[CANCER RESEARCH 38, 6-12, January 1978]

Physicochemical Approach to the Purification of Human «,-Fetoproteinfrom the Ascites Fluid of a Hepatoma-bearing Patient1Phil Gold,2 Aurora Labitan, H. C. George Wong, Samuel 0. Freedman, John Krupey, and Joseph Shuster3

Division of Clinical Immunology and Allergy /P. G., S. O. F.. J. S.¡and McGill University Medical Clinic, Montreal General Hospital Research Institute [P.G., A. L., H. C. G. W., S. O. F., J. K., J. S.¡,Montreal, Quebec H3G 1A4, Canada

ABSTRACT

A method for the purification of human o.-fetoproteinfrom the ascites fluid of a hepatoma-bearing patient isdescribed that is capable of yielding large quantities ofpure <*,-fetoprotein within a relatively short period of time.The technique is based entirely on the physicochemicalproperties of the <*,-fetoprotein molecule and uses sequential purification steps: ion-exchange chromatographyon DEAE-Sephadex A-50, molecular-sieve chromatography on Sephadex G-200, negative-affinity chromatographyon Sepharose-Blue Dextran, positive-affinity chromatography on concanavalin A-Sepharose and, finally, molecular-sieve chromatography on Sephadex G-100. The efficiency of the entire procedure in its present form is 15%of the a,-fetoprotein activity of the starting preparationfrom ascites fluid. The purity of the final product wasshown by polyacrylamide gel electrophoresis, radioim-munoelectrophoresis, and determinations of the MHz-terminal and COOH-terminal amino acid residues of the a,-fetoprotein isolated. Amino acid analysis of the finalproduct revealed a composition very similar to thosereported for a-fetoprotein preparations that have beenpreviously isolated by the use of ¡mmunochemicaltechnology.

INTRODUCTION

The oncofetal nature of AFP4 and its phylogenetic reten

tion have provided a great deal of impetus for the study ofits chemistry. Human AFP manifests a number of physico-chemical similarities to human albumin, and the 2 molecules are often described as structural and functional analogs, although this has yet to be proved. AFP has a molecular weight of approximately 70,000 daltons and an electro-phoretic mobility in the immediate postalbumin region onvirtually all supporting media (1, 11). Unlike the strictlyprotein nature of albumin, however, AFP is glycoprotein innature with a monosaccharide content of approximately4% and an average sialic acid content of 2 residues/molecule. At least some of the electrophoretic heterogeneity ofhuman AFP may be due to the variation in the sialic acidcontent of the different molecules within any single purifiedAFP preparation (1, 11). Nevertheless, differences in othersugar constituents and the presently undetermined variations in the protein core of AFP may well give rise to the

1This investigation was supported by grants from the Medical Research

Council of Canada, Ottawa, Ontario, Canada, and the National CancerInstitute of Canada. Toronto, Ontario, Canada.

2 Associate of the Medical Research Council of Canada.3 Clinical Research Associate of the National Cancer Institute of Canada.'The abbreviations used are: AFP, tt,-fetoprotein; SBD, Sepharose-Blue

Dextran; Con A. concanavalin A; PAG, polyacrylamide gel.Received December 20, 1976; accepted September 27, 1977.

microheterogeneity that has been observed (1).Despite the persistent interest in AFP chemistry, relatively

little is known about the structure of AFP. The major reasonfor the lack of structural data concerning this molecule hasbeen the technical difficulty of obtaining acceptably purepreparations of this material in large enough quantities forappropriate studies. The basis for this difficulty, in turn,has been that of separating AFP from albumin, with whichAFP shares so many physicochemical properties. Furthermore, in those biological solutions from which AFP may beappropriately isolated, the albumin concentration is almostinvariably present in concentrations some orders of magnitude greater than that of AFP.

