The myeloid leukemia-associated protein SET is a potent inhibitor of protein phosphatase 2A.

Two potent heat-stable protein phosphatase 2A (PP2A) inhibitor proteins designated I1PP2A and I2PP2A have been purified to apparent homogeneity from extracts of bovine kidney (Li, M., Guo, H., and Damuni, Z. (1995) Biochemistry 34, 1988-1996). N-terminal and internal amino acid sequencing indicated that I2PP2A was a truncated form of SET, a largely nuclear protein that is fused to nucleoporin Nup214 in acute non-lymphocytic myeloid leukemia. Experiments using purified preparations of recombinant human SET confirmed that this protein inhibited PP2A. Half-maximal inhibition of the phosphatase occurred at about 2 nM SET. By contrast, SET (up to 20 nM) did not affect the activities of purified preparations of protein phosphatases 1, 2B, and 2C. The results indicate that SET is a potent and specific inhibitor of PP2A and suggest that impaired regulation of PP2A may contribute to acute myeloid leukemogenesis.

Two potent heat-stable protein phosphatase 2A (PP2A) inhibitor proteins designated I 1 PP2A and I 2

PP2A
have been purified to apparent homogeneity from extracts of bovine kidney (Li, M., Guo, H., and Damuni, Z. (1995) Biochemistry 34, 1988Biochemistry 34, -1996. N-terminal and internal amino acid sequencing indicated that I 2 PP2A was a truncated form of SET, a largely nuclear protein that is fused to nucleoporin Nup214 in acute non-lymphocytic myeloid leukemia. Experiments using purified preparations of recombinant human SET confirmed that this protein inhibited PP2A. Half-maximal inhibition of the phosphatase occurred at about 2 nM SET. By contrast, SET (up to 20 nM) did not affect the activities of purified preparations of protein phosphatases 1, 2B, and 2C. The results indicate that SET is a potent and specific inhibitor of PP2A and suggest that impaired regulation of PP2A may contribute to acute myeloid leukemogenesis.
A number of defined chromosomal translocations occur in specific subtypes of myeloid leukemia indicating that these translocations play an important role in the process of leukemogenesis (1,2). As a result of translocation, nearby oncogenes and other genes involved in the control of proliferation or differentiation can be activated through alterations in regulatory DNA sequences that leave the encoded protein intact (e.g. Myc) or through formation of fusion genes that encode chimeric proteins (e.g. Bcr-Abl, E2A-Pbx, and Pml-RAP␣) (1,2). However, although several of the chromosomal translocations and the resulting fusion genes that occur in leukemia have been identified, often the function of the individual proteins encoded by the fusion transcripts has not been determined (1,2).
Protein phosphatase 2A (PP2A) 1 is a major mammalian protein serine threonine phosphatase that regulates diverse cellu-lar processes (3)(4)(5)(6). In cells, PP2A is thought to exist as heterotrimeric forms termed PP2A 1 and PP2A 0 and composed of a catalytic C subunit and A and B (B, BЈ, or BЉ) subunits (3)(4)(5)(6). A dimeric form of PP2A, termed PP2A 2 , has also been isolated from numerous sources and is composed of the A and C subunits (3)(4)(5)(6). However, this enzyme is thought not to exist in vivo because the missing B subunit may have been lost during the isolation procedures (3)(4)(5)(6). Two forms of the A and C subunits exhibiting apparent M r values of ϳ65,000 and 36,000 and 86% (7,8) and 97% identity (9 -13) in their predicted amino acid sequences, respectively, have been identified by molecular cloning methods. In addition, several distinct B subunits of apparent M r ϳ54,000, 55,000, 56,000, 72,000, and 130,000 have also been identified (e.g. see . Although the significance of the different A, B, and C subunits is not well understood, the distinct substrate specificities of the various trimeric PP2A forms appear to be conferred, at least in part, by the variable B subunit (3)(4)(5)(6).
