Monoclonal antibody analysis of the proliferating cell nuclear antigen

(PCNA)

Structural conservation and the detection of a nucleolar form

NAUSHIN H. WASEEM and DAVID P. LANE

ICRF Clare Hall Laboratories, South Mimms, Potters Bar, Herts EN6 3LD, UK

Summary

The proliferating cell nuclear antigen, PCNA, hasrecently been identified as the polymerase S acces-sory protein. PCNA is essential for cellular DNAsynthesis and is also required for the in vitro repli-cation of simian virus 40 (SV40) DNA where it acts tocoordinate leading and lagging strand synthesis atthe replication fork. The cDNA for rat PCNA wascloned into a series of bacterial expression vectorsand the resulting protein used to immunize mice.Eleven new monoclonal antibodies to PCNA havebeen isolated and characterized. Some of the anti-bodies recognize epitopes conserved from man tofission yeast.

Immunocytochemical analysis of primate epi-thelial cell lines showed that the antibodies recog-nized antigenically distinct forms of PCNA and thatthese forms were localized to different compartmentsof the nucleus. One antibody reacted exclusively withPCNA in the nucleolus. These results suggest that thePCNA protein may fulfil several separate roles in thecell nucleus associated with changes in its antigenicstructure.

Key words: proliferating cell nuclear antigen, monoclonalantibodies, nucleolus, nucleoplasm, yeast PCNA, insect PCNA.

Introduction

The proliferating cell nuclear antigen (PCNA) (Takasakiet al. 1984), also known as cyclin (Bravo and Celis, 1980;Bravo et al. 1981; Mathews et al. 1984) or as auxiliaryprotein for polymerase 8 (Tan et al. 1986; Prelich et al.19876; Bravo et al. 1987), is required for simian virus 40(SV40) DNA replication in vitro (Prelich et al. 1987a).Experiments using anti-sense oligonucleotides and micro-injection of antibodies strongly suggest that PCNA is alsoessential for cellular DNA synthesis (Jaskulski etal. 1988;Zuber et al. 1989a). It is required for leading strandsynthesis in the SV40 system where it probably acts as anauxiliary protein for polymerase 5, coordinating leadingand lagging strand synthesis and rendering the polym-erase more processive (Prelich and Stillman, 1988). PCNAappears to be a cell cycle-regulated protein when exam-ined by immunofluorescence (Celis and Celis, 1985). Whencells are fixed using organic solvents PCNA staining islocalized at the intranuclear sites where DNA synthesis istaking place (Bravo and Macdonald-Bravo, 1987). Thepredominant distribution of PCNA appears to change withthe stage of the cell cycle. In early S phase PCNA has avery granular distribution and is absent from the nucleoli,while at late times in S phase prominent staining of thenucleoli is evident. In cells fixed using aldehydes, however,the distribution of PCNA is different and intense diffusenuclear staining is evident throughout the cell cycle(Bravo and Macdonald-Bravo, 1987). This difference hasbeen explained by the proposal that there are two forms ofPCNA: a soluble form lost on organic solvent fixation and

Journal of Cell Science 96, 121-129 (1990)Printed in Great Britain © The Company of Biologists Limited 1990

not involved in replication, and an insoluble form associ-ated with the sites of on-going DNA synthesis (Bravo andMacdonald-Bravo, 1987). Consistent with this hypothesisthe total concentration of PCNA varies at the most by onlytwo- to threefold during the cell cycle but there is a greaterfraction of PCNA that is insoluble due to chromatinassociation in S phase than in other phases of the cell cycle(Morris and Mathews, 1989). The concentration of PCNApresent in the cell nucleus seems to be in excess of thatrequired for its replicative function (Morris and Mathews,1989).

PCNA is a very conserved protein as can be seen by theamino acid sequence homology between mammalianPCNAs (Mouriuchi et al. 1986; Matsumato et al. 1987;Almendral et al. 1987; Ogata et al. 1985). The gene forPCNA is present not only in mammals (Matsumato et al.1987; Almendral et al. 1987) but also in plant cells (Suzukaet al. 1989). Recently a gene closely homologous to PCNAwas found in the genome of baculovirus Autographacalifornica (O'Reilly et al. 1989).

Three per cent of the patients suffering from systemiclupus erythematosus, an autoimmune disease, have anti-bodies against PCNA (Miyachi et al. 1978) and it wasthrough the use of sera from these patients that theprotein was first defined. These polyclonal antibodies havebeen used for in vitro and in vivo studies of PCNA (Wong etal. 1987; Tan et al. 1987). An autoantibody against PCNA,AK, capable of immunoprecipitating PCNA from cellextracts had been shown to inhibit DNA polymerase 6activity (Tan et al. 1987; Zuber et al. 1989a). It recognizesepitopes localized at the N terminus of PCNA (Ogata et al.

121

19876). Two monoclonal antibodies designated 19A2 and19F4 have been raised against rabbit PCNA. By Westernblotting they recognize PCNA from different mammalianspecies (Ogata et al. 1987a) and from Xenopus (Zuber et al.19896). However, they are unable to recognize PCNA fromSaccharomyces cerevisiae (Bauer and Burgers, 1988). Theyare also unable to immunoprecipitate PCNA from anysource (Tan et al. 1987) and they do not inhibit plasmidreplication (Zuber et al. 1989a). Their epitopes have beenlocalized to the central region of the PCNA molecule(Ogata et al. 19876J.

Here we report the expression of recombinant PCNAgenes in Escherichia coli, and the production and charac-terization of 11 new monoclonal antibodies, which defineat least three distinct epitopes on primate PCNA. Severalof these antibodies were able to immunoprecipitate thenative protein and recognize it in different species includ-ing Schizosaccharomyces pombe and Spodoptera frugi-perda (fall armyworm) IPLB-SF21 cells (insect cells).Immunocytochemical studies of PCNA using these newantibodies suggest that the protein exists in multiplestates in the cell nucleus. The results help to resolve someof the complexities of the cell staining pattern observedwith polyclonal anti-PCNA sera. Several of the antibodieswere able to react with PCNA in routine formalin-fixedhistological sections, implying that they may be usefulreagents for the analysis of the proliferative status oftumour tissues.

