A monoclonal antibody recognizes a human nuclear protein resembling

Xenopus oocyte nucleoplasmin

JANET LORD, GEOFFREY BROWN*, CHRISTOPHER BUNCE and PAUL NATHAN

Department of Immunology, University of Birmingham Medical School, Vincent Drive, Edgbaston, Birmingham BIS 2TJ, L'K

* Author for correspondence

Summary

The monoclonal antibody 10BG2 (IgG3) was de-rived from a mouse immunized with human preB cells. In immunofluorescence studies the anti-body revealed a human nuclear-associated deter-minant, which in interphase cells was entirelyrestricted to the nucleus. In metaphase cells10BG2 antigen was detected throughout the cyto-plasm with intensified staining at the peripheryof chromosomes. 10BG2 antibody stained all hu-man normal and transformed cells tested. Incontrast, the antibody did not stain mouse 3T3cells or Chinese hamster ovary (CHO) cells. Elec-trophoresis under denaturing and non-denaturingconditions revealed the 10BG2 antigen to be a130 X103 to lWxlO'M,. protein with a subunit

molecular weight of 295X103. This suggeststhe protein is at least a tetramer. On two-dimen-sional gels the 10BG2 antigen had a streakedappearance and separated into two isoelectricvariants (pi 5*2, 5*4). The protein was also shownto be phosphorylated and thermostable, andremained in solution at pH3-6. 10BG2 antigenwas highly soluble in aqueous buffers andco-migrated on non-denaturing gels with thenucleosome-assembly protein, nucleoplasmin,purified from Xenopus oocytes. The similarityof 10BG2 antigen to nucleoplasmin is discussed.

Key words: nucleoplasmin, nuclear protein, monoclonalantibody.

Introduction

The nucleosome represents the fundamental DNApacking unit in the eukaryotic chromosome. Conse-quently an understanding of nucleosome assembly andfunction is a prerequisite to a fuller understanding ofgene regulation and function. Fractionation of a cell-free nucleosome assembly system from eggs of Xenopuslaevis has led to the identification of a protein thatfacilitates nucleosome core assembly when histones andDNA are rapidly mixed in vitro (Laskey et al. 1978).The protein, which has been termed nucleoplasmin,represents approximately 10% of the total Xenopusoocyte nuclear protein and is present within the nu-cleus at similar concentrations to those at whichnucleosomes assemble in vitro (Krohne & Franke,1980a; Mills el al. 1980). Thus it is possible thatnucleoplasmin also plays a role in nucleosome assemblyin vivo. Nucleoplasmin purified fromXevopus oocyteshas a native molecular weight of 145xlO3A/r andis composed of five 29xl0 3 -30xl0 3 M r subunits

Journal of Cell Science 87, 713-722 (1987)Printed in Great Britain © The Company of Biologists Limited 1987

(Earnshaw et al. 1980) and is highly phosphorylated(Krohne & Franke, 1980«). A high content of poten-tially charged amino acids gives the protein an acidicisoelectric point (Mills et al. 1980). The protein isfurther characterized by being thermostable up to 80°Cand by its extreme solubility in aqueous buffers, and itwill remain in solution when the pH is lowered to 3-6with acetic acid (Mills et al. 1980).

Biochemical and immunohistological studies usingpolyclonal antisera raised against Xenopus nucleoplas-min have revealed an entirely nuclear restriction of theprotein (Krohne & Franke, 19806). Polyclonal antiserahave been used to demonstrate that nucleoplasmin isalso present in Xenopus, chicken, rodent and humansomatic cells (Krohne & Franke, 19806). However, inthese studies biochemical analyses were not used todemonstrate that the antigen recognized in these cellswas nucleoplasmin or some antigenically similar pro-tein. The availability of monoclonal antibodies reactivewith nucleoplasmin would permit more precise study

713

of its subcellular localization, its biochemistry, and itsrole in nucleosome assembly.

In this paper we describe the identification andcharacterization of the monoclonal antibody 10BG2,which recognizes a human nuclear protein with exten-sive biochemical similarities to nucleoplasmin. The10BG2 antibody has been used to investigate thesubcellular and tissue distribution of the nucleoplas-min-like protein.

Materials and methods

Cells and culturesThe following cell lines were maintained in RPMI 1640medium (Gibco, Paisley, Scotland, UK) supplemented with10% (v/v) foetal calf serum (FCS) (Gibco, Paisley, Scot-land, UK), penicillin (lOOunitsml"1) and streptomycin(50jJgmr') (Gibco UK, Paisley, Scotland): the promyeloidcell lines HL60 (Collins et al. 1977) and ML1 (Takeda et al.1982); the early erythroid cell line KS62 (Lozzio & Lozzio,1964); the B cell lines EB2 and ARH77 (Epstein el al. 1964;Burke* a/. 1978); and the epithelial cell line H.Ep2 (Toolan,19S4). A simian virus 40 (SV40)-transformed human foetalkeratinocyte cell line SV6-IBAM/HFK (provided by DrP. H. Gallimore, Department of Cancer Studies, Universityof Birmingham) was maintained in Dulbecco's modificationof Eagle's minimal essential medium (DME) supplementedwith 10% (v/v) FCS (Sera Laboratories, Crawley Down,UK). The keratinocyte cell line SCC2S (provided by DrE. K. Parkinson, Department of Cancer Studies, Universityof Birmingham) was maintained in DME supplemented with20% (v/v) FCS and 0-4jtgml~' hydrocortisone. Details ofculture conditions for primary cultures of fibroblasts (pro-vided by Dr A. M. R. Taylor, Department of CancerStudies, University of Birmingham), umbilical cord endo-thelial cells (provided by Ms C. Garner, Department ofInvestigative Pathology, University of Birmingham) andepidermal keratinocytes (provided by Dr E. K. Parkinson)have been described (Brown et al. 1985). Normal lympho-cytes were prepared from peripheral blood using Ficoll-Triosil (Pharmacia AB, Uppsala, Sweden). Cells werewashed three times in balanced buffered saline solution(BBSS) (Shortman et al. 1972) containing 0-1% bovineserum albumin (BSA) (Sigma, Poole, UK).