However, despite progress on the structure of PP2A, little information is available on the regulation of this enzyme. Nonetheless, evidence has emerged that PP2A is inactivated by phosphorylation (18,19) and activated by methylation (20) of its C subunit. In addition, we recently purified to apparent homogeneity from extracts of bovine kidney two heat-stable and PP2A-specific inhibitor proteins designated I 1 PP2A and I 2 PP2A (21). These inhibitor proteins act noncompetitively and exhibit apparent K i values in the nanomolar range (21). The inhibitor proteins appear to be PP2A-specific because, in contrast to PP2A, they do not affect the activities of PP1 C , PP2B, and PP2C (21), the other major mammalian protein serine threonine phosphatases (3)(4)(5)(6). Furthermore, I 1 PP2A and I 2

PP2A
did not affect the activities of 11 different protein kinases (21). Because the purified preparations exhibited distinct peptide patterns following cleavage with Staphylococcus aureus V8 protease, I 1 PP2A and I 2 PP2A may be the products of distinct genes (21). However, direct evidence for this possibility was not provided.
This study was undertaken to determine the identity of I 2 PP2A on a firm basis. In this communication, we show that the purified bovine kidney preparations of I 2 PP2A represent a truncated form of SET (22), a largely nuclear protein also termed PHAP-II (putative class II human histocompatibility leukocyte-associated protein II) (23) and TAF (template-activating factor) (24). In acute non-lymphocytic myeloid leukemia, the SET gene, which is located on chromosome 9q34 centromeric to c-abl and nup214, is fused to nup214 (also termed can) apparently as a result of translocation (22). This SET-Nup214 fusion gene encodes a 5-kilobase transcript that contains a single open reading frame predicting a chimeric SET-Nup214 protein of apparent M r ϳ150,000 (2,22). The results suggest that fusion of SET with Nup214 in acute myeloid leukemia may impair the normal regulation of PP2A and contribute to leukemogenesis.
Generation, Purification, and Sequencing of Tryptic Peptides-Aliquots (2 g) of purified bovine kidney I 2 PP2A were subjected to SDS-PAGE, followed by electrophoretic transfer onto Immobilon TM P membranes. After staining with Ponceau S, bands corresponding to I 2

PP2A
were cut out and subjected to overnight incubation at 37°C with Ntosyl-L-phenylalanine chloromethyl ketone-treated trypsin (0.5 mg) as described (34). Tryptic peptides were then resolved by high pressure liquid chromatography reversed-phase chromatography on an Aquapore RP-300 column (1 ϫ 250 mm) equilibrated with 0.1% (v/v) trifluoroacetic acid. After washing with this solution, the column was developed, at a flow rate of 0.15 ml/min, with a linear gradient from 0.1% trifluoroacetic acid to 0.08% trifluoroacetic acid containing 38.5% (v/v) acetonitrile in 30 min, followed by a linear gradient from 0.08% trifluoroacetic acid containing 38.5% acetonitrile (v/v) to 0.08% trifluoroacetic acid containing 59.5% (v/v) acetonitrile in 10 min. Fractions (0.045 ml) were collected, and the absorbance at 214 nm was determined. The amino acid sequences of peptides eluting at about 38% (peptide I) and 48% (peptide II) acetonitrile and the N terminus of I 2 PP2A were determined using standard reagents for gas phase chemistry on an automated protein sequencer (Applied Biosystems) equipped with an on-line UV detector to identify phenylthiohydantoin derivatives.
Generation of SET cDNA-First strand SET cDNA was generated at 42°C for 30 min in 0.02 ml of 10 mM Tris-HCl, pH 8.3, containing 50 mM KCl, 5 mM MgCl 2 , 20 units of RNase inhibitor, 1 mM of each dNTP, 0.5 g of human kidney polyadenylated RNA (Clontech), 50 pmol of the 3Ј SET-based oligonucleotide (GGATCCTAGTCATCTTCTCCTTC), and 2.5 units of murine leukemia virus reverse transcriptase (Perkin-Elmer). After heat denaturation at 99°C for 5 min, a 0.005-ml aliquot of the mixture was subjected to amplification by polymerase chain reaction in 0.1 ml of 20 mM Tris-HCl, pH 8.3, containing 10 mM KCl, 10 mM (NH 4 ) 2 SO 4 , 0.1% Triton X-100, 2 mM MgCl 2 , 0.1 mg/ml bovine serum albumin, 100 pmol of the 3Ј and 5Ј (CCGCGGAGCAGCCATAT-GTCGGC) SET-based oligonucleotides, 0.2 mM of each dNTP, and 5 units of recombinant pfu DNA polymerase (Stratagene). After heat denaturation for 5 min at 94°C, polymerase chain reaction amplification for 30 cycles was performed as follows: 30 s at 94°C, 30 s at 55°C, and 2.5 min at 72°C except that, in the last cycle, extension was carried out for 7 min. The amplified cDNA (854 base pairs) was subcloned into the SmaI site of pUC18 and used to transform Escherichia coli ONE SHOT TM as recommended by the manufacturer (Invitrogen). After growth at 37°C in Luria-Bertani medium (35) containing 100 g/ml ampicillin, the pUC18 containing SET cDNA was purified from the cells by the alkaline lysis method (35). The identity of the purified cDNA with the coding region of SET (22) (GenBank accession no. M93651) was confirmed by sequencing both strands using M13 sequencing primers and the dideoxynucleotide chain termination method (U. S. Biochemical Corp. TAquence sequencing kit and DuPont NEN [␣-35 S] dATP).
