MOLECULAR AND CELLULAR BIOLOGY, June 1994, p. 3752-3762 Vol. 14, No. 60270-7306/94/$04.00+0Copyright C 1994, American Society for MicrobiologY

PRL-1, a Unique Nuclear Protein Tyrosine Phosphatase,Affects Cell Growth

ROBERT H. DIAMOND," 2 DREW E. CRESSMAN,' THOMAS M. LAZ,1'2tCHARLES S. ABRAMS,2 AND REBECCA TAUB' 2,3*

Departments of Genetics' and Medicine2 and Howard Hughes Medical Institute,3 University of PennsylvaniaSchool of Medicine, Philadelphia, Pennsylvania 19104-6145

Received 9 November 1993/Returned for modification 21 December 1993/Accepted 1 February 1994

PRL-1 is a particularly interesting immediate-early gene because it is induced in mitogen-stimulated cellsand regenerating liver but is constitutively expressed in insulin-treated rat H35 hepatoma cells, whichotherwise show normal regulation of immediate-early genes. PRL-1 is expressed throughout the course ofhepatic regeneration, and its expression is elevated in a number of tumor cell lines. Sequence analysis revealsthat PRL-1 encodes a 20-kDa protein with an eight-amino-acid consensus protein tyrosine phosphatase(PTPase) active site. PRL-1 is able to dephosphorylate phosphotyrosine substrates, and mutation of theactive-site cysteine residue abolishes this activity. As PRL-1 has no homology to other PTPases outside theactive site, it is a new type of PTPase. PRL-1 is located primarily in the cell nucleus. Stably transfected cellswhich overexpress PRL-1 demonstrate altered cellular growth and morphology and a transformed phenotype.It appears that PRL-1 is important in normal cellular growth control and could contribute to thetumorigenicity of some cancer cells.

The liver constitutes one of the few normally quiescenttissues in the adult body that have the capacity to regenerate(3, 10, 33). As a result, it provides a unique multicellular,physiologically normal system in which to study the mitogenicresponse of epithelial cells. In the rat, following a 70%hepatectomy, the cells in the remaining, intact, liver lobesrapidly resume proliferation. They initiate the first round ofDNA synthesis within 12 to 16 h postsurgery and continue totraverse the cell cycle until the liver regains its initial mass inabout 10 days, whereupon the cells again become quiescent.Although it is composed mainly of hepatocytes, the liver has acomplex multicellular architecture, implying that intercellularcommunications must exist during regeneration. Multiple fac-tors, including circulating hormones, growth factors, and ner-vous input, participate in the regulation of this response, butthe actual mechanism remains incompletely understood.

Immediate-early genes, which are induced by mitogens inthe absence of de novo protein synthesis, encode proteinswhich play an important role in the regulation of cell growth (2,19, 29, 36, 55). They include proteins of diverse functions(transcription factors and growth factors, etc.), and it is clearthat during the growth response, multiple pathways must besimultaneously activated. The mRNA and protein levels ofmany of these genes remains elevated for extended periodsduring cell growth, and their regulatory role extends beyondthe immediate Go/Gl transition. At the cellular level, proteinsencoded by immediate-early genes may help control progres-sion through G, and later phases of the cell cycle. To begin tounderstand the control of liver regeneration, we identifiedimmediate-early growth response genes in regenerating liverand insulin-treated H35 cells, a minimal-deviation hepatomacell line that has many properties of regenerating liver. H35cells become quiescent under serum deprivation and reenterthe cell cycle in response to physiologic concentrations of

* Corresponding author. Mailing address: Department of Genetics,University of Pennsylvania School of Medicine, Philadelphia, PA19104-6145. Phone: (215) 898-9131. Fax: (215) 573-2195.

t Present address: Synaptic Pharmaceuticals, Paramus, NJ 07652.

insulin (50). In the two liver systems, 70 total and 41 novelimmediate-early genes were identified (36). Although themajority of the immediate-early genes are common to bothliver systems, one gene, PRL-1 (referred to as SL-314 inreference 36), was of particular interest, because it is inducedas an immediate-early gene in regenerating liver and mitogen-treated fibroblasts but constitutively expressed in H35 cells.Upon analysis of the sequence of PRL-1, we noted that it has

the signature sequence for a protein tyrosine phosphatase(PTPase) (5, 11, 24, 42, 52). This was of significance becausecontrol of phosphorylation of cellular proteins plays a criticalrole in cell growth (4). It is known that a variety of tyrosinekinases, including transmembrane receptors such as the epi-dermal growth factor and insulin receptors, membrane-associ-ated kinases of the Src family, and more recently tyrosine-threonine kinases like mitogen-activated protein (MAP)kinase kinase, have specific intracellular substrates in the signaltransduction pathways that regulate cell growth and metabo-lism. Initially, it was felt that PTPases played housekeepingroles, but it is now known that they have specific intracellulartargets and function in cellular growth control (5, 11, 24, 42,52). Although H35 cells have many liver-like properties, theyare tumorigenic. We wondered if PRL-1 is important fornormal cell growth and if PRL-1 overexpression in H35 cellscould contribute to their tumorigenicity. Here we demonstratethat PRL-1 is a PTPase and that it is highly expressed ingrowing hepatic cells and some tumor cell lines. PRL-1 islocated primarily in the cell nucleus, and its overexpressionmodulates the growth of cells.

MATERIALS AND METHODS

Cell culture and animals. H35 cells were grown as describedpreviously (50). Following serum deprivation, cells weretreated with various agents (insulin [10-8 M; Sigma], fetalbovine serum [FBS; 20%], and cycloheximide [10 ,ug/ml;Sigma]) as described previously (51). BALB/3T3 cells weregrown and induced as described previously (36). Cells wereharvested by scraping in 4 M guanidine thiocyanate buffer, and

3752

PRL-1, A NUCLEAR PROTEIN TYROSINE PHOSPHATASE 3753

A.

3.

E.':, . ,zil o

ff 2.5-T1 3T3

0 .5 1 2 4 8 16 h post-insulin

H35 Cells

C.

h post-hepatectomy

Regenerating LiverFIG. 1. PRL-1 is an immediate-early gene in regenerating liver and

serum-stimulated fibroblasts but is constitutively expressed in H35cells. (A) Northern blot showing expression of PRL-1 mRNA beforeand 3 h after the application of a mitogenic stimulus in H35 cells(insulin), regenerating liver (RL) (70% hepatectomy), and BALB/c3T3 (3T3) fibroblasts (serum) in the presence of cycloheximidepretreatment. SFM, serum-free medium; 3CI, 3 h after insulin treat-ment of serum-starved H35 cells; Q, quiescent liver; 3CH, 3 h afterhepatectomy; SD, serum-deprived 3T3 cells; 3CS, 3 h after serum

treatment of quiescent 3T3 cells. Sizes are shown in kilobases. (B)Time course of expression of PRL-J in H35 cells before and at varioustime points before and after the addition of insulin to serum-starvedquiescent cells. The ethidium-stained 28S rRNA subunit from the gelis shown for normalization. (C) Time course of expression of PRL-1 inregenerating liver at various time points before and after a 70%hepatectomy. The ethidium-stained 28S rRNA subunit from the gel isshown for normalization.

RNA was prepared as described previously (43). For regener-ating liver, Fischer rats (160 to 200 g; Bantin-Kingman) wereether anesthetized and subjected to midventral laparotomywith approximately 70% liver resection (left lateral and medianlobes) (20). For cycloheximide-treated samples, rats were

treated with cycloheximide (50 mg/kg; 5% solution in phos-phate-buffered saline [PBS]) intraperitoneally 15 min prior tosurgery as described elsewhere (2).