10BG2 monoclonal antibodyThe monoclonal antibody 10BG2 (IgG3) was derived from afusion experiment in which a Balb/c mouse was immunizedwith the pre-B cell line SMS/SB derived from a patient withacute lymphoblastic leukaemia. Mice were injected intraper-itoneally with 1X 107 cells and three weeks later boosted with1-5X 107 cells intraperitoneally and l x 107 cells intravenously.Three days later the spleen was removed for cell fusion.Hydridoma supernatants were screened by indirect immuno-fluorescence against air-dried acetone-fixed multispot prep-arations of the immunizing cell line. The 10BG2 antibody-stained nuclei of SMS/SB cells and was selected for detailed

analysis. Procedures for cell fusion, cloning and characteriz-ation of hybridoma cultures have been described in detail(Brown et al. 1981).

Detection of 10BG2-reactive nuclear antigenThe presence of 10BG2-reactive antigen within various celltypes was investigated by indirect immunofluorescence.Cytocentrifuge (Shandon cytocentrifuge II, Runcorn, UK)preparations of cells were obtained by centrifugation at500 revsmin"1 for 3 min and either: (1) thoroughly air driedand fixed in acetone at room temperature for 3 min; or (2)fixed for 30 min at room temperature in 3-7% formaldehydeprepared in 0 1 M-phosphate buffered saline, pH7-3 (PBS),and then rinsed with PBS; or (3) air dried and fixed inmethanol for 5 min at —20°C. Frozen sections of tonsil andmuscle, cut at 5 jim, were fixed in acetone for 10 min at roomtemperature prior to staining. 10BG2 antibody was applied astissue culture supernatant and revealed using fluorescein-conjugated sheep antibody to mouse immunoglobulin. Thestaining procedure has been described in detail (Brownet al. 1985). Slides were mounted in buffered glycerol(pH8-6) containing 25gI"1 l,4-diazobicyclo-(2,2,2)-octane(DABCO) to retard fading of fluorescence during microscopy(Johnson et al. 1982) and studied using a Leitz microscopeequipped with a HBO50 mercury burner for incident illumi-nation. Monoclonal antibodies that react with histone H2B(HBC-7 (Turner, 1982)), DNA (33; a gift from Drs A.Morgan and N. Staines, Department of Immunology, KingsCollege, London) and a nuclear envelope protein (AGF 2 3 ;Brown et al. 1985) were used as controls. Nuclei isolatedfrom HL60 cells by using 2% Tween 40 to disrupt the cellmembrane (Standring & Williams, 1978) were also incubatedin suspension with 10BG2 antibody and the above controlantibodies. Fluorescein-labelled sheep antibody to mouseimmunoglobulin was used as second antibody. 10BG2 anti-body binding to isolated HL60 nuclei, which had been fixedin 0-25 % glutaraldehyde, was assessed using an indirectradioactive binding assay and 125I-labelled (Fab')2 rabbitanti-mouse immunoglobulin (5^Ci/ig~') as the second anti-body (Brown et al. 1986). The fixation procedure andbinding assay as used with whole cells have been described indetail (Morris & Williams, 1975).

Immunoprecipitation, electrophoresis andimmunostainingHL60 cell proteins were internally radiolabelled with[35S]methionine as described (Browne/ al. 1985). [35S]meth-ionine-labelled extracts were also prepared from the cells ofthe murine plasmacytoma cell line NSI, which do not expressthe 10BG2 antigen. These extracts were used as controls inimmunoprecipitation experiments using 10BG2 antibody.HL60 cells were also labelled with 32P to determine whetherthe 10BG2 antigen was phosphorylated. Cells were washedtwice in PO4-free Locke's solution supplemented with 5 mM-Hepes (pH7-4), glucose (Znigml"1) and BSA (2mgml~')and preincubated in this buffer at 3x l0 6 ml~ ' for 30min at37°C. 32PO4 ( lOmCimr1) was added to the incubation flaskto a final concentration of 200^Ciml~' and the incubationcontinued for 3 h. Cells were then washed twice in ice-coldPCvfree Locke's solution containing 5 mM-Hepes (pH7-4).The cell pellet was taken up in ice-cold 10BG2 extraction

714 J.Lordetal.

buffer (20 mM-Tris • HC1, pH 8-0, containing 0-2 mM-EGTA,5mM-NaF, 1 mM-phenylmethylsulphonyl fluoride (PMSF)and 0-1 % Triton X-100), pipetted vigorously and the super-natant was collected using a Beckman microfuge and stored at—20°C. The 32P-labelled proteins were analysed by non-denaturing gel electrophoresis and were also used in immuno-precipitation experiments. The procedure for immuno-precipitation of Triton-soluble proteins has been described indetail (Brown et al. 1985).

