Pituitary tumor transforming gene (PTTG) transforming and transactivation activity.

Pituitary tumor transforming gene (PTTG) is a newly identified transforming gene, the functional mechanism of which is little understood. Computational analysis reveals a C terminus rich in Glu and Pro, a known characteristic of transcriptional activation domains. We report here that murine PTTG indeed possesses transactivation ability, which correlates highly with its transforming properties. Pro(139), Ser(159), Pro(157)-Pro(158)-Ser(159)-Pro(160) (PPXP motif), and Leu(120)-Asp(121)-Phe(122)-Asp(123)-Leu(124) were found to be important for transactivation. Mutation to Ala at a key Pro(139) residue not only disrupted the transactivation function but also resulted in the loss of transforming ability in NIH3T3 cells. A murine PTTG cDNA that encodes a variant C-terminal tail (Gly-Lys-Gly-Val-Arg-Ser-Asn-Gly-Cys-Lys-Asp-Leu-Val-Thr) was cloned. This novel PTTG is devoid of transactivation and transforming ability; deletion of its variant C-terminal tail restores both transactivation and transforming ability. These results show a high correlation between the transforming and transactivation functions of PTTG and also indicate that the novel PTTG variant may function as an endogenous competitor to wild-type PTTG.

sion in 3T3 cells resulted in increased expression of basic fibroblast growth factor-2, a potent angiogenic growth factor (3). PTTG has also been found to participate in early stage development of prolactinoma and is highly expressed in experimental and clinical pituitary tumors (4,5). PTTG recently was shown to behave as a vertebrate sister-chromatid separation inhibitor providing a potential mechanism for PTTG to mediate aneuploidy and genetic instability, thus contributing to cell malignancy (6).
Human PTTG has several gene homologues comprising a gene family including PTTG1 (most homologous to rat PTTG), PTTG2, PTTG3, and PTTG4 (7). We have also cloned the murine PTTG cDNA (most homologous to human PTTG1) and its promoter (8). In a detailed computational analysis of human, rat, and murine PTTG protein sequences, Glu and Pro residues were observed to be abundant in the C-terminal region, a known characteristic of transactivation domains in transcription factors and/or co-activators. We therefore fused murine PTTG with a GAL4 DNA binding domain and found that PTTG indeed possesses transcriptional activation ability; mutation analysis identified several residues and regions important for this activation, and soft agar assay suggested a high correlation between transactivating and transforming abilities of PTTG. Furthermore, a novel murine PTTG alternative transcript encoding a variant protein containing a different Cterminal tail was detected. This variant PTTG is devoid of both transactivation function and transforming ability. Deletion of the variant C-terminal tail restores both transactivation and transforming ability. These results imply that PTTG-mediated transactivation correlates with its transforming ability, and the variant C terminus may function as an endogenous competitor to wild-type PTTG.

EXPERIMENTAL PROCEDURES
Plasmids and Antibodies-pcDNA3.1 and pCRII vectors were from Invitrogen, and murine PTTG was cloned in this laboratory (8) and subcloned into pcDNA3.1. pGAL4 vector, which contains the GAL4 DNA binding domain, and pGAL4-VP16 as a positive control vector were from Stratagene. pLuc vector, which contains five copies of GAL4 DNA element in front of a minimal promoter, and the luciferase gene were also from Stratagene. Polyclonal antibody against a 20-amino acid peptide of rat PTTG was generated commercially (Research Genetics).
Cell Culture-Mouse fibroblast NIH3T3 (ATCC CCL-92) cells were maintained in Dulbecco's modified Eagle's medium low glucose medium (Life Technologies, Inc.) with 10% fetal bovine serum, and mouse MOP8 (ATCC CRL-1709) and mouse embryonal carcinoma F9 (ATCC CRL-1720) were maintained in Dulbecco's modified Eagle's medium high glucose medium (Life Technologies, Inc.) with 10% fetal bovine serum. All culture media were supplemented with standard antibiotics, and cells were passaged twice weekly.
Transcriptional Activation Assay-Murine PTTG cDNA was fused in-frame with pGAL4, designated pGAL4-mPTTG and used as template for all deletion and mutation analysis. pGAL4-VP16 was used as a positive control. Experimental plasmids were co-transfected with pLuc and pCMV-␤-Gal (as internal control). Cell lysates were prepared 48 h after transfection and assayed for luciferase activity as described (9,10).
Rapid Amplification of 3Ј cDNA Ends (3ЈRACE) and mPTTG Variant cDNA Cloning-3ЈRACE was performed as suggested by Roche Molecular Biochemicals. Total RNA derived from F9 cells was used for reverse transcription, and one PTTG gene-specific primer (5Ј-GCTC-CAGCCGTGCCTAAAGCCAG-3Ј) and universal primer were used in polymerase chain reaction. A probe derived from unique sequence in the alternative transcript was used to screen a F9 cDNA library, and full-length mPTTG variant cDNA was cloned and sequenced. * This work was supported by National Institutes of Health Grant CA75979 and the Doris Factor Molecular Endocrinology Laboratory. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s)AF071209.
Mutagenesis and Deletion-Mutagenesis and deletion of PTTG using pGAL4-mPTTG as template were performed as suggested using the ExSite polymerase chain reaction-based method (Stratagene). All mutants were sequence-verified. A total of 37 primers were used for mutation and deletion, and their sequences are available upon request.
Soft Agar Assay-The assay was performed as described previously (2). 5000 cells were seeded in triplicate for each cell line in the top layer in the presence of 0.5% agar in a 35-mm dish. 3 weeks after seeding, cell morphology was observed, and numbers of transformed colonies (Ͼ16 cells per colony) were counted. MOP8 cells were used as a positive control, and NIH3T3 cells were used as a negative control.