Northern (RNA) and Southern blots. RNA preparation,Northern blot analyses, and labeling of recombinant PRL-1plasmids have been described elsewhere (36). Hybridizationbuffer consisted of 10% dextran sulfate, 40% formamide, 0.6 MNaCl, 0.06 M sodium citrate, 7 mM Tris (pH 7.6), 0.8xDenhardt's solution, and 0.002% heat-denatured, sonicatedsalmon sperm DNA. Northern blots were hybridized at 42°Cfor 16 h and washed twice for 30 min each time at 60°C in 0.015M NaCl-0.0015 M sodium citrate-0.1% sodium dodecyl sul-fate (SDS) prior to exposure to film (43). Genomic DNA wasextracted from normal rat liver, H35 cells, and human HeLacells (43), and Southern transfers were performed. Genomic

TABLE 1. Expression of PRL-1 mRNA in tissues and cells

Tissue or cell line Expressiona

TissuesBrain ............................................ +++Intestine.Spleen.............................................+/-Heart.Kidney.............................................+/-Thymus.............................................+/-Lung.Muscle ............................................ +++Stomach.Liver.............................................+/-

Cell linesHeLa ............................................ ++++HepG2 ............................................ +++Cvi ............................................ +++Burkitt lymphoma ................ ............................ + +Lymphoblast .............................................+/-Fibroblast .............................................+/-a Relative to peak level in regenerating liver (+ + + +) or normal liver (+/-),

based on Northern blots of 10 ,ug of total RNA hybridized with a PRL-1 probe.Tissues were from normal rats, and all cell lines were actively proliferating.

blots were washed twice for 30 min each time at 60°C in 0.0075M NaCl-0.00075 M sodium citrate-0.1% SDS prior to expo-sure to film.

Purification of bacterially expressed PRL-1 and phos-phatase assays. A 543-bp insert encompassing the entirePRL-1 open reading frame was amplified by PCR, and theinsert flanked by XhoI-BamHI sites was ligated into the PETvector (39) and then transformed into BL-21 cells. Log-phasecultures were induced for 2 h with 1 mM isopropylthiogalac-topyranoside (IPTG), and the cells were pelleted and resus-pended in a total volume of 10 ml of 50 mM Tris (pH 7.6)-100mM NaCl. The cells were then lysed by gentle sonication, andthe soluble fraction was collected after the insoluble materialwas pelleted. The protein was purified by using a nickel-agarose column (Qiagen) and nondenaturing conditions asdescribed previously (21). The purity of the protein wasconfirmed by running a sample on an SDS-polyacrylamide geland staining with Coomassie blue. Protein concentrations weredetermined by the Lowry method. For phosphatase assays,equal volumes of corresponding column fractions were usedfor both PRL-1 and control proteins. In some cases, theconcentration of the control protein in the equivalent volumewas less than that for PRL-1. In such cases, appropriateamounts of bovine serum albumin (Sigma) were added so thatcontrols consisted of the same total amount of protein in thesame volume of the same column fraction as was used in thePRL-1 assay. The mutant protein was created by using amutant oligonucleotide (GCTGTCCATTCTGTCGCAGGCCTTGGCAGA) in a PCR. The sequences of the wild-type andmutant plasmids were confirmed.

Transient transfections. Ten micrograms of pCMV-1-PRL-1(open reading frame of PRL-1 cloned into a pCMV vector byPCR, with sequence confirmed) and control pCMV weretransfected into NIH 3T3 cells as we have described previously(22), and at 48 h after removal of coverslips for immunofluo-rescence (see below), total cell lysates were prepared (Fig. 4A).The cells were serum deprived 24 h prior to harvesting.Transfection efficiency was assessed relative to a ,-galactosi-dase control plasmid. Immunoblots were prepared and ana-lyzed with anti-PRL-1 antibodies by the enhanced chemilumi-nescence (Amersham) technique as described below. The same

VOL. 14, 1994

3754 DIAMOND ET AL.

A. CACAATCTTCAATGAGTAGACATATTCCTCAGTTCTGTGGTGTTCTCGGTCACACATTTATGGAGTTTCTGAAGGGCAGTGGAGACTACTGCCAGGCACAGCACGACCTCTATGCAGACAAGTGAACTGTAGAAATTCATTACTACTCCACCAAGAAGCCCCCATAAGAGTGGATAGCCTGGACACAGTCGTGTTGAATTGAAATCTGCAGAGCATTTTCCAAGAGCTCAGACCTGGATGGGGTAAACCTCAGTGCACTTCCTCTGTATCGCCTCAGTATTCCTGGATTGAAGAGTCACTGCTTCTTGTGAGGAGGTTCATTTCATTGCCCGTTTCTCCCGACTCATACTCAAAGCACTGAGAATTTCAAGTGGAGTATATTGAATATTGAAGTAGACTTCAGGTTGTTTTTTGGTTTTGTTTTGGTTTTTT

TGTATTCAATTTTTTAATTCTTTCATAACCCTATTGGGTGTTTTTTTAAACTAA

ATTAACATGGCTCGAATGAACCGCCCTGCTCCTGTGGAAGTCACATACAAGAAC 594MetAlaArgMetAsnArgProAlaProValGluValThrTyrLysAsn 16

ATGAGATTTCTTATTACACACAATCCAACCAATGCGACCTTAAACAAATTTATA 648MetArgPheLeuIleThrHisAsnProThrAsnAlaThrLeuAsnLysPheIle 34

GAGGAACTTAAGAAGTATGGAGTTACCACAATAGTAAGAGTATGCGAAGCAACT 702GluGluLeuLysLysTyrGlyValThrThrIleValArgValCysGluAlaThr 52

TACGACACTACTCTTGTGGAGAAAGAAGGCATTCATGTTCTTGACTGGCCTTTT 7562yrAspThrThrLeuValGluLysGluGlyIleHisValLeuAspTrpProPhe 70

GATGATGGTGCACCACCATCCAACCAGATTGTTGATGACTGGTTAAGTCTTGTG 810AspAspGlyAlaProProSerAsnGlnIleValAspAspTrpLeuSerLeuVal 88

AAGATTAAGTTTCGTGAAGAACCTGGTTGCTGTATTGCTGTCCATTGTGTCGCA 864LysIleLysPheArgGluGluProGlyCysCysIleAlaValHisCysValAla 106

GGCCTTGGCAGAGCTCCGGTGCTTGTTGCCCTAGCATTAATTGAAGGTGGAATG 918GlyLouGlyArgAlaProValLeuValAlaLeuAlaLeuIleGluGlyGlyMet 124

AAATATGAAGATGCAGTACAATTCATAAGACAAAAGCGGCGCGGAGCTTTTAAC 972LysTyrGluAspAlaValGlnPheIleArgGlnLysArgArgGlyAlaPheAsn 142

AGCAAGCAACTTCTGTACCTGGAGAAGTACCGTCCTAAAATGCGGCTCCGCTTC 1026SerLysGlnLeuLeuTyrLeuGluLysTyrArgProLysMetArgLeuArgPhe 160

AAGGATTCCAACGGTCATAGAAACAACTGTTGTATTCAATAAAACTGGGGTGCC 1080LysaAspSerAsnGlyHisArgAsnAsnCysCysI1eGln 173