Expression and Purification of SET-SET cDNA was excised from pUC18 by incubation with BamHI and NdeI and then isolated with GLASSMILK ® (Geneclean II, BIO 101, Inc.) from low melting agarose following gel electrophoresis. pET-21a DNA (50 ng) (Novagen) was linearized with BamHI and NdeI and dephosphorylated with calf intestinal alkaline phosphatase. This DNA was then incubated overnight at 16°C with 10 ng of the purified cDNA in 0.01 ml of 20 mM Tris-HCl, pH 7.6, containing 1 mM ATP, 5 mM MgCl 2 , 5 mM dithiothreitol, 50 g/ml bovine serum albumin, and 4 units of T4 DNA ligase (Invitrogen) as recommended (35). The mixture was then used to transform E. coli BL21(DE3) (Novagen). A 5-ml overnight culture of these cells was then inoculated into 500 ml of Terrific Broth containing 100 g/ml ampicillin. After growth at 37°C to log phase, IPTG was added (final concentration, 0.5 mM), and the cells were grown for an additional 2 h. Unless indicated otherwise, all subsequent operations were performed at 4°C. After centrifugation for 15 min at 3,000 ϫ g in a Beckman JA-10 rotor, cells were resuspended in buffer A (25 mM Tris-HCl, pH 7.4, containing 10% glycerol, 4 mM EDTA, 1 mM benzamidine, 0.1 mM phenylmethanesulfonyl fluoride, and 14 mM ␤-mercaptoethanol) and lysed at 1200 p.s.i. with a French ® press. After centrifugation at 39,000 ϫ g for 30 min in a Beckman JA-20 rotor, pellets were discarded, and to the supernatant was added 30 volumes of buffer B (buffer A containing 1 mM instead of 4 mM EDTA). The mixture was applied onto a column (2.5 ϫ 8.5 cm) of poly-L-lysine-agarose equilibrated with buffer B. After washing with 500 ml of buffer B, the column was developed with a 600-ml linear gradient from 0 to 1.0 M NaCl. Inhibitor activity was recovered at about 0.65 M NaCl. Active fractions were pooled and mixed at room temperature with 2 volumes of a 99% ethanol solution. After centrifugation at 30,000 ϫ g for 10 min in a Beckman JA-14 rotor, pellets were resuspended in buffer B and subjected to gel permeation chromatography on a calibrated column (2.5 ϫ 95 cm) of Sephacryl S-200 equilibrated with buffer B containing 0.1 M NaCl and 0.01% Brij 35. A major peak of inhibitor protein activity emerged at V 0 (ϳ220 ml) from Sephacryl S-200. This activity peak was not detected following chromatography on Sephacryl S-200 of extracts from mock-transformed or transformed cells that had been incubated in the absence of IPTG. Fractions containing this inhibitory activity were pooled, and 20 g/liter trichloroacetic acid was added with stirring for 10 min. After centrifugation, the supernatant was discarded, and pellets were resuspended in a solution of 70% ethanol. The mixture was centrifuged and the supernatant was discarded. This procedure was repeated three times. The final pellets were resuspended in buffer C (buffer A containing 0.1 mM instead of 4 mM EDTA) and dialyzed, with three changes, against 20 volumes of this buffer in 16 h. The preparations were aliquoted and stored at Ϫ70°C. A second peak of unidentified inhibitor activity emerged from Sephacryl S-200 with apparent M r ϳ50,000. This inhibitor activity was detected following gel permeation chromatography of extracts from transformed and mock-transformed cells that had been incubated in the absence or presence of IPTG and was therefore discarded.

RESULTS AND DISCUSSION
Amino Acid Sequencing-To establish the identity of I 2 PP2A , we set out to determine the amino acid sequence of this protein.
