Detection of Mouse Osteopontin by Western Blotting(1).pdf

Detection of Mouse Osteopontin by Western BlottingSusan R.Rittling1and Fei FengDepartment of Cell Biology and Neuroscience,Rutgers University,Piscataway,New Jersey 08854Received August 14,1998Several different antibodies to mouse and/or rat os-teopontin have been developed,and some antiseraraised against human osteopontin have been shown toreact with mouse osteopontin.We have taken advan-tage of the lack of osteopontin protein in mice with atargeted disruption of the osteopontin gene to charac-terize the reactivity and the specificity of several ofthese antibodies with mouse osteopontin by Westernblotting.Our results demonstrate that,with the excep-tion of the rat monoclonal antibody MPIIIB10,whichdoes not recognize mouse osteopontin on Westernblots,all the tested reagents do react with mouse os-teopontin,but their sensitivity varies widely,and insome cases there is significant cross-reactivity of theantibodies with other proteins found in mouse tissueextracts.1998 Academic PressThe secreted phosphoprotein osteopontin(OPN)isexpressed in a wide variety of tissues in mammals,asassessed by in situ hybridization and analysis ofmRNA levels(1,2).As a secreted molecule,however,the location of the protein itself may not always coin-cide with the site of synthesis.Immunohistochemicalexperiments have demonstrated the presence of theprotein in a variety of tissues,which in addition to bone(3,4),most notably include those with substantial pop-ulations of epithelial cells such as mammary gland,kidney,and gall bladder(1).The expression of theprotein is increased in a variety of pathologies,includ-ing tumorigenesis(57),a variety of renal syndromes(810)and in numerous forms of tissue injury(1114).The high level of expression of OPN in these situationsimplies that the protein is playing an important role,but its exact function remains unclear.We have devel-oped a strain of mice with a targeted disruption of theOPN gene(15).These animals cannot produce fulllength osteopontin protein,and so provide a valuablesystem for revealing the function of OPN in normal andabnormal physiology.To assess the level of OPN present at steady state inmouse tissues,we have used Western blotting to detectOPN in tissue extracts.Murine OPN has a predictedmolecular weight of 35,000 kDa(6)and is phosphory-lated and glycosylated(16).These modifications plusthe general acidic nature of the protein result in anom-alous migration of the protein on SDSPAGE gels.Inaddition,the apparent molecular weight of OPN canvary from 44 to 65 kDa depending on the tissue fromwhich the protein was prepared,and the exact runningconditions.Thus,on Western blots it can be difficult toidentify OPN,particularly if the immune reagents cancross react with proteins other than OPN in this mo-lecular weight range.We have taken advantage of the lack of osteopontinin mice with a targeted disruption of the OPN gene(Opn2/2mice)to assess the reactivity of different im-munologic reagents with murine osteopontin.In thework described here,we have assessed some of theconditions for analysis of OPN by Western blotting,and characterize the sensitivity and specificity of aseries of different antibodies with mouse OPN.Theseresults clarify and define the parameters of antibodyreactivity with mouse osteopontin,providing impor-tant information for those working with these tools inthe mouse system.MATERIALS AND METHODSProtein extracts.Tissue protein extracts were prepared as hasbeen described(5)briefly,frozen powdered tissues were suspendedin RIPA buffer,and the protein concentrations of the supernatantsdetermined.Bone extracts were prepared by extraction of powderedbone tissue in 6M guanidine,50 mM Tris,pH 7.5,0.5 M EDTA,for48 