Materials and methods

Preparation ofcDNA constructsThe cDNA of rat PCNA, PCR-1 in pBR322 was obtained fromMoriuchi (Matsumato et al. 1987). It was subcloned into the/3-galactosidase expression vector pUR288 (Ruther and Muller-Hill, 1983). In order to do this, pUR288 was partially digestedwith EcoBl, filled with Klenow fragment and digested again withXbal. An Xbal and Hindi digestion of PCR-1 gave two fragmentsof 1.0 and 1.1kb (lkb=105 base-pairs). The 1.0kb fragmentcontaining the PCNA cDNA was purified and ligated into Xbaland the blunt-ended pUR288 vector to give pC288.

A second construct was made in the protein A expression vectorpRIT2T obtained from Pharmacia (Nilson et al. 1985). A 728 bp(base-pair) fragment containing part of the PCNA cDNA inpC288, obtained by BomHI and Pstl digestion, was subcloned intopRIT2T to give pC2T. This construct lacked the 3'-end 93nucleotides of PCNA. E. coli N4830-1 cells, which contain a heat-inducible lamda repressor, were transformed with pC2T. Thetransformed cells, when induced at 42 °C, produced a partialPCNA polypeptide fused to protein A. The third plasmid used waspT7.7 (Tabor and Richardson, 1985). It was digested with HindBl,made blunt-ended by Klenow fragment, cut again with BaniHI,ligated with a 946 bp BarniU. and EcoRV fragment of pC288 togive pC10. E. coli p lys S BL21(DE3) cells (a gift from W. Studier,Brookhaven Laboratories, USA) (Studier and Moffatt, 1986) weretransformed with pClO.

Purification of the fusion proteinsProtein A fusion protein. A bacterial culture of E. coli N4830-1

cells containing pC2T was grown and induced according to themanufacturer's instructions. The cells were pelleted at 7000 revsmin"1 in a GSA 600 rotor in a Sorvall RC-5B centrifuge forIfimin The pellet was suspended in 10mM sodium phosphatebuffer containing 150 mM sodium chloride, pH7.4 (PBS), andtreated with lmgrnl"1 lysozyme. After sonication for 20s atmaximum amplitude it was spun at 30 000 revs min"1 in a 60Tirotor in a Beckman L8-55M centrifuge for 30 min. The super-natant was applied to a column of human IgG-Sepharose 6 FastFlow (Pharmacia). The column was washed with 50 vols of PBS

containing 1 M sodium chloride and 0.1% Nonidet P40 (NP40).The bound protein A-PCNA fusion protein was eluted with100 mM glycine-HCl, pH3.0. The fractions were immediatelyneutralized with 2 M Tris-base.

77 fusion protein. A culture of pCIO in E. coli p lys SBL21(DE3) cells was grown and induced as previously described(Studier and Moffatt, 1986). The cells were pelleted at 7000 revsmin"1 and frozen rapidly. Thawing the pellet in 50 mM Tris-HCl,1 mM EDTA, pH 8.0, released intracellular lysozyme resulting inthe lysis of the cells. The lysed cells were sonicated for 20 s andspun at 30 000 revs min"1 for 30 min (as above). The pellet waswashed twice with 1 % NP40 and stirred in 1 M-urea for 1 h at 4°Cand then centrifuged at 30 000 revs min"1 for 30 min. The pelletwas resuspended in 8 M urea, spun at 30 000 revs min"1 for 30 minand the supernatant dialysed gradually against PBS to yield asoluble PCNA fusion protein.

Monoclonal antibody productionAt intervals of 21 days six mice were injected intraperitoneallywith 50 ^g of protein A-PCNA fusion protein (obtained frompC2T) in PBS emulsified with an equal volume of Freund'scomplete adjuvant. After three injections the titres of the serawere monitored by Western blotting and the mouse with thehighest titre was boosted with two injections of 10 /ig of PCNAfusion protein obtained from pCIO. These final boosts were carriedout on two successive days: one intraperitoneally and the otherintravenously. Three days after the final intravenous boost themouse was killed and the spleen removed. The spleen cells werefused with mouse myeloma cells Sp2/0-Agl4 using 50% poly-ethylene glycol 1500 (Harlow and Lane, 1988). The cells weregrown in Dulbecco's modified Eagle's medium (DMEM) contain-ing 20 % foetal calf serum (FCS) and supplemented with inter-leukin 6 at 4ngml~1 as growth factor (E. Weiss, personalcommunication). Hypoxanthine and azaserine were used to selectspecifically the hybridoma cells. The fusion was screened by cellstaining of CV-1 cells (an immortalized line of monkey kidneyepithelial cells) and Western blotting of HeLa S-100 fraction (seebelow).

IsotypingThe antibodies were isotyped using the Serotec Mouse IsotypingKit (Serotec, Oxford, UK).

Western blottingSoluble extracts (SlOO) from HeLa cells were prepared accordingto our laboratory's standard protocol (Gannon and Lane, 1987).The samples for Western blotting from insect cells were preparedby adding SDS sample buffer directly to the cell pellet. The pelletwas then sonicated and boiled for 5 min. SDS-polyacrylamide gelelectrophoresis was performed according to Laemmli (1970).Electrophoretic transfer of proteins to nitrocellulose was per-formed according to Towbin et al. (1979). The nitrocellulosemembrane was blocked with a 20% (w/v) solution of driedskimmed milk for 1 h at room temperature followed by incubationwith the different monoclonal antibodies. After washing with 1 %NP40 in PBS, the blot was incubated for 1 h at room temperaturewith alkaline phosphatase-conjugated rabbit anti-mouse immu-noglobulin (diluted 1:500 in PBS containing 10% FCS). The blotswere washed again and incubated with bromochloroindoyl phos-phate and nitrobluetetrazolium substrate (Harlow and Lane,1988). The reaction was stopped with 1 % acetic acid. For Westernblotting of the cell extract of Schizosaccharomyces pombe, abiotinylated sheep anti-mouse antibody at 1:500 dilution wasused. After incubating with horseradish peroxidase-conjugatedstreptavidin the blot was developed with the DMB(diaminobenzidine)/H2O2 substrate (Harlow and Lane, 1988).

