J. Mol. BioE. (1991) 220, 1041-1053

Effect of Single Amino Acid Replacements on the Thermodynamics of the Reactive Site Peptide Bond Hydrolysis in Ovomucoid Third Domain Wojciech Ardeltt

and Michael Laskowski Jr

Department of Chemistry, Purdue University West Lafayette, IN 47907, U.S.A. (Received 26 Septem,ber 1990; accepted 24 April

1991)

We have measured equilibrium constants, Khyd, at pH 6 for the hydrolysis of the reactive site peptide bond (bond between residues 18 and 19) in 42 sequenced variants (39 natural, 3 semisynthetic) of avian ovomucoid third domains. The values range from 64 to approximately 35. In 35 cases the effect of a single amino acid replacement on Khyd could be calculated, 13 are without effect and 22 range from a factor of 1.25 to 55. Several, but not all, of the effects can be rationalized in terms of residue-residue interactions that are affected by the reactive site hydrolysis. As the measurements are very precise it appears that additional measurements on designed rather than natural variants should allow for the precise measurement of side-chain-side-chain interaction energies. Keywords: proteinase inhibitors; Kazal inhibitors; ovomucoids; peptide bond hydrolysis equilibria; amino acid replacements

1. Introduction Frequently hydrolysis of a single peptide bond in a globular protein does not produce two independent fragments. Instead the two fragments remain combined either by covalent crosslink such as disulfide bridge(s) or by non-covalent interactions. In such cases we can usefully define an equilibrium constant: ,n*\

where (P) and (P*) are the concentration of intact protein and of the protein with a single, specified peptide bond hydrolyzed, respectively. The value of Khyd is of interest to those who carry out enzymatic synthesis of peptide bonds (for a review, see Fruton, 1982), and to those who wish to measure the thermodynamic strength of localized interactions in proteins (e.g. see Laskowski t Scheraga, 1956). The value of Khyd is of special concern in the study of protein inhibitors of serine proteinases, which obey “the standard mechanism” (Laskowski & Kato, 1980). For such inhibitors we write: E+I

ko.

g,C ‘; E+I*, 011

(2)

where E is the enzyme and C is the stable complex between enzyme and inhibitor. I and I* are the t Present address: Alfacell Corporation, Bloomfield, N

(6)

where c(,~ is a fraction of modified inhibitor at the equilibrium. I and I* are virgin and modified inhibitors, with t’he reactive site peptide bond intact and. hydrolyzed, respectively. Time-course of the hydrolysis (virgin inhibitor used as a substrate) or resynthesis (modified inhibitor as substrate) was followed by a modification of the met,hod described by Ardelt & Laskowski, 1982). A Mono Q column within a fast protein liquid chromatography system (FPLC, Pharmaria) was used rather than the

t In the majority of cases, the cleavage proceeded at pH 1.6 to 1.8 but for OMCHAS( -), a very weak inhibitor of SGPB, pH of the incubation mixture had to be elevated to 30, and lower yield of the modified inhibitor was obtained. For three 3rd domains (Australian ca 1 are largely due to changes in residue-residue interactions upon peptide bond hydrolysis. The conclusion that Pi residues may exert interaction-independent effects is an unhappy one as in many ovomucoid third domains Pz Thr interacts with P; Glu in a side-chain-side-chain hydrogen bond. There is also a weak intraresidue H-bond within the P; Glu19 residue (Fujinaga et al., 1982; Bode et al., 1985). As would be expected, and is actually found in OMSVP3( - ) (Musil et al., 1991)) Table 5 Effect of P, replacement8 on K,,,,,, Amino

acid

IRU

Ala Met Hse V&l “By definition as this residue is present substance OMTKY3( -) see Fig. 3. b Table 4. ‘Table 3.