h as described(15).SDSPolyacrylamide gels.Proteins were separated on 12%SDSpolyacrylamide gels(29:1 acrylamide:N9,N9-bis-methylene-acrylamide)at pH 8.8 using a mini-gel format(Bio-Rad,Hercules,CA).Stackinggels were 4%polyacrylamide at pH 6.8.Running buffer was 3 g/literTris,14.4 g/liter glycine,1 g/liter SDS.Separation was at 180 V forabout 1 h.Transfer to various substrates was in 25 mM Tris,pH 8.3,192 mM glycine,20%MeOH at 100 V constant voltage for 1.5 h.PVDF1To whom correspondence should be addressed at Department ofCell Biology and Neuroscience,Division of Life Sciences,RutgersUniversity,Nelson Laboratories,604 Alison Rd.,Piscataway,NJ08854-8082.Fax:732-445-0104.E-mail:rittlingmbcl.rutgers.edu.BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS250,287292(1998)ARTICLE NO.RC9893062870006-291X/98$25.00Copyright 1998 by Academic PressAll rights of reproduction in any form reserved.membranes(Immobilon,Millipore,Bedford,MA)were wet in MeOHprior to use.Phosphate labeling of RAW cell OPN.RAW 264.7 transformedmacrophage cells(ATCC#TIB-71,(17)were cultured in DMEMcontaining 10%calf serum as described(18).Cells were grown to8090%confluence in 100 mm dishes,washed with PBS,and incu-bated with 300mCi of32PO4in 1.25 ml of PO42free DMEM at 37Cfor 3 h with occasional shaking.Conditioned medium was collectedand Na2HPO4(pH 7.4)added to 50 mM,and PMSF to 0.1 mg/ml.Free radioactivity was removed by desalting on a PD10 column(Pharmacia,Piscataway,NJ),and the radioactivity in individualfractions determined.3000 cpm of labeled protein were precipitatedwith barium citrate by addition of 1/10 volume of 15%barium chlo-ride followed by 1/10 volume of 0.2 M Na citrate.The resultingprecipitate was collected by centrifugation at 5000g for 10 min,andwashed with 15%barium chloride and with water.OPN was elutedfrom the precipitate by boiling for 2 min in SDS sample buffercontaining 0.2 M Na citrate prior to separation on 12%SDSPAGEgels.As indicated,eluted samples were mixed with 5mg kidneyextract.Western blotting.Following transfer,PVDF membranes wereblocked in 1%nonfat dry milk in TTBS(0.1 M Tris,pH 7.5;0.15 MNaCl;0.1%Tween 20)for 1 h at RT.Antibodies were diluted asindicated in 1%gelatin in TTBS,and incubated with blots for 1 h.Following washing with TTBS(5 3 3 min)horseradish-peroxidase(HRP)labeled secondary antibodies diluted in TTBS were added andincubated for 1 h.Antibody reactivity was visualized by enhancedchemiluminescence(Amersham,Chicago,IL),and blots were imme-diately exposed to XAR-5 film for 30 sec to 3 min.Blocking experi-ments were performed by incubating the indicated antiserum with 3mg GST-OPN for 1 h at RT in a total volume of 100ml 1%gelatin inTTBS.This material was then diluted to the indicated final concen-tration and incubated with the membranes as described above.Immunological reagents.Antiserum 199 was raised in goatagainst rat OPN purified from media conditioned by pup smoothmuscle cells(19).It was supplied as purified IgG at 23 mg/ml,andused at a dilution of 1:1500.The mouse monoclonal antibodyMPIIIB10was raised against a rat bone extract fraction.It wasobtained from the Developmental Studies Hybridoma Bank main-tained by the Department of Pharmacology and Molecular Sciences,Johns Hopkins University School of Medicine,and the Departmentof Biological Sciences,University of Iowa,under Contract N01-HD-2-3144 and used at a dilution of 1:1000.Antisera LF 123 and LF 124were raised in rabbits against COOH and NH2 fragments respec-tively of recombinant human OPN(20),and were used at dilutionsof 1:3000 and 1:750,respectively.Goat anti-mouse and anti-rabbitIgG labeled with horseradish peroxidase were from Bio-Rad andwere used at a dilution of 1:3000.HRP-labeled rabbit anti-goat IgGwas from Sigma(St.Louis,MO),and