Cystatin Mimicry by Synthetic Peptides

BioL Chem. Hoppe-Seyler Vol. 373, pp. 465-470, July 1992

Cystatin Mimicry by Synthetic Peptides F. GAUTHIER a G. LALMANACH a, T. MOREAU a, F. BORRAS-CUESTA b AND J.HOEBEKE a a

Laboraloire d' Enzymologie el de Chimie des Prot6ines, U.R.A. 1334 du Centre National de la Recherche Scientifique, Universilo Francois Rabelais, Faculte" de M&lecine, 2bis Bd Tonnelte 37032 TOURS cedex, France b

Departamento de Medecina Iniema, Faculdad de Medecina, Universidad de Navarra, E-31080 PAMPLONA, Espana

Summary Synthetic peptides which tentatively mimic the cystatin inhibitory surface were used to study the mechanism of inhibition of cysteine proteinases by their natural inhibitors. The inhibitory properties of these peptides depend mainly on the presence of the QxVxG consensus sequence. N and C-terminal peptide derivatives bearing large hydrophobic groups showed dramatically improved inhibition. Molecular dynamic studies after energy minimization showed that the non covalent interaction between these hydrophobic groups induced the formation of a loop structure which probably favours inhibition. Antibodies were raised against one of these peptides, which recognized kininogens in the serum of all mammal species tested, but not cystatins from family two.

Introduction The mechanism of inhibition of cysteine proteinases by their natural inhibitors of the cystatin superfamily is not yet fully understood. There is evidence, however, that it differs from the standard mechanism reported for serine proteinases [1]. This interaction does not require a functional cysteinyl residue at the proteinase active site but occurs through extended contact surfaces involving several conserved regions of cystatin molecules [2-5]. These conclusions were drawn from the results of several approaches used to investigate the mechanism of inhibition. These included the elucidation of the crystallographic structures of free and proteinase-bound inhibitors [6,7], simulation of enzyme-inhibitor interactions by molecular modeling [8], the use of recombinant cystatins modified by site-directed mutagenesis [9-13], kinetic analysis of the interaction between cysteine proteinases and isolated inhibitory cystatin-like fragments [14-16] and the use of synthetic peptides corresponding to putative inhibitory fragments of cystatins [17-19]. We have developed and extended this last approach by constructing synthetic peptides which mimic the supposed inhibitory surface of cystatin C-like molecules. Such cystatin mimotopes were designed to be used as cysteine proteinase inhibitors and to induce antibodies which could recognize the reactive inhibitory site of natural inhibitors.

Design and synthesis of cystatin-mimicking peptides A series of peptides was constructed starting from the informations provided by the 3D structure of chicken cystatin [6]. The model peptide (K12K) included the three segments thought to intervene in the mechanism of inhibition: the N-terminal fragment which possibly accomodates the S1-S2 subsites of the proteinase [5], the QxVxG consensus sequence, and the PW dipeptide present as ß-hairpin loops in cystatins [2,4]. This 12-residue model peptide, designated as K12K [20], was substituted essentially within its N and C-terminal segments, to give a series of 31 peptides. These peptides have been screened for their papain inhibitory properties by incubating papain (0.8 nM) with an excess of peptides in the wells of a microtitration plate, and reading the

Brought to you by | Universite de Tours Authenticated Download Date | 1/29/16 10:03 AM

466

F. Gauthier et al.

Vol. 373 (1992)

released fluorescence from a N-methyl coumarylamide substrate [Lalmanach et al, in preparation]. All peptides but those with substitutions in the three central residus of QVVAG segment showed inhibitory properties. Representative peptides were selected and their inhibition constant Ki was measured. The results are summarized in table 1. They show that the QVVAG segment is essential for the interaction, since substitutions in the QVVAG sequence lead to a complete loss of inhibition (peptide n°27). Peptides substituted for their N-terminal Gly residue (n°30) or for their C-terminal PW dipeptide (n°29) remained inhibitory. However the Clengthening of the N-terminal segment (n°5,21,16) or the introduction of an Ala residue as a spacer between the QVVAG and the PW dipeptide (n°21/n°22) resulted in a significant improvement of the Ki (Table 1).

