12 Pancreatic Ribonuclease

Pancreatic R ibmuclease PETER BLACKBURN

STANFORD MOORE

I . Introduction . . . . . . . . . . . . . . . . . . . . . . . . . I1 . Preparation . . . . . . . . . . . . . . . . . . . . . . . . . 111. Chemical Properties . . . . . . . . . . . . . . . . . . . . . . A . Modification of Functional Groups . . . . . . . . . . . . . . B . Roles of Residues Near the NHZand COOH Termini . . . . . C . Chemical Synthesis . . . . . . . . . . . . . . . . . . . . D . Biosynthesis . . . . . . . . . . . . . . . . . . . . . . . E . Immunochemistry . . . . . . . . . . . . . . . . . . . . . IV. Physical Properties . . . . . . . . . . . . . . . . . . . . . . A . X-Ray Diffraction . . . . . . . . . . . . . . . . . . . . . B . Nuclear Magnetic Resonance . . . . . . . . . . . . . . . . C . Optical Properties . . . . . . . . . . . . . . . . . . . . . D . TheFoldingPathway . . . . . . . . . . . . . . . . . . . . V. Species Variations . . . . . . . . . . . . . . . . . . . . . . . A . Variations in Amino Acid Sequence . . . . . . . . . . . . . B . Variations in Carbohydrate Moieties . . . . . . . . . . . . . VI . Bovine Seminal Plasma RNase . . . . . . . . . . . . . . . . . VII . Cytoplasmic RNase Inhibitor . . . . . . . . . . . . . . . . . . A . Purification and Chemical Properties . . . . . . . . . . . . . B . Studies on in Virro Protein Synthesis . . . . . . . . . . . . VIII Catalytic Properties . . . . . . . . . . . . . . . . . . . . . . A.Assay . . . . . . . . . . . . . . . . . . . . . . . . . . B . Inhibitors and Activators . . . . . . . . . . . . . . . . . . C.Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . D. Mechanism of Catalysis . . . . . . . . . . . . . . . . . . IX . Research Applications . . . . . . . . . . . . . . . . . . . . .

.

.

1

317 318 320 320 345 359 361 362 364 364 366 382 385 397 398 407 411 416 416 423 424 424 426 428 430 433

Introduction

Bovine pancreatic ribonuclease. the first enzyme for which the chemical structure could be written. has been the subject of extensive structure-function studies . The literature to 1970 was reviewed by 3 17 THE ENZYMES. VOL . XV Copyright @ I982 by Academic Press. Inc . All rights of reproduction in any form reserved . ISBN C-12-122715-4

3 18

PETER BLACKBURN AND STANFORD MOORE

Richards and Wyckoff ( I ) in Volume IV of this series. The present chapter covers some of the research on the enzyme in the subsequent decade. A characteristic of the research since 1970 is that it has broadened in scope. The subject has been reviewed within the context of mammalian nucleolytic enzymes by Sierakowska and Shugar ( 2 ) . When the structural work on RNase was begun in the 1950s, the enzyme was viewed as a catalyst of rather limited physiological interest; it was recognized as one of the enzymes of the digestive tract. Neutral RNases of similar molecular design are now known to be present in many tissues, and the specific cytoplasmic inhibitor of enzymes of this type has been characterized in considerable detail. The basic chemical information on the bovine pancreatic enzyme has facilitated studies of the biochemistry of a number of members of the class of catalysts (EC 3.1.27.5) defined as those which, at near neutral pH, cleave RNA endonucleolytically to yield 3’-phospho-, mono-, and oligonucleotides ending in Cp or Up, with 2’,3’-cyclic phosphate intermediates. In most of this review the term RNase refers to bovine pancreatic RNase A; in the extension of the discussion to other members of the series, the enzyme is defined in context. RNase S refers to the enzyme cleaved primarily between residues 20 and 21 by subtilisin ( I ) . RNase A, an unusually well-defined enzyme, has been a test protein in the study of a wide variety of chemical and physical methods of protein chemistry. The volume of the literature has necessitated a selection of topics for this chapter. II. Preparation

The starting product for the chromatographic purification of bovine pancreatic RNase has usually been the enzyme prepared as described by Kunitz and McDonald ( 3 ) . The previously reviewed ( 1 ) methods of ion exchange chromatography have proved effective in yielding homogeneous preparations of the protein, including the separation of RNases that differ in the extent of glycosylation. Such a fractionation is exemplified by the isolation of ovine RNases from pancreatic juice by Becker et a / . ( 4 ) . The introduction of affinity chromatography has simplified the isolation of pure RNases by providing a highly efficient method for separating active enzymes from molecules that do not have an affinity for the coupled substrate analog; the technique can be used as an early step in purificaI. 2. 3. 4.

Richards, F. M., and Wyckoff, H. W. (1971). “The Enzymes,” 3rd. ed., Vol. IV, p. 647. Sierakowska, H . , and Shugar, D . (1977). f r o g r . Nircleic Acid Rrs. M o l . B i d . 20, 59. Kunitz, M., and McDonald, M. R . (1953). Biocliem. f r e p . 3, 9. Becker, R. R., Halbrook, J. L., and Hirs, C. H. W. (1973). JBC 248, 7826.

12. PANCREATIC RIBONUCLEASES

3 19

tion or as a final step after ion exchange chromatography and gel filtration have been used to isolate an RNase fraction of given charge and size. Wilchek and Gorecki ( 5 ) introduced the use of 5’-(4-aminophenylphosphoryl)uridine-2’(3‘)-phosphate-Sepharose4B (pup-Sepharose) for this purpose; the purified enzyme was adsorbed at pH 5.2 in 0.02 M sodium acetate buffer and was eluted with 0.2 M acetic acid. Under these conditions, some subsequent users of the method have observed adsorption of proteins other than RNase and difficulty in the elution of the enzyme with acetic acid. Both of these problems have been overcome by maintaining a sufficient cation concentration in the loading buffer and in the eluent to nullify the properties of the acidic adsorbent as a nonspecific cation exchanger. Stewart and Stevenson ( 6 ) , in the course of preparing bison RNase, found that the positively charged heterocyclic base piperazine was preferable to Na’ for reduction of extraneous protein adsorption: they added the sample in 0.025 M piperazine-HC1 buffer at pH 5.3. Elution was with 0.25 M sodium phosphate buffer at pH 3: phosphate was chosen for the eluent since this anion has an appreciable affinity for RNase. In their isolations of pancreatic RNase from a variety of animal species, Beintema and his associates ( 7 - 1 0 ) have generally used acid extraction and (NH&S04 precipitation and applied the RNase-containing solution to pup-Sepharose in 0.23 M acetate buffer at pH 5.2; elution was with the same buffer, 4 M in NaC1. Alternatively, a linear salt gradient from 0.2 to 6 M NaCl was used for elution. In the isolation of human pancreatic RNase, Weickmann P t ul. (I I ) combined acetone precipitation, chromatography on phosphocellulose, and adsorption on pup-Sepharose. A long column (0.4 x 72 cm) of pup-Sepharose was used by Wang and Moore (12) to remove RNase from preparations of pancreatic DNase; the RNase content was reduced to less than 1 part per 10 million. Smith ef d. (13) used uridine 5’-triphosphate-hexane-agaroseas an affinity adsorbent with the same buffer systems as used by Beintema et ul. They found that a change to pH 5.5 for the eluent buffer gave slightly 5. Wilchek, M . , and Gorecki, M. (1969). EJB 11, 491. 6. Stewart, G. R., and Stevenson, K . J. (1973). BJ 135, 427. 7. Wierenga, R . K . , Huizinga, J. D . , Gaastra, W., Welling, G. W., and Beintema, J. J. (1973). FEES L e t f . 31, 181. 8. Gaastra, W., Groen, G., Welling, G . W., and Beintema, J. J . (1974). FEES L e u . 41, 227. 9 . Gaastra, W., Welling, G. W., and Beintema, J. J. (1978). EJB 86, 209. 10. Havinga, J . , and Beintema, J . J. (1980). EJB 110, 131. 1 1 . Weickmann, J. L., Elson, M., and Glitz, D. G. (1981). B i o c h e r n i s r ~ .20, 1272. 12. Wang, D., and Moore, S . (1978). JBC 253, 7216. 13. Smith, G. K., Schray, K. J., and Schaffer, S . W. (1978). A n d . Biochem. 84, 406.

320

PETER BLACKBURN AND STANFORD MOORE

sharper elution. The 5’-UTP derivative is less stable than pup-Sepharose, columns of which can be used repeatedly without loss of effectiveness ( 5 ) . Scofield et (11. (14) synthesized N4-(aminohexanoy1aminoheptyl)cytidine 2’(3‘)-phosphate-Sepharose4B by first using a bisulfite-induced transamination to introduce an alkyl diamine at the 4 position of cytidine. The product was compared with pup-Sepharose for the chromatography of RNase. The two adsorbents performed similarly and bound about 5 mg of RNase per milliliter of settled bed. 111.

Chemical Properties

Discussion of the results of chemical modification of bovine pancreatic RNase is made with reference to the sequence (15) in Fig. 1. The geometry is considered in reference to the three-dimensional structure derived by Richards and Wyckoff ( 1 , 16) for RNase S (Fig. 2,Ref. 160). A. MODIFICATION O F FUNCTIONAL GROUPS 1. Amino Croirps

Many of the reactions that modify the lysine residues of RNase A have been summarized in Table VI of Richards and Wyckoff’s review ( I ). Of the ten lysine residues, Lys-41 has been placed at the active center of the enzyme by both chemical and physical studies (1, 16-21); Lys-7 is nearby. Pyridoxal phosphate has been found to form a Schiff base with the c-NH2group of either Lys-41 or Lys-7 ( 2 2 , 2 3 )and with the a-NH2group of Lys-1 (24); reduction of the Schiff bases with borohydride yields stable adducts. The products of the reaction were separated and identified by 14. Scofield, R. E., Werner, R. P., and Wold, F. (1977). A n d . Biochem. 77, 152. 15. Smyth, D. G., Stein, W. H., and Moore, S. (1963). JBC 238, 227. 16. Richards, F. M., and Wyckoff, H. W. (1973). In “Atlas of Molecular Structures in Biology,” Vol. I , Ribonuclease S (D. C. Phillips and F. M. Richards, eds.), p. 1. Clarendon, Oxford. 16a. Cantor, C. R., and Schimmel, P. R. (1981). “Biophysical Chemistry,” p. 930. Freeman, San Francisco, California. 17. Glick, D. M., and Barnard, E. A. (1970). BBA 214, 326. 18. Brown, L. R., and Bradbury, J. H. (1975). EJB 54, 219. 19. Brown, L. R., and Bradbury, J . H. (1976). EJB 68, 227. 20. Jentoft, J . E . , Jentoft, N.,Gerken, T. A., and Dearborn, D. G. (1979).JBC 254,4366. 21. Wodak, S . Y., Liu, M. Y.,and Wyckoff, H. W. (1977). JMB 116, 855. 22. Means, G. E. , and Feeney, R. E. (1971). JBC 246, 5532. 23. Raetz, C. R. H., and Auld, D. S. (1972). Biochemistry 11, 2229. 24. Riquelme, P., Brown, W. E., and Marcus, F. (1975). Int. J . Peptide Protein Res. 7, 379.

4

3 2

fi

5

6

7

9

8

10

12

11

13

14

16

15

17

18

19

20

A l a - A l a ~ . y s ~ ~ e - G 1 ~ - * ~ ~ ~ & Z - H i s -Asp M e.ct ~er?fje

Ala t f

Tbr t Glu

r"b

+"pGlu

t

-

75

Ser- 9 r -

JiYS

nnl -

-

55

l-Ala+ Cys-Glu-G+p

N n P-

115

116 117 118 119 120

~ - ? L P - ~ ~ 1 - P r o - i ~ - ~ - ~ ~ ~ ~ ~ - ~ ~

VII

100

48

47

46

45

44

43

42

41

40

11

94

39

38

37

36

35

34

33

FIG.1 . The sequence of amino acid residues in bovine pancreatic RNase A. From ( I S ) , based upon the combined researches of the several laboratories referred to therein.

FIG.2. Three-dimensional structure of RNase S, based upon the data of Richards and Wyckoff (I). From (160); illustratian caavrieht. Irvine Geiss.

12. PANCREATIC RIBONUCLEASES

323

peptide mapping and amino acid-analyses. They had 0, 17, and 58% of the activity of RNase, respectively. Pyridoxal phosphate is a competitive inhibitor with respect to the substrate cyclic 2‘ ,3’-CMP (23) whereas pyridoxal itself does not react with the enzyme (22). Circular dichroism measurements in the far-UV and thermal transition profiles measured by CD suggest that alkylation of Lys-7 or Lys-41 via pyridoxal phosphate does not significantly affect the conformation of the molecule (25). Results on the binding of 3’-CMP and the kinetics of the hydrolysis of cycIic 2‘,3‘-CMP agree with those of Riquelme et al. (24) and indicate that the eNH2 group of Lys-7, although located in the region of the active site, is not directly involved in catalysis. The inactive derivative formed by modification with this reagent at the e-NH2 group of Lys-41, however, no longer binds nucleotides (25); RNase modified at the +NH2 group of Lys-41 by dinitrofluorobenzene, although inactive, still binds substrate analogues. Pares et ul. (26) have shown that at pH 5.5, 6-chloropurine 5 ’ ribonucleotide monophosphate and 8-bromoadenosine 5’-monophosphate bind specifically to RNase with affinities similar to those of 5’-AMP and 5’-GMP. At pH 7.3 and 40°, specific alkylation of the m-NHz group of Lys-1 was observed with a 60-fold molar excess of the 6-chloro derivative (27). The modified enzyme is only slightly less active than RNase. The authors suggest (28) that their results are compatible with a third basebinding site on RNase A. From X-ray studies on RNase complexes with analogues of UpA ( I ) and CpA ( 2 I ) , the binding is considered to have at least 5 centers, Bl, R1,pl, Rz, B2 (B and R for base and ribose). Since p u p binds more strongly than 2’(3‘)pU, Irie and associates ( 2 9 , 2 9 0 )propose the existence of a po site to account for the influence of the 5’-phosphate. The binding of ApUp and GpCp is stronger than that of GpC (30); Pares et al. (28) conclude that they are observing an extension of the binding site to include pz, R3, and B3 in the vicinity of Lys-I. Reductive alkylation of RNase A with formaldehyde (31-33) has been 25. Dudkin, S. M., Karabachyan; L. V., Borisova, S. N., Shyiapnikov, S. V., Karpeisky, M. Ya., and Geidarov, T. G. (1975). BBA 386, 275. 26. Pares, X., Arus, C . , Llorens, R . , and Cuchillo, C. M. (1978). BJ 175, 21. 27. Pares, X., Puigdomenech, P., and Cuchillo, C. M. (1978).Int. J . Peptide Protein Res. 16, 241. 28. Pares, X . , Llorens, R., Arus. C., and Cuchillo, C. M. (1980). EJB 105, 571. 29. Sawada, F., and Irie, M. (1969). J . Biochem. (Tokyo) 66, 415. 29a. Mitsui, Y., Urata, Y., Torri, K . , and Irie, M. (1978). BBA 535, 299. 30. White, M. D., Bauer, S., and Lapidot, Y. (1977). Nucleic Acids R e s . 4, 3029. 31. Means, G. E., and Feeney, R. E. (I%@. Biochemisfry 7,2192. 32. Pa&, W. K . , and Kim, S. (1972). Biochemistry 11, 2589. 33. Means, G. E. (1977). “Methods in Enzymology,” Vol. 47, p. 469.