Purification of the monoclonal antibodiesThe antibodies were purified on a protein A-Sepharose CL-4Bcolumn according to standard protocol (Harlow and Lane, 1988).

Biotinylation of antibodiesPurified antibodies were biotinylated using biotin iV-hydroxy-

122 N. H. Waseem and D. P. Lane

succinimide ester as previously described (Harlow and Lane,1988). The molar ratio of biotin Af-hydroxysuccinimide to proteinwas 20:1.

ImmunoprecipitationFour 15 cm plates of CV-1 cells at 70% confluency were eachlabelled with 250 /<Ci of [36S]methionine overnight. The cells werewashed once with PBS and removed with 5 mM EDTA in PBS at37 °C. They were washed again with 150 mM NaCl, 50 mM Tris-HC1, 5mM EDTA, pH8.0 (NET), and lysed in 1% NP40 in NETcontaining 1 mM-phenylmethylsulphonyl fluoride and 1 mM dith-iothreitol. After spinning at 100 000 revs min"1 in a TL-100.2rotor in Beckman TL-100 ultracentrifuge for 30 min, the super-natant was preabsorbed with protein G—Sepharose 4 Fast Flow(Pharmacia) for 30 min at 4°C. The suspension was spun at7000 revs min"1 in a Sorvall RC-5B centrifuge for 5 min and thesupernatant was divided into four aliquots per plate. An equalvolume of hybridoma tissue culture supernatant containing eachtest antibody was added and incubated overnight at 4°C. Anti /S-galactosidase monoclonal antibody BG2 was used as a control. A50 jd sample of protein G-Sepharose 4 Fast Flow was added toeach tube and the incubation continued for 30min at 4°C. Theprotein G beads were washed with 250 mM NaCl, 1% NP40 inNET and boiled for 5 min in sample buffer. Half of the sample wasapplied to a 12.5% SDS-polyacrylamide gel. After electrophor-esis at 100 V the gel was fixed in 50 % trichloroacetic acid (TCA)for 15 min and incubated in Enhance (Du Pont) for 15 min.Enhance was removed and the gel was washed with water anddried. The autoradiogram was developed overnight at roomtemperature.

Enzyme-linked immunosorbent assay (ELJSA)For the competition assay, protein at a concentration oflO/igml"1 obtained from pCIO was coated onto a 96-well plasticmicrotitre plate by overnight incubation at 4°C. The plate wasblocked with 10 % bovine serum albumin (BSA) in PBS for 3 h atroom temperature. Competing unlabelled monoclonal antibodiesat different concentrations were added and incubated overnightat 4°C. The wells were washed with 0.1 % NP40 in PBS and 100 jdof biotinylated anti-PCNA antibody (8/jgml"1) was added. Afterincubation for 3h at 4°C, streptavidin-horseradish peroxidase at1:1000 dilution in PBS containing 1 % BSA, was added for 15 min.Colour was developed with 3',3',5',5'-tetramethylbenzidine(TMB) and H2O2, and read on an ELJSA reader (MolecularDevices, Vmax) at 450 run after stopping the reaction with 50 /d of1 M H2SO4.

For the sandwich ELJSA an extract of CV-1 cells was preparedin 1 % NP40 in NET. Different affinity-purified monoclonalantibodies at Sjjgml"1 were added to the wells of a 96-wellmicrotitre plate and incubated overnight at 4°C. After incubatingwith a 10 % BSA blocking solution for 4 h at room temperaturedifferent dilutions of CV-1 cell extracts were added and incubatedovernight at 4°C. The wells were washed with 0.1 % NP40 in PBSand biotinylated anti-PCNA antibody was added and incubatedfor 3h at 4°C. After repeating the above washing the wells wereincubated with horseradish peroxidase-labelled streptavidin for15 min at room temperature. After washing away the unboundstreptavidin the colour was developed with TMB and H2O2 asbefore.

Cell stainingCV-1 cells were grown on coverslips in DMEM containing 10%FCS. They were fixed in a 1:1 (v/v) mixture of acetone-methanolfor 5 min at room temperature and air dried. They were incubatedwith hybridoma tissue culture supernatant containing anti-PCNA antibodies for 1 h at room temperature. FITC-conjugatedrabbit anti-mouse antibody was then added and incubated for 1 hat room temperature. After washing away the unbound antibodythe coverslips were mounted in Gelvatol (Monsonto chemicals) onglass slides and dried.

Tissue stainingFormalin-fixed paraffin-embedded tissue sections were deparaffi-

nized in xylene and passed through a graded ethanol series. Theywere fixed in acetone for 5 min and incubated with 1 % NP40 for5 min. After adding the monoclonal antibody the sections wereincubated overnight at 4°C in humidified chamber. After wash-ing, the slides were incubated with F1TC-conjugated rabbit anti-mouse antibody for l h at room temperature. A third layer ofF1TC-conjugated swine anti-rabbit antibody was sometimes usedto enhance the staining. The slides were mounted in Gelvatol anddried.

Results

Preparation and purification of fusion proteinThe aim of this study was to produce a library of mono-clonal antibodies against native PCNA that would beuseful immunological reagents for the further analysis ofPCNA activity in vivo and in vitro. An obvious source ofthe antigen would be PCNA purified from eukaryotic cells,but the complexity of the purification procedure and pooryield of the protein encouraged us to explore alternativeways of obtaining the antigen without compromising itspurity. We therefore investigated the possibility ofexpressing the rat PCNA cDNA in E. coli. The /3-galactosi-dase-PCNA fusion protein produced by pC288 was foundto be insoluble in PBS and Tris/HCl buffers. We thereforedecided to subclone the cDNA of PCNA into pRTT2T, sinceprotein A fusion proteins obtained from this vector werereported to be soluble (Nilson et al. 1985).

