HAPTENINHIEITION OF PRECIPITATION OF ANTIBODIES AND POLYHAPTENS
Dec., 1942
[CONTRIBUTION FROM
THE
3013
GATESAND CRELLIN LABORATORIES OF CHEMISTRY, CALIFORNIA INSTITUTE OF TECHNOLOGY, No. 9071
The Serological Properties of Simple Substances. IV. Hapten Inhibition of Precipitation of Antibodies and Polyhaptenic Simple Substances BY DAVIDPRESSMAN, DAVIDH. BROWN, AND LINUSPAULING
As part of a series of studies of the antigenantibody reaction by use of polyhaptenic simple substances as antigen~,’J~~ we have made a quantitative investigation of the ability of various substituted phenylarsonic acids to inhibit the precipitation of compounds containing two or more phenylarsonic acid groups with antiserum to sheep serum coupled with diazotized parsanilic acid. The investigation included study of the effect of various concentrations of harken XXIII (HOC)NN-SO~HI) on the amount of precipitate obtained with antiserum and various concentrations of antigen VI / R O H R, where R is the p-azophenylarsonic acid
W \H
group); the effect of changed conditions of precipitation for the same system; the effects of four haptens on the precipitation of each of five antigens and each of two pools of antiserum; and the effects of 24 different haptens on one precipitin reaction. The results are discussed in this paper. Experimental Methods Simple Antigens and Haptens.-The substances used are the polyhaptenic compounds OH
OH
.IR I
I11 OH
”(>::
HO
R”(CH2)zR” XV
R’
=
”
0
“ with R = NN-sOaH2,
N N ~ N N C > A S O ~ H ~ ,
__
XXIII
HOCI)..
XXVIII
or’,
“2
XXIX 0 S 0 3 H 2 , and twenty others given in Table VI. The preparation of most of these substances has already been described‘ or will be discussed el~ewhere.~
p-Benzoylaminophenylarsonic acid was prepared from p-arsanilic acid and benzoyl chloride by the SchottenBaumann reaction. The product was precipitated with hydrochloric acid and purified by extraction with boiling ethanol. p-Acetaminophenylarsonic acid was prepared from p arsanilic acid and acetic anhydride in basic solution. The compound was purified by precipitation with hydrochloric acid and recrystallization from boiling water. Antisera to sheep serum coupled with diazotized arsanilic acid were prepared as previously described, with use of the same rabbits. Method of Analysis.-Each precipitate was centrifuged and washed thoroughly with three 10-ml. portions of 0.9% saline solution a t room temperature. The amount of protein in the precipitate was determined with the FolinCiocalteu reagenta by a modification to be discussed elsewhere. B&er.-The borate buffer of pH 8.0 was prepared by adding 0.16 N sodium hydroxide solution to 0.2 M boric acid in 0.9% sodium chloride solution. The antigen and hapten solutions were all diluted with this buffer.
Time at room temp.
R‘ XI
and XX R
0 S 0 3 H 2 ,
TABLE I EFFECTOF STANDING AT ROOM TEMPERATURE AND IN THE REFRIGERATOR ON AMOUNT OF PRECIPITATE OBTAINED IN THE PRESENCE OF HAPTEN Analyses for 5-ml. aliquots of a mixture of equal volumes of serum S, antigen VI (25 pg./ml.), and hapten X X I I I (60/pg.(ml.). pH of all supernates 8.2.
V O H R VI
R
R” = -CONH C S ) A s O d L and the haptens X X I
and
(1) L. Pauling, D.Pressman, D. H. Campbell, C. Ikeda, and M. Ikawa, THIS JOURNAL, 64,2994 (1942). (2) L. Pauling, D.Pressman, D. H. Campbell, and C. Ikeda, ibid., 64,3003 (1942). (3) L.Pauling, D.Pressman, and C. Ikeda, ibid.. 64, 3010 (1942). (4) The first experiments on the precipitation of polyhaptenic simple substances by antisera and its inhibition by haptens were carried out by K . Landsteiner and J. van der Scheer, J. Ex9. M c d . , S6, 399 (1932).
