THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 251, No. 21, Issue of November 10, pp. 6798-6806, 1976

Printed in U.S.A.

Studies on the Structural Localization of Rabbit H Chain Allotypic Determinants Controlled by the a Locus

PURIFICATION AND IMMUNOLOGICAL PROPERTIES OF AN IMMUNOPEPTIDE BEARING a3 ALLOTYPIC DETERMINANT(S)

(Received for publication, January 21, 1976)

AFTAB A. ANSARI,* MARIELLA CARTA~ORCINI,$ ROSE G. MAGE,* AND ETTORE APPELLA$, 7

From the *Laboratory of Immunology, National Institute of Allergy and Infectious Diseases, $.Istituto Superiore Sanita, Viale Regina Elena, Rome, Italy, and the 0 Laboratory of Cell Biology, National Cancer Institute, National Institutes of Health, Bethesda, Maryland 20014

An immunopeptide bearing a3 allotypic determinant(s) was isolated from the y chain of an a3 homozygous rabbit (G222-2) immunized with type III pneumococcal vaccine. Immunological properties of peptides were studied using a radioimmunoassay that involved inhibition by these peptides of a reaction between ““I-labeled anti-a3 antibody and Sepharose-bound a3 immunoglobulin G (IgG). The y chain was isolated from IgG of restricted heterogeneity and then citraconylated and digested with trypsin. The tryptic digest (TDl) was passed through an anti-a3 immunoabsorbent column either directly or after an intermediate step of Sephadex G-75 chromatography. The bound peptides (Tl) were eluted with 0.1 M acetic acid and further digested with tryps@. The digest (TD2) was again run on the anti-a3 immunoabsorbent column to purify the bound immunopeptide T2. In the radioimmunoassay this immunopeptide was found to have major a3 det,erminant(s). Its molecular weight was found to be approximately 6,000, which decreased to about 3,000 after reduction and alkylation. These data, together with NH,- and COOH-terminal analyses and cysteine peptide mapping, demonstrated that T2 is composed of two polypeptide chains linked by a disulfide bond, one from the cysteine 22 region having lysine at the COOH terminus and the other from the cysteine 92 region having arginine at the COOH terminus. The lysine peptide was separated from the arginine peptide and its NH,-terminal sequence was found to be Gly-Asx-Glx-Ser-Thr-Cys. Since the cysteine is at position 22, the lysine peptide starts at position 17. It has approximately 22 residues. The framework sequence from 17 to 20 is different from those reported so far. In addition, the heavy chain used in these studies has some other unusual features including a histidine, probably in the first hypervariable region. The presence of histidine in the first hypervariable region of rabbit heavy chain has not been reported previously.

The other peptide which is about 30 amino acids in length and ends with arginine 94, probably includes positions 67, 70, 71, 84, and 85 that are believed to have substitutions correlating with a allotypes.

In a hypothetical three-dimensional model of the F, portion of rabbit anti-S111 antibody BS-5, residues 17 to 33 of the lysine peptide and 67 to 79 and 84 to 85 which may be present in the arginine peptide are fully exposed on the surface and are far removed from the antibody combining site.

The antigenic determinants of rabbit IgG’ include isotypic specificities common to all rabbits and allotypic specificities by which rabbits can be subdivided into several groups (1). The heavy chain allotypes al, a2, and a3, detectable by precipitat- ing alloantisera, behave in breeding studies as if controlled by three alleiic genes at an autosomal locus designated as the a locus (reviewed in Refs. 2 and 3). Several studies support the

7 To whom correspondence should be addressed

view that the genes of the a locus are controlling or structural genes for the variable region, but because of the considerable sequence heterogeneity of this region it has been difficult to establish the exact chemical nature of the sequences which define the allotypes (4, 5). Chemical analyses have established that there are reproducible total compositional differences in the Fd fragments and in the heavy chains from al, a2, and a3 normal and antibody IgG pools (6-11). Most of the composi- tional differences are accounted for as differences in the

‘The abbreviations used are: IgG, immunoglobulin G; DITC, p-phenylene diisothiocyanate; Fd, Fd fragment (NH,.terminal half) of

sequence of the first 94 residues from the NH, terminus of the

the heavy chain; H, heavy chain; V,, variable region of H; TPCK, heavy chain (4, 10). Approximately 16 positions in the variable L-1-tosylamido-2.phenylethyl chloromethyl ketone; dansyl, 5-di- regions of heavy chains from pooled IgG have been correlated methylaminonaphthalene-1-sulfonyl. with a locus specificities (4). Work with rabbit antibodies to

6798

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

a3 Immunopeptide 6799

bacterial polysaccharides with limited heterogeneity has con- firmed this correlation for some, but not all of these positions (12-17).

Structural definition of the antigenic determinants con- trolled by the a locus is of decisive importance for understand- ing the genetics of heavy chain variable regions and interpret- ing breeding studies such as those which suggest that recombi-

nation occurs between closely linked genes coding for the variable and constant portions of heavy chains (18, 19). Correlation of structure with specificities can be misleading because the genes for heavy chains are closely linked. Direct proof for the localization of an antigenic determinant could come from isolation of peptides carrying a allotypic specifici- ties. Florent and co-workers (20) succeeded in isolating such peptides from normal heavy chains but because of their heterogeneity, could not determine their sequence or localiza- tion. In the present work, we have confirmed and extended their observations. We have found that y chains from anti- bodies of restricted heterogeneity elicited by immunization with pneumococcal vaccines facilitate the study of allotype- related peptides by eliminating some of the problems inherent in the use of heterogeneous pools.

EXPERIMENTAL PROCEDURE

MuteriaIs-Ribonuclease and insulin B chain were obtained from Sigma Chemical Co. (St. Louis, MO.). TPCK-trypsin, carboxypepti- dase A, and carboxypeptidase B were purchased from Worthington (Freehold, N. J.). Guanidine hydrochloride and urea were ultrapure grade from Schwarz/Mann (Orangeburg, N. Y.). Aminopropyl glass and p-phenylene diisothiocyanate were from Pierce (Rockford, 11.). Polyamide sheets were obtained from Cheng Chin Trading Co., Ltd., No. 75 Sec. 1, Hankow St. Taipei, Taiwan.