324

PETER BLACKBURN AND STANFORD MOORE

applied with substitution of borohydride by cyanoborohydride, which is more specific for the reduction of Schiff bases and improves the efficiency of alkylation (34).Borohydride will reduce disulfide bonds and can cleave peptide linkages (35);cyanoborohydride does neither, and can be used at physiological pH (34). Cyanoborohydride tends to favor dialkylation of the protein NH2 groups (36). Reductive methylation of proteins permits the lysine residues to be studied by NMR, either from the proton resonances of the N-methyl groups (18, 19) or from the 13Cresonances of the N-methyl groups when [ 13C]formaldehydeis used for the modification (20, 34). Such studies with RNase A indicate that the fully modified protein retains its native conformation, and that Lys-41 is the only lysine residue that the chemical shift alters on binding of phosphate (18, 19) or 3’-CMP (20). Ligand binding was found to perturb the p K, of dimethylated Lys-41 (37). Feeney and his associates (36) have demonstrated the reversible reductive alkylation of RNase A and other proteins by use of the a-hydroxyaldehyde or ketone compounds, glycolaldehyde and acetol. Reductive alkylation of the monosubstituted amine is reversed by periodate oxidation to yield the primary amine; the dialkyl derivative is not labile to periodate oxidation. Bello et al. (38, 39) have studied the reaction of the arylating reagent 2-carboxy-4,6-dinitrochlorobenzene(CDNCB) with a number of model compounds and RNase A. The reagent reacts with imino, amino, and sulfhydryl groups; at pH 8.2 sulfhydryl groups react much faster than amino groups. With RNase A, CDNCB reacts preferentially with the e N H 2group of Lys-41 at 450 times the rate it reacts with the e N H 2group of a-N-acetyllysine. The CDNP-derivatives have absorption spectra typical of nitroanilines with A,, at 368-370 nm at pH 7.0, and at 345-350 nm in 0.1 M HCl. With RNase A, there was a small amount of product modified only on the a-NH2 group (7%); this product was fully active toward yeast RNA. Reaction at e N H 2 groups other than that of Lys-41 was not observed, even with a 6-fold molar excess of reagent over the enzyme concentration. The product modified at Lys-41 had only 0.6% of the activity of native RNase A. Reaction of CDNCB with N- 1-carboxymethyl-His-119-RNase A was much slower than with the native enzyme. The anionic carboxymethyl 34. 35. 36. 37. 38. 39.

Jentoft, N., and Dearborn, D. G. (1979). JEC 254, 4359. Crestfield, A. M., and Moore, S. (1963). JEC 246, 831. Geoghegan, K. F., Ybarra, D. M., and Feeney, R. E. (1979). Biochemistry 18,5392. Jentoft, J. E., Gerken, T. A., Jentoft, N., and Dearborn, D. G. (1981). JEC 256,231. Bello, J., Iijima, H . , and Kartha, G. (1979). Int. J . Peptide Protein Res. 14, 199. I&na, H . , Patrzyc, H., and Bello, J . (1977). EBA 491, 305.

12. PANCREATIC RIBONUCLEASES

325

group on His-119 is presumed to inhibit binding and/or orientation of CDNCB at the active site. The reaction at Lys-41 is inhibited by 2'(3')UMP (38). CDNP-RNase. has been crystallized in the presence of phosphate. X-Ray diffraction data were collected at 3 A resolution. The difference electron density map indicates no overall change in the protein conformation (38). The CDNP group is situated in the active site but does not directly occupy the pyrimidine or ribose binding sites; it is situated in the wider space leading to the substrate binding region, in the same general region as the DNP group of DNP-Lys-41-RNase S (40). The a-NH2 group is a weaker nucleophile than the e N H 2 group of Lys-4 1. The low p K , of the a-NH2group results from the inductive effect of the peptide carbonyl group. The e N H 2group of Lys-41 has a lowered pK, as a result of neighboring positive charges; Carty and Hirs (41) suggested that a neighboring arginine residue was responsible for this shift. Similarly, Migchelsen and Beintema (42) propose that the higher p K , values, obtained by proton NMR studies, for the active-site histidines of rat pancreatic RNase, as compared to those of bovine RNase A, result from substitution of Arg-39 in the bovine enzyme by Ser-39 in the rat enzyme. These authors (42) also suggest that this substitution explains the much lower rate of reaction of fluorodinitrobenzene with the e N H 2group of Lys-41 of the rat enzyme (43)compared to that of the bovine enzyme. Modification of Arg-39 and Arg-85 by kethoxal (3-ethoxy-2-ketobutanal) reduces the reactivity of Lys-41 to CDNCB (39). Modification of the guanidino groups of these residues by kethoxal lowers the pK, of the group to about 6; the decreased reactivity of Lys-41 with CDNCB at pH 7 and 8 is compatible with an increase in the p K , of the c-NH2 group of Lys-41 by about 1 pH unit (38, 39). Other positively charged groups near the active site include the e N H 2 groups of lysine residues 7 and 66. Walter and Wold (44) have acetylated RNase in the presence of an RNA digest and 2'(3')-CMP substrate analogues. They found that no single lysine residue was protected by the substrate analogues from acylation, and suggested that the sum of the residual amounts of lysine residues 7, 41, and 66 correlated fairly well with the residual enzymatic activity. Their acylation reactions were performed at 4" with an excess of acetic anhydride at pH 8.7 in the presence of 1 M sodium acetate plus 0.5 M borate buffer. 40. 41. 42. 43. 44.

Allewell, N. M . , Mitsui, Y., and Wyckoff, H. W. (1973). J5C 248, 5291. Carty, R. P., and Hirs, C. H. W. (1968). JBC 243, 5254. Migchelsen, C., and Beintema, J. J. (1973). J M B 79, 25. Gold, M . H. (1971). Ph.D. Thesis, SUNY at Buffalo, Buffalo, New York. Walter, B., and Wold, F. (1976). 5iuchemisfr.~15, 304.

326

PETER BLACKBURN AND STANFORD MOORE TABLE I THEPROTECTIVE EFFECTOF PoLY(A) ON THE AMIDINATION OF LYSINE RESIDUES IN RNASEA

Lysine residue Protection afforded (%)"

1 <5

7 51

31 25

37 36

41 100

61 100

66 <5

91

40

98 <5

104 43

" Calculated from the relative recovery of individual peptides and their overlap peptides, the sum of which equals 100%. From Blackburn and Gavilanes ( 4 4 ~ ) . Blackburn and Gavilanes (440) have studied the protective action of poly(A) toward amidination of lysine residues by methyl acetimidate at pH 8.5. Tryptic hydrolysis of the amidinated and performic acid oxidized protein, and separation of the peptides by reversed phase high-pressure liquid chromatography ( 4 9 , permitted identification of the protected lysine residues (Table I). Only two lysine residues were completely protected from amidination, Lys-41 and Lys-61. Other lysine residues were protected to different degrees. In the study by Walter and Wold (44), it was not possible to distinguish between modification of Lys-61 and Lys-66. The results of amidination in the presence of poly(A) show Lys-66 to be fully available. Amidination of lysine residues 7 (49%) and 66 (>95%) occurred with no loss of enzymatic activity; this result demonstrates that a modification of these residues, which retains a positive charge, does not affect activity. Lys-41, however, has a more sensitive role at the active site, since amidination (45-47) or guanidination ( I ) leads to inactivation. RNase A is inactivated by methylthioinosinedicarboxyaldehyde,the periodate oxidation product of /3-~-ribosyl-6-methylthiopurine(48). Inactivation results from formation of a Schiff base, presumably with Lys-41. This reagent is a potent antitumor and immunosuppressive agent that is presumed to act via formation of Schiff bases with essential amino groups of proteins. Inactivation of RNase A after reaction with N-acetoxy-2fluorenylacetamide (N-acetoxy-2-FAA), an activated metabolite of the carcinogen, N-hydroxy-2-fluorenylacetamide, has been reported (49). N-Acetoxy-2-FAA modifies the e N H 2 groups of the protein primarily by acetylation rather than by arylamidation, and results in the inactivation of the enzyme. 44a. Blackburn, P., and Gavilanes, J. G. (1982). JBC 257, ,316. 45. Blackburn, P., and Gavilanes, J. G. (1980). JEC 255, 10959. 46. Reynolds, J. H. (1968). Biochemistry 7, 3131. 47. Blackburn, P., and Jailkhani, B. L. (1979). JEC 254, 12488. 48. Spoor, T. C,, Hodnett, J. L., and Kimball, A. P. (1973). Cancer Res. 33, 856. 49. Barry, E. J., and Gutmann, H. R. (1973). JBC 248, 2730.

12. PANCREATIC RIBONUCLEASES

327

The differential reactivities of the amino groups of RNase A have been demonstrated by reaction with phthalyl-4-isothiocyanate (50). At p H 7.2 the reaction is specific for the w N H 2 group of the enzyme. The adjacent carboxyl groups of the phthalyl group act as a strong binding site for lanthanides. Bradbury et al. (50) propose to use this derivative in NMR studies of the NH2 terminus of RNase A. Garel (51) has coupled fluorescein-isothiocyanate to the a-NHzgroup of RNase A by reaction at pH 6. The pK, of the fluorescein group, initially at 6.2, is sensitive to conformational changes in the protein. The molar absorption coefficient for the derivative at 495 nm, is 72,000 cm-' A ! - ' ; the maximum change in molar absorbance at 495 nm upon titration of the fluorescein group, k 4 9 5 is 50,000 cm-' M - I . The change in €495 due to change of the p K, of the fluorescein reporter group has been used to study the binding of 2'-CMP and the thermal unfolding of the protein. Reaction of RNase A with the heterobifunctional reagent ethyl bromoacetimidate has been reported (52). At pH 9.0 rapid amidination of the amino groups occurs with inactivation of the enzyme as a result of reaction with Lys-41. As the pH is lowered, alkylation of a single histidine residue occurs, with cross-linking between Lys-41 and, primarily, His119. The pH optimum for this cross-linking alkylation is at pH 5.6, the same as that exhibited by the reaction of RNase A with haloacetates (cf., I, 53). The dimeric structure of the ribonuclease of bovine seminal plasma (see Section VI) and the cytotoxic properties thereof prompted studies on the preparation of cross-linked dimers of RNase A. The reaction between a diimido ester and the NH2 groups of the protein was studied (54, 55) in terms of the yield of a cross-linked dimer with maximum activity toward poly(A). poly(U). The reaction had been examined earlier by Hartman and Wold (56), primarily in reference to intramolecular cross-linking. Dimer formation was favored (55) at pH 7.7-8.0 at 21" with 1.25 equiv of dimethyl suberimidate and a protein concentration of 6%; the product obtained in 20% yield had 19 unmodified NH2 groups out of a theoretical 20 for a dimeric molecule, in which 2 such groups are involved in the crosslinking. The activity of the cross-linked dimer (55) toward poly(A). poly(U) in 0.14 M salt was 8.5 times that of the monomeric enzyme toward 50. 51. 52. 53. 54. 55. 56.

Bradbury, J. H.. Howell, J. R., Johnson, R. N . , and Warren, B. (1978). EJB 84,503. Garel, J.-R.(1976). WE 70, 179. Diopoh, J . , and Olomucki,M. (1979). Hoppe-Seyler's Z . Physiol. C h e m . 360, 1257. Plapp, B. V. (1973). JBC 248, 4896. Bartholeyns, J., and Moore, S. (1974). Science 186, 444. Wang, D., Wilson, G . , and Moore, S. (1976). Biochemistry 15, 660. Hartman, F. C., and Wold, F. (1967). Biochemistry 6, 2439.

