Journal of Clinical InvestigationVol. 42, No. 8, 1963

CHARACTERIZATION OF ANTIBODIES IN NORMALHUMANURINE *

By EZIO MERLER,t JACK S. REMINGTON,f+t MAXWELLFINLAND, ANDDAVID GITLIN WITH THE TECHNICAL ASSISTANCE OF CORNELIAPROCTOR

(From the Departments of Pediatrics, Biochemistry, and Medicine, Harvard Medical School,and the Thorndike Memorial Laboratory, Second and Fourth [Harvard] Medical

Services, Boston City Hospital, Boston, Mass.)

(Submitted for publication November 6, 1962; accepted May 2, 1963)

Certain of the physiochemical and immuno-chemical properties of the y-globulins of normalhuman urine have been reported by Webb, Rose,and Sehon (2) and by Franklin (3). These uri-nary proteins were found to be electrophoreticallyand antigenically closely related to serum y-globu-lins, but smaller in size; their molecular weightsranged from 10,600 (2) to 38,000 (3). Sedi-mentation coefficients S,,20 for the major con-stituents of the urinary y-globulins were reportedto be 1.14 S (2) and 1.6 S (3). Although Webband associates (2) and, more recently, Rowe andSoothill (4) did not demonstrate 7 S material intotal urine concentrates, Franklin (3) noted asmall amount of a rapidly sedimenting componentwith Sw,20 of about 7, in addition to the major 1.5to 1.6 S peak in y-globulins from a number of dif-ferent urines.

Recently, precipitating and viral neutralizingantibodies were shown to be present in the y-glob-ulins of normal human urine (5, 6). The anti-body activity was considered to be probably re-siding in the 7 S fraction. Porter (7), has shownhowever, that nonprecipitating fragments with anapproximate molecular weight of 55,000, obtainedby enzymatic degradation in vitro of rabbit serumantibodies, retain the ability to combine specificallywith homologous antigen. The present studieswere therefore undertaken to determine whetherthe low-molecular-weight y-globulins of normalurine possess antibody activity, and preliminaryobservations (8) suggested that they do.

* Presented in part before The Federation of AmericanSocieties for Experimental Biology on April 14, 1962 (1).Aided by grants A-251 and E-23 from the National Insti-tutes of Health, Bethesda, Md.

+ Present address: Palo Alto Medical Research Foun-dation, Palo Alto, Calif.

t Postdoctoral Research Fellow, National Institutes ofHealth.

METHODS

Collection and treatment of urine. Urine was collectedin clean containers without preservative. The followingprocedures were employed to concentrate the urine. 1)Lyophilization. After filtration through Whatman 12filter paper, the urine was dialyzed through 18/32-inchVisking 1 tubing against either distilled water or runningtap water at 100 C for 48 to 72 hours and lyophilized.2) Pervaporation followed by Iyophilization. After filtra-tion and dialysis, urine was poured into a reservoir con-nected to twelve 6-foot columns of 8/32-inch Viskingtubing, which presented a large surface area per unitvolume. The lower ends of the tubing were immersedin running tap water at 10° C while the upper portionswere exposed to three table fans. From 5 to 10 L ofurine could thus be concentrated 100-times in less than8 hours. The concentrated urine, usually dark brown incolor, was then lyophilized. Sufficient 0.15 M NaCl wasadded to bring the powder into solution; this was centri-fuged at 3,000 rpm for 30 minutes, and the supernatantfluid was stored at -200 C. 3) Passage through Sepha-dex.2 The method used was a modification of the pro-cedure described by Flodin, Gelotte, and Porath (9).After filtration through Whatman 12 filter paper, theurine was added to Sephadex G-25. The thick suspen-sion was thoroughly stirred, allowed to stand at roomtemperature for 15 to 30 minutes, and then filtered withsuction through Bfichner funnels with interposed fine-mesh metal screen (10). The efficiency of this filtrationwas further enhanced by a sheet of rubber or Parafilm 3placed over the top of the funnels. Filtrates were thenreconcentrated in the same manner to about 1/500th to1/1,000th of the original volume of urine; 2 to 3 L ofthe original urine was thus reduced in 3 or 4 steps to1 to 3 ml.

7 S y-globulin. Human y-globulin was obtained frompooled plasma of normal adults by low-temperatureethanol-water fractionation (Squibb lot 330-2). Ap-proximately 96%o of this material settled as sediment inthe ultracentrifuge with an S, 2o of 6.8 at a protein con-centration of 1%. Less than 4%o0 of molecules settled assediment at 9 to 11 S, and no detectable amount did so

1 Visking Co., Division of Union Carbide Corp., Chi-cago, Ill.

2 Pharmacia, Uppsala, Sweden.3 Marathon Co., Menasha, Wis.

1340

CHARACTERIZATIONOF ANTIBODIES IN NORMALHUMANURINE

at 19 S. The material had the electrophoretic mobility ofnormal human -y-globulin and gave a single precipitin lineupon immunoelectrophoresis with an antiserum to normalhuman serum.

Antisera. A pool was made of sera obtained from 4rabbits immunized with human serum 7 S y2-globulins.Antiserum4 13,411 from a horse immunized with wholenormal human serum was used.

Immunoelectrophoresis. The micromethod of Scheideg-ger (11) was employed, with 1%o agar in borate buffer atpH 8.6 (/A 0.064). The precipitin lines appeared to beat optimal intensity after 24 to 72 hours. The agar waswashed in 0.15 MNaCl; the precipitin lines were stainedwith Nigrosin or Ponceau S.

