[25]
AFFINITY
LABELING
469
chromatography and in labeling experiments and the capacity for the detection and characterization of interaction in reactivity probe studies. Acknowledgments This article is dedicatedto Professor E. M. Crook in the year of his retirement. I thank the ScienceResearchCouncilfor support of this work, and CarolineReddickand Joy Smithfor the rapid productionof the typescript.
[25] A p p l i c a t i o n o f A f f i n i t y L a b e l i n g f o r S t u d y i n g Structure and Function of Enzymes
By BRYCE V. PLAPP Affinity labeling is a popular method that may be used to determine topography of active sites, to elucidate enzyme mechanisms, and to provide new, rationally designed drugs. But despite the promise of the approach and the investment of much empirical work, our great expectations have not yet been fully realized. This is due, in part, to our lack of knowledge about the specific enzymes we wish to modify and to our ignorance of fundamental principles that explain the reactions of proteins with small molecules. But it is also partly due to the limited goals of some researchers who only wish to prove that they have prepared an active-sitedirected reagent (or to invent a new type of inactivation) when the ultimate goal is to learn something about the enzyme. As William H. Stein would ask when discussing research results presented to him, "What did you learn about Nature?" There have been many reviews on affinity labeling, including a volume in this series. 1 Here we will discuss some of the uses of active-sitedirected reagents and consider the design and evaluation of such reagents for the benefit of those investigators who may wish to use this approach in a critical manner. Since "suicide" or "kcat" inactivators have been recently reviewed, we will not discuss them even though study of their chemical transformations also leads to mechanistic information.2,a "Photoaffinity" labeling has also been reviewed. 4-6 1 This series, Vol. 46. 2 N. Seiler, M. J. Jung, and J. K o c h - W e s e r , " E n z y m e - A c t i v a t e d Irreversible Inhibitors." Elsevier, A m s t e r d a m , 1978. a C. Walsh, Horiz. Biochem. Biophys. 3, 36 (1977). 4 H. Bayley and J. R. K n o w l e s , this series, Vol. 46, Article [8]. V. C h o w d h r y and F. H. W e s t h e i m e r , Annu. Rev. Biochem. 48, 293 (1979). 6 W. F. Benisek, J. R. Ogez, and S. B. Smith, Adv. Chem. Set. 198, 267 (1982).
METHODS IN ENZYMOLOGY, VOL. 87
Copyright © 19~2by AcademicPress, Inc. All rights of reproduction in any form reserved. ISBN 0-12-181987-6
470
INITIAL RATE AND INHIBITOR METHODS
[25]
U s e s of Affinity L a b e l i n g
Identification o f E s s e n t i a l A m i n o Acid R e s i d u e s In the characterization o f an e n z y m e , it is interesting to try to identify amino acid residues that are at the active site and possibly involved in catalytic activity. An investigator may attempt to design and synthesize a new active-site-directed reagent. In my judgment, a more rewarding approach is to evaluate simple chemical reagents first. This avoids the delay of perhaps 1 man-year for the synthesis o f a well-designed active-site-directed reagent and the disappointment if the compound is inactive. Furthermore, the methodology for using simple rea g e n t s - - e . g . , bromoacetic a c i d - - i s well established. 7's Finally, such studies are useful for determining what kinds o f functional groups to place on an active-site-directed reagent and are essential for determining whether the reagent reacts in a facilitated manner, as it should. Of course, it usually turns out that a simple reagent modifies more than one amino acid residue o f the e n z y m e during complete inactivation, and the investigator may consider several possible explanations: (a) only one residue is " e s s e n t i a l " for catalysis, (b) modification of several residues in the active site interferes with substrate binding, or (c) modification o f several residues denatures the enzyme. (In this context, an " e s s e n t i a l " residue is one that cannot be modified without " c o m pletely" inactivating the e n z y m e - - e . g . , to less than 1% residual activity. This is probably still more activity than the e n z y m e would have if a residue directly involved in catalysis were modified. Thus, alcohol dehydrogenase c a r b o x y m e t h y l a t e d on Cys-46 with about 2% residual activity, 9 or liver alcohol dehydrogenase with a phosphopyridoxyl group on Lys-228 and 10% activity, 1° or ribonuclease c a r b o x y m e t h y l a t e d on His-12 with about 5% activity 11 are partially active and apparently not modified on " e s s e n t i a l " residues, whereas chymotrypsin methylated on His-57 with about 0.05% activity TM may be said to be modified on an essential residue.) In order to distinguish among the possibilities, "differential labeling ''~3 may be used, where the active site is first protected with the bound ligand while the e n z y m e is modified with one reagent, and then the ligand is r e m o v e d and the e n z y m e is modified at the active site with another reagent. 7 G. E. Means and R. E. Feeney, "Chemical Modificationof Proteins." Holden-Day, San Francisco, California, 1971. s This series, Vol. 11. 9 C. H. Reynolds and J. S. McKinley-McKee, Arch. Biochem. Biophys. 168, 145 (1975). lo D. C. Sogin and B. V. Plapp, J. Biol. Chem. 250, 205 (1975). 11E. Machuga and M. H. Klapper, Biochim. Biophys. Acta 481, 526 (1977). 12R. Henderson, Biochem. J. 124, 13 (1971). 13A. T. Phillips, this series, Vol. 46, Article [7].
[25]
AFFINITY LABELING
471
In favorable cases, a simple reagent may modify just one amino acid residue, if the environment of the residue makes it especially reactive or if the "simple" reagent binds and reacts like an active-sitedirected reagent. For instance, the reactions of liver alcohol dehydrogenase 14 or pancreatic ribonuclease 15 with iodoacetate seem to fit the criteria normally required for affinity labeling (see later), apparently because of ionic interactions with the enzyme. Whenever a reagent, by whatever experimental method, is found to give essentially complete inactivation with nearly stoichiometric modification, the modified amino acid residue should be identified by protein sequence analysis. Amino acid analysis is not sufficient to conclude that a particular amino acid is involved in binding some part of a ligand. When the three-dimensional structure of the enzyme is determined, assignment of a function and comparison of results from solution and crystal studies become possible if the amino acid is firmly identified. Without the identification, one can only speculate. Active-site-directed reagents are useful in identifying amino acid residues in active sites, and many examples are compiled in Table I. Topography of Active Site With respect to the goal of "mapping" an active site, it is clear that X-ray crystallography provides more detailed (and higher resolution) three-dimensional information about more amino acid residues than affinity labeling ever will. The problem with an active-site-directed reagent is that one does not k n o w for sure how it binds to the enzyme while it is modifying a residue; affinity labeling is inherently a "low-resolution" approach. Thus an enzymologist may find that trying to crystallize an enzyme will be more rewarding than trying to make specific reagents. Although it has been argued that the structure in the crystal is "static" and not necessarily the same as the "dynamic" structure in solution, I a m not aware of any evidence from chemical modification studies that establishes a different structure in solution. Of course enzyme structures are dynamic, but other tools, such as N M R , are more appropriate for describing such dynamics. Nevertheless, when the three-dimensional structure of the enzyme is not known, mapping of active sites is useful. It is interesting, in fact, that in 1963 it was possible to describe the structure of the active site of chymotrypsin by use of three different affinity-labeling agents that placed functional groups in different regions of the substrate binding 14 C. H. Reynolds, D. L. Morris, and J. S. McKinley-McKee, Eur. J. Biochem. 14, 14 (1970). 15 B. V. Plapp, J. Biol. Chem. 248, 4896 (1973).
TABLE I IDENTIFICATION OF AMINO ACID RESIDUES IN ENZYMES MODIFIED BY AFFINITY LABELING 1. Chymotrypsin, bovine pancreas a. Diisopropylphosphorofluoridate: Ser-195 a p H 7, 25 °, 45 M -1 sec -lb b. Tosyl-L-phenylalanylchloromethane: His-57 c p H 7, 25 °, 7.7 M -1 sec -la c. p-Nitrophenylbromoacetyi-a-aminoisobutyrate: Met-192 e d. Photolysis of diazoacetyl-Ser-195 led to O-carboxymethyl-Tyr-146 of another molecule of enzyme f 2. Trypsin, bovine pancreas a. ct-N-Tosyl-L-lysylchloromethane: His-46~ p H 7, 25 ° K1 = 0.21 mM, ka = 2.6 x 10 -a sec -1 b. p-Guanidinophenacyl bromide: Ser-183 h p H 7.1, 25 °, KI = 1.63 mM, ka = 1.3 × 10-a sec -a 3. Trypsinogen, bovine pancreas Diisopropylphosphorofluoridate: Set-183 ~ p H 7, 25 °, ka/Ki = 6.8 x 10-4 M -1 sec -1 (Note: trypsin is inactivated with k = 5.0 M -1 sec -~) 4. T h r o m b i n a-N-Tosyi-L-lysylchloromethane: His-43 j p H 7.5, KI = 2.3 mM, k a = 9 x 10-4 sec -1 5. Thermolysin N-Chloroacetyl-D,L-N-hydroxyleucine methyl ester: Glu-143 k p H 7.2, 25 ° , KI = 7.5 mM, ka = 7.5 x 10-a sec -~ 6. Carboxypeptidase-Av L~u, bovine pancreas N-Bromoacetyl-N-methyl-L-phenylalanine: Glu-270 ~ (Side reactions: a - N - A s p - l , His-13) p H 7.5, 25 ° , KI = 4.8 raM, ka = 3.2 x 10-a s -1 7. Carboxypeptidase B, bovine pancreas a. c~-N-Bromoacetyl- D-arginine: Thr-Phe-Glu*-Leu-Arg-Asp-Lys-Gly-Arg-Tyr-Gly-Phem (Homologous to Glu-270 in carboxypeptidase A) b. 4-(Bromoacetamido)butylguanidine: Thr-Ile-Tyr*-Pro-Ala-Ser-Gly-Gly-Ser-Asp-Asp-Trp" (Homologous to Tyr-248 in carboxypeptidase A) p H 8, Inactivation 15 times faster than with bromoacetamide 8. o-Alanine carboxypeptidase, Bacillus subtilis Penicillin G: Ser-36 (Leu-Pro-Ile-Ala-Ser*-Met)° (Homologous to Set-44 in/3-1actamase) p H 7.5, 25 °, K1 large, k3/K1 = 440 M -1 sec -lp 9. fl-Lactamase, B. cereus 6fl-Bromopenicillanic a c i d - - " s u i c i d e substrate": Phe-Ala-Phe-Ala-Ser*-Thr-Tyr-Lys (Set-44 = Ser-70) q.... p H 7.5, 30 °, KI > 2 mM, ka/Ki = 1.8 × 104 M - ' sec -1 10. Staphylococcal nuclease t Deoxythymidine derivatives, B r A c N H = bromoacetamido, Ph = p-phenyl, P = phosphate or phosphodiester I. 3'-BrAcNHPhP, 5'-P 1 m o l / m o l incorporated Ki ~ 1 I.tM, ka/Kl = 0.3 M -~ sec -~ 0.36 Lys-48, 0.40 Lys-49 pH 9.4 0.15 T y r - l l 5 472
TABLE I
(Continued) 2 mol/n)o! incorporated 0.8 Tyr-85, but e n z y m e active with higher Km 3 m o l / m o i incorporated including Lys-24 (or Met-26?) 0.9 Tyr-115 Tyr-85 0.5 His-46 0.5 Trp-140 (not in active site)
II. 5 ' - B r A c N H P h P K~ ~ 2 m M
III. 3 ' - B r A c N H P h P K~ ~ 20 m M , b o u n d to subsite? IV. 3'-N~PhP, 5'-P V. 5'-N2PhP 3'-N2PhP b o u n d to subsite 11. Ribonuclease, bovine pancreas a. 5'-(p-Diazophenylphosphoryi)uridine 2'(3')-phosphate :Tyr-73 ~ Not at active site b. 2'(3')-O-Bromoacetyluridine: His-12" p H 5.5, 30 ° , KI = 87 m M , k3 = 3.5 x 10 -3 sec -t
~VI.
