Biochirnica et Biophysica Acta, 420 (1976) 246-257

© Elsevier Scientific Publishing Company, Amsterdam- Printed in The Netherlands BBA 37241 F A C T O R S A F F E C T I N G T H E P R O P E R T I E S OF I N S O L U B I L I Z E D A N T I BODIES

S. COMOGLIO, A. MASSAGLIA, E. ROLLERI and U. ROSA Nuclear Research Center, SORIN- Dept. of Radioehemistry, Saluggia and Clinical Physiology Laboratory, CNR, Pisa (Italy)

(Received June 2nd, 1975)

SUMMARY I g G separated from an antiserum to estradiol was coupled under various experimental conditions to Sepharose activated either with CNBr or by conversion into a long-armed derivative (the N-hydroxysuccinimide ester). The conjugates were characterized by measurement of the binding parameters, in order to evaluate separately the loss of sites and the loss of affinity. The cross-reactivity with estriol and estrone was measured to obtain information on the occurrence of structural alterations of the antibody site. The results show that the loss of immunoreactivity varies in extent (from 95 to less than 10~o) and in nature (loss of sites or of affinity or a combination of both effects) depending on the coupling conditions. The use of a hydrocarbon extension to keep the protein distant from the matrix does not prevent the loss of active sites but is effective in safeguarding the affinity of the residual sites. The loss of sites can be substantially reduced by coupling at a p H value around neutrality and by keeping the protein/matrix mass ratio low. At a coupling p H of 6.4 and at a mass ratio of 0.1-0.2 nmol IgG/mg of Sepharose, the antibodies were insolubilized with a negligible loss of sites and affinity; on increasing the mass ratio (up to 10 nmol IgG/mg Sepharose) there is a progressive loss of sites accompanied by a substantial lowering of the affinity of the residual sites. On the basis of the above-mentioned findings, the nature of the effects occurring when antibodies are transferred from solution onto a solid matrix is discussed.

INTRODUCTION Immobilization of antibodies by covalent coupling to a solid matrix is accompanied by substantial losses of immunoreactivity [1, 2, 3]. Little is known, however, about the role played by the coupling conditions in determining the behaviour of the insolubilized protein. Moreover, most of the available data are expressed in terms of overall loss of immunoreactivity (loss of titer), as is usual in radioimmunoassay, while more valid information could be derived by measuring variation of binding parameters, i.e. by separately evaluating the loss of binding sites and the loss of affinity. In the present work we have investigated the role of various factors, such as the

247 nature and conditions of the coupling reaction, the interposition of a hydrocarbon extension (arm) between the protein and the binding groups of the matrix and the extent of substitution (the protein-matrix mass ratio), on the loss of immunoreactivity associated to the insolubilization. This was done using IgG separated from an antiserum to estradiol which was conjugated under various experimental conditions to an agarose matrix. The conjugates were characterized by measurement of binding parameters and the results were correlated to those obtained with the same antibodies in solution. Antibodies to estradiol were chosen for these studies for the following reasons: the small size of the hapten, whose diffusibility through the agarose matrix should facilitate the interpretation of the results; the univalence of the antibody site and the absence of aspecific binding of estradiol on the Sepharose matrix, which simplifies the measurement of the binding parameters; the availability of a highly specific antiserum [4], which implies that the cross-reactivity of the insolubilized antibodies with estriol and estrone can yield information on the occurrence of structural alterations at the level of the antibody site. As for the agarose matrix, this was activated either by CNBr or by conversion into the N-hydroxysuccinimide ester according to the procedure described by Cuatrecasas and Parikh [5]. This procedure should offer some specific advantages in studies such as those reported in the present paper. First of all, model experiments with amino acids [5] have shown that it should be possible to favour a more selective coupling of a protein to the active groups of the agarose ester by varying the pH and the composition of the reaction medium. Furthermore, the succinyl moiety has been successfully used as a chemical arm in affinity chromatography to minimize the specific contribution of the matrix microenvironment [6] and the coupling by the active ester occurs rapidly under mild conditions (aqueous medium at 4 °C). Since the present study is based on the measurement of the binding parameters, the question arises whether quantitative data can be reliably obtained when one of the binding partners is immobilized on an insoluble matrix. Gawronski and Wold [7] have shown that this approach was perfectly valid in the case of ribonuclease S-protein-S-peptide interaction, provided that free (unbound) component is totally removed from the insoluble phase at equilibrium. In our case the problem is considerably simplified by the absence of aspecific interaction of estrogens with the agarose matrix, which permits elution of the free labelled steroid from the agarose beads at equilibrium, thereby avoiding misclassification of the free and bound fraction. EXPERIMENTAL Chemicals 17fl-2,4,6,7-[3H]estradiol (specific activity 85 Ci/mM), 2,4,6,7-[3H]estriol and 6-7-[3H]estrone, were supplied by CEA-IRE-SORIN; estradiol, estriol and estrone were from Steraloids, Inc. Pawling, N.Y. ; Sepharose 4B CNBr-activated was obtained from Pharmacia (Uppsala, Sweden); 3-3'-diaminodipropylamine from Eastman Kodak (Rochester, N.Y.); N-hydroxysuccinimide from Fluka A G (Switzerland) and N-N'-dicyclo hexylcarbodiimide from K & K (Plainview N.Y.). All reagents were of analytical grade.