Double diffusion in agar gel. The technique of Ouchter-lony (12) was used. In some instances, double diffusionin tubes was used as described by Preer (13). Quanti-tative estimation of urinary y-globulins was performed asdescribed by Feinberg (14).

Immunization. Two persons with normal blood ureanitrogen and urine that was free of protein by the sulfo-salicylic acid test were given booster doses of tetanustoxoid. Ten days later a serum sample was drawn; col-lection of total urine output was then begun and continuedfor 10 days in one subject and for 14 days in the other.

Volunteers immunized with poliovirus vaccine werethe same as those used previously (Subjects 1 and 5) (6).

Chromatography. Fractionation of the urinary proteinswas performed on columns of DEAEcellulose 5 at roomtemperature. All buffer solutions were saturated withtoluene. Concentrated urine was dialyzed against 0.0175Mphosphate buffer at pH 6.3, and insoluble residues wereremoved by centrifugation at 2,000 rpm for 20 minutes.The clear, dark, supernatant fluid was placed on the col-umn previously equilibrated with starting buffer untilthe effluent had a pH of 6.3.

Columns were connected to a continuous-flow ultra-violet absorption meter, and the optical density of theeluates at 254 m/t was recorded directly. Elution with0.0175 M phosphate buffer at pH 6.3 was continued untilthere was no further ultraviolet absorption. A lineargradient (phosphate, 0.0175 M at pH 6.3 to 0.4 M at pH5.2) was then used to separate the remaining urinaryproteins.

Contents of tubes comprising portions of the elutionpatterns within each peak were combined, dialyzed againstdistilled water, and lyophilized. After the addition of 1or 2 ml of 0.15 MNaCl to each lyophilized fraction, smallamounts of insoluble cellulose resin were separated bycentrifugation at 15,000 rpm for 20 minutes.

Enzymatic digestion and peptide patterns. Proteinswere denatured before enzymatic digestion either by heat-ing at pH 8 for 45 minutes at 1000 C, or by oxidationwith performic acid (15). Enzymatic hydrolyses wereperformed in the pH-stat at 370 C (16). Twice-crystal-lized trypsin0 was used in a 1:20 enzyme: substrate ratio.

4 No. 13,411, Institut Pasteur, Paris, France.5 Selectacel 70 Standard grade, Brown Co., Berlin, N. H.6 Worthington Biochemical Company, Freehold, N. J.

The peptides obtained from each digestion were separatedon paper in two dimensions, first by electrophoresis inone direction, followed by ascending chromatography inthe second, as described elsewhere (17). Peptide pat-terns were stained specifically for tyrosine by the methodof Acher and Crocker (18).

Ultracentrifugation. Preparative ultracentrifugationwas performed in a Spinco model L ultracentrifuge witha swinging bucket rotor (SW-39) with gradients ofsucrose 1 to 20%o (19). In most instances, tubes werecut 1.5 and 3 cm from the top of the sucrose column togive three fractions.

Analytical ultracentrifugal analyses were performed ina Spinco model E ultracentrifuge equipped with Schlierenoptics. Sedimentation velocity experiments were per-formed at 52,640 rpm in a double-sector 12-mm cell, withrotor temperature maintained at 200 C. Sedimentationcoefficients were corrected for the density of the buffer,but not for protein concentration, which never exceeded1%.

Weight-average molecular weights were determined bythe method of Archibald as modified by Klainer andKegeles (20). Measurements were made at short enoughintervals of centrifugation to avoid any redistribution ofsolute. The temperature of the rotor was maintained at200 C. The Schlieren optical system was used with aphase plate at the diaphragm, which was maintained at800 to keep the curve outline sharp and to minimize the,uncertainty of the curve position at the meniscus. Runswere made with a 12-mm cell using a 20 sector at 35,600rpm. Measurements near the cell bottom were not used.The concentration of the original sample was determinedin arbitrary units in a synthetic boundary cell. Photo-graphs were made on Kodak metallographic plates.Measurements of the photographic plates were made ina Mann microcomparator.

The concentration of the protein at the meniscus (Cm)was calculated from the measurements using the equation(20):

Cm=CO- F0.01 nX2ZFX'm n=o

where C. is the original concentration (determined in thesynthetic boundary cell); 0.1 cm is the comparator in-terval along the X axis; F is the enlargement factor; Z.is the ordinate (proportional to concentration gradient,dc/dx); and nD is the number of comparator intervalsneeded to bring the ordinate to zero. The solvent cor-rection was checked and found to be negligible.

The weight average molecular weight (M.) was thendetermined from the formula:

M RT (dc/dx)mU -( Vp)Cw2 XmCm

with w the angular velocity. The partial specific volume vfor the urinary y-globulin was not determined; the valueaccepted for serum 'y-globulin was used, 0.736 ml per gDiffusion values were measured in a synthetic boundarycell. Fujita's (21) correction was not applied to thediffusion constant.