Most attempts at AFP purification have involved the useof immunochemical technology (9) to avoid the problemsintroduced by the physicochemical similarities betweenAFP and albumin and to take advantage of their antigenicdifferences in xenogeneic animals. Such procedures haveused anti-AFP antibody preparations, either in liquid orsolid phase, as probes for the isolation of AFP (5, 11).Such approaches have provided adequate quantities ofpurified AFP for radiolabeling and, hence, for the establishment of radioimmunoassays for this material (5, 11, 14). Byuse of affinity chromatography, in some laboratories it hasbeen possible to obtain adequate quantities of AFP forreproducible structural studies. However, the method is,initially, costly and time consuming, as it requires theproduction of large volumes of relatively high-liter and low-affinity anti-AFP antiserum and also has the inherent difficulties associated with immunoadsorbent affinity chromatography.

The present study describes a method for the bulk purification of human AFP from the ascites fluid of a patientsuffering from primary hepatoma. The technology has beenbased entirely on the physicochemical rather than on theimmunochemical properties of the molecule and allows forthe preparation of large quantities of highly purified AFPwithin relatively short periods of time.

MATERIALS AND METHODS

Ascites Fluid

An aliquot of 1500 ml of AFP-positive ascites fluid wasobtained under sterile conditions by repeated drainage ofthe peritoneal cavity of a patient suffering from primaryhepatoma. The peritoneal drainage was required to relieverespiratory embarrassment suffered by the patient due tothe accumulation of the ascites fluid.

Assays for AFP and Albumin

The AFP content of the ascites fluid and of the sequentially purified AFP-containing material was determined by

CANCER RESEARCH VOL. 38

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Purification of Human AFP

either radioimmunoassay (14) or by the Laurell (6) rockettechnique, as previously described. The albumin concentration in the various fractions was measured by the Laurellrocket technique, with human albumin and anti-humanalbumin (Behringwerke AG and Hoechst Pharmaceuticals,Montreal, Quebec, Canada).

Purification of AFP from the Ascites Fluid

Ion-Exchange Chromatography on DEAE-Sephadex A-50. The ascites fluid was divided into 200-ml aliquots anddialyzed against 0.1 M phosphate buffer, pH 7.O. Each 200-ml aliquot of equilibrated ascites fluid was first subjected tobulk ion-exchange Chromatography on 90 g of DEAE-Sephadex A-50 (Pharmacia Canada Ltd., Dorval, Quebec, Canada) equilibrated with the same buffer, packed under gravity in a cylindrical glass funnel 13 cm in diameter, andfitted with a sintered glass filter. Chromatography wasperformed by gravity flow, with 3 liters of the phosphatebuffer described above used as the initial buffer and thenwith 5 liters of 0.1 M phosphate buffer, pH 5.5. Elution wasfinally performed with 2.5 liters of 0.1 M sodium acetate-0.5M NaCI buffer, pH 4.O. On the basis of the AFP assaysperformed, the fraction eluted with the final buffer anddesignated DE-1 was subjected to further purification. Fraction DE-1 was concentrated in an Amicon Model 2000concentrator (Amicon Corp., Lexington, Mass.) fitted witha PM-30 filter, dialyzed thoroughly against 0.1 M phosphatebuffer (pH 7.0), and rechromatographed on a DEAE-Sephadex A-50 column (5 x 39 cm) equilibrated previously withthe same buffer. After application of the sample, elutionwas performed by ascending Chromatography at 80 ml/hr,with a linear gradient consisting of phosphate and sodiumacetate-NaCI solution buffers. The mixing chamber contained 2 liters of 0.1 M phosphate buffer (pH 5.5), and thereservoir contained an equal volume of 0.1 M sodiumacetate-0.5 M NaCI (pH 4.0). During this procedure, as in allof the elution steps that followed, the Chromatographiepattern was monitored at 280 nm with an LKB Uvicord 2(Fisher Scientific Co. Ltd., Montreal, Quebec, Canada) or aZeiss PMQ II spectrophotometer equipped with a flow-through cell (Carl Zeiss Canada Ltd., Montreal, Quebec,Canada). Fractions 23 ml in volume were collected andassayed for AFP. On the basis of the assays performed, thefractions that were eluted at conductance values between5 and 28 mmhos were collected, pooled, and concentratedin an Amicon concentrator as described above. This pooledand concentrated material was designated Fraction DE-2.