248

Antiserum The antiserum to 17fl-estradiol-6-(O-carboxymethyl)oxime-bovine serum albumin was raised in rabbit following the immunization scheme proposed by Lindner [8].

Separation of the IgG from the antiserum IgG was prepared from rabbit antiserum by an (NH4)2SO4 precipitation followed by purification with DEAE-cellulose according to the method described by Stanworth [9]. Immunoelectrophoretic analysis with guinea pig anti-rabbit serum and anti-rabbit y-globulins revealed the final solution to be composed entirely of y-globulins. Analytical ultracentrifugation in a Spinco model E machine failed to reveal any 19 S-type component. The purified IgG was freeze-dried and stored at 4 °C.

Preparation of N-hydroxysuccinimide ester of Sepharose The method described by Cuatrecasas and Parikh [5] was used. CNBractivated Sepharose 4B beads were treated with 3,3'-diaminodipropylamine and the resulting product was succinylated. The succinylamine-dipropylamine Sepharose was washed with 0.1 M N a O H at room temperature to remove labile carboxyl groups. The succinylated agarose was washed with distilled water followed by extensive washing with dioxane, then suspended in the same solvent and treated with N-hydrosysuccinimide and N-N'-dicyclohexylcarbodiimide. The N-hydroxysuccinimide ester so obtained was exhaustively washed with dioxane and then freeze-dried. The product used in the present studies contained from 50 to 60 #mol of reactive N-hydroxysuccinimide residues per g of freeze-dried material. This was calculated by measuring the amount of N-hydroxysuccinimide formed after complete hydrolysis of the ester; 40-mg samples of the lyophilized material were suspended in 5 ml of a 0.5 M NaHCO3 solution and stirred overnight at 4 °C. After centrifugation and washing, the amount of N-hydroxysuccinimide in the supernatant was measured spectrophotometrically at 260 nm against a standard solution of N-hydroxysuccinimide.

Coupling of lgG to N-hydroxysuccinimide ester of Sepharose The coupling reaction was carried out in 5-ml plastic syringes modified by forcing a 3 mm thick porous plastic disc down to the bottom (Vyon F, Porvair Ltd. 20-50/~m pore diameter), at pH 6.4 (0.1 M phosphate/citrate buffer) or at pH 8.6 (0.1 M NaHCO3 solution) varying the relative proportions of IgG and Sepharose. The activated Sepharose (4 mg) was added to 4.5 ml of a solution of IgG in the proper buffer (from 0.2 to 35.10 -6 M assuming a molecular weight of 160 000 for IgG) and allowed to react under gentle stirring for 14 h at 4 °C. The unreacted IgG solution was discharged and the solid phase was washed with buffer until no detectable protein concentration was found (A280 < 0.01). Uncoupled IgG was measured by the method of Lowry et al. [10] in the combined discharged solution, previously ultrafiltered against saline to remove N-hydroxysuccinimide. The amount of coupled IgG was calculated by the difference. The conjugate was resuspended in 4.5 ml of coupling buffer containing 1 M glycine and stirred for 2 h at room temperature. This was done to ensure total masking of the unreacted active ester groups [5]. The conjugate was then extensively washed with the same buffer to remove glycine and suspended in 0.04 M Tris-acetate buffer, pH 7.4.

