif on the hScrib protein, and this interaction appears to be required for hScrib's ability to downregulate ERK phosphorylation. In addition, hScrib also controls the pattern of PP1γ localization, where loss of hScrib enhances the nuclear translocation of PP1γ. Furthermore, we also show that the ability of hScrib to interact with PP1γ is important for the ability of hScrib to suppress oncogene-induced transformation of primary rodent cells. Taken together, these results demonstrate that hScrib acts as a scaffold to integrate the control of the PP1γ and ERK signaling pathways and explains how disruption of hScrib localisation can contribute towards the development of human malignancy.


Introduction
The control of cell polarity and the maintenance of tissue architecture are intimately related and are, in part, controlled by a tri-partite macromolecular signaling complex consisting of the Scrib complex, the Par complex and the Crumbs complex [1,2]. Through a series of antagonistic interactions the components of these three complexes control a variety of downstream signaling pathways that, in turn, directly contribute to the regulation of cell polarity and cell proliferation [3]. It is now clear that the loss of control of these pathways is a common event during the development of diverse human malignancies [1,[4][5][6][7]. These defects are particularly evident at the later stages of malignant progression, and a variety of studies in both Drosophila and transgenic mice have provided additional supporting evidence of tumour suppressor activity for the various components of these signaling complexes [8][9][10][11].
The hScrib complex consists of three proteins, hScrib, hDlg1 and Hugl-1/2. In Drosophila, loss of either Scrib or Dlg produces imaginal disc overgrowth with invasive characteristics [8] [12], phenotypes that can be functionally complemented by the mammalian equivalents [13][14][15]. More recently Scrib has been implicated in the control of the JNK and ERK signaling cascades, and loss of hScrib appears to enhance the effects of the Ras and Myc oncogenes, and can contribute to mammary tumour development [16][17][18][19][20][21]. Recent studies have also demonstrated that hScrib can interact directly with ERK, and control both ERK activation and its nuclear translocation [19]. However, the physical interaction between ERK and hScrib is not sufficient to explain the inactivation of ERK, since high levels of hScrib appear capable of directly reducing the levels of ERK phosphorylation [19]. Since hScrib has no known phosphatase activity itself, it therefore seemed possible that a protein phosphatase might be recruited by hScrib to fully inactivate the ERK signaling pathway.
Control of ERK activation reflects an exquisite balance between the activities of the activating kinases and the de-activating protein phosphatases. Activated ERK can translocate to the nucleus, where it activates several transcription factors and also phosphorylates cytoplasmic and nuclear kinases [22][23][24]. Since phosphorylation of both the threonine and tyrosine residues of ERK is required for its activation, dephosphorylation of either is sufficient for its inactivation [25]. There are several reports demonstrating that dephosphorylation of active ERK can be achieved by tyrosine-specific phosphatases, by serine/threonine-specific phos-phatases or by dual specificity (threonine/tyrosine) protein phosphatases [26][27][28][29]. One of the important negative regulators of the ERK signaling pathway is PP2A, a member of the PPP family of protein serine/threonine phosphatases which also includes PP1 [30,31]. However, PP2A is thought to exert its activity mainly upon other activating kinases within the cascade, rather than upon ERK itself [32][33][34]. In addition, recent studies have also shown that hScrib can directly regulate the Akt signaling cascade by recruitment of the protein phosphatase PHLPP1 to the plasma membrane, thereby resulting in de-phosphorylation of Akt [35]. Here, we have used a proteomic approach to extend our investigations into the regulation of the ERK signaling cascade by hScrib. We now show that hScrib interacts with PP1c, and that this association correlates with the ability of hScrib to downregulate ERK activation. We also provide compelling evidence that hScrib directly contributes to the regulation of PP1c function by controlling its translocation between the cytoplasm and the nucleus. Thus, loss of hScrib expression results in both ERK activation and aberrant nuclear translocation of PP1c.

