![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Previous Article | Next Article 
Journal of Virology, April 2008, p. 3612-3623, Vol. 82, No. 7
0022-538X/08/$08.00+0 doi:10.1128/JVI.02435-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.
Huijun Wei,,
Tung-Chin Hsieh,
and
David C. Pallas*
Department of Biochemistry and Winship Cancer Institute, Emory University School of Medicine, Atlanta, Georgia 30322
Received 12 November 2007/ Accepted 16 January 2008
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
E4orf4 induction of cell death in transformed cells requires an interaction with the serine/threonine protein phosphatase PP2A (24, 43). PP2A is a highly conserved, ubiquitous eukaryotic protein phosphatase that has broad biological roles in important cellular functions, such as metabolism, initiation of DNA replication, cell cycle control, and apoptosis, and has been implicated in the development of human cancers (20, 21, 37). The core enzyme of PP2A is a heterodimer consisting of a catalytic C subunit and a structural A subunit. This core enzyme can associate with one of a number of distinct B-type subunits to form a wide variety of heterotrimeric holoenzymes (21). In mammalian cells, there are three B-type subunit families (B/B55, B'/B56, and B"/PR72/130) and a putative fourth family (striatin family), each of which has multiple isoforms. In S. cerevisiae, PP2A C subunits are encoded by two genes, PPH21 and PPH22, whereas the PP2A A subunit is encoded by a single gene, TPD3 (54). Moreover, two B-type families are represented: a single PP2A B/B55 subunit encoded by CDC55 (18, 57) and a single B'/B56 subunit encoded by RTS1 (45).
Of the different PP2A heterotrimers, only PP2AB
AC heterotrimers have been shown to be critical for E4orf4-induced growth inhibition and cell killing of transformed cells. E4orf4 interacts with the PP2ABAC holoenzyme complex through direct interaction with the B subunit (24). Expression of PP2A B antisense inhibits the induction of cell death by E4orf4 (43), and overexpression of B increases cell death (44). In addition, E4orf4 mutants with defects in complex formation with PP2A B cannot induce cell death (31). Thus, the PP2A B interaction with E4orf4 is clearly required for E4orf4-induced cell death.
Expression of E4orf4 also inhibits the growth of S. cerevisiae, resulting in irreversible arrest of cells in G2/M and elongated cell morphology (41). Multiple observations suggest that the mechanism of E4orf4-induced growth inhibition in S. cerevisiae is very similar to the mechanism of E4orf4-induced cell death in mammalian cells. First, E4orf4 arrests both mammalian and yeast cells in G2/M (25, 41). Second, E4orf4 interacts with the PP2A B subunit in mammalian cells and the PP2A B subunit, Cdc55p, in S. cerevisiae, and this interaction is required for E4orf4 function in both cell types (24, 41, 44). Third, E4orf4 mutants that cannot bind to the PP2A B subunit and cannot induce cell death in mammalian cells are also unable to induce growth inhibition in yeast and vice versa (2, 41). Fourth, E4orf4 induces accumulation of reactive oxygen species both in mammalian cells and in yeast cells (25, 27). Therefore, analysis of the mechanism of E4orf4-induced growth inhibition in yeast is likely to provide insight into the mechanism of E4orf4-induced cell death in mammalian cells.
The carboxy terminus of the PP2A C subunit is highly conserved from yeast to humans and is modified by both methylation and phosphorylation. In both mammalian and yeast cells, addition and removal of a single methyl group on the carboxy-terminal leucine of the PP2A C subunits differentially regulates the formation of certain PP2A heterotrimers (16, 50, 58-60). In yeast, deletion of the gene for the PP2A methyltransferase Ppm1p almost abolishes (
20-fold reduction) stable formation of PP2A heterotrimers containing the Cdc55p subunit, while PP2A heterotrimers containing Rts1p are reduced less than 2-fold (40%) (16, 58, 59). Mutation of the catalytic subunit carboxy terminus to mimic phosphorylation or to remove the site of methylation has also been shown in yeast to disrupt PP2ABAC heterotrimers and, in some cases, PP2AB'AC heterotrimers, providing useful tools for investigating PP2A function in E4orf4-induced growth inhibition and cell killing (10, 16, 58, 59).
Although it has generally been assumed that E4orf4 functions through the PP2ABAC heterotrimeric complex because of the importance of the PP2A B subunit for E4orf4-induced growth inhibition/cell killing in both mammalian and yeast cells, little is known about the involvement of the A and C subunits. Previous results from yeast using a nonquantitative assay indicated that Tpd3p is required for full E4orf4-induced growth inhibition (25), but the relative importance of Tpd3p and Cdc55p has not been quantitatively compared. Moreover, a requirement for the PP2A C subunit in E4orf4-induced growth inhibition and cell killing has not been demonstrated. Thus, it is possible that the B subunit (Cdc55p) mediates the growth inhibition function of E4orf4 in part through a pathway that is independent of the A subunit (Tpd3p), C subunit (Pph21p/Pph22p), or both.