Rabbit Antisera-Rabbits of known allotype were bred and main- tained in our own colonies at the National Institutes of Health, Bethesda, Md. They were immunized with type III (285RE-3 al, AH80-5 a3, G222-2 a3) or Type VIII (F196-1 a2) pneumococcal vaccines as previously described (21). Intravenous injections of increasing doses of formalinized Pneumococci were given three times a week for 4 weeks (Course 1). Rabbits were then maintained on an immunization schedule of 0.5 ml of vaccine (0.25 mg of bacterial protein) adminis- tered once a week. All animals were bled at weekly intervals, before antigen injection. Some animals were rested and subsequently restimulated by a course of immunization identical to the first one.

Sera were examined by microcellulose acetate electrophoresis for evidence of IgG components of limited heterogeneity (22). Isoelectric focusing in 5% polyacrylamide gel using ampholines (pH 3 to 10) was performed by the method of Awdeh et al. (23). After electrophoresis, immunoglobulin in the gels was precipitated in 18% Na,SO, and gels were fixed, stained, and dried as previously described (21, 24).

Antisera against purified normal a3 IgG were raised in a*cz2b4b4 rabbits as described by Dray et al. (25).

IgG Fractions-IgG was prepared from antisera by precipitation with 18% sodium sulfate, followed by chromatography on DEAE-cel- lulose (26). The conditions for elution of components of limited heterogeneity depended upon their particular charge properties. Typi- cally, the crude globulin fraction was applied in 0.01 M sodium phosphate buffer, pH 6.8, and buffers of increasing ionic strength (e.g. 0.005 M increments) and pH (to pH 7.5) were used to elute bound components.

Heavy Chains and Dyptic Digestion-Heavy and light chains were isolated from IgG by gel filtration on Sephadex G-100 in 6 M urea containing 1 M acetic acid, after oxidative sulfitolysis of interchain disulfide bonds as described previously (27). For citraconylation (28), the heavy chains were suspended in water, the pH adjusted to 8.0 with 0.5 N NaOH, and a 50.fold molar excess of citraconic anhydride over the lysine content was added in 50.~1 portions with stirring at room temperature. The pH was maintained between 7.5 and 8.0 by addition of 5 N NaOH with a pH-stat. When the base uptake ceased, the clear solution was dialyzed against several changes of 0.1 M NH,HCO,, pH 8.0 at 5”. Tryptic digestion was carried out with TPCK-treated trypsin (Worthington) in 0.1 M NH,HCO, at 37” with an enzyme to substrate ratio of 1:lOO (w/w). The reaction was stopped after 12 h by

lyophilization. To remove citraconyl groups (29), the tryptic digest was suspended in 0.04 M pyridine/acetate buffer, pH 3.5, and left at room temperature for 24 h with stirring. The suspension was then lyophil- ized.

Affinity Chromatography-For isolation of active peptides from tryptic digests by affinity chromatography, we first coupled purified normal a3 IgG to Sepharose by the cyanogen bromide method as described by Goldman and Mage (30). The antigen/Sepharose column was then used to prepare pure anti-a locus allotype antibodies from pooled antiserum. The purified antibodies were in turn coupled to Sepharose 2B by a similar procedure with the following slight modifications. The Sepharose was activated with CNBr for 1 ‘h h; the ratio of IgG to Sepharose was about 3 mg of IgG/ml of activated Sepharose; after coupling for 2 h at room temperature and overnight in the cold, the antibody/Sepharose was washed with borate buffer and allowed to react with 1 M ethanolamine for 2 to 3 h at room temperature with stirring, to block any activated sites still available on the Sepharose. The immunoabsorbent was then washed extensively with 0.1 M NH,HCO, and packed in a chromatographic column. Before each experiment, the column was washed with 0.1 M acetic acid and then reequilibrated with 0.1 M NH,HCO,. Usually, 40 to 50 mg of tryptic digest of heavy chain was dissolved in 3 ml of 0.1 M NH,HCO, and applied to an antibody/Sepharose column containing 250 mg of bound antibodies. The nonbound peptides were washed through with 0.1 M NH,HC03, pH 8, at a flow rate of approximately 30 ml/h. The bound, active fraction was eluted with 0.1 M acetic acid, pH 2.9 and an increased flow rate of 100 ml/hr. The eluted fractions (immunopeptide Tl) were lyophilized and examined for antigenic activity. The anti- body/Sepharose columns showed considerable loss of capacity to absorb active peptides if used more than five times.

The immunopeptide (Tl) was further digested with TPCK-trypsin (TD2) and the active peptide (T2) purified on the immunoabsorbent column as described above.

Gel Diffusion-The mixture of tryptic peptides as well as subse- quently isolated fractions were initially screened for antigenic activity by an Ouchterlony-type immunodiffusion test in which control antigen and antibody are arranged to produce a symmetrical pattern. This pattern is distorted by asymmetric introduction of an inhibiting or precipitating antigen (20). Where possible, as specified under “Results,” several antisera produced in different homozygous and heterozygous recipients were utilized for activity tests, and controls to test for nonspecific inhibition of inappropriate allotype/anti-allotype precipitation were included.

Radioimmunoassay-Radiolabeling of purified anti-a3 antibody was done by the lactoperoxidase method (31) at pH 5.5 (0.05 M sodium acetate buffer). The protein (2 x 10-O mol) was mixed with 2 x 10-l’ mol of lactoperoxidase, 2 x lo-’ mol of sodium iodide (carrier), and 0.2 mCi of “‘1 in the form of NaI in a total volume of 480 ~1. Iodination was started by the addition of 20 ~1 of 10m3 M H,O,. After 1 h at room temperature, the solution was dialyzed in the cold against several changes of 0.05 M sodium acetate buffer, pH 5.5, and finally against 1% NH,HCO,. After dialysis, 10% bovine serum albumin in 1% NH,HCO, was added to the solution and the counts were diluted by the addition of unlabeled antibody to a specific activity of 30,000 cpm/pg. About 40% of the added iodide was determined to have bound to the protein, and 99% of the counts in the protein sample were precipitable by 10% trichloroacetic acid. The direct binding curve (see below) between a3 IgG immunoabsorbent and ‘251-labeled purified anti-a3 antibody showed that only 25% of the antibody bound to the immunoabsorbent, probably because of exposure to low pH values during antibody purification and exposure to oxidizing agent and high radioactivity during its labeling. The labeled antibody was repurified by mixing it with a3 IgG-immunoabsorbent for 1 h followed by centrifugation and washing two times with 1% NH,HCO,. The antibody was then eluted sequentially wit\ glycine/HCl buffer (ionic strength O.l), pH 2.8, followed by pH 2.4. Only 0.3% of the bound counts were eluted at pH 2.8, whereas 8.3% eluted at pH 2.4. The latter fraction was used in the radioimmunoassays. Its maximum binding to a3 IgG immunoabsorb- ent was found to be 80%.

