Antibody Combining Sites to a Series of Peptide Determinants of Increasing Size and Defined Structure*

to which a series of peptides of defined length and structure have been attached were used in this study as immunogens and as cross-reacting antigens in order to evaluate to what extent the hapten inhibition method reflects the size and nature of the antigenic determinant to which the animal was exposed. Peptides of the structure @Ala),-Gly (n = 1 to 4) were coupled to RNase and rabbit serum albumin in a one-step synthesis. Peptides of the structures (DAla), s), (Dua),-Gly (n = 1 to 4), and @Ala),-Gly-s-ami-nocaproic acid (n = 1 to 3) were used as inhibitors of the immunological reactions. From cross-precipitation and inhibition experiments it was concluded that the size of the combining sites was in all cases such as to accommodate amino acid residues, and thus the antigenic determinant is a tetrapeptide. Direct is presented that the lysine residue in the protein the antigenic determinant only when the hapten attached is smaller a tetrapeptide.

Proteins to which a series of peptides of defined length and structure have been attached were used in this study as immunogens and as cross-reacting antigens in order to evaluate to what extent the hapten inhibition method reflects the size and nature of the antigenic determinant to which the animal was exposed.
Peptides of the structure @Ala),-Gly (n = 1 to 4) were coupled to RNase and rabbit serum albumin in a one-step synthesis.
From cross-precipitation and inhibition experiments it was concluded that the size of the combining sites was in all cases such as to accommodate 4 amino acid residues, and thus the antigenic determinant is a tetrapeptide.
Direct evidence is presented that the lysine residue in the protein carrier participates in the antigenic determinant only when the hapten attached is smaller than a tetrapeptide. The most exposed portion of the determinant (NH2-terminal amino acid residue) plays an immunodominant role. An increase in the inhibitory capacity in a series of related haptens does not necessarily reflect the properties of the combining site. Nevertheless, the inhibition technique can be used as a tool for estimating the size and nature of the combining sites of antibodies, provided that the structure of the antigenic determinants is known.
The capacity to inhibit antigen-antibody reactions by means of oligosaccharides (2), oligonucleotides (3), and oligopeptides (4)(5)(6)(7)(8) has led to considerable insight into the size and nature of the combining sites in the antibody molecule.
Previous studies of the capacity of alanine peptides to inhibit the poly-L-alanyl and poly-D-alanyl immune systems led to the conclusion that the size of the combining region of the antibodies is such as to accommodate up to 3 to 4 alanine residues (6). Moreover, from * This research was sponsored in part by United States Public Health Service Grant AI-04715 from the National Institute of Allergy and Infectious Diseases, Nat.ional Institutes of Health. A preliminary report of this work was presented at the 38th meeting of the Israel Chemical Society in October 1968 (1). the extent of inhibition of the stereospecific antigen-antibody reactions with alanine peptides containing both L and D residues at defined positions, we reached the conclusion that the region of the antigenic determinant furthest removed from the protein carrier is of paramount importance in determining the specificities of the antibodies formed.
The polypeptidyl proteins used in the above studies as immunogens and as cross-reacting antigens were prepared by polymerization techniques (reaction of the protein with N-carboxy-c-amino acid anhydrides) and, thus, the peptides attached to the protein are of somewhat different sizes. This heterogeneity is relatively restricted, because of the nature of the polymerization reaction (9,10). Such antigenic determinants, differing in their size, may lead to a more heterogeneous population of antibodies.
In order to circumvent this, and to find out to what extent the hapten inhibition data reflect the size and nature of the antigenic determinants to which the animal was exposed, we have now used as immunogens and as cross-reacting antigens proteins to which peptides of defined length and structure have been attached.
A series of peptides of the structure @Ala),-Gly (n = 1 to 4) were coupled to ribonuclease in a one-step synthesis.
The conjugates obtained elicited in rabbits peptide-specific precipitating antibodies.
Results of inhibitions of the precipitin reactions with various chemically related peptides showed that the size of the combining sites of antibodies accommodates 4 amino acid residues, irrespective of the size of the conjugated hapten.
Thus, when the attached peptide was composed of 4 (or more) amino acid residues, the antipeptidyl antibodies did not show any carrier' specificity.
On the other hand, when the hapten attached was smaller than a tetrapeptide the antipeptidyl antibodies exhibited definite carrier specificity.
This study confirms also that the inhibition technique can be used as a tool for estimating the size and nature of the combining site of antibodies, provided that the structure of the antigenic determinants is known.
Otherwise, it is possible to reach erroneous conclusions.