Nights in refrigerator
0
1
0
2 1 2
hour hour 2 hours 2 hours Overnight Overnight a/4
1
2 0 1
Amount of precipitated antibody, pg.
125 131 138 131 131 144 125 125 131 131 144 144 106 106 144 106 119 125 125 125 138 63 63 63 94 100 100 106
Hapten replaced by buffer
2 hours
2
810
( 5 ) D. Pressman and D. H. Brown, to be published. (6) 0. Folin and V. Ciocalteu, J. Bid. Chcm., ‘73,627 (1927).
1 ) ~ v r nPRESSMAN, DAVID 13. BROWN A N D LINUSI'AULING
301f3
The Effect of C u e d Conditions on Amount of Precipitate.-In order to observe the effect of standing on the amount of precipitate obtained in the presence of hapten, equal volumes of serum S , antigen VI (25 Fg./ml.), and hapten XXIII (60 pg./ml.) were mixed and 5-ml. aliquots of the mixture were permitted to stand zero, three-quarters, or two hours or overnight a t room temperature and were then placed in the refrigerator at 3 O over zero, one, or two nights. The amount of precipitate obtained in each test was about 15% of that obtained in the absence of hapten The results are given in Table I. Only when the tubes stood overnight a t room temperature without subsequent refrigeration was a significant decrease in the amount of precipitate observed. This effect was also observed with specific precipitates in the absence of hapten.2 A slight decrease was observed when the tubes stood overnight a t room temperature and then overnight in the refrigerator.
TABLE 11 ANTISERUM s AND ANTIGEN VI BY HAPTEN XXIII Antigen solution, hapten solution, and antiserum, 1 ml. each; 2 hours a t room temperature, overnight in refrigerator. Blanks of antiserum and buffer: 0, 0, 0 pg. fiH of all supernates 8.2. INHIBITION OF PRECIPITATION O F
Amount of hapten, pg.
0
Amount of antigen, pg. 6.3 12.5 25 50 100 Amount of antibody precipitated, pg.a
122 (131) 125
4.1 8.3 16.5 31 63 125 250 500
335 262 212 163
88 41
84
660 356 285 178 116 31 19 3 6
305 181 147 109 72 35 12 0 0
(144) 131
Vol. 64
The Effect of Various Amounts of Hapten EUII on Amount of Precipitate Obtained with Various Amounts of Antigen VI.-The results of the study of inhibition by hapten XXIII of the precipitation of antigen VI are given in Table 11. As previously observed with phenylarsonic acid as the hapten,l the optimum zone does not shift significantly with increasing amounts of hapten. Experiments on the Relative Inhibiting Powers of Different H a p t e n s . S i c e the effect of hapten inhibition is most easily given theoretical interpretation in the antigenantibody equivalence zone, experiments were carried out with antiserum pools S and T to determine the optimum zones (which may be taken as the equivalence zones) for antigens 111, VI, XI, XV, and XX. The results are given in Table 111. It is seen that antiserum S contained much more antibody than T. As previously observed,1 antigen XI, with long haptenic groups, gave much larger amounts of precipitate than the others. It is interesting, as an example of the variability of antisera, that antigen XX gave more precipitate than antigens 111, VI, and XV with antiserum T but less with antiserum s. Hapten inhibition experiments were then carried out with haptens XXI, XXIII, XXVIII, and XXIX and these five antigens in the optimum zones for the individual antigens with each of the antisera S and T. The results are given in Tables I V and V. Experiments were also carried out with antiserum S and antigen VI a t the optimum concentration in the presence of each of twenty-four haptens a t various concentrations, with the results shown in Table VI.
106
Discussion
103
(31) 9 41 6 6 12 0 li 3 3 0 0 ' Values are averages of duplicate analyses, with mean deviation of 1 5 % from the averages, except for the values in parentheses, which represent single analyses.