Triton-soluble extracts were also prepared from unlabelledcells as follows. Exponentially grown cells (1X107) werewashed twice in cold BBSS and the pellet taken up in 100^tlof 10BG2 extraction buffer. Twenty five units of micro-coccal nuclease (Sigma, Poole, UK) were added and thecells incubated for 5 min at 37°C, after which the samplewas placed on ice and 400/JI of an appropriate electro-phoresis sample buffer was added. For non-denaturing gelelectrophoresis sample buffer contained 012mM-Tns-HCI,pH 6-7, 20 % glycerol and 0-0025 % Bromophenol Blue (Bio-Rad, Watford, UK). For electrophoresis of samples underdenaturing conditions 2-5 % SDS and 10% /3-mercaptoetha-nol (Bio-Rad, Watford, UK) were also added to the samplebuffer. Extracts taken up in denaturing sample buffer wereincubated in a boiling water bath for 5 min. Insolublematerial was removed from all electrophoresis samples usinga Beckman microfuge. Electrophoresis of cell extracts in gelscontaining 4—12% polyacrylamide was performed as de-scribed by Laemmli (1970) with the omission of SDS fromboth the gels and buffers in the case of non-denaturingelectrophoresis. For immunoblotting proteins were electro-phoretically transferred from non-denaturing gels onto nitro-cellulose filters (BA 85, Schleicher and Schuell, Dassel, W.Germany) by the procedure of Towbin et al. (1979), butexcluding methanol from the transfer buffer. Nitrocellulosefilters were then immunostained for the presence of 10BG2antigen using a 1Z5I-labelled (Fab')2 rabbit anti-mouse im-munoglobulin as the second antibody. Xenopus oocyte nu-cleoplasmin (kindly supplied by Dr M. A. Sheehan), whichhad been affinity purified using a specific mono-clonal antibody (PA1C2), was electrophoresed on non-denaturing polyacrylamide gels in parallel with Triton ex-tracts of HL60 cells. After blotting the proteins onto nitrocel-lulose filters, individual lanes were cut and immunostainedusing monoclonal antibodies specific for Xenopus nucleoplas-min (PB2E11 and PA3C5, kindly provided by Dr S. Dil-worth) or 10BG2 antibody. Two-dimensional gel electro-phoresis was also performed on a variety of extracts usingprepoured isoelectric focusing gels (LKB Ampholine PAGPlates, PH range 3-5—9-5 and 4 0 - 6 5 ) in the first dimensionand 10 % SDS—polyacrylamide gels in the second dimension.

Partial purification of 10BG2 antigenIn order to assess the stability of the 10BG2 antigen to heatand acid, the antigen was partially purified on a sucrosegradient. Nuclei prepared from 10 HL60 cells wereextracted for 10BG2 antigen as described above. A 1 mlsample of the extract was loaded onto a 13 ml sucrose gradient(10% to 40%) and spun at 37 500 revs min"1 (MSE65ultracentrifuge, 6x14 ml swing-out rotor) for 165h at 4°C.Fractions (05 ml) were collected from the sucrose gradientand were analysed for the 10BG2 antigen by inhibition of

10BG2 antibody binding to isolated nuclei as assessed by theindirect radioactive binding assay (Brown et al. 1986).Proteins of known molecular weight and sedimentation co-efficient were applied to duplicate gradients in order todetermine the sedimentation co-efficient of the 10BG2 anti-gen. The proteins used were: haemoglobin, 4 6 S ; urease,8-5 S and 17 S; and catalase 11-2S. All the gradient fractionswere heated for 10 min at 80°C, following the addition of0025 vol. of PMSF ( 6 m g m r 1 in ethanol), to assess ther-mostability. Acid stability was determined by lowering thepH of the gradient fraction to 3 6 with acetic acid. 10BG2andcontrol (RC2.43) immunoprecipitates of 3SS-labelled HL60proteins were treated in the same way. Precipitated materialwas removed after both treatments using a Beckman micro-fuge, the pH of the acid-treated fractions was returned toneutral with 3M-Tris'HCl, pH8-8. Treated and untreatedextracts were analysed on SDS-polyacrylamide (10%) gels.

Results

Subcellular and tissue distribution of the J0BG2antigen

Immunofluorescence studies using cytocentrifugepreparations of HL60 cells fixed with acetone, formal-dehyde or methanol showed that in interphase cells10BG2 antigen was located entirely within the nucleus.As shown in Fig. 1A, in which formaldehyde prep-arations were stained, the staining pattern was mottledin appearance and nucleoli were unstained. Gradualfocusing adjustments using a X95 objective lens re-vealed that the staining was distributed throughout thenucleus. Similar results were obtained with acetone-and methanol-fixed cytocentrifuged cells. Metaphasecells showed intense diffuse cytoplasmic staining(Fig. IB). The chromosomes of cells in metaphasewere unstained though there was increased fluor-escence at the periphery of individual chromosomes.

