Immunochemical studies of pancreatic colipase-lipase interaction employing immobilized synthetic peptides

Vol. 189, No. 3, 1992 December 30, 1992

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS Pages 1374-1381

IMMUNOCHEMICAL STUDIES OF PANCREATIC COLIPASE-LIPASE INTERACTION EMPLOYING IMMOBILIZED SYNTHETIC PEPTIDES Nathalie Rugani, Laurence de la Four&e, Robert Julien, Louis Sarda*+ and Jdlle Rathelot Institut de Chimie Biologique, Faculte Saint-Charles, Place V.Hugo 13331 Marseille cedex 3, France

Received

October

4,

1992

In view to study the possible participation of the sequence portions of colipase including or close to the free carboxyl groups at positions 15 and/or 72 to the binding with pancreatic lipase, we have used three synthetic peptides matching portions a-1659-67 and 67-72 of the amino acid sequence. Polyclonal rabbit anticolipase immune serum, which cross-reacts with peptides in ELISA, was fractionated on columns of peptide coupled to Sepharose. Of the three fractions of antibodies, only that interacting with peptide 8-16 had the capacity to inhibit colipasedependent lipase activity by specifically preventing the association of lipase with its protein cofactor previously bound to lipid.We conclude that the region spanning residues 8-16 of colipase is of importance for colipase-lipase interaction in the active complex formed at Q 1992Academic Press, 1°C. interface.

Pancreatic colipase is a small non enzymatic protein acting as a specific cofactor by restoring the activity of pancreatic lipase inhibited by bile salt (1). Bile salt prevents the binding of lipase at interface and colipase functions by anchoring lipase to its water-insoluble substrate. The enzyme and its cofactor form an active complex in a one to one molar ratio (2). Colipase is secreted in a proform (procolipase) of 95 amino acid residues with 5 disulfide bridges. Procolipase is converted in vivo to a short form (90 amino acid residues) by trypsin which specifically cleaves the Argg-Glyg bond under mild conditions (3). The short form of the cofactor (activated colipase) has enhanced lipid binding capacity and is considered as the form of colipase acting in vivo (3). There is also evidence that the amino terminal activation pentapeptide cleaved by trypsin has regulatory properties for fat intake (4). The biological function of colipase requires specific protein-lipid and protein-protein interaction.Then, two distinct surface binding sites, a lipid and a lipase binding site have been postulated on the cofactor molecule. Studies with physico-chemical techniques including spectrophotometry and NMR have shown that highly hydrophobic regions 52-59 and 6-10, conserved in colipases from vertebrates, are part of the lipid binding site of the cofactor (5-7). *To whom all correspondence should be addressed. + Supported by CNRS (ERS No26 Lipolyse Enzymatique) and by the Commission of the European Communities (BRIDGE T-Project PL 890328). 0006-291X/92

$4.00

Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.

1374

Vol.

189,

No.

3, 1992

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH

COMMUNICATIONS

Identification of amino acid residues involved in colipase-lipase interaction has come first from studies with porcine colipase inactivated by selective chemical modification at free carboxyl groups (8,9) and lipase fragments prepared by limited proteolysis with chymotrypsin (10,ll). It was shown that dicarboxylic acid residues at positions 15 and/or 72, present in homologous proteins, might play a role in colipase-lipase binding and that interaction involved the Cterminal domain of lipase. Recently, determination of the three dimensional X-ray structure of the colipase-lipase complex formed in solution in absence of interface, has revealed that residues 44-46,65-67 and 89 of the cofactor interact with amino acids from various P-strands in the C-terminal non catalytic domain of lipase (12). Predictive conformational studies on colipases from several mammalian species have indicated that free carboxyl groups at positions 15 and 72 are likely located in major antigenic regions and, therefore, that specific antibodies directed to these regions could be obtained (13). In previous reports (14,15) we have described the preparation and properties of anticolipase polyclonal antibodies and the separation of fractions of antibodies specifically directed to the lipid and lipase binding sites of colipase , respectively. Synthetic peptides have been successfully used as probes for studying the antigenic structure of proteins in relationship with biological function (16- 18). In this communication, we report studies carried out with three synthetic peptides matching the regions of colipase which include residues at or close to positions 15 and 72. Immobilized

peptides were used to separate

fractions of anticolipase polyclonal antibodies. Studies of the inhibitory properties of these fractions support the conclusion that the sequence region of colipase around glutamic acid at position 15, which includes the free hydrophobic

amino terminal part of the polypeptide,

interacts with lipase in the active complex formed at lipid-water interface. MATERIALS