Palladium on powdered charcoal (10% catalyst) was obtained from Matheson, Coleman and Bell. Complete Freund's adjuvant was obtained from Difco. All solvents used were of analytical grade.

GdLKhN~S
(339) Calculated: N 12.4,S 9.4 Found : N 12.1,S 9.5 Synthesis of Peptides General Procedure-The H. (Ala),-Gly .OH peptides were prepared according to the following scheme: Benzyloxycarbonyl-(o-alanyl),-glycine-A solution of 10 mmoles of the succinimide ester of Z-n-Ala in 30 ml of dioxane was added to a solution of 10 mmoles of free peptide (or glycine) and 20 mmoles of sodium bicarbonate in 30 ml of water.
After 18 hours the clear mixture was concentrated under vacuum, mainly to remove dioxane.
Water was added to 100 ml, and acidification to pH 2 with 6 N HCl caused precipitation of the N-blocked peptide (except for Z-DAla-Gly).
After 4 hours at room temperature the precipitate was collected on a sintered glass filter, washed with 0.01 N HCI and water, and then dried under reduced pressure over KOH and HtSOa. The yields were in the range 2   of 85 to 90%. In the case of Z-oAla-Gly, upon acidification of the aqueous solution an oily precipitate was formed. This was extracted into ethyl acetate; the extract was dried over Na2S04 and concentrated.
The blocked peptide was precipitated by adding absolute ether, collected on a sintered glass filter, and dried under reduced pressure. Yield of Z-DAla-Gly, 55%. Analytical data of the peptides are given in Table I. Benzyloxyoarbonyl Peptides Containing e-Aminocaproic Acicl-Z-Gly-e-AC was prepared by coupling Z-Gly-OSu with e-AC in a dioxane-water mixture, as described above. The other peptides, Z-@Ala),-Gly-C-AC (n = 1 to 3) were prepared analogously by reacting R-n-Ala-GSu with the appropriate free peptide. The yields were in the range of 70 to 85%. Analytical data of the peptides are given in Table I.
The purity of beneyloxycarbonyl peptides was determined in three ways: (a) lack of'reaction of the blocked peptides (20 to 50 pmoles) with ninhydrin; (b) determination of neutral equivalents; (c) nitrogen analysis (Table I).
l+ee Peptides-Benzyloxycarbonyl peptides were dissolved in acetic acid (10 to 100 ml per g of peptide, according to solubility) with heating, if necessary. Water (10% by volume) and palladium on charcoal (lo'%, 200 mg) were added, and the hydrogenation was carried out at room temperature and pressure during 12 hours. The catalyst was filtered off, and the solution was concentrated to a few milliliters in a rotary evaporator at 40' (peptide @Ala) rGly came out of solution).
Water was added, and the volume was again reduced to a few milliliters.
On the addition of ethanol and ether the peptides precipitated out, were collected on a sintered glass filter, washed with ethanol and ether, and dried under reduced pressure. The yields were 90 to 100%.
The peptides nAla-Gly, (nAla)zGly, and (nAla)3-Gly were recrystallized from water-ethanol mixture with 86 to 90% yields; nAla-Gly-e-AC was recrystallized from water-ethanolacetone mixture with 50% yield. The peptide Gly-e-AC precipitated as a brown substance; it was dissolved in hot 80% ethanol treated with activated charcoal (Darco G60, Fluka), the charcoal was filtered off, and crystals came out from the clear filtrate (70 y0 yield).
Microanalysis indicated that the free peptides contained varying amounts (0 to 60/, w/w) of bound water. No effort was made to remove it by drying at elevated temperatures.
The purity of the peptides was determined in three ways. (a) One was by high voltage paper electrophoresis at pH 1.4, 3000 volts for 4 hours, where the peptides are separated according to their size (12). Thus contamination by the reactants in the peptide synthesis would show up as additional spots. With ninhydrin reagent only one spot was detected on the electrophoresis sheet at the appropriate distance from the application point (300 mpmoles of peptide load).
(b) By chromatography in n-butyl alcoholacetic acid-water, one spot was detected.
(c) It was also done by determining the ratio of total nitrogen (micro-Kjeldahl) to amino nitrogen (Van Slyke), which should be equal to n, where n = number of amino acid residues in the peptide.
The data for c as well as the specific rotations of the peptides are given in Table II.
The peptide Gly-E-AC fulfilled the criteria a and b for purity, but in Van Slyke's analysis the amount of amino nitrogen found was 30% higher than calculated.