12. Carbonic a n h y d r a s e , erythrocyte w Bromoacetazolamide: H u m a n type C, His-64 Bovine type B, His-64 N-Bromoacetylacetazolamide: H u m a n type B, His-67 13. L y s o z y m e , chicken eggwhite 2 ' , 3 ' - E p o x y p r o p y l E-glycoside of di(N-acetyl)-D-glucosamine: Asp-52 ~ 14. Sucrase-Isomaltase, rabbit small intestine ~'~ 1-D-1,2-anhydro-myo-inositol (active isomer of conduritol-fl-epoxide) Sucrase: Iie-Asp*-Met-Asn-Gln-Pro-Asn-Ser-Ser p H 6.8, 37°, 0.027 M -~ sec -l Isomaltase: Ile-Asp*-Met p H 6.8, 37 °, 0.21 M -~ sec -~ 15. fl-Galactosidase, E. coli ~-D-Galactopyranosylmethyl-p-nitrophenyltriazene: Met-500 aa'b~ p H 7, 25 °, Kl = 0.48 m M , k3 = 9.8 x 10 -3 sec -~ 16. Galactosyltransferase, bovine colostrum Uridine 5'-diphosphate cleaved by periodate and irreversibly attached to e n z y m e with borohydride reduction: Ser-Gly-Lys *cc 17. Triose p h o s p h a t e isomerase, rabbit muscle 3-Chloroacetolphosphate: Glu-16 5aa'ee p H 6.5, 2 °, 2.3 × 10a M -~ sec -1 18. G l u c o s e p h o s p h a t e isomerase, rabbit muscle N - B r o m o a c e t y l e t h a n o l a m i n e phosphate: Val-Leu-His*-Ala-Glu-Asn-Val-Aspit p H 8, 37 ° , KI = 0.056 m M , k3 = 1.8 × 10-3 sec -~ 19. Phosphoglycerate m u t a s e , rabbit muscle N - B r o m o a c e t y l e t h a n o l a m i n e phosphate: ( T r p , L y s , C y s * , A s p , S e r , G l u 2 ,Gly,Ala,Leu,Phe~) og p H 7, r o o m temp., KI = 0.32 m M , k3 = 6.8 × 10 -3 sec -1 20. Ribulosebisphosphate carboxylase, spinach N - B r o m o a c e t y l e t h a n o l a m i n e phosphatehn: Tyr-Gly-Arg-Pro-Leu-Gly-Cys*-Thr-Ile-Lys*-Pro-Lys Trp-Ser-Pro-Glu-Leu-Ala-Ala-Ala-Cys*-Glu-Val-Trp-Lys pH 8, 30 ° +5 m M Mg z+, KI : 3.0 raM, ka = 4.8 × 10 -4 sec -~ L y s p H 8, 30 °, no Mg 2+, /(i = 0.8 raM, k~ = 8.4 x 10-5 sec -1 2 C y s
(continued) 473
TABLE I
(Continued)
21. Aldolase, rabbit muscle N-Bromoacetylethanolamine phosphate ~ p H 8.5, room temp., KI = 0.76 mM, ka = 1.4 × 10-4 sec -1, Lys-146 p H 6.5, room temp., K1 = 0.87 mM, k3 = 3.3 x 10 -5 sec -~, His-359 22. 2-Keto-3-deoxygluconate-6-phosphate aldolase, Pseudomonas putida Bromopyruvate: Glu-56 jj,kk pH 6, 24.5 °, K~ = 1 mM, kz = 0.011 sec -~ 23. Aspartate aminotransferase, pig heart cytoplasmic and mitochondrial /3-Chloro-L-alanine (suicide inactivator): Lys-258 tz (same lysine as binds pyridoxal phosphate) pH 7.4, 18°, 3.2 M formate, KI = 0.2 M, ka = 0.33 sec -~ 24. Aspartate aminotransferase, pig heart cytoplasmic only 4'-N-(2,4-dinitro-5-fluorophenyl)pyridoxamine-5'-phosphate: Lys-258 m'n 25. Tryptophan synthase, f12 subunit, E. coli Bromoacetylpyridoxamine phosphate: Cys-230 n" (Or Cys-62 if Cys-230 is first blocked) pH 8, 37 °, K~ = 0.6 mM, ka = 0.0023 sec -~ 26. Formylglycinamide ribonucleotide amidotransferase, Salmonella typhimurium L-azaserine: Ala-Leu-Gly-Val-Cys *°° 27. Anthranilate synthetase component II, Serratia marcescens L-(aS,5S)-aAmino-3-chioro-4,5-dihydro-5-isoxazoleacetic acid: Cys-83 pp p H 5.3, room temp., 0.5 m M chorismate, KI = 0.14 mM, ka = 0.052 sec -~ 28. Thymidylate synthetase, Lactobacillus casei 5-Fluorodeoxyuridine 5'-phosphate: Ala- Leu-Pro-Pro-Cys* -His-Thr-Leu-Tyr ~ Labeling by iodoacetate in presence o f 5,10-methylene tetrahydrofolate and F d U M P protects CySH 29. L-Isoleucine tRNA ligase, E. coli L-Isoleucylbromomethane: Ile-Glu-Ser-Met-Val-Ala-Asp-Arg-Pro-Asn-Trp-Cys*-Ile-Ser-Argrr pH 7.5, 27 °, Ko = 0.7 mM, ka = 1.8 × 10-3 sec -~ 30. Phosphofructokinase, sheep heart p-Fluorosulfonylbenzoyl-5 '-adenosine: Asn-Phe-Ala-Thr-Lys*-Met-Gly-Ala-Lys, allosteric site s~ 31. Phosphofructokinase, rabbit muscle N~-(6-Bromoacetamidohexyl)-AMP-PCP: Cys*-Lys-Asp-Phe-Arg~, ATP site 32. Mitochondrial F~-ATPase, bovine p-Fluorosulfonylbenzoyl-5'-adenosine:/3 subunit, Ile-Met-Asp-Pro-Asn-Ile-Val-Gly-Ser-Glu-His-Tyr*-Asp-Val-Ala-Arguu p H 7, 20% glycerol, ka/[I] = 0.09 M -1 sec -~ 33. cAMP-dependent protein kinase II, porcine skeletal muscle p-Fluorosulfonylbenzoyl-5'-adenosine: catalytic subunit Glx-Ile(Asp2,Met, Thr,Ser,Glu,Gly,Leu,Phe,His,Lys*) w p H 7, 37 °, 10% glycerol, KI = 57/xM, ka = 6 × 10-4 sec -~ 34. cAMP-dependent protein kinase II, porcine heart 8-Azidoadenosine 3' : 5'-monophosphate: regulatory subunit Lys-Arg-Asn-Ile- Ser-His-Tyr*-Glu-Glu-Gln-Leu-Val-Lys-Met~ Photoaffinity labeling, Ko about 70 nM, 0.5 mol/mol
474
inhibitory
TABLE I
(Continued)
35. AS-3-Ketosteroid isomerase, Pseudomonas testosteroni 3-Oxo-4-estren- 17/3-yl acetate Photoinactivation: Asp-38 ~ Ala-38 x/ 36. G l u t a m a t e d e h y d r o g e n a s e , bovine liver Iodoacetyldiethylstilbestrol: Cys-89, allosteric site uu p H 7.6, 37 °, 1 m M N A D H , KI = 10/xM, ka = 0.0021 sec -1 37. Aspartate-fl-semialdehyde d e h y d r o g e n a s e L-2-Amino-4-oxo-5-chloropentanoic acid: Phe-Val-Gly-Gly-Asp-His*-Thr-Val-Ser~z p H 7.2, 0 °, Kz = 0.2 m M , ka = 5.1 × 10 -3 sec -1 38. Alcohol d e h y d r o g e n a s e a. Nicotinamide-5-bromoacetyl-4-methylimidazole dinucleotide H o r s e Liver: Cys-174 (Zn ligand in nicotinamide site!) aaa p H 6.5, 37 °, Kx = 0.7 m M , ka = 1.6 × 10-3 sec -~ Yeast: Cys-43 (Zn ligand homologous to Cys-46 in horse e n z y m e , in nicotinamide site) b~ b. 4-(3-Bromoacetylpyridino)butyl d i p h o s p h o a d e n o s i n e ccc H o r s e liver: Cys-46 p H 6.2, 25 ° , KI = 1 m M , ka = 1.7 x 10-3 sec -a Yeast: Cys-43 p H 6.6, 25 °, KI = 4 m M , k3 = 2.8 × 10-z sec -~ R. A. O o s t e r b a a n , P. K u n s t , J. van Rotterdam, and J. A. C o h e n , Biochim. Biophys.