Initially, the amino acid sequences of two tryptic peptides (LNEQASEEILK, peptide I) and (QHEEPESFFTWFTDH, peptide II), generated and purified as described under "Experimental Procedures," were determined. Comparison with amino acid sequences available at the PIR, SwissProt, and GenBank data bases indicated that peptides I and II were identical to residues 45-55 and 182-196 predicted for human SET, respectively (22). Earlier, we reported (21) that the N-terminal amino acid sequence of I 2 PP2A (SDGADATSTK) showed 70% identity (differences are underlined) with residues 17-26 of SET (22). Because the purified preparations of bovine kidney I 2 PP2A exhibit an apparent M r ϳ20,000 (21), whereas SET exhibits an apparent M r ϳ39,000 as estimated by SDS-PAGE (23,24), the results suggested that I 2 PP2A may have been derived from SET by proteolysis possibly during the purification procedure and that the three amino acid differences noted in the N terminus may be species and/or tissue related. In this connection, it is pertinent that the human SET gene encodes two transcripts of 2.0 and 2.7 kilobases that result from the use of alternative polyadenylation sites. However, both transcripts contain identical open reading frames (2,22).
Expression and Purification of SET-To test whether SET inhibits PP2A in a manner corresponding to I 2 PP2A , a cDNA coding for human SET was generated and placed into a pET-21a vector under the control of the T7lac promotor as described under "Experimental Procedures." Consistent with the possibility that SET inhibits PP2A, IPTG-induced expression of this cDNA in bacteria resulted in about a 10-fold increase in PP2A inhibitor activity as determined following poly-L-lysine-agarose chromatography (Fig. 1). Because of interference from an unidentified endogenous inhibitor(s), differences in PP2A inhibitor activity in bacterial extracts from control and IPTG-treated cells could not be distinguished. This (these) endogenous inhibitor(s) was (were) not a product(s) of the introduced cDNA because similar activity was detected in extracts of bacteria SET Is a Potent Inhibitor of Protein Phosphatase 2A 11060 that had been mock-transformed or transformed with the vector alone.
To more directly test the possibility that SET inhibits PP2A, a procedure was developed to purify the IPTG-induced PP2A inhibitor to apparent homogeneity from the bacterial extracts as described under "Experimental Procedures." This procedure was based on the one employed previously in the purification of I 2 PP2A from extracts of bovine kidney (21) and included chromatography of the extracts on poly-L-lysine-agarose, precipitation with ethanol, and gel permeation chromatography on Sephacryl S-200, followed by trichloroacetic acid and ethanol precipitation. The purified preparations of the recombinant protein consisted of a single Coomassie Blue staining polypeptide of apparent M r ϳ39,000 as estimated by SDS-PAGE (Fig.  2). N-terminal amino acid sequencing of these preparations (to the 10th residue) confirmed that the purified protein was SET. Typically, about 3 mg of this protein was obtained from a 500-ml culture of transformed cells.
Effect of SET on PP2A Activity-The effect of the purified recombinant SET preparations on PP2A activity was examined next. These preparations inhibited PP2A potently (Fig. 3). Half-maximal inhibition of the phosphatase occurred at about 2 nM (Fig. 3), similar to the potent inhibition obtained with purified preparations of bovine kidney I 2 PP2A (21). Previously, we showed that, by contrast to PP2A, purified preparations of PP1 C , PP2B, and PP2C were unaffected by I 2 PP2A (21). Similar experiments revealed that recombinant human SET also exhibited little or no effect on the activities of PP1 C , PP2B, and PP2C (Fig. 3). Together, these results indicate that SET is a potent and specific inhibitor of PP2A. Because the recombinant SET preparations inhibited PP2A 1 (Fig. 3) PP2A 2 (not shown), and PP2A C (not shown), SET is analogous to I 2 PP2A in that it appears to act by binding to the C subunit of the phosphatase.
Earlier studies suggested a role for SET as a transcriptional activator because it enhanced adenovirus core DNA replication in HeLa cell extracts (24). The molecular basis of this effect was not determined, although our results raise the possibility that it may have been a consequence of PP2A inhibition. A role for SET in antigen-mediated responses was also suggested because this protein appeared to bind to a peptide (CFIIKGLRK-SNAAERRGPL) patterned on an amino acid sequence present in the cytoplasmic C-terminal region of the DR2␣ chain of human histocompatibility class II receptor (23). However, the functional significance of this interaction and whether it occurs with the intact receptor were not determined. Based on the results presented herein, we recommend that SET be renamed I 2 PP2A to indicate its function and to distinguish it from I 1 PP2A , which is the product of a distinct gene. 2 This revised nomenclature is used in the remaining discussion.