328

PETER BLACKBURN AND STANFORD MOORE

the same substrate. The dirneric derivative has been studied in terms of its tumorostatic properties (57, 58). The cross-linking reaction has been extended to the preparation of poly-sperrnine-RNase (59) (see Sections VI and VIILB). Wang (60) prepared the bifunctional enzyme RNase-DNase by a procedure similar to that introduced by King and Kochoumain (61). The initial step with each enzyme was thiolation by reaction of NH2 groups with N-acetylhornocysteine thiolactone. The SH group added to DNase was covered by preparing the derivative in the presence of 4,4’dithiodipyridine. The cross-linkage was accomplished by thiol-disulfide interchange. The hybrid enzyme (yield 25-33%, based upon the DNase used) contained one molecule each of DNase and RNase bridge; the combination had 75 and 40% cross-linked by one - S - S of the activities of the parent enzymes, respectively, toward DNA and RNA. Reductive glycosarnination of lysine residues can be used to obtain derivatives with carbohydrate moieties attached to NH2groups. The conditions described by Gray (62) for the coupling of oligosaccharides to proteins in the presence of cyanoborohydride have been used by Wilson (63, 64) to attach lactose or rnannobiose to RNase cross-linked dimer. In 24 hr at pH 7 and 37”, in the presence of phosphate to protect Lys-41, the disaccharide was attached to an average of five lysine residues per dimer without significant decrease in enzymic activity. The disaccharides were chosen in order to obtain derivatives [e.g., N‘- 1-(l-deoxylactitolyl-lysproteins] with terminal galactopyranose or mannopyranose rings to study the respective selective uptakes (63, 6 4 ) of injected synthetic glycoconjugates by the known receptors for the given sugars in hepatocytes or cells of the reticuloendothelial system. Baynes and Wold (65) conducted uptake studies with the naturally occurring bovine RNases B, C, and D, which differ in the carbohydrate side chains attached to Asn-34 (66, 67). Biondi 57. Tarnowski, G. S., Kassel, R. L., Mountain, I. M., Blackburn, P., Wilson, G., and Wang, D. (1976). Cuncer Res. 36, 4074. 58. Bartholeyns, J., and Zenebergh, A. (1977). B t r . J . Currcer 15, 85. 59. Wang, D., and Moore, S. (1977). Biochemistry 16, 2937. 60. Wang, D. (1979). Biochemistry 18, 4449. 61. King, T. P., Li, Y.,and Kochoumian, L. (1978). Biochemistry 17, 1499. 62. Gray, G. R. (1974).A B B 163, 426. 63. Wilson, G. (1978). JBC 253, 2070. 64. Wilson, G.(1979). J . Gen. Ptiysiol. 74, 495. 65. Baynes, J. W., and Wold, F. (1976). JBC 251, 6016. 66. Plummer, T. H., Jr., and Hirs, C. H. W. (1%4). JBC 239, 2530. 67. Plummer, T. H . , Jr. (1968). JBC 243, 5961.

12. PANCREATIC RIBONUCLEASES

329

et 01. (68)acylated the e N H 2groups of RNase with gluconyl-glycine azide to introduce polyhydroxyalkyl side chains. 2. Histidin e Residues The initial studies on modification of the imidazole side chains of histidine residues of RNase A have been reviewed by Richards and Wyckoff (I). In an extension of the studies on the reaction of bromoacetate with the enzyme, Lennette and Plapp (69, 70) showed that carboxymethylation of N-3 of His-12 and N-1 of His-119 occurs, respectively, 120 and 1000 times faster than with the corresponding imidazole nitrogens of histidine hydantoin (70). The increased rate of alkylation of the active-site histidines of RNase A by bromoacetate, compared to the carboxymethylation of histidine hydantoin, corresponds to a difference in the free energy of activation (AC) of 3.2 to 4.2 kcaYmo1. This results primarily from a decreased enthalpy of activation (AH of 5.5 kcal/mol) and not from the entropy of activation (AS). From crystallographic studies on RNase S, the imidazole side chain of His-119 can occupy at least four different positions in the crystal structure, depending on the nature of ligands bound at the active site ( I ) . Lennette and Plapp (70) suggest that bromoacetate binds to a particular conformation of the enzyme and stabilizes the position of the imidazole of His-119, permitting a hydrogen bond between the /3-carboxyl of Asp-121 and N-3 of His-1 19. In this way, the nucleophilicity of N-1 of His-1 19 would be greatly increased by the inductive effect resulting from the hydrogen bond with N-3 of His-1 19. Such a hydrogen bond was postulated to exist by Sacharovsky et ( I / . (71) to explain the increased pK, values of the active-site histidine residues in des-( 121- 124)-RNase A, and has similarly been invoked by Antonov et al. (72)to fix the positively charged imidazole of His-119 into the catalytically active position in the complex between RNase S and 3' ,5'-2-deoxy-2-fluoro-UpA (73). Santoro et nl. (74) found no evidence for a hydrogen bond between 68. Biondi, L . , Filira, F., Giorrnani, V., and Rocchi, R. (1980). f n t . J . Pepptide Protein Res. 15, 253.

69. Lennette, E . P., and Plapp, B. V. (1979). Biochrmis/r.v 18, 3833. 70. Lennette, E. P., and Plapp, B . V. (1979). Eiorlirmistry 18, 3938. 71. Sacharovsky, V. G., Chervin, I. I., Yakovlev, G. I., Dudkin, S. M . , Karpeisky, M. Ya., Shliapnikov, S . V., and Bystrov, V. F. (1973). FEES Lett. 33, 323. 72. Antonov, I . V., Gurevich, A. Z., Dudkin, S. M . , Karpeisky, Y. Ma., Sacharovsky, V. G . , and Yakovlev, G. I. (1978). EJB 87, 45. 73. Pavlovsky, A . G . , Borisova, S. N . , Borisov, V. V., Antonov, I . V., and Karpeisky, M . Ya. (1978). FEES Lett. 92, 258. 74. Santoro, J . , Juretschke, H. P., and Ruterjans, H . (1979). BEA 578, 346.

330

PETER BLACKBURN AND STANFORD MOORE

Asp-121 and His-1 19 in their NMR studies of the carboxyl groups of RNase A. Walters and Allerhand (75)found that only His-119 exists in the Nc2-H (i.e,, N-3-H) tautomeric form most commonly found for nonhydrogen-bonded histidine residues. Their 13C NMR studies were performed in 0.2 M acetate. It seems unlikely that bromoacetate or iodoacetate binds to the active site of RNase A much differently than acetate. Indeed, acetate competitively inhibits the alkylation of the active-site histidine residues (70). The proximal positioning of the imidazole side chain of His-1 19 near the phosphate of dinucleotide phosphates appears to result primarily from binding of the base at the purine binding site ( 2 0 . While it is tempting to invoke a hydrogen bond between N-3 of His-119 and the P-carboxyl of Asp-121 to explain the nucleophilicity of N-1 of His-119, no direct NMR evidence for such a hydrogen bond exists. (See neutron diffraction studies, Section IV,A.) Alkylation of the active-site histidine residues of RNase A with iodoacetamide occurs preferentially with His-12, although the rate of alkylation is 7 times slower than that with iodoacetic acid (76). The rate of inactivation by iodoacetamide shows a bell-shaped pH dependence, with midpoints at pH 3.8 and 6.2. It has been proposed from NMR studies that RNase A undergoes a pH-dependent conformational transition (77-79) involving hydrogen bonds between the P-carboxyl group of Asp-14 and Tyr-25 or His-48 (74, 80), with the midpoints of the transitions at pH values 4.2 and 6.2. Such a conformational change, which affects the locale of His- 12, might explain the pH-dependent alkylation with iodoacetamide. The preferential alkylation of N-3 of His-12 by the neutral iodoacetamide (76)is consistent with the respective p K, values of the active-site histidine residues, 5.8 for His-12 and 6.2 for His-119 (Table VIII, Section IV,B). Clearly, the negative charge on the carboxyl group of iodoacetate or bromoacetate is an important factor in determining both the rate and selectivity of the alkylation of the active-site histidine residues of RNase A. Orientation and facilitation of alkylation of the active-site histidine residues by bromoacetate or iodoacetate does not seem to depend upon the positively charged e N H 2 group of Lys-41. Modification of Lys-41 with iodoacetate or iodoacetamide does not substantially inhibit the alkylation 75. Walters, D. E., and Allerhand, A . (1980). JBC 255, 6200. 76. Fruchter, R. G., and Crestfield, A. M. (1967). JBC 242, 5807. 77. Ruterjans, H., and Witzel, H. (1969). EBJ 9, 118. 78. Patel, D. J . , Camel, L. L . , and Bovey, F. A. (1975). Biopdymers 14, 987. 79. Cohen, J. S., and Shindo, H. (1975). JBC 250, 8874. 80. Lenstra, J. A . , Bolscher, B. G . J. M., Stob, S., Beintema, J . J . , and Kaptein, R. (1979). EJB 98, 385.

12. PANCREATIC RIBONUCLEASES

33 1

of the active-site histidine residues even though it does inactivate the enzyme (81). The selectivity of the alkylation of RNase A with bromoacetate at pH 5.5 decreases with prolonged times of reaction (1 to 42 days) (82). Alkylation of residues proceeds in the following sequence: (1) N-1 of His-119; (2) Met-30; (3) N-3 of His-12; (4) N-3 of His-105 and N-3 of N1-carboxymethyl-His-119;and ( 5 ) a-NH2of Lys-1. Both His-12 and His119 of the same active site are reported to be ultimately alkylated. Pincus and Carty (83) showed that treatment of RNase A with 2’(3’)0-bromoacetyluridine results in carboxymethylation of the active site histidine residues. The ratio of N-3-carboxymethyl-His- 12 to NI-carboxymethyl-His- 119 products is approximately 6 : 1 (83, 84). The carboxymethyl derivatives arise from the rapid hydrolysis of the parent uridine carboxymethyl RNase A ester (83). Binding of the nucleoside portion of the alkylating ligand orients the bromoacetyl moiety in such a way that it lies close to N-3 of His-12. The small amount of alkylation of N-1 of His-1 19 of RNase by 2’(3’)-0-bromoacetyluridinemay arise from hydrolysis of the ester bond generating some bromoacetate, which then preferentially alkylates His- 119. With 2‘-bromoacetamido-2’-deoxyuridine at pH 5.5, RNase A reacts rapidly with absolute specificity for N-3 of His-12 (85). A comparison of the rate constants for the alkylation of RNase A and free L-histidine with 2’(3’)-0-bromoacetyluridine,2‘bromoacetamido-2’-deoxyuridine,bromoacetic acid and iodo- and bromoacetamide (86) are shown in Table 11. Taking into account the difference in the relative rates of reaction of bromoacetamide and 2’(3’)-0bromoacetyluridine with free L-histidine, the enhancement of the rate of alkylation that results from nucleoside binding to the active site is about a factor of 25. The overall rate of reaction of 2’(3’)-0-bromoacetyluridine with RNase A is 3100 times the rate with free L-histidine; thus, the nucleophilicity of the imidazole N-3 of His-12 is increased approximately 120-fold (86), a value similar to that obtained by Lennette and Plapp (70). Ferrate ion is a powerful oxidant and an analog of phosphate (87); ferrate has been demonstrated to inactivate RNase A as a result of reaction at His-119 (88). The specificity of the reaction is pH-dependent. At 81. 82. 83. 84. 85. 86. 87. 88.

Heinrikson, R. L. (1966). JBC 241, 1393. Bello, J . , and Nowoswiat, E. F. (1971). EJB 22, 225. Pincus, M . , and Carty, R. P. (1970). BBRC 38, 1049. Machuga, E . , and Klapper, M. H . (1975). JBC 250, 2319. Lan, L. T., and Carty, R . P. (1972). BBRC 48, 585. Pincus, M . , Thi, L. L . , and Carty, R. P. (1975). Biochemistry 14, 3653. Lee, Y. M . , and Benisek, W. F. (1976). JBC 251, 1553. Steczko, J . , Walker, D. E . , Hermodson, M., and Axelrod, B. (1979). JBC 254, 3254.

332

PETER BLACKBURN AND STANFORD MOORE TABLE I1 COMPARISON OF THE KINETIC CONSTANTS FOR T H E ALKYLATION OF RNASEA A N D FREEL-HISTIDINE’ Reaction

2’(3’)-O-Bromoacetyluridine+ RNase A 2’(3’)-O-Bromoacetyluridine+ L-histidine 2’-Bromoacetamido-2‘-deoxyuridine + RNase A 2’-Bromoacetamido-2’-deoxyuridine + L-histidine Bromoacetic acid + RNase A Bromoacetic acid + L-histidine Iodoacetamide + RNase A lodoacetamide + L-histidine Bromoacetamide + RNase A Bromoacetamide + r-histidine

(x

k3/ Ka or krpObs lo4M - ’ sec-’)

403 0.I29 518 0.113 89.0 0.037 0.48 0.0052 1.9 0.015

From Pincus et cil. (86) and references therein. [Reprinted with permission from Biochemistry 14, 3653-3661. Copyright (1975)American Chemical Society.]

pH 5.0 the reaction is specific for His-119; at pH 7.0 two tyrosine and two lysine residues were also modified. The enzyme is protected by active-site ligands from ferrate inactivation. Reaction of His-I 19 with ferrate was followed by specific titration of the histidine residues with diethylpyrocarbonate and measurement of the absorption of the product, ethoxyformylhistidine, at 240 nm (89-91 ). RNase A forms stable charge-transfer complexes between chloropentammineruthenium”’ dichloride and the imidazole moiety of histidine residues (92). Based on the kinetics of incorporation, three of the four histidine residues are reactive; the three derivatives that contain a single ruthenium-histidine complex were separable from other reaction products by ion exchange chromatography (93). These complexes have absorption bands arising from charge-transfer transitions between the imidazole and ruthenium ion, which depend upon the state of protonation of N-1 of the histidine; the protonated form has absorption maxima at 303 and 450 nm, the unprotonated form has absorption maxima at 365 and 600 nm, and the pK, of N-1 of the imidazole in the complex is 8.8 (94). His-105 89. Ovadi, J., Libor, S.,and Elodi, P. (1967).Artu Biochim.Biophys.Acad. Sri. Hung. 2, 455.

90. Melchior, W.B., Jr., and Fahrney, D. (1970).Biochemistry 9, 251. 91. Roosemont, J. R. (1978).Anal. Biochern. 88, 314. 92. Matthews, C. R., Erickson, P. M., Van Vliet, D. L., and Petersheim, M. (1978). JACS 100, 2260. 93. Matthews, C. R., Erickson, P. M., and Froebe, C. L. (1980).BBA 624, 499. 94. Sundberg, R. J., and Gupta, G. (1973).Biuinorg. Chem. 3, 39.

