Oncogene (2006) 25, 3247–3257 & 2006 Nature Publishing Group All rights reserved 0950-9232/06 $30.00 www.nature.com/onc ORIGINAL ARTICLE The human homologue of the RNA polymerase II-associated factor 1 (hPaf1), localized on the 19q13 amplicon, is associated with tumorigenesis

N Moniaux1,4, C Nemos1,4, BM Schmied2, SC Chauhan1, S Deb1, K Morikane2, A Choudhury1, M VanLith2, M Sutherlin2, JM Sikela3, MA Hollingsworth1,2 and SK Batra1,2

1Department of Biochemistry and Molecular Biology, University of Nebraska Medical Center, Omaha, NE, USA; 2Eppley Institute, University of Nebraska Medical Center, Omaha, NE, USA and 3Department of Pharmacology, University of Colorado Health Sciences Center, Denver, CO, USA

The 19q13 amplicon in pancreatic cancer cells contains a somes, chromosomal translocations, and amplifi- novel pancreatic differentiation 2 (PD2) gene (accession cations, induce a transformed phenotype leading to number AJ401156), which was identified by differential cancer. These genetic alterations constitute key events screening analysis. PD2 is the human homologue of the contributing to tumor progression and metastasis. They RNA polymerase II-associated factor 1 (hPaf1). In yeast, are often stabilized when they confer a growth or Paf1 is part of the transcription machinery, acting as a survivaladvantage to the cells(Lengauer et al., 1998). docking in between the complexes Rad6-Bre1, Gene amplification (HSR, homogeneously staining COMPASS-Dot1p, and the phosphorylated carboxyl region and DM, double minute) is one of the most terminal domain of the RNA polymerase II. As such, important mechanisms leading to the alteration of gene Paf1 is directly involved in transcription elongation via expression in solid tumors. The best example studied is histone H2B ubiquitination and histone H3 methylation. the amplification of HER2 on the chromosomal locus The PD2 sequence is highly conserved from Drosophila to 17q12 in primary breast cancer. HER2 is named for humans with up to 98% identity between rodent and human epidermalgrowth factor receptor-2 and belongs human, suggesting the functional importance of PD2/ to the epidermalgrowth factor receptor (EGFR) family hPaf1 to maintain cellular homeostasis. PD2 is a modular of receptor tyrosine kinases (RTKs). In total, 25–30% of protein composed of RNA recognition motif, DEAD- women with breast cancer demonstrate amplification boxes, an aspartic/serine (DS)-domain, a regulator of the and overexpression of the HER2 gene (Slamon et al., condensation domain and myc-type helix– 1987). Amplification/overexpression of HER2 induces loop–helix domains. Our results further showed that PD2 the phosphorylation of p27Kip (inhibitor of the cdk2) on is a nuclear 80 kDa protein, which interacts with RNA threonine 157 by AKT and its retention in the cytoplasm polymerase II. In addition, we have demonstrated that the (Blain and Massague, 2002; Shin et al., 2002), leading overexpression of PD2 in the NIH 3T3 cells result in enhanced the cells to enter the cell cycle and proliferate. growth rates in vitro and tumor formation in vivo.Altogether, Amplicons often contain several syntenic , such this paper presents strong evidence that the overexpression of as an amplification of the 11q13 locus that encompasses PD2/hPaf1 is involved in cancer development. FGF3, FGF4, and CYCD1 (Gaudray et al., 1992) or the Oncogene (2006) 25, 3247–3257. doi:10.1038/sj.onc.1209353; amplification of the locus 12q13 that encompasses published online 20 February 2006 MDM2, CDK4, and GLI (Khatib et al., 1993). We now recognize that amplicons are stabilized in the Keywords: hPaf1; hPAF1 complex; amplicon; pancrea- when they encompass severalgenes that tic cancer provide a selective growth advantage to the cell. In some cases, the stabilization of an amplicon is driven by amplification of nonsyntenic regions such as the co- amplification of the loci 12q13 and 2p24 that encom- passes MYCN (Khatib et al., 1993). Introduction Although DNA amplification is found in a wide range of tumor types, recurrent amplification of a particular The accumulation of genetic alterations in somatic cells, locus seems limited to specific tumors, such as glioma, such as mutations, changes in the number of chromo- breast and ovarian cancers. In an effort to identify differentially expressed genes that may play important Correspondence: Dr SK Batra, Department of Biochemistry and roles in pancreatic tumor growth and progression, Molecular Biology, University of Nebraska Medical Center, 985870 our laboratory pointed out the existence of a double Nebraska MedicalCenter, Omaha, NE 68198-5870, USA. minute amplification corresponding to the chromosomal E-mail: [email protected] 4These authors contributed equally to this work. locus 19q13 (Batra et al., 1991a, b). This genomic Received 25 August 2005; revised 8 November 2005; accepted 10 amplification is now recognized as an underlying cause November 2005; published online 20 February 2006 of many low-frequency genetic events (being amplified Human RNA polymerase II-associated factor 1 N Moniaux et al 3248 in 10–20% of the cases) in pancreatic adenocarcinoma total poly(A+) (Griffin et al., 1994; Ruggeri et al., 1998; Altomare et al., RNA RNA 2003). The 19q13 locus is also amplified in several other cancers such as follicular lymphoma (Werner et al., 1997), Mantle cell lymphoma (Werner et al., 1997), Burkitt’s lymphoma, small-cell lung cancer (Ried et al., CD11/HPAFPanc1 CD11/HPAFPanc1 1994; Petersen et al., 1997), non-small-cell lung cancer (Petersen et al., 1997; Bjorkqvist et al., 1998), breast 2.4 kb carcinoma (Kallioniemi et al., 1994), and uterine cervix cancer (Heselmeyer et al., 1997). Two amplified genes have been characterized at the PD2 19q13 locus, including the ribosomal protein, rpS16, and AKT2 (Batra et al., 1991b; Cheng et al., 1996). AKT2 1.4 kb belongs to a family of three members, AKT1, 2, and 3, which share a strong homology with the PKC and PKA families (Bauer and Baier, 2002). Activation of AKT2 promotes a variety of biological activities involved in tumorigenesis, such as cell survival and cell-cycle progression. In addition, AKT2 overexpression is β-actin reported to be associated with an increase in the aggressiveness of pancreatic cancer cells (Cheng et al., Figure 1 Northern blot analysis of PD2 cDNA with RNA from 1996). For these reasons, AKT2 is considered as the poorly and well-differentiated pancreatic cancer cell lines. Total main target gene associated with the 19q13 amplifi- RNA and purified poly(A) RNA were fractionated by electro- cation. phoresis on 1.2% agarose gelcontaining 0.66 M formaldehyde and transferred onto nitrocellulose via capillary blotting. Lanes 1 and 2 We hypothesized that the stabilization of the 19q13 contained totalRNA from CD11/HPAF and Panc1, respectively. amplicon, detected in 10–20% of the pancreatic Lanes 3 and 4 contained poly(A þ ) RNA from CD11/HPAF and adenocarcinomas, required the presence of another gene Panc1, respectively. The membrane was hybridized with 106 cpm/ favoring tumor development and progression and acting mlof a-32P-labeled PD2/Paf1 cDNA probe. After phospho-imager scanning, the membrane was stripped with 0.1%SDS/0.1 Â SSC synergistically with AKT2. In the present paper, we and re-hybridized with 106 cpm/mlof a-32P-dCTP labeled b-actin describe the identification and characterization of such a probe. novelgene, pancreatic differentiation factor 2 ( PD2). We show that the PD2 gene product is the human homologue of the RNA polymerase II-associated factor 1 (hPaf1), a key component of a multifunctional The chromosomal localization of the PD2 gene was transcriptional complex that is involved in the regula- determined by screening both the CEPH megabase-insert tion of gene transcription. The domain-structure of PD2 YAC DNA and the Coriell human X rodent somatic cell is conserved in evolution from yeast to human with up hybrid DNA pools. PD2 was mapped to the short arm to 98% identity between rodent and human. We of between the q13–qter regions. The demonstrate clearly that PD2/hPaf1 possesses trans- NCBI (NationalCenter of BiotechnologyInformation) forming activity both in vitro and in vivo. human genome resources database was used to resolve its precise location. The PD2 gene was located on chromosome 19 in the q13.2 (loc126275) region, oriented from centromere to telomere, between the Results IgG binding protein gene (FcGBP) and the zinc-finger protein 36 gene (ZFP36). Interestingly, both PD2 and Identification of the human homologue of yeast Paf1 AKT2 presented the same chromosomal localization, as part of the 19q13 amplicon 19q13.2 (Figure 2). Since AKT2 is known to be present To identify markers of differentiation for pancreatic in an amplicon, we hypothesized that PD2 might be cancer, single-stranded cDNA probes synthesized from carried on the same chromosomalamplification. South- both poorly and well-differentiated tumor cells, Panc1 ern blot analysis was performed on Panc1 and CD11/ and CD11/HPAF, respectively, were used to screen a HPAF cells (Figure 3), and revealed that Panc1 cells lgt11 phage cDNA library of Panc1 cells (Batra et al., contained a 30-fold gene amplification of PD2 as 1991a, b). In total, 17 clones were isolated, subcloned compared to CD11/HPAF. The ubiquitous expression into the pBluescript vector, and sequenced. Among the of PD2/hPaf1 in human, as revealed by blast scan using clones isolated, three were previously reported as the partialPD2 cDNA sequence, and presence of DNA- human ribosomalprotein S16 (Batra et al., 1991b), the and RNA-binding domains within its sequence, led us to human ribosomalprotein rpL17 (Batra et al., 1991a), hypothesize that PD2 might act in concert with AKT2 and AKT2 (Cheng et al., 1996). Another clone from this for the stabilization of the 19q13 amplification. screening, called PD2, showed an overexpression of To validate this hypothesis, we decided to extend and approximately 30-fold in Panc1 as compared to CD11/ fully characterize the PD2 partial sequence. The PD2 HPAF (Figure 1). insert was used to screen a lgt11 phage cDNA library