The same value was obtained after repeated recrystallizations from water-ethanol. However, potentiometric titration gave a satisfactory result. Thus, 20.5 pmoles of peptide (41 pmoles of nitrogen, by Kjeldahl analysis) contained: carboxyl group, 19.9 pmoles; amino group, 21 pmoles (pK, = 4.5; pKb = 8.2). Apparently, amino-terminal glycine in this peptide is not determined accurately by the Van Slyke nitrous acid method.
Upon acidification of the reaction mixture the product came out as a thick gel. It was dried and recrystallized from methanol-petroleum ether. The yields were in the range of 50 to 80%. Neutral equivalents of the products agreed with the calculated molecular weights within 3%.
Succinimide Esters of N-o-Nitrophenylsulfenyl-(o-alanyl),glycine-N, N'-Dicyclohexyl carbodiimide (10 mmoles) was added to a solution containing 10 mmoles of Nps-peptide (except for Nps-(DAla)c-Gly, see below) and 10 mmoles of N-hydroxysuccinimide in dimethylformamide (20 to 100 ml, according to solubility of the Nps-peptide), with cooling. After 18 hours at 4" the dicyclohexylurea formed was filtered off and washed with dimethylformamide. Gpon addition of ether to the filtrate the active ester was precipitated, collected on a sintered glass filter, washed with ether, dried under reduced pressure, and used without further purification.
The yields were between 60 to 80%.
Nps-@Ala) rGly is hardly soluble. Five millimoles of peptide were dissolved in 200 ml of dimethylformamide, and 5 mmoles of N-hydroxysuccinimide and of N, N'-dicyclohexyl carbodiimide were added with cooling.
After 18 hours at 4" the reaction mixture was in the form of a gel which liquefied at room temperature. It was used as such for coupling with proteins.
After 18 hours at 4" the reaction mixture was dialyzed against 0.05 M sodium bicarbonate, then dialyzed against water, and lyophilized. The yield of [Nps-@Ala),-Gly],-RNase was 1.1 to 1.2 g. Removal of the Nps-blocking group was accomplished by dissolving 1.0 g of the protein derivative in 60 ml of glacial acetic acid (except for [Nps-vAla-Gly],-RKase for which 98% formic acid was used) containing 2 g of %-LTry, and 10 ml of 2 N HCl in dioxane were added. After 1 hour at room temperature 400 ml of ether were added and the precipitate formed was collected on a sintered glass filter, washed with absolute ethanol and ether, and dried under reduced pressure. Yield of [HCl. @Ala),-Gly],-RXase was 0.85 to 0.95 g. The material was dissolved in 10% aqueous acetic acid (25 ml) and chromatographed on a Sephadex G-25 fine (Pharmacia, cppsala) column (600.ml packing volume, 1 meter long, flow rate 40 ml per hour) equilibrated with the same solvent.
The effluent was monitored by absorption at 280 ink. The effluent containing the protein (which was eluted after 260 ml) was lyophilized.
Yields of the RSA derivatives were similar to the yields of the RSase derivatives.
As the T\jps-@Ala) 4-Gly-OSu was not isolated for preparation of [@Ala) rGly],-l>rotein derivatives, the crude preparation in dimethylformamide (see above, the synthesis of this active ester) was used for coupling to RSase or RSA, as described for the other protein derivatives.
The [Xps-(nAla)h-Gly],-RXase (or RSA) obtained formed a heavy precipitate, and the reaction mixture was dialyzed against 0.05 M sodium bicarbonate and then against water and lyophilized.
In the case of the RNase derivative, removal of the Nps-blocking groups (in 98yi formic acid) and chromatography on Sephadex G-25 in 1O70 acet.ic acid were carried out as usual. From 1 g of RXase, 0.9 g of bhe final product was obtained.
When the lyophilized [Nps-(nAla)4-Gly],-RSA was dissolved in 987, formic acid for the removal of the Xps-blocking groups, the solution obtained was very viscous. Sfter treatment with HCl and precipitation with ether, the hydrochloride derivative gave a very viscous solution in 10% acetic acid. The latter solution was centrifuged, the thick gel which settled down was discarded, and only the supernatant was chromatographed on the Sephadex G-25 column. From 1 g of RSA, 0.5 g of the final product was obtained.
The average number of peptides attached to a protein molecule was determined in three ways: (a) by amino acid analysis of the [@Ala),-Gly],-protein; (b) by the desamination method used for polyalanyl ribonuclease (lo), where lysine residues to which a peptide was coupled are not desaminated.