Acta 27, 549, 556 (1958). b D. E. F a h r n e y and A. M. Gold, J. Am. Chem. Soc. 85, 997 (1963); or A. M. Gold, this series, Vol. 11, Article [83]. c E. B. Ong, E. Shaw, and G. Schoellmann, J. Biol. Chem. 240, 694 (1965). d E. Shaw and J. Ruscica, Arch. Biochem. Biophys. 145, 484 (1971). e W. B. L a w s o n and H.-J. S c h r a m m , Biochemistry 4, 377 (1%5). t C . S. Hexter and F. H. W e s t h e i m e r , J. Biol. Chem. 246, 3928 (1971). E. Shaw and G. Glover, Arch. Biochem. Biophys. 139, 298 (1970). h D. D. Schroeder and E. Shaw, Arch. Biochem. Biophys. 142, 340 (1971). i p. H. Morgan, N. C. Robinson, K. A. Walsh, and H. Neurath, Broc. Natl. Acad. Sci. U.S.A. 69, 3312 (1972). J G. Glover and E. Shaw, J. Biol. Chem. 246, 4594 (1971). k D. Rasnick and J. C. Powers, Biochemistry 17, 4363 (1978). t G. M. H a s s and H. Neurath, Biochemistry 10, 3535, 3541 (1971). " M. T. K i m m e l and T. H. Plummer, Jr., J. Biol. Chem. 247, 7864 (1972). n T. H. Plummer, Jr., J. Biol. Chem. 244, 5246 (1969). o D. J. W a x m a n and J. L. Strominger, J. Biol. Chem. 255, 3964 (1980). P P. M. Blumberg and J. L. Strorninger, Proc. Natl. Acad. Sci. U.S.A. 68, 2814 (1971). q M. J. L o o s e m o r e , S. A. C o h e n , and R. F. Pratt, Biochemistry 19, 3990 (1980). S. A. C o h e n and R. F. Pratt, Biochemistry 19, 3996 (1980). V. K n o t t - H u n z i k e r , S, G. Waley, B. S. Orlek, and P. G. S a m m e s , FEBS Lett. 99, 59 (1979). t p. C u a t r e c a s a s and M. Wilchek, this series, Vol. 46, Article [38], and references therein. u M. Gorecki, M. Wilchek, and A. Patchornik, Biochim. Biophys. Acta 229, 590 (1971). v M. Pincus, L. L. Thi, and R. P. Carty, Biochemistry 14, 3653 (1975). u, D. L. C y b u l s k y , S. I. Kandel, M. Kandel, and A. G. Gornall, J. Biol. Chem. 248, 3411 (1973). x y . E s h d a t , J. F. M c K e l v y , and N. Sharon, J. Biol. Chem. 248, 5892 (1973). A. Quaroni and G. Semenza, J. Biol. Chem. 251, 3250 (1976).
476
INITIAL RATE AND INHIBITOR METHODS TABLE I
[25]
(Continued)
H. Braun, G. Legler, J. Deshusses, and G. Semenza, Biochim. Biophys. Acta 483, 135 (1977). aa A. V. Fowler, I. Zabin, M. L. Sinnott, and P. J. Smith, J. Biol. Chem. 253, 5283 (1978). 0b M. L. Sinnott and P. J. Smith, Biochem. J. 175, 525 (1978). cc j. T. Powell and K. Brew, Biochemistry 15, 3499 (1976). aa F. C. Hartman, Biochemistry 10, 146 (1971). ee F. C. Hartman, this series, Vol. 46, Article [10]. rrD. R. Gibson, R. W. Gracy, and F. C. Hartman, J. Biol. Chem. 255, 9369 (1980). go F. C. Hartman and I. L. Norton, J. Biol. Chem. 251, 4565 (1976). ha j. V. Schloss, C. D. Stringer, and F. C. Hartman, J. Biol. Chem. 253, 5707 (1978). " F. C. Hartman and J. P. Brown, J. Biol. Chem. 251, 3057 (1976), JJ H. P. Meloche, C. T. Monti, and R. A. Hogue-Angeletti, Biochem. Biophys. Res. Commun. 84, 589 (1978). kk N. Suzuki and W. A. Wood, J. Biol. Chem. 255, 3427 (1980). u y . Morino, A. M. Osman, and M. Okamoto, J. Biol. Chem. 249, 6684 (1974). mmF. Riva, D. Carotti, D. Barra, A. Giartosio and C. Turano, J. Biol. Chem. 255, 9230 (1980). nn E. W. Miles and W. Higgins, Biochem. Biophys. Res. Commun. 93, 1152 (1980); W. Higgins, E. W. Miles, and T. Fairwell, J. Biol. Chem. 255, 512 (1980). o o T. C. French, I. B. Dawid, and J. M. Buchanan, J. Biol. Chem. 238, 2186 (1963). PP J. Y. Tso, S. G. Bower, and H. Zalkin, J. Biol. Chem. 255, 6734 (1980). qq R. L. Bellisario, G. F. Maley, J. H. Galivan, and F. Maley, Proe. Natl. Aead. Sci. U.S.A. 73, 1848 (1976). rr p. Rainey, E. Holler, and M.-R. Kula, Eur. 7. Biochem, 63, 419 (1976). 8s L. Weng, R. L. Heinrikson, and T. E. Mansour, J. Biol. Chem. 255, 1492 (1980). tt K. Nagata, K. Suzuki, and K. Imahori, J. Biochem. (Tokyo) 86, 1179 (1979). uu F. S. Esch and W. S. Allison, J. Biol. Chem. 253, 6100 (1978); 254, 10740 (1979). w M. J. Zoller and S, S. Taylor, J. Biol. Chem. 254, 8363 (1979). wwA. R. Kerlavage and S. S. Taylor, J. Biol. Chem. 255, 8483 (1980). ~ J . R. Ogez, W. F. Tivol, and W. F. Benisek, J. Biol. Chem. 252, 6151 (1977). ~u F. Michel, M. Pons, B. Descomps, and A. Crastes de Paulet, Eur. J. Biochem. 84, 267 (1978). zz J.-F. Biellmann, P. Eid, C. Hirth, and H. J6rnvall, Eur. J. Biochem. 104, 59 (1980). a~a C. Woenckhaus and R. Jeck, Hoppe-Seyler's Z. Physiol. Chem. 352, 1417 (1971). booH. Jfrnvall, C. Woenckhaus, and G. Johnscher, Eur. J. Biochem. 53, 71 (1975). c~ C. Woenckhaus, R. Jeck, and H. J6rnvall, Eur. J. Biochem. 93, 65 (1979).
pocket. ~6 As shown in Fig. 1, Met-192, Ser-195, and His-57 could be identified. A little model building produced a picture of the active site that fits the three-dimensional structure as determined later by X-ray crystallography of chymotrypsin and some of its inhibited forms. ~7,18 ~ B. R. Baker, "Design of Active-Site-Directed Irreversible Enzyme Inhibitors." Wiley, New York, 1967. ~r D. M. Blow, in "The Enzymes" (P. D. Boyer, ed,), 3rd ed., Vol. 3, p. 185. Academic Press, New York, 1971. ~s D. M. Segal, J. C. Powers, G. H. Cohen, D. R. Davies, and P. E. Wilcox, Biochemistry 10, 3728 (1971).
[25]
AFFINITY LABELING
477
O
H
Ser-195 Met-192
C--R 3
L
His-57
R~CONH--C--H R2
H
O
I
tl
Ser-195--O
O
[[
P--F
Ser-195--O--C
I
0/~'0
Met-192--S
I
I
CH3
CH3
I
CH~CH CH3CH
CH~
CH2CONHC--CH3
q
Br
I
CH3
His-57 C1 O CH3--~,,
~_._2/
C--CH2
N~ . . . N - - H
/)--SNH--C--H
II
0
I
CH2
FIc. 1. Topography of active site of chymotrypsin as determined by sequence analysis of protein labeled with the indicated reagents. See Table I, item 1, for references.
In contrast, the careful studies of Cuatrecasas and co-workers on the affinity labeling of staphylococcal nuclease gave a picture of the active site that was quite different from the picture determined independently by X-ray crystallography (Fig. 2). Although the affinity labeling allowed the identification of several amino acid residues that are near the active site, only one of these (Tyr-85) is in contact with the parent inhibitor, deoxythymidine 3', 5'-diphosphate, as determined in the crystal. Interestingly, two residues in the active site (Lys-84 and Tyr-ll3) did not appear to react. Furthermore, two labeled residues (Lys-24, Trp-140) were so far from the "active site" that it was possible that the reagents reacted in a subsite. More importantly, it was concluded from the labeling studies that Tyr-85 was near the 5'-phosphate and possibly involved in catalytic action. But the crystallography places that residue near the 3'-phosphate and Tyr-ll3 near the probable cleavage site on the 5'-phosphate. It may be concluded that the flexibility of the reagents or the protein led to many of the apparent differences in active site topography, but the lesson to be learned is that affinity-labeling results need to be interpreted cautiously. An extensive series of steroid affinity labels have also been used to
478
INITIAL RATE AND INHIBITOR METHODS
[25]
Affinity Labeling T ,|
~
Tyr-85 II,V ,---- ~ ( ~\~---PJ 5'
~
I
III Lys-24 or Met-26? Lys-48, Lys-49 VI His-46, Trp-140
X-ray Crystallography T Arg-35 Arg-87 / ~ Tyr-85
® CaZ÷
O3PO~,/~ Tyr- 113
OPO3 (~) Lys-84
FIG. 2. Topography of active site of staphyloccal nuclease as derived by affinity labeling studies (see Table I, item 10, for reagents and references) or by X-ray crystallography [A.
Arnone, C. J. Bier, F. A. Cotton, V. W. Day, E. E. Hazen, Jr., D. C. Richardson, J. S. Richardson, and A. Yonath, J. Biol. Chem. 246, 2302 (1971)].
study the topography of the active site of human placental 17/3-dehydrogenase, ~aa° and 3a,20/3-hydroxysteroid dehydrogenase, 6 and a number of different amino acid residues have been modified. Since the reactive functional group is bromoacetyl, which has the flexibility to move several angstroms, the exact positioning of the amino acids is difficult. 21 It will be interesting to determine the structures by X-ray crystallography. I n v e s t i g a t i n g Catalytic M e c h a n i s m s
The identification of "essential" amino acid residues opens the way for speculation a b o u t the roles o f the amino acid side chains in the binding o f substrates and in the catalytic m e c h a n i s m . Unfortunately, modification o f a side chain in the active site with a bulky (or e v e n a small) group can inactivate for steric reasons, and thus it is difficult to determine if the modified group directly participates in catalysis. Despite the ambiguity, affinity labeling can yield s o m e information. Once the reactive amino acids h a v e been identified, the active-sitedirected reagent can be used to determine h o w environmental factors 19 M. Pons, J.-C. Nicolas, A.-M. Boussioux, B. Descomps, and A. Crastes de Paulet, Eur. J. Biochem. 68, 385 (1976). 20 C.-C. Chin, P. Asmar, and J. C. Warren, J. Biol. Chem. 255, 3660 (1980). 21j. C. Warren and J. R. Mueller, this series, Vol. 46, Article [50].