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3249 Chromosome 19 was flanked by a Kozak consensus sequence and a 30-untranslated region of 138 bp. The noncoding 30 19p13.3 region contained the polyadenylation signal, AATAAA. An open reading frame, from bp 157 to 1752, yields a 19p13.2 predicted translation product of 59.9 kDa. A compar- ison of the genomic sequence (Human Genome Re- HNRPL 19p13.12 sources Database) with the cDNA sequence provided SIRT2 the genomic structure of PD2 (Figure 4). The PD2 gene 19p13.11 NFKBIB was 5178-bp long and contained 14 exons of sizes 19p12 FBG4 F-box protein ranging from 30 to 551 bp (Table 1). The first exon 0 19p11 AKT2 contained the 5 -untranslated region, the translation 19q11 MAP3K10 start site, and the first 16 amino-acid residues. The last PSMC4 exon contained sequences that coded for a domain rich 19q13.11 in serine and aspartic acid residues and the 30- FCGBP untranslated region. The length and class of each exon 19q13.13 PD2/hPaf1 and intron are summarized in Table 1. Introns ranged in 19q13.2 ZFP36 size from 78 to 1539 bp. The sequences at the exon/ 19q13.32 RPS16 intron boundaries were highly conserved with respect to SUPT5H canonicalacceptor/donor site (AG/GT). 19q13.41 The predicted translation product of the PD2 cDNA DLL3 was examined for the presence of signature motifs, using 19q13.43 PROSITE software at the Expert Protein Analysis Figure 2 A scheme showing the localization of AKT2 and PD2 on System proteomics server of the Swiss Institute of the chromosomallocusof 19q13.2. The direction of arrows Bioinformation. This analysis revealed a high degree indicates the orientation of transcription of the genes on the of similarity between PD2 and the functional domains chromosome. of DNA- and RNA-binding . A schematic representation of the predicted translation product is presented in Figure 4. A number of domains were a Panc1 CD11/HPAF b identified, including three myc-type helix–loop–helix, a leucine zipper, a DEAD-box subfamily ATP-dependant RI dIII mHI RI dIII HI Eco Hin Ba Eco Hin Bam 12345 6 helicase domain, one eukaryotic RNA recognition motif (RRM) RNP-1 region signature, and a regulator of chromosome condensation (RCC1) signature. Taken 12.2 kb together, the presence of these domains suggested that putative functions for the PD2 protein include DNA or RNA binding. Two specific domains were also posi- 6.1 kb tioned toward the carboxyl-terminus: a glutamic acid- 5.1 kb rich domain, from amino acid 358 to 452, and a serine/ aspartic acid-rich domain, from amino acid 402 to 531. 4.0 kb An additional sequence analogy analysis was per- formed using the database from the NCBI (National 3.0 kb Center of Biotechnology Information). This revealed a high degree of similarity between PD2 and sequences from mouse (AK017762, AB041615, BC010317), Droso- Figure 3 Southern blot analysis of genomic DNA (10 mg) from phila melanogaster (AY070561.1, AC008139, and Panc1 and CD11/HPAF pancreatic tumor cell lines. After digestion AE003605), Caenorhabditis elegans (NP_505925, with the indicated enzymes, the DNA was separated on 0.8% NM_073524 and CAB02869.1), Schizosaccharomyces agarose geland Southern blottedas per standard methods. Lanes pombe (CAB65804), and Saccharomyces cerevisiae 1, 2, and 3 are Panc1 and lanes 4, 5, and 6 are CD11/HPAF DNAs, respectively, digested with EcoRI, HindIII, and BamHI. The blot (P38351). All homologue sequences were compared was hybridized to a-32P-dCTP labeled PD2 cDNA probe in (a). The and aligned as displayed in Figures 4 and 5. The human agarose gelbefore blottingwas stained with ethidium bromide ( b) PD2 showed 98% and 50% similarity to its rodent and to demonstrate an equal DNA loading in various lanes. Drosophila counterparts, respectively. The S. cerevisiae homologue corresponded to the widely investigated Paf1 protein, which showed 22% identity with the human constructed from normalfetalpancreas-derived mRNA. sequence, and 44% similarity for a segment of 333 A larger clone (1.9 kb) was isolated, subcloned into the amino-acid residues that excluded the carboxyl termi- pBluescript vector, and sequenced. The complete PD2 nus. Altogether, 22 amino acids were totally conserved cDNA sequence was submitted to the GenBankt/EBI among human, rodent, Drosophila, and S. cerevisiae, Data Bank under the accession number, AJ401156. This including two tyrosine residues. The least conserved sequence contained a 50-untranslated region of 156 bp domain was found in the carboxyl-terminal region. This upstream of the ATG translation initiation codon, and domain is very rich in serine and aspartic acid residues