TGATGCCATTGCCTTGGAAGAGGAACTTCAGATGGGACCTGATTTGTTATTTAC 1134CCAATGTGTCCACTTACCTGTGGAAGCTCCAGGGGAATATTGAAAAAGTTTTAC 1188CAGGCCACAAGCTTGACAGAATTGCAACCTCTATAATTGGGCTATGATCAACAC 1242GTTTGGACACTTAGCAAAAGATTTTTGCTGGTCAGCATTTAAAATGTGCTTATT 1296ATTTGTACCAATTGACCTTTCCTAAAATAAGGTATTGAGTAATGTCATTAAATG 1350TACTCCTGTGCCAGAATATTATTAGTCTATAAGGAATTTAGAAGGATTAGGTGC 1404CAAAATACCCAGCACAATACTTGTATATTTTTAGCATCATACAGAACCAAAATT 1458CCAAGAACTAAGAACTCTCCAGACCTTCCATGGTGTATTCCTTCAGTCATTTCA 1512AACACCGCAGGGCTTCTCTTGTTATCTGCCTGCTCACTCTATGTTTACATCTCC 1566CACACTTACACCAGAACACATCAGGTTTGCTTAGCTATCTTTTAAGTCTTGCAA 1620TGATTATTTAATGTCTCTGTCTTATTTTGTGCTGTTTTGGGAAACCTCCATTTG 1674AAAATCAACTTTGTTACAGAAGCACATATCTTCAATAATGTCTCCAGACAAAAA 1728GCCTTATAGTTAATTTAATGTTTGCACTCGGGTGCAACCTGACAGGGAGGGCCT 1782GAACAAGAAAGGAGAGGAGGCTATTAAATATTTTTAGTAATATGTTGCCTTTGT 1836CTTGTGCAGAACATGTAGAGTATGCTCTTTAATTTAGTAAATATTTTTAAGACG 1890TAGAGATACATTGTTGTAGCTAACCACTTAATCAAAATTTCTGAAATTCTTGTG 1944TTTTCCATACCTATCTGAGGTTTTCCAACTTGTTTGAATTATGGTTTTCCCCTT 1998CTCTTCCCAATCTCTTGCAAAAAAGTAAAAGTGGGATCTGCTAGTGAACTGAGC 2052AGAAATATTTTATACGCCTTTTGAGCTATGTAACTTAATAATTGGATACTTGAT 2106CATTTGTTTTATTATGTAATCGATAAAATGGTGATGTGTATTAATGTTAGTTCA 2160ACCATATATTTATACTGTCTGGGAATGTGTGGTTATAGTTCTGTGGGAGAAATA 2214GTTTGTCAGTGTTCACCAGCTTGTAAAAACTTAGTGCGAGAGCTTCAACATCTA 2268AATAAATGATGAAACGCATTCGTCACTGAGGTCACTTTGCTTAAAATTAACTTA 2322ATTTGTAGAAAACAGTGGATTCAATTATTATCATTTCAGTTTATGGACAAATTT 2376GTTAGGGTTACCAAGTGCGTTTAAAAATTGCTCTTTAAAGGTCTAGATAATTGT 2430GAATCAATTGAATGTTGGGTACCAAGGGAAAACGGTTTGTAATAGTTGATGACC 2484TTGATTTTTAATTCAATTCCACCAGTCACTTGTAGCTTTATGCAGTTTCCAATC 2538CACTTTTCTCATTTTTAAGTTTATTACTTACCTGTATATATTTTGAAATTAATT 2592TGAACCTGCGTATTTGGCACATGATGGCTTATAAATTTTAACTTTC 2642

B. consensus VHCXAGXXR

PRL-1 CCIAVHCVAGLGRAP

hPTP hpItVHCsAGaGRtg

VVH1 p v 1 V H C a A G v n R a g

Hscdc25 i i I v f H C e f S a e R g P

FIG. 2. (A) Sequence of the PRL-1 cDNA. The amino acid se-

quence comprising the PTPase active site consensus sequence is shownin boldface, the basic region is underlined, and the potential tyrosinephosphorylation sites are shown in boldface italic. A full-length cDNAwas obtained from the regenerating liver cDNA library (36). Thesequence of the clone was determined by the Sanger dideoxy sequenc-ing method (44) and was confirmed through bidirectional sequencing.(B) Alignment of PRL-1 amino acid sequence with sequences of theactive sites of the three known classes of PTPases (5, 11, 13-15, 24, 34,42, 52). hPTP, human PTPase; VVH1, vaccinia virus H1; Hscdc25,human sequence for cdc25. (C) Genomic Southern blot hybridizedwith the full-length PRL-1 cDNA probe.

transfected NIH 3T3 cells grown on glass coverslips were

stained with anti-PRL-1 antibodies for immunofluorescence.Antibodies, immunofluorescence, and immunoblots. Poly-

clonal anti-PRL-1 antibodies were prepared by Cocalico Bio-logicals, Reamstown, Pa., against bacterially expressed purifieddenatured PRL-1 protein described in the legend to Fig. 3.Immunoblotting was performed as we have described previ-ously (51), using a 1:1,000 dilution of anti-PRL-1 (ot-PRL-1)antisera followed by a 1:10,000 dilution of goat anti-rabbit seraand the chemiluminescence detection system (Amersham).Immunofluorescence was performed essentially as describedelsewhere (18). Cells were grown to 80% confluence on glasscoverslips and serum deprived for 2 days in 1% FBS-Dulbeccomodified Eagle medium (Fig. 4B and 6) or treated as describedabove for transient transfections. All coverslips were washedwith PBS, fixed and permeabilized with 50% methanol-50%acetone, and then treated with affinity-purified anti-PRL-1antibodies at 4 ng/pl, preimmune sera, or anti-PRL-1-depleted(a-PRL-1-depleted) antisera at an equivalent immunoglobulinconcentration (dilutions were made with PBS-5% fetal calfserum-0.01% sodium azide). ot-PRL-1-depleted antisera was

prepared from the flowthrough of the PRL-1 affinity columnand contained very low residual anti-bacterial PRL-1 activity.Following incubation, cells were rinsed in PBS and antibodybuffer and covered for 30 min with rhodamine-conjugated goatanti-rabbit immunoglobulin G diluted 1:100 in antibody buffer.Coverslips were rinsed in PBS, dried, and mounted on glassslides, using 50% PBS-50% glycerol-0.25% phenylenediaminedihydrochloride. Fluorescence was examined in a Nikon Mi-crophot-FXA fluorescence microscope.

Cell extracts. For cellular extracts (Fig. 5), growing cellswere treated as described previously (54) by sonication inlow-salt buffer to prepare the soluble fraction, resuspension ofthe 100,000 x g pellet in 1% Triton to prepare the Triton-soluble fraction, and resuspension of Triton-insoluble materialin loading dye containing 4% SDS and 10% P-mercaptoetha-nol. Liver was extracted and fractionated as described previ-ously (27, 51). Nucleus preparation was verified by microscopy,as we have done previously (22, 51), and was 99% pure.

54108162216270324378432486540

C.

MOL. CELL. BIOL.

PRL-1, A NUCLEAR PROTEIN TYROSINE PHOSPHATASE 3755

aIa.0

1.A. B. C.

0.8 0.4-0.15

0.6 0.3-0.1.

0A4 0.2,

oControl 0.05,* .

0.2.

p pR== --

100 200 300 400 500

Protein ttghni20 40 60

tme (min)2 4 6 8 10

PNPP (mM)

0.3-

0

8 0.2-£A00

< 0.1-

0+ Active + PRL-1/PRL-1 Vanadate

E.