The data on hapten inhibition given in the foregoing tables can be most effectively discussed by comparison with the simple theory developed in a preceding paper of this series.2 I t was shown that for the system postulated, with both antigen and antibody assumed to be bivalent, a plot of the amount of antibody precipitated, for the case
TABLE I11 PRECIPITATION OF ANTISERA s AND T AND FIVEANTIGENS Antigen solution 0.5 ml.; antiserum, 0.5 ml.; 1 hour a t room temperature and overnight in refrigerator. ___..
Antiserum S Antigen I11 VI XI XV
xx
.
4.7
(1.2
8.C
194 101 147 (138) (69)
228 157 226 160 100
303 278 291 (187) 97
Amount of antigen used, pg. 10.4 13.5 17.5 22.8 Amount of a n t i b o d y precipitated, *g n
(275) 426 438 178 107
306 372 603 125 54
226 263 859 115 47
191 (156) 1070 (75) 28
?!LR
38.5
169 85 1410 47 28
135 69 1340 32 22
1300
469
359
50
31
50
(106) 32 28 22
Blanks of antiserum and buffer: 0, 0, 0. PH of supernates 8.2 Antiserum T Antigen 111 VI XI
16 0 42 10 253 319 438 500 460 472 (187) 19 22 13 13 xv 63 56 81 41 44 28 xx 63 Blanks of antiserum and buffer: 0, 0, 0. PH of supernates 8 . 3 Averages of duplicate analyses, with mean deviation *7%; single analyses in parentheses. 10 47 153 0 88
Dec., 1942
HAPTEN INHIBITION OF PRECIPITATION O F ANTIBODIES AND
3017
POLYHAFTENS
TABLEIV ANTISERUM s AND ANTIGEN Antigen solution, hapten solution, and antiserum, 1ml. each; overnight a t room temperature and overnight in refrigerator. Blanks of antiserum and buffer: 0, 0, 0 pg. PH of all supernates 8.2. HAFTEN
INEIBITION O F ~ E C I P I T A T I O NO F
0
Antigen, pg.
Amount of hapten,
1.95
3.9
Hapten
7.8
15.6
pg.
31.3
Amount of antibody precipitated, p g . 5
62.5
522 XXI 540 488 413 394 365 (456) 488 468 XXIII 285 257 181 (438) 541 XXVIII 410 276 563 479 (406) 538 528 XXIX 432 525 507 463 500 XXI 556 432 369 266 191 147 VI 363 254 127 91 47 XXIII 485 25 397 122 XXVIII 457 319 213 60 519 XXIX 528 362 294 488 441 XXI 2310 2290 2160 (2070) 1935 1640 1400 XI 2255 XXIII 2130 1955 1170 1585 794 59 2285 XXVIII 2190 1860 1320 1800 970 2410 2330 2280 2115 XXIX 2210 (2060) 416 225 XXI 386 347 xv 307 332 228 366 131 XXIII 347 79 25 37 XXVIII 310 360 178 88 225 363 XXIX 388 402 347 (400) (425) (419) XXI 167 194 125 XX 178 178 169 147 135 XXIII 125 88 37 16.7 50 6 141 125 106 XXVIII 63 37 6 181 181 XXIX 147 190 141 125 a Averages of duplicate analyses, with mean deviation' f3%; single analyses in parentheses.
125
250
113 219 403
241 50 166 378
101 6 28 222
97 0 6 197
1360 416 663 1740
1090 194 435 1490
188 16 41 366
119 0 13 385
94
60 0 0 94
I11
25
TABLE V HAPTENINHIBITION O F PRECIPITATION OF 'ANTISERUM T AND ANTIGEN Antigen solution, hapten solution, and antiserum, 3 ml. each for antigen VI, 2 ml. for XI and XX; 2 hours at room temperature and overnight in refrigerator. Blanks of antiserum and buffer: 0, 0, 0 pg. PH of all supernates 7.8-7.9. Antigen, pg./ml. Hapten
VI 10
XXI XXIII
61
XXIX XXI XXIII
353 353 303 356
1320
XXIX 6.3
XXI
XXIII XXVIII
XXIX
700 1
1
1
I
I
I
I
1
Amount of antibody precipitated, pg.5
373
XXVIII
XX
tion we attribute to the heterogeneity of the antiserum, which may contain antibodies with greatly varying combining powers.
