Eur. J. Biochem. 64, 369- 372 (3976)

Binding of Human Hemoglobin and Its Polypeptide Chains with Haptoglobin Coupled to an Agarose Matrix Andreas TSAPIS, Monique ROGARD, Annette ALFSEN, and Constantin MIHAESCO Laboratoire d'Immunochimie, Unite 108 d e l'lnstitut de la Sante et d e la Recherche Medicale, Institut de Recherches sur les maladies du sang, H6pital Saint Louis, Paris, and Laboratoire des Etats Lies Moleculaires, Centre National de la Recherche Scientifique, Universite Rene Descartes, Paris (Received July 1 1 , 1975/February 4, 1976)

The interactions of human haptoglobin covalently linked to agarose with human hemoglobin and with p-chloromercuribenzoic-acid-treated a and [I' chains (a* and ,f?*chains) were studied by flow chromatography and equilibrium binding. The results indicate that in solid state, haptoglobin maintains the same binding characteristics as in solution, the order of binding affinities being : hemoglobin >a* chain >/I* chain. The study of the binding parameters of the a* chain shows an heterogeneity of binding sites on the haptoglobin and an average affinity constant K , of 3.6 x lo4 l/mol.

Haptoglobins are a,-glycoproteins present in the blood plasma of many animal species and posses high binding affinities for the liganded form of hemoglobin. The mechanism of interaction between these two molecules has been largely studied in aqueous solution [l - 31. Under physiological conditions, the overall stoichiometry of the reaction can be schematically described as follows:

While the detailed tridimensional structure of the Hb molecule is known at the atomic scale, considerably less information on the haptoglobin structure is available. The data on the detailed molecular location of the respective combining sites present on either of these reactant molecules are very scarce. It is known that the heavy chains of haptoglobin contain four binding sites, i . r . one pair of sites for each afl hemoglobin dimer [4]. Moreover the binding affinity for the x hemoglobin chain is considerably higher than that for the /?chain [2]. We thought that the use of the haptoglobin niolecule covalently bound to a solid matrix could help the study of this type of interaction. In the present study we have examined the binding characteristics of an haptoglobin-agarose affinity adsorbent previously developed by one of us [5]. The results show that in the Abbreviufrons. Hb, human liganded hemoglobin; Hp, human haptoglobin, ; Aga-Hp, haptoglobin,-, coupled to Sepharose 4 B; a* chain, hemoglobin n chain treated with p-chloromercuribenzoic acid; /J* chain, hemoglobin /i'chain treated with p-chloromercuribenzoic acid.

solid state the haptoglobin molecule preserves essentially the same interacting characteristics as in solution. MATERIALS AND METHODS Preparation of H p s x* Chains, fi* Chains and of Hb Hb was obtained from lysates of human erythrocytes freshly prepared by ion-exchange chromatography on CM-Sephadex [6]. H b concentration was determined by the cyanmethemoglobin spectrophotometric method [7] and expressed as heme equivalents. Throughout this study the Hb was used in the liganded form. The dissociation into a and p chains was carried out during the reaction of H b with p-chloromercuribenzoic acid [8]. The cx and p hemoglobin chains were then separated following the method described by Yip et ul. [9]. The ''C-labelled a* and ,f?* chains were obtained by the same method using p-chloromercuri['4C]benzoic acid with a specific activity of 145 mCi/ mol and 827 mCi/mol. 14C-labelled a* chains with a specific activity of 120 rnCi/mol and 600 mCi/mol were obtained. The specific activity of 14C-labelled p* chains was 210 mCi/mol. The purity of the Hb and of a and fi chain preparations was tested by disc polyacrylamide gel electrophoresis. Hb as well as the a* and p* chains were stored for a short period of time in dialysis bags equilibrated against 100 mM potassium phosphate buffer pH 7.4, 1 mM in EDTA at + 4 "C. The 14C-labelled H b was obtained by reacting a 1.2 mM Hb solution with i~do['~C]acetamide (82 mCi/ mol) at a final concentration of 50 mM at pH 7.4 for