One possibility for a PP2A C subunit-independent pathway could be that the B subunit (Cdc55p) provides a function to E4orf4 through association with another PP2A family catalytic subunit. In mammalian cells, PP2A A and B subunits have been reported to form trimeric complexes with PP5, the mammalian homolog of the S. cerevisiae phosphatase Ppt1p (28). In S. cerevisiae, the PP2A A subunit (Tpd3p) and B subunit (Cdc55p) have also been proposed to interact with Sit4p, the mammalian homolog of protein phosphatase 6 (PP6) (4), to form a heterotrimeric complex (38). Thus, the possibility exists that the B subunit may mediate E4orf4-induced growth inhibition and cell killing through multiple PP2A family C subunits and perhaps even independently of PP2A C subunits.
In the present study we have quantitatively compared the effects of PP2A A subunit (TPD3) deletion and PP2A B subunit (CDC55) deletion on E4orf4 growth inhibition in S. cerevisiae in order to determine whether E4orf4 mediates some of its effects via Cdc55p independently of Tpd3p. We have also used PP2A C subunit mutants and deletion of the PP2A methyltransferase gene PPM1 to disrupt PP2ACdc55p/Tpd3p/Pph21(22)p heterotrimer formation (and thus the association of E4orf4 with Pph21p/Pph22p) to determine the importance of Pph21p/Pph22p association with E4orf4 for E4orf4-mediated growth inhibition. Moreover, we have assayed whether E4orf4 associates with Sit4p in S. cerevisiae and have analyzed the effects of deletion of genes for various PP2A family phosphatase catalytic subunits, including Sit4p on E4orf4-induced growth inhibition, to determine whether E4orf4 might function through, or genetically interact with, any of these phosphatases. Our observations support a new model in which E4orf4 mediates growth inhibition and cell killing both through PP2ABAC heterotrimers and through a B regulatory subunit-dependent pathway that is independent of the PP2A C subunit. Our analysis also indicates that Ppm1p has a function other than regulating the assembly of PP2A heterotrimers. Finally, our results suggest that selective PP2A trimer inhibitors and PP6 inhibitors may be useful as adjuvant anticancer therapies.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
TPD3 cells were obtained from the laboratory of Richard Hallberg. DEY3 cells were originally created by Evans and Stark (11). These cells are deleted for PPH21, PPH22, and PPH3 and contain a URA3 maintenance plasmid that expresses Pph22p from its own promoter (11). CDC55 TPD3 cells were constructed by crossing ADR496 and TPD3 cells and screening for colonies able to grow on medium lacking histidine or leucine; then, double deletions of the CDC55 and TPD3 genes were confirmed by PCR. PCR-based gene deletion (5) utilizing a set of previously described vectors (46) was used to make additional knockout strains in S. cerevisiae; all primers used are listed in Table 2, and the markers used are indicated in Table 1. The primers in Table 2 used for deletion of the CDC55 gene in DEY3 cells were previously published by Roopchand and colleagues (41). Strains deleted for RTS1 demonstrated temperature sensitivity at 37°C, consistent with previous data (45). The double deletion strain SIT4 CDC55 was created by deletion of CDC55 in the SIT4 strain obtained from Research Genetics. Each deletion strain was confirmed by PCR, at a minimum with a sense primer upstream of the deleted gene and an antisense primer internal to the drug resistance marker inserted. In addition, deletion of CDC55 in the SIT4 CDC55 strain was confirmed by Western blotting for Cdc55p.
|
|
Plasmids.
pRS424 (2µm TRP1 A) (8) and pRS314 (CEN6 TRP1 A) (46) were gifts from the laboratory of Anita Corbett. To provide appropriate selection markers for PCR-based gene disruption of yeast with limited selection markers, two plasmids were used. pFA-KanMX (56) was used for PCR of KAN cassettes, and pAG32 (hphMX4) (17) was used for PCR of hygromycin B (hph) cassettes. pRS314-PPH21 wild-type and T364A, T364D, Y367E, Y367F, and L369 plasmids were subcloned from corresponding pRS316 plasmids that were described previously (58). The Pph21p mutants are referred to by the position mutated preceded by the wild-type amino acid (single letter code), followed by the introduced residue. pYES2 and pYES2-GAL-HA-E4orf4 have been previously published (41). Yeast transformations were performed as previously described (7).
Plasmid shuffle method.
DEY3 cells, which contain a Pph22p-expressing plasmid with a URA3 selectable marker, were transformed with TRP marker-containing pRS314 plasmids expressing wild-type Pph21p or one of the Pph21p mutants, T364A, T364D, Y367E, Y367F, or L369, and then grown on SC-Trp plates. Colonies were picked and streaked on 5-FOA plates to select for cells without the URA3-containing plasmid. Cells grown on 5-FOA plates were then streaked on SC-Ura and SC-Trp plates to confirm the loss of the URA3 plasmid and the presence of the TRP1 plasmid.