[25]
AFFINITY LABELING
479
affect reactivity. Thus the pH dependence of modification can be used to give pK values for the reactive amino acid, or at least for the system of which it is a part. (The ionization of another nonmodifiable residue can affect the binding of the reagent, the state of the residue to be modified, or the structure of the enzyme.) For instance, the alkylation of His-57 in chymotrypsin by tosyl-L-phenylalanylchloromethane depends on a basic group with a pK of 6.8 and an acidic group with a pK of 8.9. The pH dependencies for catalytic action show similar pK values, and it was concluded that His-57 has the pK of 6.8. 22 But note that acylation of Ser-195 by phenylmethanesulfonyl fluoride also depends on a group with a pK of 7. 23 The use of very well-designed active-site-directed reagents that closely resemble a substrate, bind tightly to the enzyme, and react with considerable facilitation may also lead one to attempt to position the modifiable residue near a part of the substrate and to assign a catalytic role. The first compelling example of this was the use of tosyl-Lphenylalanylchloromethane to modify His-57. 24 One could then imagine that the imidazole group could participate as an acid-base catalyst for the scission of the amide bond of a substrate. The identification of Ser195 as the site of acylation by diisopropylphosphorofluoridatez5 or phenylmethanesulfonyl fluoridez6 is less compelling because of the uncertainty of the mode of binding of these simpler reagents and the possibility that the acyl group might have been transferred from another group (e.g., the histidine) that was initially acylated. This discussion illustrates the value of using an active-site-directed reagent that is isosteric with a substrate. A reagent that closely resembles a substrate should produce more facilitation of the reaction and give one more confidence in assigning a catalytic function to the modifiable residue. Tosyl-L-phenylalanylchloromethane is almost isosteric with a substrate. If we assume that one of the structures on the catalytic pathway is A shown below (the imidazole is about to accept a proton from a serine hydroxyl), then B might be the ground-state structure leading to attack by the imidazole on the --CH2CI. However, it appears to me that the imidazole might be as much as 2/~ away (the diameter of a hydrogen atom) from its optimum position for alkylation. Perhaps the flexibility of the enzyme or reagent is sufficiently high so that reactivity is close to maximal. But then one won22 F. J. K6zdy, A. Thomson, and M. L. Bender, J. Am. Chem. Soc. 89, 1004 (1967). 23 A. M. Gold and D. Fahrney, Biochemistry 3, 783 (1964). 24 E. B. Ong, E. Shaw, and G. Schoellmann, J. Biol. Chem. 240, 694 (1965). 25 R. A. Oosterbaan, P. Kunst, J. van Rotterdam, and J. A. Cohen, Biochim. Biophys. Acta 27, 549, 556 (1958). 26 A. M. Gold, Biochemistry 4, 897 (1965),
480
I N I T I A L RATE A N D INHIBITOR M E T H O D S Ser-195 o--n
His-57 ..... N ~ / N - - H
Ser-195
[25]
His-57
O - - H ..... N ~ / N - - H H
r
R--C--N--R'
II
O
R--C--C--C1
I
Ir
H
O A
r
H B
ders why the serine hydroxyl is not alkylated significantly. A possible explanation is that the reagent forms a hemiketal with the serine before it alkylates the histidine, as suggested on the basis of high-resolution Xray crystallography of subtilisin inactivated with halomethyl ketones. 27 Nevertheless, benzyloxycarbonyl-L-phenylalanyl-L-alanylchloromethane reacts with a cysteine (Cys-25) in papain rather than with the histidine. 28 Of course, a sulfhydryl is more reactive than a hydroxyl group, but tosyl-L-phenylalanylchloromethane reacts with papain 100 times faster than it reacts with chymotrypsin29 or 2000 times faster than it reacts with cysteine) ° Other chloromethylketones, tosyl-L-lysylchloromethane, 3° and benzyloxycarbonyl-L-phenylalanylchloromethane 31 react even faster with papain. Furthermore, trypsin reacts with p-guanidinophenacyl bromide to form an ether linkage with the active-site serine. 32 Other examples have been summarized previously. 3~ Thus, one must conclude that chloromethylketone reagents are not completely specific. It would be interesting to design other functional groups for these active-site-directed reagents that are closer to being isosteric. In this regard, it is impressive that peptidyl diazomethylketones are highly specific and greatly facilitated (101°-fold) in their reactions with thiol proteinases.34 It may be noted that suicide, or mechanism-based, inactivators probably must be close to being isosteric with a substrate, since the enzyme presumably acts on the reagent to convert it to a reactive species. Even these species often seem to have a poor shape or be the width of one hydrogen atom too far away, however. To reiterate, if we wish to understand enzyme mechanisms, in par27 T. L. Poulos, R. A. Alden, S. T. Freer, J. J. Birktoft, and J. Kraut, J. Biol. Chem. 251, 1097 (1976). zs j. Drenth, K. H. Kalk, and H. M. Swen, Biochemistry 15, 3731 (1976). z9 M. L. Bender and L. J. Brubacher, J. Am. Chem. Soc. 88, 5880 (1966). a0 j. R. Whitaker and J. Perez-Villasefior, Arch. Biochem. Biophys. 124, 70 (1968). al R. Leary, D. Larsen, H. Watanabe, and E. Shaw, Biochemistry 16, 5857 (1977). 32 D. D. Schroeder and E. Shaw, Arch. Biochem. Biophys. 142, 340 (1971). 3a E. Shaw, this series, Vol. 11, Article [80]. 34 G. D. J. Green and E. Shaw, J. Biol. Chem. 256, 1923 (1981).
[25]
AFFINITY LABELING
481
ticular how enzymes recognize and bind substrates, the design of active-site-directed reagents should be relatively sophisticated. Affinity labeling can also be used to analyze the "inherent reactivity" of amino acid side chains, which could be a source of catalytic power for enzymes. Of course, one must separate the inherent reactivity from the reactivity due to the "circe" effect. 35 This will be discussed later under the heading o f " Facilitation." Considerations in the Design of Active-Site-Directed Reagents Specificity What characteristics allow a reagent to react specifically and rapidly? The affinity group is very important in this respect, and one should examine the structures and affinities of substrates and inhibitors to find out (a) groups that are essential for the binding, (b) tolerance for adding or removing groups, and (c) groups that might be on the enzyme and could be modified. It is often assumed that the affinity group should resemble a substrate, product, or inhibitor. But remember that "resemblance" is in the eye of the biochemist, whereas the enzyme :apparently uses more sophisticated criteria to determine how well it binds a ligand. (We cannot easily predict how well an enzyme will bind some structure that we design.) Therefore, extensive empirical studies are required to determine the size, shape and functionality of the best ligands. Baker TM has emphasized that studies on "bulk tolerance" are prerequisites for successful design. But we should also look for the minimum structure that is required for good binding. Often, one can learn much about the active site by examining space-filling models of compounds that are known to bind to the enzyme. In evaluating binding, it is necessary to do proper kinetic studies to be able to decide which part of the active site a ligand binds to and what the dissociation constant is. It is not sufficient to determine an Is0 (concentration of inhibitor giving 50% inhibition in a particular assay). The type of inhibition against one of the substrates should be determined; if it is competitive, the inhibitor probably binds to the same site as the substrate and the Ki is the dissociation constant. If the inhibition is noncompetitive or uncompetitive, different sites may be involved and the apparent Ki usually should be corrected--e.g., for the effect of nonsaturation by the nonvaried substrate(s) or for the simplification of the kinetic equation. 36 With such kinetic constants in hand, it is possias w. P. Jencks, Adv. Enzymol. 43, 219 (1975). 36I. H. Segel, "Enzyme Kinetics." Wiley,New York, 1975.
482
[25]
INITIAL RATE AND INHIBITOR METHODS TABLE II ROLE OF REAGENT STRUCTURE IN DETERMINING RATE OF INACTIVATION OF ACETYLCHOLINESTERASEa
Structure no.
Reagent
KI
k3
ka/ K1 (M -1
Ki, inhib.
(mM)
(sec -1)
sec -1)
(mM) --
1
CH3SO~F
Large
--
2.5
2
CH3SO2--O,,~
0.4
8.3 x 10-4
2
®N(CH3)a 3
CH3SO2--O..~
0.053
®N(CH3)3 0.1
5 × 10-3
50
I
~/~
0.11
f
CH3
CHs
4
CHsSOz--O-..~.~
5
(9 CH3SO2--OCH2CH2N(CHa)3
No inactivation
6
CH3SO2--O..~
No inactivation
(CHa:)3NCH2
~,~
0.02
5.5 × 10-4
27
® N ( C H 3 ) 41.2
"a~
a Rates of inactivation are taken from R. Kitz and I. B. Wilson [J. Biol. Chem. 237, 3245 (1962)]. Inhibition constants are from R. Kitz and I. B. Wilson [J. Biol, Chem. 238, 745 (1963)], who also showed that quaternary ammonium salts can stimulate the rate of inactivation by CHaSO~F by up to 33-fold, and the stimulation has subsequently been studied by others [B. Belleau and V. DiTullio, J. Am. Chem. Soc. 92, 6320 (1970); M. R. Pavli~, Biochim. Biophys. Acta 327, 393 (1973)]. b l e to r a t i o n a l i z e b i n d i n g affinity in t e r m s o f s t r u c t u r e a n d to b e g i n to design an appropriate reagent. F o r e x t e n s i v e s e r i e s o f c o m p o u n d s , H a n s c h ' s c o r r e l a t i o n a n a l y s i s is r e q u i r e d . T h i s a p p r o a c h h a s b e e n a p p l i e d to t h e m o n u m e n t a l r e s u l t s from Baker's laboratory and shows, among other things, that related
[25]
AFFINITY LABELING
483
enzymes may have similar binding sites and that the specificity of the binding site must be explored with carefully chosen substituents? 7,3s Eventually an affinity group should be found that binds to the enzyme with a dissociation constant less than or equal to 1 mM (1/xM for some photoaffinity reagents), since the tighter the compound binds, the more facilitated the reaction should be. If the dissociation constant is 10 or 100 mM, the reagent may be quite nonspecific because it binds by simple ionic or hydrophobic interactions at various sites on the enzyme. The importance of the structure of the affinity group for obtaining specific modification directed towards particular amino acid residues was illustrated with the studies in Figs. 1 and 2. Furthermore, as shown in Table II, the location of the functional group on the affinity portion of the reagent is critical for determining whether a facilitated reaction is obtained. Although the affinity group for reagent 2 binds tightly, the compound does not inactivate acetylcholinesterase any faster than 1, which has no "affinity group". Compounds 3 and 4 have more suitable designs and react in a moderately facilitated manner. In contrast, 5 and 6 do not react measurably. Just how these compounds bind into the active site is not yet known. It is clear that simply attaching a functional group somewhere onto the affinity group is not good enough. If the group is attached in the wrong location, the compound could still bind to the enzyme, but then it could protect the enzyme against a bimolecular attack by another molecule of reagent. As discussed later, this mechanism of reaction can also result in saturation kinetics and can deceive the investigator about the nature of the reaction. An impressive example of the role of the affinity group in providing rate-enhancement facilitation and selectivity is the series of studies by Kettner and Shaw, illustrated in Table III. By varying the peptide structure, reagents that are selective enough to be used in vivo may be obtained. Covalent Chemistry In choosing the functional group to place onto the affinity group, the knowledge obtained from the use of simple reagents on the enzyme is valuable. A great variety of alkylating, acylating, photolabile groups, and other groups have been used as can be seen by perusal of Volume 46 of this series, on affinity labeling. 37c. Silipo and C. Hansch,J. Med. 3s M. YoshimoIoand C. Hansch,J.
Chem. 19, 62 (1976). Med. Chem. 19, 71 (1976).