Therefore, the moles of lysine residues per mole [(oAla),-Gly],-protein found should be equal to the moles of peptide attached; (c) by measuring the absorbance of the [Nps-@Ala),-Gly],-RSA at 390 rnp in a solution of 0.14 M NaCl-0.01 M NaZHP04-0.0034 M HCl, pH 7.2 (with the value of E = 3660 at 390 rnp for the molar absorbance of Nps-D-alanine dicyclohexylammonium salt in the same solution).
The [Nps-@Ala),-Gly],-RNase derivatives are not soluble in the above medium, and their absorbance could not be measured.
The analytical data for the protein derivatives are given in Table III.
Amino &id Analysis-The protein samples were subjected to hydrolysis with 6 M hydrochloric acid in sealed tubes at 110" for 24 hours. The amino acids were then determined by quantitative analysis (16), with the Beckman-Spinco automatic amino acid analyzer, model 120 B.
High Voltage Electrophoresis-Electrophoretic separation of peptides was carried out as described before for alanine peptides (12)~

Spectrophotometric
Measurements-Absorbances of proteins were measured in a Zeiss spectrophotometer, model PM &II. Immunization Procedure-Randomly bred rabbits (2.5 to 3.5 kg) of both sexes were used. The antigens were incorporated in a water in oil adjuvant mixture (17). Equal parts of 2% antigen solution in aqueous 0.9% sodium chloride and complete Freund's adjuvant were homogenized by repeated filling and forcible ejection from a syringe.
After bleeding prior to immunization, 1 ml of the adjuvant-antigen mixture was administered intramuscularly into the thighs of the hind legs of the animal. Four injections were given at lo-day intervals.
Animals were bled 5 days after the third and fourth injections, and then at weekly intervals for another 2 months.
Studies were performed on sera of individual rabbits, and for this purpose the sera from each rabbit were pooled separately.

Quuntifative Precipitin
Studies-To a constant volume of antiserum (0.2 to 0.5 ml) increasing amounts of the precipita.nt dissolved in 0.9y0 sodium chloride were added (the amount of precipitant was determined by Kjeldahl nitrogen analysis, with the factor of 6.25 for conversion from micrograms of nitrogen to micrograms of protein), and the final volume was brought to 1.7 ml with 0.9% sodium chloride.
The reaction mixture was incubated at 37" for 45 min and then kept at 5" for 18 hours. Precipitates formed were separated by centrifugation, washed twice with 2.0 ml of 0.9(J& sodium chloride, and dissolved in 1.1 ml of 0.1 N sodium hydroxide, and the absorbance at 280 rnp was determined. Quantitative Inhibition studies-Inhibition was estimated by measuring the absorbance of the precipitate (in 0.1 N NaOH) formed in a mixture which contained peptide, antiserum, and precipitant.
Constant amounts of antiserum and precipitant were used. These were chosen from the optimal zone of the precipitin curve to yield a precipitate which in 0.1 N NaOH solution would have an absorbance, at 280 rnp, in the range of 0.5 to 0.7. The reagents were reacted as follows: varying amounts of peptide dissolved in 0.9% sodium chloride were added to a constant volume of serum, the volume was adjusted to 1.5 ml with 0.9% sodium chloride, the mixture was incubated at 37" for 45 min, a constant amount of precipitant in 0.2 ml of 0.9% sodium T~RLE III ilnalytical data of (DAla),-Gly-proteins Peptidyl protein DAla-Gly-RNase.. a Determined by quantitative amino acid analysis, and calculated on the basis of 12 moles of alanine residues and 3 moles of glycine residues per mole of RNase, and of 54 moles of alanine residues and 18 moles of glycine residues per mole of RSA.
6 Average of the values calculated for excess moles of alanine residues divided by the number of alanine residues in one peptide chain, and of excess moles of glycine residues.
c Obtained by determining the lysine content after desamination of the peptidyl protein with nitrous acid (9).
d Calculated from the molar absorbance of the Nps-group at 390 rnw (e = 3660).
additional 45 min at 37" and then at 4" for 18 hours. The precipitates formed were separated by centrifugation and mere treated as described above for quantitative precipitin.

Synthesis and Characterization of Peptidyl Proteins
For the purpose of the present study a series of peptidyl proteins were prepared, to be used as immunogens or cross-reacting antigens (precipitants).
The peptides were prepared by stepwise synthesis.
We used n-alanine in the synthesis of the peptides as such determinants would be more resistant to alteration in the organism, and in order to prevent their proteolytic digestion by serum (6). The COOH-terminal residue of the peptides attached to proteins was always glycine, in order to avoid possible racemization during activation of the carboxyl function.