249

Coupling of IgG to CNBr-activated Sepharose The method described by Wide [11] was used. The activated Sepharose (4 mg) was added in a plastic syringe to 4.5 ml of a solution of IgG (1.5-10 -6 M ) in 0.1 M sodium phosphate buffer, pH 8.4, and the reaction mixture was stirred gently overnight at 4 °C. The unreacted IgG solution was discharged and the solid phase was washed with the buffer until no proteins were detectable (A280 < 0.01). The amount of unreacted IgG was measured as described above. A coupling yield of about 90 ~o was obtained. The conjugate was treated with 1 M ethanolamine at p H 8 for 1-2 h. The solid phase was then extensively washed at pH 8 to remove unreacted ethanolamine, then suspended in 0.04 M Tris-acetate buffer, pH 7.4, containing 0.25 ~ lysozyme.

Estradiol-binding activity of unconjugated IgG The gel-equilibration method of Pearlman and Crepy [12] was used. 300 mg of Sephadex G-25 ("coarse") were suspended in test tubes with 1.5 ml of 0.05 M phosphate/citrate buffer, pH 7.4, containing 0.5 ~ lysozyme. Gel swelling was carried out by gently stirring the suspension overnight at 4 °C. To each tube were added in turn: 0.1 ml of IgG solution at the proper dilution; 0.1 ml of estradiol standard solution at concentrations ranging from 1 to 50.10-9 M; 0.1 ml of tracer solution corresponding to 0.5.10 - 9 M of 17fl-2,4,6,7-[3H]estradiol. The tubes were shaken for 2 h at 4 °C and then centrifuged for 5 min at 2000 × g. A 0.5 ml aliquot of the supernatant was transferred for counting. The partition factor K' (i.e. ratio of external to internal volume) was determined in each experiment by quadruplicate measurements with tracer in the absence of antibody. The value obtained, K' = 1.85 ~ 0.05, was found to agree closely with the theoretical partition factor determined by using blue dextran [12], thus indicating the absence of non-specific adsorption in our experimental conditions. The bound and free fractions were evaluated from the counts related to total radioactivity and radioactivity in the external volume, by applying the formulae: Lto t =

L 1

(1)

-]- L,; L¢ = F~ + B

F~

-- K-:

B

=Le--Fe~Ltot

(2)

L__A_~

Zt ot

(I__ ~_7__~ L_ )/K

(3)

where L t o t = total ligand concentration, L~ ~ ligand concentration in the internal volume, Le = ligand concentration in the external volume (bound and free), Fe = concentration of unbound ligand in the external volume, B = concentration of ligand bound to antibody.

Estradiol-binding activity of IgG-Sepharose conjugates This was measured as follows: in 3-ml plastic test tubes were added 0.5 ml of the conjugate suspension at the proper dilution, 0.1 ml of the tracer solution, 0.3 ml of 0.04 M Tris-acetate buffer, pH 7.4, and 0.1 ml of estradiol standard solution at concentrations ranging from 1 to 50.10 -9 M. The tubes were then mounted in a ro-

250 tating stirrer and kept at 4 °C for 12 h. The Sepharose suspension was centrifuged for 10 min at 2000 × g. A 0.05 ml aliquot of the supernatant was transferred for counting. The F fraction was directly calculated from the counting results, while the B fraction was calculated by difference. Control experiments established that there was no detectable binding of estradiol either to Sepharose 4B or to a conjugate of 7 S v-globulins (Pentex Inc., Kankakee, Ill. U.S.A.) prepared in the same way as the immuno IgG-Sepharose conjugates. Essentially the same procedure was followed in cross-reactivity experiments, using the systems [3H]estriol-estriol or [3H]estrone-estrone.