12. PANCREATIC RlBONUCLEASES

333

is the predominant singly labeled derivative; the change in extinction at 370 nm of this derivative during thermal- and urea-induced unfolding of the protein has been followed (93). This derivative had 66% of the activity of native RNase A toward cyclic 2',3'-CMP; the other two singly modified derivatives had 89% and 53% activity, respectively (93). The effects of the paramagnetic ruthenium on the NMR spectrum of the histidine C-2 and C-4 protons have been used to confirm the assignments of these resonances in the derivative labeled at His-105 ( 9 5 ) , and have similarly confirmed His-105 and His-48 as the two histidine residues not fully exposed to solvent in thermally unfolded RNase A (96, 97). The paramagnetic hexacyanochromate [CI-(CN)~]~ion binds to the active site of RNase A and selectivity broadens the C-2 and C-4 proton resonances of His-12 and His-119 (98). 3. Arginine Residues Takahashi (99) reported that 2 to 3 arginine residues of RNase A were modified by reaction with phenylglyoxal at neutral pH and room temperature. The principal residues modified were Arg-39 and Arg-85; activity loss correlated closely with modification of Arg-39. Residues Arg-10 and Arg-33 were generally unreactive. The facilitated alkylation of the activesite histidine residues by iodoacetate was largely unaffected by the modification of the arginine residues. Yankeelov (100) reported that after more than three of the four arginine residues of RNase A were modified by oligomers of 2,3-butanedione, 45% of the activity of the native enzyme, measured at pH 5.2 with RNA as substrate, was retained. Patthy and Smith (101, 102) found that reaction of RNase A with 1,2-cyclohexanedione in borate buffer at pH 8 to 9 results in modification of 2 to 3 arginine residues and 90% loss of enzymatic activity, measured at pH 7.2 with RNA as substrate. Arg-39 reacts rapidly and its modification contributes mostly to the loss of enzymatic activity. Arg-85 also reacts rapidly, Arg-10 reacts slowly, and no reaction was observed with Arg-33. Removal of the blocking groups with hydroxylamine results in complete recovery of enzymatic activity, except when Arg-10 has also been modified, in which case recovery is 80%. Similar results on the loss of enzy95. Matthews, C. R . , Recchia, J . , and Froebe, C. L. (1981). A n d . Biochem. 112, 329. 96. Matthews, C. R . , and Westmoreland, D. G . (1975). Biochemistry 14, 4532. 97. Matthews, C. R., and Froebe, C. L. (1981). Mncromolecules 14, 452. 98. Inagaki, F., Watanabe, K., and Miyazawa, T. (1979). J . Biochem. 86, 591. 99. Takahashi, K. (1968).JBC 243,6171. 100. Yankeelov, J. A., Jr. (1970). Biochemistry 9, 2433. 101. Patthy, L., and Smith, E. L. (1975). JBC 250, 557. 102. Patthy, L., and Smith, E. L. (1975). JBC 250, 565.

334

PETER BLACKBURN AND STANFORD MOORE

matic activity toward cyclic 2‘,3’-CMP at pH 6.5, and the number of arginine residues modified upon reaction of RNase A with 1,2cyclohexanedione were observed by Blackburn and Jailkhani (47);reaction with 2,3-butanedione, under conditions described by Riordan (103), modified four arginine residues resulting in 95% loss of enzymatic activity (47). Kethoxal (3-ethoxy-2-ketobutanal) reacts with the guanidino groups of arginine residues and to some extent with e N H 2groups of lysine residues (39). The predominant sites of modification were Arg-39, Arg-85, and 85% of Arg-10; no reaction occurred with Arg-33. With two arginine residues modified, Arg-39 and Arg-85, the derivative had 90% of the activity of native RNase A toward RNA at pH 5.0, but only 20 to 25% of the activity of native RNase A toward cyclic 2’,3‘-CMP at pH 7.0. Modification of Arg-10 results in further loss of activity. The pH dependence of the activity of RNase A modified with kethoxal relative to the activity of native RNase A exhibits a titration curve with a midpoint at pH 5.8 in 0.1 ionic strength buffer. Modification of the guanidino group of a-N-acetylarginine with kethoxal lowers the pKa of the modified guanidino group to about 6.0. Modification of the arginine residues by kethoxal had little effect on carboxymethylation of the active-site histidine residues with bromoacetate or with 2’-O-bromoacetyluridine. Reaction of the r-NH2 group of Lys-41 in kethoxal-modified RNase by the arylating reagent 2carboxy-4,6-dinitrochlorobenzeneoccurs at 25 and 2% of the rate with the native enzyme at pH 8.5 and 7.5, respectively, corresponding to an increase in pKa of the r-NH2 group of Lys-41 by 1 pH unit (38, 39). The differences observed in the effects of arginine modification on the activity of RNase, according to Iijima et al. (39),may be explained by the different pH values chosen for assay of the various derivatives. It is possible that the pH dependence of the relative activity of kethoxal-modified RNase A with respect to native RNase is related to the lowered pK, of the modified guanidino groups. Similarly, extensive modification of RNase A with phenylglyoxal yields a protein with a much lower isoionic point, with little or no migration exhibited at pH 6.5 (99). Also, the derivative produced by reaction of arginine with 2,3-butanedione elutes from a short column of sulfonated polystyrene ahead of lysine at pH 5.25. close to the breakthrough of this buffer (103). These results suggest that the pKa of the guanidino groups modified with phenylglyoxal and a,P-diketones is most likely similar to that obtained after modification with kethoxal, i.e., close to 6.0. Iijimaet al. (39) suggest that the decreased activity of kethoxal-modified 103. Riordan, J. P. (1973). Biochemistry 12, 3915.

12. PANCREATIC RIBONUCLEASES

335

RNase A results from the modification of both Arg-39 and Arg-85, and not simply from modification of Arg-39. Arg-39 and Arg-85 are situated near the active site of the enzyme ( 1 ) . Arg-85 is invariant among the species of pancreatic RNases studied (Section V), with the exception of the mouse enzyme in which it is substituted by a histidine residue. Arg-39 is also generally conserved, but is substituted by serine in mouse and rat enzymes, by tyrosine in the muskrat and chinchilla, by lysine in bovine seminal plasma RNase, and is deleted in the RNases of the dromedary and bactrian camels and the horse. The activities of three of these enzymes toward cyclic 2',3'-CMP relative to that of the bovine enzyme (45) are: Mouse, 0.24; rat, 0.52; dromedary, 1.18. The enzyme that exhibits the lowest specific activity has substitutions at both Arg-39 and Arg-85. 4. Aspartate and Asparagine, Glutamate and Glirtamine Residues RNase A has 11 carboxyl groups: 5 aspartate, 5 glutamate, and the a-carboxyl of the terminal Val-124. Modification of these residues in general leads to decreased enzymatic activity, which appears to result from conformational changes that occur with progressive modification of the carboxyl groups ( I ) . Potential interactions involving carboxyl groups, the modification of which could lead to a loss of activity, include the hydrogen bonds between Asp-14 and His-48 or Tyr-25 (74, 7 7 4 0 ) , Glu-2 and Arg-10 (104-106), Val-124 and His-105 ( 7 3 , and possibly Asp-121 and His-119 (69, 71 -73). Glu-1 11 hydrogen bonds to N-1 of the purine base of dinucleotide substrates (2f). Esterification occurs most rapidly at Asp-53 ( 1 07) and Asp-49 (108, 109). The derivative esterified with methanol at both of these residues exhibits no gross conformational change, but the product has only 65% of the activity of the native enzyme toward cyclic 2' ,3'-CMP (109). Although the K, is unaltered, after limited digestion by subtilisin the dimethyl S-protein exhibits a fourfold weaker interaction with the S-peptide, which may explain the reduced activity (109). Modification of Asp-53 with diazoacetyl glycinamide has no effect on the enzymatic activity (107). In 104. Marchiori, F., Borin, G., Moroder, L., Rocchi, R . , and Scoffone, E. (1972). BBA 257, 210. 105. Hofmann, K . , Visser, J. P., and Finn, F. M. (1970). JACS 92, 2900. 106. Hofmann, K . , Andreatta, R., Finn, F. M., Montibeller, J . , Porcelli, G . , and Quattrone, A. J . (1971). Bioorg. Ckem. 1, 66. 107. Riehm, J . P., and Scheraga, H. A. (1965). Biorhemistry 4, 772. 108. Acharya, A. S . , and Vithayathil, P. J. (1975). I n t . J . Pepride Protein R e s . 7 , 207. 109. Acharya, A. S., Manjula, B . N . , and Vithayathil, P. J. (1978). EJ 173, 821.

336

PETER BLACKBURN AND STANFORD MOORE

the presence of unmodified S-protein, inactive RNase methylated at five carboxyl groups (I 10) exhibits 30% of the activity of the native enzyme; esterification of the carboxyl groups leads to a weaker interaction between the S-peptide and S-protein segments of the RNase molecule and permits hybrid formation. The derivative esterified at Asp-49 and Asp-53 does not show increased activity in the presence of added S-protein (109). In strong acid, RNase A undergoes deamidation. A monodeamidated product accumulates, designated RNase Aa,; subsequent deamidation occurs more slowly (I If). The monodeamidated product has full enzymatic activity; further deamidation is associated with loss of enzymatic activity. Hydrolysis of the monodeamidated derivative with subtilisin gives only 50% conversion to monodeamidated RNase S (I I2). Tryptic hydrolysis and peptide mapping indicates that one of the four amides in the peptide, corresponding to residues 67 through 8 5 , is the primary site of deamidation (111). Spectroscopic and immunological techniques indicated that RNase Aal has a conformation close to that of native RNase A (I I1, 113). The conformation of RNase Aal is, however, less thermostable than that of RNase A, as judged by susceptibility to trypsin and reaction of methionine residues with o-benzoquinone (I 14). 5 . Methionine Residues

RNase A has four methionine residues at positions 13, 29, 30, and 79, which in general are invariant among pancreatic enzymes of different species, except for a few cases where they are conservatively substituted by valine, isoleucine, or leucine. They are buried in the interior of the protein and cannot usually be modified unless the molecule is denatured (I). Alkylation of RNase A with methyl iodide at low pH specifically modifies Met-29, with no effect on the enzymatic activity (115); in 8 M urea at low pH all four methionine residues are alkylated (I I 6 ) . Separation of the reaction products and identification of the sites of modification showed that alkylation at Met-29 or Met-79, or both Met-29 and Met-79 yield fully enzymatically active species upon removal of urea. Their re110. Acharya, A. S., Manjula, B. N., Murthy, G. S., and Vithayathil, P. J . (1977). Inr. J. Peptide Prorein Res. 9, 213. 1 1 1 . Manjula, B . N . , Acharya, S. A . , Vithayathil, P. J . (1976). Inr. J. Peptide Protein Res. 8, 275. 112. Manjula, B . N . , Acharya, A. S. , and Vithayathil, P. J. (1977). BJ 165, 337. 113. Das, M . K . , and Vithayathil, P. J . (1978). BBA 533, 43. 114. Das, M. K . , and Vithayathil, P. J . (1978). i n r . J . Peptide Protein Res. 12, 242. 115. Link, T. P., and Stark, G. R. (1968). JBC 243, 1082. 116. Stark, G. R . , and Link, T. P. (1975). Biochemistry 15, 3476.

12. PANCREATIC RIBONUCLEASES

337

spective transition temperatures are 58", 43",and 36", compared to 63" for native enzyme. Methylation of Met-13 or Met-30 prevents refolding to an active conformation. Alkylation of methionine converts the residue to a positively charged sulfonium ion and introduces an alkyl group on the side chain, which can apparently be accommodated only at Met-29 and Met-79. Upon photoxidation of RNase A in the presence of hematoporphyrin in 10% acetic acid, only Met-29 is oxidized to the sulfoxide; at higher concentrations of acetic acid (10 to 50%) Met-13 is also oxidized (117). As with the alkylated derivatives, oxidation at Met-29 does not inactivate the enzyme; the derivative oxidized at both Met-29 and Met-13 had low activity. Modification of Met-13 in S-peptide by oxidation to the sulfone, or alkylation with iodoacetate or iodoacetamide to produce a sulfonium derivative, significantly reduces the strength of binding between S-peptide and S-protein; however, the complexes, when formed, are fully active (1, 118). Richards and Wyckoff suggested that rotation about the a-/3 carbon-carbon bond of Met-13 can allow the charged sulfur atom to have access to solvent without a change in the conformation of the S-peptide (1). In RNase A, the conformation of the S-peptide is more constrained because of the unbroken peptide bond between residues 20 and 21 (119), and it is less likely that the charged sulfur on Met-13 can be similarly accommodated (116). The reaction of o-benzoquinone with methionine at acidic pH produces the 3,4-dihydroxyphenylmethioninesulfonium salt (120).The accessibility of the methionine residues to o-benzoquinone has been used as a probe of protein conformation after esterification of the carboxyl groups of RNase A and RNase S (110). Modification with this reagent introduces a chromophore that enables the number of modified methionine residues to be determined spectrophotometrically (120-122). 6. Disuijide Bonds The four intramolecular disulfide bonds of RNase A are invariant features of the sequence among all of the mammalian pancreatic RNases; they are between residues 26 and 84,40 and 95,58 and 110, and 65 and 72. These four intrachain disulfide bonds contribute to the overall conforma117. Jori, G., Galiazzo, G., Tamburro, A. M., and Scoffone, E. (1970). JBC 245, 3375. 118. Vithayathil, P. J . , and Richards, F. M. (1960). JBC 235, 2343. 119. Carlisle, C . H . , Palmer, R. A . , Mazumdar, S . K., Gorinsky, B. A . , and Yeates, D. G. R. (1974). J M B 85, 1 . 120. Vithayathil, P. J . , and Murthy, G. S. (1972). Nature N e w B i d . 236, 101. 121. Gupta, M. N . , and Vithayathil, P. J. (1980). f n t . J . Peptide Protein Res. 15, 236. 122. Gupta, M. N., Murthy, G. S . , and Vithayathil, P. J. (1980). Znt. J . Peptide Protein Res. 15, 243.