PTTG Possesses Transcriptional Activation Properties-
PTTG does not exhibit significant similarity to a known protein, but computational analysis shows that the N-terminal half (1-100 aa) is rich in alkaline residues (21/100), whereas the C-terminal half is rich in acidic residues, especially glutamic acid (8/96) and proline (17/96). Because an abundance of Glu and/or Pro is characteristic of transactivation domains for transcription factors and/or co-activators, we tested whether PTTG demonstrated transcriptional activation. Fig. 1 shows that PTTG exhibits transcriptional activation; the mutant human PTTG-M9 (replacing region 163-166 from Pro-Pro-Ser-Pro to Ala-Leu-Ala-Leu), which is devoid of transforming ability (3), also does not activate transcription. This result suggests that transforming and transcriptional activation functions of PTTG are distinct and correlate with each other.
cDNA Cloning of a Novel Murine PTTG Variant-Murine PTTG shows high homology with human and rat PTTG. Nevertheless, on Northern blot analysis, a second band (ϳ1.6 kb) was detected in RNA derived from embryonic F9 cells (8). To determine the presence of this alternative PTTG transcript, 3ЈRACE was performed and a ϳ1.3 kb band was amplified in addition to the band corresponding to wild-type mPTTG (Fig.  2a). Sequencing revealed that this transcript differed from wild-type mPTTG after exon 4. Using the distinct sequence obtained using this 3ЈRACE product as probe, an F9 cDNA library was screened and a ϳ1.7-kb mPTTG alternative transcript was identified, which encoded a 188-aa murine PTTG cDNA variant that has been deposited in GenBank TM (accession number AF071209). Comparison of this mPTTG variant with wild-type is shown in Fig. 2b, and major differences are apparent in the C-terminal tail.
Correlation between Transactivation and Transformation Induced by PTTG-Understanding PTTG structure-function relationships in terms of its transforming ability is important for unraveling its mechanism of action as an oncogene. However, establishing stable transfectant cell lines and performing anchorage-independent soft agar assays is time-consuming and not suitable for pilot scale screening. As mutant human PTTG-M9 is devoid of both transforming and transactivation ability, we reasoned that these two functions might correlate, and therefore we tested transactivating activity to identify amino acid residues important for this function. Transactivation-related Glu and Pro, phosphorylation-related Ser and Tyr, conserved acidic and/or alkaline residues, other conserved regions, and the novel mPTTG variant C-terminal tail were mutated and/or deleted and tested for transactivating activity ( Table I). Deletion of the first 100-aa mPTTG residues does not alter transactivation, indicating that this function is inherent within the C-terminal half of the mPTTG protein. Several residues were found to be important for transactivation including Pro 139 , Ser 159 , Pro 134 , and Glu 169 , and two regions including Leu 120 -Asp 121 -Phe 122 -Asp 123 -Leu 124 and Pro 157 -Pro 158 -Ser 159 -Pro 160 (PPXP motif). Mutation of Pro 139 to Ala or Leu resulted in complete loss of transactivation, whereas mutation of Pro 139 to Asn did not disrupt transactivation, showing that the Pro 139 tolerates a hydrophilic residue but not hydrophobic substitution for its activity. Mutation of Ser 159 to Ala or Ser 159 to Leu abrogated transactivation by 70 and 80%, respectively, whereas mutation of Ser 159 to Thr only reduced transactivation by 30%. Thus, maintaining the Ser 159 position hydrophilic appears critical, and this residue may be a potential phosphorylation site for regulation. Mutation of Glu 169 to Gln resulted in loss of 75% transactivation activity, showing that the acidic residue is important at this position. Two regions were also found to be important for transactivation, including a highly conserved region, Leu 120 -Asp 121 -Phe 122 -Asp 123 -Leu 124 , which when completely deleted disrupted transactivation. The other region, a PXXP motif from aa 157 to 160 is also important for the activity. Moreover, the mPTTG cDNA variant did not exhibit transactivating ability, whereas deletion of the variant tail Gly-Lys-Gly-Val-Arg-Ser-Asn-Gly-Cys-Lys-Asp-Leu-Val-Thr (equivalent deletion of 171-196 aa in wild-type mPTTG) restored transactivation.
Although these results provided useful insights for the trans-