AND METHODS

Chemicals and reagents :Tween 20, bovine serum albumin (BSA), casein and 2,2’-azinobis-(3ethylbenzothiazoline sulfonate) (ABTS) were from Sigma (St-Loais,Mo,USA). Microtitration plates were purchased from Falcon (Oxnard,Ca,USA). Peroxydase conjugated antibodies against rabbit immunoglobulins were from Biosys (Compibgne, France). Phosphate buffered isotonic saline (PBS) Dulbecco”A”, pH 7.2, was obtained from Oxoid Ltd (Barinstocke, England).CNBr-activated Sepharose CH4B was purchased from Pharmacia-France (StQuentin-Yvelines, France). Proteins : horse pancreatic lipase, free of colipase, and porcine colipase (procolipase form) were prepared at the laboratory according to previously described procedures (2,19). Protein determination: concentration of solutions of porcine colipase, horse lipase and immunoglobulins was estimated spectrophotometrically, at 280 nm, using absorbance coefficients (At’:)

of 3.5, 13.3 and 13.5, respectively.

Assay of colipase activity : colipase was assayed at 25% and pH 9.0 with the potentiometric method, using the triolein-deoxycholate lipolysis system and colipase-free horse lipase (20). Colipase activity was expressed in units. Under the standard assay conditions, one colipase unit corresponds to the liberation of one microequivalent fatty acid per minute by colipaseactivated lipase. Synthetic peptides : peptides YlOL and Y9L, corresponding to sequence regions of porcine colipase 8-16 (Be-Ile-Asn-Leu-Asp-Glu-Gly-Glu-Leu) and 59-67 (Tyr-Cys-Cys-Pro-Cys-Glu1375

Vol. 189, No. 3, 1992

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

Arg-Gly-Leu) were synthetizedby NeosystemLaboratory(Strasbourg,France).PeptideY 11V correspondingto sequenceregion 67-76 (Leu-Thr-Cys-Glu-Gly-Asp-Lys-Ser-Leu-Val)was synthetizedin Dr.Van Rieschoten’slaboratory(Marseille,France).One residueof tyrosine,not presentin colipase,was addedat the N-terminusof peptidesY 1OLand Y 11V to allow better spectrophotometric detectionandiodinationby the lactoperoxidase method(21). The purity of eachpeptidewas over 85%. In peptidesY9L and Y 1lV, free sulfhydryl groups, correspondingto half-cystine residuesin the original protein, were blocked by reaction with mercuric acetateto preventthe formation of disultide bonds.Peptideswere storedat -8O’C in dry state. Coupling of synthetic peptides to carrier bovine serum albumin: five mg of peptidetogether with iodinatedpeptide(1.7 105CPM) were m ixed with BSA (molar ratio BSA / peptide:l/28) in lml of PBS pH 7135. An equalvolume of glutaraldehyde(2% in water) was then addedto the m ixture and, after 24 hours incubation at room temperature,the peptide-serumalbumin conjugate was dialyzed againstone liter of the same PBS buffer. The yield of the coupling reaction was determined from the radioactivity incorporated in serum albumin coupled peptide.Tbepeptideconjugatewas kept in solutionat - 2O’C. Antipeptide antibodies : immune serawere producedin rabbitsby four injections of a m ixture of 0.4 mg of free peptidewith 0.6 mg of BSA-coupledpeptide emulsified in 1 m l of Freund’s completeadjuvant.Animals werebleededoneweek after the last injection. Anticolipase immune serum:the antiserumwas obtainedas describedpreviously (22), freezedried and stored at -80°C. The immune serum was reconstitutedby dissolving 80 mg of lyophilized materialin one m l of distilled water and the solution was filtered ( M illipote filters with 0.45 pm pores). Immunoa@zity chromatography on immobilized synthetic peptides : eachpeptide (5mg), was