1341

EZIO MERLER, JACK S. REMINGTON, MAXWELLFINLAND, AND DAVID GITLIN

Electrophoresis. Cellulose-acetate strips (22) wereused with a borate buffer at pH 8.6 (ja 0.064). Afterdrying, the strips were stained with either Ponceau S orNigrosin. Starch-block electrophoresis was performedwith hydrolyzed starch 7 in borate buffer at pH 8.6 (a0.064) (23). The separated proteins from serum orurine were eluted from 1-cm segments of the blocks.Urinary protein fractions with electrophoretic mobilitiescorresponding to those of individual serum proteins werepooled and dialyzed against distilled water. The frac-tions were lyophilized, resuspended in 0.15 M NaCl, andcentrifuged at 15,000 rpm for 20 minutes to remove in-soluble starch granules. The processed eluates from 1-cmsegments of protein-free starch blocks contained solublestarch in quantities sufficient to appear as a peak in theSchlieren pattern upon ultracentrifugation. Consequently,a polyvinyl chloride preparation 8 to replace starch wasused for the electrophoretic separation of the samplesanalyzed by ultracentrifugation.

Polio virus neutralization. Neutralizing antibody topolio viruses was measured9 by the immunoinactivationprocedure as previously described (6).

Urinary excretion of radioiodinated proteins after ivinjection of labeled 7 S y2-globulins. A 'y-globulin prep-aration (15 mg in 3 ml of carbonate buffer at pH 9.2)obtained as fraction II-1,2 by Cohn's method 9 was labeledwith I"31 as described elsewhere (24). The labeled y-globulin was dialyzed against 10% sucrose in 0.15 MNaCl. The protein was then centrifuged at 35,000 rpmfor 18 hours in a gradient of sucrose (10 to 40%o) to as-sure ultracentrifugal homogeneity of the material to beinjected. The sucrose gradient was sliced in six equalsegments. The solution from each segment was assayedfor radioactivity in a 3- X 3-inch NaI crystal scintillatorand for protein concentration by reading the opticaldensity at 280 m/A. The middle two segments with thehighest counts and optical density were combined anddialyzed against 0.15 MNaCl at 2° C. The nondialyzablesolution (15 ml) was sterilized; 83 fsc of 131 was injectedintravenously into a normal person and a patient knownto have agammaglobulinemia.

Blood was drawn at 12 minutes, 12 hours, and 36 hoursafter iv administration of the iodinated -y-globulin. Totalurinary output was collected for the first 12 hours andfor the three 24-hour periods thereafter. Urine sampleswere concentrated by pervaporation, followed by lyophili-zation, and a sample was separated by zone electrophoresison starch. Materials from each 1-cm segment of theblocks were assayed for radioactivity; protein concentra-tion was estimated by measurement of optical densityat 280 m/A.

7 Connaught Medical Laboratories, University of To-ronto, Toronto, Canada.

8 Pevikon, C-870, Stockholms Superfosfat Fabriks Ak-tiebolag, Stockholm, Sweden.

9 By Dr. A. Martin Lerner.

RESULTS

Fractionation of urinary proteins. A represen-tative pattern of separation of the concentrated(lyophilized) urinary proteins by starch-block

electrophoresis is shown in Figure 1. This patterncompares favorably with those obtained by moving-boundary electrophoresis (25). In the presentstudies, however, optical densities at 280 m/L maynot necessarily represent true protein concentra-tions of the various fractions. Pigments, whichare bound to urine proteins, possess a strong ultra-

.30

.20'

.10AE

00

.50

I-

zW .40a

,_ .30

0.

.20

.10

0.

SERUM

U R N E

FRACTIONSI If III IV

. . t . . . . .I.-5 -3 -I 0 2 4 6 8 10 12

DISTANCE FROM ORIGIN, CM.

FIG. 1. SEPARATIONBY STARCH-BLOCKELECTROPHORESISOF NORMALHUMANSERUMAND DIALYZED, LYOPHILIZED,NORMALHUMANURINE.

violet absorption and probably contribute to theoptical densities observed. By double diffusionin agar gel and immunoelectrophoresis, the eluatesfrom the area corresponding to fraction I of thestarch block were shown to possess antigenicproperties common to those of serum 7 S y-globu-lins.

Figure 2 shows the elution pattern of the uri-nary proteins obtained from the DEAE cellulosecolumn, and Figure 3, the immunoelectrophoreticpatterns of the fractions. The fraction eluted fromthe DEAE cellulose resin with 0.0175 M phos-phate buffer at pH 6.3 contained proteins with

1342

CHARACTERIZATIONOF ANTIBODIES IN NORMALHUMANURINE

the same antigenic specificities and electrophoreticmobilities as fraction I by starch-block electro-phoresis. Figure 3A depicts the immunoelectro-phoretic pattern obtained when an antiserum pre-pared against normal human serum reacted witha sample of concentrated urine proteins; the reac-tion of this antiserum with normal human serumis also shown. At least five precipitin lines, cor-responding to five, electrophoretically distinguish-able proteins of urine with antigenic sites similarto those of serum, are evident. Figure 3B showsthe precipitin lines obtained with fraction I (pre-pared by chromatography on DEAEcellulose) ofthe same urine. Two lines are visible (one ofthem very faintly), both with a mobility similar tothat of serum y-globulins, but appearing to mi-grate toward the anode at a faster rate. y-Globu-lins were not confined to fraction I, but were elutedalso in fractions II and III (Figure 3C, D), al-though in these two fractions they possessed eitheran atypical y mobility, or they were contaminatedwith other urinary proteins. The immunoelec-trophoretic patterns of fractions IV, V, and VI(Figure 2) are depicted in Figure 3E, F, and G,respectively.

Electrophoretic patterns of urine concentratedby lyophilization and of normal serum with cellu-lose-acetate strips are shown in Figure 4. In anumber of instances, separation of urinary pro-teins by cellulose-acetate or starch-gel electropho-resis revealed a component that migrated furthertoward the cathode than did serum y-globulins(arrow in Figure 4B). A precipitin line did notresult from the reaction of this slow-moving com-ponent with antisera prepared in rabbits againstserum 7 S y2-globulins.