484
INITIAL RATE AND INHIBITOR METHODS
[25]
TABLE III RATE ENHANCEMENT SPECIFICITY AND SELECTIVITY OF PEPT1DYL CHLOROMETHYL KETONES,a P-Arg-CH2CI 10-4 X kobs/[I] (M -~ min -l, 25°, pH 7.0)
P
Thrombin
Plasma Kallikrein
Piasmin
Urokinase
Val- lie-ProVal-ProIle-ProDns-Glu-GlyPhe-Alalle-LeuGlu-GlyPro-GlyAc-Gly-GlyAla-PhePro-Phe-
73 54 42 26 8 5.2 1.9 1.2 0.74 0.17 0.12
2.2 2.9 2.0 140 0.86 8.9 16 3.3 1.4 440 150
0.35 0.35 0.31 28 0.09 0.36 1.3 0.091 0.053 14 3.7
0.18 0.54 0.39 4.2 0.35 0.014 20 0.79 2.6 0.0059 0.0015
a The estimated bimolecular rate constants for inactivation, kobs/[I], were taken from several publications by C. Kettner and E. Shaw, Biochemistry 17, 4778 (1978); in "Chemistry and Biology of Thrombin" (R. L. Lundblad, K. G. Mann, and J. W. Fenton, eds.), p. 129. Ann Arbor Sci. Publ., Ann Arbor, Michigan, 1977; Biochim. Biophys. Acta 569, 31 (1979). The most reactive inactivator of thrombin is D-Phe-LPro-L-Arg-CH2CI, which has a kobs/[I] value of 6.8 x l0 s M -1 min -1 [C. Kettner and E. Shaw, Thromb. Res. 14, 969 (1979)].
Depending on the purpose of the reagent, one could choose either an "exo" or "endo" type of affinity labeling. 39 When Baker 16 originally chose to use "exo" affinity labeling, he postulated that enzymes from different sources (e.g., normal and cancerous tissues) would have the same amino acid residues at the active site, but might differ in nonessential residues outside the active site. However, even quite large and sophisticated exo-site reagents may inactivate a variety of homologous enzymes. 4° Furthermore, we now know that enzymes catalyzing the same reactions (e.g., serine proteases or NAD-dependent dehydrogenases) may have very similar tertiary structures, even with several differences in the amino acids within the active site. These differences frequently can be exploited by varying the chemistry and location of the functional group. Thus the impetus for using "exo" affinity labeling is somewhat reduced. Nevertheless, we have used such reagents in attempting to modify the only alkylat39 M. Cory, J. M. Andrews, and D. H. Bing, this series, Vol. 46, Article [9]. 40 D. J. Robinson, B. Furie, B. C. Furie, and D. H. Bing, J. Biol. Chem. 2,55, 2014 (1980).
[25]
AFFINITY LABELING
485
able residue in the substrate binding pocket of liver alcohol dehydrogenase. A methionine residue located 14/~ from the catalytic zinc ion was the target, and we found that reagents of just the right length, size, and shape would inactivate the enzyme in a facilitated manner. The reagents that were too short, too long, or too rigid reacted with less facilitation and presumably with less specificity. 41 One difficulty with using very long and flexible groups onto which is attached the reactive functional group is that one does not necessarily known how these groups are bound by the enzyme. The flexibility reduces the resolution with which one can map the active site. " E n d o " alkylators should be utilized if one wishes to identify amino acid residues that may participate in catalysis and also if one hopes to obtain catalysis by the enzyme of the chemical reaction. " E n d o " alkylators should be isosteric, if possible, with a natural substrate, and the functional group could replace some portion of the normal substrate. Suicide inactivators, or enzyme-facilitated or enzyme-catalyzed modifications, are examples of this type of " e n d o " modification. To illustrate the difficulties that one may encounter in trying to use isosteric reagents, the work by Woenckhaus and coworkers on liver alcohol dehydrogenase may be cited (Table I, item 38). NAD analogs in which either the nicotinamide ring or the adenine ring contain the reactive functional groups each modify cysteine residues that are ligands to zinc in the nicotinamide binding pocket. Obviously the enzyme finds it difficult to distinguish the two ends of the NAD analogs. The chemistry of the leaving group is important in designing reagents. We found that variation of the leaving group (CI, Br, I, tosyl) did not affect the relative facilitation of carboxymethylation of pancreatic ribonuclease as compared to reaction with a model compound, whereas with pancreatic deoxyribonuclease, smaller leaving groups produced more facilitation (Table IV). We attribute this to an interaction of deoxyribonuclease with the leaving group itself. 15 With tosyl-L-phenylalanylchloromethane analogs, the sulfonate esters appear to have enhanced reactivities with chymotrypsin (Table IV). One of the reasons for studying the reaction of ribonuclease with tosylglycolate (carboxymethyltosylate) was that we thought that the tosyl group would bind into the active site, but then be displaced when the carboxymethyl group reacted with the enzyme. In other words, the affinity group is itself the leaving group. This experimental design was also used by Nakagawa and Bender to methylate His-57 in chymotrypsin with ~1 W.-S. Chen and B. V. Plapp, Biochemistry 17, 4916 (1978).
486
INITIAL RATE AND INHIBITOR METHODS
[25]
TABLE IV E F F E C T OF LEAVING GROUP ON REACTIVITYa
X-CH2COOHb X I Br Tosyl CI Condition
Cbz-Phe-CH2Xc
NBPa Rel. rate
RNase M-1 sec -1
DNase M-1 sec -~
52 32 6.3 0.63
0.050 0.085 0.0083 0.0028 pH 5.5, 37°
0.014 0.116 0.00036 0.0085 pH 7.2, 25°, + 4 mM CuCI2
X
Chymotrypsin M-~ sec -1
Mesyl Br Tosyl CI
1800 790 7400 69 pH 6.8, 25°
a The rates of reaction are the pseudobimolecular rate constants, k J K i ; see "Kinetics" in next section. b B. V. Plapp, J. Biol. Chem. 248, 4896 (1973). c D. Larsen and E. Shaw, J. Med. Chem. 19, 1284 (1976). a Relative reactivities with 4-(p-nitrobenzyl)pyridine in 75% 2-methoxyethanol at pH 4.2 and 37°, using the procedure of B. R. Baker and J. H. Jordaan [J. Heterocycl. Chem. 2, 21 (1965)]. The numbers are expressed as the change in absorbance at 570 nm (1-cm path) per rain divided by the final molarity of the reagent in the reaction mixture. For comparison, other compounds were: BrCH~CH2OH, 0.3; CHaI, 45. An extensive series of aziridines and related compounds have been studied previously IT. J. Bardos, N. Datta-Gupta, P. Hebborn, and D. J. Triggle, J. Med. Chem. 8, 167 (1965), and references therein]. m e t h y l b e n z e n e s u l f o n a t e42 a n d b y W h i t e a n d B r a n c h i n i to e t h y l a t e leuciferase w i t h a n o t h e r s u l f o n a t e ester. 43 I n t h e s e c a s e s , the c h a n g e s in activity o f the e n z y m e are n o t d u e to the i r r e v e r s i b l e b i n d i n g o f the affinity g r o u p into the a c t i v e site; r a t h e r , t h e y r e s u l t f r o m m o d i f i c a t i o n b y v e r y small s u b s t i t u e n t s . S u c h r e a g e n t s c o u l d also be a p p l i e d to o t h e r e n z y m e s in o r d e r to d e t e r m i n e w h e t h e r the a m i n o acid r e s i d u e t h a t is modified is really i n v o l v e d in c a t a l y t i c a c t i v i t y o r is j u s t so close to the a c t i v e site t h a t its m o d i f i c a t i o n i n t e r f e r e s with s u b s t r a t e b i n d i n g .
E v a l u a t i o n of A c t i v e - S i t e - D i r e c t e d R e a g e n t s After designing and synthesizing an affinity-labeling reagent, we s h o u l d d e t e r m i n e w h e t h e r it is a c t i v e - s i t e - d i r e c t e d a n d c o l l e c t the d a t a req u i r e d to l e a r n a b o u t the e n z y m e s t r u c t u r e a n d f u n c t i o n . A c c o r d i n g l y , the f o l l o w i n g e x p e r i m e n t a l criteria for affinity l a b e l i n g will b e d i s c u s s e d , in o r d e r to e x p l a i n w h y c e r t a i n q u e s t i o n s s h o u l d be a s k e d a n d h o w to u s e the answers. 42 y. Nakagawa and M. L. Bender, Biochemistry 9, 259 (1970). 43 E. H. White and B. R. Branchini, this series, Vol. 46, Article [61].
[25]
AFFINITY LABELING
487
Kinetics Do the kinetics of inactivation of the enzyme by the reagent show saturation behavior, as predicted by the mechanism of affinity labeling? The kinetics should be studied under conditions that maintain the activity of the enzyme in the absence of reagent. Usually the concentration of reagent should exceed the concentration of the enzyme, so that pseudofirst-order kinetics may be observed. Note that it is insufficient to simply report the molar ratio of the reagent to the enzyme because (a) it is the concentration of the reagent that determines the rate of inactivation when reagent exceeds enzyme concentration, and (b) it is often difficult to determine the molarity of the enzyme. If the kinetics of inactivation are pseudo-first-order, then the concentration of reagent should be varied and the pseudo-first-order rate constants for inactivation calculated, so that one can determine the kinetic constants that characterize the reaction of an active-site-directed reagent. If the kinetics of inactivation are not pseudo-first-order, the investigator may have to distinguish among a variety of possible explanations: reagent instability, partial activity of modified enzyme, etc. Kitz and Wilson44 derived an equation that is based on the appropriate assumption that the enzyme is a reagent, as contrasted to the less rigorous assumption of steady state .45 (The steady-state concentration of the reversible enzyme-activator complex does change as a function of time.) Thus for affinity labeling, E + I . kl,. E ' I - ~ k2
E-X,
Kl = k2/kl
it is assumed that formation of the reversible complex (E-I) is in rapid equilibrium compared to the formation of the irreversibly inactivated enzyme (E-X); that [I] > > [E]; and that when the reaction mixture is assayed for enzymatic activity, E and E.I produce full activity while E - X is inactive. Then
d[E-X]/dt = k3[E-I], KI = [E][I]/[E-I] [E]t = [E] + [E.I] + [E-X] = [E.I]Kff[I] + [E.I] + [E-X]
d[E-X]/dt = ka([E]t - [E-X])/(1 + Kff[I]) or
dIE-X] [E]t - [EX]
k3dt (1 + KI/[I])
Integrating, ln([E]t - [E-X]) = -katl(1 + Kff[I]) + ln[E]t 44 R. Kitz and I. B. Wilson, J. Biol. Chem. 237, 3245 (1962). 45 A. Cornish-Bowden, Eur. J. Biochem. 93, 383 (1979).