separately coupled to one gram of CNBr-activated SepharoseCH-4B according to the Manufacturer’sinstructions.The extent of the coupling reactionwas followed by adding trace amountof peptidelabeledwith radioactiveiodine.Yields were 61,67 and 72% for Y lOL, Y9L and Y 1lV, respectively. Immunoaffinity chromatography separationswere carried out as follows: 4 m l of reconstitutedanticolipaseimmuneserumwere incubatedovernightat 4OCwith an amountof Sepharose-coupled peptidecorrespondingto 2.5 mg of free peptide.The m ixture was placed in a column (1.15 cm x 3 cm) and washedwith 10 m l of PBS buffer pH 7.35 so that the column output returnedto the 280 nm baseline.Bound immunoglobulinswere eluted with 10 m l of 0.1 M glycine-HCl buffer pH 2.8. The eluate was immediately neutralizedby adding an equal volume of 1M Tris-HCl buffer pH 8.4. Purified antibodiesfractions eluted from the columns were concentrated(Millipore Ultra Free MC) and kept at -20°C. About 5 nmoles,.0.75 nmoles and 1 nmoles of antibodies were recovered from the columns of immobrhzedpeptidesY lOL, Y9L andY 1lV, respectively. Immunoinactivation of colipase : the effect of antibodieson colipaseactivity was studiedunder

different conditions, as reported already (14). Experiments with fractions of anticolipase polyclonal antibodiesseparatedon immobilizedsyntheticpeptideswere carried out as follows: in a first seriesof assays,colipase(5 pg) was m ixed with purified immunoglobulins(100 pg) in 0.1 M PBS, pH 7.2, with 0.1% BSA ( final volume 150 pL). The m ixture was left for one hour at 37°C and one night at 4’C. Residualcolipase-dependent lipase activity was measured by adding an aliquot (15 pL) of the m ixture to the standardlipase assaysystem, followed by 10 pg of colipase-freelipase.Ina secondseriesof assays,0.5 pg of colipasewas preincubated with lipid substratein the standardlipase assaysystem (final concentrationof colipase: 1.5 10 -9 M) and 0.15 mg of immunoglobulinswere added(final concentrationof antibodies:30 lo-9M). After ten m in , residualcolipase-dependent lipaseactivity was measuredby adding 10 pg of colipase-free lipase. Under these assayconditions, it can be discriminated between inhibitory antibodiesreacting with regions at the lipid or lipase binding site of the cofactor, respectively. Immunoinactivationof colipasewas also studiedwith immune seraraisedagainst synthetic peptides. In this case, assayswere carried out by m ixing 10 pg of colipase with 200 pL of non-diluted serum containing about 300 ng of specific antibodies,in 0.1 M PBS, pH 7.2, with 0.1% BSA(final volume: 300 pL). Residualactivity was determinedon an aliquot (15 pL) of the incubationm ixture. 1376

Vol.

189,

No.

3, 1992

BIOCHEMICAL

AND

BIOPHYSICAL

RESEARCH

COMMUNICATIONS

EL.ZSA titrations : ELISA was usedto determinethe titer of antipeptideimmune sera.Titrations

were carried out underconditionsreportedbefore(15) with minor modifications: microtitration plates were pretreatedwith a 0.2% (w:w) solution of glutaraldehydein O.lM phosphatebuffer pH 5.0 (100 p.L per well). After washing , each well was coatedwith 50 pL of a solution of peptidein 0.1 M PBS pH 7.2 (10 pg per mL) and left for 16 hours at room temperature.Wells were washed with distilled water and the remaining protein binding sites were blocked by adding 100 pL of a 1% aqueoussolution of casein to each well. Finally, the plates were incubated with series dilutions of antipeptide immune serum for 16 hours at 37’C. Bound antibodies were determined with peroxidase-coupledgoat antirabbit immunoglobulins antibodies. The titer of the antiserum correspondedto the dilution that gave half maximal absorbanceobtained with non-diluted immune serum. ELISA was also used to probe the immunochemical reactivity of antipeptideantibodiesand fractions of anticolipasepolyclonal antibodies with colipase. Assays were performed under usual conditions (15) by coating microtitration plateswith 50 ng and lng of colipase,respectively. RESULTS