370

3 h at room temperature [lo]. The excess i ~ d o [ ' ~ C ] acetamide was then eliminated by extended dialysis at 4 "C. The specific activity of 14C-labelled H b was of 42.5 mCi/mol Hb chain ( M , 16 520). The 14C activity was counted in a TriCarb 2425 model liquid scintillation system after the peroxidation of the samples. The specific activity was measured after suitable correction for counting efficiency and quenching. Human haptoglobin of H p 2-1 phenotype was purified from pleural fluids of patients with malignant diseases, according to a previously described method [ll]. The purity of the Hp preparations was tested by polyacrylamide gel electrophoresis and by immunoelectrophoresis using a potent polyvalent rabbit antiserum against normal human serum. The H p concentration was measured by its absorbance at 280 nm using an absorption coefficient A;& = 12.0. The binding capacity of the purified H p preparations was measured by fluorescence quenching titration [2] using a Fica spectrophotofluorimeter. All the preparations tested were approximately 90 of the theoretical value of activity. Coupling of Hp to agarose (Sepharose 4B) was effected by cyanogen bromide activation of the agarose following the procedure previously described [ 5 ] . The amount of haptoglobin covalently coupled to agarose was obtained by Kjehdal nitrogen determination on 1-ml aliquots of sedimented material. For most of the preparations of agarose-bound haptoglobin (Aga-Hp) used in this study the content was of 40 + 2 nmol H p per 1.0 ml of sedimented material. This preparation was stored for 4 weeks in 100 mM potassium phosphate buffer pH 7.4, 1 mM in EDTA at 4 "C without significant loss of its binding activity.

Binding of Hb , x* and /3* Chains on Aga-Hp All binding experiments were effected in graduated chromatography columns (diameter 10 mm x 80-mm height) filled with 1-ml bed volumes of Aga-Hp settled for 24 h. In the study of the binding of the ligands to the Aga-Hp matrix two kinds of experiments were undertaken as follows. Binding Assays by Flow Chromatography. Aliquots of 1 ml of Hb, c(* and /3* chain solutions were applied on the columns. The ligand concentration varied between 0.01 mM and 0.80 mM. Subsequent washing of the column was performed with the same buffer until theeffluent reachedaminimal concentration. 0.27 pmol ''C-labelled Hb and 0.60 pmol of ''C-labelled a* chains were added to a l-ml chromatographic bed of Aga-Hp followed by washing the column with' the buffer previously described. Aliquots of 1 ml were collected. Solutions of H b and /3* chains (0.121 mM) were used for displacement experiments.

Binding Characteristics of Haptoglobin Coupled to Agarose

Equilibrium Binding Experiments. Binding experiments at equilibrium were performed by adding 1-ml volumes of 14C-labelled M* chains to l-ml bed volumes of Aga-Hp in chromatographic tubes. The ligand concentration varied between 1.66 pM and 0.66 mM. The tubes were stoppered, gently mixed for 30 min at 4 'C and then spun for 10 min at 500 rev./min in a centrifuge at + 4 "C. The volume of the eluted solution was carefully measured and the concentration of the ''C-labelled a* chains determined by counting the I4C radioactivity. The bound a* chains were calculated as difference between the total amount of a* chains added and the free a* chains measured. Suitable corrections were introduced for non-specific binding and liquid trapping in the solid matrix by using the same amounts of human IgG coupled to agarose in the same conditions. All experiments were prepared in triplicate samples. The amount of bound ligand (Hb or isolated chains) was measured by Kjeldahl nitrogen determination with an automaticTechnicon system, using aconversion factor of 6.60. Alternatively the amount of the bound ligand was obtained by the difference between the protein content before and after binding to the AgaHp : (a) by spectrophotometric measurements at 540 nm of the cyanmet form of the ligand, and (b) by liquid scintillation counting of 14C-labelled proteins.