Ultracentrifugal analyses were performed oneleven urines of seven persons. Figure 5A showsa representative Schlieren diagram obtained withconcentrated whole urine. Sw,20 was 1.3 ± 0.1,significantly lower than the value of Webb andassociates (2), and it did not vary significantlywhether the urine samples were concentrated bylyophilization, by dialysis and pervaporation, orwith Sephadex.

Although ultracentrifugal patterns of total urineconcentrates failed to reveal components with Sw, 20over 2, the concentrated y-globulins from the sameurines usually contained both a major componentwith an average S,,20 of 1.5 ± 0.3 and a minor

GRADIENT TO P04 0.4 M, pH 5.2

."

EZIO MERLER, JACK S. REMINGTON, MAXWELLFINLAND, AND DAVID GITLIN

FIG. 3. IMMUNOELECTROPHORESISOF CHROMATOGRAPHIC(DEAE) FRACTIONSOF NORMALHUMANURINE.(See Figure 2.) s = Normal human serum, u = normal human urine. Center well contained horseantinormal human serum. A, lyophilized normal human urine; B, fraction I; C, fraction II; D, fractionIII; E, fraction IV, F, fraction V; and G, fraction VI.

one with an average S,,20 of 5.7 + 0.5 (Figure5B). These values represent the arithmeticmeans of the sedimentation coefficients of urinaryy-globulins from each of five persons. In all in-stances, the major 1.5 S peak was broad. Evenafter prolonged centrifugation, complete separa-tion of this peak from the meniscus was not enough

to indicate the distribution of molecular weightsin the system.

Centrifugation in a sucrose gradient was usedto separate the two differently sedimenting com-ponents of urine y-globulins. Materials from thetop and bottom fractions of the sucrose gradientswere analyzed by sedimentation velocity for mo-

1344

CHARACTERIZATIONOF ANTIBODIES IN NORMALHUMANURINE

A

FIG. 4. SEPARATION BY CELLULOSE-ACETATE ELECTRO-PHORESIS OF A) NORMALHUMANSERUM AND B) DI-ALYZED AND LYOPHILIZED HUMANURINE. Arrow pointsat urinary components migrating toward cathode fasterthan y-globulin.

lecular size, and by immunoelectrophoresis anddouble diffusion in agar for antigenic properties.Ultracentrifugation of the top fraction revealedonly low-molecular-weight material with an Sw,20of 0.92 (Figure 5C).

Ultracentrifugation of the y-globulins from thebottom fractions of the sucrose gradients showeda single peak with an Sw,20 of 7.1. These largey-globulins were present in all but two of the bot-

tom fractions of the urine concentrates tested fromtwelve different persons. The finding of a large-molecular-weight material in this protein fractionof normal urine differs from that of Webb andassociates (2) and Rowe and Soothill (4), butcorresponds to that of Franklin (3).

Figure 6 shows the precipitin lines obtainedwhen the top and bottom fractions of two prepara-tions of urinary -y-globulins separated in sucrosegradients reacted in agar with an antiserum pre-pared in rabbits against human serum 7 S y2-glob-ulins. The material from the top fraction gave areaction of partial identity with serum 7 S y-glob-ulins (Figure 6, left). The bottom fraction ofone preparation contained a single component thatgave a reaction of complete identity with serumy-globulins (Figure 6, left), whereas in anotherbottom fraction, (Figure 6, right) there werethree lines suggesting either that the faster sedi-menting component was not antigenically homog-enous, or that the low-molecular-weight com-ponent of urine y-globulins was polydisperse;complete separation was not obtained in thispreparation.

Results by the Ouchterlony method suggestedthat the two precipitin lines visible in the im-munoelectrophoretic pattern of the urinary y-glob-ulins (Figure 3B) might correspond to the twodifferently sedimenting components. This sug-gestion was confirmed by immunoelectrophoresis;the top fraction from the gradient formed a pre-

FIG. 5. ULTRACENTRIFUGEPATTERNS OF A) TOTAL CONCENTRATEDURINE, B) URINARY y-GLOBULINS (DEAE), ANDC) TOP OF SUCROSEGRADIENT.

1345

:.::., k4e.t

-P v:- ;.'; .f:

EZIO MERLER, JACK S. REMINGTON, MAXWELLFINLAND, AND DAVID GITLIN

FIG. 6. DOUBLEDIFFUSION IN AGAR. G= human serum7 S y-globulin, A = rabbit antihuman 7 S -globulin, T= urine fraction from top of sucrose gradient, and B= urine fraction from bottom of sucrose gradient.

cipitin line corresponding to the line closest to theantibody well, and the bottom fraction, to the onefarthest from it (Figure 3B). Although this bot-tom fraction constituted only 10% or less of thetotal urinary y-globulins (8), it formed a muchclearer precipitin line with the antibody than didthe low-molecular-weight material.