Serial dilution spotting assay for measuring E4orf4-induced growth inhibition of S. cerevisiae cells.
Growth inhibition by E4orf4 was quantitatively assessed using serial dilution cell growth assays. Cells were grown at 30°C in 2% glucose medium for 2 days or to saturation. Cells were then diluted with sterile H2O to the concentration of 2 x 106 cells/ml. Four 10-fold serial dilutions were then made with sterile H2O, and 5 µl of each dilution was spotted on agar plates containing 2% glucose or 2% galactose. If not indicated in the figure legends, photographs were taken 2 to 3 days after spotting. In some cases, cells were grown on 2% raffinose medium for 2 h before dilution and spotting.
Antibodies.
Anti-hemagglutinin (HA) tag antibodies 16B12 and F-7 were obtained from Berkeley Antibody Company and Santa Cruz Biotechnology, Inc. PP6 antibody directed against a C-terminal peptide was obtained from Stratagene. Sit4p antibody was a kind gift from Yu Jiang (38). An affinity-purified rabbit polyclonal antibody specific for Pph21p and Pph22p was used to visualize Pph21p proteins in cell lysates (16). Ppm1p rabbit polyclonal antiserum (RK3332) was raised to two Ppm1p peptides (SKEDTAKSPFLIDQGRYKC and QSNLKESRNLEMPTLMTYNSKEKYASRC), corresponding to Ppm1p residues 153 to 170 and 248 to 274, respectively, with an additional carboxy-terminal cysteine for coupling to keyhole limpet hemocyanin using the Pierce Imject kit. The specificity of this antibody was evaluated using affinity purification and wild-type and PPM1 cells (not shown). Cdc55p antibody, a gift from the laboratory of Egon Ogris, has been previously described (58).
Preparation of cellular extracts, immunoprecipitation, and Western blotting. Yeast cells were harvested and lysed in lysis buffer (100 mM Tris-Cl pH 7.5, 200 mM NaCl, 1 mM EDTA, 5% glycerol, 0.5 mM dithiothreitol) containing 1 mM phenylmethylsulfonyl fluoride and yeast cocktail inhibitor (Sigma) as previously described using glass beads and vortexing (49, 58) or a mini-beadbeater (BIOSPEC). The lysate protein concentration was determined using Bio-Rad DC protein assay reagent, and then a one-half volume of radioimmunoprecipitation assay buffer (50 mM Tris-Cl, pH 7.5, 200 mM NaCl, 1% Triton X-100, 0.5% Na-deoxycholate, 0.1% sodium dodecyl sulfate [SDS]) was added. Immunoprecipitation of E4orf4-PP2A complexes via HA-tagged E4orf4 was then performed using anti-HA tag monoclonal antibody (16B12 or F-7) plus protein A-Sepharose beads and incubation at 4°C for 1 h 30 min. After immunoprecipitation, the samples were boiled in gel loading buffer and analyzed by 10% or 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE). Then, proteins were transferred to nitrocellulose membranes and immunoblotted. Bands were visualized using horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence. Bands were quantitated using a Fluor S-Max chemilumimager (Bio-Rad, Richmond, CA) that directly measures band intensities without the use of film via a supercooled charge-coupled-device camera that provides linear data over 4.8 orders of magnitude. This method yielded highly reproducible results that did not vary with image capture times.
Cell culture and creation of stable lines that express vector control or PP6-directed shRNAs. Human 293T cells for generating lentiviral stocks and H1299 human lung cancer cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum in 10% CO2 at 37°C. To create H1299 lines that stably express shRNA vector (vector control) or PP6-directed shRNA, cells were infected with shRNA lentiviruses. These lentiviruses were generated using a three-plasmid-based lentivirus system (34) (available from The RNAi Consortium [TRC] at the Broad Institute). The targeted sequence in the PP6 shRNA lentivirus we used was 5'-CCAGAACGACAACGCCATATT-3' (TRC ID TRCN0000002765). After infection with the appropriate lentivirus, infected cells were selected with puromycin (2 µg/ml) until a control dish of uninfected H1299 cells was completely dead. After incubating at least one additional day to allow the cells to recover from puromycin treatment, immunoblotting was used to confirm the knockdown of PP6. Cells were then used for experiments, and aliquots of cells were frozen for later use. Although PP6 knockdown was quite stable (data not shown), cells were used within 10 days after selection and thawing.