RESULTS AND DISCUSSION

Binding Assuys by Flow Chromatography The binding of Hb and of the a* and fl* chains to the Aga-Hp was studied. The curves obtained (Fig. 1) show firstly that the addition of increasing amounts of H b leads to a saturation plateau at a value of 1.5 mol Hb chain per mol Hp. The quantitative data differ since the stoichiometry of the saturated Hp . Hb complex in solution was shown to be of 1 mol H b or 4 mols x chain per 1 mol Hp, while the calculated data for H b fixed as Aga-Hp corresponded to only 0.39 mol H b per mol of Hp.The difference in the binding capacity observed could be due either to H p denaturation during coupling or to the Hp binding sites being hidden. The latter could occur during the H p coupling due to the chemical binding and to the orientation of the Hp molecule on agarose. Secondly, the addition of similar concentrations of a* chains resulted in smaller amounts of bound ligand and a saturation plateau not being reached even at 0.60 mM. Thirdly, the p* chain binding was so weak that it was measurable only at high concentrations (0.50 mM and 0.73 mM). These results indicate that the order of binding offinity of Aga-Hp for the ligands is: H b > x chains >/3 chains. The same order was observed in previous studies in solution [2].

A. Tsapis, M. Rogard, A. Alfsen, and C. Mihaesco

371

//-

-

0

0

0.1

0.2

0.3

- - -- - - - .

-fP

,---

0.4

-/

0.5

0.60.73

ilgand added (rnM)

Fig. 1 . Binding of H b , r* and [j* chains to Ago-Hp. Aliquots o f 1 ml of ligand were applied on 1 ml o f Aga-Hp. The excess (non-bound) ligand was eluted by the bufer described in Materials and Methods. chains ( 0 ) Binding of Hb ( x ) , a* chains (0)and

a*

loo

J

Fraction number Fig. 2. Elurron profile of excess &and upplied on I-tnl column. ''C-labelled I* chains on Sepharose 4 B (@), ''C-labelled Hb on Aga-Hp ( x ) and 14C-labelled c(* chains on Aga-Hp (0)

Studying the elution profile of excess H b or a* chains applied on the column, we found as shown in Fig. 2 that the H b excess was eluted from Aga-Hp as a symmetrical peak similar t o that obtained on pure agarose with no trailing. The a* chain elution profile showed considerable trailing and did not reach the base line even at high volumes of elution. These data provide additional evidence of the high binding affinity of Hb for the Aga-Hp, contrasting with the weaker interaction between this adsorbent and the a chains. In addition, the a chain trailing profile

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 Fraction number

Fig. 3. Displacement assay.s ~ f " ~ C - l c ~ h e l l e dchains a * bound to Axa- H p by 100 t ~ ~ ~ p o l a . s ~ s i u m p h o . ~ huffer. p } l a i e pH 7.4 ( x ), hy Hh solution in (he sarne huffi.v (0.121 m M ) (a) and h?. /j* chains solution in thr .wnw huffcv (0)

suggests a conspicuous degree of heterogeneity of the affinity constants of this system. In displacement experiments, using Aga-Hp on which a* chains were bound at or near saturation and subsequently submitted to elution with B* chains and Hb tetramer solutions, approximately 50 yd of the a* chains bound to Aga-Hp are displaced by addition oftheB* chains. Theremaininga* chainsare irreversibly bound to the affinity adsorbent and can not be chased by the H b solution. Nevertheless in the absence of added p* chains, H b was able to displace all the bound a* chains (Fig. 3).