For direct binding curves, 50 ~1 of the labeled antibody solution in 10% albumin (pH adjusted to 8) containing about 15 ng of antibody (4,500 cpm) were mixed in microfuge tubes with different volumes of 50. to 100.times diluted a3 IgG-immunoabsorbent (approximately 1 mg of IgG/ml of gel), and the volume made up to 200 ~1 with 1% NH,HCO,. The reaction was carried out at room temperature with mixing on a rotary wheel for 3 to 4 h. The tubes were then centrifuged at room temperature and 100 ~1 of supernatant was transferred to a

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

6800 a3 Immunopeptide

clean tube. Radioactivity was determined in each supernatant and in the portion remaining in each tube and the percentage of counts calculated. A blank was also run in which no immunoabsorbent was added. Labeled antibody solutions were centrifuged immediately before use in the assay in order to obtain low blank values (+2%). Solutions used without prior centrifugation gave high blank values (>5%).

The concentrations of all the inhibitors used in the inhibition assay were determined by amino acid analysis. For inhibition assays, varying amounts of inhibitor and a fixed amount of labeled antibody (50 al, 15 ng, 4,500 cpm) were mixed in microfuge tubes in a total volume of 185 ~1. The reactants were incubated with mixing overnight at room temperature in order to allow sufficient time for the inhibitors to react with the antibody. Fifteen microliters of the immunoabsorbent (an amount giving approximately 60% of the maximum binding as determined from the direct binding curve) were added and stirring continued for 6 h. The tubes were centrifuged and 100 ~1 of supernatant was withdrawn and the per cent of counts per min bound was calculated. Each reaction mixture contained 2.5% albumin. For each inhibitor concentration, four tubes were set up: two of them received 15 pl of immunoabsorbent while the other two (blanks) did not receive any immunoabsorbent, and thus served as controls to check if there was any precipitation of the inhibitor.antibody complex. Among the inhibitors tested, IgG, CH, TDl, and TD2 gave less than 5% blank values in the concentration range studied. The remaining two inhibi- tors, H and TD2, also gave low blank values ( <5%) except that H gave 10 to 12% precipitation of labeled antibody in the concentration range 40 x IO-’ to 170 x lo-” nmol and TD2 gave 8 to 11% precipitation at 100 x 10.’ and 200 x lo-” nmol concentrations. All blank values were subtracted from the respective test sample values. A control was also set up in which no inhibitor was added. The percentage of inhibition was calculated as follows. % Inhibition = (A - P) x 100/A where A and Pare the per cent of counts per min bound in absence and in presence of inhibitor, respectively.

DEAE-cellulose Chromatography-The immunopeptide T2 was reduced and ‘“C-carboxymethylated and chromatographed on a DEAE-cellulose column (1 x 30 cm) in the presence of 2 M urea. The starting buffer was 0.01 M Tris/HCl, 2 M urea, pH 8.35. A gradient from 0.01 to 0.02 M Tris/HCl buffer was generated with the help of an Ultrograd (LKB) using a 4-h setting at a flow rate of 25 ml/h. The fractions were analyzed by measuring radioactivity.

Determination of Molecular Weight-The molecular weights of fragments before and after full reduction and alkylation were esti- mated by gel filtration on an agarose (Bio-Gel A-1.5) column (250 x 0.8 cm) in 5 M guanidine hydrochloride containing 0.2 M NH,HCO,. Full reduction and alkylation of fragments was performed in 0.5 M Tris, 5 M

guanidine hydrochloride, pH 8.2, using 0.01 M dithiothreitol for the reduction, and 0.022 M iodoacetamide containing iodo [14C]acetamide (1.5 x lo6 cpm/pmol at 80% counting efficiency) for the alkylation. The guanidine/agarose column was calibrated with the following reduced and carboxymethylated standards of known molecular weights: mouse y heavy chain (53,000), mouse light chain (23,000), ribonuclease (13,700), and insulin B chain (2,900). For normalization of the gel filtration data we have used an expression in which relative elution volumes (R,) of the peptides were calculated by dividing the elution volumes (V,) of the peptides by the elution volume of low molecular weight salts which equals void volume (VJ plus inner volume (V,); thus R, = V,I(V, + VJ.

Amino Acid Analysis-Amino acid analyses were performed on a Durrum D-500 amino acid analyzer after hydrolysis with 6 N HCl for 24 h in evacuated tubes at 105”.

NH,-terminal Analysis-Approximately 2 to 4 nmol of peptides in 5 al of 0.1 M NaHCOs were dansylated (32) by the addition of 5 /.d of dansyl chloride reagent (1 mg/ml in acetone). The reaction was carried out at 37” for 1 h and then dried under vacuum. A 5 to 10 ~1 volume of 6 N HCl was added to the tube and hydrolysis was performed after sealing the tube for 16 h at 105”. After hydrolysis the tube was opened and its contents were evaporated to dryness. The sample was trans- ferred with ethyl acetate to the corner of a polyamide sheet (5 x 5 cm) for chromatography. The sheet was developed first with a mixture of heptane, butanol, and formic acid (1O:lO:l) and then with 0.15 M NH,OH in the vertical direction. The dansylated amino acids were identified by comparing with standard derivatives of the amino acids chromatographed on the other side of the polyamide sheet. The spots were visualized under ultraviolet light.