Adenovirus infection and WST-1 cell viability assay. A 96-well plate was seeded with 100 µl of H1299 cells per well at a concentration of 2.5 x 104 cells/ml. The next day, cells were infected with control adenovirus (CMVrtTA) at a multiplicity of infection (MOI) of 10 or with E4orf4-expressing adenovirus (CMVrtHA-E4orf4) and CMVrtTA adenovirus, each at an MOI of 5, and then E4orf4 expression was induced by the addition of 1 µg/ml doxycycline. Four days after infection, cell viability was assayed in a water-soluble tetrazolium salt (WST-1) killing assay (Roche).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To test this possibility, serial dilution cell growth assays were carried out to compare the effects of deletions of TPD3 and CDC55 on E4orf4-induced growth inhibition of W303 S. cerevisiae cells. Wild-type, TPD3, and CDC55 cells were transformed with a galactose-inducible E4orf4 plasmid, and cell growth was compared on glucose and galactose media. Figure 1 shows that when E4orf4 was induced on galactose plates, the growth of the wild-type cells was strongly inhibited. Deletion of CDC55 could partially rescue E4orf4-induced slow growth, confirming the previous finding that E4orf4 functions through Cdc55p-dependent and Cdc55p-independent pathways (41). Deletion of TPD3 partially rescued E4orf4-induced slow growth to a similar extent as deletion of CDC55, suggesting that Cdc55p does not provide a function to E4orf4-induced growth inhibition that is independent of the A subunit.
|
TPD3 cells grow more slowly under normal growth conditions than CDC55 cells (54), we used a second approach to compare the importance of Tpd3p and Cdc55p for E4orf4-induced growth inhibition. We reasoned that if Cdc55p provided a Tpd3p-independent function to E4orf4, then double deletion of CDC55 and TPD3 should rescue E4orf4-mediated growth inhibition more effectively than deletion of TPD3 alone. CDC5 TPD3 cells and TPD3 cells grow at a similar rate and thus should provide a fair comparison that is not affected by the general health of the cells. Therefore, we compared the growth rates in liquid medium of wild-type, TPD3, and CDC55 TPD3 cells expressing E4orf4 or containing an empty vector (control). The average doubling time for cells with or without E4orf4 was calculated from three independent experiments, and the ratios of doubling times of cells with E4orf4 to cells without E4orf4 were compared. As indicated in Table 3, E4orf4 caused an 58% increase in the doubling time of wild-type cells. Consistent with the results from serial dilution cell growth assays, indicating that deletion of TPD3 can partially rescue E4orf4-induced slow growth (Fig. 1), E4orf4 caused only an 11% increase in the doubling time of tpd3 cells. Double deletion of TPD3 and CDC55 did not further diminish the growth inhibition effect of E4orf4 compared to deletion of TPD3 alone (Table 3), indicating that Cdc55p does not mediate an E4orf4-induced growth inhibition function in these cells through a pathway independent of Tpd3p.
|
PPH21 PPH22 cells in a W303 background grow very poorly on glucose-containing medium and are inviable on galactose-containing medium, even in the absence of E4orf4 expression (reference 40 and unpublished data). To circumvent this problem, we used alternative approaches to disrupt Cdc55p (and thus E4orf4) association with the PP2A catalytic subunit that did not involve deletion of the genes for both catalytic subunits. As a first approach, we tested whether Pph21p mutants defective in PP2ACdc55p/Tpd3p/Pph21(22)p complex formation could rescue E4orf4-induced growth inhibition. We reasoned that if E4orf4 mediated its Cdc55p-dependent growth inhibitory effects solely via this trimeric complex, then disruption of Cdc55p (and thus E4orf4) binding to the PP2A catalytic subunit with these mutants would rescue E4orf4-mediated growth inhibition to a similar extent as deletion of CDC55.
We previously demonstrated that certain mutations in the C terminus of Pph21p abolish yeast PP2ABAC heterotrimeric complex formation (15, 16, 58). Specifically, Pph21p mutants with acidic substitutions of threonine 364 or tyrosine 367, termed T364D and Y367E, do not associate stably with Cdc55p, while mutants with conserved neutral substitutions at these same positions, T364A and Y367F, associate with Cdc55p at near-wild-type levels. In addition, a Pph21p mutant lacking the carboxy-terminal leucine, L369, did not form PP2A heterotrimers containing Cdc55p (15, 16). In this study, we directly tested for the ability of these mutants to associate with E4orf4 by examining whether they could be coimmunoprecipitated with epitope-tagged E4orf4. The results shown in Fig. 2 indicate that while wild-type Pph21p and the neutral substitution mutant controls T364A and Y367F were easily detected in E4orf4 immunoprecipitates, T364D, Y367E, and L369 did not specifically associate with E4orf4.
|
to the same level as wild-type cells, indicating that the formation of stable PP2ACdc55p heterotrimers may not be required for Cdc55p-mediated E4orf4-induced growth inhibition in yeast. Moreover, although in repeated experiments cells expressing T364D were more resistant to E4orf4-induced slow growth than wild-type cells (see Fig. 4, below, for another example), they were not nearly as resistant to E4orf4-induced growth inhibition as cells deleted for CDC55 (Fig. 3A). Thus, three separate PP2A catalytic subunit mutants that prevent stable association with E4orf4 are unable to provide the same rescue from E4orf4-induced growth inhibition that deletion of CDC55 provides.