Cells and treatments
HEK293 (human embryonic kidney cells) and HaCaT (Human keratinocytes) were obtained from ATCC [36,37]. HEK293, HaCaT and Baby Rat Kidney (BRK) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, penicillin-streptomycin (100 U/mL) and glutamine (300 mg/mL) in a humidified 5%CO 2 incubator. Transfection was carried out using calcium phosphate precipitation as described previously [37] or using Lipofectamine 2000 (Invitrogen) according to the manufacturer's protocol. The depleted Scribble cell lines were generated as described previously [19]. Cell transformation assays were done using BRK cells obtained from 9 day old Wistar rats with a combination of HPV-16 E7 and EJ-ras, plus the appropriate hScrib and PP1c expression plasmids. Cells were placed under G418 selection for three weeks, and then fixed and stained.

Plasmids
The wild type pCDNA3-HA-PP1c was the kind gift of Dr. Wilhelm Krek (Swiss Federal Institute of Technology (ETH) Zurich). The wild type HA-tagged pcDNA hScrib expression plasmid and the truncated mutant pGEX hScrib PDZ1-C, PDZ1-4, S1445A, S1445D, and CT expression plasmids have been described previously [19]. The L1266Y1268RAA mutation (KADA) to doubly change the Leucine (L) and Tyrosine (Y) residues to Alanine (A) in hScrib was done using the QuikChange site-directed mutagenesis kit from Stratagene Cloning Systems (Celbio) according to the manufacturer's instruction. The mutants were confirmed by DNA sequencing. See Figure S1 for a detailed description of the location of the different hScrib mutations.

Immunofluorescence and Microscopy
For immunofluorescence cells were grown on glass coverslips and fixed in 3.7% paraformaldehyde in PBS for 20 mins at room temperature. After washing in PBS the cells were permeabilised in PBS/0.1% Triton for 5 mins, washed extensively in PBS and then incubated with primary antibody diluted in PBS for 1 hour followed by the appropriately conjugated secondary antibodies. Secondary antibodies conjugated to Alexa Fluor 488 or 548 were obtained from Invitrogen. The cells were then washed several times in water and mounted on glass slides. Cells were visualized by using a Zeiss Axiovert 100 M microscope attached to a LSM 510 confocal unit.

siRNA transfection
HEK293 cells were seeded on 6 cm dishes and transfected using Lipofectamine 2000 (Invitrogen) with control siRNA against Luciferase (siLuc), or siRNA against hScrib and PP1c sequences (Dharmancon). 48 hours post-transfection cells were harvested and total cells extracts or cell fractionated extracts were then analysed by western blotting.

Fusion protein purification and in vitro binding assays
GST-tagged fusion proteins were expressed and purified as described previously [19]. Proteins were translated in vitro using the Promega TNT kit and radiolabelled with ( 35 S) cysteine or ( 35 S) methionine (Perkin Elmer). Equal amounts of in vitro-translated proteins were added to GST fusion proteins bound to glutathione agarose (Sigma) and incubated for 1 hour at 46C. After extensive washing with PBS containing 0.25% NP-40, or as otherwise indicated, the bound proteins were analysed by SDS-PAGE and autoradiography.

In vitro phosphorylation
Purified GST fusion proteins were incubated with commercially purified ERK1 (Cell Signaling Technology) or PKA (Promega) for 20 mins at 30uC in phosphorylation buffer (0.25 M Tris pH7.5, 1 M MgCl 2 , 3 M NaCl, 0.3 mM aprotinin, 1 mM Pepstatin) or using the kinase buffer supplied by New England Biolabs supplemented with 56 nM ( 32 P) c-ATP (Perkin Elmer) and 10 mM ATP following the manufacturer's instruction. After extensive washing, the phosphorylated proteins were monitored by SDS-PAGE and autoradiography.

Mass spectrometry analysis
HEK293 cells were transfected with HA-tagged Scrib and after 24 hours the cells were extracted in mass spectrometry lysis buffer (50 mM Hepes pH 7.4, 150 mM NaCl, 50 mM NaF, 1 mM EDTA, 0.25% NP40) and extracts incubated with anti-HA beads (Sigma) for 2-3 hours on a rotating wheel at 4uC. The beads were then extensively washed with PBS, dried and the immunoprecipitated proteins were subjected to proteomic analysis as described previously [38].