COOH-terminal Analysis-For COOH-terminal analysis (33), 1 to 4 nmol of peptide in 0.2 M N-ethylmorpholine buffer, pH 8.3, were

digested with carboxypeptidase B (enzyme to substrate ratio, l/50, w/w) for 2 to 4 h at 37”. A blank sample, enzyme alone, was also analyzed. The reaction was stopped by the addition of 0.1 ml of glacial acetic acid. It was then lyophilized and subjected to amino acid analysis on a Durrum D-500 amino acid analyzer in order to identify the released amino acids. In some cases, carboxypeptidase A was also added along with carboxypeptidase B to release amino acids other than arginine and lysine.

Cysteine-Peptide Map--S- [“CJcarboxymethyl cysteine-containing peptides were identified as outlined by Mole et al. (34). The peptide material was digested with trypsin plus chymotrypsin for 6 h in 1% NH,HCO, at 37”. The digest was lyophilized, dissolved in water and electrophoresed on Whatman No. 3 paper at pH 3.5. Radioautography on Kodak x-ray film Blue Brand BB54 was performed after electropho- resis and the ‘Y-containing peptides were identified on the basis of their mobilities given by O’Donnell et al. (35).

Attachment of Lysine Peptide to Glass Support and Solid Phase Edman Degradation-The peptide sample (35 nmol) was dissolved in 250 ~1 of a buffer containing water, pyridine, and N-methylmorpholine (20:30:3). Forty-five milligrams of DITC (p-phenylene diisothiocya- nate) aminopropyl glass (36) were mixed and stirred under nitrogen at 45” for 1 h. Free sites were then blocked by the addition of 250 ~1 of ethanolamine followed by stirring under nitrogen at 45” for another hour. The derivatized material was washed three times with dimethyl- formamide and three times with methanol. After vacuum drying, more than 50% of the peptide was found to have coupled with the glass support. The dried support was poured into a column (3 x 100 mm) and subjected to degradation on a Sequamat 12K Sequencer (Sequa- mat Inc., Boston). Liberated thiazolinones were counted for ‘“C in order to identify “C-labeled S-carboxymethyl cysteine and were back hydrolyzed with HI (37) for 4 h at 150” for identification by amino acid analysis on Durrum D-500 analyzer.

Liquid Phase Edman Degradation-Automated Edman degradation of reduced and S-[‘“Clcarboxymethylated T2 was carried out with a Beckman Sequencer model 890-B. The liberated thiazolinones were counted for ‘“C to identify cysteine. For some steps an aliquot of the thiazolinone was converted to the phenylthiohydantoin derivative and analyzed by gas chromatography (38). The remainder of the thiazoli- none sample was regenerated to the free amino acid by HI hydrolysis (37) and chromatographed on a Durrum D-500 analyzer.

Tryptophan Content-The tryptophan content of the immunopep- tide T2 was determined by hydrolysis with 4 N methanesulfonic acid in the presence of 0.2% 3-(2-aminoethyl)indole in uacuo (10 mm Hg) at 115” for 24 and 72 h (39). The hydrolysates were analyzed on the basic column (20 cm) of a JEOL amino acid analyzer at 57”.

RESULTS

Antipneumococcal antisera from rabbits homozygous at the a locus were selected because they appeared partially or highly restricted in heterogeneity by microcellulose acetate electro- phoresis (22). IgG fractions were prepared from these sera and examined by isoelectric focusing (Fig. 1). Heavy chains were prepared from fractions containing major components of lim- ited heterogeneity, citraconylated, and digested with trypsin.

Separation of Active Peptides by Sephadex G-75 Gel Filtra- tion-Initial attempts to separate tryptic peptides from a3 heavy chain of rabbit G222-2 (Sample I, a3 in Fig. 1) were made utilizing Sephadex G-75 in 0.1 M NH,HCO,. Fig. 2 shows a typical separation. The mixture of tryptic peptides and each of the major ultraviolet absorbing fractions eluted from the Sephadex column were assayed for antigenic activity. Results similar to those shown in Fig. 3 were obtained using six different anti-a3 antisera made in a2b4, a2b5, alb4 and ala2b5 recipients. The controls shown in the figure were a3b4 anti-a2 and a3b4b5 anti-al antisera reacting with a2 and al control antigens (IgG), respectively. It can be seen in Fig. 3 that specific inhibitory activity was clearly present in the whole tryptic digest and in Sephadex G-75, peak IV. Some activity was also present in peaks II and III. The inhibitory activity was specific for the a3 allotype since no inhibition was observed of precipitation of al with anti-al or a2 with anti-a2 (Fig. 3).

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

a3 Immunopeptide 6801

peptides were bound and eluted from specific immunoabsorb- ents. In the present paper, further studies done on the pep- tides from G-222-2 heavy chain (a3) are reported.

Peak IV (Fig. 3) which displayed strongest inhibitory activ- ity was rechromatographed on the same Sephadex column and its purity checked by isoelectric focusing. Although a single major band was observed, a scan of cysteine-containing pep- tides after tryptic and chymotryptic digestions gave clear evidence for contamination by peptides from the constant region of the heavy chain. Moreover, automated sequence analysis of the peptide also revealed a sequence starting with alanine 124 in the heavy chain constant region.

Immunopeptide Tl-Further purification of the active pep- tide was achieved with an immunoabsorbent column prepared by coupling anti-a3 antibody to Sepharose 2B. The active fraction (peak IV) from the Sephadex G-75 column, or the whole tryptic digest of heavy chain was applied on the immunoabsorbent column. Unbound peptides were washed through with 0.1 M NH,HCO,, pH 8, and the bound active fraction was eluted with 0.1 M acetic acid, pH 2.9. The efficiency of the immunoabsorbent decreased with reuse. The yields of active peptides (immunopeptide Tl) from immuno- absorbent columns decreased from 0.8 to 0.1 mol/mol of heavy chain on repeated chromatographies of whole tryptic digests of heavy chains on the same immunoabsorbent. The yield was somewhat lower when the intermediate step of G-75 purifica- tion was employed.