The carboxyl function was activated in the form of succinimide ester, as such peptide derivatives are known to react well in aqueous solutions (Is), which had to be used because of the protein moiety.
Indeed, the coupling reaction was very efficient, as apparent from data given under "Materials and Methods" and from Table III. For blocking of the amino function of the peptides, Nps-was used, as it is effectively removed under mild acidic conditions (14). Benzyloxycarbonyl-L-tryptophan was added to the reaction mixture at the stage of the Nps-removal, as the Npsreleased might react with some amino acid residues in the protein.
Thus, it is known that tryptophan and cysteine residues react in aqueous acid with sulfenyl halides (18).
The amino acid analysis of the peptidyl proteins synthesized shows that the number of amino acid residues in excess of those present in the original protein is in good agreement with that expected from the composition of the peptides attached (Table  III).
Moreover, it is possible to calculate from the amino acid analytical data the average number of peptide chains per protein molecule.
The results (Table III) are in reasonable agreement with those obtained by two other methods, namely, desamination of the peptidyl proteins with nitrous acid (10) and spectral measurements at 390 rnp of the Nps-peptidyl RSA derivatives. The agreement between the results obtained by three independent Precipitant (pg 1 methods shows that, in every case, only one peptide was attached to an e-amino group of lysine residue in the protein. The peptides used in inhibition studies (Table II) were prepared via benzyloxycarbonyl derivatives (Table I) as described under "Materials and Methods." Their concentrations were determined by nitrogen (Kjeldahl) analysis.

Immune Response to Peptidyl Proteins
In view of the possible differences in the combining sites of antibodies of the same specificity elicited in different rabbits, the present study was carried out in each case with sera from individual rabbits.
Sera collected before immunization did not precipitate with any of the immunogens or the cross-reacting antigens.
Groups of six rabbits were immunized with the following peptidyl derivatives of RNase: [nAla-Gly]ii-RNase, [(nAla)*-Gly]i,.,-RNase, [(DAla)a-Gly]io-RNase, and [(nAla)rGly]l.?-RNase. The presence of antibodies to the peptidyl moiety was followed by the precipitin reaction of the sera with the respective peptidyl derivatives of RSA. Typical precipitin reactions are shown in Fig. 1. In the same figure results are also given of cross-precipitin reactions with poly-n-alanyl-RSA (the poly-Dalanyl chains contained, on the average, 8.2 n-alanine residues). The absolute amount of antibodies precipitated from sera of individual rabbits by the respective peptidyl derivatives of RSA varied between 0.35 mg and 2.1 mg of antibody per ml of serum (most animals had antipeptidyl antibodies in the range of 0.8 to 1.2 mg per ml of serum), but the ratio of the amount of antibodies precipitated by the poly-n-alanyl-RSA and by the peptidyl-RSA was in each case very similar. Poly-n-alanyl-RSA cross-reacted only to a limited extent with anti-nAla-Gly antibodies, whereas it cross-reacted completely with anti-(nAla)rGly antibodies. As seen in Fig. 1, the extent of cross-reaction increased in parallel with the increment in the length of the alanyl peptide in the immunogen.
The above results indicate, as expected (6)) that the combining sites of the antibodies investigated are complementary to a peptide length greater than 3 amino acid residues. In these experiments the peptide reacting with the antibody was attached to a protein carrier, whose vicinity may have affected the results.
For a more precise analysis, experiments with peptides as inhibitors were performed.

Inhibition Studies
None of the peptides used in inhibition studies was degraded when incubated with rabbit sera under the conditions of inhibition experiments, as followed by high voltage electrophoresis (6). Peptides used in the inhibition studies did not cause any precipitation with either sera from before immunization or with immune sera.
The peptides used for the inhibition of the immunospecific precipitin reactions were of three categories, possessing the general formulae @Ala), (n = 2 to 5), @Ala),-Gly (n F 1 to 4), and (nAla),&ly-e-AC (n = 1 to 3). The alanine peptides are representative of the NH*-terminal part of the antigenic determinant . The peptides with COOH-terminal glycine are of the same structure as the peptides attached to the protein carrier in 1 the immunogen.
Peptides with COOH-terminal e-aminocaproic I I I too-I ' +-lacid should reflect any carrier specificity, as E-AC is chemically related to the side chain of lysine to which the peptides are O--/h . coupled in the immunogen.
Inhib&on of gala-Gly System-Results of the inhibition of the precipitin reaction between nAla-Gly-RSA and an anti-nAla-.
Gly-RNase serum are shown in Fig. 2. While the data given in 75-/I---. . Fig. 2 were obtained with the serum of an individual rabbit, closely similar findings were observed when antisera of three 2 .! / other animals were tested.