Calculation of the binding data The equilibrium constant (K0) and the molar concentration of the combining sites lab0] were calculated from the binding data using the Sips' formula [13]. BIF a= Ko(Abo-- B)

(4)

where B and F are the usual notations for free and antibody-bound ligand and c~ is a heterogeneity index which reduces to unity when a single family of sites is present. In the latter case, Eqn. 4 coincides with the usual Scatchard relationship. In the case of multiple site equilibrium, the Sips' treatment yields intrinsic binding data defining the average behaviour of the system under study on the assumption that the site energy distribution closely resembles a Gauss function. In order to circumvent the effect of a possible dependence of the K0 value on the dose range, the Sips' equation was applied in all the experiments within the same interval of ligand concentration. RESULTS

The reaction of lgG with the active ester of agarose Some preliminary experiments were carried out to define the conditions for the preparation of conjugates with different IgG/agarose mass ratios. In the first place the effect of pH was investigated since the results reported by Cuatrecasas and Parikh for aminoacids [5] suggest that it could affect the coupling reaction of a protein in two ways. On the one hand, the hydrolysis of active agarose ester increases with increasing pH so that the coupling yield should decrease with increasing pH. On the other hand, since the coupling reaction occurs with the unprotonated form of amino groups, both the reaction rate and the yield should be favoured at higher pH values. The coupling reaction was studied at pH 6.4 and 8.6, the same values adopted by Cuatrecasas and Parikh for their experiments. Fig. 1 shows the coupling yields for different initial values of the IgG/Sepharose ratio and for a reaction time of 10 h. The degree of substitution, expressed as nmol of IgG bound per mg of Sepharose, is lower at pH 8.6 than at pH 6.4 for the entire range of IgG concentration considered. For both pH conditions the degree of substitution attains its maximum value for an initial mass ratio of 20-25 nmol IgG/mg matrix. The time course of the reaction was then studied at both pH values, using an initial IgG/matrix mass ratio of about 20 nmol/mg. As it is shown in Fig. 2 the reaction is complete within 1 h at pH 8.6 while it takes more than 4 h at pH 6.4.

251

0 10 (n

~O-----~--O

o/O ~ ° ~

x

oJ

0 E ~3 __= CI

G. 0 ¢.1

I

I

0

I

10

I

I

I

20

30

40

R E A C T E D I g G ( m o l e ~ 1 0 - 9 / m g Sepharose)

Fig. 1. Amount of IgG coupled to the N-hydroxysuccinimide ester of Sepharose after a reaction time of 10 h, at different values of the initial IgG/Sepharose mass ratio. The experiments were carried out at 4 °C in 0.1 M NaHCOa at pH 8.6 (e) or in 0.1 M phosphate/citrate buffer at pH 6.4 (O). "

8

o

o~O.....--o -----o 'o

4

/

2

/~0

0 E

''~0

•

a o.

~0 (J

I 60

I 120

I 180

I 240

REACTION

I 300

TIME

360

(min)

Fig. 2. Time course of reaction of N-hydroxysuccinimide ester of agarose with IgG at 4 °C in: (0) 0.1 M NaHCO3 buffer, pH 8.6 and (C)) 0.1 M phosphate/citrate buffer, pH 6.4. The data refer to an initial mass-ratio IgG/Sepharose of 18.10 -9 mol IgG/mg Sepharose assuming an average molecular weight of 160 000 for IgG. These results suggest that at p H 8.6 the increased availability o f u n p r o t o n a t e d e - a m i n o g r o u p s on the p r o t e i n is n o t able to c o m p e n s a t e for the greater rate o f hydrolysis o f the active ester. T h e occurrence o f this c o m p e t i t i v e effect m a k e s the prep a r a t i o n difficult o f conjugates at p H 8.6 with a prefixed I g G / a g a r o s e mass ratio, unless the s a t u r a t i o n value o f 1.5-1.6 n m o l o f I g G / m g o f agarose is a t t a i n e d (see Fig. 1). F o r this reason the a b o v e - m e n t i o n e d value was a d o p t e d to p r e p a r e the conjugates for the c o m p a r a t i v e experiments r e p o r t e d below.