Chromatography on Sephadex G-200. Fraction DE-2 wasthoroughly dialyzed against 0.05 M Tris-0.5 M NaCI buffer,pH 8.0, and was then chromatographed by ascending flowon a Sephadex G-200 column (5 x 85 cm) (PharmaciaCanada) equilibrated previously with the same buffer andcalibrated with ferritin and bovine serum albumin as molecular markers of 500,000 and 68,000 daltons, respectively.An elution flow rate of 40 ml/hr was used, and 10-mlfractions were collected and assayed for AFP activity. Onthe basis of these assays, those fractions appearing between elution volumes of 1000 and 1200 ml were pooled,concentrated in an Amicon concentrator as describedabove, and subjected to further purification.

Affinity Chromatography on SBD. The AFP-containingfraction obtained by Sephadex G-200 Chromatography wasdialyzed against 0.05 M Tris-0.5 M NaCI buffer, pH 8, andthen chromatographed by descending gravity flow at a rateof approximately 220 ml/hr on an SBD (Pharmacia Canada)column (5 x 35 cm) equilibrated previously with the samebuffer (15). Virtually all of the AFP activity was found in theeffluent fraction with 0.05 M Tris-0.5 M NaCI buffer. Thebound material was eluted with 6 M urea, and the columnwas reequilibrated with the starting buffer for repeated use.

Affinity Chromatography on Con A-Sepharose. Afterconcentration the AFP-containing fraction from the SBDcolumn was dialyzed against 0.1 M sodium acetate-1.0 MNaCI buffer (pH 6.0) containing 1.0 rriM CaCL, 1.0 mMMgCI2, and 1.0 mM MnCL. The dialyzed residue was thenchromatographed on a Con A-Sepharose (Pharmacia Canada) column equilibrated previously with the same buffer(10). The material that was not bound by the Con A waswashed through with this buffer, whereas the retainedmaterial was subsequently eluted with 2% «-methyl-D-glu-coside dissolved in the same buffer.

Chromatography on Sephadex G-100. The materialeluted from the Con A-Sepharose column by 2% «-methyl-D-glucoside (Aldrich Chemical Co., Inc., Milwaukee, Wis.)was concentrated and dialyzed against the 0.15 M NaCI-0.05 M phosphate buffer, pH 5.5. Molecular-sieve Chromatography was performed on a Sephadex G-100 column (2.5x 85 cm) equilibrated previously with the same buffer.Elution was performed by ascending flow at a rate of 15ml/hr. The AFP-containing fractions were pooled and concentrated as described above.

Analytical Procedures

PAG. PAG electrophoresis was carried out in 7% PAG(Eastman Kodak Co., Rochester, N. Y.), with a discontinuous buffer system as described by Davis and Ornstein (4).

Determination of the NH2-terminal Amino Acid Residueof the Purified AFP. Through the courtesy of Dr. A. C.Wang, an attempt was made to determine the NH.-terminalamino acid residue of the purified AFP preparation by useof the automated Edman degradation method (18). In addition, dansylation was used to identify the NH2-terminalresidue (8).

Determination of the COOH-terminal Amino Acid Residue of the Purified AFP. The COOH-terminal amino acidresidue of the purified AFP preparation was determined bymodifications of the methods described by Ambler (2) andWhite ef al. (19). Five mg AFP were dissolved in 1 mlmorpholine acetate and subjected to carboxypeptidase Adigestion. Aliquots containing 2 mg equivalents of AFPwere removed at 1, 5; 15, 30, 60, and 120 min. The freeamino acid residues obtained were examined in a Beckman120C automatic amino acid analyzer (Beckman Instruments, Inc., Montreal, Quebec, Canada) as previously described (7). The sample size and sensitivity of the aminoacid analytical method used was such that contaminatingmaterial representing 5% of the initial sample could bedetected.