The cDNA cloned in pRIT2T lacked 93 nucleotides fromthe 3' end of the PCNA coding region and therefore theexpected size of the protein A—PCNA fusion proteinobtained from this construct would be approximately 54K(K=103Afr). The actual fusion protein produced by pC2Twas about 68K and was found, significantly, to be com-pletely soluble. Since the fragment of protein A present inthis vector was about 27K, the molecular weight of thefusion protein obtained in this study represents an over-estimation of about 14K. This discrepancy in the molecu-lar weight of the fusion protein could be due to the lack of atermination codon in the cDNA construct of PCNA result-ing in the use of a translational termination codon presentdownstream from the multiple cloning site in pRTT2T. Thediscrepancy may also be due, in part, to the previouslyreported anomalous behaviour of PCNA in SDS—poly-acrylamide gels (Matsumato et al. 1987). The recovery ofthe protein A-PCNA fusion protein was about 80 %. Thepreparation was contaminated with a minor band of about70K and by some lower molecular weight species. Thesmall species reacted with alkaline phosphatase-conju-gated rabbit IgG, suggesting that they were derived fromproteolytic cleavage of the protein A-containing section ofthe fusion protein.

The PCNA fusion protein produced in pCIO had anadditional 20 amino acids at its N terminus derived fromthe amino terminus of the bacteriophage T7 gene 10protein. The protein produced by pCIO in p lys SBL21(DE3) cells was insoluble in PBS and Tris-HClbuffers. It was found, however, to be soluble in 8 M urea. Toour surprise the fusion protein remained soluble atlmgml" 1 when the urea concentration was graduallylowered by slow dialysis against PBS over a period of16-24 h. The protein obtained by this simple inclusion-body solubilization procedure was at least 80 % pure basedon the Coomassie Blue staining of SDS-polyacrylamidegels (Fig. 1). This preparation was used without furtherpurification.

Proliferuting cell nuclear antigen 123

200.1 2 3 4 5 6 7 8 9 10 11 12

« PCNA

Fig. 1. SDS-polyacrylamide gel electrophoresis of proteinpurified from E. coli transformed with the pCIO PCNAexpression plasmid. Lane 1: total cell extract of p lys SBL21(DE3) cells containing the parent plasmid pT7.7. Lane 2:total cell extract of p lys S BL21(DE3) cells containing pCIO.Lane 3: purified PCNA fusion protein from p lys S BL21(DE3)cells containing pCIO.

Production of monoclonal antibodies against PCNAThe purified protein A-PCNA fusion protein was used toimmunize six mice. Dilutions of sera were tested for theirability to detect PCNA on Western blots of HeLa S100fraction. All the sera reacted strongly with a single 36Kband. One mouse had a titre of 1:10 000 in this assay andwas used for the fusion. Of the 36 wells that initiallycontained anti-nuclear antibody in the cell-staining assay,11 clones were established as stable anti-PCNA antibody-producing clones. All antibodies were positive on Westernblots of HeLa S100 fraction where they reacted with asingle band of 36K. Some antibodies reacted more stronglythan others: in particular, PC3 and PC6 gave very strongreactions while PC7 reacted very weakly. Two antibodies,PC2 and PCIO, also reacted with a faint band of about120K besides reacting with 36K protein band (Fig. 2).

Characterization of monoclonal antibodiesA summary of the characteristics of the anti-PCNA mono-clonal antibodies are listed in Table 1.

Immunoprecipitation. The antibodies were tested for

87,

88-

45 _

Fig. 2. Western blotting of HeLa S100 with different anti-PCNA antibodies. The antibodies were: lanes 1, PCI; 2, PC2; 3,PC3; 4, PC4; 5, PCS; 6, PC6; 7, PC7; 8, PC8; 9, PC9; 10, PC10;11, PC11; 12, PBS control. The prestained molecular weightmarkers (xlO~3) are (from the top): myosin, phosphorylase b,bovine serum albumin, ovalbumin, cr-chymotrypsinogen andlactoglobulin.

their capacity to immunoprecipitate PCNA from a [^S]-methionine-labelled cell extract of CV-1 cells. Ten of the 11antibodies were capable of immunoprecipitating PCNAfrom the cell extract (Fig. 3). The antibodies PCI (lane 2),PC2 (lane 3), PC3 (lane 4), PC7 (lane 8), PC8 (lane 9) andPC10 (lane 11) immunoprecipitated substantial amountsof PCNA. The remaining antibodies reacted more weaklyand PC11 (lane 12) appeared completely negative. Theidentity of PCNA in these immunoprecipitation exper-iments was confirmed by Western blotting the [ S]-methionine-labelled immunoprecipitated bands (data notshown). Equal loading of the gel can be observed from theintensity of the 68K band, present in all samples. No otherobvious bands co-precipitated with PCNA in all lanes.

Competition assays. To determine whether the 11 dif-ferent antibodies bind to sterically discrete epitopes theywere examined in competition assays and sandwichELJSA. For the competition assays four of the antibodieswere biotinylated (PC3, PC8, PC9 and PCIO). The abilityof all 11 unlabelled antibodies to block the binding of thesefour labelled antibodies to solid-phase PCNA was quanti-fied. The results of this analysis using antibody PC3 aslabel are illustrated in Fig. 4 and the complete results aresummarized in Table 2. Although this analysis is notcomplete a minimum of three and a maximum of four

Table 1. Summary of the characteristics of the new anti-PCNA monoclonal antibodies

Western blotting

Monoclonalantibodies Isotype HeLa

S. pombe(yeast)

S. frugiperda*(insect)

Immuno-precipitation

Cellstaining

Tissuestaining

PCIPC2PC3PC4PC5PC6PC7PC8PC9PCIOPC11

IgGlIgGlIgG2bIgG2aIgGlIgGlIgGlIgG3IgGlIgG2aIgGl

• Spodoptera frugiperda.