488
INITIAL RATE AND INHIBITOR METHODS
[9.5]
If the logarithm of enzyme activity [E]t - [E-X] -- [E] + [E-I] is plotted against time (most conveniently on semilog paper), the observed firstorder rate constant, kobs, may be calculated from the slope or by kobs ---0.693/t1/2, and is related to the desired constants by Eq. (1). kobs = k~[I]/(Ki + [I])
(1)
Since this equation predicts hyperbolic saturation kinetics, a plot of
1/kobs against 1/[I] allows a graphical estimation of kz and KI. For better estimates, one can use the programs of Cleland, 4n which provide the values and their standard errors. Note that when [I] < < KI, kobs equals the pseudobimolecular rate constant, kz/Ki, which has units o f M -1 sec -1 and is used to compare reactivities of various reagents--e.g., as a measure of the extent of facilitation obtained with active-site-directed reagents as compared to other reagents. We usually assume that an active-site-directed reagent should give saturation kinetics, but note that the reaction of human serum cholinesterase with diisopropylfluorophosphate did not give saturation kinetics, whereas another organophosphate, malaoxon, did. 47 Saturation kinetics may not arise if the rate of dissociation (k2) for the complex and the rate of the unimolecular reaction (k3) are relatively rapid. This makes it difficult to use concentrations of reagent approaching Kt, since the rate of reaction gets too fast to measure. If saturation kinetics are not observed and k3 and KI cannot be determined accurately, the apparent bimolecular rate constant, kJKi, can be calculated from the slope of a plot of kobs against [I]. The KI usually is taken as a measure of the affinity of the reagent for the enzyme, and it is gratifying when the K~ agrees with the inhibition constant determined for the reagent from competitive inhibition kinetics. However, the KI determined from the inactivation kinetics may be larger than that observed from competitive inhibition kinetics if the reagent binds in a less favorable way when it undergoes the chemical reaction than when it binds as a reversible inhibitor. On the other hand, if the KI values agree, it may be because one molecule of active-site-directed reagent binds to and protects the active site against bimolecular reaction by a second molecule of the reagent. E+I E+I
kb
~E-X
, K ' , E.I
Saturation kinetics are also observed, as the equation for this "self-protection mechanism ' ' ~ is kinetically equivalent to Eq. (1). 48 W. W. Cleland, this series, Vol. 63, Article [6]. 4¢ A. R. Main, Science 144, 992 (1964).
[25]
AFFINITY LABELING kbK~[I] kobs = K~ + [I]
489 (2)
This mechanism can be rendered unlikely if the following criteria are met.
Reactivity How much faster does the reagent react with the enzyme than does a simple reagent that does not have the affinity group? If the affinity label had a bromoacetyl function, bromoacetate or bromoacetamide could be used as the simple reagent for determination of the bimolecular rate of inactivation. If the pseudobimolecular rate constant, ks/K~, for reaction of the enzyme with the active-site-directed reagent is considerably faster than the bimolecular rate constant for the simple reagent, it is reasonable to conclude that true affinity labeling is occurring. On the other hand, if the affinity reagent reacts more slowly than the simple reagent, one cannot conclude that self-protection is occurring since the (usually large) affinity reagent could have reduced reactivity because of electronic or steric effects. In doing this experiment, a good control is to test also a mixture of two compounds that separately represent the affinity and functional groups of the active-site-directed reagent, for instance bromoacetate and a substrate or substrate analog. The binding of a substrate may induce a conformational change that exposes an amino acid side chain for reaction. In a similar way, reaction of a reagent that resembles one substrate of the enzyme may be stimulated by the presence of one or more of the other substrates of the enzymes. For instance, inactivation of lactate dehydrogenase by bromopyruvate, which is actually a substrate, was stimulated fivefold by NAD. 4a Another criterion for true affinity labeling is that active-site-directed reagents with the functional group in different positions on the affinity group should have different rates of inactivation, as reflected in ka/K~. If the functional group of a reagent is not juxtaposed to react with an amino acid side chain while the affinity group is bound, that reagent should not react with facilitation, but could still react by the bimolecular mechanism at a rate that may approach the rate with a simple reagent, or could react more slowly if the affinity group hinders access of the reagent to the site of reaction. Interpreting these experiments requires some information about the relative reactivities of the various reagents with an amino acid side chain: incorporating a functional group into a more complex structure could conJ. BerghAuser, I. Falderbaum, and C. Woenckhaus, Hoppe-Seyler's Z. Physiol. Chem. 352, 52 (1971).
490
INITIAL RATE AND INHIBITOR METHODS
[25]
ceivably alter its reactivity because of electronic or steric effects. The model nucleophile, 4-(p-nitrobenzyl)pyridine, which reacts to form a colored alkylated product, is useful for this purpose (see Table IV). It would be better, of course, to compare the rate constants for reactions of the reagent with the same type of group that reacts in the protein. Table V presents selected data on the reactivities of simple reagents with functional groups of amino acids and proteins. These data illustrate the range of possible reactivities and show that the microscopic environment of the protein can depress or considerably enhance the rate of reaction of simple reagents. This fact makes it difficult to eliminate the self-protection mechanism if the rates of reaction of simple reagents with the enzyme are not determined. Inactivation Does the active-site-directed reagent completely inactivate the enzyme? If an essential residue is being modified, one expects no residual activity. However, many investigators fail to follow a reaction after enzyme activity is reduced to less than 10% of the initial activity, probably because the assays are inconvenient. Moreover, the kinetics may start to deviate from first order, and the investigator chooses not to be concerned with this complication. Critical information is lost thereby which could be obtained by simply starting a reaction mixture with a concentration of enzyme that is l0 or 100 times higher than is normally used and making dilutions in order to follow the reaction from zero to more than 99% inactivation. If the reaction begins to slow down substantially, more reagent can be added as a check for reagent decomposition. (One can also analyze for reagent, of course, or determine its rate of decomposition in the reaction medium.) If the enzyme cannot be completely inactivated, it may be because (a) some impurity or decomposed reagent binds tightly to the active site and protects against reagent, or (b) the modified enzyme has residual activity, or (c) the enzyme preparation is heterogeneous, containing some unmodifiable isoenzyme, or (d) other reasons. As a start toward distinguishing among these possibilities, the enzyme can be isolated from the reaction mixture and re-treated with reagent, or its kinetics of action on substrates can be studied. If (b) or (c) hold, the kinetics may be significantly different than those for native enzyme. For instance, chymotrypsin alkylated on Met-192 with bromoacetyl-aaminoisobutyrate has 20% of the residual activity of native chymotrypsin in a standard assay, but its Vmax is increased by 1.4-fold (accompanied by a 10-fold increased Km) with acetyltyrosine ethyl ester as substrate. 4a On 49w. B. Lawsonand H.-J. Schramm,Biochemistry 4, 377 (1965).
TABLE V SELECTED DATA ON REACTIVlTIES OF SIMPLE REAGENTS WITH MODEL COMPOUNDS AND ENZYMES a 1. C y s t e i n e s u l f h y d r y l a. C I C H ~ C O N H ~ G l y - C y s - G l y , 30 °,/~ = 0.27 M -1 s e c -~, pK 9.0 b F i c i n , 30.1 °, k = 16 M -1 sec -1, p K 8.3 c P a p a i n , 30.5 ° k = 6.2 M -1 s e c -x, p K 8.5 a b. I C H z C O N H 2 G l u t a t h i o n e , 25 °, p H 11.2 ( m e r c a p t i d e ) , ko = 27 M -~ sec - l e T h i o l s u b t i l i s i n , 25 °, p H 7 (ion-pair), ko = 7.2 M -~ sec -~e P a p a i n , p H 5.5 (ion-pair), ko = 14 M -1 sec -~e P a p a i n , p H 10 ( m e r c a p t i d e ) , ko = 976 M -1 sec -~e L i v e r a l c o h o l d e h y d r o g e n a s e , 25 °, p H 7.2, 0.021 M -~ sec -1~ Y e a s t a l c o h o l d e h y d r o g e n a s e , 25 °, p H 7.6, ko = 0.43 M -~ sec - ~ Y e a s t h e x o k i n a s e B, 35 °, k = 14 M -~ sec -~, p K 10 h G l y c e r a l d e h y d e - 3 - p h o s p h a t e d e h y d r o g e n a s e 25 °, k = 280 M -1 sec -~, p K = 8.2 ~ c. B r C H 2 C O N H ~ L i v e r a l c o h o l d e h y d r o g e n a s e , 25 °, p H 8, ko = 0.027 M -1 s e c - ~ Y e a s t a l c o h o l d e h y d r o g e n a s e , 25 °, p H 7.9, ko = 0.37 M -~ sec -~k d. C H 3 F G l u t a t h i o n e , 25 °, p H 11.2 ( m e r c a p t i d e ) , ko = 0.92 M -~ s e c -1 T h i o l s u b t i l i s i n , 25 °, p H 7 (ion-pair), ko = 4.2 M -~ sec -~ P a p a i n , 25 °, p H 5.5 (ion-pair),/co = 0.028 M -~ sec -1 P a p a i n , 25 °, p H 10 ( m e r c a p t i d e ) , ko = 0.6 M -x sec -1 e. N - E t h y l m a l e i m i d e G l u t a t h i o n e , 25 °, p H 6.5, ko = 263 M -~ sec -~l P a p a i n , 25 °, p H 6.4, ko = 2.5 M -1 sec -~l Y e a s t a l c o h o l d e h y d r o g e n a s e , 20 °, p H 7.0, ko = 0.22 M -1 s e c -lm f. A c r y l o n i t r i l e G l u t a t h i o n e , 30 °, k = 0.59 M -1 sec -~, p K = 8.6 n B o v i n e s e r u m albumil~ ( r e d u c e d ) , 30 °, p H 7, 6 M u r e a , k o = 0.01 M -1 sec -~° 2. M e t h i o n i n e t h i o e t h e r a. I C H 2 C O O H P a n c r e a t i c r i b o n u c l e a s e , 25 °, p H 3.5, 8 M u r e a , ko = 7.2 x 10 -4 M -~ sec - ~ b. CH3 Ip a - N - A c e t y l m e t h i o n i n e , 25 °, p H 3.0, ko = 5 × 10 -4 M -1 sec -1 P a n c r e a t i c r i b o n u c l e a s e , 25 °, p H 3.3, ko = 8 × 10 -4 M -~ sec -~ c. BrCHI~CONH-C6H5 q c ~ - N - A c e t y l m e t h i o n i n e , 25 °, p H 6.0, 10% e t h a n o l , ko = 1.3 x 10 -3 M -1 sec -~ c ~ - C h y m o t r y p s i n - M e t - 1 9 2 , 27.5 °, p H 6.0, 10% e t h a n o l , ko = 0.26 M -1 s e c -~ 3. H i s t i d i n e i m i d a z o l e a. B r C H 2 C O O H " H i s t i d i n e h y d a n t o i n (N-1 + N-3), 25 °, p H 7.72, ko = 5.9 × 10 -~ M -~ sec -1, ( p K = 6.4) P a n c r e a t i c r i b o n u c l e a s e (His-12 + H i s - l l 9 ) , 25 °,/~ = 2.6 x 10 -~ M -~ see -~, p H o p t i m u m = 5.5, p K v a l u e s = 4.7, 6.3 b. B r C H 2 C H 2 C O O H ~ H i s t i d i n e , 25 °, p H 5.5, 2.3 x 10 -s M -~ sec -~ P a n c r e a t i c f i b o n u c l e a s e - H i s - l l 9 , 25 °, p H 5.5, k0 = 6.3 x 10 -4 M -~ sec -~ c. B r C H s C O C O O H ' P a n c r e a t i c r i b o n u c l e a s e - H i s - l l 9 , 25 °, p H 5.5, ko = 9.1 x 10 -2 M -~ s e c -~