Cross-reactivity of antipeptide antibodies : the reactivity of immune seradirectedagainstthe

threesynthetic peptidesmatchingportionsof the sequenceof colipasewas studiedwith ELISA. Microtitmtion plateswere coatedwith peptidesor porcinecolipase(procolipaseform). Results of ELISA titrations are presentedin Figure 1. Curves of Figures Ia to lc indicate that antipeptide antibodies reactedwith peptideantigensand, to a lesser extent, with the protein cofactor boundto plastic plates.This bringsevidencethat selectedpeptidesequencesam part of exposedantigenic regionsof colipase.Titers of antiseradirectedagainstpeptidesY lOL, Y9L, and Y 11V estimatedfrom datashown in Figure 1 were 1: 50, 1: 800 and 1: 100, respectively. Immunoinactivationexperimentscarriedout with antipeptideimmune serafurther indicatedthat none of the antibodies had the capacity to inhibit colipase-dependentlipase activity at concentrationusedin assay. Cross-reactivity of mticolipasepolyclonal antibodies : the reactivity of antiporcineprocolipase polyclonal antibodies was also studied by ELISA. Microtitration plates were coated with amounts of peptide ranging from 1 to 1O‘tng per well (Figure 2). Data of Figure 2 show that anticolipase antibodies react with the three peptides bound to microtitration plates thus indicating that subfractions of immunoglobulins can be separated by immunoaffinity chromatographyon immobilized peptides.

Figure Enzyme-linked immunosorbent assay of antipeptide immune sera. Wells of Fhe microtitration plates were coated with 500 ng of peptide (full circles) or with 50 ng of porcine colipase (full squares) and incubated with 50 pL of series dilutions of antisera. Experiments were carried out with antisera against peptides Y lOL(a), Y9L(b), Y 1 lV(c), respectively. 1377

Vol.

189,

No.

BIOCHEMICAL

3, 1992

1

AND

lLl

BIOPHYSICAL

100

1cfJo

RESEARCH

COMMUNICATIONS

loo00

Peptide (ng per well)

Figure Enzyme-linked immunosorbent assay of anticolipase polyclonal antibodies with synthetic peptides.Microtiter plateswere coatedwith increasingamountsof peptidesY lOL(a), Y9L(b), Y 1 lV(c) and incubatedwith 50 pL of reconstitutedanticolipaseantiserum containing 250 ng of specific antibodies.

Separation andproperties of subfractions of onticolipase polychal antibodies : three columns prepared with Sepharose-coupled peptides Y lOL, Y9L and Y 1 IV, respectively, were loaded with the same amount of anticolipase polyclonal antibodies. After elution of immunoglobulins adsorbed on peptides, each fraction was assayed for its capacity to react with porcine procolipase in ELISA and to inhibit colipase-dependent lipase activity under standard assay conditions. Results of the ELBA titrations are reported in Table 1. Data of Table 1 indicate that the three fractions of anticolipase polyclonal antibodies separated on immobilized peptides had retained their capacity to react with specific antigenic regions of colipase. Results of the immunoinactivation

experiments are reported in Table 2. It appears from the data of Table 2

that only immunoglobulins separated on Sepharose-coupled peptide Y 1OL which matches the sequence portion 8-16 of colipase, inhibited colipase-dependent lipase activity. Inhibition occurred either after mixing antibodies with colipase in absence of interface, or with colipase previously bound to emulsified lipid. Then, this fraction contains antibodies

having the

capacity to hinder colipase-lipase association by interacting specifically with antigenic region in or at vicinity of the lipase binding site of colipase, at the N- terminus of the protein.