Control PRL-1(C104S)

FIG. 3. PTPase activity of purified PRL-1. (A) Hydrolysis of PNPP as a function of protein concentration. (B) Hydrolysis of PNPP as a functionof time. In this experiment, 100 jig of PRL-1 per ml and 20 mM PNPP were used. (C) Hydrolysis of PNPP as a function of varying substrateconcentrations. For this experiment, 200 jLg of PRL-1 per ml was used. Hydrolysis of PNPP was done in 300 ,lI of solution containing 50 mMN-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES; pH 7.5), 0.1% ,B-mercaptoethanol, and the indicated concentrations of PNPP andenzyme at 37°C for 60 min. The reaction was stopped by the addition of 6 ,ul of 10 N NaOH, and A410 was measured. PRL-1 protein and controlprotein IKBa (51) were expressed as fusion proteins linked to six histidines in E. coli, using the PET system (39). (D) Hydrolysis oftyrosine-phosphorylated PRL-1 by PRL-1. c-Src isolated by immunoprecipitation from platelets (23) was added to 2.5 ,ug of inactive PRL-1 (no3-mercaptoethanol) and 20 ,uCi of [y-32P]ATP in 30 p.l of kinase buffer at room temperature for 15 min (23), after which the reaction was stopped

by the addition of EDTA (12 mM). The reaction was then split into equal samples. Tube 1 served as the control; 15 plI of 2x phosphatase buffer(200 mM Tris [pH 7.6], 0.2 M 3-mercaptoethanol) and 2.5 ,ug of active PRL-1 were added to tube 2 (+Active PRL-1), and 3.10 mM sodiumorthovanadate was added to tube 3 (+PRL-1/Vanadate). The tubes were then incubated at 37°C for 1 h, and reactions were stopped by the additionof 30 pul of loading dye. Samples containing labeled Src were incubated with phosphatase buffer and control bacterial protein and showed nodephosphorylation. Samples were fractionated on an SDS-polyacrylamide gel, which was then dried and exposed to X-ray film. The results shownare averages of three separate experiments and based on the level of 32P in the Src and PRL-1 phosphoproteins as determined by densitometry.(E) The relative activities of control (IKBa) protein, PRL-1 mutant protein C104S, and wild-type PRL-1 (215 ,ug of each protein per ml) weremeasured by using the PNPP assay as for panel A. In this experiment, the protein was stored at -20°C in a modified storage buffer containing 20mM Tris (pH 7.9), 100 mM KCl, 5 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride, 80 mM imidazole, 50% glycerol, 2.5 mM EDTA, and200 ,ug of acetylated bovine serum albumin per ml.

Cytoplasmic material was further fractionated by a high-speed(100,000 x g) spin, followed by resuspension of the pellet in0.5% Triton and resuspension of the final Triton-insolublepellet in loading buffer. Nuclear material was further fraction-ated by treatment in 0.5% Triton to remove the outer nuclearmembrane from intact nuclei, lysis of nuclei by Douncehomogenization, and recovery of nuclear extract. The nuclearextract was subjected to high-speed centrifugation, the pelletwas resuspended in 0.5% Triton to recover the Triton-solublefraction, the Triton-insoluble pellet was resuspended in 1.5 MKCl to recover the salt-extractable material, and the remaininginsoluble material was resuspended in loading buffer. Thenuclear pellet and KCl fraction represented about 3% of thetotal cellular protein.

Stable pCMV-PRL-1 cell lines. pCMV-PRL-1 and pCMVstable transfectants were made with 10 ,ug ofDNA plus 1 ,ug of

neomycin control plasmid, using standard methods (54). Thetransfection efficiencies for pCMV control and pCMV-PRL-1were 10-3 and 10-4, respectively, and approximately 750 totalpCMV Neo+ and 75 pCMV-PRL-1 Neo+ clones were ob-tained in two separate transfection experiments. SpecificpCMV-PRL-1 clones were selected for further study on thebasis of detecting a relatively high level of PRL-1 protein byimmunofluorescence of very early colonies. Immunofluores-cence was performed simultaneously in all cell lines withidentical reagents, exactly as described for Fig. 6. For conflu-ence studies, cells were grown in Dulbecco modified Eaglemedium-10% FBS to confluence, and the number of cells wasdetermined. Cells were fed regularly and allowed to growpostconfluence. Soft agar cloning was performed as describedpreviously (54) after seeding of duplicate plates with 5 x 104cells. Soft agar colonies were scored relative to negative [NIH

100

0e'Ea

IL

CMco

VOL. 14, 1994

3756 DIAMOND ET AL.

A.

-43

-29

PRL-1--

-18

*.14

BAffinity purified ct-PRL- 1 Pre-immune x-PRL-1 depleted sera

I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~q

N~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~A~~6jA V

.1~~~~~~~~~~~~~~~~~~~j~~~~~~~~~~~~~1.~~~~~~~~~~~~~~~~~~~.~~~~~~~~~~~~~~~~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ .~a/~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~4

ofpMVPLJ-rnfetdcel roe.wt.nt-R-1atsea.. adb r hefurecn adpas hoorph,rspciel.Peimnanisrashwe n secfi saiin, ndth cntol(pMV sowd lw evl f ucea ad ytplsmc tanig no sow).(BImmnolurececeofserm-epivd talytrasfctd CM-PL.(.D.).cells

MOL. CELL. BIOL.

PRL-1, A NUCLEAR PROTEIN TYROSINE PHOSPHATASE 3757

29-

PRL-1 -_-

18-

H35 Cells

..iii

1 2 3

Normal Liver

5h Posthepatectomy

4 5 6 7 8 9 10 11

-*-PRL-1

FIG. 5. Growing H35 cells (lanes 1 to 3), normal liver (right upper panel), and regenerating liver (5 h posthepatectomy; right lower panel) wereanalyzed by immunoblotting with anti-PRL-1 antibodies. Lane 1, soluble protein fraction; lane 2, Triton-soluble fraction; lane 3, Triton-insolublepellet resuspended in loading dye; lane 4, soluble cytoplasmic extract; lane 5, Triton-soluble cytoplasmic material; lane 6, Triton-insolublecytoplasmic pellet in loading buffer; lane 7, outer nuclear membrane; lane 8, nuclear extract; lane 9, Triton-soluble nuclear material; lane 10, 1.5M KCl-extracted material from Triton-insoluble nuclear pellet; lane 11, residual nuclear pellet in loading buffer. For lanes 4 to 11, the positionof PRL-1 is indicated relative to the migration of a PET-PRL-1 protein standard. Thirty micrograms of protein was loaded per lane and verifiedby Ponceau staining of the membrane. Nonspecific sticking of PRL-1 antibodies to Ponceau-positive bands in some lanes was observed in bothnormal and regenerating liver fractions.

3T3; pCMV(AD6)] and positive (v-Src-transformed 3T3) con-trol cells.

Nucleotide sequence accession number. The GenBank ac-cession number for the sequence shown in Fig. 2A is L27843.