Antibody activity. Concentrates of urine hadbeen shown to contain neutralizing activity againstpoliovirus types I and III (6). It was thereforeof interest to determine if this activity were asso-ciated with the 7 S fraction or the low-molecular-weight material in these urines. The urine pro-teins were separated in fractions by starch-blockelectrophoresis and cellulose chromatography.Analysis by double diffusion in agar of fractionI (DEAE) and of materials eluted from the starchblocks with an electrophoretic mobility matchingthat of serum 7 S y-globulin indicated that onlymaterials giving reactions of partial identity with7 S y-globulin were present in these urines. Theprecipitation lines corresponded to those obtainedfrom the upper fractions of the separation by su-crose gradient. In the ultracentrifuge, fraction Ifrom these urines gave only one peak with an S, 20of 1.70 and did not contain the minor, 7.1 S com-ponent ordinarily found in concentrated urinaryy-globulins. Virus-neutralizing activity resided infraction I (DEAE) of these urines and was foundonly in urines from subjects with neutralizing ac-tivity in the corresponding serum. In some in-stances., antibody could be demonstrated in theurine only a week or more after booster doses of

vaccine and only after the antibody titer in the se-rum had increased. No comparison of titers ob-tained with urine y-globulins with those of thecorresponding sera has been attempted.

Precipitating antibodies to tetanus toxoid weredemonstrated in subjects' urines after booster dosesof that antigen. The activity of these antibodieswas also associated with the low-molecular-weightcomponent of the y-globulins from the top seg-ment of the sucrose gradients. The y-globulinscontaining these antibodies were shown by doublediffusion in agar to give a single precipitin linethat showed a reaction of partial identity withserum 7 S y-globulin (Figure 6, A-T). Thefaster sedimenting component, from the bot-tom segment, did not precipitate with the antigen,perhaps owing to its low concentration in theurine. The precipitin lines from the reaction ofserum and the urinary low-molecular-weighty-globulins with tetanus toxoid in Preer tubes areshown in Figure 7. Precipitation of the serumantibodies with homologous antigen occurredwithin 4 to 7 days, whereas precipitin lines withthe /-globulins of urine took up to 3 or 4 weeksto appear. Precipitin lines could not be demon-strated with these low-molecular-weight urinaryy-globulins when they reacted in agar with diph-theria toxoid, type-specific pneumococcal polysac-charides, normal human serum, or 0.15 M NaCl.,-Globulins prepared from the urines of personswhose sera did not precipitate with tetanus toxoiddid not react visibly with that antigen.

The material from the top segment of the su-crose gradient that precipitated with tetanus toxoidhad a molecular weight of 12,900 which decreasedwith time, indicating the polydisperse nature ofthese preparations (Table I). The diffusion con-stant was 7.4 x 10-7 as compared to 4.0 x 10-7

TABLE I

Archibald measurements of molecular weights as a functionof time for StD,20 of 0.92 urinary -y-globulin

Time ZnZnXn2 Cm AIW

min1 68.936 0.0266 13,1046 83.318 0.0249 12,437

10 93.144 0.0237 12,14314 102.118 0.0226 11,67026 120.515 0.0204 10,74930 115.081 0.0211 10,420

Average Co = 0.0349.

1346

CHARACTERIZATIONOF ANTIBODIES IN NORMALHUMANURINE

I

FIG. 7. DOUBLE DIFFUSION IN AGAR (TUBES). p = Pre-cipitin lines, t = tetanus toxoid, s = human antitetanus serum,and u = low-molecular-weight urinary 'y-globulin.

for serum 7 S y-globulins (26). S,, 20 was 0.92.S successively increased for materials from areasof increasing sucrose concentration.

Sedimentation coefficients of the total urinaryy-globulins from the various fractions from theseparation by sucrose gradient were nearly con-stant in different experiments at the concentra-tions used (0.5 to 1 g per 100 ml); measurementsat lower concentrations indicated that dependenceon concentration was small for a 17o% change inthe concentration of protein. Values for molecu-lar weight obtained by extrapolation to infinitedilution should therefore not differ greatly fromthose obtained at the concentration used in theseexperiments.

Quantitation of urinary y-globulins. In fractionI (DEAE) this was done by titrating increasingdilutions of a known weight of this material andof a standard solution of human 7 S y-globulinby gel diffusion in an Ouchterlony plate against astandard solution of the anti-7 S y-globulin anti-serum (14). With the greatest dilution of anti-gen (serum 7 S y-globulin or urinary y-globulin)that gave a visible precipitin line with antiserum asan endpoint, such lines were not discernible withless than 1.5 ytg of urinary y-globulins or 0.78 ugof serum 7 S y-globulin. The extinction coeffi-

cient of total urinary y-globulins E'9-m was 4.47 ascompared to 14.9 for serum y2-globulins.

Analysis of peptides released by enzymatic hy-drolysis of serum and urinary y-globulins. Thepeptides obtained from enzymatic digestion ofheat-denatured (HD) total urinary y-globulinsand of the HD y-globulins of 0.92 S were com-pared with those from HDserum 7 S y-globulins.When trypsin hydrolysates were adjusted, afterdigestion, to pH 4.2, about 15% of the weight of7 S y-globulin precipitated, whereas little or no

TABLE II

Relation between the number of micromoles NaOHper micro-mole protein used during enzymatic digestion and the

number of peptides revealed

;&MolesOH/

jsmole No. ofSample* protein peptides

Serum ey-globulin (HD) 67 55DEAEfraction I (HD)t 5.4 32

Segment 1 of fraction I (HD)t 5.2 11Serum -y-globulin (PAO) 130 70

DEAEfraction I (PAO)t 7.3 30

* HD = heat-denatured; PAO = performic-acid oxi-dized.

t An average mol wt of 30,000 was assumed for thisfractionwof the urinary 7-globulins.

t'S.,20 ='0.92.