Amino Acid Analysis of the Purified AFP Preparation.Amino acid analysis of the purified AFP preparation was

JANUARY 1978

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

P. Gold et al.

performed, after hydrolysis in 6 N MCI, in a Beckman 120Csingle-column automatic amino acid analyzer.

Radioimmunoelectrophoresis. The purified AFP preparation was labeled with I25I (Charles E. Frosst & Co.,Montreal, Quebec, Canada) by the chloramine-T technique.

The purified radiolabeled preparation was mixed with untreated ascites fluid, which served as an AFP carrier. Thismixture, with normal human serum serving as a control,was then examined against both specific anti-AFP andunabsorbed anti-human fetal cord blood antiserum to de

termine the homogeneity of the purified AFP preparationas previously described (20).

RESULTS

Purification of AFP. By the Laurell rocket technique, theinitial 1,500 ml AFP-positive ascites fluid was found to

contain a total of 885 mg AFP activity and 10,800 mgalbumin. After bulk ion-exchange chromatography onDEAE-Sephadex A-50, an apparent increase in total AFP

activity of the preparation was observed, in that 1448 mgAFP appeared to have been recovered in Fraction DE-1. On

concentration and column chromatography of this fractionon a DEAE-Sephadex A-50 column, the AFP activity of the

preparation appeared in the first peak after application ofthe linear gradient of 0.1 M sodium acetate-0.5 M NaCI

buffer to the 0.1 M phosphate buffer with which the columnhad initially been equilibrated (Chart 1). Conductance studies revealed that the AFP-active fraction was eluted by the

linear gradient at conductance values between 5 and 28mmhos. The total AFP recovered in this peak and designated DE-2 was 513 mg.

On chromatography of DE-2 on Sephadex G-200, theAFP-active material appeared in the elution volume between

1000 and 1200 ml (Chart 2). This elution volume corresponded to a molecular-weight zone of approximately

68,000 daltons. The total AFP activity recovered from thiscolumn was 461 mg.

On affinity chromatography of the AFP-active fractionfrom the Sephadex G-200 column on SBD, virtually all of

the AFP activity was recovered in the unbound or effluentpeak (Chart 3). The AFP activity recovered at this steptotaled 350 mg. The second stage of affinity chromatography on Con A-Sepharose yielded no AFP activity in theeffluent, but a sharp AFP-containing peak was eluted withthe application of 2% a-methyl-D-glucoside solution (Chart

4). This peak contained a total of 228 mg AFP activity. At

100

TUBE NUMBER

H.H. 500,000 H.H. 68,000

o-W/-80

TUBE NUMBER

Chart 2. Molecular-sieve chromatography of DE-2 on Sephadex G-200.The AFP-containing fractions appear in the molecular weight range of68,000

0.9-1

0.6 -

0.3-

0.

AFP

20 30 45 75

TUBE NUMBER

90

Chart 1. Chromatographie pattern of Fraction DE-1 on DEAE-SephadexA-50. The AFP-containing fractions appeared after the application of thelinear gradient at conductance values between 8 and 22 mmhos.

Chart 3. Removal of albumin from the AFP-containing material by affinitychromatography on SBD. AFP and some apparently pigment-containingalbumin are not bound by the SBD. The remainder of the albumin, however,is retained by the SBD and may be subsequently eluted with 6 M urea.

the end of this step of purification, no albumin could bedetected.

The final Chromatographie step on Sephadex G-100 gave

rise to a single peak that contained virtually all of the AFPactivity (Chart 5). The total quantity of the final product ofAFP purification was 133 mg. Hence, the overall efficiencyof the procedure was found to be approximately 15%. Therecoveries of AFP at the various stages of purification andthe corresponding contamination with albumin are shownin Table 1.