In contrast to monoclonal antibodies to histone 2B,DNA and a nuclear envelope protein, 10BG2 antibodyfailed to stain air-dried and unfixed cytocentrifugepreparations of HL60 cells. Air-dried cytocentrifugepreparations of cells that had been pre-incubated withBBSS for 10 min at room temperature, prior to fixationin acetone, were unstained by 10BG2 antibody. HL60nuclei isolated using Tween 40, were also unreactivewith 10BG2 antibody when stained in suspension, butwere intensely stained by monoclonal antibodies toDNA. In contrast, 10BG2 antibody bound well toisolated HL60 nuclei that had been fixed in 0-25%glutaraldehyde, giving l l x 103ctsmin~' per 106 nucleiin indirect radiobinding assays over a background ofl-5xl03ctsmin~' per 106 nuclei. Glutaraldehyde-fixed nuclei were also reactive with antibodies to DNAand histone 2B. These results suggest that the 10BG2antigen can readily diffuse out of the nucleus unlessimmobilized by fixation. The possibility that 10BG2

Nucleoplasmin-like protein in human cells 715

antibody can only recognize a protein when altered byfixation is excluded by the negative result obtainedwhen cytocentnfuged cells were preincubated withPBS and subsequently fixed in acetone prior to stainingwith 10BG2 antibody and also by precipitation studies,which showed that the antibody was able to recognizesoluble antigen (see below). As shown in Table 1, allhuman cell lines tested and cells from primary culturesof fibroblasts, endothelial cells and keratinocytes wereintensely stained by 10BG2 antibody. 10BG2 stainedlymphocytes and monocytes in samples of blood mono-nuclear cells. All lymphoid areas of cryostat sections oftonsil were stained together with fibroblasts in areas ofconnective tissue and tonsil epithelial cells. The nucleiof muscle cells were strongly stained. In contrast,10BG2 antibody did not stain Chinese hamster ovarycells (CHO-Kl) and Balb/C 3T3 fibroblasts. Thus to

date the 10BG2 antibody appears to recognize ahuman-specific determinant.

The biochemical nature of the 10BG2 antigenThe native molecular weight of the 10BG2 antigen wasdetermined using Ferguson plots (Hedrick & Smith,1968). For this analysis carbonic anhydrase (30X103

iV/r), ovalbumin (45xlO3./V/r), bovine serum albumin(66xl03yV/r) and phosphorylase B (185XlO3A/r) wereelectrophoresed in non-denaturing gels containing 4, 6,8, 10 and 12% polyacrylamide. The retardation co-efficients (A"r) obtained for these proteins were plottedas a function of molecular weight (see Fig. 2A). Theresultant line was then used to derive the molecularweight of the 10BG2 antigen using KT values obtainedin parallel. In four independent determinations, one ofwhich is shown in Fig. 2A, the 10BG2 antigen was

Fig. 1. Expression of the 10BG2antigen by cells during interphaseand metaphase. Cytocentrifugepreparations of the humanpromyeloid cell line HL60 werefixed in formaldehyde and stainedwith 10BG2 antibody by indirectimmunofluorescence as described inMaterials and methods. The nuclearrestriction of the 10BG2 antigenduring interphase and its absencefrom nucleoli are clearly shown (A).10BG2 antigen was expressedthroughout the cytoplasm of cells inmetaphase (B). In these cellschromosomes were not stained by10BG2 antibody but the staining wasintensified at their periphery. X950.

716 J.Lordetal.

Table 1. Reactivity of J0BG2 antibody with variouscell types

Cell type

Human haemopoietic cells

Normal cells:Blood lymphocytes, monocytesBlood crythrocytesTonsil lymphocytesf

Cell lines:HL60 (myeloid)KS62 (early erythroid)EB2, ARH-77 (B cell)SMS/SB (pre-B)

Human non-haemopoietic cells

Normal cells:Tonsil epithelium')'Tonsil connective tissucfMuscle ccllsj

Primary cell cultures:KcratinocytcsUmbilical cord endothelial cellsEmbryo lung fibroblasts

Cell lines:H.Ep2 (carcinoma of larynx)SCC25 (squamous cell carcinoma)SV6-1BAM/HFK (SV40-transformed

foetal keratinocytes)

% Positivenuclei*

>99<1

>99

>99>99>99>99

>99>99>99

>99>99>99

>99>99

>99

Non-human cell lines

CHO (Chinese hamster ovary) <13T3 fibroblasts <1

• Cytocentnfuged cell lines were fixed with acetone and stainedusing an indirect immunofluorescence.

t Studied using cryostat sections of tissue.

calculated to have a native molecular weight in therange from 130X103 to 14OX103.

10BG2 antibody failed to bind to Triton X-100-soluble HL60 proteins when electrophoresed in de-naturing gels and blotted onto nitrocellulose. There-fore, the subunit molecular weight of the 10BG2antigen was determined using protein immunoprecipi-tated with 10BG2 antibody (see Fig. 2B). Precipitatesfrom HL60 cells gave a single sharp band of29-5xl03iWr, which was not seen in control immuno-precipitates prepared either with an irrelevant mono-clonal antibody or with 10BG2 and an extract frommouse NSI cells, which do not express 10BG2 antigen.It was noted that 10BG2-precipitated proteinaccounted for approximately 1-7% of the total 3SSactivity in extracts from whole HL60 cells.