Subcellular Fractionation assays
Differential extraction of HEK 293 cells to obtain cytoplasmic, membrane, cytoskeleton, and nuclear fractions was performed using the Calbiochem ProteoExtract Fractionation Kit according to the manufacturer's instructions. To inhibit phosphatase activity during the preparation of cell lysates, phosphatase inhibitors (1 mM Na 3 VO 4 , 1 mM b-Glycerophosphate, 2.5 mM Sodium Pyrophosphate, 1 mM Sodium Fluoride) were also included.

Immunoprecipitation and Western blotting
Total cellular extracts were prepared by directly lysing cells from dishes in SDS lysis buffer. Alternatively cells were lysed in either E1A buffer (25 mM HEPES pH 7.0, 0.1% NP-40, 150 mM NaCl, plus protease inhibitor cocktail; Calbiochem) or RIPA buffer (50 mM Tris HCl pH 7.4, 1% NP-40, 150 mM NaCl, 1 mM EDTA, plus protease inhibitor cocktail; Calbiochem) and the cell extracts were analysed by SDS-PAGE and western blotting. For immunoprecipitations, total cell lysates were transferred into a tube of equilibrated EZview Red Anti-HA Affinity Gel beads (Sigma), and incubated for 2 hours at 4uC. Immunoprecipitates were extensively washed four times in lysis buffer and solubilised in SDS-PAGE sample buffer. For western blotting, 0.45 mm nitrocellulose membrane (Schleicher and Schuell) was used and membranes were blocked for 1 hour at 37uC in 10% milk/PBS followed by incubation with the appropriate primary antibody diluted in 10% milk/0.5% Tween 20 for 1 hour. After Figure 1. Interaction between hScrib and PP1c in vivo. A) Results from the mass spectroscopy analysis of hScrib containing immunoprecipitates identified 6 peptides (indicated) corresponding to PP1c. B) In vitro translated PP1c (upper panels) and PP2A subunit A (lower panels) were incubated for 1 hour at 4uC with purified GST-hScribP1-C or GST alone immobilized on Glutathione agarose. After extensive washing, the bound proteins were analysed by SDS-PAGE and autoradiography which are shown in each of the upper panels. The gels were rehydrated and stained with Coomassie to show equal levels of GST loading in the respective lower panels. C) Endogenous PP1c was immunoprecipitated from HaCaT (upper panels) and HEK 293 cells (lower panels), with pre-immune antibody used as control. The immunoprecipitated proteins were then analysed by western blotting using anti-hScrib and anti-PP1c antibodies. doi:10.1371/journal.pone.0053752.g001 several washings with PBS 0.5% Tween 20, secondary antibodies conjugated with HRP (DAKO) in 10% milk/0.5% Tween 20 were incubated for 1 hour. Blots were developed using Amersham ECL reagents according to the manufacturer's instructions.

PP1c is a direct binding partner of hScrib
Based on our previous studies we reasoned that down-regulation of ERK phosphorylation by hScrib might involve the recruitment of a protein phosphatase [19]. To investigate this possibility we performed proteomic analyses to identify additional interacting partners of hScrib. HEK293 cells were transfected with an HAtagged hScrib expression plasmid and after 24 hours the cells were extracted, and hScrib-bound protein complexes were immunoprecipitated with anti-HA agarose beads and then subjected to mass spectroscopy analysis. Several previously reported interacting partners were identified, including vimentin. However, of the novel interacting partners, the most prominent phosphatase identified was the catalytic subunit of PP1c ( Figure 1A), a major eukaryotic serine/threonine protein phosphatase. To investigate whether hScrib can interact with PP1c, an in vitro pull-down assay was performed using purified GST-hScrib P1-C fusion protein and in vitro translated radiolabeled PP1c. For comparison a similar assay was also done using in vitro translated radiolabeled protein phosphatase 2A (PP2A). After extensive washing the bound PP1c and PP2A were detected by SDS PAGE and autoradiography, and the results in Figure 1B demonstrate strong interaction between hScrib and PP1c. In contrast, no interaction was observed between hScrib and PP2A, confirming the specificity of the association between hScrib and PP1c. To determine whether endogenous hScrib and PP1c could exist in a complex in vivo, immunoprecipitations were performed on cell extracts from HEK293 and HaCaT epithelial cells using anti-PP1c antibody. Co-immunoprecipitated hScrib was then detected by western blotting, and the results in Figure 1C show a significant degree of co-immunoprecipitation of hScrib with PP1c in both cell lines. Taken together, these results demonstrate that hScrib and PP1c can exist as a complex in vivo.