124 N. H. Waseem and D. P. Lane

2 3 4 5 6 7 8 9 10 11 12

2 0 0 -

97 _

68 - «>-

45 -

25 -

Fig. 3. Immunoprecipitation of [3BS]methionine-labelled CV-1cells with anti-PCNA antibodies. The antibodies were: lanes 1,BG2; 2, PCI; 3, PC2; 4, PC3; 5, PC4; 6, PC5; 7, PC6; 8, PC7; 9,PC8; 10, PC9; 11, PCIO; 12, PC11.

Table 2. Competitive inhibition between all 11 anti-PCNA antibodies and biotinylated antibodies PC3, PC8,

PC9 and PCIO

0.001 0.01 0.1Antibody concn

10

Fig. 4. Competition ELISA of unlabelled antibodies withbiotinylated anti-PCNA antibody PC3. Different dilutions ofunlabelled antibodies were added to the PCNA fusion protein-coated plate and probed with a fixed volume of biotinylatedPC3. The graph is split into two for clarity.

Comantibody

PCIPC2PC3PC4PCSPC6PC7PCSPC9PCIOPC11BG2

PC3

0.40.30.8—

4.00.8—

0.3-

0.10.4-

CM* for biotinylated antibody

PCS

1.91.01.2-

30.0—-

2.0-

0.31.2-

PC9

- t——

_

—_0,4——-

PC10

2.10.21.6_

25__

5.0-

0.10.4-

•Micrograms of competing antibody to give 50% inhibition of labelledantibody.

t Too high to be significant

sterically separate epitopes on PCNA can be defined. PCI,PC2, PC3, PC5, PC8, PC10 and PC11 fall into one group inthat they all block the binding of the PC3, PC8 and PC10to PCNA, albeit with widely differing efficiencies. Inparticular, PC10 is a very good inhibitor of the binding ofall three of these antibodies and indeed is more efficient atblocking PC3 and PC8 binding than the appropriatehomologous antibody. This suggests that PC10 has aparticularly high affinity for PCNA. PC9 clearly recog-nizes a second discrete epitope, since, apart from itself,none of the antibodies can inhibit its binding to PCNA. Atleast one and possibly two further epitopes are defined byPC4 and PC7, since these antibodies do not block thebinding of any of the labelled antibodies to PCNA. ThePC6 antibody further refines the epitope map, since it isable to block partially the binding of PC3 but not PC8 orPC10, implying that even in the first large group ofcompeting antibodies some differences in the binding siteon PCNA exist. A diagrammatic representation of theseresults is shown in Fig. 6 (below).

Sandwich ELISA. The biotinylated antibody PC9 wasused to try and detect PCNA captured by other anti-PCNAantibodies in a sandwich ELISA. The antibodies that wereable to form a sandwich were PC2, PC3, PC5, PC6, PC8,PCIO and PC11 (Fig. 5). The best pair of antibodies for thequantitative assay of PCNA by sandwich ELISA is prob-ably PC8 (capture) and PC9 (label). PC9 could not detectPCNA captured by antibodies PC4, PC7 and PCI.

Immunofluorescence: nucleolar and nucleoplasmicPCNA staining. In the earlier studies of Celis (Celis andCelis, 1986) immunofluorescence of methanol-fixed humanamnion cells was used to define a series of distinct stainingpatterns for PCNA that changed through the cell cycle.This complex set of staining patterns seen with polyclonalanti-PCNA antibodies represents the location of the insol-uble fraction of PCNA and is localized at sites of DNAsynthesis. The monoclonal antibodies 19A2 and 19F4 havebeen reported to give similar results (Ogata et al. 1987a;Madsen et al. 1987). We compared the staining patternsobtained with these two commercially available anti-bodies with the 11 new anti-PCNA antibodies on culturesof CV-1 monkey epithelial cells fixed in acetone-methanolor methanol alone. The PC7 antibody failed to stain thesecells at all. The majority of the new antibodies showed thesame patterns as seen with 19A2 and 19F4 (data notshown). Specifically, the antibodies PCI, PC2, PC3, PC5,

Proliferating cell nuclear antigen 125

2 4 6Antigen dilution

4 6Antigen dilution

10

Fig. 5. Sandwich ELISA of PCNA with biotinylated PC9 asprobe. Fixed concentration of different 'capture' antibodies wereadded to coat the plate. Different dilutions of CV-1 extract werethen added to the plate and probed with labelled PC9.

Fig. 6. Diagrammatic representation of the epitope mappingresults. Epitopes are represented by circles and overlaps implystearic interference.

Fig. 7. Immunofluorescence staining of CV-I cells with anti-PCNA antibodies: A, PC9; B, PC9; C, PC4. Bars: 10 /JM (A andC); 30/JM(B).

PC6, PC8, PC10 and PC11, like 19A2 and 19F4, showedgranular staining throughout the nucleus. In some cellsthe Celis pattern Sb/Sc was evident in that the nucleo-plasm was stained but the nucleoli were not. Other cellsshowed staining of both the nucleoplasm and nucleoli(pattern Sd/Se) and, finally, a fraction of the cells showedvery weak staining of the nucleoplasm but strong stainingof the nucleoli (pattern Sf). PC4 showed a subtle differencein that cells with the Sb/Sc and Sd/Se patterns wereclearly visible (Fig. 7C); however, no cells with the Sfpattern of predominantly nucleolar staining weredetected. The most surprising reaction, however, was withthe antibody PC9. This antibody stained only the nucleoli(Fig. 7A and B) of the CV-1 cells. Most cells were stronglypositive for this nucleolar staining. There was no evidence

126 N. H. Waseem and D. P. Lane

for any nucleoplasmic staining with PC9 although in asmall fraction of the cells the nucleolar staining was weak.