491
(continued)
492
I N I T I A L RATE A N D I N H I B I T O R M E T H O D S TABLE V
[25]
(Continued)
d. ICH2CONH2 Histidine, 25 °, p H 5.3,/co = 1.2 x 10-6 M -1 sec -it Pancreatic ribonuclease, 25 °, pH 5.3 (pH optimum) ko = 1.1 x 10-4 M -~ sec -it Pancreatic trypsin, 25 °, p H 7.0,/co = 5.3 x 10 -6 M -~ sec -~ (pK = 6.7) u e. 1-Fluoro-2,4-dinitrobenzene v a-N-Acetylhistidine, 20 °, k = 7.4 x 10 -4 M -~ sec -~, pK 7.2 a-Chymotrypsin, 20 °, k = 7.5 x 10-3 M -~ sec -~, p K 6.8 4. Lysine, ~-amino a. BrCI-~COOH Pancreatic ribonuclease-Lys-41, 25 °, p H 8.5, k0 = 2.6 x 10-3 M -~ sec -lw b. 1-Fluoro-2,4-dinitrobenzene x Gly-Lys, 15°, k = 0.22 M -1 sec -~, p K = 10.1 Ribonuclease-Lys-41, 15°, k = 0.44 M -~ sec -~, p K = 8.9 c. Acrylonitrile • -Aminocaproic acid, 30 °, k = 1.96 x 10-3 M -~ sec -~, p K 10.6" 5. Peptidyl, a-amino a. l-Fluoro-2,4-dinitrobenzene Gly-Gly, 30 °, 6% dioxane, p H 10, k0 = 0.37 M -~ sec - ~ Valyl terminal o f streptomyces griseus trypsin, 20 °, /~ = 6.7 x 10-3 M -1 sec -~, p K 8.1 z b. N-Ethylmaleimide H2N-Val-Leu-Ser . . . . 25 °, p H 7.4, k0 = 6 x 10 -3 M -1 sec -laa c. Acrylonitrile Tetraglycine, 30 °, k = 5.7 x 10 -3 M -~ sec -~, p K = 7.6 ~ 6. Serine hydroxyl CH3SO2F: chymotrypsin-Ser-195, 25 °, p H 7, k0 = 0.021 M -~ sec -~°b a Since the unprotonated forms of the amino acid side chains usually react m u c h faster than the protonated forms, the p H dependence of the rate constant should be measured. Where this was done, the pH-independent rate constant k is reported, together with the apparent p K value. In other cases the observed rate constant ko is given, along with pH values. In general, reactions were studied in buffers of about 0.1 ionic strength, but the original papers should be consulted for specific details and other examples. b H. Lindley, Biochem. J. 82, 418 (1962). c j. R. Whitaker and L.-S. Lee, Arch. Biochern. Biophys. 148, 208 (1972). d I. M. Chaiken and E. L. Smith, J. Biol. Chem. 244, 5087 (1969). e p. Hahisz and L. Polg~lr, Eur. J. Biochem. 71,563,571 (1976). I N . Evans and B. R. Rabin, Eur. J. Biochem. 4, 548 (1968). g E. P. Whitehead and B. R. Rabin, Biochem. J. 90, 532 (1964). h j. G. Jones, S. Otieno, E. A. Barnard, and A. K. Bhargava Biochemistry 14, 2396 (1975). L. Polg~tr, Eur. J. Biochern. 51, 63 (1975). J R. W. Fries, D. P. Bohlken, R. T. Blakley, and B. V. Plapp, Biochemistry 14, 5233 (1975). k B. V. Plapp, C. Woenckhaus, and G. Pfleiderer, Arch. Biochem. Biophys. 128, 360 (1968). t B. L. B. Evans, J. A. K n o p p , and H. R. Horton, Arch. Biochem. Biophys. 206, 362 (1981).
[25]
AFFINITY LABELING
493
j. R. Heitz, C. D. Anderson, and B. M. Anderson, Arch. Biochem. Biophys. 127, 627 (1968). n M. Friedman, J. F. Cavins, and J. S. Wall, J. Am. Chem. Soc. 87, 3672 (1965). o j. F. Cavins and M. Friedman, J. Biol. Chem. 243, 3357 (1968). P T. P. Link and G. R. Stark, J. Biol. Chem. 243, 1082 (1968). E. W. Bittner and J. T. Gerig, J. Am. Chem. Soc. 92, 2114 (1970). r E. P. Lennette and B. V. Plapp, Biochemistry 18, 3933, 3938 (1979). R. L. Heinrikson, W. H. Stein, A. M. Crestfield, and S. Moore, J. Biol. Chem. 240, 2921 (1965). t R. G. Fruchter and A. M. Crestfield, J. Biol. Chem. 242, 5807 (1967). u T. Inagami and H. Hatano, J. Biol. Chem. 244, 1176 (1969). v W. H. Cruickshank and H. Kaplan, Biochem. J. 130, 1125 (1972). w R. L. Heinrikson, J. Biol. Chem., 241, 1393 (1966). x A. L. Murdock, K. L. Grist, and C. H. W. Hirs, Arch. Biochem. Biophys 114, 375 (1966). J. T. Gerig and J. D. Reinheimer, J. Am. Chem. Soc. 97, 168 (1975), z R. G. Duggleby and H. Kaplan, Biochemistry 14, 5168 (1975). aa D. G. Smyth, O. O. Blumenfeld, and W. Konigsberg, Biochem J. 91, 589 (1964). bb A. M. Gold, this series, Vol. 11, Article [83]. m
other substrates, keat is increased as much as eightfold. 5° D-Amino acid oxidase treated with the suicide substrate D-propargylglycine also appears to have residual activity, and its Km values for amino acids appear to be differentially altered. ~1 Protection
Do substrates, products, or reversible inhibitors protect against inactivation? Such protection is the usual evidence that reaction occurs at the active site, although another possible explanation is that the "active-sitedirected" reagent reacts at an "allosteric" site. The protective agent (L) should give competitive inhibition with an inhibition constant of KL against inactivation by the reagent, according to the following equation: ka[I] kobs = KI(1 + [L]/KL) + [I] Therefore, one should do an experiment with varied concentrations of I and L, in order to evaluate k3, K,, and KL by the usual procedures used in enzyme kinetics? 2 If the inhibition pattern is. not competitive or if KL is not of about the same magnitude as the constant determined by other methods (equilibrium dialysis, inhibition of enzyme activity), inactivation 5o F. J. K~zdy, J. Feder, and M. L. Bender, J. Am. Chem. Soc. 89, 1009 (1967). 51 p. Marcotte and C. Walsh, Biochemistry 17, 2864 (1978). ~2 j. A. Todhunter, this series, Voi. 63, Article [15].
494
INITIAL RATE AND INHIBITOR METHODS
[25]
by the reagent may not be due to reaction at the active site. It should be apparent from this discussion that just using one (high) concentration o f a substrate to test for protection is p o o r experimental design; e n z y m e s may have, in addition to the active site, low-affinity sites where substrates or inhibitors can bind. Specificity Does modification of one amino acid residue correlate with complete inactivation? This is the ultimate criterion for the specificity o f labeling and the efficacy o f the active-site-directed reagent. Usually the investigator varies the extent o f inactivation by treatment for varied times o f reaction or with varied concentrations o f inactivator, isolates the e n z y m e from the reaction mixture, and determines the incorporation o f reagent. Typically, a plot o f residual activity against incorporation deviates from linearity (1:1 stoichiometry) and shows that some secondary sites are reacting before the active site is completely modified. In some unusual cases, modification o f one site appears to inactivate two sites ("half-ofthe-sites" reactivity). 53-~ After finding the reagent or conditions that give specific labeling, the amino acid residues that are modified should be identified. Amino acid analysis can often be used to determine which amino acid is modified, either because o f loss o f an amino acid or because of conversion to a derivative. Studies with model compounds are essential for identification and have been carried out with many reagents. Amino acid sequence analysis o f labeled peptides is then required. If the three-dimensional structure has been determined, X-ray crystallography can be used to identify the modified sites (e.g., Walder e t a1.56). Sequence analysis is often difficult, but the investigator should try to determine the extent o f labeling of each o f the various sites modified and the r e c o v e r y o f label throughout the purification o f labeled peptides. Note that it was the careful analysis of the products o f carboxymethylation o f pancreatic ribonuclease that led to the concept that two histidine residues were involved in catalytic activity o f that enzyme. 57 Quantitative analyses on the c a r b o x y m e t h y l a t e d derivatives o f S t r e p t o c o c c u s f a e c i u m dihydrofolate reductase also allowed identification o f three methionine residues that could be essential when about 1.5 methionines reacted in a differential labeling experiment. ~8 The three s3 s. A. Bernhard and R. A. MacQuarrie, J. Mol. Biol. 74, 73 (1973). W. B. Stallcup and D. E. Koshland, Jr., J. Mol. Biol, 80, 41, 63, 77 (1973). 55A. Levitzki, J. Mol. Biol. 90, 451 (1974). 56j. A. Walder, R. Y. Walder, and A. Arnone, J. Mol. Biol. 141, 195 (1980). 5r A. M. Crestfield, W. H. Stein, and S. Moore, J. Biol. Chem. 238, 2413 (1963). 5s j. M. Gleisner and R. L. Blakley, Fur. J. Biochem. 55, 141 (1975).
[25]
AFFINITY LABELING
495
methionines m o d i f i e d - - r e s i d u e s 28, 36, and 5 0 - - c o r r e s p o n d to Leu-27, Val-35, and Phe-49 in the homologous sequence of the L a c t o b a c i l l u s c a s e i enzyme, for which X-ray crystallography identifies Leu-27 and Phe-49 as part of the binding site for the inhibitor m e t h o t r e x a t e ? 9 The reaction of human serum prealbumin with N-bromoacetyl-L-thyroxine also revealed several sites of modification, giving a rough map of the binding site even while showing how nonspecific affinity labels can be. 6° Since it is useful to know which secondary sites react, for new insights into s t r u c t u r e - f u n c tion relationships, investigators should try to characterize the products of reaction as completely as possible, rather than to presume that modification is absolutely specific. E v e n tosyl-L-phenylalanylchloromethane, which alkylates predominately His-57 in chymotrypsin, may modify 0.3 residues o f Met-192 per molecule. 61 Although photoaffinity labeling appears, in principle, to be a good method to label many different residues in the active site, the low yields of each product make sequence analysis difficult. A possible complication of photoaffinity "labeling" is that the reagent may not b e c o m e attached to the protein, and it could b e c o m e difficult to locate the site of reaction. Ogez e t al. 62 have found that Asp-38 in As-3-ketosteroid isomerase is converted to an alanine upon photoinactivation in the presence of 3-oxo-4-estren-17/3-yl acetate. A review of other photo-labeling studies suggests that this complication may occur in other cases. 6 On the other hand, if the enzyme can be inactivated without attaching a sterically perturbing group, conclusions about the functional role of the modifiable amino acid may be more definitive. Not many sites o f labeling by " s u i c i d e " reagents have been identified, but there appear to be two problems with these reagents. The activated reagent may react with a c o e n z y m e at the active site rather than with an amino acid residue, or the reagent may diffuse away from the active site and react with other sites on the same, or other, proteins, a Facilitation of R e a c t i o n A c h i e v e d with A c t i v e - S i t e - D i r e c t e d Reagents H o w much faster do affinity labeling agents react with an e n z y m e than do simple reagents without the affinity groups? That is, what is the ratio of the pseudobimolecular rate constant for the reaction o f the specific re59D. A. Matthews, R. A. Alden, J. T. Bolin, D. J. Filman, S. T. Freer, R. Hamlin, W. G. J. Hol, R. L. Kisliuk, E. J. Pastore, L. T. Plante, N. Xuong, and J. Kraut, J. Biol. Chem. 253, 6946 (1978). oo S.-Y. Cheng, M. Wilchek, H. J. Cahnmann, and J. Robbins, J. Biol. Chem. 252, 6076 (1977). ol K. J. Stevenson and L. B. Smillie, Can. J. Biochem. 46, 1357 (1968). 02j. R. Ogez, W. F. Tivol, and W. F. Benisek, J. Biol. Chem. 252, 6151 (1977).