hScrib interacts with PP1c through a conserved RVxF motif
The PP1 holoenzyme is composed of a catalytic subunit and several regulatory subunits, which target the catalytic subunit to specific subcellular locations. The RVxF motif is a short conserved PP1-binding motif initially identified in previous studies showing that these residues can block the interaction of regulatory subunits with the PP1 catalytic subunit [39]. As shown in Figure 2A [40,41] spanning residues 1265-1268. This sequence is also highly conserved in mammalian Scrib proteins, but is absent in Drosphila. Based on previous studies, mutation of the L and Y residues would be expected to severely perturb the interaction with PP1 [39][40][41][42]. To investigate whether this KLDY motif is responsible for the capacity of hScrib to bind to PP1c, a panel of GST-hScrib fusion proteins consisting of the full length (FL), two truncated proteins encompassing PDZ domains 1-4 (P1-4) and the carboxy terminal third of hScrib (CT), plus a full length hScrib with the KLDY/KADA mutation, were used in pull-down assays with in vitro translated radiolabeled PP1c. The levels of bound PP1c were then assessed by SDS PAGE and autoradiography and, as can be seen from Figure 2B, PP1c binds to the carboxy terminal region of hScrib which contains the predicted PP1 binding motif. Furthermore the KLDY/KADA mutation significantly decreases the capacity of PP1c to interact with hScrib, confirming that the major site of interaction is through the KLDY consensus motif.