Purity of the AFP Preparation. Analytical alkaline PAGelectrophoresis of the purified AFP preparation revealed asingle band in the immediate postalbumin area (Fig. 1).Radioimmunoelectrophoresis of the 125l-labeled AFP prep

aration, with the native ascites fluid as carrier, gave rise toa single, homogeneous radioactive arc against both specific anti-AFP and unabsorbed anti-human cord serum anti-

sera (Fig. 2).Chemical Analysis of the Purified AFP Preparation. The

amino acid composition of the purified AFP is shown inTable 2. The concentrations of the various amino acidresidues are very similar to those that have been reportedby other workers (5, 13) who used immunochemical procedures for the purification of AFP.

CANCER RESEARCH VOL. 38

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Purification of Human AFP

0.9-1

0.0

(•AFF-H

25

TUBE NUMBER

50

Chart 4. Final step in the separation of albumin from AFP by affinitychromatography on Con A-Sepharose. None of the albumin is retained bythe column, whereas all of the AFP is bound and may subsequently beeluted with 2% a-methyl-D-glucoside.

Attempted analysis of the NhL-terminal amino acid resi

due of the purified AFP by use of both the Edman degradation procedure and the dansylation method produced dan-syl derivatives of the e-amino group of lysine but no demonstrable NH,,-terminal product, suggesting possible blockade

of the NHo terminus of the purified material. Enzymaticdegradation of the AFP by carboxypeptidase A demonstrated the presence of only valine at the COOH-terminal

position. As a control for this study the COOH terminus ofhuman serum albumin was determined and shown to beleucine, as previously reported (3).

DISCUSSION

The method for the purification of human AFP fromascites fluid, which has been described, is capable ofyielding large quantities of AFP within a relatively shortperiod of time. The procedure used is based on the physi-

0.5-,

"•0.3-

0.1 J

1 .21

0.8-

0.4

o.o-W/—r—l 20

lìX

40TUBE NUMBER

60

Chart 5. Final step in the AFP purification procedure with molecular-sieve chromatography on a Sephadex G-100 column.

Table 2Amino acid composition of human AFP from different sources

AFP prepared by the current method (A) is compared to AFPpurified by immunochemical methods [B, C (5), and D (13)] and isexpressed in moles of amino acid per 67,000 g of peptide.

moles/mole protein

Ascites fluid Fetal serum

AminoacidAspartic

acidThreonineSerineGlutamic

acidProlineGlycineAlanineHalf-cystineValineMethionineIsoleucineLeucineTyrosinePhenylalanmeLysineHistidineArginineA48.1839.1145.2389.0630.0648.6076.1927.1233.736.7724.9560.4111.4624.0936.8912.6515.34B49.0035.0036.00104.0022.0027.0049.0026.0029.006.0026.0054.0016.0029.0035.0012.0017.00C49.0036.0037.00110.0021.0026.0050.0022.0027.004.0025.0053.0016.0027.0036.0012.0017.00D42.5038.0033.5092.1024.7035.2049.9028.3031.306.4033.1057.9017.9027.9048.7017.3020.30

Table 1AFPpurificationVolumeAFP-containing

materialsAscites

fluidAfterbulk DEAE-SephadexA-50chromatography

(DE-1)AfterDEAE-SephadexA-50column

chromatography(DE-2)AfterSephadexG-200chromatographyAfter

SBD affinitychromatography(effluent)After

Con Aaffinitychromatography(eluate)After

SephadexG-100chromatography(mi)15003859080501845AFP(mg/ml)0.593.685.605.767.0012.662.95Total

AFP(mg)8851447.6513.4460.8350228133Albumin(mg/ml)7.210.810.263NilNilTotalalbumin(mg)10,8004,158918480100NilNil

JANUARY 1978

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

P. Gold et al.

cochemical properties of the AFP molecule and uses sequential purification steps of ion-exchange, molecular-sieve, and negative- and positive-affinity chromatography,which are simple to perform and may be readily applied invirtually any biological laboratory. The present methodoffers an attractive alternative to the isolation of AFP byaffinity chromatography with "monospecific" anti-AFP an