|
|
CDC55 background compared to cells expressing wild-type Pph21p (i.e., expression of these mutants should also prevent deletion of CDC55 from providing full rescue). We therefore compared E4orf4-induced growth inhibition of CDC55 cells expressing wild-type, T364D, Y367E, or L369 Pph21p. The results in Fig. 3B show that deletion of CDC55 rescued cells expressing wild-type or T364D Pph21p to the same extent, indicating that T364D has no Cdc55p-indpendent effects that enhance E4orf4-induced growth inhibition. On the other hand, cells expressing Y367E or L369 Pph21p consistently showed a slightly reduced rescue upon deletion of CDC55 (Fig. 3B and data not shown), indicating that these mutants have a Cdc55p-independent effect that enhances E4orf4-induced growth inhibition. Together, the results in Fig. 3A and B indicate that disruption of stable formation of PP2ABAC heterotrimers by PP2A catalytic subunit mutation partially rescues E4orf4-induced growth inhibition but, in the case of Y367E and L369, this effect is not seen because they have an additional, unknown effect that enhances E4orf4-induced growth inhibition. Nevertheless, in all cases deletion of CDC55 substantially rescued E4orf4-induced growth inhibition of DEY3 cells expressing these mutants (Fig. 3B), strongly supporting the existence of a Cdc55p-dependent, PP2AB(Cdc55p)AC heterotrimer-independent effect(s) important for E4orf4-induced growth inhibition.
Downregulation of Rts1p function enhances E4orf4-induced cytotoxicity.
What might the Cdc55p-independent effect of the Y367E and L369 Pph21p mutants (but not T364D) be that enhances E4orf4-induced growth inhibition even in the absence of CDC55 (Fig. 3)? Previous analyses showed that T364D forms complexes with Rts1p (60% of wild-type level) but is unable to associate with Cdc55p, while Y367E and L369 are unable to form complexes with either Rts1p or Cdc55p (16, 58). We therefore hypothesized that greatly reduced complex formation of Pph21p with Rts1p (and thus reduced Rts1p function) in DEY3 cells expressing Y367E and L369 enhanced E4orf4-induced growth inhibition. To test this hypothesis we deleted RTS1 in DEY3 cells and compared the E4orf4-induced growth inhibition of these cells with that of the parental DEY3 cells to see if loss of Rts1p indeed made cells more sensitive to E4orf4-induced growth inhibition. Both cells expressing wild-type Pph21p and cells expressing the T364D mutant were analyzed. Consistent with our results shown in Fig. 3, wild-type DEY3 cells expressing the T364D mutant showed a partial rescue from E4orf4-induced growth inhibition relative to wild-type DEY3 cells expressing wild-type Pph21p, while deletion of CDC55 gave full rescue (Fig. 4A). However, RTS1 deletion made both wild-type and T364D Pph21p-expressing cells more sensitive to E4orf4-induced growth inhibition (Fig. 4A), effectively masking any rescue by T364D. This result indicates that loss of Rts1p function does indeed enhance E4orf4-induced cytotoxicity. Moreover, we found in separate experiments that in the absence of Rts1p, cells expressing T364D are inhibited to a similar extent by E4orf4 as cells expressing the two mutants lacking Rts1p binding, Y367E and L369 (Fig. 4B). Together, these results support our hypothesis that differences in Rts1p association may explain why the T364D Pph21p mutant, but not the Y367E and L369 mutants, provides a partial rescue from E4orf4-induced growth inhibition in wild-type DEY3 cells.
Disruption of PP2ABAC complexes by deletion of the PP2A methyltransferase gene, PPM1, does not rescue E4orf4-induced growth inhibition.
Our results with PP2A catalytic subunit mutants unable to form stable PP2ABAC heterotrimers suggest that there is a Cdc55p-dependent function(s) in E4orf4-induced growth inhibition that does not require stable Cdc55p association with Pph21p/Pph22p. To test this idea by an independent approach, we deleted the PP2A methyltransferase gene, PPM1, to disrupt E4orf4 association with PP2A catalytic subunits. Loss of PP2A methylation due to deletion of PPM1 nearly abolishes the stable formation of PP2ACdc55p heterotrimers when assayed in a coimmunoprecipitation assay (58, 59) or an in vivo association assay (16). To determine whether Pph21p can specifically associate with E4orf4 in PPM1 cells, HA-tagged E4orf4 was expressed in wild-type and PPM1 cells, the HA-tagged E4orf4 was immunoprecipitated, and the immunoprecipitates were analyzed for the presence of Pph21p. While Pph21p could be readily detected in the E4orf4 immunoprecipitate in wild-type cells, it was greatly reduced in the E4orf4 immunoprecipitate from PPM1 cells (Fig. 5A). Quantitation indicated that the levels of immunoprecipitated Pph21p from PPM1 cells are only about 5% of the amount from wild-type cells.
|
PPM1 cells with a plasmid expressing galactose-inducible E4orf4 and tested the effect of the deletion of PPM1 on E4orf4-induced growth inhibition. When E4orf4 expression was induced, it inhibited growth of wild-type and PPM1 cells to the same extent (Fig. 5B), further supporting the idea that substantial interaction of E4orf4 with the PP2A catalytic subunits, Pph21p/Pph22p, is not required for E4orf4 to inhibit cell growth in S. cerevisiae.