hScrib and ERK are substrates of PP1c
We have previously shown that hScrib is a substrate for both PKA and ERK. Furthermore, hScrib can downregulate ERK activation through a direct protein-protein interaction [19], although the precise mechanism by which hScrib can achieve this is still unknown. We therefore wanted to determine whether phosphorylation of hScrib by either PKA or ERK1 could influence the ability of hScrib to interact with PP1c and, furthermore, whether hScrib itself was a substrate of PP1c. To do this, purified GST-hScrib fusion protein was subject to phosphorylation by either PKA or ERK1 in the presence of non-radiolabeled ATP, and after extensive washing binding assays  . hScrib is a substrate of PP1c. A) Purified GST-hScrib fusion protein was in vitro phosphorylated with purified PKA or ERK1 as described previously (19) and then incubated with PP1c for 20 mins at 30uC. Bound PP1cwas detected by western blotting with anti PP1c antibody. The lower panel shows the ponceau stain of the nitrocellulose, and the upper right panel shows the quantitations from three independent experiments. Note that hScrib phosphorylated by PKA exhibits increased association with PP1c. B) Purified PP1c was incubated with purified full length wild type GST-hScrib fusion protein (P1-C), the mutants S1445A, S1445D or GST alone as a control. After extensive washing the bound PP1c was ascertained by western blotting. The upper panel shows the result of the western blot, with the 20% input of PP1c also shown for comparison. The lower panel shows the ponceau stain of the nitrocellulose. The histogram shows the quantitation from three independent experiments. C) Purified GST-hScrib wild type and PKA phospho-site mutants of hScrib were in vitro phosphorylated with purified PKA in the presence of radiolabeled ATP as described previously (19) and incubated with PP1c for 20 mins at 30uC. The remaining level of phosphorylated hScrib was then determined following SDS PAGE and autoradiography. The two right-hand lanes show lack of phosphorylation of hScrib in the absence of PKA, whilst the lower panels show the Coomassie stain of the gel demonstrating equal levels of the GST-hScrib fusion protein throughout. The quantitation of hScrib phosphorylation from three independent experiments is also shown. doi:10.1371/journal.pone.0053752.g003 hScrib Regulates ERK Signalling PLOS ONE | www.plosone.org were performed using commercially purified PP1c. The bound protein was then detected by western blotting using anti-PP1c antibodies. The results in Figure 3A demonstrate a number of important features. In the absence of phosphorylation there is a strong interaction between hScrib and the purified PP1c, demonstrating that the interaction between hScrib and PP1c is indeed direct. However, there is also a clear increase in the association between hScrib and PP1c when hScrib is phosphorylated by PKA, but not when it is phosphorylated by ERK1. We had previously shown that the major PKA phosphorylation site on hScrib was S1445 [19]. Therefore, to further confirm that phosphorylation of hScrib by PKA at S1445 can influence its capacity to interact with PP1c, we repeated the pull down assays using the phospho-mimic mutation of hScrib, S1445D. As can be seen from Figure 3B, the S1445D mutant exhibits a significantly increased capacity to interact with PP1c, which is similar to that seen following phosphorylation by PKA. These results demonstrate that phosphorylation of hScrib by PKA at S1445 can indeed increase the ability of hScrib to directly interact with PP1c.
We then analysed whether hScrib was a potential substrate of PP1c. Purified GST-hScrib fusion protein was subjected to in vitro phosphorylation with purified PKA and radiolabeled ATP. After extensive washing the radiolabeled hScrib fusion protein was incubated with purified PP1c, and the amount of phosphorylated protein determined by SDS PAGE and autoradiography. The results obtained in Figure 3C demonstrate that the level of phosphorylated hScrib is decreased following incubation with PP1c, demonstrating that hScrib is a potential substrate of the phosphatase and, furthermore, that hScrib can directly recruit active PP1c. Also shown are the non-phosphorylatable mutants of hScrib, confirming the specificity of the phosphorylation reaction.
We then proceeded to determine whether the interaction of hScrib with PP1c might be involved in the capacity of hScrib to downregulate ERK activation. Cells were transfected with control siRNA against luciferase or against PP1c, and after 24 hours the cells were then transfected with an hScrib expression plasmid. After a further 24 hours the cells were extracted and the levels of activated phospho-ERK analysed by western blotting. The results obtained are shown in Figure 4A. As can be seen, in the absence of hScrib, siRNA PP1c has minimal effect on the levels of phospho-ERK ( Figure 4A upper three panels). In contrast, overexpression of hScrib significantly reduces the levels of phospho-ERK ( Figure 4A lower five panels), and this is in agreement with previous studies [19]. However, the ability of hScrib to downregulate the levels of phospho-ERK is largely abolished following treatment with siRNA PP1c, suggesting that this activity of hScrib is in part PP1c-dependent. To further investigate this, we repeated the assay using the PKA phospho-mimic mutant (S1445D) and the non-PP1c binding mutant (KADA) of hScrib. After 24 hours the levels of phospho-ERK were analysed by western blotting and the results obtained are shown in Figure 4B. As can be seen the wild type and S1445D mutant of hScrib both strongly inhibit the levels of phospho-ERK, whilst the non-PP1c binding mutant of hScrib is decreased in this activity. Taken together these results demonstrate that the ability of hScrib to interact with PP1c correlates with its ability to down-regulate the levels of phospho-ERK.