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3250 a Mouse

Drosophila

MYC type Helix-loops-helix DEAD-box Ser, Asp Leucine zipper Glu rich RRM domain domain RCC1

Human

b

c Ser, Arg rich domain Caenorhabditis elegans

Figure 4 Schematic representation of the modular structure of PD2 and its homologues. (a) Comparison of human, mouse and Drosophila PD2. The domain structure and the relative size of the molecule are conserved. (b) Schematic representation of the genomic organization of PD2.(c) Schematic representation of the Caenorhabditis elegans sequence corresponding to Paf1.

Table 1 Characteristic of the exon–intron boundaries of the human PD2/hPaf1 gene Exon Intron

No Size (bp) 50-Splice donor Name Size (bp) Class 30-Splice acceptor

1 >211 CCACAGgtaggcgcca 1 530 2 tctccactagGCCCAA 230TGAGAGgtgagcccca 2 98 2 atgttcctagGTCTGG 3 93 GAACAGgtgagactaa 3 307 2 ccccacccagGT T C G T 4 122 C C A A T G gtgtgtgggg 4 93 0 ctctccttagTTCTTC 587CAAGAGgtgagtgggt 5 76 2 cccaccccagATCCCA 6 102 G G T C A A gtaagttact 6 83 2 tcattcctagGATTGG 7 106 A A A T C A gtaattgagg 7 77 0 gtcccaccagATCTCA 8 69 TTTAAGgtcaggccca 8 81 0 atctccctagATGTGG 9 104 G A T T A G gtaggtggtc 9 85 2 ccccaaccagGGGCAT 10 117 T G A T G T gtgcgtgggt 10 92 2 tcatcttcagGTATGA 11 128 A A C C A G gta c t c a g c a 11 1539 2 c c c c c t g c agGGTCCG 12 106 G C T C A G gtaacaagca 12 121 0 ctgcttccagGAGGCA 13 91 G C T C A G gtaaatctgg 13 78 1 tcccttccagATGAGG 14 551

Upper case, exonic sequence; lower case, intronic sequence; bold underlined letters are 50 donor and 30 acceptor sites, respectively.

for human PD2, rich in serine and arginine for tions) were isolated and fused with the SP2/0 mouse Drosophila, and absent in the yeast Paf1. myeloma cell line. Hybridomas were screened for PD2 reactivity by ELISA. The antibody that exhibited the strongest reactivity (F169-3B2) was expanded and used PD2/hPaf1 is a nuclear protein that interacts with RNA in subsequent experiments. polymerase II (RNAPII and DNA) The subcellular distribution of PD2 in Panc1 and To explore the biological characteristics of PD2, we CD11/HPAF cells was investigated using immunoblot- generated monoclonal antibodies specific for a 22- ting and confocal microscopy. Immunoblotting analyses residue peptide (conjugated to KLH), derived from identified a protein of B80 kDa in the nuclear and total position 327–348 of PD2. Balb/c mice were immunized cellular fractions of Panc1, whereas no signal was with the peptide, and splenocytes from mice producing obtained from the cytoplasmic extract (Figure 6a). A specific antibodies (following six to seven immuniza- similar pattern was also observed for CD11/HPAF cells