Table Enzyme-linked immunosorbent assay of subfractions of antiporcine procolipase polyclonal antibodies separated on immobilized synthetic peptides. Titrations were performed by coating the plates with Ing of porcine procolipase per well. The same amount of immunoglobulins (100 ng per well) was used in’all assays. Absorbance was measured at 405 nm.

Antibodies adsorbed on peptide Y 1OL 0.337

Antibodies adsorbed on peptide Y9L

Antibodies adsorbed onpeptideY11V

0.362

0.356

1378

Total colipase polyclonal antibodies 0.535

Vol.

189,

No.

3, 1992

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH

COMMUNICATIONS

Table 2. Immunoinactivation of colipase-dependent lipase activity by fractions of anticolipase polyclonal antibodies separated on Sepharose-coupled synthetic peptides. Results are expressed as percent of residual colipase activity.

Antibodies adsorbed on peptide Y 1OL

Antibodies adsorbed on peptide Y9L

(1)

(2)

(1)

30

40

100

(2) loo

Antibodies adsorbed onpeptideY11V (1)

(2)

100

100

(1) Antibodies preincubated with colipase in absence of lipid. (2) Antibodies mixed with colipase in presence of lipids.

DISCUSSION The interaction of pancreatic lipase with water-insoluble triacylglycerol substrate, in presence of bile salt, specifically requires colipase. The protein cofactor binds to interface and anchors lipase by forming a mole to mole active complex.The first attempts to characterize the amino acid residues of the particular domain of the cofactor involved in colipase-lipase association at interface were made by a chemical approach. Results have led to the conclusion that regions including or near free carboxyl groups at positions 15 and/or 72 play a critical role in the stabilization of the active complex. Later, evidence was gives that the small non catalytic Cterminal domain of pancreatic lipase (12KDa ) separated by limited proteolysis of the enzyme protein (SOKDa),had the capacity to bind with colipase in solution and it was proposed that this domain had the same function in the intact enzyme. From the recent determination of the three _ dimensional structure of the pancreatic procolipase-lipase complex formed in solution, it appears that residues 4446,65-67 and 89 of procolipase are interacting with amino acid of various p strands in the C-terminal domain of lipase. However, there is no indication from the X-ray structure that regions around residues Glu 15 and/or Asp 72 are involved in procolipaselipase binding. Predictive conformational studies on colipase from several mammalian species have shown that residues Glu 15 and Asp 72, conserved in homologous proteins are located in exposed regions of colipase corresponding to major antigenic sites. On the basis of this observation , we have used three synthetic peptides matching sequence portions of colipase including or close to dicarboxylic acid residues at positions 15 or 72 to separate fractions of antiporcine procolipase polyclonal antibodies and we have studied the capacity of each fraction to inhibit colipasedependent lipase activity under different experimental conditions. Results reported here indicate that the fraction of antibodies adsorbed on peptide Y 1OL matching sequence portion 8- 16 of the porcine cofactor inactivates colipase previously bound to lipid by interacting specifically with accessible antigenic region located in or close to the lipase binding site . By contrast, antibodies adsorbed on peptides Y9L and Y 1 lV, matching fragments 59-67 and 67-72 of colipase, respectively, had no inhibitory effect under the same experimental conditions. Our results 1379

Vol.

189, No. 3, 1992

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

support the hypothesis according which the peptide segmentat positions 8-16 of colipase participatesto colipase-lipasebindingin the activecomplexformedat interface. Previousstudieson colipase-lipaseassociationhave shown that the two proteins interactsin solution with an apparentbinding constant(Ka) in the rangeof 106M while the value of Ka is increasedby severalordersof magnitude(Ka=lOlO)in presenceof a lipid-water interface (2, 23).Thehigher affinity betweencolipaseandlipaseat interfacesuggeststhat additionalproteinprotein interactionsam formed,likely as a result of the conformationalchangesinducedby the interfacial binding of lipase.Actually, it is well establishedfrom the work by Brozozowskiet al (24) that the helical flap structureof lipaseis displacedduringconformationalchangeassociated with interfacial binding. From the resultsreportedhere,it can be speculatedthat the N-terminal regionof the proteincofactorinteractswith lipasein its interface-activated conformation. REFERENCES 1. Borgstrom, B. and Erlanson-Albertsson, C. (1984) in Lipases ( Borgstrom, B. and Brockman, H.L. eds), pp. 151-183,Elsevier, Amsterdam. 2. Rathelot, J., Julien, R., Bose-Bieme,I.,Gargouri, Y., Canioni, P. and Sarda,L. (1981) Biochimie