Loss of hScrib enhances PP1c nuclear localization
Having found that PP1c plays a role in hScrib regulation of ERK signaling, we were next interested in determining whether hScrib could also potentially affect PP1c localisation. Therefore, we first analysed the pattern of PP1c expression in human keratinocytes after stably silencing hScrib expression in these cells. The distribution of PP1c in control and shScrib HaCaT cells were analysed by immunofluorescence. The results in Figure 5A and Figure 5B, show that most of the PP1c localises in the nucleus, although some also co-localises with hScrib at the plasma membrane and within the cytoplasm. More importantly, however, upon loss of hScrib expression there is a significant increase in the amount of nuclear PP1c, with a corresponding decrease in the cytoplasmic pool. In order to verify these results we also performed . PP1c is required for hScrib-induced de-phosphorylation of ERK. A) HEK 293 cells were transfected with PP1c si RNA or si Luc RNA as control (CTL) and after 24 hours were then transfected with a plasmid expressing HA-tagged hScrib. After a further 24 hours the cells were extracted and levels of phospho and total ERK determined by western blot analysis. The upper three panels shows the changes in the ERK profiles when cells were transfected with siRNA PP1c alone, whilst the lower set of five panels show the effects in the presence of ectopically expressed hScrib. B) HEK 293 cells were transfected with HAtagged wild type hScrib, or the S1445D and KADA mutants. Total cell extracts were then made after 48 hours and the hScrib, phospho-ERK and total ERK were detected by western blotting. doi:10.1371/journal.pone.0053752.g004 a series of transient siRNA experiments, where hScrib levels were ablated in 293 cells, and the levels of PP1c, both in total cell extracts or in the respective cellular fractions (Fig. 5C), were analysed by western blotting. As can be seen, loss of hScrib resulted in decreases in the cytoplasmic and membrane pools of PP1c, but a corresponding increase in the amounts of the nuclear form of the protein.
To investigate the pattern of pERK expression following hScrib depletion we repeated the immunofluorescence assays staining for hScrib, PP1c and pERK. The results obtained are shown in Figure 6. As can be seen, under conditions of hScrib depletion there is a marked increase in the levels of both nuclear and cytoplasmic pERK, consistent with previous observations [19]. This is also accompanied by an increase in the levels of nuclear PP1c. Cells were either extracted in SDS PAGE sample buffer (Total lysate) or were fractionated into cytoplasmic (F1), membrane (F2) and nuclear (F3) pools (the example shows the integrity of a typical extraction procedure) and then PP1c was detected by western blotting. p84 was used as a loading control for the nuclear fraction, cadherin was used as a loading control for the membrane fraction and a-tubulin was used as the loading control for the cytoplasmic fraction and total cell extracts. Note the relative increase in nuclear PP1c following hScrib knockdown but no overall change in total PP1c levels. doi:10.1371/journal.pone.0053752.g005 hScrib Regulates ERK Signalling PLOS ONE | www.plosone.org hScrib tumour suppressor activity requires an intact PP1c binding motif We have previously shown that hScrib can suppress cell transformation induced by EJ-ras and Human Papillomavirus (HPV)-16 E7 [19]. To determine whether the interaction between hScrib and PP1c was physiologically relevant in this context, primary BRK cells were transfected with HPV-16 E7 plus EJ-ras in the presence or absence of the hScrib wild type and KLDY/KADA mutant hScrib expressing plasmids, with or without the PP1c expression plasmid. After 3 weeks the cells were fixed and stained and the numbers of colonies counted. As can be seen from Figure 7, co-expression of wild type hScrib and PP1c strongly inhibits the oncogene cooperation between E7 and EJras, whilst the KADA mutant of hScrib is compromised in this activity. These results demonstrate that the hScrib-PP1c interaction is functionally relevant in an assay of oncogene cooperation.