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3251 human MAPTIQTQAQREDGHRPNSHRTLPERSGVVCRVKYCNSLPDIPFDPKFITYPFDQN-RFV mouse MAPTIQTQAQREDGHRPNSHRTLPERSGVVCRVKYCNSLPDIPFDPKFITYPFDQN-RFV drosophila MPPTINNSAVNSAAEKRPQRQTERK-SEIICRVKYGNNLPDIPFDLKFLQYPFDSH-RFV C. elegans MPQESSDAKRIEPPRKVDFMLKPRFTNTVPDVPFDAKFMTCPFVPLGRFV S. cerevisiae MSKKQEYIAPIKYQNSLPVPQLPPKLLVYPESPETNAD pombe MQENRRQDYILRVRYHNPLPPPPFPPKLINIPNPVK-QYA human QYKATS---LEKQHKHDLLTEPDLGVTIDLINPD-TYRIDPNV-LLDPADEKLLEEE--- mouse QYKATS---LEKQHKHDLLTEPDLGVTIDLINPD-TYRIDPNV-LLDPADEKLLEEE--- drosophila QYNPTS---LERNFKYDVLTEHDLGVTVDLINRE-LYQADSMT-LLDPADEKLLEEE--- C. elegans EFQPAA---IYRDYKHAVICDDDMGLNVDLIDLK-KYDEDPIETEIDEKDNILLEDD--- S. cerevisiae SSQLINSLYIKTNVTNLIQQDEDLGMPVDLMKFPGLLNKLDSKLLYGFDNVKLDKDDRIL pombe LPNFVST--LVQEKKIPIENDIELGMPLDLAGITGFFEGDTSWMHSDLSSVNLDPID--- human IQAPTSSKRSQQHAKVVPWMRKTEYISTEFNRYGISN-EKPEVKIGVSVKQQFTEEEIYK mouse IQAPTSSKRSQQHAKVVPWMRKTEYISTEFNRYGISN-EKPEVKIGVSVKQQFTEEEIYK drosophila TLTPTDSVRSRQHSRTVSWLRKSEYISTEQTRFQPQNLENIEAKVGYNVKKSLREETLYL C. elegans GAAKLIAKRSQQHSKLVPWMRKTEYISTEFNRFGVTA-DRQETKLGYNLKKNQQVEDMYR S. cerevisiae LRDPRIDRLTKTDISKVTFLRRTEYVSN-----TIAAHDNTSLKRKR-RLDDGDSDDENL pombe RSLLKVAGGSGSTHLEVPFLRRTEYISSEVARAASNR-GNLRLTASTSKALAEQRGRSLR human DRDSQITAIEKTFEDAQKSIS--QHYSKPRVTPVEVMPVFP-----DFKMWINPCAQVIF mouse DRDSQITAIEKTFEDAQKSIS--QHYSKPRVTPVEVMPVFP-----DFKMWINPCAQVIF drosophila DREAQIKAIEKTFSDTKSEIT--KHYSKPNVVPVEVLPIFP-----DFTNWKFPCAQVIF C. elegans DKQSQIDAINKTFEDVRKPVK--EHYSKKGVKAVEESFVFP-----DFDHWKHLFAHVQF S. cerevisiae DVNHIISRVEGTFNKTDKWQ----HPVKKGVKMVKKWDLLP------DTASMDQVYF pombe EVPKQLEAINESFDVVQQPLEQLKHPTKPDLKPVSAWNLLPNTSMAGIQHLMLRVADDLS human D---SDPAPKDTSGAAALEMMSQAMIRGMMDEEGNQFVAYFLPVEETLKKRKRD-QEEEM mouse D---SDLAPKDTSGAAALEMMSQAMIRGMMDEEGNQFVAYFLPVEETLKKRKRD-QEEEM drosophila D---SDPAPAGKNVPAQLEEMSQAMIRGVMDESGEQFVAYFLPTEQTLEKRRTD-FINGE C. elegans D---GDTITTEFEEEDERQQARESSVIKAMEFEDQKFAAVFVPTIGCLTSFMDD-LELER S. cerevisiae I----LKFMGSASLDTKEKKSLNTGIFRPVELEEDEWISMYATDHKDSAILENELEKGMD pombe ERSHSYSSLVNLQEGHNLTKRHEVALFMPSSAEGEEFLSYYLPSEETAEEIQAKVNDASA human DYAPDDV--YDYKIAREYNWNVKNKASKGYEENYFFIFREGDGVYYNELETRVRLSKRRA mouse DYAPDDV--YDYKIAREYNWNVKNKASKGYEENYFFIFREGDGVYYNELETRVRLSKRRA drosophila LYKEEEE--YEYKIAREYNWNVKTKASKGYEENYFFVMRQ-DGIYYNELETRVRLNKRRV C. elegans PFDEDMK--YEFLLSREYTFKMEH--LPPRDRDVFIMYHRNNVFQYNEVDCNVKMTRKPK S. cerevisiae EMDDDSHEGKIYKFKRIRDYDMKQVAEKPMTELAIRLNDKDGIAYYKPLRSKIELRRRRV pombe DVHEPFVYNHFRNFDASMHVNSTGLEDLCLTFHTDKDHPEANQVLYTPIYARSTLKRRHV human KAGVQSGT----NALLVVKHRDMNEKELEAQEARKAQLENHEPEEEEEEEMETEEKEAGG mouse KAGVQSGT----NALLVVKHRDMNEKELEAQEARKAQLENHEPEEEEEEEMEAEEKEAGG drosophila KVGQQPN-----NTKLVVKHRPLDSMEHRMQRYRERQLEVPGEEEEIVEEVREEEQMQII C. elegans MALSRK------SKLTLTYRNPSELEQKDMNKREAELYEQPKTRKQEILEKIQEK-KEE S. cerevisiae NDIIKPLVKEHDIDQLNVTLRNPSTKEANIRDKLRMKFDPINFATVDEEDDEDEEQPEDV pombe RAPVSLDA----VDGIELSLRDLND-EESLQLKRARYD--TFGLGNIKDLEEEEEKLRSV human SDEEQEKGSSSEKEGSEDEHSGSESEREEGDRDEASDKSGSGEDESSEDEARAARDKEEI mouse SDEEQEKGSSSEKEGSEDEHSGSESDREEGDRDEASDKSGSGEDESSEDEARAARDKEEI drosophila GETEKTSEDAAVGAQAASGADSPAQVARDRQSRSRSRTRSGSSSGSGSGSGSRASSRSKS C. elegans GGDSSDQSSDSDDDKPQKSRSDSSSDVSSDDDSPRKKEPTVDSDSD S. cerevisiae KKESEGDSKTEGSEQEGENEKDEEIKQEKENEQDEENKQDENRAADTPETSDAVHTEQKP pombe EGSLN--EELSEEEKPAESREQLESAEQTN-GVKPETQAQNMSASESQANSPAPPVEEGN human FGSDADSEDDADSDDEDRGQAQGGSDNDSDSGSNGGGQRSRSHSRS----ASPFPSGSEH mouse FGSDADSEDDADSDDEDRGQAHRGSDNDSDSGSDGGGQRSRSQSRSRSRSASPFPSGSEH drosophila GSRSGSGSRSRTNSPAGSQKSGSRSRSVSRSRSRSKS-GSRSRSRSRSKSGSRSRSGSRS C. elegans S. cerevisiae EEEKETLQEE pombe TQPSPVEQLQNEED human SAQEDGSEAAASDSSEADSDSD mouse SAQEDGSEAAASDSSEADSDSD drosophila GSGSRSPSRSRSGSPSGSGSSSGSASDE C. elegans S. cerevisiae pombe Figure 5 Comparison of human PD2 with its mouse, Drosophila, Caenorhabditis elegans, S. pombe, and S. cerevisiae homologues. Identical amino acids are indicated by black letters on a gray background. The KKRK nuclear localization sequence is shown in bold letters.

(Figure 6a). In confocalimmunofluorescence micro- was pulled down using the F169-3B2 antibody, resolved scopy, F169-3B2 strongly reacted to the nuclei of Panc1 on 10% acrylamide gel, and immunoblotted with anti- cells (Figure 6c). The staining presented a punctuate RNA polymerase II rabbit polyclonal (Santa Cruz) distribution throughout the nuclei except for the antibody. The immunoblot revealed that RNAPII co- nucleoli and the perinuclear chromatin, which remained immunoprecipitated with PD2 (Figure 7). Two bands unlabelled. This pattern of labeling is specific for migrating at very close molecular weight were detected proteins implicated in the transcriptional process on the western blot. This result is in accordance with the (Grande et al., 1997). For CD11/HPAF cells, a mild specificity of the polyclonal anti-RNAPII antibody used reactivity was detected in the nuclei. To verify that the in this study that recognizes the tandem repeat domain reactivity of F169-3B2 was truly specific to PD2, this of the large subunit of the polymerase. The CTD antibody was preincubated with increasing concentra- (CarboxylTerminalDomain) of the RNAPII in mam- tions of the PD2327–348 peptide before immunoblotting malis composed by 52 repetitions of the YSPTSPS (Figure 6b). Binding of F169-3B2 to the 80-kDa band motif, which is phosphorylated on Ser2 and Ser5. was blocked by the PD2327–348 peptide. Likewise, Therefore, PD2 interacts with both Ser2- and Ser5- preincubation of F169-3B2 with 1 mg of PD2 peptide phosphorylated forms of the RNAPII. PD2 was also also blocked its reactivity in immunofluorescence assays detected in the fraction immunoprecipitated using the (data not shown). In addition, Panc1 protein extract RNAPII antibody (Figure 7).