was used at 1:15,000.RESULTSOsteopontin is the major phosphoprotein secreted bycells,and is made at high levels by a number of trans-formed cells(21,22).The protein can be labeled effi-ciently by incubation of cells with32PO4.We used OPNlabeled in this way to assess the ability of the protein tobind to various substrates.Proteins secreted by thetransformed macrophage cell line RAW264.7 were la-beled with32PO4and enriched for osteopontin by bar-ium citrate precipitation.Bound proteins were sepa-ratedona12%SDSpolyacrylamidegel,andsubsequently transferred to different substrates.Whilethe labeled protein was readily detected in the driedgel(Fig.1A),very little radioactivity was detected afterthe proteins were transferred to nitrocellulose(Fig.1D).Transfer to either PVDF membranes or Zetaprobeon the other hand resulted in excellent recovery of theprotein on the membrane(Figs.1B and 1C),and thebinding of the protein to the membrane was not af-fected by the presence of other proteins in the sample(compare lanes 1 and 2).We conclude that OPN bindspoorly to nitrocellulose and that PVDF is the preferredsubstrate for Western blotting of this protein.We have gone on to assess the ability of severalanti-OPN antibodies to react with murine OPN onWestern blots.Various forms of OPN,purified or ex-tracted from mouse tissues,were separated and trans-ferred to PVDF membranes.Tissue extracts from wildtype and Opn2/2mice were examined in parallel.Therat monoclonal antibody MPIIIB10was raised against afraction of proteins extracted from rat bone,recognizesrat OPN(23),and has been suggested to react withmurine OPN.Our results demonstrate however thatthis antibody cannot recognize mouse OPN protein,atleast by Western blotting(Fig.2).Neither purifiedrecombinant mouse OPN(GST-OPN,Fig.2,lane 1)northe post-translationally modified form of OPN,such asthat extracted from bone(compare Fig.2,lane 3 andFig.3,lane 2)is detected by this antibody.MPIIIB10isa mouse monoclonal antibody so the secondary anti-FIG.1.Transfer of32PO4labeled OPN to different substrates.OPN secreted from RAW 264.7 cells was labeled by incubation of thecells with32PO4.This conditioned medium was precipitated withbarium citrate,and bound proteins were eluted and separated on12%SDSPAGE gels.One set of lanes was fixed and the dried gelexposed directly to film(A).The other three sets of lanes weretransferred to various substrates:PVDF(B);Zeta-probe(C)andnitrocellulose(D).For each panel,lane 1 is conditioned medium,200ml,precipitated with barium citrate,and eluted with SDSPAGEloading buffer containing 0.2 M citrate,and lane 2 is the samematerial as in lane 1,mixed with 5mg kidney extract.The lanelabeled M is14C-labeled molecular weight markers.The sizes of themarker bands in kDa is shown on the left.Vol.250,No.2,1998BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS288body used(HRP-labeled goat anti-mouse IgG)also de-tects endogenous murine IgG in the tissue extracts(bands labeled mIgG).In addition,in most tissue ex-tracts,an approx.90 kDa protein is recognized by thisantibody.This protein is not osteopontin,as deter-mined by comparison of extracts from wild-type ani-mals with those from mice with a targeted disruptionof the OPN gene(Fig.1,compare lanes 58).In addi-tion,preincubation of MPIIIB10with an excess of mouseGST-OPN failed to reduce the antibody reactivity withthis protein(data not shown).Additional evidence thatthis protein is not OPN derives from the failure of otheranti-OPN antibodies to detect a protein of this size(Figs.3 and 4).While MPIIIB10does recognize rat OPNpurified from bone(lane 2),in the normal rat kidneythe predominant band recognized is the 90-kDa speciesthat is not OPN;indeed OPN is not detected by thisantibody in this kidney