1347

EZIO MERLER, JACK S. REMINGTON, MAXWELLFINLAND, AND DAVID GITLIN

protein precipitated from""the''urinary y-globutin.Some of the peptides in .the patterns of urinaryy-globulins appeared to be present in those of se-rum 7 S y-globulins. Most of the peptides of theurinary y-globulins, however, did not overlap thoseof serum y-globulins. More peptides were dis-tinguishable in patterns of the serum 7 S y-globu-lins than in those of the total urinary y-globulins.Only eleven peptides were distinguishable in hy-drolysates of the 0.92 S urinary y-globulins.

In contrast to the results obtained with serum7 S y-globulins, the number of spots evident onthe chromatograms of urinary y-globulins wasgreater than the number of moles of hydroxylions needed per mole of urinary proteins duringenzymic digestion to keep the pH constant at 8.0(Table II).

0.2

=L.0. I.E0ODcm 0-

z

a-I

4

At

I-0

The peptides obtained from enzymatic diges-tion Of 'performic-acid-oxidized (PAO) total uri-nary, y-globulins. were compared with those fromPAOserum 7S y}-globulins. Fewer peptides werenoted on the patterns of PAOurinary y-globulinsthan those obtained with the serum 7 S y-globulins(Table II). As observed with HD y-globulin,only a few peptides from the PAOurinary y-glob-ulin were superimposable on those from serum7 S y-globulins. Performic-acid oxidation of theurinary y-globulins before enzymatic digestion re-sulted in a peptide pattern with only two fewerspots than that obtained by enzymatic digestion ofHD urinary y-globulins. Considerably fewer ty-rosine residues were found in the peptide patternsof urinary y-globulins as compared to those ofserum y2-globulins.

16 la to

70

60

50

40

-30

020 o

z10 -D

Io:w

-I

, a

FIG. 8. SEPARATIONBY STARCH-BLOCKELECTROPHORESISOF NORMALHUMANSERUMAND PERVAPORATED,LYOPHILIZED NORMALHUMANURINE. Solid line: optical density at280 m/u. Dotted line: counts per minute corrected for background and decay.

-6 -4 -2 0 2 4 6 8 10 12 14DISTANCE FROM ORIGIN, CM.

1348

CHARACTERIZATIONOF ANTIBODIES IN NORMALHUMANURINE

TABLE III

Activity of urinary proteins after injection of I'3l-labeled serum 7 S y-globulins

Activity*Time of urine

collection (after % in -y-Subject injection) Totalt y-Globulins globulins

hours cpm cpm %Normal 0-12 471 166 35

12-36 62 25 4036-60 244 94 3960-84 349 184 53

Agammaglobulinemict 0-12 308 192 6212-36 139 82 5936-60 85 58 6860-84 198 119 60

* Corrected for background and radioactive decay.t Total counts per minute recovered from all fractions of the starch block; recovery of 1131 from amount applied

ranged from 95 to 98%. Varying amounts of urine concentrates from the time intervals described were placed on theblocks.

I White man 18 years old with congenital agammaglobulinemia. At the time of this study, serum -y-globulin wasless than 100 mg per 100 ml.

Reduction of urinary y-globulins.10 A prepa-ration of urinary y-globulins from the top of aseparation by sucrose gradient had an apparentmolecular weight M, of 9,800. These proteinswere reduced under mild conditions with mercap-toethanol to determine if easily accessible disul-fide groups were indeed present in these y-globu-lins. Reductions were done in 0.02 M mercap-toethanol at pH 8.0. Samples were kept at 20 Cfor various periods, but never less than 24 hours.From the approach to equilibrium during sedi-mentation as an estimation of molecular weightafter reduction, these urinary y-globulins had aMa of 6,200.

Urinary excretion of radioiodinated proteinsafter iv injection of labeled 7 S y2-globulins. Arepresentative pattern of the starch-block electro-phoresis of concentrated urinary proteins from anormal person given IJ31-labeled 7 S y2-globulinsis shown in Figure 8. The urine was concen-trated by pervaporation followed by lyophilization.This pattern differed markedly from that obtainedwhen the urine was concentrated by lyophilizationalone (Figure 1). Approximately one-third toone-half of the protein-bound radioactivity in thenormal urine was present in proteins with theelectrophoretic mobility of y-globulin (Table III).

10 This study was performed in collaboration with Dr.Karl Schmid of the Massachusetts General Hospital, Bos-ton, Mass., who kindly determined the molecular weights.

More than 50% of the protein-bound radioactivitywas found in similar y-globulin areas of starchblocks of the urinary proteins from the patientwith agammaglobulinemia.

Na 131 in Na2SO3 was added to two urine sam-ples, in one instance before and in the other afterconcentration by pervaporation. After dialysis toremove free iodide, the urines were submitted tostarch-block electrophoresis. Binding of the ra-dioactive atoms to urinary proteins was limitedto materials possessing the electrophoretic mo-bilities and immunologic specificity of serum,8-globulins and albumin.

In a separate experiment, concentrated urinefrom the normal person was applied to a columnof DEAE cellulose, and materials eluted by0.0175 M phosphate buffer at pH 6.3 (fractionI) were concentrated and separated in a gradientof sucrose (1 to 20%). A comparable amountof radioactivity was found in the top and bottomfractions of the gradient; a larger amount of ra-dioactivity was found in the middle fraction. Theprotein concentration, as determined by dryweight, was greatest at the top of the gradientand decreased toward regions of higher sucroseconcentration.