338

PETER BLACKBURN AND STANFORD MOORE

tional stability of the RNase molecule without determining the final overall conformation of the protein (123). Reduction of the disulfides between residues 65 and 72, and 58 and 110 (124-126) and protection of the sulfhydryls with phosphorothioate (126) yields a species with full activity toward RNA and enhanced activity toward cyclic 2’,3’-CMP. Spectrophotometric titrations (126) and CD measurements (127) show little difference from native conformation. Incubation of RNase A with 0.1 M mercaptoethanol at pH 8.5, upon exposure to the air over a period of days, yields a mixture of chromatographically separable, metastable products with enhanced activity toward cyclic 2’,3’-CMP (128). The fraction with the highest activity (fourfold) had no detectable free sulfhydryl groups; the protein has probably formed mixed disulfides with mercaptoethanol. Under nondenaturing conditions, at pH 8.7 and 25”, Creighton (129) found the half-life for the reduction of native RNase A by 10 mM dithiothreitol to be approximately 10 hours. The primary product was fully reduced RNase A, with no significant accumulation of species with 1,2, or 3 disulfide bonds. Sperling et al. (130) studied the reduction of the disulfide bonds with both dithiothreitol and dithioerythritol. After 1 hour at pH 8.0, with a fivefold molar excess of reducing agent over disulfide, partial reduction of RNase A occurs at the disulfide bond between residues 65 and 72 with no loss of enzymatic activity. The ability to selectively and partially reduce the disulfide bonds of RNase A depends upon the nature and concentration of the reducing agent, the pH, and the folded state of the protein. The ability to form linear mixed disulfides can be important for production of species with enhanced enzymatic activity. Dithiothreitol has little tendency to form mixed disulfides, since its greater reducing potential arises from its ability to form a thermodynamically stable intramolecular, disulfide-linked, six-member ring (131). In the presence of 6 M guanidinium chloride, at pH 8.5 and 25”, RNase 123. Scheraga, H. A. (1980). I n “Protein Folding” (R. Jaenicke, ed.), p. 261. Elsevier, Amsterdam. 124. Sela, M . , White, F. H., Jr., and Anfinsen, C. B. (1957). Science 125, 691. 125. Resnick, H., Carter, J . R . , and Kalnitsky, G. (1959). JBC 234, 1711. 126. Neumann, H., Steinberg, I. Z., Brown, J . R. Goldberger, R. F., and Sela, M. (1967). EJB 3, 171. 127. Tamburro, A. M., Boccu, E., and Celotti, L. (1970). Int. J . Peptide Protein Res. 2, 157. 128. Watkins, J. B., and Benz, F. W. (1978). Science 99, 1084. 129. Creighton, ‘r. E. (1979).J M E 129, 411. 130. Sperling, R., Burstein, Y., and Steinberg, I. 2. (1969). Biochemistry 8, 3810. 131. Cleland, W. W. (1964). Biochemistry 3, 480.

12. PANCREATIC RIBONUCLEASES

339

A is completely reduced by 20 mM 0-mercaptoethanol. Lower concentrations of P-mercaptoethanol only bring about partial reduction of the disulfide bonds under these conditions (132). After net reduction of one disulfide bond and S-alkylation with N-ethylmaleimide, all enzymatic activity is lost. The sulfhydryl groups of reduced RNase react with selenium (133) and mercury (130, 1-34), which can be incorporated into (and elongate) the disulfide bonds. Elongation of the disulfide bond between residues 65 and 72 has no effect on the enzymatic activity (130). Sperling and Steinberg (134)introduced a single mercury atom into all four disulfides of RNase A; the derivative had 25% of the activity of native enzyme toward cyclic 2’,3’-CMP, and 5% toward RNA. The re-formation of RNase disulfides during refolding of the protein is dealt with in Section IV,D. 7. Tvrosine Residues RNase A has six tyrosine residues at positions 25, 73, 76, 92, 97, and 115 in the sequence (15). Early spectrophotometric titrations at alkaline pH suggested that three of the residues titrate with normal pK, values of about 10.2 and three titrate with abnormal ply, values above 12; these phenolic groups were described as “accessible” and “buried,” respectively (1). Subsequent proton (80) and 13C NMR (135) titration studies indicate only two tyrosine residues with abnormal p K, values. Moreover, the X-ray crystallographic data for RNase S ( 1 ) show that only two tyrosine residues are truly buried, residues 25 and 97. Iodination of RNase A permitted identification of the normal and abnormal tyrosine residues (1). Tyr-25 and Tyr-97 are not iodinated; Tyr-92 is iodinated at pH 9.4 (136, 137) but not at pH 6.4 (138), when only residues 73, 76, and 115 are iodinated. Tyr-92, although situated on the surface of the molecule (I), titrates with an abnormally high pK,; its hydroxyl is involved in a hydrogen bond with the carbonyl group of the amide bond between Lys-37 and Asp-38. A local conformational change in the protein may break this hydrogen bond and lead to “normalization” of the behavior of Tyr-92. For example, in horse pancreatic RNase, deletion of residue 39 introduces steric constraints that do not permit the 132. Garel, J.-R. (1977). FEBS Lett. 79, 135. 133. Ganther, H . E . , and Corcoran, C . (1969). Biochemistry 8, 2557. 134. Sperling, R . , and Steinberg, I. Z. (1971). JEC 246, 715. 135. Egan, W., Shindo, H . , and Cohen, J. S. (1978). JBC 253, 16. 136. Cha, C.-Y., and Scheraga, H . A. (1963). JEC 238, 2958. 137. Cha, C.-Y.,and Scheraga, H . A. (1963). JBC 238, 2965. 138. Woody, R . W., Friedman, M . E., and Scheraga, H. A. (1966). Eiodwmistry 5, 2034.

340

PETER BLACKBURN AND STANFORD MOORE

hydrogen bond to involve Tyr-92; Tyr-92 of this protein has a normal pK, (139). Acetylation of the tyrosine residues at pH 7.5 with acetic anhydride or N-acetylimidazole prevents their cleavage by N-bromosuccinimide. By this approach Burstein and Patchornik (140) found that residues 73, 76, and 115 were available to acylation and residues 25, 92, and 97 were unavailable. The availability of the tyrosine residues to modification reactions depends upon the protein’s conformation. In des-( 121- 124)-RNase A, Fujioka and Scheraga (141) found Tyr-25 unavailable to iodination at pH 9.5. The peptide mapping data were not definite on the availability of Tyr-92 and Tyr-97. The CD meaururements of Taniuchi (142) and Puett (143) in the near-UV indicated only a small difference between the conformation of the aromatic residues in RNase A and des-(121-124)-RNase A. Ultraviolet absorption spectra on the native and denatured protein suggest between two and three “buried” tyrosine residues in des-(121124)-RNaseA and two in des-(119-124)-RNase A (144).Tyr-97 is the least accessible tyrosine residue of RNase A; exposure of this residue would require a substantial change in the protein conformation (1 ). Most probably, it is Tyr-92 that is exposed upon removal of the last 4 to 6 carboxylterminal residues of RNase A. Nitration of tyrosine residues with tetranitromethane (145, 146) has been applied to RNase A (146, 147). The modification reaction can be followed spectrophotometrically (146) and introduces a nitro group orrho to the phenolic hydroxyl, lowering its p K, from 10.2 to about 6.8. Reduction of the 3-nitro group to the amine with bisulfite raises the p K, of the hydroxyl to about 10; the pK, of the aromatic amine is about 5 (148). At pH 8, 3.1 tyrosine residues were reported to be nitrated, based on the extinction at 428 nm; 2.6 nitrated tyrosine residues were indicated by amino acid analysis (146).Beaven and Gratzer (147) reported that a maximum of three tyrosine residues could be nitrated; the product remained enzymatically active. Reaction of proteins with tetranitromethane can lead 139. 140. 141. 142. 143. 144. 145. 146. 147. 148.

Scheffer, A . J., and Beintema, J. J . (1974). E J E 46, 221. Burstein, Y.,and Patchornik, A. (1972). Biochemisrr.~11, 2939. Fujioka, H . , and Scheraga, H. A. (1965). Eiocliemistry 4, 2206. Taniuchi, H. (1970). JEC 245, 5459. Puett, D. (1972). Bioc/iemisrry 11, 1980. Puett, D. (1972). Biochemistry 11, 4304. Riordan, J . F., Sokolovsky, M., and Vallee, B. L. (1966). JACS 88, 4104. Sokolovsky, M . , Riordan, J. F.. and Vallee, B. L. (1966). Biochemistry 5, 3582. Beaven, G. B., and Gratzer, W. B. (1%8). EEA 168,456. Sokolovsky, M., Riordan, J. F., and Vallee, B. L. (1967). BBRC 27, 20.

12. PANCREATIC RIBONUCLEASES

34 1

to cross-linking (149, I N ) , resulting in poor correlation between spectrophotometric measurements and amino acid analyses; in that case, the sum of Tyr + 3-nitro-Tyr is less than the total tyrosine content of the native protein. The mechanism of the nitration reaction involves the phenoxide ion and proceeds via formation of free radicals that lead to cross-linking (151) and possibly other side reactions (152, 153). Thus, interpretation of the results of the reaction of proteins with tetranitromethane requires careful characterization of the reaction products. Garel and Baldwin (154)found that nitration of RNase followed first-order kinetics, indicating similar reactivities for the reacting groups. About 10% of cross-linked material was found using the procedure according to Sokolovsky er a / . (146). The monomeric fraction had 2.7-2.8 3-nitro-Tyr residues per molecule; the modified protein was similar to the native enzyme with respect to T,, K,,,, and kcat values toward cyclic 2',3'-CMP. Van der Zee r f 01. (155) have purified, by isoelectric focusing, two species from the reaction products of tetranitromethane with RNase A. One was nitrated at Tyr-115, the other at Tyr-76 and Tyr-115; both were fully active toward cyclic 2',3'-CMP. Seagle and Cowgill (156) reported on the nitration of RNase A and conversion of the nitro groups to the amine. The procedure described by Beaven and Gratzer (147) was used to modify three tyrosine residues under nondenaturing conditions. The same procedure under denaturing conditions permits modification of all six tyrosine residues. Prior acetylation (157) of the three normal tyrosine residues enabled only the three abnormal tyrosine residues to be modified (156). Seagle and Cowgill (156) report that after only one net tyrosine residue was nitrated and reduced to the amine, the product had 70% enzymatic activity toward cyclic 2',3'-CMP. The fluorescence emission maximum of this derivative was at 395 nm, higher than values observed for aminotyrosyl residues in peptides (I%),and suggests this aminotyrosyl residue is located in an unusual environment (156). The products of the nitration reaction were not analyzed for possible cross-linking; under the conditions of Beaven and Gratzer (147) as used by Seagle and Cowgill (156), this un149. 150. 151. 152. 153. 154. 155. 156. 157. 158.

Vincent, J. P., Lazdunski, M., and Delaage, M. (1970). EJE 12, 250. Hugli, T. E., and Stein, W. H . (1971). JBC 246, 7191. Bruice, T. C., Gregory, M . J., and Walters, B. L. (1968). JACS 90, 1612. Walters, S. L., and Bruice, T. C. (1971). JACS 93, 2269. Jewett, S. W., and Bruice, T. C. (1972). Biochemistry 11, 3338. Garel, J.-R., and Baldwin, R. L. (1975). J M B 94, 621. van der Zee, R.,Duisterwinkel, F. J . , and Welling, G . W. (1977). EJB 77, 125. Seagle, R. L . , and Cowgill, R . W. (1976). BBA 439, 470. Riordan, J. F., Wacker, W. E. C., and Vallee, B. L. (1965). Biochemistry 4, 1758. Seagle, R. L., and Cowgill, R. W. (1976). BBA 439, 461.

342

PETER BLACKBURN AND STANFORD MOORE

doubtedly occurred. Since van der Zee et al. (155) demonstrated Tyr-115 as a sole site of nitration in one of their reaction products, perhaps Tyr-115 is the 3-amino-Tyr residue reported by Seagle and Cowgill (156) to be in an unusual environment. Van der Zee et al. (155) suggest that this result may arise from the interaction between the neighboring side chains ( 1 ) of modified Tyr-115 and unmodified Tyr-73. These authors also observe that correct spectrophotometric quantitation of 3-nitro-Tyr can be best obtained at 381 nm between pH 4 to 10, where the absorption spectrum exhibits an isosbestic point (146). The diazonium salt of 5’-(4-aminophenylphosphoryl)uridine 2’(3’)phosphate reacts stoichiometrically with RNase A and modifies only Tyr-73 (159). The reaction is inhibited by the competitive inhibitor cytidine 2’(3’), 5‘-diphosphate. The modification does not affect the activity toward RNA but weakens 40-fold the binding constant toward cyclic 2‘ ,3’-CMP. By contrast, reaction withp-diazophenyl phosphate modifies 2 lysine, 1 histidine, and 3 tyrosine residues of RNase A (159). The peptide modified by reaction with 4-diazophenyl-pup is isolated in good yield from a tryptic digest of the reduced S-carboxymethylated RNase by affinity chromatograppy on Sepharose-RNase A. The K , values for 4-nitrophenyl-pup and 4-aminophenyl-pup determined by the inhibition of hydrolysis of cyclic 2’,3’-CMP at pH 7.0 and 22” are 43 and 70 p M , respectively (159). 8. Reactions wit17 Radicals (1. Radiolysis. In aqueous solutions RNase is inactivated by hydrogen atoms generated externally (160, 161), by steady (162) or pulse radiolysis (163, I @ ) , and by OH radicals and hydrated electrons generated by pulse radiolysis (165). The chemical evidence (160-162), and that from pulse radiolysis kinetic spectroscopy (163-165), indicate the occurrence of intramolecular radical chain reactions with cystine and methionine linked with modification of tyrosine and possibly phenylalanine residues (160166). Reaction with OH radicals leads to greater damage of surface residues and is associated with extensive cross-linking of the protein. Di-

159. Gorecki, M., and Wilchek, M. (1978). BBA 532, 81. 160. Holmes, B. E., Navon, G., and Stein, G. (1%7). Nature (London) 213, 1087. 161. Shapira, R., and Stein, G. (1968). Science 162, 1489. 162. Mee, L. K . , Adelstein, S. J., and Stein, G. (1971). Radiat. Res. 47, 349. 163. Mee, L. K., Adelstein, S. J., and Stein, G. (1972). Radiat. Res. 52, 588. 164. Lichtin, N . N . , Ogdan, J., and Stein, G. (1971). BBA 263, 14. 165. Lichtin, N. N . , Ogdan, J., and Stein, G. (1972). BBA 276, 124. 166. Stein, G. (1968). In “Energetics and Mechanisms in Radiation Biology” (G.0. Phillips, ed.), p. 407. Academic Press, London.