RESULTS

Identification of PRL-1. As with many immediate-earlygenes, PRL-1 mRNA is generally induced in mitogen-treatedcells, including regenerating liver following partial hepatec-tomy and 3T3 cell fibroblasts following serum stimulation (Fig.1). PRL-1 is an immediate-early gene, as indicated by itsinduction in the presence of a cycloheximide-mediated proteinsynthesis blockade (Fig. 1A). In H35 rat hepatoma cells,however, PRL-1 is constitutively expressed. Even under serum-starved quiescent conditions, PRL-1 is expressed in H35 cellsat a level comparable to that seen at various time points afterthe addition of insulin (Fig. iB), and there is no PRL-1superinduction with cycloheximide pretreatment (Fig. 1A).Unlike expression of some immediate-early genes, PRL-1expression is not limited to the GI phase of the cell cycle, asPRL-1 mRNA remains elevated throughout the growth re-sponse (up to 72 h posthepatectomy) (Fig. 1C). PRL-1 expres-sion is elevated in both parenchymal and nonparenchymal cellsin the regenerating liver (17), and it is expressed throughoutthe course of hepatic development (unpublished data). PRL-1is expressed in brain and skeletal muscle but not in a numberof other tissues. High levels ofPRL-1 mRNA are also found inseveral tumor cell lines, but the level of this mRNA is relativelylow in proliferating, nontumorigenic cells (e.g., fibroblasts andlymphoblasts) (Table 1).The full-length PRL-1 cDNA clone corresponds in size to

the 2.7-kb mRNA (Fig. 1A) and contains an additional poly-adenylation signal at position 2269 (Fig. 2A). This more 5'polyadenylation signal accounts for the 2.5-kb mRNA and wasused in one of the cDNA clones that we isolated. The 519-bpopen reading frame encodes a 173-amino-acid protein with aprojected molecular mass of 19.8 kDa. The open reading frameis flanked by a 546-bp 5' untranslated region and a 1.8-kb 3'untranslated region that contains several ATlTA sequenceswhich could decrease mRNA stability (45). We confirmed that

the long 5' untranslated region is part of the PRL-1 mRNA byusing a probe from this region in a Northern blot analysis (notshown). There are multiple stop codons in each frame bothbefore and after the open reading frame. The initial startcodon is in a good Kozak consensus sequence site, with both anA at position -3 and a G at position +4 (26). Presumablybecause of the long 5' untranslated region, synthetic PRL-1mRNA translates weakly but detectably in rabbit reticulocytelysate, migrating as a 20-kDa protein (not shown).While PRL-1 is a novel gene, we noted that the deduced

amino acid sequence contains the VHCXAGXXR signaturesequence for the active site of PTPases (Fig. 2B) (5, 11, 24, 42,52). The carboxyl terminus of PRL-1 contains a few aminoacids, including Val-130, Arg-134, Gln-145, and Leu-146, thatare found at similar positions in other classes of PTPases. Thebalance of the amino acid sequence does not have homology toany previously described PTPases, including 3CH134 (6)(RL-30 in our regenerating liver study [36]), a VH1 homologthat is also induced in the growth response. PRL-1 is a basicprotein (pl 9.53) with a highly basic region toward the carboxylterminus. It has two potential tyrosine phosphorylation sitesnear the amino terminus of the protein (Fig. 2A) and weakconsensus sites for protein kinase C and casein kinase II.Hybridization of a PRL-1 probe to genomic Southern blotsreveals the presence of multiple bands, even after washingunder stringent conditions at high temperature (Fig. 2C).These bands are highly conserved between rat, H35 cell, andhuman (HeLa cell) DNAs. Both full-length and coding-regionPRL-1 probes hybridize to multiple genomic fragments, sug-gesting that PRL-1 is a member of a new gene family and/orhybridizes to multiple pseudogenes.

Demonstration of PRL-1's PTPase activity. To demonstratePRL-l's PTPase activity, we expressed the protein as a histi-dine fusion protein in Escherichia coli, purified it over a nickelchelate affinity column, and used it in a p-nitrophenyl phos-phate (PNPP) assay (8, 15). We found that PRL-1 has low butdefinite activity in this system (Fig. 3A) compared with acontrol recombinant protein isolated from the correspondingfraction from an affinity column. Phosphatase activity is linearwith respect to enzyme concentration (Fig. 3A) and time (Fig.3B) and obeys first-order kinetics when the concentration of

VOL. 14, 1994

3758 DIAMOND ET AL.

pCMV(AD6)Control

pCMV-PRL- 1(CD 1)

pCMV-PRL- 1(BA 1)

pCMV-PRL- 1(BD5)

pC MV - PR L -- 1

(BA2)

FIG. 6. PRL-1 levels are elevated in stably transfected pCMV-PRL-1 cells. Immunofluorescence analysis of serum-deprived (1% FBSfor 2 days) pCMV and pCMV-PRL-1 stable cell lines, using affinity-purified anti-PRL-1 antisera, was performed simultaneously withidentical reagents. Photographic exposures times were identical for allcells. Magnification, x312.5.

substrate is varied (Fig. 3C), with a predicted Vm,, of 2.2nm/min/mg under these conditions. This activity is lower thanthose of other PTPases and could be accounted for by im-proper folding of the bacterial protein or absence of an

important regulatory subunit. We also found that a reducingagent such as dithiothreitol or ,-mercaptoethanol is requiredfor PRL-1 phosphatase activity (data not shown), the enzyme

activity is heat sensitive, and the pH optimum is 7.3 to 7.5.PRL-1 activity is inhibited by sodium orthovanadate (60% at 1mM). Unlike the control phosphatase (potato acid phos-phatase), PRL-1 was unable to remove phosphate residuesfrom a serine-threonine-phosphorylated substrate (casein)

(not shown). Because we found tyrosine phosphorylation sitesin PRL-1, we tested whether immunoprecipitated Src couldphosphorylate inactive PRL-1 and found that PRL-1 wasreadily phosphorylated in vitro along with the Src kinase itself.We then arrested the kinase reaction with EDTA and addedactive PRL-1. In three separate experiments, active PRL-1dephosphorylated tyrosine-phosphorylated PRL-1 morereadily than it dephosphorylated Src (up to 95% versus 40%),and this dephosphorylation was inhibited by orthovanadate(Fig. 3D). Other phosphotyrosine substrates that were tested,including cdc2 and angiotensin (8) peptides, were readilydephosphorylated by control phosphatase but not by PRL-1.Finally, to show that the phosphatase activity of PRL-1 isdependent on the cysteine residue present in the active site, wetested the activity of wild-type PRL-1 and C104S PRL-1 in thePNPP assay (Fig. 3E). As with other PTPases, the phosphataseactivity of PRL-1 is completely abolished by the cysteinemutation (16).

Subcellular localization of PRL-1. To begin to assess thecellular function of PRL-1, we initiated studies to determine itsintracellular location. As the normal cellular level of PRL-1 islow and was not readily detectable, we overexpressed PRL-1 inNIH 3T3 cells by using transient transfection. Immunoblottingof serum-deprived pCMV-PRL-1-transfected cells with anti-PRL-1 antibodies revealed a 21-kDa protein not present incontrol cells (Fig. 4A). Immunofluorescence of these samecells demonstrated grossly positive cells of slightly abnormalmorphology that were not present in transfection controls.Staining was present in both the nucleus and cytoplasm of cells,whereas background cells showed only a small amount ofnuclear staining. To further document the integrity of theanti-PRL-1 antibodies and demonstrate subcellular localiza-tion of PRL-1 in cells that overexpressed the protein less thantransiently transfected cells, we performed further immunoflu-orescence studies. Immunofluorescence of H35 cells (notshown) or pCMV-PRL-1 stably transfected NIH 3T3 cells withaffinity-purified PRL-1 antibodies demonstrated staining inboth the nucleus and cytoplasm, but staining was predomi-nantly nuclear (Fig. 4B). This staining was not detected bypreimmune sera or ot-PRL-1-depleted sera (Fig. 4B).We used immunoblots to further substantiate the cellular

localization of PRL-1. As H35 cells express high levels ofPRL-1 mRNA constitutively, we performed immunoblots ingrowing H35 cells. As we could not detect PRL-1 in total cellextracts, we fractionated the extracts into soluble, membrane,and Triton-insoluble fractions. We clearly detected PRL-1 inthe insoluble fraction (Fig. 5, lane 3), although there was somePRL-1 in the soluble fraction as well in other experiments.PRL-l's presence in the insoluble fraction was surprising sinceit is readily soluble in E. coli. We reasoned that in cells, PRL-1must be associated with other proteins that render it insoluble.In the presence of SDS without reducing agents, PRL-1 stillmigrated at 21 kDa, indicating that these interactions do notdepend on disulfide bonds. In the presence of high concentra-tions of salt (1.5 M KCl), PRL-1 was shifted into the solublefraction, whereas intermediate filament proteins (vimentin)remained insoluble (not shown), suggesting that interactionsbetween PRL-1 and other proteins can be disrupted. We thenperformed further nuclear and cytoplasmic fractionation stud-ies in normal and regenerating liver (Fig. 5, lanes 4 to 11). A21-kDa protein was detected primarily in the Triton-insolublenuclear fraction of the regenerating liver. As in H35 cells,PRL-1 was solubilized in high salt.