In all cases nAla-Gly-r-AC was the z 50-c most efficient inhibitor (50% inhibition at 0.10 to 0.12 mM 2 peptide).
The peptide attached to the protein, nAla-Gly, was 2 1 far less inhibitory (50% inhibition at 2.5 to 4.0 InM peptide). All of the other peptides were either similar in their inhibitory capacity to nAla-Gly, or less efficient.
In the (DAla), series, an increase in n (from 2 to 4) did not improve inhibitory efficiency, in agreement with the structure of the peptide attached, which had only 1 n-alanine residue at its NH2 terminus.
The efficiency of inhibition of the nAla-Gly peptide was considerably increased only when e-AC was attached Thus, the size of the combining site in this case corresponds to that of 4 amino acid residues. Elongation of the nAla-Gly at the NH2 terminus did not improve the inhibitory efficiency, and, in the case of the best inhibitor, nAla-Gly-e-AC, the attachment of additional n-alanine residues at the NH2 terminus decreased dramatically the capacity of the peptides to inhibit.
These data indicate that the interaction of the peptide inhibitor with the combining site "starts" from the NH&erminal portion of the peptide. Inhibition of (oLila)&+ System-Results of the inhibition of the precipitin reaction between (nAla)2-Gly-RSA and an anti-(nAla)-Gly-RNase serum are shown in Fig. 3. As seen in Fig. 3, (nAla)&ly-e-AC was the most efficient inhibitor. The values for 50% inhibition varied for antisera of four different rabbits between 0.22 and 0.34 mM peptide.
(nAla)&ly-c-AC was 4 to 13 times more inhibitory than (nAla)&ly, indicating that also in this system the protein moiety contributed to the specificity of antibodies formed.
In contrast to the nAla-Gly system, in the (nAla)z-Gly system the efficiency of the inhibition by means of @Ala), peptides increased with n. (nAla)4 was in one animal less inhibitory than (nAla)&ly (Fig. 3) ; however, in three other animals it was 2 to 3 times more inhibitory.
Nevertheless, in all cases the most efficient inhibitor was @Ala)-Gly-e-AC (see Table IV).
It seems that this peptide contains all of the components of the complete antigenic determinant in the correct order.
The attachment of an additional alanine residue at the NH2 terminus of (nAla)rGly-e-AC resulted in a peptide, @Ala) %-Gly-r-AC, which was less inhibitory than (nAla)&ly-e-AC with the sera of all four animals tested. This is in agreement    protein in the immunogen shows that (oAla)3 was less inhibitory than (DAla)rGly, and (DAla)z was a very poor inhibitor. Thus, it seems that the antibody active site is complementary in this case to the tetrapeptide @Ala) rGly. with the interpretation given above for the interaction of the @Ala),,-Gly-e-AC peptides with anti-DALGly-RNase sera.
Inhibition of (~Ah)~Gly System-Results of the inhibition of the precipitin reaction between (DAla)rGly-RSA and an anti-(DAla)3-Gly-RNase serum are shown in Fig. 4. In this system, for antisera of all four animals tested, (nAla)rGly-e-AC (50% inhibition at 0.10 to 0.15 mM peptide) was not significantly better as an inhibitor than (DAla)&ly (50% inhibition at 0.15 to 0.28 mM peptide), suggesting that in this case the protein does not contribute to the specificity of the antigenic determinant.
-4 comparison of the inhibitory capacity of peptides resembling the NHz-terminal portion of the haptenic peptide attached to the The data given in Fig. 4 permit an evaluation of the capacity of the antibody to distinguish between glycine and alanine residues. Although (nAla)z is a poor inhibitor, it is better than DAla-Gly. @Ala)3 is a better inhibitor than (DAla)zGly. On the other hand, when glycine is in the fourth and fifth positions from the NH2 terminus ((DAla) s-Gly and (DAla)rGly), its replacement by alanine ((oAla)d and (DAla)s) is not reflected in significant changes in inhibitory efficiency. Inhibition of (~AZu)rGly Syskm-Results of the inhibition of the precipitin reaction between (nAla)rGly-RSA and an anti-(nAla)b-Gly-RNase serum are shown in Fig. 5. In this system  the inhibitory capacity increases in the order @Ala)*, (nAla)3, (nAla)4, and the inhibition with (nAla)rGly is either the same as with @Ala) 4 or at the extreme it is a-fold better.