Fig. 3 shows that the 10BG2 activity purified on thesucrose gradient corresponded to a major protein bandon SDS-polyacrylamide gels of MT 32x 103. From twodeterminations the sedimentation co-efficient of theprotein was estimated to be 9-2 S. The sucrose gradient

fractions were also assessed for heat and acid stabilityand the 32xlO3A7r protein band shown in Fig. 3 wasstable to both treatments (data not shown). The valuefor molecular weight of the monomer here is slightlyhigher than that gained in earlier experiments and mayreflect a change in the proteins used as standards. Thisis not an uncommon problem and has been commentedon by other authors (Dingwall et al. 1982).

The isoelectric point of 10BG2 antigen was deter-mined using the partially purified protein obtainedfrom the sucrose gradient. All fractions containing10BG2 activity were subjected to isoelectric focusing;Fig. 4 shows a two-dimensional gel of one of the peakfractions. The proteins indicated by arrows were theonly ones found in each 10BG2 fraction and in additionwere the only proteins stable to heat and acid treat-ments. Furthermore, a band containing 10BG2 ac-tivity, as identified by immunoblotting, was excisedfrom a non-denaturing gel and the protein eluted fromthe gel was separated by two-dimensional electrophor-esis. The gel contained proteins that electrophoresed inthe same position as those indicated in Fig. 4. We aretherefore sure that the 10BG2 antigen contains isoelec-tric variants, as indicated in Fig. 4, with pi values of5-2 and 5-4. The streaked appearance of the 10BG2antigen is very similar to that of nucleoplasmin and thepresence of the pentameric form of the protein in thesecond dimension reflects a degree of resistance to theaction of SDS. This resistance was also occasionallyseen in denaturing gels of 10BG2 immunoprecipitates;Fig. 5 (lanes C and E) demonstrates the multibandpattern sometimes seen, which may represent thedifferent polymeric forms of the 10BG2 antigen. Incontrast Fig. 2B and Fig. 5, lane F, show that themonomer alone can be produced and imply thatthorough treatment of samples with SDS is required toensure complete denaturation of the protein to itsmonomeric form.

Fig. 5 shows that the 10BG2 antigen is phosphoryl-ated as a phosphorylated protein from Triton X-100extracts of 32P-labelled HL60 cells (Fig. 5, lane B)runs parallel, in a non-denaturing gel, to the proteinrevealed by immunoblotting with 10BG2 antibody(Fig. 5, lane A). Immunoprecipitates of Triton X-100extracts from S-labelled HL60 cells were stable toheat (lane E) and acid (lane F) treatments. No labelledproteins were seen in immunoprecipitates produced byan irrelevant antibody to rotavirus (RC 2.43; Fig. 5,lane D).

The properties of the 10BG2 antigen bear a strikingresemblance to those of the nucleosome-assembly pro-tein nucleoplasmin. This similarity was confirmed bycomparison of the electrophoretic mobility of the10BG2 antigen with that of nucleoplasmin purifiedfromXenopus oocytes. As shown in Fig. 5 (lanes G andH), immunostaining of nitrocellulose blots for the

Nucleoplasmin-like protein in human cells 717

presence of Xenopus nucleoplasmin and 10BG2 antigenrevealed that both proteins had migrated the samedistance on non-denaturing, 7-5% polyacrylamidegels. Similarly, nucleoplasmin and 10BG2 co-migratedin separate non-denaturing gels containing 6% and10% polyacrylamide. These data confirm that 10BG2antigen has a similar molecular weight to Xenopusnucleoplasmin and also reveal that the proteins carry asimilar charge. 10BG2 antibody did not cross-reactwith Xenopus nucleoplasmin (data not shown) andsimilarly the PA3C5 and PB2E11 antibodies toXenopus nucleoplasmin did not bind to proteins inTriton extracts of HL60 cells (data not shown). Fur-thermore, in immunofluorescence studies none of theanti-nucleoplasmin monoclonal antibodies was seen tobind to acetone-fixed HL60 cytocentrifuge prep-arations.

Discussion

Immunofluorescence studies showed that the 10BG2monoclonal antibody recognized a protein that ininterphase cells was entirely restricted to the nucleus.In these cells the 10BG2 staining was mottled inappearance and distributed throughout the nucleus,with the exception that nucleoli were completely

unstained. In metaphase cells 10BG2 antigen was seenthroughout the cytoplasm but was concentrated at theperiphery of chromosomes, though the chromosomesthemselves were not stained. Similar staining patternswere seen in all human cells tested. These includednormal haemopoietic cells and cell lines, normal andtransformed epithelial cells, fibroblasts and keratino-cytes. The presence of 10BG2 antigen in the nuclei ofsuch diverse cells suggests a key housekeeping functionfor this protein. In contrast, 10BG2 antibody failed tostain either mouse 3T3 fibroblasts or Chinese hamsterovary (CHO) cells.