Effect of PRL-1 on cell growth. Given PRL-l's inductionduring the GO/G1 transition in normal cells, its constitutiveexpression in H35 cells, and its high level of expression in a

MOL. CELL. BIOL.

PRL-1, A NUCLEAR PROTEIN TYROSINE PHOSPHATASE 3759

A.

Days

FIG. 7. Some pCMV-PRL-I cells show grossly abnormal morphology, and viable clones show enhanced growth rate. (A) Photograph of apCMV-PRL-1 cell. (B) Growth curves of two pCMV-PRL-1 lines (BD5 and CD1) and the control (AD6). Growth rates for cell lines weredetermined by measuring the number of cells from duplicate 35-mm-diameter dishes on consecutive days after plating at a low cell number.

number of tumor cell lines, we hypothesized that PRL-1 couldbe a modulator of cell growth. To test this theory, we stablytransfected pCMV-PRL-1 and a neomycin gene into NIH 3T3cells, and scored for Neo+ colonies. Notably, in two separatetransfection experiments, we scored only 10% as many pCMV-PRL-1-positive clones (approximately 75 total colonies) aspCMV control clones (approximately 750 total colonies). Thisfinding suggested that very high levels of PRL-1 could be toxicto cells. pCMV-PRL-1 clones were selected for further studyon the basis of the high level of PRL-1 protein determined byimmunofluorescence of cells during the initial passage. Subse-quent studies were performed in early-passage cells. In serum-deprived cells, by immunoblotting (not shown) and immuno-fluorescence (Fig. 6), PRL-1 levels were three- to fivefoldhigher in pCMV-PRL-1 cells than in pCMV controls. Byimmunofluorescence, in all cells, PRL-1 was primarily nuclearand showed a speckled pattern. Both immunofluorescence andimmunoblotting produced patterns similar to those for H35and NIH 3T3 cells, suggesting that in stably transfected cells,PRL-1 compartmentalizes properly. The majority of coloniesin the pCMV-PRL-1 transfection grew up rapidly with no delayas compared with control cells. However, a few of the pCMV-PRL-1 colonies contained cells of grossly abnormal morphol-ogy, including multinucleated giant cells (Fig. 7A). Most ofthese grossly abnormal clones did not propagate sufficiently tobe selected. All pCMV-PRL-1 clones that were studied hadrapid growth rates (maximum doubling time of approximately20 h, compared with 24 h for the control) (Fig. 7B).

Several of the pCMV-PRL-1 lines demonstrated unusualmorphology and enhanced saturation density (Fig. 8). Thenumber of cells was measured out to 6 days postconfluence todetermine saturation density, which was >105/cm2 for allpCMV-PRL-1 cells and 2 x 104/cm2 for pCMV(AD6) (con-trol). The lines that were the most altered morphologically(CD1 and BA1) or had the highest growth rate (BD5) formedcolonies in soft agar at a rate comparable to that for v-Src-transfected cells, while BA2 gave greatly increased coloniesrelative to control cells. v-Src-transfected cells were muchmore able to form tumors in nude mice than the pCMV-PRL-1lines that were tested. It is possible that the high level of PRL-1in the BA2 cells was relatively toxic and contributed to thelower rate of soft agar cloning.

DISCUSSION

Prior to this study, three distinct families of PTPases hadbeen identified, including the large family characterized byPTPa (5, 11, 12, 24, 42, 46, 52), the vaccinia virus phosphatasefamily (6, 15, 38), and the cdc25 family (8, 13, 14, 34). Amongmembers of each of these families, there is significant homol-ogy in regions outside the active site. As judged from its lack ofhomology to any of these families and the establishment of itsenzymatic activity, PRL-1 is a novel type of PTPase. Thus far,PRL-1 shows most activity as an autophosphatase, but its trueintracellular substrates are not yet identified. Although studiesto date have been negative, it is not clear whether in someinstances PRL-1 could have serine-threonine phosphataseactivity as well. PRL-1 is a particularly small enzyme, and it islikely that it is associated in the cell with one or moreregulatory subunits that could enhance its phosphatase activity.The identification of regulatory subunits and/or specific sub-strates will be important in delineating its specific activity.

Since it is clear that PTPases and tyrosine kinases playcentral roles in the regulation of cell growth and division (5, 11,24, 25, 42, 52), it is intriguing that we have identified a novelnuclear PTPase induced in the G1 phase of the growthresponse that alters cell growth. In this regard, we and othershave identified members of the vaccinia virus PTPase familythat are also induced as part of the growth response in cells andapparently show nuclear localization (6, 38). Recently, amember of this family has been shown to have a negative effecton cell growth (37). Other phosphatases have also been shownto modulate cell growth. Some PTPases may actually serve toincrease the level of phosphorylation within the cell by dephos-phorylating specific regulatory sites on kinases, thus activatingthem. In this manner, it is possible to envision a significantpositive role for PTPases in the growth response. For example,it has been shown that overexpression of PTPat (54), a mem-brane-associated PTPase, results in cellular transformationthrough activation of c-Src. Another example of this type ofgrowth enhancement is seen in the cdc2-cdc25 system, in whichthe cell cycle regulatory kinase cdc2 is specifically activatedfollowing dephosphorylation by the PTPase cdc25 (25). It isknown that cdc25 PTPase plays a crucial role in governing theG2-M cell cycle transition in yeast cells and probably in

VOL. 14, 1994

3760 DIAMOND ET AL.

ie k~~#.%v O

FIG. 8. pCMV-PRL-1 cells show altered saturation density and growth in soft agar. Photographs of confluent cultures, 4-day postconfluentcultures, and soft agar clones (magnification, X25) of pCMV-PRL-1 and pCMV(AD6) (control) cell lines at 3 weeks are shown.

mammalian cells, in which a number of cdc25 homologs have level of transformation induced by PRL-1 is more modest thanbeen identified (8, 13, 14, 34). that seen with v-Src or even PTPot, which appears to actPRL-1, whose expression is elevated in several tumors, is through activation of c-Src. Consistent with this, H35 cells

able to alter the growth of 3T3 cells when overexpressed. The which constitutively express PRL-1 also display a relatively low

MOL. CELL. BIOL.

PRL-1, A NUCLEAR PROTEIN TYROSINE PHOSPHATASE 3761

level of tumorigenicity and do not grow in the absence ofserum. Some pCMV-PRL-1 cells showed grossly abnormalmorphology and did not propagate, and immunofluorescenceof some of these nascent clones indicated that levels of PRL-1were high. It appears that very high levels of PRL-1 negativelyaffect cell growth. However, it is not known what constitutesphysiologic levels of PRL-1. It is not clear whether PRL-1 canbe a physiologic inhibitor under certain cellular conditions orwhether the high levels of PRL-1 achieved in these studiesresulted in toxicity due to a loss of substrate specificity. Certainproteins like c-Myc may cause either transformation or celldeath under specific physiologic cellular conditions (9).