Cells deleted for PPM1 cannot be rescued from E4orf4-induced growth inhibition by deletion of CDC55.
We were somewhat surprised that deletion of PPM1 did not provide any rescue from E4orf4-induced growth inhibition in our experiments described above (Fig. 5B), given that it dramatically reduces the formation of PP2ACdc55p heterotrimers. One possible explanation was that the very small amount of PP2ACdc55p heterotrimers present in the PPM1 cells is sufficient to mediate E4orf4-induced growth inhibition. If this scenario were true, then additional deletion of CDC55 in the PPM1 cells should rescue these cells from E4orf4-induced growth inhibition. To test this hypothesis, we constructed a PPM1 CDC55 double mutant and compared the E4orf4-induced growth inhibition of these cells with wild-type and PPM1 cells (Fig. 5C). Consistent with our results in Fig. 5B, E4orf4 efficiently inhibited the growth of both wild-type and PPM1 cells. Surprisingly, however, the PPM1 CDC55 cells were also inhibited to a similar extent as wild-type and PPM1 cells, indicating that cells that do not express Ppm1p cannot be rescued by deletion of the gene encoding Cdc55p. Thus, Ppm1p must have another function in addition to regulating PP2ACdc55p heterotrimer formation, and this function must intersect with E4orf4 signaling.
Loss of Sit4p enhances E4orf4-induced growth inhibition. One possibility for an additional function of Ppm1p that may intersect with E4orf4 signaling is that Ppm1p might regulate another phosphatase in addition to PP2A. There are two mammalian phosphatases, PP4 and PP6, that share approximately 60% amino acid identity with PP2A and have identical carboxy-terminal amino acids, including a carboxy-terminal leucine, the known site of carboxyl methylation for both PP2A and PP4 (19). PP6 has a functional homolog in S. cerevisiae termed Sit4p (4) that shares these similarities with Pph21p/Pph22p. While it is not yet known whether PP6 and Sit4p are indeed methylated, it seems likely that they would be based on the critical similarities between their sequences and those of PP2A and PP4 C subunits. Thus, Sit4p is a potential substrate of Ppm1p. We therefore tested whether deletion of SIT4 would mimic PPM1 deletion with regard to E4orf4 growth inhibition.
To test this possibility, we exploited the existence of SIT4 cells in the S. cerevisiae knockout collection (BY4741 strain background) and compared wild-type, CDC55, SIT4, and SIT4 CDC55 cells for growth inhibition by E4orf4. The results of a typical experiment are shown in Fig. 6. Comparison of the growth inhibition of wild-type and SIT4 cells showed that loss of Sit4p greatly sensitizes cells to E4orf4-induced growth inhibition (Fig. 6). Moreover, while deletion of CDC55 partially rescued wild-type cells in this strain background, CDC55 deletion was unable to rescue the E4orf4-induced growth inhibition of the SIT4 cells (Fig. 6). Thus, deletion of SIT4 has a more severe effect than deletion of PPM1, but it is similar to deletion of PPM1 in that deletion of CDC55 shows no rescue of SIT4 or PPM1 cells in a serial dilution growth assay (Fig. 5C and 6).