Tissue staining. These antibodies were also checked fortheir reactivity on formalin-fixed paraffin-embedded ton-sil sections, as PCNA is potentially a useful marker for thehistological detection of cells with proliferative potential.PC2, PCS, PC6, PC8, PC10 and PC11 were all able to reactwith PCNA in sections of human tonsil where theypreferentially stained the nuclei of cell within the germi-nal centres. None of the other antibodies reacted withPCNA in these sections (Table 1). Some of the positiveantibodies were used to stain sections of rectal carcinomatissue. The antibodies PC2, PC10 and PC11 stained theproliferating tumour tissue in these sections (data notshown).

PCNA in insect cells as well as in S. pombeSince we had used CV-1 cells and HeLa cells for screeningthe hybridoma fusion, we wanted to know whether theseantibodies would also react with PCNA from insect cellsand S. pombe cell extract on Western blots. Of the 11antibodies tested five of them, namely, PC2, PC3, PC5,PC8 and PC10, gave a positive signal. They reacted with a34K band on a Western blot in insect cells (Fig. 8A) andwith a 36K band in S. pombe extract (Fig. 8B). Theantibody PC10 reacted particularly strongly with both theinsect and S. pombe extracts.

1 2 3 4 5 6 7 8

97-

68-

45"

25-

18-

14-

1 2 3 4 5 6 797 ~

68 -

4 5 -

Discusslon

Antibodies against proteins required for SV40 DNA repli-cation, like T antigen and DNA polymerase a, are veryuseful in understanding the various steps involved inDNA replication (Smale and Tjian, 1986; Murakami et al.1986). To understand the role of PCNA in DNA replicationit is necessary to have monospecific antibodies againstPCNA that will recognize different epitopes on nativePCNA. The monoclonal antibodies 19A2 and 19F4 wereraised against SDS-denatured rabbit PCNA (Ogata et al.1987a; Tan et al. 1987) and are unable to immunoprecipi-tate native PCNA or inhibit PCNA-dependant DNA repli-cation (Tan et al. 1987). We therefore decided to expressPCNA in bacteria, purify it under non-denaturing con-ditions and use it to raise antibodies. Consistent with themore native state of our immunogen, 10 of the 11 anti-bodies produced were able to immunoprecipitate PCNAfrom cell extracts. Little is known about the role of PCNAin DNA replication. Although it increases the processivityof polymerase 8 it has not been shown to associate with theenzyme directly. It has been speculated that PCNA in-hibits an inhibitor of DNA replication (Lee et al. 1989).Since these new antibodies are able to recognize all formsof native PCNA, they are powerful tools with which tostudy various proteins associated with PCNA and theireffect on the activity of PCNA.

These antibodies recognized different epitopes on PCNAas can be seen from the competition assay and sandwichassay data. The antibody PC9 seems to recognize acompletely discrete epitope. It did not compete with anyother antibody and binds to PCNA captured by seven otherantibodies (PC2, PC3, PCS, PC6, PC8, PC10 and PC11).When CV-1 cells were stained with PC9 only the nucleoliwere positive. This distribution was seen in all the cellsexamined though in a fraction of the cells the staining wasrelatively weak. The PC4 antibody failed to stain selec-tively the nucleoli of any cells. The remaining antibodies

2 5 -

18-

14-

BFig. 8. A. Western blot of insect cells, with: lanes 1, PCI; 2,PC2; 3, PC4; 4, PC3; 5, PC10; 6, PC5; 7, PC8; and 8, PBS.B. Western blot of S. pombe cell extract with: lanes 1, PC4; 2,PC2; 3, PC3; 4, PC10; 5, PCS; 6, PC8; and 7, PBS control. Mrvalues (xlO~3) on left.

showed a more conventional series of staining reactions, inthat cells were identified in each of the categories desig-nated by Celis and were similar to the patterns seen withthe existing anti-PCNA reagents. Despite these verydifferent staining patterns seen with PC9, and to a lesserextent PC4, the immunochemical evidence stronglysuggests that all of these antibodies are highly specific forPCNA. PC9 reacts strongly and specifically with PCNA inimmunoblotting, immunoprecipitation and ELJSA assays.The sandwich immunoassays also show that the PC9epitope is present on the same molecules as the epitopesrecognized by PC2, PC3, PC5, PC6, PC8, PC10, and PC11.If all of the staining reactions are due to PCNA, then theresults suggest that while PCNA is present in both thenucleolus and the nucleoplasm throughout the cell cycle,the epitopes available for antibody binding are different inthe two compartments. The PC9 epitope is only present onthe nucleolar form and is absent from or masked on thenucleoplasmic form. These surprising results suggest thatearlier interpretations of the nuclear location of PCNA are

Proliferating cell nuclear antigen 127

flawed because the serological reagents used were onlyable to detect a fraction of the total PCNA.

Why should PCNA be present in the nucleoli through-out the cell cycle? Autoradiography using [3H]thymidinesuggests that replication of the nucleolar DNA occurs at adenned stage late in S phase. This is consistent with theappearance of the Sf pattern seen with most anti-PCNAantibodies and the earlier reports of 'movement' of PCNAinto the nucleolus at this stage in S phase. The nucleolusmay represent a special site for DNA replication. Minutevirus of mice (MVM) replicates in the nucleolus, andnucleolar components are redistributed during adenovirusinfection (Walton et al. 1989). SV40 DNA is associatedwith the nucleoli during SV40 infections (Geuskens andMay, 1974). Alternatively, it is possible that PCNA mayplay another discrete, non-replicative role in the nu-cleolus. In this context it is interesting to note that twoother proliferative markers, the Ki-67 antigen (Verheijenet al. 1989) and numatrin (Feuerstein et al. 1988), are alsonucleolar and that the rapid onset of ribosomal RNAtranscription is a characteristic early feature of the mi-totic response following recovery from amino acid star-vation (Grummt et al. 1976). Perhaps PCNA, like severalother replication factors (DePamphilis, 1988), may also beactive in transcription.