12. PANCREATIC RIBONUCLEASES

343

merization is essentially absent upon H atom attack. The cross-linking probably arises from the abstraction of H atoms from the saturated a-carbon atom of the polypeptide backbone, and accounts for 20% of the action of OH radicals with RNase (165). The kinetics of the transitions in the absorption spectra of RNase A and RNase S-protein upon reaction with H atoms differ both qualitatively and quantitatively at different wavelengths of the spectra. Thus, the intramolecular free radical chain reactions in the two proteins are different and reflect the different conformational flexibilities of the proteins. The reactions in RNase A differ from the sum of those observed spectrally in S-peptide and S-protein (167). b. UV Irradintion. Irradiation of anaerobic solutions of RNase A at pH 5.0 and 4" with UV light at 254 nm causes inactivation of the enzyme. Inactivation correlates with the destruction of disulfides (168-171); histidine is not affected (169). The loss of activity is associated with generation of free suIfiydry1, but always less than 2 moles per mole of disulfide disrupted (171); the remaining cysteine is lost via secondary reactions generating primarily H2S and aldehyde (160, 161, 166, 172, 173). The photodestruction of RNase A disulfides is nonrandom (172); rates differ by at least a factor of 10, and may involve interactions with neighboring aromatic groups (169, 172-174). When irradiation is at 280 nm, where 90% of the absorption is due to tyrosine residues, photoinactivation of RNase A still correlates with disulfide destruction (169). Amino acid analyses of photoinactivated RNase A, with 50% residual enzymatic activity, indicates that only the cystine content is significantly reduced (169-172). Spectrophotometric evidence suggests the possible modification of tyrosine to bityrosine (175). Since reduction of the disulfide bonds between residues 65-72 and 58- 110 does not inactivate the enzyme ( I ) , photoinactivation must arise from disruption of the disulfide between either residues 26-84, or 40-95, or both. Photoinactivation of RNase A at near 280 nm most likely occurs 167. Lichtin, N. N . , Ogdan, J . , and Stein, G . (1973). Radiat. Res. 55, 69. 168. Schultz, R. M., Immartino, A. J . , and Aktipis, S . (1975). BBA 386, 120. 169. Rathinasamy, T. K . , and Augenstein, L. G. (1968). Biophys. J . 8, 1275. 170. Grist, K . L., Taylor, T., and Augenstein, L. (1965). Radiat. Res. 26, 198. 171. Augenstein, L., and Riley, P. (1964). Phororhem. Photobiol. 3, 353. 172. Risi, S., Dose, K., Rathinasamy, T. K., and Augenstein, L. (1967). Photochem. Pirotobiof. 6 , 423. 173. Shafferman, A., and Stein, G. (1974). Photochern. Photobiol. 20, 399. 174. Arian, S . , Benjamini, M . , Feitelson, J., and Stein, G. (1970). Photochem. Phorobiol. 12, 481. 175. Aktipis, S . , and Iammartino, A . J. (1972). BBA 278, 239.

344

PETER BLACKBURN AND STANFORD MOORE

via activation of buried tyrosine residues, with subsequent sensitization of disulfide bonds (176, 177). Tyr-92 and Tyr-97 are both adjacent to the essential disulfide between residues 40-95. Tyr-97 is also adjacent to the essential disulfide between residues 26-84; Tyr-25 is in the vicinity of this disulfide. The exposed Tyr-73 and Tyr-115 are adjacent to the nonessential disulfide between residues 58- 110. The nonessential disulfide between residues 65-72 is not near any tyrosine residues; Tyr-76 is not close to any disulfide. The photoinactivation of RNase A has been shown by CD measurements to be associated with an altered and less stable conformation of the protein (168, 178). Photooxidation of NE-dinitrophenyl-Lys-41-RNase A by UV irradiation in the presence of molecular oxygen leads to specific modification of one methionine, one histidine, and one tyrosine residue (179) at positions 30, 12, and 97, respectively (180). After specific reduction of the disulfide bond between residues 65 and 72 according to Sperling ef al. (130), and S-dinitrophenylation, photooxidation of the protein upon irradiation led to modification of Tyr-73 and Tyr-115 (180). These reactions permit the topological and relative spacial arrangements of residues to be probed. In general, oxidation is normally restricted to Cys, Trp, Tyr, Met, and His residues (179, /81). The covalent attachment of 4-thiouridylic acid to RNase by irradiation with UV light at 334 and 365 nm has been reported (182). Irradiation of RNase A complexes with cytidine 2’(3’),5‘-diphosphate (pCp) or uridine 2‘(3’),5’-diphosphatewith ultraviolet light >300 nm resulted in covalent attachment of the pyrimidine nucleotides to the enzyme (183). Tryptic hydrolysis and peptide mapping showed attachment in the peptide segment from Asn-67 through Arg-85 of RNase A . Matheson et al. (184) studied the more generalized labeling of RNase A with the aryl nitrene, N-(4-nitreno-2-nitrophenyl)-2-aminoethylsulfonate, generated by Aash photolysis from N-(4-azido-2-nitrophenyl)-2aminoethyl sulfonate (NAP-taurine) (185), and have used this reaction to 176. 177. 178. 179. 180.

Volkert, W. A., and Grossweiner, L. I . (1973). Plzotochem. fhotohiol. 17, 81. Setlow, R., and Doyle, B. (1957). BBA 24, 27. Aktipis, S., and lammartino, A. J. (1971). BBRC 44, 918. Scoffone, E., Galiazzo, G . , and Jori, G. (1970). BBRC 38, 16. Jori, G . , Galiazzo, G . , Marchiori, F., and Scoffone, E. (1970).l a f .J . Peptide Protein Res. 2, 247. 181. Jori, G . , Gennari, G ., Galiazzo, G . , and Scoffone, E. (1970). FEBS Letr. 6, 267. 182. Sawada, F. (1975). BBRC 64, 311. 183. Sperling, J., and Havron, A. (1976). Biochemistry 15, 1489. 184. Matheson, R. R., Jr., Van Wart, H. E., Burgess, A. W., Weinstein, L. I., and Scheraga, H. A. (1977). Biochemistry 16, 3%. 185. Staros, J. V., and Richards, F. M . (1974). Biochemistry 13, 2720.

12. PANCREATIC RIBONUCLEASES

345

study steps in the thermal unfolding of RNase A (186). Nitrenes generated in the presence of the protein are capable of inserting into carbon-hydrogen bonds to form secondary amines that are stable to acid hydrolysis (187). Modification of exposed amino acids is not loo%, and varies in extent with the type of residue for reasons that are not altogether clear (188). With RNase, a degree of saturation of labeling occurred at a reagent to protein ratio of 212: 1 (184). With mixtures of amino acids free in solution, basic amino acids are labeled more than acidic amino acids, and nonpolar amino acids with larger side chains are labeled more than those with smaller side chains. However, the selectivity of the nitrene for the amino acids in a polypeptide chain like that of RNase A is influenced by the environment around these residues in the protein.

B. ROLESOF RESIDUES NEARTHE NH2A N D COOH TERMINI 1. S-PeptideS-Protein Interaction

The S-peptide-S-protein system discovered by Richards ( 1 , 189), which results from the controlled cleavage by subtilisin primarily between Ala-20 and Ser-21, has formed the basis for extensive studies of the roles of residues in the two sections of the enzyme. The first 25 residues of the sequence contain some of the most variable positions of the mammalian pancreatic RNase molecules (see Section V), yet hybrids prepared from S-peptides and S-proteins of different species have remarkably similar properties (190-192). The S-peptide segment has been subject to continued synthetic studies ( 1 93) to identify residues that have key roles in the binding to S-protein and in the activity of the complex. With the knowledge that residues 15 to 20 are dispensable ( I ) , substitutions in residues 1 through 15 have given a number of crystalline analogues that give X-ray diffraction patterns very similar to that of RNase S (194, 195). Substitution of Asn for Asp at 186. Matheson, R. R., Jr., and Scheraga, H. A. (1979). Biochemistry 18, 2437. 187. Knowles, J. R. (1972). Accounrs Chem. Res. 5, 155. 188. Bayley, H . , and Knowles, J. R. (1978). Biochemistry 17, 2414. 189. Richards, F. M. (1955). C.R. Trav. Lab. Carisberg 29, 322. 190. Welling, G. W., Lenstra, J. A., and Beintema, J . J. (1976). FEES Lerr. 63, 89. 191. Voskuyl-Holtkamp, I., Schattenkerk, C., and Havinga, E. (1976). Int. J . Pepride Protein Res. 8, 455. 192. Voskuyl-Holtkamp, I., and Schattenkerk, C. (1977). I n / . J . Peptide Prorein Res. 10, 60;ibid. 10, 153; ibid. 11, 218. 193. Chaiken, I. M. (1978). In “Semisynthetic Peptides and Proteins” (R.E. Offord and C. DiBello, eds.), p. 349. Academic Press, New York. 194. Chaiken, I. M., Taylor, H. C., and Amrnon, H. L. (1977). JBC 252, 5599. 195. Pandin, M., Padlan, E. A . , Di Bello, C., and Chaiken, I. M. (1976).PNAS 73, 1844.

346

PETER BLACKBURN AND STANFORD MOORE

position 14 reduces the affinity of S-peptide for S-protein about 20-fold but has no effect on the enzymatic activity in the presence of an excess of this analog (1%). In S-protein, Tyr-25 titrates spectrophotometrically with a normal pK, of about 10.2 (197); addition of S-peptide to form RNase S' restores the buried characteristics of Tyr-25 (197-200). Addition of lC,7€, 10S-triguanidino-(Orn-10, Asn- lrl)-S-peptide to S-protein restores 75% of the spectral characteristics of RNase S' (1%). This result indicates that Asn-14 can hydrogen bond to the phenolic hydroxyl of Tyr-25, as does Asp-14 in RNase A and RNase S. Finn and Hofmann (201) and Hearn et al. (202) proposed that a charge interaction exists between Asp-14 and Arg-33. Filippi et al. (1%) suggested that the weaker interaction arising from substitution of Asn for Asp-14 may result in large part from loss of this charge interaction. The binding of 1€,7€,10S-triguanidino-(Orn-10, Asn-14)-S-peptide to S-protein, as followed by the change in CD at 222 nm, exhibits a titration curve with a midpoint near pH 6.0 (1%). The binding of S-peptide or lC,7'-diguanidino-S-peptide does not exhibit this pH dependence. The titration behavior of the binding of the Asn-14 S-peptide analog may be a reflection of an interaction involving His-48 expressed in the absence of a strong electrostatic interaction with Arg-33. NMR titration studies on RNase A have suggested a close association between Asp-14 and His-48 (77-79), and between Asp-14 and Tyr-25 (74, 80). However, the 13C NMR studies of Niu et al. (203, 203a), with an S-peptide analog of residues 1-15 synthesized with 13C-enriched Asp14, Met- 13, and His-12, indicate that the proposed interaction between Asp-14 and His-48 is unlikely, and favor a hydrogen bond between Asp-14 and Tyr-25. (See Section IV,B.) Met-13 contributes to the binding of S-peptide to S-protein (I). Alkyla1%. Filippi, B., Moroder, L., Borin, G . , Samartsev, M., and Marchiori, F. (1975). EJB 52, 65. 197. Shenvood, L. M., and Potts, J. T., Jr. (1965). JBC 240, 3806. 198. Woodfin, B. M., and Massey, V. (1968). JBC 243, 889. 199. Fung, D. S., and Doscher, M. S. (1971). Biochemistry 10, 4099. 200. Rocchi, R., Borin, G . , Marchiori, F., Moroder, L., Peggion, E., Scoffone, E., Crescenzi, V., and Quadrifoglio, F. (1972). Biochemistry 11, 50. 201. Finn, F. M., and Hofmann, K. (1973). Accounts Chem. Res. 6 , 169. 202. Hearn, R . P., Richards, F. M. Sturtevant, J. M., and Watt, G . D. (1971).Biochemistry 10, 806. 203. Niu, C.-H., Matsuura, S., Shindo, H., and Cohen, J. S. (1979). JBC 254, 3788. 203a. Cohen, J. S . , Niu, C.-H., Matsuura, S., and Shindo, H. (1980). In "Frontiers in Protein Chemistry" (T.-Y. Liu, G . Mamiya, and K . T. Yasunobu, eds.), p. 3. ElseviedNorthHolland, New York.