extract.We have observed asimilar failure of several anti-OPN antibodies to detectOPN in normal kidney extracts by Western blotting(see below).The identity of the 90 kDa protein thatcross reacts with MPIIIB10is at present unknown.The antibody OP199 was raised in goat against ratOPN purified from rat pup smooth muscle cells(19).This antibody recognizes mouse OPN with excellentspecificity and sensitivity(Fig.3).A single major bandis detected by this antibody in several mouse tissues,with minor cross reactivity with other proteins in ex-tracts from the Opn2/2animals(lanes 3,5,7,and 9).OPN protein can be detected by this antibody in tumorextracts(lane 4),in bone(lane 2)and in extracts madefrom mammary glands two days following parturition(lane 8)(5).However,this antibody is unable to detectOPN in the kidney extracts made from normal,healthymice(compare lanes 6 and 7),and we conclude that inthe normal,nonpathologic kidney,the level of OPNprotein extractable from the tissue is below the level ofdetection of this assay.As observed previously(5),theOPN derived from tumors migrates more rapidly onthese gels than does purified OPN or that extractedfrom mammary glands.Antisera LF123 and LF124 were raised in rabbitsusing recombinant fragments of human OPN as an-tigen.The antigen used for LF123 was roughly theC-terminal half of the molecule,and that used forLF124 was the N-terminal half(20).When testedagainst mouse OPN by Western blotting,LF123,spe-cific for the C-terminus of OPN,reacts strongly andspecifically with mouse OPN derived from bone,tu-mors and mammary glands,and the sensitivity issimilar to that of antiserum OP199(Fig.4A,lanes 2,4,and 8).Only minor cross reactivity is observedwith proteins extracted from Opn2/2mice(lanes 3,5,7,and 9).That these species are not forms of OPN isdemonstrated by a failure of recombinant GST-OPNto block the reactivity of the antiserum with thesebands(Fig.4C).LF124 on the other hand,which was raised againstthe NH2terminal half of human OPN,shows a muchreduced sensitivity(Fig.4B,compare 4A and 4B,lanes1).In addition,this antiserum shows considerablecross reactivity with proteins that are not OPN.Theseinclude a species of approximately 50 kDa,which ispresent in kidney,mammary gland,and tumor ex-tracts,and to a much lesser extent,in bone,and aprotein of approximately the same size as OPN(about60 kDa)found exclusively in extracts from mouse bone(Fig.4B,lanes 2 and 3).Further confirmation thatthese proteins are not OPN comes from the failure ofrecombinant GST-OPN to block their detection byLF124(Fig.4D).While preincubation of the antiserumwith 3mg of GST-OPN protein reduces dramaticallythe intensity of the band detected in a preparation ofOPN from RAW 264.7 cells(Fig.4D,lanes 1 and 3),itFIG.3.Reactivity of antibody 199 with OPN in mouse tissues.Extracts of tissues from Opn1/1and Opn2/2mice were prepared asdescribed in methods.5mg of total tissue protein were separated on12%SDSPAGE,transferred and reacted with antibody OP199(1:1500).Lane 1:100 ng purified mouse OPN;lanes 2 and 3:extractsfrom bone;lanes 4 and 5:extracts from mouse mammary tumors;lanes 6 and 7:extracts from kidney;lanes 8 and 9:extracts fromlactating mammary gland.Lanes 2,4,6,and 8 are proteins extractedfrom Opn1/1mice,lanes 3,5,7,and 9 are proteins extracted fromOpn2/2mice.MW markers in kDa are shown on the left;the posi-tions of OPN are shown on the right.FIG.2.Reactivity of MPIIIB10antibody with mouse and rat tis-sue extracts.Proteins were separated on a 12%SDS polyacrylamidegel,transferred to PVDF membrane,and reacted with the MPIIIB10mouse monoclonal antibody at a dilution of 1:1000.Detection waswith horseradish peroxidase labeled goat anti mouse IgG,followedby ECL.Lane 1:275 ng mouse GST-OPN;lane 2:200 ng