We thank Dr T. Moriuchi for his generous gift of PCR-1plasmid, Dr E. B. Lane for tissue sections and assistance with theimmunohistochemiatry, Professor Paul Nurse for extracts of S.pombe cells, Dr W. Studier for E. coli p lys S B21CDE3) cells andthe other members of the Molecular Immunochemistry Labora-tory for helpful advice.

References

ALMENDRAL, J. M., HUEBSCH, D., BLUNDELL, P. A., MACDONAIJ>-BRAVO,H. AND BRAVO, R. (1987). Cloning and sequence of the human nuclearprotein cyclin: homology with DNA-binding proteins. Proc. natn.Acad. Sci. U.SJi. 84, 1576-1579.

BAUER, G. A. AND BURGERS, P. M. (1988). The yeast analog ofmammalian cyclin/proliferating cell nuclear antigen interacts withmammalian DNA polymerase delta. Proc. natn. Acad. Sci. U.S.A. 85,7506-7510.

BRAVO, R. AND CELIS, J. E. (1980). A search for differential polypeptidesynthesis throughout the cell cycle of HeLa cells. J. Cell Biol. 84,795-802.

BRAVO, R. AND CELIS, J. E. (1985) Changes in the nuclear distribution ofcyclin (PCNA) during S-phase are not triggered by post translationalmodification that are expected to moderately affect its charge. FEBSLett. 182, 435-440.

BRAVO, R., FEY, S. J., BELLATIN, J., LARSEN, P. M., ARBVALO, J. ANDCKUS, J. E. (1981). Identification of a nuclear and of a cytoplasmicpolypeptide whose relative proportion are sensitive to changes in therate of cell proliferation. Expl Cell Res. 136, 311-319.

BRAVO, R., FRANK, R., BLUNDELL, P. A AND MACDONALD-BRAVO, H.(1987). Cyclin/PCNA is the auxiliary protein of DNA polymerase 6.Nature, Land. 326, 515-517.

BRAVO, R. AND MACDONALD-BRAVO, H. (1987). Existence of twopopulations of cyclin/proliferating cell nuclear antigen during the cellcycle: association with DNA replication sites. J. Cell Bwl. 105,1549-1564.

CELIS, J. E. AND CELIS, A. (1986). Cell cycle dependent variations in thedistribution of the proliferating cell nuclear antigen in cultured cells:subdivision of S phase. Proc. natn. Acad. Sci. U.S-A. 82, 3262-3266.

DEPAMPHIUS, M. L. (1988). Transcriptional elements as components ofeukaryotic origins of DNA replication. Cell 52, 636-638.

FEUERSTEIN, N., CHAN, P. K. AND MOND, J. J. (1987). Identification ofNumatnn, the nuclear matnx protein associated with the induction ofmitogenesis, as the nucleolar protein B23. J. biol. Chem. 263,10608-10612.

GANNON, J. V. AND LANE, D. P. (1987). p63 and DNA polymerase acompete for the binding to SV40 T antigen. Nature, Land. 329,456-458.

GEUSKENS, M. AND MAY, E. (1974). Ultrastructural localization of SV40viral DNA in cells, during lytic infection, by in situ molecularhybridization. Expl Cell Reg. 87, 175-185.

GRUMMT, I., SMITH, V. A. AND GRUMMT, F. (1976). Amino acid starvationaffects the initiation frequency of nucleolar RNA polymerase. Cell 7,439-445.

HAHLOW, E. AND LANE, D. (1988). Antibodies, A Laboratory Manual.Cold Spring Harbor Laboratory Press, New York.

JASKULSKI, D., DERIEL, J. K., MBRCER, W. E., CALABRETTA, B. ANDBASERGA, R. (1988). Inhibition of cellular proliferation by antisenseoligodeoxynucleotides to PCNA cyclin. Science 240, 1544-1646.

LAEMMLI, U. K. (1970). Cleavage of structural proteins during theassembly of head of bacteriophage T4. Nature, Land. 227, 680-685.

LEE, S. K., KWONG, A. D., ISHIMI, Y. AND HURWITZ, J. (1989). Studies onthe DNA elongation inhibitor and its proliferating cell nuclearantigen-dependent control in simian virus 40 DNA replication in vitro.Proc. natn. Acad. Sci. U.S.A. 86, 4877-4881.

MADSEN, P., OGATA, K. AND CELIS, J. E. (1987). PCNA (cyclin)autoantibodies and monoclonal antibodies reveal similar patterns ofcyclin (PCNA) antigen staining in human cultured cells. Leukemia 1,220-226.

MATHEWS, M. B., BERSTEIN, R. M., FRANZA, B. R. JR. AND GARRELS, J. I.(1984). Identity of the proliferating cell nuclear antigen and cyclin.Nature, Land. 303, 374-376.

MATSUMATO, K., MORIUCHI, T., KOJI, T. AND NAKANE, P. K. (1987).Molecular cloning of cDNA coding for rat proliferating cell nuclearantigen (PCNA)/cyclin. EMBO J. 6, 637-642.

MIYACHI, K., FRITZLER, M. J. AND TAN, E. M. (1978). Autoantibody to anuclear antigen in proliferating cells. J. Immun. 121, 2228-2234.

MORIUCHI, T., MATSUMATO, K., KOJI, T. AND NAKANE, P. K. (1986).Molecular cloning and nucleotide sequence analysis of ratPCNA/cyclin cDNA. Nucl. Acids Symp. Ser. 17, 117-120.

MORRIS, G. F. AND MATHEWS, M. B. (1989). Regulation of proliferatingcell nuclear antigen during the cell cycle. J. biol. Chem. 264,13856-13864.