DISCUSSION

The present study describes certain of the physi-ochemical and immunochemical properties of the

1349

EZIO MERLER, JACK S. REMINGTON, MAXWELLFINLAND, AND DAVID GITLIN

low-miolecular-weight -y-globulins present in nor-mal human urine that possess both precipitatingand viral neutralizing antibody activity. Thesey-globulins at pH 8.6 are more electronegativethan normal serum 7 S y-globulins and share cer-tain antigenic determinants with serum 7 S y-glob-ulins.

Urinary y-globulins from individuals hyperim-munized with purified tetanus toxoid precipitatedspecifically with that antigen. The reason pre-cipitation took longer when tetanus toxoid reactedwith the low-molecular-weight y-globulins of urine,compared with the time for the same antigen toprecipitate with serum antibody is not clear. Thephysicochemical factors for precipitation withthese low-molecular-weight antibodies may differfrom those required for 7 S -y-globulins. Becausethe yield of low-molecular-weight y-globulins fromlarge volumes of urine was relatively small, pre-cipitation in agar gels had to be used rather thanquantitative precipitation from solutions; the agarmatrix may have been necessary for visible pre-cipitation. That precipitation of antigen with low-molecular-weight y-globulin from the top of asucrose gradient was not due to coprecipitationwith contaminating urinary 7 S y-globulins wasshown by the inability of the concentrated 7 Sy-globulins from the bottom of the same sucrosegradient to form a visible precipitate with thetetanus toxoid. Moreover, material from the topof separations by sucrose gradient of whole hu-man antiserum centrifuged in parallel with theurine did not precipitate with the antigen, whereasthe bottom fractions did.

Titration by immunoprecipitation (14) indi-cated that about 50% of the urinary -y-globulinsreacted with the antiserum and thus had antigenicsites common to serum 7 S y-globulins. Separa-tion by starch-block electrophoresis of the urinaryproteins of a patient with agammaglobulinemiarevealed patterns indistinguishable from those ofnormal subjects; however, materials eluted fromthe starch blocks of this patient that correspondedto the area of electrophoretic mobility of serumy-globulins did not react with an antiserum againstserum 7 S y-globulins.

The low value for the extinction coefficient ofthe urinary y-globulins suggests that either theseproteins are poor in aromatic amino acid resi-dues, or they contain proteins poor in tyrosine and

tryptophane. The low number of tyrosine resi-dues in the pattern of peptides after enzymatic di-gestion of these proteins favors the first possi-bility but does not exclude the second.

In peptide patterns of fraction I (DEAE) andin patterns for the protein of the top of separa-tions by sucrose gradient of fraction I, only a fewof the peptides had a mobility similar to that ofpeptides released by enzymatic digestion of serum7 S y-globulins. Similarly, the mobility of themajority of peptides obtained after enzymatic di-gestion of PAOurinary y-globulins differed fromthat of peptides from PAO serum 72-globulins.These results suggest the possibility that the pri-mary structure of urinary y-globulins may differin portions from that of serum 7 S y2-globulins;hence, at least part of the urinary y-globulins maynot be derived from the 7 S serum 72-globulins.The results of studies of the excretion of I131-la-beled y-globulin reported by Webb, Rose, andSehon (2) and Franklin (3) were interpreted bythese authors as showing that urinary y-globulinsare derived, at least in part, from serum 7 Sy-globulins. They reported a significantly greater,relative amount of radioactivity in the urinaryy-globulins than was found in the present study;the reason for this discrepancy is not clear. Thelong half-life of serum y-globulins and their highconcentration in serum compared to that in urinemake it difficult to interpret the results obtainedwith labeled y-globulins. Since proteins contain-ing radioactive iodine were found in almost allsections of the starch-block fractionation of urinein this study, it is possible that the degradationof the iodinated, isologous, serum 7 S -y-globulinseither released iodide that was then bound to pro-teins unrelated to y-globulins, or yielded frag-ments with different electrophoretic mobilities.The first possibility is unlikely, since control ex-periments showed that adsorption of iodide byurinary proteins is not random but preferential;the second possibility cannot be ascertained withthe present data. The reason for the higher per-centage of counts found in the proteins with amobility similar to that of serum y-globulins inthe urine of the agammaglobulinemic patient com-pared to that of the normal volunteer is not clear,but agrees with the immunochemical evidence thatthe urinary y-globulins are not homogeneous andmay have multiple origins.

1350

CHARACTERIZATIONOF ANTIBODIES IN NORMALHUMANURINE

The lack of relation between the amount of ra-dioactivity and the protein concentration in thedifferent sucrose-gradient fractions of the urinaryy-globulins from the normal individual may berelated to the results found by analyses of pep-tides, which suggested that the relative tyrosinecontent of the urinary y-globulins may vary withthe molecular weight of the protein. These re-sults appear to differ from those of Stevenson(27), who may, however, have been dealing witha different group of urinary y-globulins.

The low-molecular-weight y-globulins of urinemay be derived in whole or in part from the low-molecular-weight y-globulins described in normalhuman serum by Berggard (28). The origin ofthe low-molecular-weight y-globulins of serum isnot known. It is possible that low-molecular-weight y-globulins of serum are normally producedindependently of the 7 S y-globulins and, owing totheir small size, are cleared rapidly by the kidney."1

SUMMARY

Urines from persons immunized with tetanustoxoid or poliovirus vaccine were shown to con-tain low-molecular-weight antibodies specific forthe immunizing antigens. Neutralizing activityto poliovirus was found in y-globulins with anSw, 20 of 1.7. Precipitating antibody activityagainst tetanus toxoid was found in proteins witha molecular weight of 12,900 and an S,,20 of 0.92.