Binding data for the unconjugated lgG U s i n g a gel-equilibration p r o c e d u r e to evaluate the free a n d b o u n d fraction, the I g G was t i t r a t e d with estradiol. A n o n linear B/F vs B p l o t was o b t a i n e d , indicating a certain degree o f heterogeneity. The curves were linearized b y i m p o s i n g a ~-0.88. T h e value o f the average e q u i l i b r i u m constant, K0, o f the I g G fraction used was

252 found to be 4.8.109 M -1 while a value of 5.0.1011 antibody sites per nmol of IgG was calculated assuming a molecular weight of 160 000 for IgG. The ~o cross-reaction was 2.05 for estrone and 0.45 for estriol. These results confirm that the antibodies used in the present work are able to recognize changes in the functional groups of the Dring of estrogens, which makes the measurement of cross-reactivity a particulary suitable index for monitoring possible variations of site-specificity associated to the insolubilization.

Comparative binding studies on IgG conjugates Binding studies were carried out on conjugates to N-hydroxysuccinimide ester of Sepharose prepared at pH 6.4 and 8.6. It was also felt of interest to compare the results with those obtained using conjugates prepared by direct coupling at pH 8.6 to CNBr-activated Sepharose. Such a comparison should yield indirect information on the role played by the hydrocarbon extension in determining the properties of the conjugated antibodies. As a matter of fact, Sepharose activated with CNBr or with the N-hydroxysuccinimide ester reacts specifically, although through different reaction mechanisms with the protein amino groups. Moreover, the distribution of the active groups within the Sepharose beads must be similar, since CNBr-activated Sepharose is the starting product for the preparation of the active ester. Within the limits set by the difficulty of controlling the coupling yield, the experiments were performed using conjugates with similar IgG/Sepharose mass ratios (approx. 1.5 nmol IgG/mg Sepharose). The binding data are reported in Table I. The residual immunoreactivity (Ko × Abo) is normalized to that of the original lgG fraction and expressed as percentages. It is interesting that, despite the similarity in the total amount of immobilized IgG in the conjugates coupled by different procedures or by the same procedure at different pH, there are important differences in the binding parameters of these derivatives. In comparison with the original antibodies, the insolubilization at pH 8.6 with Sepharose activated by CNBr or by the N-hydroxysuccinimide ester, leads to the same large loss of active sites. The residual sites, however, show an affinity constant very close to that of the original antibodies for conjugates prepared from the active ester. With CNBr-activated Sepharose the affinity constant of the residual sites is approximately three times lower than that of the original antibodies. This accounts for the larger loss of immunoreactivity induced by the coupling to CNBr-activated Sepharose. A considerably higher immunoreactivity is retained by the conjugates prepared at pH 6.4 from the Sepharose active ester. Analysis of the binding data shows that this is due to a six-fold increase of the concentration of residual sites in comparison with the conjugates prepared from the same matrix at pH 8.6. Table I also shows that the specificity of antibodies in solution and in solidphase, as indicated by the cross-reactivity of estriol and estrone, is essentially the same. Binding studies on conjugates with different lgG/Sepharose mass ratio These studies were limited to the conjugates prepared at pH 6.4 using the Nhydroxysuccinimide ester. A summary of the properties of seven different preparations is given in Table II showing the effect of the degree of substitution on the binding data. It is seen that the affinity constant of the residual sites is progressively lowered

253 TABLE I P R O P E R T I E S O F I g G - S E P H A R O S E C O N J U G A T E S P R E P A R E D BY D I F F E R E N T T E C H N I Q U E S , K E E P I N G T H E I g G / M A T R I X M A S S R A T I O C O N S T A N T ( ~ 1.5 n m o l O F l g G / m g SEPHAROSE) T h r e e different preparations of conjugate were studied for each set o f coupling conditions. Since they slightly differed in the I g G / m a t r i x m a s s ratio (from 1.41 to 1.55 n m o l o f I g G / m g Sepharose) it was preferred to give in the table the binding d a t a for each preparation. T h e heterogeneity coefficient a n d the cross-reactivity were m e a s u r e d only for one p r e p a r a t i o n o f each group. State of the I g G