FIG. 1. Isoelectric focusing patterns of rabbit antisera and corre- sponding IgG fractions. al, rabbit 285RE-3 immunized with tvne III pneumococcal vaccine; a2, rabbit F196-1 immunized with type VIII pneumococcal vaccine; ~3, Samples I and II, rabbits G222-2 and AH80-5 immunized with type III pneumococcal vaccine.

Immunopeptide Tl was inhibitory in the gel diffusion tests described for peak IV. Its molecular weight as determined by chromatography on a calibrated agarose column in 5 M

guanidine hydrochloride was found to be approximately 9,000 (Fig. 4).

EFFLUENT VOLUME ( ml)

Zmmunopeptide TZ-Exposure of the Tl peptide to low pH values during its purification through the immunoabsorbent column also hydrolyzed off citraconyl blocking groups and generated free amino groups of lysine. Thus it was now subjected to a second tryptic digestion to cleave newly exposed lysine peptide bonds. The resulting digest (TD2) was passed through the anti-a3-immunoabsorbent column to purify the bound immunopeptide T2. This immunopeptide also had inhibitory activity detected by the immunodiffusion method. Its activity was also assayed by a radioimmunoassay in which inhibition by the immunopeptide of the binding of ‘Y-labeled anti-a3 antibody to Sepharose-bound a3 IgG was measured quantitatively. The inhibition curves for the different inhibi- tors tested are shown in Fig. 5, which also includes points for heterologous inhibition by a2 IgG, H chain, and H chain tryptic digest. In each case, the heterologous inhibition was less than 10%. Duplicate or triplicate measurements of counts per min bound generally agreed within 2%. The straight lines in Fig. 5 were drawn by the method of least squares.

FIG. 2. Elution profile from a Sephadex G-75 column (2.5 x 1OQ cm) in 0.1 M NH,HCO, of the tryptic digest (200 mg) of G222-2 a3 citraconylated H chain.

Active peptides bearing a3 activity could also be isolated from the tryptic digest of a second anti-S-III heavy chain (AH 80-5 Fig. 1, Sample II, a3). Following similar procedures, peptides bearing al and a2 activity were isolated from the heavy chains of the rabbits 285RE-3 (al) and F196-1 (a2), respectively. In each case, at least one peak on G-75 Sephadex chromatography showed strong inhibitory activity. Immuno-

The molecular weight of T2 was found to be approximately 6,000 (Fig. 4). The amino acid composition of T2 is given in Table I (first column). After full reduction and alkylation, a single peak of approximately 3,000 molecular weight was obtained (Fig. 4). NH,-terminal analysis of T2 by the dansyla- tion method as well as by automated Edman degradation gave two amino acids, glycine and threonine, in equal amounts, and a third amino acid, serine, in traces. COOH-terminal analysis of T2 by carboxypeptidase B treatment for 4 h gave two amiho acids, arginine and lysine, in about 90% yield. These data suggest that the immunopeptide T2 is composed of two chains of about 3,000 daltons each, held together by a disulfide bond.

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

6802 a3 Immunopeptide

FIG. 3. Gel diffusion test for the inhibitory activity of the tryptic digest of the citraconylated H chain (a3H) and the fractions (I to VII) obtained from the Sephadex G-75 column (Fig. 2). The upper panel shows specific inhibition of a3-anti-a3 reaction. The lower panels show heterologous inhibitions (specificity controls) of al-anti-al reaction and a2-anti-a2 reaction.

Convincing evidence as to the localization of the immunopeptide in the heavy chain sequence was obtained by radioautographic analysis of labeled half-cystine-containing peptides. After electrophoresis at pH 3.5 and radioautography, “C-containing peptides were identified on the basis of their mobilities (35). In the radioautograph shown in Fig. 6, no half-cystine-peptides with mobilities characteristic of those from the constant regions were found. The only half-cystine- peptides found had mobilities compatible with peptides con- taining cysteines 22 and 92 of the variable region of heavy chain. The two bands in the cysteine 22 region indicate some sequence heterogeneity as has been observed previously with normal immunoglobulins (34). Further evidence that cysteine 92 is indeed present in the T2 fragment was obtained from the kinetic analysis of carboxypeptidase A plus B digestion of reduced and carboxymethylated T2 and a peptide isolated from the reduced and carboxymethylated T2 by DEAE-cel- lulose chromatography (see below) that had lysine at the COOH terminus. When the latter peptide was digested with carboxypeptidase A plus B, the following amino acids were

liberated in decreasing order of yield: lysine, valine, serine, threonine, tyrosine, and phenylalanine. (Tryptophan was also liberated but its amount could not be quantitatively esti- mated.) Digestion of the total reduced and alkylated T2 (mixture of both peptides) liberated, in addition to the above amino acids, arginine, alanine, and carboxyamidomethyl cys- teine. When the values for the amino acids liberated from the lysine peptide were substracted from the values of the amino acids liberated from equimolar amount of T2, the following COOH-terminal sequence for the arginine peptide became evident: Tyr-Phe-Cys-Ala-Arg. These amino acids correspond to the sequence of the rabbit heavy chain from position 90-94.

Separation of the two polypeptide chains from fully reduced and carboxymethylated immunopeptide T2 was attempted on a DEAE-cellulose column. The resulting elution profile is shown in Fig. 7. The peaks were pooled separately and analyzed for amino acid composition and COOH- and NH,-ter- minal residues. COOH-terminal analysis with carboxypepti- dase B gave 1 mol of lysine/mol of each peptide peak. Dansylation showed dansyl glycine as a major spot for each

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

a3 Immunopeptide

peak. Thus it appeared that several variants of the peptide containing COOH-terminal lysine had been separated. The

6803

chain shows variable methionine residues at positions 34 and 79. The amino acid composition of the CNBr fragment contain- ing residues 1 through 34 of the G222-2 H chain isolated according to Friedenson et al. (40) is also given in Table I. It is noteworthy that this fragment has a histidine residue, as does the lysine peptide. The latter peptide, as is seen below, starts at position 17 and does not have any histidine up to position 22. This would mean that the histidine residue is present between 23 and 34.