It seems thus that also in this case the antibody active site is complementary to a tetrapeptide. Similar to the (nAla)3-Gly system, the replacement of alanine by glycine in the second and third position from the NH2 terminus in the inhibitor decreased the inhibitory efficiency in the (nAla)rGly system, whereas replacement in the fourth and fifth position did not affect the inhibition significantly. Comparison of Homologous and Heterologous Systems-The amounts of antibodies precipitated with the original immunogen, as well as with RXase alone and with the respective peptidyl-RSA, are given in Table V. Since the antibodies precipitated with RNase alone are certainly not of antipeptidyl specificity, it can be seen from Table V that the peptidyl-RSA  derivatives  precipitated at least 70 to 90% of antipeptidyl antibodies. Indeed, in two cases in which sera previously absorbed with RNase were reacted with the appropriate peptidyl derivatives of RNase and RSA, similar amounts of precipitates were formed.
Studies on inhibition with peptides were carried out on the antisera previously absorbed with RNase, but once using as precipitant the homologous antigen and once the appropriate RSA derivative.
As seen in Table VI, the inhibitory capacity of the peptides is similar.
Thus, in the (nAla)3-Gly system, the peptide concentration causin, v 50% inhibition when the RNase derivative was used as precipitant was 50 to 65% of that found when the RSA derivative was used for precipitation. However, the relative potencies of the different peptides within each experiment were quite similar, and thus led to the same conclusions concerning the size and nature of the combining sites.
The participation of the protein lysine residues in the antigenic determinant is reflected to the same extent, independently of whether the homologous or the cross-reacting antigens are used (Figs. 6 and 7). DISCUSSIOX Both the immunogens and the cross-reacting precipitants used in this study were prepared by covalent attachment of peptides to proteins.
The attachment was carried out via N-blocked peptide succinimide ester. The first account of chemical attachment of a peptide to a protein was given in an immunochemical study by Landsteiner and Van der Scheer (2&22), who have coupled the peptides via azo links to tyrosine residues in the protein.
Recently, several reports have appeared in which the attachment of peptides to proteins was performed making use of water-soluble carbodiimide derivatives (23)(24)(25). The main purpose of the present investigation was to elucidate to what extent inhibition studies may reflect the structure of the hapten to which the animal was exposed and the size and nature of the combining sites of the antibodies formed.
For a more precise evaluation of the results we attached to RNase as haptens a homologous series of peptides of defined amino acid sequences. The reactions of antisera against the various immunogens with cross-reacting precipitants, in which the same haptens were attached to RSA, were inhibited by different related peptides, and the main results are summarized, for the sake of comparison, in Table IV. The difference in inhibitory capacity between two peptides which was less than a factor of 2 was not considered significant, because of thermodynamic considerations (6). It must be remembered that the exploration of the combining site is based on the strength of the interaction between a series of haptens and the antibody combining site; the precipitating antigen serves only as a tool for estimating its degree of saturation with the hapten.
In fact, any molecule which binds to the antibody combining site would serve this purpose, provided that its degree of binding can be measured.
The properties of the combining site are deduced from the correlation between the extent of antibody-hapten interaction and structural features of the hapten.
At least 70 to 90% of the antipeptidyl antibodies formed by injecting peptidyl-RNase derivatives were precipitated by the corresponding peptidyl-RSA derivatives (Table V). Moreover, the relative efficiency of inhibition with different peptides was similar in each case tested, whether the homologous or the crossprecipitating system was used (Figs. 6 and 7; Table VI). The efficiency of inhibition increases in each system with the increment in the length of the peptide resembling the hapten in the immunogen, starting from the NH2 terminus.
As seen in Table IV, a bigger peptide of related composition is not necessarily a better inhibitor.
When peptides of the same size are compared, the peptide possessing the amino acid sequence identical with that starting from the NH2 terminus of the hapten attached to the immunogen is in each case the most efficient inhibitor. Thus, e.g. in the DAla-Gly system, the peptide DAla-Gly-e-AC (E-AC corresponding in size to a dipeptide) is a better inhibitor than (DAla)rGly, whereas in the (DAla)3-Gly system, the peptide (DAla)3-Gly inhibits better than DAla-Gly-e-AC.
In all of the four immune systems investigated, the size of the antigenic determinant to which the antibody site is complementary seems to consist of 4 amino acid residues. Thus in the (DAla) rGly system, (DAla) rGly is not significantly better as an inhibitor than (DAla)4, whereas (DAla)4 is much more inhibitory than (DAla)3 ( Fig. 5; Table IV).