K0" 10 -9 ( M - 1)

Combining sites a ( x 1011)

Residual immunoreactivity b

Heterogeneity coefficient c

(~) C o u p l e d to C N B r activated Sepharose at p H 8.6 C o u p l e d to the N-hydroxysuccinimide ester o f S e p h a r o s e at p H 8.6 C o u p l e d to t h e N-hydroxysuccinimide ester o f S e p h a r o s e at p H 6.4 In solution, in 0.05 M phosphate/citrate buffer, p H 7.4

~ cross reaction (estradiol 100 ~ ) Estrone

Estriol

1.4 1.6 1.3 4.3 4.7 4.4

0.82 0.78 0.65 0.78 0.75 0.81

5 8 4 16 18 21

0.81

2.25

0.56

0.80 --

2.34 ---

0.54 ---

4.7 4.8 4.7

4.01 4,55 4.25

63 71 68

0.83 ---

2.46 ---

0.53 ---

4.8

5.01

100

0.88

2.35

0.49

a N u m b e r o f a n t i b o d y sites per n m o l o f I g G as calculated a s s u m i n g for I g G Mr = 160 000. T h e residual i m m u n o r e a c t i v i t y was calculated for each conjugate as K0 x Abo (see Experimental), n o r m a l i z e d to that o f t h e I g G in solution a n d expressed as percentage. c Value that m u s t be i m p o s e d to the coefficient a in Eqn. 4 (see experimental) to linearize the B / F vs B plot.

T A B L E II EFFECT OF VARYING THE IgG/MATRIX MASS RATIO ON THE PROPERTIES OF IgG C O N J U G A T E D A T p H 6.4 T O T H E N - H Y D R O X Y S U C C I N I M I D E E S T E R O F S E P H A R O S E See t h e f o o t n o t e s to Table I for the interpretation o f the results reported in the last three columns. Conjugate (N)

I II III IV V VI VII

nmol IgG/mg Sepharose

0.10 1.0 1.4 2.3 4.7 7.4 10.2

Average equilibrium constant K0 x 10 -9 (M -1)

Number of combining sites ( x 1011)

Residual immunoreactivity (~)

4.6 4.3 4.1 3.2 2.5 2.0 1.7

4.8 4.3 4.0 3.9 3.2 2.9 2.2

92.00 77.04 68.33 51.91 33.33 24.15 15.58

Cross-reaction estradiol = 1 0 0 ~ Estrone

Estriol

2.41

0.52 --0.50 --0.45

-2.55 --2.3

254 when increasing quantities of IgG are bound to Sepharose. This is accompanied by a loss of binding sites, so that the residual immunoreactivity is of the order of 15 at a substitution degree approximately ten times larger than that of the conjugates used for the experiments reported in Table I. On the contrary, if the substitution degree is kept much lower, i.e. about one tenth that used in the preceeding experiments, the loss of binding sites is less than 10 ~ in comparison to the original IgG. Therefore, the antibodies can be insolubilized with a negligible loss of the immunoreactivity by controlling the degree of substitution. It was thought of interest to compare the rate of reaction with estradiol of the lgG in solution with that of conjugates to increasing degrees of substitution. The results represented in Fig. 3 show that the rate of binding of estradiol to the IgG decreases with the insolubilization of the protein and it is significantly affected by the degree of substitution. B Bma~ ~

/°/ //° ~]

e

I 4

I 8

INCUBATION

I 12

16

TIME (hr)