Unlike other PTPases which are associated with soluble andmembrane fractions, the majority of PRL-1 is found in the cellnucleus. However, PRL-1 itself is a small soluble protein thatcould diffuse through nuclear pores, and it is not surprisingthat PRL-1 can distribute between nuclear and cytoplasmicfractions, particularly when highly overexpressed. Interest-ingly, like PRL-1, two proteins involved in cell growth, PP1, aserine-threonine phosphatase, and RB, a protein regulated byserine-threonine phosphorylation, are associated in part withinsoluble portions of the cell nucleus such as chromatin or thenuclear envelope. The association of RB with insoluble pro-teins may be regulated during the cell cycle (1, 27, 35).The identification of PRL-l's potential substrates will be of

importance in defining its cellular function. At this time, thereare three known classes of nuclear tyrosine-phosphorylatedproteins that are important in cell growth: MAP kinases (4,30), cdc2 family proteins (32), and some transcription factors(28, 40, 41, 47-49). The activity of these tyrosine-phosphory-lated proteins may be precisely regulated by specific PTPasesthat are themselves nuclear. MAP kinase is activated by boththreonine and tyrosine phosphorylation and can then move tothe cell nucleus, where it phosphorylates substrates such asspecific transcription factors and histone H3, thereby modu-lating their activity. Recently, it has been shown that a nuclearPTPase in the vaccinia virus family can efficiently dephosphor-ylate MAP kinase, resulting in its inactivation (7, 53). It is nowclear that there are several cdc2-like molecules (cdks) that arealso regulated by tyrosine phosphorylation (32). Some cdksand members of the cyclin family are expressed specifically inthe GI phase of the mammalian cell cycle and thus may beimportant in controlling transition through GI (31). Althoughcurrent evidence suggests that cdc2 family proteins are regu-lated by cdc25-type PTPases, other PTPases could be involvedin their regulation as well. Finally, following stimulation witheither interferons, cytokines, or growth factors, the transcrip-tion factor ISGF-3 (Stat-91) becomes phosphorylated on atyrosine residue and translocates to the nucleus, where it isable to bind to and regulate target promoters (28, 40, 41,47-49). Linkage of PRL-1's activity to any of these nucleartyrosine-phosphorylated proteins or other as yet unknownpotential substrates will be important in elucidating PRL-1'srole in cellular growth control.

ACKNOWLEDGMENTS

We thank Tom Brautigan for helpful discussions, Victor Garsky andPaul Friedman for peptides, and Laurie Zimmerman for help withmanuscript preparation.

This work was supported in part by funds from NIH grant DK44237to R.T. R.H.D. is supported in part by NIH grant K1t DK02096, andC.S.A. is supported in part by NIH grant Kit HL02464.

REFERENCES1. Alberts, A. S., A. M. Thorburn, S. Shenolikar, M. C. Mumby, and

J. R. Feramisco. 1993. Regulation of cell cycle progression and

nuclear affinity of the retinoblastoma protein by protein phos-phatases. Proc. Natl. Acad. Sci. USA 90:388-392.

2. Almendral, J. M., D. Sommer, H. MacDonald-Bravo, J. Burck-hardt, J. Perera, and R. Bravo. 1988. Complexity of the earlygenetic response to growth factors in mouse fibroblasts. Mol. Cell.Biol. 8:2140-2148.

3. Arias, I. M. 1988. Introduction: organizational principles, p. 3-6.In I. M. Arias, W. B. Jakoby, H. Popper, D. Schachter, and D. A.Shafritz (ed.), The liver: biology and pathobiology. Raven Press,New York.

4. Blenis, J. 1993. Signal transduction via the MAP kinases: proceedat your own RSK. Proc. Natl. Acad. Sci. USA 90:5889-5892.

5. Brautigan, D. L. 1992. Great expectations: protein tyrosine phos-phatases in cell regulation. Biochim. Biophys. Acta 1114:63-77.

6. Charles, C. H., A. S. Abler, and L. F. Lau. 1992. cDNA sequenceof a growth factor-inducible immediate early gene and character-ization of its encoded protein. Oncogene 7:187-190.

7. Charles, C. H., H. Sun, L. F. Lau, and N. K. Tonks. 1993. Thegrowth factor-inducible immediate-early gene 3CH134 encodes aprotein-tyrosine-phosphatase. Proc. Natl. Acad. Sci. USA 90:5292-5296.

8. Dunphy, W. G., and A. Kumagai. 1991. The cdc25 protein containsan intrinsic phosphatase activity. Cell 67:189-196.

9. Evan, G. I., A. H. Wyllie, C. S. Gilbert, T. D. Littlewood, H. Land,M. Brooks, C. M. Waters, L. Z. Penn, and D. C. Hancock. 1992.Induction of apoptosis in fibroblasts by c-myc protein. Cell 69:119-128.

10. Fausto, N., and J. E. Mead. 1989. Regulation of liver growth:protooncogenes and transforming growth factors. Lab. Invest.60:4-13.

11. Fischer, E. H., H. Charbonneau, and N. K. Tonks. 1991. Proteintyrosine phosphatases: a diverse family in intracellular and trans-membrane enzymes. Science 253:401-406.

12. Frangioni, J. V., P. H. Beahm, V. Shifrin, C. A. Jost, and B. G.Neel. 1992. The nontransmembrane tyrosine phosphatase PTP-1Blocalizes to the endoplasmic reticulum via its 35 amino acidC-terminal sequence. Cell 68:545-560.

13. Galaktionov, K., and D. Beach. 1991. Specific activation of cdc25tyrosine phosphatases by B-type cyclins: evidence for multipleroles of mitotic cyclins. Cell 67:1181-1194.

14. Gautier, J., M. J. Solomon, R. N. Booher, J. F. Bazan, and M. W.Kirschner. 1991. cdc25 is a specific tyrosine phosphatase thatdirectly activates p34cdc2. Cell 67:197-211.

15. Guan, K., S. S. Broyles, and J. E. Dixon. 1991. A Tyr/Ser proteinphosphatase encoded by vaccinia virus. Nature (London) 350:359-362.

16. Guan, K., and J. E. Dixon. 1991. Evidence for protein-tyrosine-phosphatase catalysis proceeding via a cysteine-phosphatase inter-mediate. J. Biol. Chem. 266:17026-17030.

17. Haber, B. A., K. L. Mohn, R. H. Diamond, and R. Taub. 1993.Induction patterns of 70 genes during nine days after hepatectomydefine the temporal course of liver regeneration. J. Clin. Invest.91:1319-1326.

18. Harlow, E., and D. Lane. 1988. Antibodies: a laboratory manual.Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y.

19. Herschman, H. R. 1991. Primary response genes induced bygrowth factors and tumor promoters. Annu. Rev. Biochem. 60:281-319.

20. Higgins, G. M., and R. M. Anderson. 1931. Experimental pathol-ogy of the liver. I. Restoration of the liver of the white ratfollowing partial surgical removal. Arch. Pathol. 12:186-202.

21. Hoffman, A., and R. G. Roeder. 1991. Purification of his-taggedproteins in non-denaturing conditions suggests a convenientmethod for protein interaction studies. Nucleic Acids Res. 19:6337.