Table 1. Ki values for the interaction of papain with peptides.

Peptide

Ki (μΜ)

KGAGQWAGPWK [K12K] KGAGQWAGLLK KGAGQDLDGPWK KLLLQWAGPWK VGGQWAGPWK LVGGQWAGPWK GAQWAGAPWK VGGQWAGAPWK [K13K] LVGGAQWAGAPWK

1 29 27 30 22 20 5 21 16

140 150

n.s 180 200 110 60 40 30

All peptides were first screened for their inhibitory properties towards papain, and Ki values of selected samples were determined according to Moreau et al [20].

Analysis of the time-dependence of inhibition revealed that almost complete enzymatic activity was recovered after about two hours of preincubation with most of the peptides tested. This suggests that these peptides may be cleaved by papain, and was confirmed by reverse phase h.p.l.c.. A single cleavage site was found in all peptides assayed. Unexpectedly, this cleavage site was located at the Ala-Gly bond of the QVVAG segment. There was no significant cleavage in the N-terminal segment, which includes the conserved Gly residue and has been previously described as a substrate-like region [3,19].

Table 2.

Peptide

[Kl 2 Κ] [K13K] [K13CK]

Ki values for the interaction of papain with substituted and derivatized K12K and K13K peptides. Ki (μΜ)

KGAGQWAGPWK 140 Mns-KGAGQWAGPWK-Mns 0 .5 VGGQWAGAPWK 40 KVGGQWCacmGAPWK 15 F1-KVGGQWC_„GAPWK-F1 0 .5


Vol. 373 (1992)

Cystatin Mimicry by Synthetic Peptides

467

Two sets of experiments were used to improve the stability of cystatin-mimicking peptides and thus make them better inhibitors than substrates. First the PI Gly residue in the QVVAG segment was substituted to tentatively reduce the rate of cleavage by papain [21] without modifying the interaction. Second the N and C-terminals were derivatized with large fluorescent residues, since it has been suggested that this improves substrate affinity for the enzyme [22]. For this purpose, the Gly residue of the QVVAG segment of K13K (n°21) was replaced by a Cysacm (S-acetamidomethyl-L-cysteine) to give the so called K13CK peptide. This peptide was derivatized using fiuorescein isothiocyanate and the difluoresceinylated derivative was tested. Peptide K12K (n°l) was disubstituted with mansyl groups as previously described [20]. There was a dramatic decrease in Ki values after derivatization of N and C-terminal ends of both peptides by hydrophobic groups (Table 2). Synthetic peptides which include the QVVAG consensus segment of cystatins may therefore be used as inhibitors of cysteine proteinases. However, the conformation of this segment in synthetic peptides differs from that of natural inhibitors since cystatins do not appear to be cleaved by papain. It could be that derivatization by hydrophobic groups stabilizes the QVVAG in a conformation closer to that of cystatins. This can be predicted by molecular modeling analysis.

Cystatin-mimicking peptide modeling and molecular dynamic studies In order to understand the structure-function relationship for peptides derivatized with large hydrophobic groups, we simulated their structure by molecular modeling and dynamic studies after energy minimization. Peptides were constructed using the Biopolymer module of Insight II software (Biosym). Energy minimization and molecular dynamics were studied using Discover (Biosym). The difluoresceinylated derivative of the K13CK peptide is represented in Fig. 1 as a minimized extended structure and after molecular dynamic simulation at one of the lowest energy conformations.

Figure 1. Molecular dynamic study of the FI-K13CK-F1 conformations.

Energy minimization was done with 500 iterations of the steepest descent algorithm, followed by 500 iterations of conjugate gradient. The molecular dynamic simulation was run at 300K for 50 ps. Conformalional stales were captured every 0.5 ps and the total energy was represented as a function of time. A and B represent the minimi/.cd extended conformation and the predicted conformation at one of the lowest levels of energy respectively. Small arrows indicate the positions of fluorcsccinyl groups.