K hYd 1GO” 1.55 b 1.64’

206' 2.42' in

the

reference

Peptide Bond Hydrolysis these interactions break upon reactive site hydrolysis. (There is no comparable interaction in OMJPQ3 because the P2 residue is Pro). The residue replacement data (Tables 3 and 4) strongly support the P,-P; interaction. Replacing P2 Thr17 by Arg17 increases Khyd by a factor of 2.87. Replacing P; Glu19 by Leul9 (unfortunately along with a PI2 Asp7 + Asn7 replacement) leads to a 1.84-fold increase in Khyd. In order to make this reasoning more rigorous we would need more variants but it seems likely that, the equilibrium constant for the Thrl7-Glul9 H-bond in virgin ovomucoid third domains is about 1 to 2, which is equivalent to saying that Thr17 and Glu19 are hydrogen bonded to each other 50 to 67% of the time. Other results from our laboratory (T. Bigler, M. A. Qasim & M. Laskowski Jr. unpublished results) confirm this conclusion. (k) The P’, 5 residue The P’i5 residue is Asn33 in both OMJPQ3( -) and OMSVPS( - ). In both structures the two hydrogen atoms of ND2 of this residue are involved in hydrogen bonds with Pro17 0 and Thr17 0 and with Asp17 0 and Glu19 0, respectively. The Asn33 OD2 is also hydrogen bonded with Asn36 N. Upon conversion to modified inhibitor the two H-bonds involving Asn33 ND2 break in OMSVPS( -) as residues 17 (P,) and 19 (Pi) separate (Musil et al., 1991). We have data on the replacement of Asn33 by Ser33 (Table 3). It is likely that in this derivative there is only an H-bond to Asn36 N and the two others are broken. The finding of a 2.4-fold increase in KW upon replacement is consistent with this view. However, the equilibrium constant for forming this set of bonds seems quite small, just 1.4, yet Asn33 is very strongly conserved both in ovomucoid third doma,ins and in Kazal family inhibitors in general. (I) The P; residue The effects discussed above were relatively simple t,o explain because they involved interactions that coupled the peptide chain on both sides of the reactive site bond. This is not the case with Tyr20 (Pi). The only important interaction made by Tyr20 is that, one of the faces of its ring is stacked with the ring of Pro22 (Pi). The other face is in contact with the solvent. Somewhat surprisingly Tyr20 and Pro22 remain stacked in the modified forms of both OMJPQ3*( -) and OMSVP3*( - ) (Musil et al., 1991). Yet replacing Tyr20 by either His20 or by Asp20 leads to a more than fivefold increase in Khyd. Tt is quite interesting that in spite of a great difference between His and Asp the effect of replacing Tyr20 by either is essentially the same. This lends credence to a notion (supported by the study of many other variants on their association with proteinases (Park, 1985)) that it is the Tyr and not t,he replacing residue t,hat is responsible for the