496
INITIAL RATE AND INHIBITOR METHODS
[25]
agent (kalKi) to the bimolecular rate constant for the simple reagent (kb)? We assume that large values of this ratio will yield more specific reaction, and thus this ratio should be determined as part of the characterization of the reagent. Unfortunately, many investigators just report that a simple reagent did not inactivate under the same conditions employed for the specific reagent. Furthermore, there is too little information available on the rates of reaction of simple reagents with compounds that model amino acid side chains (Table V) to serve as a basis for comparison. How much facilitation could be expected for an active-site-directed reagent? If it is assumed that such reagents bind as do normal substrates, the extent of facilitation could approach the amount of catalytic power exerted by the enzyme as a result of binding of the reactive groups in stereochemically correct juxtaposition. The magnitude of this catalytic factor has been estimated in a variety of ways, depending on the presumed origin of catalysis: "propinquity" (Bruice and Benkovic63), "proximity and orientation" (Koshland64), or "Circe effect" (Jencks35). In the most favorable cases, we expect the facilitation ratio to be at least 103. Of course, the facilitation obtained can be much less if the reagent is poorly designed and binds so that the reactive groups are not properly positioned for reaction. As discussed above, there would be no facilitation if the reagent bound only in a nonproductive mode. Clearly, the structure of the reagent is critical for achieving facilitation. This can be demonstrated by considering the results of Shaw et al. on serine proteases. As shown in Table VI, the magnitude of facilitation varies from about 10a to about 106 with a homologous series of chloromethyl ketones. The best reagent is 1700 times more reactive than the poorest with chymotrypsin. Now, these ratios are calculated with reference to the rate of reaction with o~-N-acetylhistidine; another point of reference should be the rate of reaction of something like CHaCOCHzCI with chymotrypsin. This number would allow one to estimate the "inherent" reactivity of His-57. The acylation of Ser-195 in chymotrypsin by sulfonyl fluorides is facilitated l(P-fold by the addition of a phenyl ring to methanesulfonyl fluoride. 65 More facilitation would be expected with more specific reagents. This discussion raises an old issue in protein chemistry: is the reaction of an amino acid residue with a reagent facilitated because of binding of reagent to the enzyme (according to the classical theory of affinity labeling) or because the amino acid is "hyperreactive," due to its microenvironment? It is well known that simple reagents can react faster by an ea T. C. Bruice and S. J. Benkovic, "Bioorganic Mechanisms," Vol. 1. Benjamin, New York, 1966. D. E. Koshland, Jr., J. Theor. Biol. 2, 75 (1962). 6~ D. E. Fahrney and A. M. Gold, J. Am. Chem. Soc. 85, 997 (1963); or A. M. Gold, this series, Vol. 11, Article [83].
[25]
AFFINITY
LABELING
497
T A B L E VI EFFECT OF REAGENT STRUCTURE ON MAGNITUDE OF FACILITATION OF REACTION OF CHYMOTRYPSIN a k (M -1 sec -1)
R-Phenylalanyl-CH~-X R-
-X
Chymotrypsin
HNa-Formyl N~-Tosyl N~-Benzyloxycarbonyl N~-Formyl N~-Benzyioxycarbonyl -
-CI -C1 -Ci -CI -Br -Br
0.041 1.35 7.7 69 26 790
N~-Acetylhistidine
4.5 × 10-5 1.7 x 10-4
a At p H 7, 25°; c h y m o t r y p s i n in 5 - 1 0 % methanol, N~-acetylhistidine in 80% methanol [E. Shaw and J. Ruscica, Arch. Biochem. Biophys. 145, 484 (1971)].
order of magnitude or more with certain groups on a protein than with other groups of the same type on the protein or on free amino acids. In one such case, fluorodinitrobenzene reacts selectively with Lys-41 in RNase A because the pK value of the e-amino group is depressed about 1.4 units, although the pH-independent rate constant is similar to that for a normal primary amine (Table V, item 4b). In a contrasting case, the pHindependent rate constant for reaction of fluorodinitrobenzene with His57 in chymotrypsin is 10 times that for a-N-acetylhistidine, apparently because the carboxylate of Asp-102 increases the nucleophilicity of His-57 (Table V, item 3e). It is apparent from these examples that one cannot determine whether a reagent is "active-site-directed" simply by comparing the rates of reaction of the enzyme and a model amino acid. In order to differentiate between possible mechanisms of facilitation (that is, affinity labeling or hyperreactivity), we have explored the use of a simple active-site-directed reagent, bromoacetate, which reacts about 103 times faster with histidine residues 12 and 119 in ribonuclease than with a model histidine compound66 and binds with a dissociation constant of about 23 mM at pH 5.5 and 37° before reaction. 1° Stark e t a l . 67 explained the facilitation of the reaction by postulating that the protonated imidazole nitrogen of one histidine attracted the carboxylate of the haloacetate, orienting the reagent for nucleophilic attack by the unprotonated nitrogen of the other imidazole. This mechanism is consistent with the three-dimensional structure of the enzyme. 6s Jencks 3~ has explained the facilitaE. P. L e n n e t t e and B. V. Plapp, Biochemistry 18, 3933, 3938 (1979). 67 G. R. Stark, W. H. Stein, and S. Moore, J. Biol. Chem. 236, 436 (1961). 68 F. M. Richards and H. W. W y c k o f f in " T h e E n z y m e s " (P. D. Boyer, ed.), 3rd ed., Vol. 4, p. 647. A c a d e m i c Press, N e w York, 1971.
498
INITIAL RATE AND INHIBITOR METHODS
[25]
tion by saying that the interaction of ribonuclease and the haloacetate serves to decrease the free rotation and translation of the reagent. These explanations are equivalent to saying that the haloacetate is an affinity label and imply that the binding of reagent results in a more favorable entropy of reaction. An alternative explanation is that the environments of the histidines make them hyperreactive, which would imply a more favorable enthalpy of reaction. (In our definition, a hyperreactive group has enhanced inherent reactivity due to electronic effects and would be more reactive even with a reagent that did not bind as an affinity label.) Based on these ideas, we proposed that transition-state analysis could be used at least to characterize and possibly to differentiate between the mechanisms of facilitation.66 If the entropy of activation (AS*) were more favorable, affinity labeling is implied; a more favorable enthalpy of activation (AH*) implies hyperreactivity. In order to have a valid basis of reference, the activation parameters for the reaction of bromoacetatc with ribonuclease were compared to the parameters for the reaction of bromoacetate with histidine hydantoin, which is a good model of a histidine residue in a protein. Since the imidazole ring has two nitrogens that can react, the rates of reaction of each nitrogen were determined. The surprising finding was that the enhanced reactivities of histidine 12 and 119 in ribonuclease are due almost entirely to more favorable magnitudes of AH*. The reactivity of His-12 could be increased because of hydrogenbonding to the carbonyl oxygen of Thr-45, and the reactivity of His-119 because of contact with the carboxylate of Asp-121. If the two imidazoles have increased nucleophilicities, they should also be hyperreactive with a simple alkylating agent such as iodoacetamide. Although His-12 does react readily with iodoacetamide, His-ll9 does not. 69 Thus we must assume that the carboxylate group of the haloacetate is involved in the reaction mechanism, and we proposed that the binding of bromoacetate stabilizes the enzyme in a particular conformation, with His-ll9 fixed in a position that confers upon it hyperreactivity, as shown in the following scheme: \ CH2C\o Asp-121
_
H--N.....~N His-119
CHiC Br
®
H
H'-'O=C / \
His-12
Thr-45
Of course, this explanation implies that binding energy is used to orient the reagent, and thus we could expect changes in AS*. However, differen6a R. G. Fruchterand A. M. Crestfield,J. Biol. Chem. 242, 5807 (1967).
[25]
AFFINITY LABELING
499
tial solvation effects on the c a r b o x y m e t h y l a t i o n o f histidine hydantoin as c o m p a r e d to ribonuclease could result in c o m p e n s a t i o n b e t w e e n the changes in AH* and AS*. Complications of this sort can limit the utility o f activation p a r a m e t e r s . N e v e r t h e l e s s , we believe such data are required to provide an experimental basis for theories being d e v e l o p e d to explain enz y m e catalytic power.
Concluding Remarks It is clear that affinitylabeling has bccn useful for studying enzymes, cvcn though considerable effort is required to design, synthesize, and evaluate such reagents. Unfortunately, many of the reagents that have bccn prepared have not been dcrnonstratcd to bc significantlymore reactive or specificthan simpler chemical reagents. This has led to the publication of many incomplete studics. Progress in this area would bc facilitated if investigators would carefully assess their objectives and cautiously apply the approach as one of several parallelcourses of study. It is especially important that investigatorsbc more criticalof their results and perform more of the experiments required to evaluate the candidate active-site-directedreagents. Despite the work involved and the difficultiesthat have been encountered, the applications of affinity labeling reagents arc potentially cxtrcrncly valuable. In addition to their uses for studying enzyme structure and function, they offer one of the best avenues for the rational developrncnt of spccific chemotherapeutic agents. Unfortunately, so little is known about the detailed intcractions of small molecules with largc molecules that extensive empirical studies arc required before effective compounds can bc prepared. In this context, study of enzymes can serve as models for more complicated regulatory systems. Furthermore, affinity reagents arc bcing applied in cvcr rnorc sophisticated ways to study receptor systems. Thus, I bclicvc that significanteffortsin the dcvclopmcnt of affinitylabeling reagents arc justified and will bc rewarding. Acknowledgments This work was supported by research grant AA00279 from the National Institute on Alcohol Abuse and Alcoholism. The editorial assistance of Rosemary K. Plapp was appreciated.