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3252 IP a T b Panc1 N 1 N 1 C 1 T PD2 peptide Input PD2 KLH Panc Panc Panc CD11/HPAFCD11/HPAF NCD11/HPAF C 81 µg unblocked10 ng 100 ng1 PD2 RNAPII 220 KDa 40 PD2

PARP IP PGK Input RNAPII β globulin β-actin β-actin c PD2 80 KDa

Figure 7 Interaction between PD2 and RNAPII. The nuclear extract of Panc1 cells was immunoprecipitated with 1 mg/mlof anti- PD2 (F169-3B2) or anti-KLH monoclonal antibodies. Recovered Panc1 Panc1 Panc1 proteins were immunoblotted on 10% SDS–PAGE and incubated with anti-RNAPII rabbit polyclonal antibody C21 (Santa Cruz). A band at a molecular weight above 200 kDa (consistent with the expected molecular weight of RNAPII) was visualized after immunoprecipitation using anti-PD2 antibody (top panel). The same extract was used for immunoprecipitation using the anti- RNAPII antibody or b-globulin. The immunblotting with PD2/ CD18/HPAF CD18/HPAF CD18/HPAF hPAF1 antibody revealed an 80 kDa band (lower panel). As a PD2 Propidium iodide Overlay controlfor PD2/hPAF1 and RNAPII detection, the crude extracts were run on the same immunoblot (input). Figure 6 Biochemical characteristics and subcellular distribution of PD2. (a) The F169-3B2 PD2-specific antibody was used to analyse PD2 expression by immunoblotting of Panc1 (Left panel) and CD11/HPAF (right panel) cell lysates. Protein lysates were the 19q13 amplification. This result was confirmed by resolved on a 10% SDS–PAGE and transferred onto a PVDF confocal analysis. The NIH 3T3 parental cells presented membrane. After blocking with 5% BSA, F169-3B2 was used at a mild expression of PD2 while the sense-transfected 1 mg/ml. The N, C, and T after the cell line represent the nuclear, cytoplasmic, and total protein lysates, respectively. A protein with cells presented a strong expression of the exogene within an apparent molecular weight of 80 kDa was detected in Panc1 the nucleus of the cells (Figure 8b). Using this cellular nuclear (Panc1N) and total cell extracts (Panc1T). A similar but model, the in vitro effect of PD2 overexpression on less intense band was detected for CD11/HPAF. The PARP [poly growth kinetic was investigated. For this purpose, the (ADP-ribose) polymerase] was used as a positive control for cells were seeded at a low density on a six-well plate and nuclear fractionation, and the PGK (phosphoglycerol kinase) for cytoplasmic fractionation. (b) The F169-3B2 antibody was expanded for a period of 1 to 8 days. The PD2-sense preincubated with increasing doses (0.01, 0.1, and 1 mg/ml) of transfected cells showed faster growth with the doubling PD2 peptide and used to probe immunoblots of the Panc1N cell time of 18 h as compared to 24 and 22 h in PD2- extract. The 80 kDa band reactive to the F169-3B2 antibody antisense and control(mock-transfected), respectively disappeared in the presence of increasing concentrations of the peptide. (c) To further examine the subcellular distribution of PD2, (Figure 8c). Under an in vivo condition, the PD2-sense immunofluorescence microscopy was performed on Panc1 and transfected cells implanted at the subcutaneous site CD11/HPAF cell lines using the PD2 monoclonal antibody F169- showed altered growth in vivo compared to the PD2- 3B2. PD2 was found confined to the nucleus in more than 90% of antisense and the control (mock transfected) cells. The the methanol-fixed Panc1 cells. Lower reactivity was observed tumors of PD2-sense transfectants showed more tumor within the nucleus for CD11/HPAF cells. Preincubation of F169- 3B2 with 1 mg/mlof PD2 peptides inhibited the nuclearsignal(data volume (Figure 8d) and weight (Figure 8e) at any given not shown) (magnification  400). time point. After 20 days post-implantation, the PD2- sense transfected tumors showed a three-fold higher tumor weight and volume over the control and the PD2- The role of PD2 in the transforming phenotype antisense transfected cell tumors. To monitor the oncogenic potentialof the PD2 gene, the immortalized NIH 3T3 cells were transfected with the full-length PD2 cDNA in sense and antisense orienta- tions in the pcDNA 3.1 vector. The expression of the Discussion PD2 exogene was confirmed by Northern blotting (Figure 8) and confocalanalyses. The NIH 3T3 PD2 The acquisition of the tumor phenotype with the sense-transfected cells presented a four- to five-fold development of macroscopic tumor masses is a multi- overexpression of PD2 as compared to the NIH 3T3 step process with origin alteration of the genetic milieu. parental and antisense-transfected cells (Figure 8a). The These alterations lead to constitutive activation of level detected in the sense-transfected cells was still far oncogenes or inactivation of tumor suppressor genes. below the level detected in Panc1 cells, resulting from The constitutive activation of oncogenes is the result of

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3253

a e b tisense sens

NIH3T3Panc1NIH3T3/PD2NIH3T3/PD2 an NIH3T3 NIH3T3/PD2 sense

PD2

control

c Growth Curves for PD2 Transfected NIH3T3 Cell Lines d

3000000

2500000

Control ) 3 2000000 Antisense 2000 sense antisense * Sense 1500 1500000 Control 1000000 1000 Cell Number * 500000 500 *

0 0 12 3 45678 Tumor volume (mm 0 5 10 15 20 Days of growth Days

e 2500 2000 1500 1000 500 0

Tumor weight (mg) control sense antisense Figure 8 PD2 possesses transforming activity in NIH 3T3. PD2 sense and antisense were overexpressed in the NIH 3T3 murine fibroblastic cell line. (a) Northern blot analysis of RNAs extracted from NIH 3T3 single-cell clones transfectetd with PD2 construct. The NIH 3T3 transfected cells presented a five-fold overexpression of PD2 compared to parental and antisense cells. The Panc1 protein lysate was used as a positive control. The 28 s RNA stained with ethidium bromide was used as a loading control. (b) PD2 (green fluorescence) overexpression in NIH 3T3 transfected cells was confirmed by confocal analysis. For subcellular localization, the nucleus of the cell was labeled in red using propidium iodide. (c)5Â 104 cells were seeded in six-well plates and grown for zero to 8 days. PD2 transfected cells showed faster growth with a doubling time of 18 h as compared to 24 and 22 h in PD2 antisense and control cells. When subcutaneously implanted in nude mice, the PD2 transfected cells presented a three-fold increase in tumor volume (d) and weight (e). chromosomaltranslocation, amplification, and/or muta- et al., 1999). The PI3-K pathway is frequently activated tion within the gene itself that affects the amino-acid in numerous cancers following amplification or muta- residues regulating the functional activity of the gene tion of receptor tyrosine kinases, amplification of PI3-K product. One example of oncogene activation is the itself, or its downstream effectors such as AKT2. For 19q13.2 amplification of the AKT2 gene. AKT2 is these reasons, AKT2 is still known as the main target reported to be overexpressed in 12.1% of ovarian and gene associated with the 19q13 amplification. In 2.8% of breast tumors (Bellacosa et al., 1995). addition, overexpression of AKT2 in NIH 3T3 cells Amplification or overexpression of AKT2 was especially presents transforming activity in vitro and in vivo (Cheng detected in undifferentiated tumors (Bellacosa et al., et al., 1997) and an increase in the aggressiveness of 1995). AKT2, also known as protein Kinase B, is a breast cancer cell lines when implanted in nude mice serine/threonine kinase of 57 kDa (Bellacosa et al., 1991; (Arboleda et al., 2003). Jones et al., 1991). AKT2 is a downstream effector of In the present paper, we emphasize the presence the PI3-K (phosphatidylinositol 3-kinase) and mediates of another gene within the 19q13 amplicon that important cellular mechanisms such as cell survival and possesses transforming activity in the NIH 3T3 cells inhibition of apoptosis (Datta et al., 1997; Kennedy in vitro and in vivo. This gene, first named PD2, was