Fig. 3. Binding of [3H]estradiol to IgG coupled to the N-hydroxysuccinimideester of Sepharose with different IgG/matrix mass ratio: (A) 1.4 nmol of IgG/mg Sepharose; ([~) 4.7 nmol of IgG/mg of Sepharose; (©) refers to unconjugated IgG. The reaction was carried out in 0.04 M Triss-acetate buffer, pH 7.4; 0.05 nM of [3H]estradiol and 5 nM of IgG were used in all the experiments. The binding values are normalized to the fraction of bound radioactivity corresponding to an incubation time of 20 h (Bmax). DISCUSSION The results reported show that the nature of the coupling reaction, the reaction conditions, the interposition of a chemical arm between the protein and the matrix and the protein/matrix mass ratio in the conjugate, are all factors which play a role in determining the loss of immunoreactivity associated with the insolubilization of antibodies. An unequivocal assignment of an order of importance to these factors is difficult since, depending on the experimental conditions, the loss of immunoreactivity can be due either to a loss of active sites without modification of the affinity of the residual sites, or to a combination of both effects. The results reported in Table I show that in the conjugates with similar IgG/ agarose mass ratios prepared at the same pH (8.6) from the N-hydroxysuccinimide ester or from CNBr-activated agarose, the loss of binding sites is essentially the same. In both cases the concentration of residual sites has been decreased to as little as 10-15~o of that existing in the original antibodies. With the active ester, however,

255 one has a particular "all or none" situation where some sites are essentially intact in terms of affinity, while others are completely abolished. Using CNBr-activated agarose, the residual sites of the conjugates show a lower affinity constant in comparison to the original antibodies. This leads to a higher loss of immunoreactivity when CNBractivated agarose is used as matrix. In conclusion, these results suggest that neither the direct coupling nor the coupling by interposition of the hydrocarbon extension between matrix and protein are effective in preventing losses of binding sites. The use of a hydrocarbon extension, however, seems to be effective in safeguarding the behaviour of the sites which have not been abolished by insolubilization, possibly by overcoming the steric restraints imposed by the agarose matrix. A contribution to the question about the mechanism leading to the loss of binding sites comes from the comparative experiments on conjugates prepared from the active ester of agarose at pH 8.6 and 6.4, respectively. For both pH values, the residual sites are essentially intact in terms of specifity and affinity. Their concentration, however, is about six-fold higher at pH 6.4 than at pH 8.6 for conjugates with the same IgG/Sepharose mass ratio. The loss of immunoreactivity is then much lower when the coupling is carried out at pH 6.4. The strong effect of the coupling pH on the loss of sites could be explained on the basis of the results obtained by Cuatrecasas and Parikh [5] in model experiments with amino acids. They predicted that the coupling of a protein to the active ester of Sepharose at pH 6.4, should preferentially involve the terminal a-amino groups. This was based on the finding that the reaction with the active ester occurs solely with the unprotonated form of an amino group. The high pK value of the eamino groups of lysine should then reduce the probability of coupling through these groups at pH values around neutrality. This should favour a more selective coupling and should minimize the danger of inactivation, at least for those proteins where the terminal amino groups are not structurally associated with the architecture of the active sites. It seems unlikely, however, that such an explanation may hold in the case of a protein like IgG, since the abundance of lysine residues is such that unprotonated e-amino groups are probably available also at pH 6.4 at a concentration sufficient to compete with the terminal s-amino groups. On the other hand the arguments of Cuatrecasas and Parikh [5] could be used to postulate the formation of multiple linkages between the IgG molecule and the matrix; such an effect could be more pronounced at pH 8.6 than at pH 6.4 because of the higher concentration of unprotonated e-amino groups per molecule. This could be reflected either in a steric masking of the sites and/or in direct involvement of essential lysine residues when the coupling reaction is carried out at pH 8.6. It must be noted, however, that the formation of multiple linkages cannot be invoked to explain the results of the experiments in which conjugates, prepared at pH 6.4 are compared at increasing IgG/Sepharose mass ratios. There is a strong loss of sites and a lowering of affinity of the residual sites when passing from the mass ratio of 1 nM of IgG/mg of Sepharose to l0 nM/mg, despite the fact that this should decrease the probability of the occurrence of multiple linkages on a single molecule. As shown in Table II the progressive loss of affinity of the residual sites found