The peptide patterns after tryptic digestion ofurinary y-globulins differed from those after tryp-tic digestion of serum 7 S y-globulins. Immuno-chemical and isotopic studies indicated that thelow-molecular-weight y-globulins are heterogene-ous and may have multiple origins.

REFERENCES

1. Merler, E., J. S. Remington, D. Gitlin, and M. Fin-land. Enzymatic hydrolysis of low molecularweight antibodies from normal human urine.Fed. Proc. 1962, 21, 21.

11 Recent separation of low-molecular-weight urinary,y-globulins on columns of Sephadex G-75 resin from thecomponents of higher molecular weight supports assump-tion of the value of V used in the determination of molecu-lar weight and indicates that the value reported here maybe indeed correct, and not apparently faulty because ofthe presence of lipoproteins.

2. Webb, T., B. Rose, and A. H. Sehon. Biocolloidsin normal human urine. II. Physiochemical andimmunochemical characteristics. Canad. J. Bio-chem. 1958, 36, 1167.

3. Franklin, E. C. Physiochemical and immunologicstudies of gamma globulins of normal humanurine. J. clin. Invest. 1959, 38, 2159.

4. Rowe, D. S., and J. F. Soothill. Serum proteins innormal urine. Clin. Sci. 1961, 21, 75.

5. Remington, J. S., and M. Finland. Precipitatingantibody in normal human urine. Proc. Soc. exp.Biol. (N. Y.) 1961, 107, 765.

6. Lerner, A. M., J. S. Remington, and M. Finland.Neutralizing antibody to polioviruses in normalhuman urine. J. clin. Invest. 1962, 41, 805.

7. Porter, R. R. The hydrolysis of rabbit -y-globulinand antibodies with crystalline papain. Biochem.J. 1959,-73, 119.

8. Remington, J. S., E. Merler, A. M. Lerner, D.Gitlin, and M. Finland. Antibodies of low mo-lecular weight in normal human urine. Nature(Lond.) 1962, 194, 407.

9. Flodin, P., B. Gelotte, and J. Porath. A method forconcentrating solutes of high molecular weight.Nature (Lond.) 1960, 188, 493.

10. Shapiro, J. High-rate laboratory filtration withBuchner funnels. Science 1961, 133, 1828.

11. Scheidegger, J. J. Une micro-methode de l'immunoelectrophorese. Int. Arch. Allergy 1955, 7, 103.

12. Ouchterlony, 0. Antigen antibody reactions in gels.IV. Types of reactions in coordinated systems ofdiffusion. Acta path. microbiol. scand. 1953, 32,231.

13. Preer, J. R., Jr. A quantitative study of a techniqueof double diffusion in agar. J. Immunol. 1956, 77,52.

14. Feinberg, J. G. Identification, discrimination andquantification in Ouchterlony gel plates. Int. Arch.Allergy 1957, 11, 129.

15. Hirs, C. H. W. The oxidation of ribonuclease withperformic acid. J. biol. Chem. 1956, 219, 611.

16. Jacobsen, C. F., J. Leonis, K. Linderstrom-Lang, andM. Ottesen. The pH-stat and its use in biochem-istry. Methods biochem. Anal. 1957, 4, 171.

17. Gitlin, D., and E. Merler. A comparison of thepeptides released from related rabbit antibodies byenzymatic hydrolysis. J. exp. Med. 1961, 114, 217.

18. Acher, R., and C. Crocker. Reactions color~ees speci-fiques de l'arginine et de la tyrosine realisees apreschromatographie sur papier. Biochim biophys.Acta (Amst.) 1952, 9, 704.

19. Edelman, G. M., H. G. Kunkel, and E. C. Franklin.Interaction of the rheumatoid factor with anti-gen-antibody complexes and aggregated gammaglobulin. J. exp. Med. 1958, 108, 105.

20. Klainer, S. M., and G. Kegeles. Simultaneous de-terminations of molecular weights and sedimenta-tion constants. J. phys. Chem. 1955, 59, 952.

1351

EZIO MERLER, JACK S. REMINGTON, MAXWELLFINLAND, AND DAVID GITLIN

21. Fujita, H. Effects of a concentration dependence ofthe sedimentation coefficient in velocity ultracen-trifugation. J. Chem. Phys. 1956, 24, 1084.

22. Kohn, J. A micro-electrophoretic method. Nature(Lond.) 1958, 181, 839.

23. Kunkel, H. G. Zone electrophoresis in Methods ofBiochemical Analysis, D. Glick, Ed. New York,Interscience, 1954, vol. 1, p. 141.

24. Gitlin, D., C. A. Janeway, and L. E. Farr. Studies onthe metabolism of plasma proteins in the nephroticsyndrome. I. Albumin Y-globulin and iron-bind-ing globulin. J. clin. Invest. 1956, 35, 44.

25. Rigas, D. A., and C. G. Heller. The amount andnature of urinary proteins in normal human sub-jects. J. clin. Invest. 1951, 30, 853.

26. Pedersen, K. O., in Ultracentrifugal Studies of Se-rum and Serum Fractions, K. 0. Pedersen, Ed.Uppsala, Almquist & Viksells, 1945, p. 172.

27. Stevenson, G. T. Further studies of the gamma-re-lated proteins of normal urine. J. clin. Invest.1962, 41, 1190.

28. Berggard, I. On a 'y-globulin of low molecularweight in normal human plasma and urine. Clin.chim. Acta 1961, 6, 545.

1352