In the last case e-aminocaproic acid serves as a model for the participation of the protein carrier in antigenic specificity, as it resembles the lysine residue of protein to which the peptides were attached. 3 Indeed, in the (DAla)-Gly system, the peptide (DAla)rGly-e-AC inhibits better than (DAla)-Gly by one order of magnitude. The contribution of the protein carrier is even more remarkable in the DAla-Gly system (compare the efficiency of inhibition with DAla-Gly-e-AC and with DAla-Gly ; Fig. 2 and Table IV).
We may conclude that the protein carrier contributes to the antigenic determinant only when the peptide attached contains less than 4 amino acid residues. In other words, antibodies would show carrier specificity only when the hapten attached is less than a tetrapeptide.
Many haptens commonly used in immunochemical studies, such as the 2,4-dinitrophenyl group and the p-azobenzenearsonate group, are smaller in size than a tetrapeptide, and antibodies produced against them would be expected to reflect partially also the protein area to which the hapten was attached.
Indeed, Eisen and Siskind (19) showed that in the case of dinitrophenylated proteins lysine contributed to the antigenic dinitrophenyl determinant. More recently Parker,Gott,and Johnson (24) have reported, using as immuno-3 Peptides with E-aminocaproic acid at the COOH terminus have been used for a similar purpose in the study of a polyalanyl immune system (4). gen a protein to which a dinitrophenyl-tetrapeptide was attached, that antibody specificity can extend beyond the dinitrophenyl group.
In both cases the investigators did not establish limits as to the size of the dinitrophenyl-peptide antigenic site. It is pertinent to remark here that, in the poly-D-alanyl system, the combining site of antibodies of the immunoglobulin M class is also complementary to a tetrapeptide, similarly to that of antibodies of the immunoglobulin G class (26). As seen in Table IV, even though the determinant in the (DAla)z-Gly system, is (DAla)2Gly-e-AC, peptides of the structure (DAla)J and (DAla)S are better inhibitors than (DAla)zGly. If the antigenic determinant was not known, such results would lead to false conclusions about its nature.
It seems that, in this particular case, the presence of additional alanine residues enhances in a nonspecific way the binding to the antibody, similarly to the increased efficiency of binding by anti-(tobacco mosaic virus protein) of a decapeptide in which (Ala)s was attached to the specific pentapeptide, as compared to the pentapeptide alone (27). Findings of a similar nature were reported by Arakatsu,Ashwell,and Kabat (28), who have shown that, with some rabbit antisera to an isomaltotrionic acid derivative of bovine serum albumin, the efficiency of inhibition increased up to isomaltohexaose.
Nevertheless, the authors drew the conclusion that the general approach of elucidating sizes of the combining sites of antibodies is valid, although possible difficulties in the interpretation of the data may be encountered. The contribution of nonspecific hydrophobic interactions to the binding energy of haptens to antibodies has also been reported by Metzger,Wofsy,and Singer (29) and by Benjamini,et al. (30).
Conclusions concerning the size and nature of the combining sites of antibodies were first drawn from inhibition studies by Kabat (2,31,32). The present study confirms that this technique is useful for estimation of the dimensions and complementarity of the active sites of antibodies, provided that the structure of the antigenic determinants is known.
In connection with the present work it should be mentioned that in a recent study of the immunological specificity of the bradykinin system, which made use of an immunogen prepared by binding covalently bradykinin to polylgsine, replacement of glycine with alanine had a profound effect on binding capacity, and this was interpreted as an obligatory effect on conformation (33). Similarly, in a study of a major antigenic determinant of tobacco mosaic virus protein, Young,Benjamini,and Leung (34) found that replacement of alanine with glycine decreased dramatically the capacity of the antibody to bind the peptide determinant.
In the last case it is not known which amino acid residues composing the antigenic determinant are on its periphery.
In contrast to enzymes, which represent a homogeneous population of molecules possessing one type of active site, antibody combining sites to a unique determinant are known to be heterogeneous.
Nevertheless, by comparing pairs of peptides in which only 1 residue was replaced by another, or the effect of elongation of a peptide on its binding efficiency, it seems possible to evaluate the contribution of the various parts of the inhibitory molecule.
For example, in the present study the NHz-terminal D-alanine residue plays always an immunodominant role. The necessity for a good interaction to occur at the NH2 terminus is also evident from the previous findings that, in the poly-r,-alanine system, the tetra-L-alanine peptide, which is an excellent inhibitor (6), loses completely its inhibitory capacity when the NHz-terminal a-amino group is acetylated (35).
The use in the present study of a series of determinants, whose size and structure were changed systematically, indicates that detailed features of the antigenic determinants are reflected in the combining sites of the antibodies formed.