468

F. Gauthieret al.

Vol. 373 (1992)

Fluoresceinyl groups interact strongly by Hydrophobie contacts, favouring a loop conformation of the peptide which corresponds much better to that of chicken egg white cystatin [6] than the crude peptide [Lalmanach et al, in preparation]. This loop formation would also render the peptide less susceptible to hydrolysis, so improving its inhibitory properties.

Investigation of the cystatin inhibitory function by anti cystatin-mimicking peptide antibodies: Antibodies directed against epitopes which participate in the reactive inhibitory site of cystatins may help define the structure of this site. In previous attempts to study the inhibitory function of rat thiostatin (T-kininogen), we used a set of monoclonal antibodies which recognized the different cystatin-like domains of this inhibitor [23,24]. However none of these antibodies modified the inhibitory properties of thiostatin towards papain or lysosomal cathepsins, suggesting that the reactive inhibitory surface was poorly antigenic. As an alternative to this procedure, we used antibodies raised against inhibitory peptides which mimic the cystatin inhibitory surface, might recognize the inhibitory site of natural molecules and possibly discriminate between cystatin families. These antibodies were raised in rabbits against peptide K12K [25]. Specific anti-peptide antibodies were purified by affinity chromatography and assayed against several members of cystatin families 2 and 3. Human L and H-kininogens were recognized by anti-K12K antibodies, whereas chicken cystatin and rat cystatins C and S were not. Subsequent assays with the plasma of seven different mammalian species revealed that all kininogens were recognized by these antibodies with the notable exception of rat thiostatin [25]. This suggests that the inhibitory site of the kininogens have a great structural homology, and corroborates previous data on the particular behaviour of thiostatin [26,27]. Isolated inhibitory domains of kininogens resulting from enzymatic digestion [14], were also recognized by antipeptide antibodies, indicating that crosslinking of two neighbouring inhibitory sites by IgG is not essential for recognition. Since none of the family 2 members (i.e. chicken cystatin, human and rat cystatins C and S) was recognized, it may be concluded that antibodies can discriminate between cystatins belonging to family 2 and to family 3 of the cystatin superfamily [28].

I I

Human LMWK Human HMWK Rat Thiostatin Thiostatin domains Chicken Cystatin

3e-8

5e-8 Antibody concentration (M)

Figure 2. Effect of anti-K12K antibodies on the inhibitory properties of cystatin superfamily members.

Cysteine proteinase inhibitors were mixed with increasing amounts of anti-K12K antibodies as described in [25]. Papain was incubated for 10 minutes with the mixture and the residual enzymatic activity was recorded [29].


Vol. 373 (1992)

Cystatin Mimicry by Synthetic Peptides

469

Interestingly, these antibodies blocked, in a dose-dependant manner, the capacity of kininogens to inhibit papain, thus demonstrating that they are directed against the inhibitory site of these inhibitors (Fig. 2). Accordingly, the inhibitory properties of family 2 members and of rat thiostatin remained unchanged after preincubation with antipeptide antibodies. These results suggest that the mechanism of inhibition of cystatin could slightly differ from one family to another, which could help to explain the somewhat differing reports on the relative importance of the different cystatin fragments in the inhibition of cysteine proteinases [3,5,9,12,13,16].

Acknowledgements We are grateful to M. Garcia-Granero, C. Berasain [University of Navarra, Spain] and M. Femer-Di Martino for their help in peptide synthesis and M. Brillard-Bourdet for N-terminal protein sequencing. This work was supported by grant CRE 895006 from the Institut National de la Ξαηίέ et de la Recherche Medicale. GL holds a fellowship from the Region Centre. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