256 in these experiments is not accompanied by appreciable variations of specificity with respect to estriol and estrone. The discrimination among the closely related structures of the estrogens probably arises from a subtle interplay of a variety of interactions at the level of the antibody site. Thus preservation of the specificity should suggest that the loss of affinity cannot be attributed to structural alterations in the neighborhood of the binding sites. On the other hand if the mass ratio is lowered to 0.1 nM of IgG/mg of Sepharose, the insolubilization is accompanied by a negligible loss of active sites and no loss of affinity. This suggests that increasing the IgG/Sepharose mass ratio affects the accessibility of the hapten to the antibody sites. This conclusion is supported by the results given in Fig. 3 which show that the time required to reach the equilibrium, for the reaction with estradiol, increases with increasing the IgG/matrix mass ratio in the conjugate. It is likely that the loss of immunoreactivity found in these experiments is related to the peculiar structure of the Sepharose matrix. The polysaccaride chains do not form a random network but probably consist of losely packed regions and more dense regions. Increasing the protein concentration increases the probability that the IgG molecules diffuse up to the more dense regions where the strong steric restraint imposed by the matrix affects the antigen-antibody interaction. Furthermore the advantage of coupling through an arm is lost, and the higher density of active groups may facilitate both the occurrence of multiple linkages and the formation of clusters of protein molecules. Indeed, the results reported show that by keeping the IgG/agarose mass ratio low the antibodies can be transferred from solution to the solid matrix with a minimal loss of sites and affinity, and an overall loss of immunoreactivity of the order of 8 ~ . This result is much more favourable than those reported in the literature. Experiments to assess the immunoreactivity of antibodies coupled to cellulose, as compared with the same antibodies in solution, have been reported by Arends [1]; in his hands the coupling of antibodies to human chorionic gonadotropin to CNBractivated cellulose resulted in a severe lowering of the affinity and in loss of immunoreactivity of the order of 80-90 ~ ; similar results were reported by Moore and Axelrod [3] for antibodies to estradiol coupled to a synthetic polymer. These results are in agreement with those reported in Table II and previously discussed, which show that a non-optimization of the IgG/Sepharose mass ratio can result in a loss of immunoreactivity of 90 ~ or more. Recently Bolton and Hunter [15] have reported that coupling at pH 8.4 whole antisera to steroids (including estradiol) to CNBr-activated cellulose leads to a minimal loss of immunoreactivity as compared to that obtained with the purified IgG fractions. This finding can be rationalized in the light of the previous discussion. In fact, the presence of proteins like albumin, whose rate of diffusion is probably higher than IgG, can exert a protective action by competing with the IgG for the active groups of the more dense regions of the matrix. Such an effect should then provide an alternative to the use of relatively sophisticated procedures to minimize the loss of immunoreactivity. These conclusions may have some practical implications as a guideline to design experiments dealing with the preparation of insolubilized antibodies either for analytical purposes in radioimmunoassay or as specific agents to remove contaminants from the blood in detoxification studies.

257 REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Arends, J. (1971) Acta Endocrinol. 68, 425-430 Hart, I. C. (1972) J. Endocrinol. 55, 51-62 Moore, P. H. and Axelrod, L. R. (1972) Steroids 20, 199-212 Malvano, R., Rolleri, E., Gandolfi, G. and Rosa, U. (1974) Horm. Metab. Res. Suppl. 5, 12-17 Cuatrecasas, P. and Parikh, I. (1972) Biochemistry 11, 2291-2299 Cuatrecasas, P. (1972) Proc. Natl. Acad. Sci. U.S. 69, 1277-1281 Gawronski, T. H. and Wold, F. (1972) Biochemistry, 11,442-448 Lindner, H. R.. Perel, E., Friedlander, A. and Zetlin, A. (1972) Steroids 19, 357-375 Stanworth, D. R. (1960) Nature 188, 156-157 Lowry, O. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (1951) J. Biol. Chem. 193, 265275 Wide, L. (1969)Acta Endocrinol. 63 Suppl., 142, 207-246 Pearlman, W. H. and Crepy, O. (1967) J. Biol. Chem. 247, 182-189 Nisonoff, A. and Pressman, D. (1958) J. Immunol. 80, 417-428 Abraham, G. E., Odell, W. D., Edwards, R. and Purdy, J. M. (1970) Acta Endocrinol. Suppl. 147, 332-346 Bolton, A. E. and Hunter, W. M. (1973) Biochim. Biophys. Acta 329, 318-330