Concentration of hapten solution, pg./ml. 0 1.95 3.9 7.8 15.6 31.2
XXVIII
XI
0 116
491
(325) 347 178 344
254 294 57 291
197 185 22 244
(119) 63 3 166
1220 1245 1210 1155
1110 1150 844 1060
832 950 675 885
804 715 478 819
681 532 313 675
516 481 485 $79
472 495 466 485
(488) 460 422 479
438 378 354
485 400 222 446
0
* ArO>H,
I
Averages of duplicate analyses, with mean deviation "3%; single analyses in parentheses. a
of equivalent amounts of antigen and antibody, against the Of Present be a straight line. Some examples of plots of this sort are given in Fig. 1, taken from the data of Table VI, which correspond to the optimum zones. It is seen that the curves approach the theoretical linear form only at low hapten concentrations; the deviation from linearity a t higher concentra-
Equation of the earlier paper,2 giving the value of the initial slope of the hapten-inhibition cuTve,may be in the - - = _d H
d AB(pp)
with C=
s{L/a
C
K'
'
(Ks + PsZ)'/: + K S + K2sa+ (Ks+ K*s')'/:~
(1)
DAVIDPRESSMAN, DAVIDH. BROWNAND LINUSPAULING
3018
and
Here H is the amount of hapten added, AB(pp) is the amount of antibody precipitated, s is the solubility of the antigen-antibody complex, and K and K' are the bond-strength constants for the antigen-antibody bond and the hapten-antibody bond, respectively. Both C and C' for any antiserum depend only on the antigen and are independent of the hapten. The constant K' is characteristic of the hapten, indicating the strength of the bond between the hapten and the antibody. Thus for any given antiserum and antigen the reciprocal of the rate of decrease of amount of precipitate with increasing amount of hapten is a linear function of the reciprocal of the hapten constant K'. expect From the form of Equation 1 we the order of inhibitory activity of various haptens for a given antiserum to be the same for various antigens. This is the case for our experiments.
Vol. 64
With serum S the order of inhibition was XXIII > XXVIII > X X I > X X I X for each of the five antigens used (Table IV). The order observed for serum T was different; XXVIII > XXIII > X X I > XXIX; but this order was again observed for each of the five antigens (data for antigens VI, XI, and XX are given in Table V; less reliable results obtained for I11 and XV, not included in the table, also placed the haptens in the same order), I t is surprising that the amide compound m R U was more effective in inhibition with u serum T than was the azo compound H
O
O
,
since the antibodies were produced by inoculation be with an azoprotein and expected combine preferentially with azo groups. Antiserum T was consistent in its greater reactivity with amide groups than with azo groups, insofar as it also gave a larger amount of precipitate with the amide antigen than with the somewhat similar azo a x g e n s
R"C>R"
TABLEVI INHIBITION OF PRECIPITATION OF ANTISERUM S AND ANTIGENVI BY HAPTENS Antigen solution, 1 ml. (25 pg.); hapten solution, 1 ml.; antiserum, 1 ml.; 2 hours a t room temperature and 2 nights
in refrigerator.