63, 227-234.

3. Borgstrom, B., Wieloch, T. and Erlanson-Albertson,C. (1979) FEBS Let?. 108, 407-410. 4. Erlanson -Albertsson,C. (1992) Biochem. Biophys. Actu 1125, l-7. 5. Sari, H., Entressangles,B. and Desnuelle,P. (1975)Eur.J.Biochem 58, 561-565. 6. Rathelot, J., Canioni, P., Bose-Bierne, I., Sarda, L., Kamoun, A., Kaptein, R. and Cozzone, P.J. (1981) Biochem.Biophys. Actu 671, 155-163. 7. Cozzone, P.J., Canioni, P., Sarda,L. and Kaptein, R. (1981) Eur. J. Biochem. 114 ,119126. 8. Erlanson, C., Barrowman,J., and Borgstrom,B. (1977) Biochim Biophys. Acta 489, 150-162. 9. Erlansson,C. (1977) FEBS L&t. 84, 79-82. 10. MahC-Gouhier,N. and Leger, C.L. (1988) Biochem. Biophys. Actu 962,91-97. 11. Abousalham,A., Chaillan, C., Kerfelec, B., Foglizzo, B. and Chapus,C. (1992) Protein Engng. 5, 105-111. 12. Van Tilbeurgh, H., Sarda,L., Verger, R. and Cambillau, C. (1992) Nature 351, 159162. 13. Bellon, B., Dezan, C., Rugani, N. and Sarda, L. (1991) Znt. J. Peptide Prot. Res. 38,483-490.

14. Bose-Bieme, I., Perrot, C., Sarda,L. and Rathelot, J. (1985) Biochem. Biophys. Actu 827, 109-l 18. 15. Bose-Bieme,I., De La Foumibre, L., Rathelot, J., Him M .and Sarda,L.(1987)Biochem. Biophys . Acta 911,326-333.

16. Bidart, J.M., Troalen, F., Bousfield, G.R., Birken, S. and Bellet, D.H. (1988) J. BioZ. Chem. 263,10364-10369. 1380

Vol.- 189,

No.

3, 1992

BIOCHEMICAL

AND BIOPHYSICAL

RESEARCH

COMMUNICATIONS

17. Troalen, F., Bellet ,D.H., Ghillani, P., Paisieux, A., Boheron, C.J. and Bidart, J.M. (1988) J. Biol. Chem. 263, 10370-10376. 18. Savoca, R., Schwab, C. and Bosshara, H.R. (1991) J. Immunol. Methods 141, 245 2.52.

19. Canioni, P., Julien, R., Rathelot, J., Rochat, H. and Sarda, L. (1977) Biochimie 919-925.

59,

20. Rathelot, J., Julien. R., Canioni, P., Coeroli, C. and Sarda, L. (1975) Biochimie 1117-1123.

57,

21. Thorell, J.I. and Johansson,B.G. (1971) B&hem.

Biophys. Actu 251, 363-369.

22. Rathelot, J., Delori, P. and Sarda,L. (1983) Biochem. Biophys. Actu 742, 39-48. 23. Erlanson-Albertsson,C. (1983) FEBS Lett. 162, 225-229. 24. Brzozowski, A.M., Derewenda,U., Derewenda,Z.S., Dodson, G.G., Lawson, D.M., Turkenburg, J.P., Bjorkling, F., Huge-Jensen,B., Patkar, S.A. and Thim, L. (1991) Nature 351, 491-497.

1381