The results presented in this communication provide a firm basis for further characterization of the physiological role of I 2 PP2A . Tissue and species distribution studies indicate that, by analogy with PP2A (3-6), I 2 PP2A is ubiquitous (36), suggesting its potential importance for controlling the activity of the phosphatase in diverse cells and tissues. Interestingly, I 2 PP2A undergoes phosphorylation on unidentified serines in intact cells (36). However, how this phosphorylation affects I 2 PP2A activity and whether or not this regulation responds to extracellular stimuli is unknown. Further studies are also needed on the mechanism of action of I 2 PP2A . In this connection, the C terminus of I 2 PP2A (22) (and that of I 1 PP2A2 ) is highly acidic (22), and deletion of this acidic tail (residues 224 -277) abolished the stimulation by I 2 PP2A of adenovirus core DNA replication in HeLa cell extracts (24). However, whether deletion of the acidic tail abolishes the inhibition of PP2A by I 2 PP2A (and/or I 1 PP2A ) remains to be determined. In this regard, it is pertinent that, 2 M. Li, A. Makkinje, and Z. Damuni, manuscript in preparation.

FIG. 2. SDS-PAGE pattern of purified recombinant SET (1 g).
PAGE was performed in slab gels (12% acrylamide) with 0.1% SDS and Tris/glycine buffer, pH 8.3 (32). Purification of SET was as described under "Experimental Procedures. "   FIG. 1. Induction of PP2A inhibitor. Extracts from 500-ml cultures of control (A) and IPTG-treated (B) bacteria containing SET cDNA were prepared as described under "Experimental Procedures." Each extract was then applied onto a separate column (2.5 ϫ 8.5 cm) of poly-L-lysineagarose equilibrated with buffer B. Each column was washed with 500 ml of buffer B followed by buffer B containing 0.3 M NaCl. Inhibitor activity was recovered with buffer B containing 0.8 M NaCl. Fractions (3 ml) were collected, and PP2A inhibitor activity (q) was determined according to Li et al. (21). The absorbance at 595 nm (E) was determined according to Bradford (33) using 0.05-ml aliquots of the indicated fractions.

SET Is a Potent Inhibitor of Protein Phosphatase 2A 11061
based on the apparent M r (ϳ20,000) as estimated by SDS-PAGE and that only 16 amino acids are missing at the N terminus, the bovine kidney I 2 PP2A preparations appear to be proteolyzed largely at the acidic C terminus. Thus, it would appear that the acidic tail may not be important for inhibition. However, how removal of the 16 N-terminal amino acid residues affects the mobility of I 2 PP2A on SDS-PAGE is unknown. In this context, it is pertinent that, although the calculated molecular mass of I 2 PP2A is 32,100, this inhibitor protein exhibits an anomalous apparent M r (ϳ39,000) as determined by SDS-PAGE.
Previous studies have indicated a role for PP2A in tumorigenesis and the viral transformation of cells (3)(4)(5)(6). Thus, PP2A is inhibited potently in vitro (37,38) and in vivo (39) by the tumor promotor okadaic acid. In addition, the SV40 small tumor t antigen of several DNA tumor viruses replaces the B subunit and inhibits the phosphatase in transformed cells (40). The results presented in this communication suggest that the I 2 PP2A -Nup214 gene fusion, which occurs in acute non-lymphocytic myeloid leukemia (22), may also lead to altered regulation of PP2A and thus contribute to leukemogenesis. The chimeric protein predicted by this gene fusion contains residues 1-270 of the 277 amino acid residues of native I 2 PP2A and residues 813-2090 of the 2090 amino acid residues of Nup214. However, although Nup214 appears to be asymmetrically located on the cytoplasmic side of the nuclear envelope in normal cells (41,42), I 2 PP2A (23) and the I 2 PP2A -Nup214 fusion protein (41, 42) occur largely in the nucleus of normal and leukemic cells, respectively. Thus, it will be important to determine how, relative to normal cells, the activity of nuclear PP2A is affected in leukemic cells that express the I 2 PP2A -Nup214 fusion protein.