arginine-containing peptide was not recovered in this experi- ment. The total yield based on 14C counts in peaks 1 to 4 was only 52%. Table I gives the amino acid compositions of peaks 1 through 4 from the DEAE-cellulose column. The composition of a putative arginine-containing peptide is also given in Table I, as calculated by difference. Both the arginine peptide and the lysine peptide have 0.5 mol of methionine/mol of chain. The reported sequence of the variable region of an a3 heavy

IO-

8-

6-

4-

2- f 0

c (3 g '-

% O.%- $ - ti 0.6-

0.4-

0.2-

T2- Reduced and Alkylated B-Chain

Insulin

0.1' I I I I I I 0.4 0.5 0.6 0.7 0.8 0.9 1.0

Rv

FIG. 4. Semilogarithmic plot between molecular weight and R, for a Bio-Gel A-l.5 column (0.8 x 250 cm) in 5 M guanidine hydrochloride containing 0.2 M NH,HCO,. R, = V,l(V, + V,,); TI, immunopeptide Tl; T2, immunopeptide T2.

-

The major peak (pk2) from the DEAE-cellulose column was coupled to derivatized glass beads. Edman degradation carried out on the Sequamat gave evidence for Asx-Glx-Ser-Thr-Cys, in positions 2 through 6. Since the peptide was reduced and Y-carboxymethylated, cysteine could be identified in the liberated thiazolinones simply by determining radioactivity in each step. Fig. 8 shows the counts for each step. The sequence obtained clearly establishes the position of cysteine in our purified lysine peptide.

FIG. 6. Cysteine-peptide map of tryptic-chymotryptic digest of the immunopeptide T2 by the method of Mole et al. (34).

1 10 100 1000 10,000

INHIBITOR CONCENTRATION, nMOLE x lo4

FIG. 5. Inhibition of the reaction be- tween ‘2SI-labeled anti-a3 antibody and Sepharose-bound a3 IgG by different in- hibitors. The inhibitors were: IgG, im- munoglobulin G; H, heavy chain; CH, citraconylated H chain; TDI, tryptic di- gest of CH; TD2, second tryptic digest after removing the blocking citraconyl groups from the immunopeptide Tl; T2, immunopeptide T2. Heterologous in- hibition by a2 IgG (0); a2 H chain (Cl); and tryptic digest of a2 H chain (4.

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

6804 a3 Immunopeptide

TABLE I

Amino acid composition of a3 peptides

Amino acid T2 2”

Peaks CNRr peptide

3” 4” residues 1 rc, 34”

Aspartic acid 4.3 1.4 2.0

Threonine 6.0 2.2 2.2

Serine 7.5 2.0 1.6

Glutamic acid 4.5 1.8 2.0

Proline 3.1 1.8 1.0

Glycine 5.4 2.3 2.0

Alanine 3.4 1.6 1.6

Valine 4.3 1.5 1.9

Methionine 1.0 0.2 0.6

Isoleucine 1.6 0.3 0.3

Leucine 3.4 1.3 1.3

Tyrosine 2.2 0.3 0.6

Phenylalanine 2.1 0.4 0.8

Lysine 2.3 1.4 0.9

Histidine 1.2 0.6 1.0 Arginine 1.0 0.4 0.2

Half-Cystine’ 2.0 1.0 1.0 Tryptophan 1.2 N.D. N.D.

2.0

1.5

1.6 2.3

0.8 2.3

1.3

1.7

0.6

0.7 1.2

0.6

0.8

0.7

0.7 0.2

0.9

N.D.

a Peaks from the DEAE-cellulose column (Fig. 7). b Isolated from C, fragment of G222-2 H chain after reduction and

alkylation followed by Sephadex G-50 chromatography (40). ‘The average is corrected for the per cent of contribution of each

d Calculated from the difference between the compositions of T2 and average of peaks 1 to 4.

e Obtained as homoserine. ‘Determined as carboxymethyl cysteine. g N.D. Not determined. peak based on their yields (Fig. 7).

2.3 2.2 1.9 2.4

1.1 3.0 1.9 * 4.1 2.9 3.2 1.9 5.6

2.5 2.6 2.1 2.4

1.0 3.4 1.1 2.0

2.1 3.3 2.1 3.3

0.7 2.1 1.4 2.0

1.9 3.0 1.7 2.6

0.9 0.3’ 0.5 0.5

1.7 1.0 0.6 1.0

1.2 2.1 1.3 2.1

0.9 1.0 0.6 1.6

1.0 1.1 0.7 1.4

1.0 1.8 1.0 1.3 0.4 0.8 0.7 0.5

0.0 1.1 0.2 0.8

1.0 0.3 1.0 1.0

N.D N.D.g N.D.”

M 2400 -

z

1 5

3 4 /

16C0- /’ -

/

/’ /

m- *-

I

/-- _______- -*

J----l **

I I 0’

__ I ’ I

50 100 150 2co 250 300

ELUTION VOLUME, ml

FIG. 7. Elution profile of fully reduced and [“Clcarboxymethylated T2 from a DEAE-cellulose column (1 x 30 cm) in presence of 2 M urea. Details of the experimental procedure are described in the text. Yields of the different peaks, based on radioactivity, were: peak 1, 12.6%; peak 2, 20.2%; peak 3, 12.2%; peak 4, 7.4%; the last peak that emerged after passing 0.5 M NaCl was not studied further.

These results were confirmed by automated Edman degrada- tion of reduced and “C-carboxymethylated T2 on a Beckman Sequencer (liquid phase). When the liberated thiazolinones were counted for YJ, radioactivity was found in Step 6 and not before (Fig. 8).