Biochemical analyses of the 10BG2 antigen revealedextensive similarities with nucleoplasmin, a nuclearprotein that was first identified in Xenopus laevisoocytes. Nucleoplasmin is a 145xl03yV/r pentamercomposed of 29X103 to 30 X 103Mr subunits (Earnshawet al. 1980). Ferguson plot analysis of the 10BG2antigen revealed a native molecular weight in the range130X103 to 140X103, while electrophoresis of 10BG2immunoprecipitates from HL60 cells demonstrated asubunit of 29-5xlO3A/r. Thus, the native and subunitsizes of nucleoplasmin and 10BG2 antigen are similar.It is of particular significance that the relative valuesof the native and subunit molecular weights of the10BG2 antigen suggest that this protein shares withnucleoplasmin a highly unusual pentameric structure.

< 200

< 116-3

* 02-5

4 68

« 4 5

2 9 5 •

50 200150139-5 ^ ^

Molecular weight (x 10"3)

Fig. 2. Determination of native and subunit molecular weights for the 10BG2 antigen. A. The native molecular weight ofthe 10BG2 antigen was obtained by comparison of its retardation co-efficient (A"r) in non-denaturing gels (4-12%polyacrylamide), with those of carbonic anhydrase ( • ) , ovalbumin (O), bovine serum albumin (A) and phosphorylase B(A) . In the experiment shown the A"r of the 10BG2 antigen was found to be 14-7, corresponding to a relative molecularweight of 139-5X103. B. Triton X-100 extracts prepared from HL60 cells cultured in medium containing [35S]methioninewere immunoprecipitated with 10BG2 antibody to determine the subunit molecular weight of the 10BG2 antigen (lanes 1and 2) and also with an irrelevant monoclonal antibody to rotavirus (RC2.43) as a control (lanes 3 and 4). As shown inlanes 1 and 2, electrophoresis of 10BG2 immunoprecipitates in denaturing gels revealed a single band corresponding to asubunit molecular weight of 29-5X 103 and no band was seen in control immunoprecipitates.

718 J.Lordetal.

100

I 50

17S

4II 2S 8-2S 4-6S

10 15 20

Fraction number

25

Fig. 3. Partial purification of 10BG2 antigen as a sucrose gradient. A 1 ml sample of a Triton X-100 extract of 109 HL60nuclei was separated on a 13ml sucrose gradient (10% to 40%) by centrifugation at 37 500 revs min"1 for 16-5 h at 4°C.Samples (05 ml) were removed and assayed for 10BG2 activity (Brown et al. 1986). Samples of each fraction were alsoapplied to SDS-polyacrylamide (10%) gels, which were subsequently stained with Coomassie Blue. The peak of 10BG2activity corresponded to a major protein band in the acrylamide gel; the gel strip is shown below the sucrose gradientprofile. Numbers above the sucrose gradient indicate the position of marker proteins in a duplicate sucrose gradient. Fromthese values the sedimentation coefficient of 10BG2 antigen was estimated.

HX1O~3

140 •

4•

PH6•

t

j

Fig. 4. Separation of 10BG2 antigen fromsucrose gradient by two-dimensionalelectrophoresis and estimation of isoelectricpoint. Peak fractions from the sucrosegradient were concentrated and 60 )x\ sampleswere loaded onto prepoured isoelectricfocusing gels (LKB PAG Plates). Proteinswere electrophoresed for 1 -5 h at 15kV,50 mA and 30 W at 10°C on a multiphor IIapparatus (LKB). Strips were cut out of thegel, soaked in SDS reducing buffer (62-5 mM-Tr i sHCl , pH6-8, 5 % /3-mercaptoethanol,3 % SDS, 001 % Bromophenol Blue) for10 min at room temperature then applied tothe second dimension gel, a S D S -polyacrylamide (10%) gel. The second-dimension gel was subsequently stained forproteins with Coomassie Blue. Althoughseveral proteins were seen in these gels, thosethat were stable to heat and acid (shown byarrows) had isoelectric points of 5-2 and 5-4.

Nucleoplasmin-like protein in human cells 719

A B D G H

HX1(T3

200 •

97 •

t68 •

45 •

30 •

I

Fig. 5. Biochemical characteristics of 10BG2 antigen. Lane A: Triton X-100 extract of HL60 cells electrophoresed on a7-5% non-denaturing gel and immunoblotted with 10BG2 antibody. 10BG2 antibody bound to nitrocellulose filters wasrevealed by a l2JI-labelled rabbit (Fab')2 to mouse Ig. Lane B: Triton X-100 extract of HL60 cells, labelled with 32P,electrophoresed on a 7 5 % non-denaturing gel and revealed by autoradiography. Lanes A and B are from the same gel andshow that the 10BG2 antigen (lane A) runs parallel with a phosphorylated protein in the Triton X-100 extract (lane B).Lanes C-F : heat and acid stability was assessed using Triton X-100 extracts of HL60 cells labelled with [35S]methionine,immunoprecipitated with 10BG2 antibody (lane C), and also with an irrelevant antibody to rotavirus (RC2.43) (lane D).10BG2 immunoprecipitates were heated to 80cC for lOmin in the presence of 0-025 vol. of PMSF (6mgml~' in ethanol) toaid precipitation of proteins. Precipitated proteins were removed by centrifugation and proteins in solution wereelectrophoresed in SDS-polyacrylamide (10%) gel (lane E). 10BG2 immunoprecipitate were also treated with acetic acid tolower the pH to 3 6 . Precipitated proteins were removed by centrifugation and the remaining supernatant was readjusted toneutral pH with Tris buffer prior to electrophoresis in a SDS-polyacrylamide (10%) gel (lane F). Lanes G, H: purifiedXenopus nucleoplasmin and Tri ton X-100 extracts of HL60 were electrophoresed in adjacent lanes of a 7-5 %polyacrylamide gel under conditions that would not denature the proteins (see Materials and methods). Lane G was loadedwith 120 Hg of HL60 extract and lane H with 0-5^g of Xenopus nucleoplasmin. The proteins were electrophoreticallytransferred to nitrocellulose and incubated with 10BG2 antibody and antibodies to Xenopus nucleoplasmin (PA3C5 andPB2E11), respectively. Binding of antibodies to the filters was subsequently revealed using a 12SI-labelled rabbit (Fab')2 tomouse Ig.