|
Knockdown of PP6, the mammalian Sit4p homolog, by shRNA greatly enhances E4orf4 killing of H1299 human non-small cell lung carcinoma cells. Regardless of whether Sit4p is a direct target of Ppm1p, our results show that deletion of Sit4p greatly enhances E4orf4-induced growth inhibition in S. cerevisiae. This result in yeast suggested that inhibition of PP6 in mammalian cells might have synergistic effects with E4orf4 for killing cancer cells. We therefore used lentivirus-mediated shRNAs to knock down mammalian PP6 in H1299 human lung cancer cells to determine if this would affect E4orf4-induced cell killing. As seen in Fig. 7, endogenous PP6 was almost undetectable in the PP6 shRNA H1299 cells (2.5% of control levels when quantitated). Interestingly, no increase in cell death was detected over vector control (pLKO) cells, even with this level of knockdown (Fig. 7; compare the first two columns), suggesting that PP6 may not be an essential protein in these cells. However, when wild-type E4orf4 was expressed in the PP6 knockdown and vector control cells via an adenovirus vector, PP6 knockdown clearly enhanced E4orf4 killing (Fig. 7; compare the third and fourth columns). When a mutant (R69-75A) E4orf4 defective for binding PP2A and for killing (31) was expressed in these cells, little to no cell killing was detected in the presence or absence of PP6 knockdown (Fig. 7B, fifth and sixth columns, and data not shown). Taken together, these results indicate that PP6 is important for cell viability in the presence of E4orf4 protein and that inhibitors of PP6 may be useful in E4orf4 cancer therapy or drug therapy based on the E4orf4 cancer cell toxicity mechanism.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The clearest support for a PP2ABAC heterotrimer-independent, B subunit-dependent pathway in E4orf4-induced growth inhibition comes from our results with DEY3 cells expressing the PP2A catalytic subunit mutant T364D. Expression of this mutant in place of the wild-type PP2A catalytic subunit consistently gave only a partial rescue from E4orf4 growth inhibition, while deletion of the gene encoding Cdc55p gave full rescue in the same cells. The partial rescue afforded by T364D is not likely to be due to a small amount of residual heterotrimer formation with Cdc55p (and Tpd3p) because of the following: (i) no heterotrimer formation with Cdc55p was detected for this mutant in either an immunoprecipitation or an in vivo association assay (16, 58); (ii) cells expressing this mutant in place of the wild-type PP2A catalytic subunit are defective in the spindle checkpoint, a known Cdc55p function (58); (iii) we could not detect a complex between T364D and E4orf4 in this study (Fig. 2). In addition, the fact that deletion of CDC55 could fully rescue DEY3 cells expressing T364D from E4orf4-induced growth inhibition argues against the possibility that the T364D mutation has a Cdc55p-independent growth inhibitory effect that reduces its ability to rescue cells from E4orf4-induced growth inhibition. Further support for a C subunit-independent, B/A subunit-dependent pathway in E4orf4-induced growth inhibition comes from the finding that deletion of CDC55 provided substantial rescue to DEY3 cells expressing Y367E or L369 mutant PP2A catalytic subunits (Fig. 3 and 4 and data not shown). The inability of these mutants to form complexes with Cdc55p and E4orf4 is also well documented (16, 58) (Fig. 2). Thus, our data show directly for the first time that the PP2A catalytic subunit is indeed a target for E4orf4, but that E4orf4 mediates substantial growth inhibition in a Cdc55p/Tpd3p-dependent manner that does not require stable complex formation between E4orf4 and the catalytic subunit of PP2A.
Our results also indicate that Rts1p, the S. cerevisiae PP2A B' subunit, has some function that opposes E4orf4-induced growth inhibition. The correlation between the ability of PP2A catalytic subunit mutants to partially rescue E4orf4-induced growth inhibition and their ability to bind Rts1p indicates that at least part of Rts1p's opposing function is carried out via PP2ARts1p heterotrimers. Shtrichman and colleagues previously reported that mammalian B' subunits associate with E4orf4 (44). It is tempting to speculate that E4orf4 might bind Rts1p to block its inhibitory function toward E4orf4 growth arrest. However, the association of E4orf4 with mammalian B' subunits has been challenged, and data from a study with S. cerevisiae suggest that E4orf4 may not bind PP2ARts1p heterotrimers (41). Whether or not it is a direct target of E4orf4, Rts1p may oppose E4orf4-induced growth inhibition at G2/M. Rts1p associates with the catalytic subunit of the cell cycle regulatory kinase Cdc28p and with Clb2p, a B-type cyclin that regulates Cdc28p in G2/M (53). E4orf4 has been reported to associate with the anaphase-promoting complex (APC) (25), which regulates progression through mitosis by modulating the stability of several proteins, including Pds1p (yeast securin) and Clb2p (reference 55 and references therein). Like deletion of RTS1, mutation of CDC28 has been reported to make yeast cells hypersensitive to E4orf4 (25), leading us to hypothesize that Rts1p might regulate Cdc28/Clb2p to oppose E4orf4. Regardless of the mechanism, our results indicate that Cdc55p and Rts1p have opposite effects on E4orf4-induced growth inhibition. Furthermore, they are consistent with reports that B' subunits have a prosurvival role in Drosophila melanogaster (26, 47) and mammalian cells (48).