DISCUSSION

Numerous attempts have been made to explain the genetic and structural basis for a-locus allotypes in the variable region

AWage of peaks

1 to 4’

Arginine peptided

of rabbit heavy chains. In addition to prototype sequences of al and a3 H chains (4), partial or complete sequences of the V, regions of several rabbit antibodies are available (5, 12-17). Comparisons of these sequences have suggested that three major areas in variable regions of the H chains have amino acid substitutions that seem to correlate with allotypes of the a group. The first region comprises residues 1 to 31, the second region comprises residues 63 to 73 and the third, residues 80 to 85 (4, 5, 41). Our results clearly establish that an immunopep- tide (T2) of approximately 6,000 molecular weight can be isolated and purified from a heavy chain carrying a3 allotypic determinants. This immunopeptide is composed of two chains held together by a disulfide bond. The cysteine-peptide map and partial sequences we obtained suggest that peptide T2 contains cysteine 22 and cysteine 92 of the variable region. One peptide starts at position 17 and contains COOH-terminal lysine, and a total of 22 or 23 residues. The second peptide is about 30 amino acids in length and ends with arginine 94. The fact that our lysine peptide starts at position 17 does not rule out a role for residues 1 to 16 in some allotypic determinants. Kindt et al. (42) observed differences in the ability of homoge- neous antibodies to absorb antisera against the group a allotypes. These results suggested that each group a allotype is expressed as an array of subspecificities of which only a limited number are expressed on individual antibody molecules. In our studies we have used a heavy chain preparation obtained from IgG of limited heterogeneity. Thus all of the a3 determinants might not have been present in the IgG or in immunopeptide T2. All inhibitors tested could inhibit more than 50% of the binding of the anti-a3 antibody to a3 IgG. G222-2 a3 IgG, although partially restricted in heterogeneity, was an excellent inhibitor (98%) and appears to express allotypic determinants recognized by essentially all of the anti-a3 antibodies in our purified fraction. However, it is not certain that 100% inhibi- tion could be reached with the other inhibitors. Some of the

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

a3 Immunopeptide 6805

0 2 4 6 8

STEP NUMBER

FIG. 8. Determination of ‘“C radioactivity in the thiazolinones obtained at the individual steps of automated Edman degradations. Experiment I, Edman degradation of the lysine peptide (Fig. 7, peak 2) using the solid phase Sequencer Sequamat 12K; Experiment 2, Edman degradation of the reduced and [“Clcarboxymethylated T2 using Beckman Sequencer 890 B.

anti-a3 antibodies in the purified anti-a3 fraction studied may have been directed toward determinants which were destroyed or removed at the various steps. However, since more than 50% inhibition (61%) was achieved with peptide T2, this portion of the VH region appears to contain a major antigenic determi- nant or determinants associated with the a3 allotype.

We have shown that the cyanogen bromide Fragment 1-34 of this a3 H chain has a histidine residue. The lysine peptide from T2 also has a histidine, but not in the region 17-22. This indicates that the histidine is present in the region 23-34, most probably in the first hypervariable region. The presence of histidine in the first hypervariable region of rabbit H chain has never been reported previously. However, histidine has been found in the third hypervariable region of two other rabbit H chains from antipneumococcal antibodies (16, 17).

The framework sequence 17-20 of T2 is also very different from the ones reported so far. In addition, most heavy chains so far sequenced have arginine in position 38. However, a human variable region (V H III) has recently been reported (43) to have proline at this position. It is possible that our peptide ends with lysine at position 38, replacing arginine. Carboxypeptidase A plus B digestion of the lysine peptide released lysine, valine, tryptophan, serine, threonine, tyrosine, and phenylalanine. Although the yield of tryptophan was not determined accurate- ly, the yields of the remaining released residues are compatible with the sequence Phe-Tyr-Thr-Ser-Trp-Val-Lys. This sequence shows considerable homology with published se- quences of residues 32 to 38 of al, a2, and a3 heavy chains (4).

The two other regions which are believed to have allotype- related positions (residues 63 to 73 and 80 to 85) are probably

incorporated at least in part in our arginine peptide. Compari- sons of sequences in these regions of the H chains from different allotypes suggest that substitutions at positions 65, 67, 70, 71, 84, and 85 may correlate with a allotypes. The sequence difference Glu-Ala between allotypes al and a3 at position 85 has been followed by a peptide mapping technique and has been shown to be inherited as a genetic trait (44). This chemical marker and the serological marker detected by anti-al and anti-a3 antisera appeared to correlate in the small number of animals bred and analyzed by both serological and chemical typing.

Recent reports by Strosberg et al. (45) and by Mudgett et al. (46) that all the three allotypes, al, a2, and a3, of the a group may be present in the same individual rabbits has rekindled interest in the possibility that these allotypes may not be encoded by allelic structural genes; they could rather be under control of regulatory genes. Structural studies such as ours may eventually help to better define the nature of the variable gene markers.

A three-dimensional model of the F, portion of rabbit anti- SIII antibody BS-5 (16) has been built by Davies and Padlan (47) based on the known crystal structures of the Fab fragments of mouse myeloma protein MOPC 603 (48), human protein REI (49), and the reported amino acid sequence of BS-5 (16). In this model, portions of both peptides of our T2 are brought in close proximity by noncovalent forces and the disulfide bond. The Inv determinants of human K chains also involve two re- gions of the polypeptide chain. Substitutions are found at positions 153 and 191 (50). A three-dimensional model of the light chain shows that these residues are adjacent. Residues 17 to 33 of the lysine peptide and 67 to 79 and 84 to 85 of the arginine peptide which are included in immunopeptide T2 appear to be on the surface fully exposed to the solvent and available for interaction with anti-allotype antibody. More- over, the proposed allotypic region is found on one side of the molecule far removed by an angle of approximately 90” from the surface where hypervariable regions come together to form a postulated antibody combining site.

Acknowledgments-We are grateful to Drs. D. R. Davies and E. A. Padlan for showing us and discussing their model of the F, portion of rabbit anti-S111 antibody BS-5.

REFERENCES 1. Oudin, J. J. (1956) C. R. Acad. Ski. (Paris) 242, 2606-2608. 2. Mage, R. G., Lieberman, R., Potter, M., and Terry, W. D. (1973) in

The Antigens (M. Sela, ed) Vol. 1, pp. 299-376, Academic Press, New York

3. Kindt, T. J. (1975) Adu. Immunol. 21, 35-86 4. Mole, L. E., Jackson, S. A., Porter, R. R., and Wilkinson, J. M.

(1971) 124, 301-318 5. Jaton, J.-C., Braun, D. G., Strosberg, A. D., Haber, E., and

Morris, J. E. (1973) J. Immunol. 111, 1838-1843 6. Koshland, M. E. (1967) Cold Spring Harbor S’ymp. Quant. Biol.