tibody coupled covalently to solid-phase supporting media.From the point of view of efficiency and time required andif the time necessary to raise and absorb monospecificanti-AFP antisera for affinity chromatography is taken intoconsideration, both procedures require about the sameamount of time. Furthermore, the antibody-mediated affinity chromatography approach is limited by the ability toraise adequate amounts of anti-AFP antisera that have thecapacity to both bind AFP and yet permit quantitativeelution of this molecule from the column by a chaotropicagent. In addition, nonspecific adsorption of material tothe affinity column or leeching of the anti-AFP antibodyfrom the solid matrix may present problems that usuallynecessitate an additional step of molecular-sieve chromatography. Furthermore, it is costly to raise the requisiteamount of antiserum required. For these reasons the present procedure offers a useful alternative approach to thepurification of the AFP molecule from a variety of biologicalsources and permits the handling of large initial quantitiesof AFP-containing fluids.

The purity of the final AFP preparation was demonstrated,physicochemically, by the observation of a single band inthe immediate postalbumin zone on PAG electrophoresisand, immunochemically, by the finding of a single, homogeneous radioactive precipitin band for AFP against anti-cord serum antiserum. There was no evidence of reactivityagainst anti-normal human serum antiserum by radioim-munoelectrophoresis.

The pure AFP obtained as described showed a verysimilar chemical composition to the purified AFP preparations obtained initially in relatively small yield by immuno-chemical techniques (5, 11). The COOH-terminal aminoacid residue of valine found in the preparation describedhere is in keeping with that previously reported from another laboratory that used immunochemically purified human AFP (11, 12). However, the NH, terminus of the AFPdescribed in the present report was apparently blocked.

In the process of AFP purification, several observationswere made that deserve comment. The initial step of bulkion-exchange chromatography on DEAE-Sephadex A-50(10) allows relatively large volumes of AFP-positive ascitesfluid to be processed rapidly and removes substantial quantities of albumin and of /3- and y-globulins. The apparentincrease in AFP activity that was observed at the end ofthis stage of the procedure remains unexplained. However,this phenomenon has not been observed in all preparationsprepared thusly. In the rat urine cytosol during the processof estrogen binding, however, AFP has been shown toparticipate in a noncovalent form of protein-protein interaction with another distinct but, as yet, poorly characterizedmolecule (17). Therefore, it is possible that the initial stepof bulk ion-exchange chromatography on DEAE-SephadexA-50 removes a substance that interacts with circulatingAFP and may thereby result in an increase in immunoreac-

tive AFP at this stage of the purification procedure. Thesubsequent column Chromatographie step on the sameion-exchange material, however, resulted in the major apparent loss of AFP activity observed during any of the stepsof the procedure. It may well be that the efficiency of themethod will be substantially improved if this step can becircumvented.

Molecular-sieve chromatography on Sephadex G-200 resulted in a peak of AFP activity in the appropriate molecular-weight range of approximately 68,000. At this stage, however, the AFP-containing material was still contaminatedwith an equal amount of albumin. Hence, the subsequent 2steps of affinity chromatography were critical in eliminationof the residual albumin contamination of the AFP preparation; neither procedure used alone, even with recycling,was able to eliminate the albumin contamination.

The first affinity chromatography step on SBD was basedon the observation of Travis and Pannell (16) that thismaterial has a high capacity for the binding of albumin.Indeed, when the AFP-containing fraction from the Sephadex G-200 column was passed through the SBD column,there was little loss of AFP, but there was a two-thirdsreduction in the concentration of albumin in the effluent.

The chemical basis for the effectiveness of the SBD-affinity column in selective removal of substantial amountsof albumin has not been established. It is known that thealbumin molecule binds avidly to dyes such as Evans blue.Hence, it is possible that the interaction of albumin withSBD represents a similar ligand-binding activity. An alternative explanation has, however, been proposed by Thompson ef al. (15), based on the possibility that Blue Dextrancomplexes with a large variety of proteins because it interacts specifically with a supersecondary structure designated the dinucleotide fold. This structure apparently involves about 120 amino acids that form a ß-pleatedsheetcore. The core is composed of 5 or 6 parallel strandsconnected by a-helical intrastrand loops located above andbelow the ßsheet. If this hypothesis is correct, it wouldstrongly indicate that the AFP molecule possesses topographical features that are quite distinct from those presentin albumin and might explain the marked differences in theaffinities of these molecules for SBD.