An important implication of our results is that the methyltransferase Ppm1p has a function independent of regulating PP2ACdc55p and PP2ARts1p heterotrimer formation. Especially striking is the observation that in the absence of Ppm1p, E4orf4-induced growth inhibition became Cdc55p-independent, as assayed by serial dilution cell growth assays. We considered that the inability of CDC55 deletion to rescue PPM1 cells from E4orf4-induced growth inhibition might be partially explained by the small (<2-fold) reduction in PP2ARts1p heterotrimers caused by PPM1 deletion (16). However, we have previously shown that expression of T364D Pph21p in place of wild-type Pph21p also causes the same reduction in PP2ARts1p heterotrimers as PPM1 deletion (16), yet the growth inhibition of T364D-expressing DEY3 cells by E4orf4 can be fully rescued by deletion of CDC55 (Fig. 4). Thus, it appears that Ppm1p can function independently of Cdc55p and Rts1p. It is possible that Ppm1p may have another substrate(s) in addition to the PP2A catalytic subunits and that this substrate regulates a target relevant to E4orf4-induced growth inhibition. One candidate for such an additional substrate is Sit4p, the yeast homolog of the catalytic subunit of the mammalian phosphatase PP6, which shares 60% identity with PP2A, including a carboxy-terminal leucine, the site of PP2A (and PP4) methylation. Consistent with the possibility that Sit4p might be a target of Ppm1p, the enhanced E4orf4-induced growth inhibition caused by deletion of the gene encoding Sit4p could not be rescued by deletion of CDC55. Although deletion of SIT4 enhanced E4orf4-induced growth inhibition while deletion of PPM1 did not, this difference may be due partly to the fact that Ppm1p also regulates PP2ACdc55p heterotrimer formation, which would be expected to reduce E4orf4-induced growth inhibition. Consistent with this idea, even though PPM1 cells showed no rescue in the serial dilution growth assay upon CDC55 deletion, deletion of PPM1 does result in an ability to tolerate higher E4orf4 expression (Y. Li and D. C. Pallas, unpublished results), a phenotype previously described for CDC55 deletion cells (41). Based on a previous report suggesting that Cdc55p may form a heterotrimeric complex with Tpd3p and Sit4p, it was also possible that E4orf4 might target Sit4p directly through Cdc55p. However, we have been unable to detect a physical association between Sit4p and E4orf4 (Li and Pallas, unpublished).
Given the recent report that E4orf4 hijacks Rho GTPase-dependent actin dynamics to kill cancer cells (39), it is of interest that Sit4p has been clearly implicated in the regulation of actin dynamics in yeast (3, 23, 52). SIT4 deletion increases the time necessary for repolarization of the actin cytoskeleton after cellular stress (3), and Sit4p interacts with Bem2p, a protein with GTPase-activating protein (GAP) homology that regulates actin dynamics and cellular morphogenesis. Moreover, SIT4 deletion is synthetically lethal with BEM2 deletion, and both SIT4 deletion and BEM2 mutant phenotypes can be suppressed by a common suppressor, SSD1-vl (13, 14, 23, 51). Consistent with our observation that CDC55 deletion and SIT4 deletion have opposite effects on E4orf4-induced growth inhibition, genetic data from S. cerevisiae indicate that Sit4p and Bem2p function in opposition to Cdc55p (18, 23). However, although the links to actin are intriguing, other possible mechanisms of enhancement of E4orf4-induced growth inhibition by SIT4 deletion include Sit4p effects on regulation of G1/S progression or mitosis or a combination of these pathways.
Regardless of the mechanism of enhancement of E4orf4-induced growth inhibition by deletion of SIT4, our results with shRNA-mediated downregulation of the mammalian Sit4p homolog, PP6, in H1299 human lung cancer cells indicate that downregulation of this phosphatase has similar effects in mammalian cells—enhancing E4orf4 cancer cell toxicity. It was especially striking that PP6 downregulation had no noticeable effects on cell viability in the absence of E4orf4 expression. PP6 downregulation failed to enhance the killing of H1299 cells when an E4orf4 mutant defective in binding to PP2ABAC heterotrimers was used, indicating that PP6 knockdown synergizes with the PP2A-dependent E4orf4 cell killing pathway. After more details of this pathway are elucidated for the benefit of therapeutic mimicry of the E4orf4 mechanism, our results suggest that PP6 inhibition should be considered as part of that cancer cell-killing strategy. It is also possible that PP6 downregulation may have the potential for enhancing other cancer therapeutics. Consistent with this possibility, PP6 was recently identified in an siRNA screen of human kinases and phosphatases as having a cell survival role in HeLa cells (29).
|
ACKNOWLEDGMENTS |
|---|
Under agreements between Upstate Biotechnology Inc. (Millipore), Santa Cruz Biotechnologies, Stratagene, Inc., Invitrogen Corp., Cell Signaling Technology, Inc., and Emory University, David Pallas is entitled to a share of sales royalty received by the University from these companies. In addition, this same author serves as a consultant to Millipore. The terms of this arrangement have been reviewed and approved by Emory University in accordance with its conflict of interest policies.
This work was supported by a grant to D.C.P. from the National Institutes of Health (CA57327).
|
FOOTNOTES |
|---|
Published ahead of print on 23 January 2008.
Y.L. and H.W. contributed equally to this work.
Present address: Schering-Plough Research Institute, Kenilworth, NJ 07033.
Present address: Department of Urology, George Washington University, 2150 Pennsylvania Ave., NW, Washington, DC 20037.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