32, 119-127 7. Inman, J. K., and Reisfeld, R. A. (1968) Immunochemistry 5,

415-424 8. Koshland, M. E., Reisfeld, R. A., and Dray, S. (1968) Immuno-

chemistry 5, 471-483 9. Koshland, M. E., Davis, J. J., and Fujita, N. J. (1969) Proc. Natl.

Acad. Sci. U. S. A. 63, 1274-1281 10. Wilkinson, J. M. (1969) Nature 223, 616-617 11. Wilkinson, J. M. (1969) Biochem. J. 112, 173-185 12. Fleishman, J. B. (1971) Biochemistry 10, 2753-2761 13. Jaton, J.-C., and Braun, D. G. (1972) Biochem. J. 130, 539-546 14. Friedenson, B., Appella, E., Roholt, 0. A., and Pressman, D.

(1972) Biochem. Biophys. Res. Commun. 48, 1208-1214 15. Jaton, J.-C., and Haimovich J. (1974) Biochem. J. 139, 281-283

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

6806 a3 Immunopeptide

16. 17. 18.

19. 20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31. 32.

33. 34.

J&on, J-C. (1974) Riochem. J. 143, 723-732 Jaton, J-C. (1975) Rio&em. J. 147, 2X-247 Mage, R. G., Young-Cooper, G. O., and Alexander, C. (1971)

Nature New Hiol. 230, 63-64 Kindt, T. J., and Mandy, W. ,J.(1972)5. Immunol. 108, 1110~1113 Florent, G., van der Loo, W., and Hamers, R. (1970) Eur. J.

Rio&em. 17, 178-182 Winfield, J. B., Pincus, J. H., and Mage, R. G. (1972)J. Immunol.

108 1278-1287 Pincur, ,J. H., ,Jaton, J.-C., Block, K. J., and Haber, E. (19701 J.

immunol. 104, 1143-1148 Awdeh, %. I,., Williamson, A. R., and Askonas, B. A. (1968) Nature

219, 66-67 Winfield, J. B., Mage, R. G., and Alexander, C. B. (1973) J.

Immunol. 110, 729-735 Dray, S., Young, G. O., and Gerald, L. (1963) J. Immunol. 91,

403&415 Levy, H. B., and Sober, H. A. (1960) Proc. Sot. Exp. Hiol. Med.

103, 2.50-252 Rejnek, J., Mage, R. G., and Reisfeld, R. A. (1969) J. Immunol.

102, 638-646 Dixon, H. B. F., and Perham, R. N. (1968) Biochem. J. 109,

312-314 Atassi. M. Z., and Habeeb, A. F. S. A. (1972) Methods Enzymol.

25, 546-553 Goldman, M. B., and Mage, R. G. (1971) Immunochemistry 9,

513-524 Morrison, M. and Bayse, G. S. (1970) Biochemistry 9, 2995-3000 Woods, K. R., and Wang, K.-T (1967) Riochim. Biophys. Acta

133, 369-370 Ambler, R. P. (1972) Methods Enzymol. 25, 262-272 Mole, I,. E., Geier, M. D., and Koshland, M. E. (1975) J. Zmmunol.

114, 1442-1448

XI. O’Donnell, I. J., Frangione, B., and Porter, R. R. (1970) Biochem. J. 116, 261-268

?F. Wachter, E., Machleidt, W., Hofner, H., and Otto, J. (1973) FERS Lett. 35, 97-102

37. Smithies, O., Gibson, D., Fanning, E. M., Goodfliesh, R. M., Gilman, J. G., and Ballantyne, D. L. (1971) Biochemistry 10, 4912-4921

38. Pisano, J. J., and Bronze& T. J. (1969) J. Hiol. Chem. 244, 5597-5607

39. Liu, T. Y., and Chang, Y. H. (1971) J. Biol. Chem. 246, 2842-2848 40. Friedenson, B., Appella, E., Takeda, Y., Roholt, 0. A., and

Pressman, D. (1973) J. Riol. Chem. 248, 707?-7079 41. Cannon, L. E., Margolies, M. N., Rosemblatt, M. S., and Haber,

E. (1976) Immunogenetics in press 42. Kindt, T. J., Seide, R. K., Tack, B. F., and Todd, C. W. (1973) J.

Enp. Med. 138, X-43 43. Fair, D. S., Sledge, C., Krueger, R. G., Mann, K. G., and Hood, E.

(1975) Hi0chemistr.y 14, 5561-5568 44. Mole, L. E. (1975) Hiochem. J. 151, 351-359 45. Strosberg, A. D., Casterman, C. H., van der Loo, W., and Hamers,

R. (1974) J. Immunol. 113, 1313-1318 46. Mudgett, M., Fraser, B. A., and Kindt, T. J. (1975) J. Exp. Med.

141, 1448m 1452 47. Davies, D. R., and Padlan, E. A. (1976) Proceedings oj Royal

Society Medical Symposium October 20 to 22, 1975, Rockfeller University (Raven Press, New York), in press

48. Segal, D. M., Padlan, E. A., Cohen, G. H., Rudikoff, S., Potter, M., and Davies, D. R. (1974) Proc. Natl. Acad. Sci. U. S. A. 71, 4298-4302

49. Epp, O., Colman, P., Fehlhammer, H., Bode, W., Schiffer, M., Huber, R., and Palm, W. (1974) Eur. J. Biochem. 45, 513-524

50. Steinberg, A. G., Milstein, C. P., McLaughlin, C. L., and Solomon, A. (1974) Immunogenetics 1, 108-117

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from

A A Ansari, M Carta-Sorcini, R G Mage and E Appellaimmunopeptide bearing a3 allotypic determinants.

controlled by the a locus. Purification and immunological properties of an Studies on the structural localization of rabbit H chain allotypic determinants

1976, 251:6798-6606.J. Biol. Chem. 

  http://www.jbc.org/content/251/21/6798Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/251/21/6798.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on Decem

ber 7, 2020http://w

ww

.jbc.org/D

ownloaded from