*
= AsOaH,,
K'
0.13 .13
R = NNC-)AsOzHz,
u
K
A
x~ ion
2.2 1.4 0.7 2.2 3.3 2.9 2.4 3.6 1.2 1.5 3.9 3.4 16 2.1 5.6
R" = -CONHC--)As03Ha. w
1.95
3.0
662 581 590 563 566 516 566 560 538 479 47 3 473 463 457 450 494 (475) 541 559 513 494 550 45 1
530 575 (5W 532 501 463 538 447 (494) 419 422 388 372 344 347
Amount of hapten, pg. 7.8 15.6 31.3 62.5 Amount of precipitated antibody, pg.b
500
125
250
375 269 200 (488) (463) 535 272 354 475 385 206 238 175 I17 432 385 516 (275) 475 272 210 182 379 347 .23 169 131 432 325 235 91 .26 191 447 213 154 357 266 .28 181 131 387 119 .29 306 (206) 122 429 188 147 338 241 .29 116 128 172 376 276 216 .44 128 144 354 285 216 166 ,53 103 372 85 213 173 134 .60 116 344 79 250 182 .66 (144) 294 106 119 210 166 137 .75 238 41 91 163 144 63 .80 72 275 37 188 125 100 .80 78 106 286 6.5 400 216 103 66 .80 131 94 297 5.7 404 194 169 50 .80 50 332 101 225 175 6 454 .80 29 1 78 34 10 457 194 135 .89 41 419 335 88 13 216 153 .89 6 184 119 60 0 391 310 .98 0 0 38 160 86 269 1.02 (438) 19 0 50 203 141 85 341 1.02 0 285 16 407 91 44 0 16 1.40 163 Average value for no hapten 638 pg. Control, antiserum and buffer 0, 0, 0, 0 pg. pH of all supernates 8.1-8.2. single analyses I Second acid dissociation ~ o n s t a n t . ~ Averages of duplicate analyses, with mean deviation *3%; iii parentheses.
'
Dec.. 1942
HAPTEN INHIBITIONOF PRECIPITATION OF ANTIBODIES AND POLYHAPTENS 3019
OH
OH
of the latter is probably greater than that of the former because of the stronger forces resulting from the doubled electric ~ h a r g e ,as ~ has been R R pointed out by Haurowitzs Accordingly i t would the reverse of the behavior observed for serum be expected that increase in the second dissociation constant of the acid, leading to increase in S and the antisera tested previously.' Another example of the variability in properties of the number of doubly charged ions present, would lead to increase in bond-strength constants with the antisera is that the amide antigen R ' O " the antibody. There are given in Table VI gave more precipitate than the other amide anti- values of the second ionization constant for gen R"(CH2)2R" with serum T but less with eighteen of the haptenss It is seen that the serum S. correlation with K' is not very striking. In order to obtain from the data of Table VI The ortho-substituted phenylarsonic acids are numbers representing the inhibiting power of the seen to be, with one exception (the o-nitro acid), the twenty-four haptens, the following treatment least effective of all in combining with antibody. was used. Plots were made of the amount of An a-naphthyl residue is about equivalent to an precipitated antibody for each hapten as a func- ortho-substituted benzene ring. Meta-substition of the number of moles of hapten added, as tuted compounds are somewhat more effective illustrated in Fig. 1. Smoothed curves were in general than phenylarsonic acid, and paradrawn through the points, and the initial slopes substituted compounds are still more effective. were read from the curves. Some consideration 0-Naphthylarsonic acid is intermediate between was given to the course of the curve throughout the meta- and para-substituted phenylarsonic the range of concentration covered by the experi- acids. ment in determining the value of the initial slope, All of the para-substituted phenylarsonic acids with use of a family of curves deduced from the are more effective in inhibition than phenylarsonic whole set of data. Plots against the log of acid itself. This increased effectiveness on replacehapten concentration were also used. The re- ment of the para hydrogen atom by a larger atom ciprocals of the negative slopes -d AB(pp)/d H or group is presumably the result of increased are given in Table VI, under the heading K'. van der Waals interaction between the substituent These reciprocals would be proportional to K' and the antibody. The order of effectiveness of if the value of C' in Equation 1 were negligible the para substituents in increasing the value of the compared to the other term. In any case these constant K' is numbers would be expected to represent qualita> > tively the combining power of the haptens with NO* > CHXONH > antibody. A constant factor was introduced such C1, Br, I, CHs > OH > NHz > COOH > H that the average value of K' for the two azo The same order holds also in the ortho and meta haptens became unity. positions for the substituents NOz, CHa, and NHz, Many interesting correlations of the values of data not having been obtained for others. K' with the molecular structure of the haptens Whereas replacement of a hydrogen atom in the can be observed. The effects of a substituent on para position by a substituent group always leads the bond-strength constants of the substituted to an increase in the value of K', this is not so phenylarsonic acid molecules are seen to be de- for the ortho position. Both o-methylphenylarpendent both on the position of the substituent sonic acid and o-aminophenylarsonicacid, as well in the benzene ring and on the nature of the sub- as the a-naphthyl compounds, are less effective stituent. I t would also be expected that, aside in inhibition than phenylarsonic acid itself. from direct structural effects, the constituents This dserence in behavior from the para-subwould be of influence through their effect on the stituted compounds is attributed to steric effects. second dissociation constant of the acids. The The immunizing antigen, a protein with attached antibodies, produced presumably in about neutral p-azophenylarsonic acid groups, gives rise to solution in the animal, probably have combining (7) This may be responsible for the fact that the maximum amount of precipitate with variation of pH is obtained at fiH 8, instead of at groups for both singly ionized and doubly ionized about 7, which presumably prevails at the site of antibody formation. arsonic acid molecules. The combining power (8) F. Haurowitz, Z. ghysiol. Chcm., 445, 24 (1937). and R&
V
v..