in Ovomucoid Third Domain

1051

observed effect. Our older explanations for low Khyd values for the reactive sites of protein proteinase inhibitors focused primarily on the low entropy gain realized after opening the reactive site ring (Finkenstadt et al., 1974; J. Otlewski and M. Laskowski Jr, unpublished results). These assumptions are also supported by Mu&l et al. (1991) who found that the conformation of modified inhibitor is strikingly similar to that of the native one. In particular the residues on the P’ side of the split bond undergo little change. We are inclined to attribute a good deal of this effect to the Tyr29-Pro22 stacking. In the His20 and Asp20 variants we assume that there is no such stacking and the entropy gain due to peptide bond hydrolysis is much greater. (m) The P’,, effect This can be explained in a similar manner. We know here that Lys55 + Thr55 replacement is without effect (Table 4) while Lys55 -+ Glu55 lowers Khyd from 1.5 to 1.1. In several ovomucoid third domain structures Lys55 NZ is relatively close to Tyr20 OE but not really close enough to hydrogen bond. It seems clear, however, that replacement of Lys55 by Glu55 would allow for Tyr20-Glu55 tyrosyl-carboxylate hydrogen bonding. Indeed the presence of such a hydrogen bond has been detected by comparison of fluorescence pH titration curves of OMCSNS( -) and of OMDUK(S)( -) (M. A. Qasim and M. Laskowski, unpublished results). The finding that Khyd becomes lower when the Tyr2&Glu55 hydrogen bond is present can be readily explained by further reducing the entropy gain of the P residues on peptide bond hydrolysis. (n) Other effects We have no facile explanations for the changes in K,,yd resulting from the Ala15 -+ Gly15 and Ala15 -+ Asp15 substitutions at the P4 residue. However, it is worth noting that the 16-35 disulfide bridge undergoes a conformational change upon peptide bond hydrolysis (relief of strain) (Musil et al., 1991), therefore the nature of the residue at P4 may well contribute to the strain that the P3 disulfide experiences in virgin inhibitor. We have no data to support an explanation for the threefold increase in Khyd upon Asn36 -+ Asp36 replacement at’ P;s, especially since this substitution stabilizes against denaturation but destabilizes against the reactive site peptide bond hydrolysis (Table 2). Similarly, we know little about, the causes of the P8 Tyrll -*His1 1 cahangr (Table 4). It is also interesting to comment on several cases where changes in Khyd might have been expected but were not found. The P;2 Thr30 + Ser30 replacement leads to a tenfold loss in domain stability of ovomucoid third domain (J. Otlewski & M. Laskowski Jr, unpublished results) (see Table 3). Thr30 is strongly conserved. It is present in 129 of 131 natural domains we have sequenc~ed. Yet the

1052

W. Ardelt and M. Laskowski Jr

replacement of Thr30 + Ser30 is without effect on Khyd and on the association with various enzymes (Park, 1985). Similarly C. March, J. Otlewski & M. Laskowski Jr (unpublished results) found that conversion from virgin to modified inhibitor has only scant effect on the Tyr31-Asp27 interaction as measured by the effect of modification on the pK of Asp27. This is quite consistent with the X-ray crystallographic findings (Musil et al., 1991) that the undisturbed in H-bond is Tyr31-Asp27 OMSVP3*(-). Yet the OMJPQ3*( - ) and Tyr31-Asp27 H-bond is the strongest identified interaction contributing to the domain stability (J. Otlewski & M. Laskowski Jr, unpublished results).

(0) Conclusions The measured Khyd values are extremely accurate provided that their value is relatively close to unity. This is shown by (1) the extraordinarily good agreement of the additivity cycles, (2) the finding that when an effect of a substitution is small the ratio of the two Khyd values is almost precisely 1.00, (3) the extraordinarily good fit of Khyd pH dependence to theoretical relations (e.g. Fig. 5, and Ardelt & Laskowski, 1983). In our experience Khyd values that lie near unity are the most accurate conformation-dependent parameters that can be determined in protein chemistry. This is probably due to the fact that Khyd is a directly measured ratio of concentrations of virgin and modified inhibitor in the same mixture. Most other equilibrium constant measurements involve assumptions such as two-state approximation and a method for conversion of observed signal into fraction denatured, for overall denaturation measurements. Thus, the accuracy of Kden measurements would be expected to be lower. Effects of single amino acid substitutions are readily isolated and in all the cases we have studied thus far, are strictly additive, although that additivity is not expected to persist when substitutions of each of pairwise interacting residues, e.g. Thrl7-Glul9, Tyr20-Pro22, are investigated. Isolation of single residue effects is likely to be simpler for Khyd than for Ln, since changes in a smaller number of positions affect Khyd more than Kden. In attempting to explain effects of single residue substitution on Khyd we are aided by the fact that three-dimensional structures of intact and nicked proteins could generally be obtained and indeed were obtained (see Musil et al., 1991). In many cases the logical molecular explanation could be advanced for the effects observed. The work reported here suffers (and gains) from being a survey of available natural materials. The gain is that we can talk about the range of Khyd values among natural species (Fig. 4) and ultimately about the behavior of third domain Khyd during evolution of birds (Laskowski & Fitch, 1989). The loss is clear, once an interaction between residues is suspected it can be better probed by designed variants. Such work is now going on in our laboratory.