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3254 further characterized as the human homologue of the that the transient overexpression of wild-type HRPT2 yeast RNA polymerase II-associated factor I or hPaf1. resulted in cell proliferation inhibition, while its The functionalimportance of hPaf1 is underlined by the Leu64pro missense mutant detected in parathyroid remarkable conservation of its sequence from Droso- cancer does not. For this reason, HRPT2 is presented phila to human, with up to 98% identity in between as a tumor suppressor gene (Woodard et al., 2005). The human and rodents. The yeast Paf1 was first described observation that HRPT2/hCdc73 potentially presents as a component of a complex associated with RNAPII tumor suppressor activity in opposition with the and was believed to contribute to the transcriptional transforming activity of hPAF1 suggests that the hPAF1 function of this complex. Here, we demonstrate that, complex is crucial in maintaining cellular homeostasis similar to yeast Paf1, hPAF1 shows nuclear localization and that its function may be altered by a change in the and forms a complex with RNAPII. stoichiometry of the complex or by missense mutation. RNAPII is the major protein responsible for mRNA In addition to their presence on the same amplicon, synthesis in eukaryotes. It is now known that the hPaf1 and AKT2 share similarities; both are specifically initiation, elongation and post-transcriptional regula- overexpressed in undifferentiated tumor cells and both tion of transcription are complex tasks regulated by the in vitro and in vivo possess transforming activity. In interaction between numerous protein complexes, which addition, hPAF1 is a key regulator of transcription, control the action of RNAPII. The bulk of knowledge while AKT2 is a key player in the regulation of about this process is based on investigations carried out translation initiation via the FRAP/mTOR pathway in S. cerevisiae. Complete synthesis of mRNA by (Gingras et al., 2001). As mentioned previously, RNAPII includes at least four distinct steps: (I) amplicons are stabilized when they encompass several recruitment to the promoter and initiation; (II) detach- genes, especially genes that provide a selective growth ment from the promoter after the formation of the first advantage to the cancer cells. Specific chromosomal phosphodiester bond; (III) elongation; and (IV) termi- regions are mainly amplified in certain types of tumors nation. Paf1 and the PAF1 complex, which includes at and, therefore, may carry genes that facilitate the least four other members, Rtf1, Cdc73, Leo1, and Ctr9 development of these tumors. The tissue-specific ampli- (for a review, Squazzo et al., 2002; Hampsey and fication of the 19q13 locus in ovarian, breast, and Reinberg, 2003), are also found at the initiation site. pancreatic cancers suggests that hPaf1 and AKT2 Paf1 docks on the RNAPII phosphorylated carboxyl provide growth advantage and probably increase the terminaldomain (CTD) and bridges RNAPII with aggressiveness of the cancer cells. SPT4, SPT5, and SPT16 during the elongation process In summary, we demonstrated that the human (Squazzo et al., 2002). A knockout of Paf1 in yeast homologue of the RNA polymerase II-associated factor causes a severe phenotype that includes a reduced I gene localized within the 19q13 amplicon provides a growth rate, and substantially enhanced sensitivity to growth advantage to cancer cells and might act cytotoxic agents such as 6-azauracil, mycophenolic acid synergistically with the AKT2 oncogene. and hydroxyurea (Betz et al., 2002). This phenotype has been linked, in part, to effects on the transcription of severaldifferent genes (Betz et al., 2002). For example, Materials and methods sensitivity to hydroxyurea in yeast is conferred by specific increases in RNR1 transcript abundance, and Cell lines sensitivity to 6-azauraciland mycophenolicacid is The Panc1 cell line, a poorly differentiated (morphologically conferred through the derepression of the IMD2 and phenotypically) pancreatic adenocarcinoma cell line transcript (Betz et al., 2002). It is also postulated that (Lieber et al., 1975), was obtained from the American Type generalized post-transcriptional effects on mRNA may Culture Collection (ATCC). The well-differentiated pancreatic also contribute to this phenotype. Evidence exists that tumor cell line, CD11/HPAF, was established at Duke there are post-transcriptionalfunctions for Paf1 even University (Kim et al., 1989). Cell lines were cultured in when it is uncoupled from RNAPII (Mueller et al., Dulbecco’s modified Eagle’s medium supplemented with 10% fetalcalfserum. 2004). In the light of these findings, we propose that the overexpression of hPaf1 in human cells may confer growth advantage or resistance to cytotoxic agents, such Total RNA and DNA purification as those used in chemotherapy. Thus, hPaf1 may play a Total cellular RNA was isolated by the guanidine isothiocya- centralrolein the transcription process, contributing to nate-cesium chloride cushion ultracentrifugation method (Chirgwin et al., 1979). Poly (A) mRNAs were further purified the regulation of expression of specific genes, and to on two cycles of oligo (dT) cellulose affinity chromatography. generalized aspects of mRNA stability. In this manner, Genomic DNA from CD11/HPAF and Panc1 cell lines were the overexpression of hPAF1 may substantially con- purified by the SDS-Proteinase K digestion method and then tribute to cellular development, differentiation, meta- extracted with phenol/chloroform. bolic control, transformation and drug resistance in normaland tumor cells. Differential screening Interestingly, another member of the hPAF1 complex, The Panc1 cDNA library was subjected to differential Cdc73, was recently identified in humans as the product hybridization using single-stranded cDNA probes made of the hyperparathyroidism-jaw tumor syndrome gene from mRNA of Panc1 and CD11/HPAF cells. The probes HRPT2 (Woodard et al., 2005). The authors showed were synthesized using the SuperScripttII RnaseÀ Reverse

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3255 Transcriptase system (Invitrogen, USA), random hexamer corresponding to amino acids 327–348 (NH2-Glu Thr Arg Val primers (Pharmacia, Piscataway, NJ, USA), and 40 mCI a-32P- Arg Leu Ser Lys Arg Arg Lys Ala Gly Val Gln Ser Gly dCTP. RNA in the cDNA-RNA hybrid was hydrolysed at Thr Asn Ala Leu-COOH), and conjugated to KLH. At 5 days 651C for 30 min in an equalvolumeof 0.6 N NaOH and 30 mM after the final booster, the splenocytes were isolated and EDTA. The specific activity of cDNA obtained was 0.5– fused with SP2/O mouse myeloma cell line, using polyethylene 1.5 Â 108 cpm/mg of RNA. For screening, triplicate nitro- glycol (PEG) as described earlier. Hybridoma secreting PD2- cellulose membranes were lifted and subjected to alkaline specific antibodies were cloned by limiting dilution. Cells hydrolysis and neutralization. Prehybridization, hybridization were diluted to three cells/ml and plated (100 ml/well) into and washing were performed as described previously (Lan 96-well tissue culture plates. After 8 days, clones were et al., 1990; Batra et al., 1991a). Plaques that hybridized screened by an anti-PD2 ELISA, and those yielding the strongly with the Panc1 cDNA probe, but not with the CD11/ strongest signal were expanded. Female Balb/c mice were HPAF cDNA probe, were selected. The differential reactivity given intraperitonealinjections (0.5 ml)of pristane (Sigma) was confirmed through at least two additional screening cycles. to prime the peritoneum for ascites hybridoma growth. After Using a DNA insert derived from a differentially expressed 7–10 days, primed mice were inoculated with 106 cloned cDNA clone, five additional cDNA clones were isolated from hybridoma cells. MAb-rich ascites fluid was harvested after a normalhuman fetalpancreatic cDNA library. eight to 12 days.