Laskowski M. & Kato J. Γ19801 Ann. Rev. Biochem. 49. 593 - 626. Ohkubo I., Kurachi K., Takasawa T, Shiokawa H. & Sasaki M. [1984] Biochemistry 2 3, 5691 - 569. Abrahamson M., Ritonja A., Brown M.A., Grubb A., Machleidt W. & Barrett A.J. [1987] J. Biol. £henL 2 6 2,9688-9694. Lindahl P., Alriksson E., J rnvall H. & Bj rk I. [1988] Biochemistry 2 7. 5074 - 5082. Machleidt W., Thiele U., Laber B.,Assfalg-Machleidt I., Esterl A., Wiegand G., Kos J., Turk V. & BodeW. [1989] FEBS Lett. 243. 234 - 238. Bode W., Engh R., Musil D., Thiele U., Huber R., Karshikov Α., Brzin J., Kos J. & Turk V. [1988]. EMBO J. 7, 2593 - 2599. Stubbs M., Laber B., Bode W., Huber R., Jerala R., Lenarcic B. & Turk V. Γ1990] EMBO J. 9. 1939 1947. Bode W., Engh R., Musil D., Laber B.,Stubbs M., Huber R. & Turk V. [1990] Biol. Chem. HoppeSevler371. 111- 118. Abe K., Emori Y, Kondo H., Arai S. & Suzuki K.[1988] J. Biol. Chem. 263, 7655-7659. Nikawa T, Towatari T, Ike Υ. & Katunuma N. [1989] FEBS Leu. 255. 309 - 314. Jerala R., Trstenjak - Prebanda M., Kroon-Zitko L., Lenarcic B. & Turk V. [1990] Biol. Chem. Hoppe-Sevler371. 157-160. Thiele U., Assfalg-Machleidt L, Machleidt W. & Auerswald E.A. [1990] Biol. Chem. Hoppe-Scyler 371, 125- 136. Arai S., Watanabe H., Kondo H., Emori Υ & Abe K. [1991] J. Biochem. 109, 294- 298. Salvesen G., Parkes C., Abrahamson M., Grubb A. & Barrett A.J. [1986] Biochem. J. 234. 429 - 434. \bgel R., Assfalg-Machleidt L, Esterl A., Machleidt W. & M ller-Esterl W. [1988] J. Biol. Chem. 2 6 3, 12661 - 12668. Moreau T., Gutman N., Faucher D. & Gauthier F. [1989] J. Biol. Chem. 264, 4298 - 4303. Teno N., Tsuboi S., Itoh N., Okamoto H. & Okada Υ [1987] Biochem. Biophvs. Res. Commun. 143, 749 - 752. Marks N., Berg M.J., Makofsee R.C. & Danho W. [1990] Peptides 11, 679 - 682. Grubb A., Abrahamson M., Olafsson L, Trojnar J., Kasprzykowska R., Kasprzykowski F. & GrzonkaZ. [1990] Biol. Chem. Hoppe-Sevler371. 137 - 144.


f

470

1 \ J |

20.

!

21. 22. 23. 24. 25. 26.

| | | ! I !

27. 28. 29.

EGauthieretal.

Vol. 373(1992)

Moreau T., Hoebeke J., Lalmanach G.,Hattab M. & Gauthier F. [1990] Biochem. Biophys. Res. CgmmurU67, 117-122. Koga H., Yamada H., Nishimura Y, Kato K. & Imoto X [19911 J. Biochem. 108, 976 - 982. Lowbridge J. & Fruton J.S. [1974] J. Biol. Chem. 249, 6751 - 6754 Lalmanach G., Adam A., Moreau T., Gutman N. & Gauthier F. [1991] Eur. J. Biochem. 196, 73 -78. Lesage S., Bouhnik J., Richoux J.P., Baussant T., Gauthier F., Eager K., Corvol P. & Alhenc-Gelas F. [1992] Eur. J. Biochem. in press. Lalmanach G., Hoebeke J., Moreau T., Ferrer - Di Martino M. & Gauthier F. [1992] J. Immunol. Methods in press. Kitagawa H., Kitamura N., Hayashida H., Miyata T. & Nakanishi S. [1987] J. Biol. Chem. 262. 2190 - 2198. Gauthier R, Gutman N., Moreau T & El Moujahed A. [1988] Biol. Chem. Hoppe- Sevler 369, 251 255. Barrett A.J., Rawlings N.D., Davies M.E., Machleidt W., Salvesen G. & Turk V. Γ19861 in: Proteinase Inhibitors [Barrett AJ. and Salvesen G., Eds] pp 515 - 569, Elsevier Amsterdam. Moreau T., Gutman N., El Moujahed A., Esnard F & Gauthier R [1986] Eur. J. Biochem. 159, 341 346.