(Table 111). This is
(~T>NN OCON
3020
DAVID
PRESSMAN,
DAVID H. BROWN AND LINUSPAUI.INC
antibodies which are molded to the p-azophenylarsonic acid group, and which in general do not allow sufficient space for a larger group than hydrogen in the ortho position. On the other hand, a para substituent may fit into the space provided for the azo nitrogen atoms, and so by increased van der Waals attraction for the antibody makes the hapten more effective than phenylarsonic acid itself. The most striking and unexpected observation is that the nitro group in any position in the benzene ring causes a large increase in the bondstrength constant over the value for any other substituent in the same position. For the ortho compound this increase is so great as to overcome the steric effect and make the hapten more effective than phenylarsonic acid. The reason for this large effect of the nitro group is not obvious. The second acid dissociation cannot be advanced as the cause, since the value of KAz for o-nitrophenylarsonic acid is less than that of phenylarsonic acid, whereas its value of K’ is greater. We suggest that this action of the nitro group is the result of the well-known effect of the group in aromatic compounds in causing increased van der Waals attraction for other molecules, which shows itself in the great ability of aromatic nitro compounds to form molecular compounds with other substances. It is surprising that haptens containing the amide group 4 O N H - are nearly as effective as the corresponding haptens containing the azo group -EN---. As mentioned above, for antiserum T the amide group was found lo he even more effective than the azo group. I t will he interesting to see, by experiments with other pools of antisera, to what extent the order of combining power with antibodies of the haptens listed in Table ITI varies with the antisertiin used. In the larger haptens the effect of substituents far removed from the phenylarsonic acid group is
Vol. 64
small. Thus substitution of an amino or nitro group in the para position in phenylcarbamidophenylarsonic acid leads to an increase of only 10% in the value of K’, with the nitro group showing no greater effect than the amino group. Similarly p-hydroxyphenylazophenylarsonic acid and p-aminophenylazophenylarsonic acid show nearly the same value of K’. There is now under way an investigation of the inhibition by various haptens of precipitation with simple antigens of antisera made with use of a protein with attached azophenylazophenylarsonic acid groups as the immunizing antigen. This investigation was carried out with the aid of a grant from the Rockefeller Foundation. hlrs. Elizabeth Swingle, Mr. Frank Lanni, Mr. Stanley Swingle, and Mr. Shelton Steinle assisted with the analyses.
Summary A study has been made of the inhibiting action of haptens (substituted. phenylarsonic acids) on the precipitation of polyhaptenic simple substances containing two or more phenylarsonic acid groups by antiserum obtained by inoculating rabbits with sheep serum coupled with diazotized $-arsanilic acid. It was found that the order of inhibitory activity of four haptens was the same for each of five test antigens with a given antiserum. The order depended on the antiserum used. The antiserum which showed stronger inhibition by amide haptens than by azo haptens also gave larger amounts of precipitate with amide antigens than with azo antigens. The effect of each of twenty-four haptens on one antigen-antibody reaction was studied, and the data were interpreted to give relative values of the bond-strength constant of the haptens with the antibody. These values are shown to be correlated with the structure of the haptens. PASAPI~ N A , CAIJFORNIA
RECEIVED AUGUST10, 1942