Nevertheless, some albumin was still found in the purifiedAFP preparation after affinity chromatography on SBD, andthis may perhaps be due to those albumin molecules thathad already bound pigments physiologically and that nolonger had a site for binding to the Blue Dextran. On thebasis of the observation of Page (10), the removal of thislast relatively small aliquot of albumin was achieved bypassage of the concentrated effluent from the SBD columnthrough Con A-Sepharose, which allowed the apparentlypigmented albumin molecules to escape in the effluent,although it efficiently held the AFP until this material waseluted with 2% a-methyl-o-glucoside (10). The sequentialuse of the SBD and Con A-Sepharose affinity column isunique to this method and was essential for the removal ofall of the albumin contamination of the AFP preparation.Whether it was due to the carriage of pigment or not, thefailure of SBD to bind a fraction of the albumin, even onrecycling of the preparation, indicates the presence of atleast 2 populations of albumin with distinctly different

10 CANCER RESEARCH VOL. 38

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Purification of Human AFP

affinities for SBD. The final Chromatographie procedure onSephadex G-100 was required to remove the trace quantities of impurities that were noted on sodium dodecyl sul-fate-PAG electrophoresis after Con A-Sepharose affinitychromatography.

In handling hepatitis B virus surface antigen-positiveAFP-containing hepatoma sera, a 30% ammonium sulfateprecipitation step has recently been added to the AFPpurification prior to the ion-exchange chromatography onDEAE-Sephadex A-50. The precipitate formed during thistreatment contains the hepatitis B virus surface antigenand removes a significant quantity of ß-and -/-globulin as

well as some albumin. The quantity of AFP lost in this partof the process is negligible. The subsequent Con A affinitychromatography step results in a much cleaner definitionof the AFP-containing peak and thus facilitates the subsequent purification of AFP.

REFERENCES

1. Alpert, E., and Perenevich, R. C. Human a-Fetoprotein ImmunochemicalAnalysis of Isoproteins. Ann. N. Y. Acad. Sci., 259: 131-135, 1975.

2. Ambler, R. P. Enzymatic Hydrolysis with Carboxypeptidase. MethodsEnzymol.,25: 155-166, 1972.

3. Behrens, P. Q., Spiekerman, A. M., and Brown, J. R. Structure ofHuman Serum Albumin. Federation Proc.,34: 591, 1975.

4. Davis, B. J., and Ornstein, L. Disc Electrophoresis. Rochester, N. Y.:Distillation Products Industries, 1961.

5. Hirai, H., Nishi, S., Watabe, H., and Tsukata, Y. Some Chemical,Experimental, and Clinical Investigation of a-Fetoprotein. Gann Monograph, 14: 19-33, 1973.

6. Laurell, C.-B. Quantitative Estimation of Proteins by Electrophoresis inAgarose Gel Containing Antibodies. Anal. Biochem.. 75: 45-52, 1966.

7. Moore, S., and Stein. W. H. Chromatography of Amino Acids on

Sulfonated Polystyrene Resins. J. Biol. Chem., 192: 663-681, 1951.8. Neadle, D. J., and Polliti, R. J. The Formation of 1-Dimethylamino-

naphthalene-5-Sulphonamide during the Preparation of 1-Dimethylami-nonaphthalene-5-Sulphonamino Acid. Biochem. J., 97: 607-608, 1965.

9. Nishi, S., Watabe, H., and Hirai, H. Immunological and ChemicalCorrelation between a-Fetoproteins from Human and Several Mammalian Species. Ann. N. Y. Acad. Sci., 259: 109-118, 1975.

10. Page, M. Alpha-Fetoprotein: Purification on Sepharose Linked Concan-avalin-A. Can. J. Biochem., 51: 1213-1215, 1973.

11. Ruoslahti, E., Pihko, H.. and Seppälä,M. Alpha-Fetoprotein: Immunochemical Purification and Chemical Properties. Expression in NormalState and in Malignant and Non-Malignant Liver Disease. Transplant.Rev., 20: 38-60, 1974.