PTTG Transformation and Transactivation 7460
activation function of mPTTG, we also tested these residues and/or regions for their transforming properties. Several constructs were chosen to establish stable NIH3T3 transfectants, and soft agar assays were performed to test transforming ability (Table II). As depicted, a high correlation between transactivating and transforming abilities was observed. Pro 139 is also an important residue for transforming ability, whereas the novel mPTTG variant failed to transform NIH3T3 cells. Deletion of the variant C-terminal tail restored the transforming ability of the variant cDNA. DISCUSSION PTTG is a highly conserved protein, with murine PTTG sharing 88% and 66% amino acid homology with rat PTTG and human PTTG1, respectively. Nonetheless, the C-terminal 26amino acid tail is relatively low in homology (about 31%) and seems unlikely to be important for conserved function of the molecule. Indeed, deletion of the last 26 amino acids did not affect transcriptional activation induced by mPTTG nor did it alter the transforming ability of the protein.
These results show good correlation between transforming ability and transcriptional activation mediated by murine PTTG. Similar findings have been obtained with other oncogenes, which function as transcription factors including Jun or Myb; Jun transactivates the Myb promoter via an AP-1-like sequence, and its N terminus demonstrates the transactivation property. Interestingly, the ␦ domain in human Jun, absent in the v-Jun oncogene, stabilizes the interaction of cell-specific transcriptional inhibition with activation regions (11). Expression of v-Src or oncogenic Ras disrupts the Jun-inhibitor complex, thus increasing transcriptional activity (12). The protooncogene Myb also contains DNA binding, transcriptional activation, and negative regulatory domains (13,14). In both cases inactivation of relevant transcriptional activation domains also abrogates transforming abilities. Notably, deletion of PTTG N-terminal 100 aa disrupts transformation, while not affecting transactivation. It is therefore possible that the N-terminal 100-aa region interacts with DNA element(s) or other proteins whereas the C-terminal portion mainly interacts with the RNA polymerase complex for transcriptional activation. Thus without the N-terminal 100 aa, mPTTG still maintains its transactivation ability when fused with the GAL4 DNA binding domain but is unable to interact with endogenous substrate in NIH3T3 cells to mediate its transforming effect.
The comprehensive deletion and mutagenesis analysis in the C terminal region demonstrates that it is most likely that, overall, a three-dimensional structure is important for function, and Pro 139 , Ser 159 , Pro 157 -Pro 158 -Ser 159 -Pro 160 , and Leu 120 -Asp 121 -Phe 122 -Asp 123 -Leu 124 are critical contact sites for PTTG function. We have recently demonstrated that human PTTG1 C terminus is also critical for in vitro and in vivo transforming. 2 On the other hand, whether PTTG is a true transcription factor/co-activator remains to be tested via efforts to identify its endogenous binding element and/or interacting proteins.
Identification of the new murine PTTG variant is interesting. This variant possesses a different C-terminal tail and does not exhibit transforming ability. Deletion of the variant Cterminal tail restores both transcriptional activation and transforming ability, indicating an inhibitory effect of the tail on PTTG function by conformational perturbation or interaction with specific residues. Thus the intracellular balance of PTTG and this variant PTTG may compete to determine their ultimate effects on the cell. Future comparison of gene expression profiles between wild-type PTTG and variant PTTG stable transfectants should help identify downstream target genes for PTTG.