12. PANCREATIC RIBONUCLEASES

347

tion or oxidation of Met-13 (204) or substitution by leucine (205, 206) decreases the affinity for S-protein, but has little effect on the activity of the complex once formed. In the three-dimensional structure of the RNase molecule, the methionine side chain fits into a hydrophobic pocket formed by Val-47, Leu-51, and Val-54 ( 1 ) . This interaction was observed in the I3C NMR spectrum of S-peptide-(l-15), synthesized with [13C]Met-13,as an upfield shift resulting from greater shielding by the hydrophobic pocket (203). Modifications of Met-13 that increase the hydrophilicity (e.g., formation of the sulfoxide or the sulfone) or introduce a charge by alkylation (116) alter the steric properties of this residue. Rotation about the a-P carbon-carbon bond of Met-13 permits a charged sulfur atom in the complex to have access to solvent without a change in the conformation of the S-peptide; enzymatic activity is retained but there is a decrease in binding stability through loss of the hydrophobic contacts. Substitution of His-12 by Ser in S-peptide-(1-14) (106) and by Om-12 in (Om-lO)-S-peptide-(l-20) (207) significantly decreases the affinity for S-protein, suggesting that the side chain of His-12 contributes to the binding between S-peptide and S-protein. Replacement of His-12 with P-(pyrazolyl-3)-~-alanine( 1 05) or 4-fluoro-~-histidine(208) has no effect on the binding of S-peptide to S-protein, but the complexes are inactive. The ring protons of these residues have pK, values near 2.6, and they cannot participate in acid-base catalysis at neutral pH. Competitive ligand (2’-CMP) elution affinity chromatography on Sepharose-Caminophenyl5’-phosphoryluridine 2’(3‘)-phosphate (209) has demonstrated that substitution of 4-F-His at position 12 does not affect substrate binding (210). Carboxymethylation of N-3 of the imidazole of His-12 in a peptide that contains residues 1 through 14 actually increases the affinity of this S-peptide analogue for S-protein ( 105). Substitution of His-12 by L-homohistidine lengthens the side chain of residue 12 by one methylene group, but retains the imidazole moiety with a pK, similar to that of histidine. This substitution in an S-peptide 204. Richards, F. M., and Vithayathil, P. J. (1980). Brookliuven Syrnp. B i d . 13, 115. 205. Hoes, C., van Batenburg, 0. D., Kerling, K. E. T.,and Havinga, E. (1977). BBRC 77, 1074. 206. van Batenburg, 0. D., Raap, J., Kerling, K. E. T., and Havinga, E. (1976). Rec. Truv. Chim. Pays-Bus 95, 278. 207. Borin, G., Toniolo, C., Moroder, L., Marchiori, F., Rocchi, R., and Scoffone, E. (1972). I n r . J. Peptide Protein Res. 4, 37. 208. Dunn, B. M.,Di Bello, C., Kirk, K. L., Cohen. L. A., and Chaiken, I. M. (1974). JBC 249, 6295. 209. Chaiken, I. M.,and Taylor, H. C. (1976). JBC 251, 2044. 210. Taylor, H . C., and Chaiken, I. M. (1977). JBC 252, 6991.

348

PETER BLACKBURN AND STANFORD MOORE TABLE 111 S-PROTEIN-ACTIVATING ABILITY (Amax)A N D B I N D I N G CAPACITY OF SOME RIBONUCLEASE S-PEPTIDE-(1-14) ANALOGS I N WHICH T H E ACTIVE-SITE HISTIDINE-12 IS REPLACED B Y OTHERRESIDUES" Binding capacity' Residue 12

(%)

Substrate present

100

1000

80

10

Substrate absent

Ref.

1

H

I

0.01

(211)

Immeasurably low

(212)

0.01

(213)

nd'

(217)

H

Nhi

4IJ

80

0.01

20

0.1

<1

lood

34

100

H

(3-Pyd)Ala

- C H 2 G

-N

(4-Pyd)Ala

A H 2P

w

" From van Batenburg er

N

r r / . ( 2 1 1 - 2 / 3 ) and Hoes et rrl. ( 2 f 7 ) . with yeast RNA as a substrate at pH 5.0. A,,, is expressed relative to S-peptide-(1-14) (100%).

' Measured

12. PANCREATIC RIBONUCLEASES

349

analogue of residues 1 through 14 decreases the affinity for S-protein about 100-fold, but the complex retains 80% activity toward RNA (211); the kinetics of hydrolysis of cyclic 2’,3’-CMP exhibits a fourfold lower k,,, and a 9-fold higher K , than that exhibited by RNase S (206). Substitution of 4-imidazolylglycine for His-12 reduces the side chain of residue 12 by one methylene group; the p K , for dissociation of the imidazole protons is about 4.6 (212). The affinity for S-protein of this S-peptide analog, consisting of residues I through 14, is 10”-fold lower than that of native S-peptide residues 1 through 20; the complex has a maximum activity of 80% of that of RNase S toward RNA. Thus slight modifications in the side chain of His-12, while retaining the imidazole ring, can severely impair the ability of S-peptide to bind to S-protein (211, 212) and clearly demonstrate an important role for this residue in binding to S-protein. On the other hand, these modifications only slightly affect the activity of the complex and indicate that a degree of conformational flexibility is permissible at the active site of the complex. The effects of substitution of N-1-methylhistidine or N-3-methylhistidine at position 12 on the binding of S-peptide analogs of residues 1 through 14 (213) (Table 111) are consistent with the importance of N-3 of the imidazole for the catalytic activity, and that of N-1 of the imidazole in binding to S-protein; N-1 is thought to hydrogen bond t o the peptide carbonyl of Thr-45 ( I , 16, 214). Substitution of His-I2 of this S-peptide analogue by P-(3-pyridyl)-~-alanine introduces a side chain with a single titrating nitrogen atom with p K , for dissociation of the proton of about 5.5 (21.5). This derivative demonstrates a high affinity for S-protein, 211. van Batenburg, 0. D., Raap, J., Kerling, K. E . T., and Havinga, E. (1975). TctLett. 51, 4591. 212. van Batenburg, 0. D., Kerling, K. E. T., and Havinga, E. (1976). FEES Lett. 68, 228. 213. van Batenburg, 0. D., Voskuyl-Holtkamp, l . , Schattenkerk, C., Hoes, K., Kerling, K. E. T., and Havinga, E. (1977). BJ 163, 385. 214. Patel, D. J . , Canuel, L. L., Woodward, C., and Bovey, F. A. (1975). Biopolymers 14, 959. 215. Voskuyl-Holtkamp, I . , and Schattenkerk, C. (1979). h i . J . Pepride Protein R e f . 13, 185.

rrihedroti

‘ Binding capacity in the presence of substrate was determined directly from activity assays [Berger and Levit, Ref. (22211; values are expressed relative to Speptide-(l-14) + substrate = 1000, which corresponds to a binding constant of 5 x 10” M - ’ . Binding capacity in the absence of substrate was determined by UV-difference spectroscopy: values are expressed relative to S-peptide-(l-14) + substrate = 1000. Determined by competitive inhibition experiments by using ribonuclease S’ as a standard. ‘’ nd, Not determined.

‘

350

PETER BLACKBURN A N D STANFORD MOORE

but the complex is enzymatically inactive (213). Substitution of p(4-pyridyl)-~-alanineat position 12 (216) reduces the affinity for S-protein by about 10-fold; this complex exhibits 35% the activity of RNase S toward RNA but less than 4% toward cyclic 2',3'-CMP (217). The suggestion that both N-1 and N-3 imidazole nitrogens of His-12 are essential for activity (213) seems unlikely in view of the considerable activity toward RNA displayed by the complex of S-protein with p-(4-pyridyl)-~-Ala123-peptide-(1- 14) (217). Although Gln-11 can be substituted by Glu without major effect on activity, or significant reduction of the affinity for S-protein (218, 219), the adjacent Arg- 10 is important for binding to S-protein; substitution by ornithine significantly reduces the affinity of the S-peptide for the Sprotein (219, 220). Hofmann et cd. (105, 106) were first to suggest that an interaction between Arg- 10 and Glu-2 contributes to the conformation of S-peptide required for binding to S-protein. Evidence for this interaction has come from comparisons of the 250-MH, proton NMR spectra of N-3carboxymethyl-His-12-S-peptide with those of (Orn-lO,N-3-carboxymethyl-His- 12)-S-peptide upon addition of S-protein (221). Appropriate chemical shifts of the C6-H resonances of Arg-10 are observed upon addition of S-protein; the signals from the ornithine side chain remain unchanged. Marchiori er (11. (104) showed that substitution of Ala-6 by proline (thereby shortening the a-helical portion of the S-peptide) prevents the interaction between the side chains of Glu-2 and Arg-10. This substitution decreased the catalytic activity of the complex but was without serious effect on the K, for both cyclic 2',3'-substrates and RNA (104).

The binding properties and activities of S-peptide and des-1 through des-8 S-peptide complexes with S-protein have been studied by Berger and Levit (222, 223) and are shown in Table IV. Loss of residues Lys-1 through Thr-3 causes a 450-fold decrease in affinity for S-protein. The 216. Hoes, C., Raap, J., Bloemhoff, W., and Kerling, K. E. T. (1980). R r c . Tmv. Chim. Pays-Bris 99, 99. 217. Hoes, C., Hoogerhout, P., Bloemhoff, W., and Kerling, K. E . T., (1979). Rrc. Trav. Chim. Poys-Bns 98, 137. 218. Finn, F. M . , and Hofmann, K. (1965). JACS 87, 645. 219. Scoffone, E., Rocchi, R., Marchiori, F., Marzotto, A . , Scatturin, A . , Tamburro, A . , and Vidali, G. (1967). J . Clirm. Soc. ( C ) ,p. 606. 220. Moroder, L . , Marchiori, F., Rocchi, R., Fontana, A . , and Scoffone, E. (1969).JACS 91, 3921. 221. Finn, F. M . , Dadok, J . , and Bothner-By, A. A. (1972). Biochemistry 11, 455. 222. Berger, A . , and Levit, S. (1973). In "Peptides 1971" (H. Nesvadba, ed.), Roc. 11th Eur. Peptide Symp. Vienna, 1971, p. 373. North-Holland, Amsterdam. 223. Levit, S ., and Berger, A . (1976). JBC 251, 1333.

35 1

12. PANCREATIC RIBONUCLEASES TABLE IV B I N D I N PROPERTIES G A N D ACTIVITIES OF $PROTEIN : S-PEPTIDE SYSTEMS AT pH 5.0, 27""

(9-20) (8-20) (7-20) (6-20) (4-20) (3-20) (2-20) (1-20)

3.1 2.2 3.1 5.0 3.2 7.0

x lo-" x lo-"

x x 10-8 x lo-" x lo-'

3.2 x 4.6 x 3.2 x 2.0 x 3.1 x 1.4 x

lo4 104 10'

107 lo7 10"

6.3 6.5 7.7 10.2 10.5 11.4

Not measurable Not measurable 0.21 0.29 0.48 0.56 0.86 1.oo

" From Berger and Levit (222, 223 ). I,

Measures catalytic efficiency.

three alanine residues at positions 4 3 , and 6 contribute little to the binding with S-protein (220, 223) and may be substituted by hydrophilic serine residues with little effect on the interaction or on the activity of the complex ( 2 2 4 ) . The hydrophilic side chain of Glu-9 is exposed to solvent and forms part of the a-helical segment of S-peptide. Substitution by Leu-9 in Speptide-(l-15) causes only a 3-fold reduction in the K b for S-protein; substitution by Gly-9 in S-peptide-(I- 15) causes a 22-fold reduction in the K b (225). Thus, the negative charge and hydrophilicity at position 9 are not essential. The different abilities of leucine and glycine to substitute for Glu-9 reflect their different propensities to participate in a-helical structures (225). Acylation of the NH2 groups of S-peptide ( 1 , 222, 223) does not have a determining effect on the binding efficiency. Hofmann et al. (106) demonstrated that substitution of Nle for Lys-7 of S-peptide has only a small effect on the activity of the complex. Thus, Lys-7 is not essential for catalytic activity. The presence of substrate influences the interaction between S-protein and S-peptide (201). The complexes in general exhibit greater activity toward RNA than toward cyclic 2',3'-substrates (104, 218). The ultraviolet spectral changes that accompany binding of S-peptides to 224. Borin, G . , Marchiori, F., Moroder, L., Rocchi, R., and Scoffone, E . (1971). BBA 271, 77. 225. Dunn, B. M., and Chaiken, I . M. (1975). J M B 95, 497.

352

PETER BLACKBURN AND STANFORD MOORE

S-protein (226) are proportional to the concentration of the complex ( 1 98) and result from the exclusion of Tyr-25 from solvent. In general, the application of direct spectral techniques to follow binding between S-protein and S-peptide analogs produces qualitatively similar results to those obtained by activity determinations and competition assays between S-peptide and analogs that produce inactive complexes with S-protein (227). However, quantitatively, the presence of substrate has a significant effect on the binding constant for the association of S-peptide with S-protein (Table V ; including Refs. 228-230). Gawronski and Wold (231) studied the interaction between S-peptide coupled to CNBr-activated agarose and S-protein, and reported a K d at pH 7.5 and 23" of 2 x 10-6M.Covalent attachment of the S-peptide to the insoluble matrix had only a small (fourfold) negative influence on the association. Also, below 25" the dissociation constants for RNase S' and complexes with S-protein acetylated with 13H]aceticanhydride at 1 and 9 moles per mole of S-protein were essentially identical (232). With this system, saturating concentrations of a mixture of 2'(3')-CMP had no influence on K d . Studies of the thermodynamics of the interaction suggested that the association was entropically driven. Calorimetric studies on the association of S-peptide with S-protein, performed by Hearn et a/. (202) indicated that the process is enthalpically driven. Estimates of Kd were made between 30" and 45" from assays toward cyclic 2',3'-CMP. The van't Hoff plots were nonlinear in this range, as also reported by Gawronski and Wold (232);the higher temperature, in this case, lies close to the transition temperature of S-protein. The calorimetric measurements of AH were performed in the absence of substrate. The data in Table V and the studies of Schreier and Baldwin (233, 234), show the degree to which binding of substrate influences the association of S-peptide with S-protein. Schreier and Baldwin (233, 234), taking advantage of the different amide 3Hexchange rates of S-peptide when free and bound to S-protein, studied the interaction as a function of temperature, pH, ionic strength, and the presence of 2'-CMP. They distinguished two steps in the dissociation, a partial unfolding step and a separation 226. Richards, F. M., and Logue, A. D. (1962).JEC 237, 3693. 227. Finn, F. M. (1972). Biocltemistry 11, 1474. 228. van Batenburg, 0. D. (1977). Ph.D. Thesis, University of Leiden, The Netherlands. 229. Kenkare, U. W., and Richards, F. M. (1966). JEC 241, 3197. 230. Marzotto, A., Marchiori, F., Moroder, L., Boni, R., and Galzigna, L. (1967). EBA 147, 26. 231. Gawronski, T. H . , and Wold, E (1972). Eioehernistry 11, 442. 232. Gawronski, T. H., and Wold, F. (1972). Biochemistry 11, 449. 233. Schreier, A . A., and Baldwin, R. L. (1976). J M E 105, 409. 234. Schreier, A. A., and Baldwin, R . L. (1977). Biochemistry 16, 4203.