We thank Professor G. Kalnitzky for aspergillopeptidase B, the Purdue team (Laskowski et al., 1987, 1990) for ovomucoid third domain variants, the Martinsried team (Musil et al., 1991) for the X-ray crystallography and Drs J. Otlewski and M. A. Qasim for the use of their as yet unpublished data. Dr M. Wieczorek contributed both his semisynthetic variants and incisive discussion. We were also helped in the interpetation by Dr Richard Wynn. This work was supported by NIH grant GM10831, NSF grant PCM-8111380 and by NATO travel grant 85/0323.

References Ardelt, W. & Laskowski, M. Jr (1982). Analytical ion exchange chromatography of proteins. Anal. Biochem. 120, 198-203. Ardelt, W. & Laskowski, M. Jr (1983). Thermodynamics and kinetics of the hydrolysis and resynthesis of the reactive site peptide bond in turkey ovomucoid third domain by Aspergillopeptidase B. A&z Biochem. Polon. 30, 115-126. Ardelt, W. & Laskowski, M. Jr (1985). Turkey ovomucoid third domain inhibits eight different serine proteinases of varied specificity on the same Leu”-Glu” reactive site. Biochemistry, 24, &;3-5320. Bode, W., Epp, O., Huber, R., Laskowski, M. Jr & Ardelt, W. (1985). The crystal and molecular structure of the third domain of silver pheasant ovomucoid (OMSVP3). Eur. J. Biochem. 147, 387-395. Dobry, A., Fruton, J. S. & Sturtevant, J. M. (1952). Thermodynamics of hydrolysis of peptide bonds. J. Biol. Chem. 195, 149-154. Drapeau, G. (1976). Protease from Staphylococcus aureus. Methods Enzymol. 45, 469-475. Edsall, J. T. & Wyman, J. (1958). Biophysical Chemistry, vol. 1, Academic Press, Inc., New York. Empie, M. W. & Laskowski, M. Jr (1982). Thermodynamics and kinetics of single residue replacements in avian ovomucoid third domains: Effect on inhibitor interactions with serine proteinases. Biochemistry, 21, 2274-2284. Fastrez, J. & Fersht, A. R. (1973). Demonstration of the acyl-enzyme mechanism for the hydrolysis of peptides and anilides by chymotrypsin. Biochemistry, 12, 2025-2034. Finkenstadt, W. R., Hamid, M. A., Mattis, J. A., Schrode, J., Sealock, R. W., Wang, D. & Laskowski, M. Jr (1974). Kinetics and thermodynamics of the interaction of proteinases with protein inhibitors. Proteinase Inhibitors (Fritz, H., Tschesche, H., Greene, L. J. 5, Truscheit, E., eds), pp. 389-411, Springer-Verlag, New York. Fruton. J. S. (1982). Proteinase-catalyzed synthesis of peptide bonds. Advan. Enzymol. 53, 239-306. Fujinaga, M., Read, R. J., Sielecki, A., Ardelt, W., Laskowski, M. Jr & James, M. N. G. (1982). Refined crystal structure of the molecular complex of Streptomyces griseus protease B, a serine protease, with the third domain of the ovomucoid inhibitor from turkey. Proc. Nat. Acad. Sci., U.S.A. 79, 4868-4872. Homandberg, G. A., Mattis, J. A. & Laskowski, M. Jr (1978). Synthesis of peptide bonds by proteinases. Addition of organic cosolvents shifts the peptide bond equilibria towards synthesis. Biochemistry, 17, 5220-5227.