Equilibrium Binding Experiments with a* Chains It is known [I21 that the a* chains are in monomeric state in the solvent used and behave as univalent ligands in their binding with Hp. Our studies focused on the interaction of a* chain with Aga-Hp. The values found in the saturation experiments were taken into account and the value for the functionally active Hp per 1-ml bed was taken as 0.39 x 38 nmol= 14.8 nmol. The distribution of the intrinsic association constants Ki was expressed following the Scatchard relation [12 1: Y

-=Kn-Kr c (where r represents the number of ligand molecules bound per H p molecule at c, free concentration of

372

A . Tsapis, M. Rogard. A. Alfsen, and C. Mihaesco: Binding Characteristics of Haptoglobin Coupled to Agarose

( K z = 6.23 x lo4 Ijmol). By extrapolating to the r axis, it was found that the maximum number n of sites was 4.3 and the number of the high-affinity sites was 1.5. The maximum number of sites found for Aga-Hp is in fair agreement with the generally accepted value of 4 sites per Hp molecule in solution [2,4]. Nevertheless the proportion of high-affinity binding sites found in our experiments is significantly lower than the previously recorded figure of 50 % of the total number of binding sites [ 2 ] .These results are consistent with the results obtained by flow chromatography binding assays. Taken together, these data show that when coupled to a solid matrix, the H p molecule keeps the same general binding properties for H b its a and p chains, as in solution. The Aga-Hp represents a suitable agent for the detailed study of the mechanisms of interaction between Hp and the H b molecule and its subunits. Fig. 4. Scutchurd plot of ‘‘C-lubelled a* hindinz on Agu-Hp. The procedure was carried out as described in Materials and Methods. Inset is the Sipsian representation of the same results

We are grateful to Dr Franqoise Lavialle and Dr Marcel Wiks for the stimulant discussions and the critical reading of this paper.

REFERENCES ligand and n is the maximum number of ligand molecules that can be bound per H p molecule), and plotted as shown in Fig. 4. Assuming a Sipsian distribution of the intrinsic association constants Ki, the heterogeneity index c( of the Hp binding sites was calculated according to the following equation [I 31 :

The calculated a index was 0.72 and the average affinity constant K, was 3.6 x lo4 I/mol. The shape of the Scatchard plot indicates a conspicuous heterogeneity of binding sites of Aga-Hp for a chains and two populations of sites can be distinguished : a minor population of high-affinity binding sites (with an extrapolated Kl = 5.26 x lo5 l/ mol) and a major one with low-affinity constant

1. Waks. M. & Alfsen, A. (1966) Biochem. Biophys. Res. Commun. 23, 62 - 67. 2. Chianconc, E., Alfsen, A , , Ioppolo, C., Vecchini, P., FinazziAgro, A., Wyman, J. & Antonini, E. (1968) J. Mol. Biol. 34, 347 - 356. 3. Waks, M., Alfsen, A,, Schwaiger, S. & Mayer, A. (1969) Arch. Biochem. Biophys. 132, 268 - 278. 4. Nagel, R. L. & Gibson, Q. H. (1967) J. Biol. Chem. 242, 3428 - 3434. 5. Klein, M. & Mihaesco, C. (1973) Biochem. Biophys. Res. Commun. 52,774-778. 6 . Rosa, J. (1961) Bull. SOC. Chim. Biol. 43. 479-484. 7. Kampen, E. J. & Zijlstra, W. G. (1961) Clin. Chim. Acla, 6 , 538 - 544. 8. Bucci, E. & Fronticelli, C. (1965) J . Biol. Chem. 240, PC551 PC 552. 9. Yip, Y . K . , Waks, M. & Beychok, S. (1972) J . Bid. Chem. 247, 7237 - 7244. 10. Benesch, R. E. & Benesch, R. (1962) Biochemi.stry, I , 735-738. 11. Wtiks, M. & Alfsen, A. (2966) Arch. Biochem. Biophys. 113, 304 - 314. 12. Scatchard, G . (1949) Ann. N . Y. Acud. Sci, 51, 660-672. 13. Sips, R. (1948) J . Chem. Phys. 16, 490-495.

A. Tsapis and C. Mihaesco Laboratoire d’lmmunochimie, Unite 108 de I’I.N.S.E.R.M., Centre Georges Hayem, HBpital Saint Louis, 2 Place du Docteur-Fournier, F-75475 Paris-Cedex-10, France M. Rogard and A. Alfsen Laboratoire des Etats-Lies Moleculaires, C.N.R.S., 45 Rue des Saints-Peres, F-75006 Paris, France