Furthermore, since Xenopus nucleoplasmin and10BG2 antigen co-migrated under electrophoresis innon-denaturing polyacrylamide gels the proteins mustalso carry a similar charge. In such gels both nucleo-plasmin and 10BG2 gave broad diffuse bands. Thismay be due to the presence of several different phos-phorylated forms of the proteins, which have alreadybeen described for Xenopus nucleoplasmin (Krohne &Franke, 19806). We have shown that the 10BG2

antigen is phosphorylated and separates into two iso-electric variants with pi values of 5-2 and 5-4. This mayexplain the broad band routinely seen on non-de-naturing gels of Triton X-100 extracts revealed byimmunoblotting with 10BG2 antibody. In addition10BG2 antigen gave two spots with a streaked appear-ance on two-dimensional gels as does nucleoplasmin.Other characteristics shared by 10BG2 antigen andnucleoplasmin are their stability to heat and acid

720 J. Lord et al.

conditions (Mills et al. 1980). Nucleoplasmin has beenfurther characterized by its high solubility in aqueousbuffers (Krohne & Franke, 19806). In immunofluor-escence studies, 10BG2 antibody failed to stain cyto-centrifuge preparations of HL60 cells that had beenpreincubated in buffer prior to the staining procedure.Similarly, isolated HL60 nuclei were unreactive with10BG2 antibody when stained in suspension. If, how-ever, 10BG2 antigen was immobilized by fixation ofnuclei immediately after their preparation then thebinding of 10BG2 antibody was readily demonstratedusing a radiobinding assay. In addition 10BG2 antigenwas demonstrated, by electrophoresis and subsequentimmunostaining, in the supernatant of isolated HL60nuclei that had been incubated in buffer (data notshown). Thus, several approaches have shown that,like nucleoplasmin, 10BG2 antigen is readily extract-able in aqueous buffers. Finally, the 10BG2 antigenshows a degree of resistance to SDS treatment, whichcan lead to the pentomeric and monomeric forms of theprotein being detected on SDS-polyacrylamide gels of10BG2 immunoprecipitates. Such a banding patternhas also been demonstrated for nucleoplasmin (Ding-wall et al. 1982).

In conclusion, the 10BG2 antibody recognizes anapparently human-specific determinant expressed on anuclear restricted protein. All the biochemical charac-teristics of this protein that have been investigated aresimilar to those of nucleoplasmin. Thus, it is likely thatthe protein recognized by 10BG2 antibody is humannucleoplasmin. It must, however, be emphasized thatnucleoplasmin is defined by its ability to assemblehistones and DNA into nucleosomesm vitro (Laskey etal. 1978). Consequently, to establish conclusively thatthe 10BG2 antigen represents a human nucleoplasminanalogue will require the demonstration of this activity.The 10BG2 antibody provides a means of purifyingsufficient protein for sequence analyses and compari-son with nucleoplasmin from Xenopus oocytes. If the10BG2 antigen is nucleoplasmin, it is interesting toconsider why the antibody does not stain nucleoli.Several reports in the late 1970s demonstrated thatunlike the majority of eukaryotic DNA, transcrip-tionally active nucleolar chromatin is not organized intonucleosome-like structures (see review by Franke et al.1979). Thus it is possible that a nucleoplasmin-likeprotein is not required in the nucleolus.

To our knowledge 10BG2 is the first monoclonalantibody to a human nucleoplasmin-like protein. Thisantibody has already proved valuable in the biochemi-cal characterization of the protein and will be usedfurther to study its function in human nuclear organiz-ation.

This work was supported by grants from the LeukaemiaResearch Fund and the Medical Research Council, UK. We

thank Professor R. A. Laskey for his interest and adviceduring the development of this study and preparation of themanuscript. We are grateful to Miss Diane Parchment forsecretarial assistance and Wayne Gill for photography.

References

BROWN, G., BUNCE, C. M., JOHNSON, G. D. & LORD, J.

M. (1986). Expression of a 215,000-Dalton nuclearenvelope protein decreases during cell maturation.Leukaemia Res. 10, 1175-1182.

BROWN, G., KOURILSKY, F., FISHER, A., BASTIN, J. &

MACLENNAN, I. (1981). Strategy for screening formonoclonal antibodies against cellular antigens expressedon minor cell populations or in low amounts. HumanLymphocyte Differ. 1, 167-182.