MURAKAMI, Y., WOBBE, C, WBISSBACH, L, DEAN, F. AND HURWITZ, J.(1986). Role of DNA polymerase trand DNA primase in simian virus40 replication m vitro. Proc. natn. Acad. Sci. U.SJi. 83, 2869-2873.

NILSON, B., ABRAHMSEN, L. AND UHLEN, M. (1986) Immobilization andpurification of enzyme with staphylococcal protein A gene fusionvectors. EMBO J. 4, 1075-1080.

OGATA, K., KURKI, P., CELIS, J. E., NAKAMURA, R. M AND TAN, E M.(1987a) Monoclonal antibodies to a nuclear protein (PCNA/cyclin)associated with DNA replication. Expl Cell Res. 168, 475-486.

OGATA, K , OGATA, Y., NAKAMURA, R. M. AND TAN, E M. (1985).Purification and N-terminal amino acid sequence of proliferating cellnuclear antigen (PCNA)/cyclin and development of F.T.TSA for anti-PCNA antibodies. J. Immun. 136, 2623-2627.

OGATA, K., OGATA, Y , TAKASAKI, Y. AND TAN, E. M. (19876). Epitopes onproliferating cell nuclear antigen recognized by human lupusautoantibody and murine monoclonal antibody. J. Immun. 139,2942-2946.

O'REILLY, D. R., CRAWFORD, A. M. AND MILLER, L. K. (1989). Viralproliferating cell nuclear antigen(letter). Nature, Land 337, 606.

PRELICH, G., KOSTURA, M., MARSHAK, D. R., MATHEWS, M. B. ANDSTILLMAN, B. (1987a). The cell-cycle regulated proliferating cellnuclear antigen is required for SV40 DNA replication in vitro. Nature,Land. 326, 471-475.

PRELICH, G. AND STILLMAN, B. (1988). Coordinated leading and laggingstrand synthesis during SV40 DNA replication m vitro requiresPCNA. Cell 53, 117-126.

PRELICH, G., TAN, C. K., KOSTURA, M., MATHEWS, M. B., SO, A. G.,DOWNEY, K. M. AND STILLMAN, B (19876). Functional identity ofproliferating cell nuclear antigen and a DNA polymerase-deltaauxiliary protein. Nature, Land. 326, 517-520.

RUTHER, M. AND MULLER-HILL, B. (1983). Easy identification of cDNAclones. EMBO J. 2, 1791-1794.

SMALE, S. T AND TJIAN, R. (1986). T-antigen-DNA polymerase a compleximplicated in simian virus 40 replication. Molec. cell. Biol. 6,4077-4087.

STUDIER/F. W. AND MorFATT, B. A. (1986). Use of bactenophage T7 RNApolymerase to direct selective high level expression of cloned genes. J.molec. Biol. 189, 113-130.

SUZUKA, I., DAIDOJI, H., MATSUOKA, M., KADOWAKI, K., TAKASAKI, Y.,NAKANE, P. K. AND MORIUCHI, T. (1989). Gene for proliferating-cellnuclear antigen (DNA polymerase 6 auxiliary protein) is present inboth mammalian and higher plant genomes. Proc. natn. Acad. Sci.U.SA. 86, 3189-3193.

TABOR, S. AND RICHARDSON, C. C. (1985). A bacteriophage T7polymerase/promoter system for controlled exclusive expression ofspecific genea. Proc. natn. Acad. Sci. U.S.A. 82, 1074-1078.

TAKASAKI, Y., FISHWILD, D. AND TAN, E. M. (1984). Characterization ofproliferating cell nuclear antigen recognized by autoantibodies inlupus sera. J. exp. Med. 159, 981-992.

TAN, C. K., CASTILLO, C, SO, A. G. AND DOWNEY, K. M. (1986). An

128 N. H. Waseem and D. P. Lane

auxiliary protein for DNA polymerase delta from fetal calf thymus. J.biol. Chem. 261,12310-12 316.

TAN, C. K., SULLIVAN, K., LI, X. Y., TAN, E. M., DOWNEY, K M. AND SO,A. G. (1987). Autoantibody to the proliferating cell nuclear antigenneutralizes the activity of the auxiliary protein for polymerase delta.Niicl. Acids Res. 15, 9299-9308.

TOWBIN, H., STAEHEUN, T. AND GORDON, J. (1979). Electrophoretictransfer of proteins from polyacrylamide gels to nitrocellulose sheets-procedure and some applications. Proc. natn. Acad. Sci. U.S.A. 76,4360-4364.

VERHEIJEN, R., KUUPKRS, H. J. H., SCHLJNOBMANN, R. 0., BOKHMBR, A.L. M., VAN DBIEL, R., BRAKBNHOIT, G. J. AND RAMAEKKRS, F. C. S.(1989). Ki-67 detects a nuclear matrix-associated proliferation-relatedantigen 1. Intracellular localization during interphase. J. Cell Sci. 92,123-130.

WALTON, T. H., MOEN, P. T. JR, FOX, E AND BODNAR, J. W. (1989).

Interaction of minute virus of mice and adenovirus with host nucleoli.J. Virol. 63, 3651-3660.

WONO, R. L., KATZ, M. E., OGATA, K., TAN, E. M. AND COHEN, S. (1987).Inhibition of nuclear DNA synthesis by an autoantibody toproliferating cell nuclear antigen/cyclin. Cell, lmmun. 110, 443-448.

ZUBER, M., TAN, E. M. AND RYOJI, M (1989a). Involvement ofproliferating cell nuclear antigen (cyclin) in DNA replication in livingcells. MoUc. cell. Biol. 9, 67-66.

ZUBER, M., YASUI, W., TAN, E. M. AND RYOJI, M. (19896). Quantitationand subcellular localization of proliferating cell nuclear antigen(PCNA/cyclin) in oocytes and eggs of Xenopus laevis. Expl Cell Res.182, 384-393.

(Received 23 October 1989 - AccepUd, in revised form, 31 January 1990)

Proliferating cell nuclcear antigen 129