Subcloning and sequence analysis Confocal microscopy Single phage plaques selected after differential screening were Cells (1–5 Â 105 cells) were plated on to sterile round coverslip grown to large quantities using either plate lysates or liquid (CIR 18-1 Fisher brand 12-545-10) and grown in 12-well culture followed by glycerol gradient purification (Lan et al., plates. Cells were fixed in acetone/methanol (1:1; prechilled to 1990). Inserts cut with EcoRI from purified DNA were À201C) and permeabilized in 0.1% Triton X-100 in PBS. Cells 7 subcloned into pBluescript vectors (Stratagene, La Jolla, were washed in PBS and incubated with primary and CA, USA). Sequences were determined on both strands by secondary antibodies, with washing between incubations. For conventionalmethods and by automated sequencing with an the blocking experiments, the PD2 antibody was pretreated ABI Prism model377 XL automatic sequencer. Analysesof with 1 mg/mlof PD2-specific peptide, overnight at 4 1Cina sequences were performed with the GCG software. The rotator. nucleotide sequence reported in this paper was submitted to the EMBL data bank with the accession number, AJ401156. Further analyses were undertaken with PROSITE software at Western blot analysis the Expert Protein Analysis System proteomics server of the The protocol for cytoplasmic and nuclear extract preparation Swiss Institute of Bioinformation and using the EST and was adapted from previously published procedures (Lee BLAST resources data from the NationalCenter of Biotech- and Green, 1990; Meyer et al., 2002). Cells were allowed to nology Information. swell on ice for 30 min in hypotonic cytoplasmic extraction buffer (10 mM HEPES, pH 7.4, 10 mM KCl, 0.2% NP-40, Northern and Southern blotting 0.1 mM EDTA, 10% glycerol, 1.5 mM MgCl2,1mM DTT, 1mM PMSF, 5 mM Na VO ,5mM NaF, supplemented TotalRNAs (20 mg) and/or purified poly(A þ ) RNAs were 3 4 fractionated by electrophoresis on 1.2% agarose gels contain- with complete protease inhibitor cocktail (Roche)). Cell disruption was accomplished by several passages through a ing 0.66 M formaldehyde and transferred onto nitrocellulose via capillary blotting. Genomic DNA were digested with the 25 G needle. Nuclei were collected by centrifugation at g g indicated endonuclease enzymes and separated on 0.8% 16 000 , and the supernatant further centrifuged at 16 000 agarose gel electrophoresis before transfer by capillary action to yield the final cytoplasmic extract. Nuclear pellets were incubated in hypertonic nuclear extraction buffer (20 mM to nylon membranes. The cDNAs were labeled with a32P– labeled dCTP using a random-primer labeling kit (Invitrogen). HEPES (pH 7.6), 420 mM NaCl, 1 mM EDTA, 20% glycerol, Prehybridization, hybridization, washing, and autoradiogra- 1.5 mM MgCl2,1mM DTT, 1 mM PMSF, 5 mM Na3VO4, phy or phosphor-imaging were performed as previously 5mM NaF, supplemented with complete protease inhibitor described (Batra et al., 1991b). cocktail (Roche)). Insoluble materials were precipitated by centrifugation. The supernatant was collected and used as nuclear extract. Total protein extracts were prepared by Chromosomal mapping lysing the cells in RIPA buffer (50 mM Tris-HCl, pH 7.4; The chromosomal localization of the PD2 gene was initially 0.25% Na-deoxycholate; 150 mM NaCl; 1% NP-40; 1 mM performed using the gene-based, sequence-tagged-site (STS) EDTA), freshly supplemented with 1 mg aprotinin, 1 mg/ml mapping method as published earlier (Berry et al., 1995). The leupeptine, 5 mM NaF, 5 mM Na VO , and 1 mM phenyl- 0 3 4 3 -untranslated region (UT) of the PD2 cDNA sequence was methylsulfonyl fluoride (PMSF). Equal amounts of protein used to design primers for polymerase chain reaction (PCR) (10 mg) were resolved by electrophoresis on 10% laemli screens of both CEPH mega base-insert YAC DNA pools, SDS–polyacrylamide gels. Fractionated proteins were electro- obtained from Research Genetics, Huntsville, AL, USA transferred onto polyvinylidene difluoride (PVDF) mem- (Bellanne-Chantelot et al., 1992), and Coriell human X rodent branes, which were blocked in 5% BSA in phosphate-buffered somatic cell hybrid DNA pools (Wilcox et al., 1991). To saline (PBS). Blocked membranes were reacted with anti-PD2 identify the PD2-specific locus, a walk on the chromosome was mouse monoclonal antibody at 1 mg/mlin PBS containing performed on the NationalCenter of BiotechnologyInforma- 0.05% Tween-20 (PBST) overnight at 41C. Membranes were tion server. then washed six times (ten minutes each) in PBST and incubated in TBST containing sheep anti-mouse polyclonal Generation of monoclonal antibody (Mab) against the PD2 antibody conjugated to horseradish peroxidase (1:2000 dilu- protein tion) for 1 h at room temperature. Following six rinses in Female Balb/c mice were immunized (six to seven times) PBST, ECL reagents (Amersham, USA) were applied to subcutaneously with 50 mg of the 22-residue PD2 peptide, membranes. For the blocking experiments, the PD2 antibody