Peptide Bond Hydrolysis Jurasek. L.. cJohnson, P., Olafson, R. W. & Smillie, L. B. (1971). An improved fractionation system for pronase on CM-Sephadex. Can. J. Biochem. 49, 1195-1201. Kato, I., Schrode, ,J., Kohr, J. & Laskowski, M. Jr (1987). Chicken ovomucoid: Determination of its amino acid sequence. determination of the trypsin reactive site, and preparation of all three of its domains. Biochemistry, 26. 1933201. Klemm. ,J. I).. Wozniak, J. A.. Alber, T. BEGoldenberg, I). P. (1991). Correlation between mutational destabilization of Phage T4 Lysozyme and increased unfolding rates. Biochemistry, 30, 589-594. Laskowski, M. Jr bz Fitch, W. M. (1989). Evolution of avian ovomucoids and of birds. In The Hierarchy of Lifp (Fernholm. B., Bremer, K. & Jornvall, H., eds), Chapter 27, pp. 371-387. Elsevier Science Publishers B.V.. Karlskoga, Sweden. Laskowski. ,M. Jr & Kato. I. (1980). Protein inhibitors of proteinasrs. Annu. Rev. B&hem. 49, 593-626. Laskowski. M. Jr & Scheraga, H. A. (1954). Thermodynamic considerations of protein reactions. I. Modified reactivity of polar groups. ,I. Am. Chem. Sot. 76, 630556319. Laskowski. M. Jr & Scheraga, H. A. (1956). Thermodynamics considerations of protein reactions. II. Modified reactivity of primary valence bonds. J. Am. Chum. Sot. 78. 5793-5798.

Laskowski. M. .Jr. Tashiro. M.. Empie. M. W., Park, S. J., Kato. I.. Ardelt. W. & Wieczorek, M. (1983). Relationship bet,ween the amino acid sequence and inhibitory artivity of protein inhibitors of proteinases. Prtrteinasv Inhibitors: Medical and Biological Aspecta (Katunuma. N.. Umezawa, H. & Holzer, H., Sci. S Press, pp. 55- 68. Japan L oc. rds). Tokyo/Springer Verlag, Berlin, Japan. Laskowski. M. .Jr. Kato. I.. Ardelt, W., Cook, J., Denton, A.. Empir. M. W.. Kohr, W., Park, S. J., Parks, K., Schatzley. 1~.I,., Schoenberger, 0. L., Tashiro, M., Vichot, (i.. Whatley, H. E., Wieczorek, A. & WYeczorek. M. (1987). Ovomucoid third domains from 106 avian species: Isolation, sequences, and enzyme-inhibitor hypervariability of contact residues. Niochrmistry, 26. 2022221. Laskowski. M. Jr, Park, S. J., Tashiro, M. & Wynn, R. (1989). Design of highly specific inhibitors of serine proteinasrs. Protein Recognition of Immobilized Ligands (Hutchens. T. W.. ed.). vol. 80, pp. 1499168. Alan R. Liss. New Pork. Laskowski. M. .Jr, Apostol. I., Ardelt, W., Cook, J., Giletto. A.. Kelly. C. A.. Lu, W., Park, S. J., Qasim, M. A.. Whatley. H. E.. Wieczorek, A. & Wynn, R. (1990). Amino arid sequences of ovomucoid third

in Ovomucoid Third Domain

1053

domain from 25 additional species of birds. J. Prot. Chem. 9, 715-725. Mattis, J. A. & Laskowski, M. Jr (1973). pH dependence of the equilibrium constant for the hydrolysis of the peptide bond in soybean Arts 63-Ile reactive-site trypsin inhibitor (Kunitz). Biochemistry, 12, 2239-2245.

Musil, D., Bode, W., Huber, R., Laskowski, M. ,Jr, Lin, T.-Y. & Ardelt, W. (1991). Refined x-ray crystal structures of the reactive site-modified ovomucoid inhibitor third domains from silver pheasant (OMSVP3*) and from Japanese quail (OMJPQ3*). J. Mol.

Biol.

220,

739-755.

Papamokos, E., Weber. E.. Bode, W.. Huber, R., Empie, M. W.? Kato, 1. &, Laskowski. M. Jr (1982). Crystallographic refinement of