Discussion
We have shown previously that hScrib can regulate ERK signaling in two ways. The first involves a direct protein interaction, which is mediated via two KIM binding sites located within hScrib. The second appears to involve the recruitment of a protein phosphatase [19]. In this study we provide evidence that a candidate phosphatase is PP1c. We have also found that hScrib can control PP1c sub-cellular localisation, with a loss of hScrib promoting PP1c nuclear translocation.
Regulation of the ERK signaling cascade can occur at multiple levels and can involve Raf dephosphorylation, MEK1,2 phosphorylation, and also MEK1,2 dephosphorylation [24,[43][44]. Furthermore, it has been reported that whilst the kinases in the pathway control signal amplitude, the phosphatase PP2A mediates both signal amplitude and signal duration [32][33]. Previous studies have also implicated PP1 in regulating ERK signaling through its ability to dephosphorylate Raf-1 at Ser 259 [45]. Since we have consistently observed that overexpressed hScrib results in a decrease in ERK phosphorylation, we initiated a series of studies to identify the potential phosphatases with which hScrib might interact. Using a proteomic approach we identified PP1c as a direct interacting partner of hScrib, an interaction that we could confirm both in vitro and in vivo. Analysis of the hScrib amino acid sequence identified a potential site of interaction, KLDY, mutation of which abolished the ability of hScrib to bind PP1c. Furthermore, this consensus PP1 recognition motif is conserved in mammalian forms of Scrib, but is absent in Drosophila.
We also analysed the effects of PP1c ablation upon hScrib control of ERK phosphorylation, and found that loss of PP1c greatly diminished the ability of hScrib to downregulate the levels of phospho-ERK in vivo. Furthermore, we also found that this activity of hScrib was in part dependent upon an intact PP1c binding site motif. Interestingly, we also noted that the interaction between PP1c and hScrib was increased following PKA phosphorylation of hScrib, one potential consequence of which is PP1c-mediated de-phosphorylation of hScrib. Whether this has an important role with respect to other functions of hScrib remains to be determined and is worthy of further study. Taken together these studies demonstrate that hScrib can interact with PP1c, an activity which appears to play a role in the ability of hScrib to downregulate the ERK signaling pathway. Interestingly, this regulation of ERK by hScrib has many parallels with a recent study showing that hScrib could also regulate Akt signaling [35]. This required hScrib interaction with the phosphatase, PHLPP1, resulting in the de-phosphorylation of Akt. In this case the interaction between hScrib and PHLPP1 requires sequences in the LRR region of hScrib. Thus hScrib could potentially interact simultaneously with multiple protein phosphatases to control diverse signaling pathways. It should also be emphasized that hScrib is a multifunctional protein, and loss of hScrib also results in increased levels of MEK activity, suggesting multiple mechanisms by which hScrib can control ERK signaling [46].
To investigate whether the capacity of hScrib to interact with PP1c had any physiological relevance, we made use of an oncogene cooperation assay in primary rodent cells. Previous studies had shown that hScrib could suppress cell transformation induced by HPV-16 E7 and EJ-ras in these cells, and that this activity was dependent in part upon the ability of hScrib to interact with ERK [19]. We reasoned that this activity of hScrib might also be influenced by the ability of hScrib to interact with PP1c. Indeed, both hScrib and PP1c, either alone or in combination, could dramatically decrease the levels of HPV-16 E7 and EJ-ras induced cell transformation. However, the additive effects upon the levels of cell transformation, seen with the combination of hScrib and PP1c, was abolished if a mutant hScrib defective in its ability to interact with PP1c was included in the assay. This demonstrates that, in the context of an oncogene cooperation assay, the ability of hScrib to interact with PP1c does play a role in the ability of hScrib to suppress cell transformation.
PP1c has been linked to the regulation of a variety of different cellular processes, including the DNA damage response, nuclear function and diverse aspects of the cell cycle [47][48][49][50][51][52]. One of the important aspects of PP1c regulation is believed to be related to the control of its nuclear expression, which can be mediated by proteins possessing the consensus RVxF PP1 binding motifs, and which can thereby control the correct cellular localization of PP1 [42,52]. We therefore investigated whether hScrib might have a similar potential regulatory function with respect to the pattern of PP1c localization within the cell. This was indeed found to be the case; in two different assay systems we observed that loss of hScrib resulted in an increased nuclear accumulation of PP1c, with a concomitant decrease in the levels found in membrane and cytoplasmic fractions. Thus hScrib would appear to contribute directly to the regulation of PP1c expression patterns. Whether this is related to some of hScrib's previously reported pleiotropic effects upon cell proliferation and cell survival remains to be determined. Taken together, these studies have defined PP1c as a novel interacting partner of hScrib, an interaction which correlates with hScrib downregulation of ERK signaling and suppression of oncogene-induced cell transformation. Figure S1 Schematic diagram showing the different hScrib expression constructs. The schematic shows the arrangement of the functional domains on the hScrib protein, highlighting the LRR, and PDZ domains. The putative PP1binding site, KLDY is also shown in the carboxy terminal third of hScrib. Also summarized are the results on the interaction assays with PP1c. (TIF) Figure 7. hScrib suppresses HPV-16 E7 and EJ-ras induced transformation in cooperation with PP1c in a RVxF motifdependent manner. BRK cells were transfected with EJ-ras alone, HPV-16 E7 plus EJ-ras, HPV-16 E7 plus EJ-ras and wild type hScrib, HPV-16 E7 plus EJ-ras and PP1c, and HPV-16 E7 plus EJ-ras and wild type hScrib with PP1c, and HPV-16 E7 plus EJ-ras and PP1c plus the KADA non-PP1c binding mutant of hScrib. After three weeks the dishes were fixed and stained and the colonies counted. Results represent the mean number of colonies from 3 independent assays and standard deviations are shown. doi:10.1371/journal.pone.0053752.g007