12. Ruoslahti, E., Pihko. H., Vaheri, A., Seppälä,M., Virolainen, M., andKontinnen, A. Alpha-Fetoprotein: Structure and Expression in Man andInbred Mouse Strains under Normal Conditions and Liver Injury. JohnsHopkins Med. J. Suppl., 3. 249-55, 1974.

13. Ruoslahti, E., and Seppälä,M. Studies of Carcinofetal Proteins: Physicaland Chemical Properties of Human a-Fetoprotein. Intern. J. Cancer. 7:218-225, 1971.

14. Silver, H. K. B.. Gold, P., Feder, S., Freedman, S. O., and Shuster, J.Radioimmunoassay for Human Alpha,-Fetoprotein. Proc. Nati. Acad.Sei. U. S., 70. 526-530, 1973.

15. Thompson, S. T., Cass. K. H., and Stellwagen, E. Blue Dextran-Sepha-rose: An Affinity Column for the Dinucleotide Fold in Proteins. Proc.Nati. Acad. Sei. U. S., 72. 669-672. 1975.

16. Travis, J., and Pannell, R. Selective Removal of Albumin from Plasmaby Affinity Chromatography. Clin. Chim. Acta, 49, 49-52, 1973.

17. Uriel, J., Bouillon, D.. Aussel, C., and Dupiers, M. Alpha-Fetoprotein:The Major High-affinity Estrogen Binder in Rat Uterine Cystosols. Proc.Nati. Acad. Sei. U. S., 73: 1452-1456, 1976.

18. Wang, A. C., Banjo, C., Fuks. A., Shuster, J., and Gold, P. Heterogeneityof the Protein Moiety of Carcinoembryotic Antigens. Immunol. Commun. ,5: 205-210, 1976.

19. White, W. F., Shields, J.. and Robbins, K. C. C-terminal Sequence ofCrystalline Bovine and Human Albumins: Relationship of C-terminus toAntigenic Determinants of Bovine Serum Albumin. J. Am. Chem. Soc.,77: 1267-1269, 1955.

20. Yagi, Y., Maier, P., and Pressman, D. Immunoelectrophoretic Identification of Guinea Pig Anti-insulin Antibodies J. Immunol.. 89: 736-744,1962.

JANUARY 1978 11

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

P. Gold et al.

i

Fig. 2. Immunoelectrophoresis (left) and radioimmunoelectrophoresis(right) patterns. The troughs on the left contain the same materials denotedin the corresponding troughs on the right. Wells 1. 2, 4, and 5 are purified125l-labeledAFP in carrier ascites fluid. Well 3, is purified I25l-labeled AFP innormal human serum (NHS). Examination of the radioimmunoelectrophor-etic pattern reveals a single radioactive precipitation line between thepurified 125l-labeled AFP and the unabsorbed antiserum directed againsthuman cord serum. This band corresponds in position and shape to thatobtained between the purified '"l-labeled AFP and the absorbed, specificanti-AFP antiserum. No radiolabeled bands were observed against the anti-normal human serum antiserum, indicating the absence of normal adultserum constituents from the purified AFP preparation.

B

Fig. 1. Alkaline PAG electrophoresis. A. crystalline human serum albumin; B. purified AFP preparation; C, untreated ascites fluid. The purifiedAFP preparation appears as a single sharp band in the immediate postalbu-min zone.

12 CANCER RESEARCH VOL. 38

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

1978;38:6-12. Cancer Res   Phil Gold, Aurora Labitan, H. C. George Wong, et al.   Patient-Fetoprotein from the Ascites Fluid of a Hepatoma-bearing

1αPhysicochemical Approach to the Purification of Human

  Updated version

  http://cancerres.aacrjournals.org/content/38/1/6

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/38/1/6To request permission to re-use all or part of this article, use this link

on June 29, 2020. © 1978 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from