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3256 was preincubated with 0.01, 0.1, and 1 mg/mlof PD2 peptide Abbreviations on a rotator at 41C overnight. PD2, pancreatic differentiation 2; Paf1, RNA polymerase II- associated factor; HLH, helix–loop–helix; cAMP, cyclic Immunoprecipitations adenosine monophosphate; kb, kilobase; bp, ; PCR, Immunoprecipitation was performed by adding 5 mg of the polymerase chain reaction; CGH, comparative genomic appropriate antibodies and negative controls IgG to 100 mgof hybridization; RRM, eukaryotic RNA recognition motif; protein lysates. The reaction mixture was incubated overnight RCC1, regulator of chromosome condensation domain; 1 on a rotator at 4 C, pulled down with protein G beads, and RNAPII, RNA polymerase II; GTFs, general transcription washed in TBST. Immuno-complexes were eluted from the factors. protein G beads by boiling in 6 Â loading buffer. Acknowledgements Plasmid construction and DNA transfection and tumorigenicity The full-length PD2 cDNA (1.9 kb) was cloned at the EcoRI The work reported in this manuscript was initiated in the late site in the pCDNA3.1 vector in both the sense and antisense 1980s by Drs Batra and Hollingsworth under the direction and orientation. The NIH3T3 cells were either mock transfected sponsorship of Professor Richard S Metzgar (Duke University with vector alone or with the PD2-sense and -antisense Medical Center, NC, USA), whom authors gratefully pCDNA3.1 construct using a procedure described earlier acknowledge for his thoughtful and scholarly insights, (Singh et al., 2004). The expression was driven by a strong exemplary leadership, generous support and lasting friendship. CMV promoter and the stable transfectants were selected We would like to thank Mr Jason Jokerst for making mouse using neomycin selection. The viable NIH3T3 cells (assessed monoclonal antibodies, Mr Erik Moore for technical assis- by Trypan blue dye exclusion test) transfected with vector tance, the Molecular Biology Core Facility, UNMC, for alone and PD2-sense or antisense construct were implanted oligonucleotide synthesis and DNA sequencing, Monoclonal subcutaneously in the nude mice. The body weight of the Antibody Core Facility for making antibodies and Ms Kristi animals and the size of the growing tumors were monitored LW Berger for editorialassistance. Dr Nemos acknowledges every alternate day. A cell growth curve and the cell doubling the Travel Award Fellowship from the American/French time of PD2-mock, PD2-sense, and PD2-antisense transfected Philippe Foundation. This work was supported, in part, by cells were established as reported (Singh et al., 2004). grants from the NIH (CA78590 and CA72712).

References

Altomare DA, Tanno S, De RA, Klein-Szanto AJ, Tanno S, Gingras AC, Raught B, Sonenberg N. (2001). Genes Dev 15: Skele KL et al. (2003). J Cell Biochem 88: 470–476. 807–826. Arboleda MJ, Lyons JF, Kabbinavar FF, Bray MR, Snow BE, Grande MA, Van der Kraan I, de Jong L, Van DrielR. (1997). Ayala R et al. (2003). Cancer Res 63: 196–206. J Cell Sci 110: 1781–1791. Batra SK, Metzgar RS, Hollingsworth MA. (1991a). Cell Griffin CA, Hruban RH, Long PP, Morsberger LA, Douna- Growth Differ 2: 385–390. Issa F, Yeo CJ. (1994). Genes Cancer 9: Batra SK, Metzgar RS, Hollingsworth MA. (1991b). J Biol 93–100. Chem 266: 6830–6833. Hampsey M, Reinberg D. (2003). Cell 113: 429–432. Bauer B, Baier G. (2002). Mol Immunol 38: 1087–1099. Heselmeyer K, Macville M, Schrock E, Blegen H, Hellstrom Bellacosa A, de FD, Godwin AK, Bell DW, Cheng JQ, AC, Shah K et al. (1997). Genes Chromosomes Cancer 19: Altomare DA et al. (1995). Int J Cancer 64: 233–240. 280–285. Jones PF, Jakubowicz T, Hemmings BA. (1991). Cell Regul 2: Bellacosa A, Testa JR, Staal SP, Tsichlis PN. (1991). Science 1001–1009. 254: 274–277. Kallioniemi A, Kallioniemi OP, Piper J, Tanner M, Stokke T, Bellanne-Chantelot C, Lacroix B, Ougen P, Billault A, Beaufils Chen L et al. (1994). Proc Natl Acad Sci USA 91: S, Bertrand S et al. (1992). Cell 70: 1059–1068. 2156–2160. Berry R, Stevens TJ, Walter NA, Wilcox AS, Rubano T, Kennedy SG, KandelES, Cross TK, Hay N. (1999). Mol Cell Hopkins JA et al. (1995). Nat Genet 10: 415–423. Biol 19: 5800–5810. Betz JL, Chang M, Washburn TM, Porter SE, Mueller CL, Khatib ZA, Matsushime H, Valentine M, Shapiro DN, Sherr Jaehning JA. (2002). Mol Genet Genomics 268: 272–285. CJ, Look AT. (1993). Cancer Res 53: 5535–5541. Bjorkqvist AM, Husgafvel-Pursiainen K, Anttila S, Karjalai- Kim YW, Kern HF, Mullins TD, Koriwchak MJ, Metzgar nen A, Tammilehto L, Mattson K et al. (1998). Genes RS. (1989). Pancreas 4: 353–362. Chromosomes Cancer 22: 79–82. Lan MS, Batra SK, Qi WN, Metzgar RS, Hollingsworth MA. Blain SW, Massague J. (2002). Nat Med 8: 1076–1078. (1990). J Biol Chem 265: 15294–15299. Cheng JQ, Altomare DA, Klein MA, Lee WC, Kruh GD, Lee KA, Green MR. (1990). Methods Enzymol 181: 20–30. Lissy NA et al. (1997). Oncogene 14: 2793–2801. Lengauer C, Kinzler KW, Vogelstein B. (1998). Nature 396: Cheng JQ, Ruggeri B, Klein WM, Sonoda G, Altomare 643–649. DA, Watson DK et al. (1996). Proc Natl Acad Sci USA 93: Lieber M, Mazzetta J, Nelson-Rees W, Kaplan M, Todaro G. 3636–3641. (1975). Int J Cancer 15: 741–747. Chirgwin JM, Przybyla AE, MacDonald RJ, Rutter WJ. Meyer T, Begitt A, Lodige I, van Rossum M, Vinkemeier U. (1979). Biochemistry 18: 5294–5299. (2002). EMBO J 21: 344–354. Datta SR, Dudek H, Tao X, Masters S, Fu H, Gotoh Y et al. Mueller CL, Porter SE, Hoffman MG, Jaehning JA. (2004). (1997). Cell 91: 231–241. Mol Cell 14: 447–456. Gaudray P, Szepetowski P, Escot C, Birnbaum D, Theillet C. Petersen I, Langreck H, Wolf G, Schwendel A, Psille R, Vogt (1992). Mutat Res 276: 317–328. P et al. (1997). Br J Cancer 75: 79–86.

Oncogene Human RNA polymerase II-associated factor 1 N Moniaux et al 3257 Ried T, Petersen I, Holtgreve-Grez H, Speicher MR, Schrock Squazzo SL, Costa PJ, Lindstrom DL, Kumer KE, E, du MS et al. (1994). Cancer Res 54: 1801–1806. Simic R, Jennings JL et al. (2002). EMBO J 21: Ruggeri BA, Huang L, Wood M, Cheng JQ, Testa JR. (1998). 1764–1774. Mol Carcinog 21: 81–86. Werner CA, Dohner H, Joos S, Trumper LH, Baudis M, Barth Shin I, Yakes FM, Rojo F, Shin NY, Bakin AV, Baselga J TF et al. (1997). Am J Pathol 151: 335–342. et al. (2002). Nat Med 8: 1145–1152. Wilcox AS, Khan AS, Hopkins JA, Sikela JM. (1991). Nucleic Singh AP, Moniaux N, chauhan SC, Meza JL, Batra SK. Acids Res 19: 1837–1843. (2004). Cancer Res 64: 622–630. Woodard GE, Lin L, Zhang JH, AgarwalSK, Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, Marx SJ, Simonds WF. (2005). Oncogene 24: McGuire WL. (1987). Science 235: 177–182. 1272–1276.

Oncogene