22. Hsu, J.-C., R. Bravo, and R. Taub. 1992. Interactions betweenLRF-1, JunB, c-Jun, and c-Fos define a regulatory program in theG, phase of liver regeneration. Mol. Cell. Biol. 12:4654-4665.

23. Huang, M. M., J. B. Bolen, J. W. Barnwell, S. J. Shattil, and J. S.Brugge. 1991. Membrane glycoprotein IV (CD36) is physicallyassociated with the Fyn, Lyn, and Yes protein-tyrosine kinases inhuman platelets. Proc. Natl. Acad. Sci. USA 88:7844-7848.

24. Hunter, T. 1989. Protein-tyrosine phosphatases: the other side of

VOL. 14, 1994

3762 DIAMOND ET AL.

the coin. Cell 58:1013-1016.25. Kinoshita, N., H. Ohkura, and M. Yanagida. 1990. Distinct,

essential roles of type 1 and 2A protein phosphatases in thecontrol of the fission yeast cell division cycle. Cell 63:405-415.

26. Kozak, M. 1991. Structural features in eukaryotic mRNAs thatmodulate the initiation of translation. J. Biol. Chem. 266:19867-19870.

27. Kuret, J., H. Bell, and P. Cohen. 1986. Identification of high levelsof protein phosphatase-1 in rat liver nuclei. FEBS Lett. 203:197-202.

28. Larner, A. C., M. David, G. M. Feldman, K.-I. Igarashi, R. H.Hackett, D. S. A. Webb, S. M. Sweitzer, E. F. Petricoin III, andD. S. Finbloom. 1993. Tyrosine phosphorylation of DNA bindingproteins by multiple cytokines. Science 261:1730-1733.

29. Lau, L. F., and D. Nathans. 1985. Identification of a set of genesexpressed during the GO/Gl transition of cultured mouse cells.EMBO J. 4:3145-3151.

30. Lenormand, P., C. Sardet, G. Pages, G. L'Allemain, A. Brunet, andJ. Pouyssegur. 1993. Growth factors induce nuclear translocationof MAP kinases (p42maPk and p44mapk) but not of their activatorMAP kinase kinase (p45maPkk) in fibroblasts. J. Cell Biol. 122:1079-1088.

31. Lu, X. P., K. S. Koch, D. J. Lew, V. Dulic, J. Pines, S. I. Reed, T.Hunter, and H. L. Leflert. 1992. Induction of cyclin mRNA andcyclin-associated histone Hi kinase during liver regeneration. J.Biol. Chem. 267:2841-2844.

32. Meyerson, M., G. E. Enders, C.-L. Wu, L.-K. Su, C. Gorka, C.Nelson, E. Harlow, and L.-H. Tsai. 1992. A family of humancdc2-related protein kinases. EMBO J. 11:2909-2917.

33. Michalopoulos, G. K. 1990. Liver regeneration: molecular mech-anisms of growth control. FASEB J. 4:176-187.

34. Millar, J. B. A., and P. Russell. 1992. The cdc25 M-phase inducer:an unconventional protein phosphatase. Cell 68:407-410.

35. Mittnacht, S., and R. A. Weinberg. 1991. Gl/S phosphorylation ofthe retinoblastoma protein is associated with an altered affinity forthe nuclear compartment. Cell 65:381-393.

36. Mohn, K. L., T. M. Laz, J.-C. Hsu, A. E. Melby, R. Bravo, and R.Taub. 1991. The immediate-early growth response in regeneratingliver and insulin-stimulated H-35 cells: comparison to serum-

stimulated 3T3 cells and identification of 41 novel immediate-earlygenes. Mol. Cell. Biol. 11:381-390.

37. Noguchi, T., R. Metz, L. Chen, M.-G. Mattei, D. Carrasco, and R.Bravo. 1993. Structure, mapping, and expression of erp, a growthfactor-inducible gene encoding a nontransmembrane protein ty-rosine phosphatase. Mol. Cell. Biol. 13:5195-5205.

38. Rohan, P. J., P. Davis, C. A. Moskaluk, J. M. Kearns, H. Krutzsch,U. Siebenlist, and K. Kelly. 1993. PAC-1: a mitogen-inducednuclear protein tyrosine phosphatase. Science 259:1763-1766.

39. Rosenberg, A. H., et al. 1987. Vectors for selective expression ofcloned DNAs by T7 RNA polymerase. Gene 56:125-135.

40. Ruff-Jamison, S., K. Chen, and S. Cohen. 1993. Induction by EGFand interferon-y of tyrosine phosphorylated DNA binding pro-teins in mouse liver nuclei. Science 261:1733-1736.

41. Sadowski, H. B., K. Shuai, J. E. Darnell, Jr., and M. Z. Gilman.1993. A common nuclear signal transduction pathway activated bygrowth factor and cytokine receptors. Science 261:1739-1744.

42. Saito, H., and M. Streuli. 1991. Molecular characterization ofprotein tyrosine phosphatases. Cell Growth Differ. 2:59-65.

43. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-ratory Press, Cold Spring Harbor, N.Y.

44. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencingwith chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA74:5463-5467.

45. Shaw, G., and R. Kamen. 1986. A conserved AU sequence fromthe 3' untranslated region of GM-CSF mRNA mediates selectivemRNA degradation. Cell 46:659-667.

46. Shen, S.-H., L. Bastien, B. I. Posner, and P. Chretien. 1991. Aprotein-tyrosine phosphatase with sequence similarity to the SH2domain of the protein-tyrosine kinases. Nature (London) 352:736-739.

47. Shuai, K., C. Schindler, V. R. Prezioso, and J. E. Darnell. 1992.Activation of transcription by IFN-,y: tyrosine phosphorylation ofa 91-kD DNA binding protein. Science 258:1808-1812.

48. Shuai, K., G. R. Stark, I. M. Kerr, and J. E. Darnell. 1993. A singlephosphotyrosine residue of Stat9l required for gene activation byinterferon--y. Science 261:1744-1746.

49. Silvennoinen, O., C. Schindler, J. Schlessinger, and D. E. Levy.1993. Ras-independent growth factor signaling by transcriptionfactor tyrosine phosphorylation. Science 261:1736-1739.

50. Taub, R., A. Roy, R. Dieter, and J. Koontz. 1987. Insulin as agrowth factor in rat hepatoma cells: stimulation of protooncogeneexpression. J. Biol. Chem. 262:10893-10897.

51. Tewari, M., P. Dobrzanski, K. L. Mohn, J.-C. Hsu, R. Bravo, andR. Taub. 1992. Rapid induction in regenerating liver of RL/IF-1,an IKB that inhibits NH-KB, Rel-pSO, and PHF, a novel KB sitebinding complex. Mol. Cell. Biol. 12:2898-2908.

52. Trowbridge, I. S. 1991. CD 45 A prototype for transmembraneprotein tyrosine phosphatases. J. Biol. Chem. 266:23517-23520.

53. Zheng, C.-F., and K.-L. Guan. 1993. Dephosphorylation andinactivation of the mitogen-activated protein kinase by a mitogen-induced Thr/Tyr protein phosphatase. J. Biol. Chem. 268:16116-16119.

54. Zheng, X. M., Y. Weng, and C. J. Pallen. 1992. Cell transformationand activation of pp60c-src by overexpression of a protein tyrosinephosphatase. Nature (London) 359:336-339.

55. Zipfel, P. F., S. G. Irving, K. Kelly, and U. Siebenlist. 1989.Complexity of the primary genetic response to mitogenic activa-tion of human T cells. Mol. Cell. Biol. 9:1041-1048.

MOL. CELL. BIOL.