BROWN, G., TURNER, B., MORRIS, C , BAHMAN, A.,

FISHER, A., WHITFIELD, W., DAV'IES, S., BARTHAKUR, R.

& JOHNSON, G. (1985). A monoclonal antibody identifiesa 215,000 Dalton nuclear envelope protein restricted tocertain cell types. Eur.J. Cell Biol. 39, 86-%.

BURK, K., DREWINKO, B., TRUJILLO, J. & AHEARN, J.

(1978). Establishment of a human plasma cell line invitro. Cancer Res. 38, 2508-2513.

COLLINS, S., GALLO, R. & GALLAGHER, R. (1977).

Continuous growth and differentiation of human myeloidleukaemic cells in suspension culture. Nature, Lond. 270,347-349.

DINGWALL, C , SHARNICK, S. V. & LASKEY, R. A. (1982).

A polypeptide domain that specifies migration ofnucleoplasmin into the nucleus. Cell 30, 449-458.

EARNSHAW, W., HONDA, B., LASKEY, R. & THOMAS, J.

(1980). Assembly of nucleosomes: the reaction involvingA', laevis nucleoplasmin. Cell 21, 873-883.

EPSTEIN, M., BARR, Y. & ACHONG, B. (1964). A second

virus-carrying tissue culture strain (EB2) of lymphoblastsfrom Burkitt's lymphoma. Path. Biol., Pans 12,1233-1234.

FRANKE, W., SPRING, H., TRENDELENBURG, M. &

ZENTGRAF, H. (1979). Organisation of nucleolarchromatin. In The Cell Nucleus VII (ed. H. Busch), pp.49-91. London, New York: Academic Press.

HEDRICK, J. & SMITH, A. (1968). Size and charge isomerseparation and estimation of molecular weights ofproteins by disc gel electrophoresis. Archs Biochem.Biophys. 126, 155-164.

JOHNSON, G., DAVIDSON, R., MCNAMEE, K., RUSSEL, G.,

GOODWIN, D. & HOLBORROW, E. (1982). Fading ofimmunofluorescence during microscopy: a study of thephenomenon and its remedy, jf. hninuu. Meth. 55,231-242.

KROHNE, G. & FRANKE, W. (1980O). Immunological

identification and localization of the predominant nuclearprotein of the amphibian oocyte nucleus. Proc. natn.Acad. Sci. U.SA. 11, 1034-1038.

KROHNE, G. & FRANKE, W. (19806). A major solubleacidic protein located in nuclei of diverse vertebratespecies. Expl Cell Res. 122, 167-189.

Nucleoplasmin-like protein in human cells 721

LAEMMLI, U. (1970). Cleavage of structural proteins duringthe assembly of the head of bacteriophage T4. Nature,Land. 227, 680-685.

LASKEY, R., HONDA, B., MILLS, A. & FINCH, J. (1978).

Nucleosomes are assembled by an acidic protein whichbinds histones and transfers them to DNA. Nature,Loud. 275, 416-420.

Lozzio, C. B. & Lozzio, B. B. (1964). Human chronicmyelogenous leukemic cell line with positive Philadelphiachromosomes. Blood 45, 321-334.

MILLS, A., LASKEY, R., BLACK, P. & ROBERTIS, E. (1980).

Acidic protein which assembles nucleosomes in vitro isthe most abundant protein in Xenopus oocyte nuclei. .7.molec. Biol. 139, 561-568.

MORRIS, R. & WILLIAMS, A. (1975). Antigens on mouseand rat lymphocytes recognised by rabbit antisera againstrat brain. Eur.J. Immun. 5, 274—281.

SHORTMAN, K., WILLIAM, W. & ADAMS, P. (1972). The

separation of different cell classes from lymphoid organs.V. Simple procedures for the removal of cell debris:damaged cells and erythroid cells from lymphoid cellsuspensions, jf. Immun. Meth. 1, 273-287.

STANDRING, R. & WILLIAMS, A. (1978). Glycoproteins andantigens of membranes prepared from rat thymocytesafter lysis by shearing or with the detergent tween-40.Biochim. biophys. Ada 508, 85-96.

TAKEDA, K., MINOWADA, J. & BLOCH, A. (1982). Kinetics

of appearance of differentiation-associated characteristicsin ML-1, a line of human myeloblastic leukaemia cells,after treatment with 12-O-tetradecanoylphorbol-13-acetate, dimethylsulfoxide or 1-beta-D-arabinofuranosylcytosine. Cancer Res. 42, 5152-5158.

TOOLAN, H. (1954). Transplantable human neoplasmsmaintained in cortisone-treated laboratory animals.Cancer Res. 14, 660-666.

TOWBIN, H., STAEHELIN, T. & GORDON, J. (1979).

Electrophoretic transfer of proteins from polyacrylamidegels to nitrocellulose sheets: procedure and someapplications. Proc. natn. Acad. Sci. U.SA. 76,4350-4354.

TURNER, B. (1982). Immunofluorescent staining of humanmetaphase chromosomes with monoclonal antibody tohistone H2B. Chromosoma 87, 345-357.

{Received 10 March 1987 - Accepted 7 April 1987)

722 J. Lord et al.