12. PANCREATIC RIBONUCLEASES

353

TABLE V

LITERATURE K b VALUES FOR T H E ASSOCIATION BETWEEN S-PEPTIDE-(I-20)A N D S-PROTEIN'

Substrate None

Ref.

(227) (208) (226) (198) (200) (223) (198) (223) (229) (202) (225)

1 x 10'

0.5 x 10' 1.4 x lo4 1 x lo5 0.3 x 10"

S-Protein conc. ( M ) 5.5 x 10-j 5 x lo-: 8.7 x 10-j 6.5 x lo-' 7 x lo-'

6.5 x

4 x 107 2.2 x lo6

9.3 x 10-6 5 x 10-7 2.2 x

1.2 x 107 2.8 X lo7

4.7 x 10-7

1 x 10"

pH 5.0

7.1 4.5 -

6.8 5.0 5.4

6.5 7.0 7.0

5 x 10-7

7.1

1.5 x 10-7

7.1

1.4 x 10"

5 x lo-'

(104)

1.9 x lo8 3.3 x loH

(191)

4.2 x 10'

5.2 x lo-' 8.5 x lo-' 2.5 x lo-"

5.0 5.0 6.0 5.0

(1%)

Yeast RNA

Kb(M-')*

(223) (230)

5 x 107

" From van Batenburg (228).

6Measured in the absence of substrate as determined by UVdifference spectroscopy, and in the presence of substrate as determined by activity recovery. ' Stoichiornetric binding; no calculation possible.

step. Unfolding was enhanced at lower pH and ionic strength, suggesting that it is induced by electrostatic repulsion, possibly between the positive charges of residues Lys-7, Arg-10, and His-12. The separation step was independent of ionic strength, indicating that nonionic interactions predominate. Binding of 2'-CMP had a large effect on the separation step, and the Kd became too small to measure. The equilibrium constants and thermodynamic parameters for the dissociation and partial unfolding of S-peptide bound to RNase S are shown in Table VI. Schreier and Baldwin indicate that residues 12-14 may serve as an "anchor" in the initial combination of S-peptide with S-protein; binding of 2'-CMP with His- 12 would then assist this association. Rosa and Richards (2342) have obtained increased structural resolution by combining HPLC of proteolytic 234a. Rosa, J. J., and Richards, F. M. (1979).JMB 133, 399; (1981).ihitl. 145, 835.

354

PETER BLACKBURN AND STANFORD MOORE TABLE VI

A. TEMPERATURE-DEPENDENCE OF T H E E Q U ~ L I B R CONSTANTS IUM FOR DISSOCIATION

(Kd) A N D PARTIAL UNFOLDING (K,) OF S-PEPTIDE BOUNDTO RNASEs"

4.25 4.25 4.25 6.9 6.9 6.9

B.

4.0 7.9 12.5 0 3.6 8.4

VAN'T

1.6 x 3.4 x 8.0 x 8.8 x 2.3 X 3.7 x

1.0 x 1.6 x 2.3 x 5.9 x 7.0 x 1.3 x

10-8 lo-' lo-'' lo-'"

HOFFENTHALPY A N D ENTROPY OF DISSOCIATION AND UNFOLDING OF S-PEPTIDE BOUNDTO RNASEs'

OF

10-3 10-3 10-3 lo+ lo-'

PARTIAL

PH

Temp. range ("C)

Reaction

AH (kcaYmol)

AS (euimol)

4.25 4.25 6.9 6.9

4- 12.5 4-12.5 0-8.4 0-8.4

Dissociation (Kd) Partial unfolding (K,) Dissociation ( K d ) Partial unfolding (K,)

29 14 28 13

71 39 57 24

' Reprinted with permission from Schreier and Baldwin (234), Biorhemistry 19, 42034209. Copyright (1977) American Society.

digests performed at pH 2.8 to identify residues undergoing tritium exchange in such experiments with S-peptide and S-protein. Niuet al. (235) used a synthetic S-peptide-(l- 15) labeled at His-12 with to obtain by NMR measurements a K d of 0.2 x lO-'M, which was greater by a factor of 5 than the Kd determined for the S-peptide-( 1-20) by competition experiments. On thermodynamic grounds they propose a hydrogen bond between Ser-16 and His-48 ( 1 ) to account for the difference in binding enthalpies between S-peptides-(1-20) and (1- 15). Schreier and Baldwin (234) suggest that the low K b values obtained by spectrophotometric techniques (Table 111), may reflect the presence of some molecules that are not completely native and dissociate more readily than RNase S. S-protein tends to aggregate at low pH values (236). At pH 4.5, Gawronski and Wold (231) reported anomalous binding data for the titration of agarose-S-peptide with S-protein. By attaching S-protein to the agarose matrix to prevent low pH aggregation of S-protein and titrating with 235. Niu, C.-H., Shindo, H., Matsuura, S . , and Cohen, J. S. (1980). JBC 255, 2036. 236. Allende, J . E . , and Richards, F. M. (1962). Biochemistry 1, 295.

12. PANCREATIC RIBONUCLEASES

355

S-peptide, normal titration behavior was observed. The dissociation constant of the aggregated S-protein dimer was reported to be near M at pH 7.5, but less than M at pH 4.5. They conclude that the pHdependent association of S-protein may affect the values obtained for K d of RNase S' derived by direct spectrophotometric titrations, especially those at lower pH values since relatively high concentrations of S-protein are required by this procedure (see Table V). Such an influence would tend to increase the values obtained for Kd.Indeed, Dunnet af. (208) were unable to derive a binding constant for the association of S-peptide with S-protein by spectrophotometric titration at pH 7, where the binding was apparently stoichiometric. Hearn et al. (202), however, found no effect of S-protein concentration on AH for the association of S-peptide with S-protein, and concluded that either the aggregation of S-protein does not involve significant enthalpy changes or that aggregation is insignificant in the concentration range 2 x 1O-j to 2 x lop3M. Gawronski and Wold (232) indicated that AH of dissociation of S-protein dimer is close to zero, but that A S of dissociation was - 18.5 ca1 deg-' mole-'. NMR studies have demonstrated that the active-site histidine residues of RNase A, S, and S' occupy the same chemical environment (237-239), and that the dominant conformation of free S-peptide in aqueous solution is that of a random coil (221, 240). Upon binding to S-protein, the S-peptide undergoes a coil-to-helix transition (221). By incorporating 13Clabeled glycine at position 6 and either I9F as p-fluorophenylalanine or 13C-labeledphenylalanine at position 8 of S-peptides of residues 1- 15, the coil-to-helix transition undergone by S-peptide upon binding to S-protein has been observed by NMR spectroscopy (241. 242). This transition in unfolded RNase A, with its native disulfide bonds intact, is thought to be an early event during refolding of the molecule (243). The interactions between S-peptide and S-protein help to stabilize intermediates during refolding of the molecule (244-246). These findings are in contrast to those 237. Meadows, D. H . , Jardetzky, O . , Epand, R. M., Riitejans, H. H., and Scheraga, H. A. (1968). P N A S 60,766. 238. Cohen, J. S., Griffin, J. H., and Schechter, A. N. (1973). JBC 248, 4305. 239. Shindo, H., and Cohen, J. S. (1976). JBC 251, 2648. 240. Silverman, D. N . , Kotelchuck, D., Taylor, G. T., and Scheraga, H . A. (1972). ABB 150,757. 241. Chaiken, I. M. (1974). JBC 249, 1247. 242. Chaiken, I. M . , Freedman, M. H . , Lyerla, J. R., Jr., and Cohen, J . S. (1973). JBC 248, 884. 243. Blum, A. D., Smallcombe, S. H . , and Baldwin, R. L. (1978). J M B 118, 305. 244. Schmid, F. X., and Baldwin, R. L. (1979). J M B 135, 199. 245. Labhardt, A. M., and Baldwin, R. L . (1979). J M B 135, 231. 246. Labhardt, A. M., and Baldwin, R. L. (1979). J M B 135, 245.

356

PETER BLACKBURN AND STANFORD MOORE

reported by Klee (247, 248) who estimated 10- 15% helical content of the peptide based on CD measurements. Scoffone et al. (249), on the other hand, concluded on the basis of CD and ORD measurements with S-peptide analogs that the dominant conformations were disordered. The contributions of S-peptide residues to the refolding of RNase A are discussed in Section IV,D. A new approach to the study of the S-peptide-S-protein interactions has been reported by Hoogerhout et a / . (250).Their approach is to prepare synthetic extensions at the NH2terminus of acetimidyl-blocked S-protein. This approach should reveal any influence that the covalent linkage between residues Ala-20 and Ser-21 exerts in native RNase A. Conformational restrictions associated with the covalent attachment of the S-peptide fragment to the S-protein portion have already been indicated (1, 119). Modifications to Met-13 such as, photooxidation (117, 118) and alkylation (1 16, 118),which inactivate RNase A, can be accommodated by RNase S’; even though the interactions of the modified S-peptides with S-protein are weaker, the complexes are active. ) obtained enzymatic synthesis Homandberg and Laskowski ( 2 5 0 ~have (yield 50%) of RNase A$ (a mixture of RNase A, des-Ser-21 RNase A, and possibly Ser-21A RNase A) from RNase S by the use of subtilisin in 90% (v/v) glycerol.

-

2. Modijication Near the COOH Terminus Anfinsen (251) showed that removal of the tetra peptide at the COOH terminus of RNase A by limited peptic hydrolysis resulted in a derivative with little or no activity. Lin (252) was able to demonstrate 0.5% of the activity of RNase A toward 2’,3’-cyclic CMP. The susceptibility of His-12 and His-119 to alkylation by iodoacetate is still present but reduced in rate; the alkylation favors His-12 rather than His-1 19, which is preferentially modified in the native enzyme. The transition temperature drops from 61” for RNase A to 44” for RNase-(1-120). Taniuchi (142) and Puett (143) found that the derivative maintains much of the original conformation, as judged by CD spectra. After the -S-Sbonds are split by 247. Klee, W. A. (1968). Biochemistry 7, 2731. 248. Brown, J. E., and Klee, W. A. (1971). Biochemistry 10, 470. 249. Scoffone, E., Marchiori, F., Moroder, L . , Rocchi, R., and Borin, G . (1973). In “Medicinal Chemistry 111” (P.Pratesi, ed.), p. 83. Buttenvorths, London. 250. Hoogerhout, P., Bloehmhoff, W., and Kerling, K. E. T. (1979). Rec. Trav. Chim. Pays-Bas 98, 515. 250a. Hornandberg, G . A . , and Laskowski, M . , Jr. (1979). Biochemistry 18, 586. 251. Anfinsen, C. B. (1956). JBC 221, 405. 252. Lin, M. C. (1970). JBC 245, 6726.

12. PANCREATIC RIBONUCLEASES

357

reduction (142) the reoxidized protein is largely disordered. In terms of the process of biosynthesis, the chain is thus not programmed for proper folding until the synthesis has proceeded beyond residue 120. When Phe- 120 is removed by controlled hydrolysis with carboxypeptidase A at pH 5 (2521, all evidences of catalytic activity and native structure are lost. The transition temperature of the des-( 120- 124) derivative is lowered to 34". The loosened structure is highly susceptible to proteolysis by trypsin at 25". CD spectral studies by Puett (144) show the decreased conformational stability of this derivative by unfolding experiments in guanidinium chloride solutions. The phenylalanine residue, which in the three-dimensional structure fits into a hydrophobic pocket, thus has an important role in maintenance of the native conformation of the molecule. The further removal of His-1 19 can be accomplished by carboxypeptidase A action at pH 7.6 to give RNase-( 1- 118). In a collaborative study initiated by Lin and Gutte (25.3) a synthetic peptide corresponding to the 14 amino acid residues at the COOH terminus, synthesized by the solid phase method, was added to RNase-( 1- 118);the peptide's presence led to regeneration of 90% of the activity toward 2',3'-cyclic CMP and 70% toward yeast RNA at a peptide-to-protein ratio of 3. If both the first 20 and the last 6 residues were removed from RNase, a three-component system consisting of S-peptide (3.7 equivalents), a 21-118 residue core, and residues 111-124 (3.4 equivalents) gave 30% activity toward the synthetic substrate. In this case both of the two histidine residues that are near the active center were supplied by adsorbed peptides. These studies were extended (254) to include synthetic COOH-terminal peptides of varying lengths. There was negligible reactivation until the chain length reached 9 residues, which gave 60% activity. The Kd values for the added peptides and RNase-(1-118) were 2.5 X M at 9 residues and 2 x lo-' M at 14 residues. The role of Phe-120 was examined by substituting other residues at this position in the synthetic peptides (25s). Leu, Ile, or Trp was inserted at position 120. The maximum regenerable activities, calculated according to Berger and Levit (222), were 98, 13, 12, and 0.5%, respectively, for the peptides with Phe, Leu, Ile, or Trp at position 120 (256). The order of the Kd values was 2 X 2.5 x 4x and 3.5 x 10-jM. The K , values for the complexes with 2',3'cyclic CMP (255) were all near the value for the native enzyme (about M ) . The lowered enzymatic activities appear to result from misalign253. Lin, M. C., Gutte. B., Moore, S., and Merrifield, R. B. (1970). JBC 245, 5169. 254, Gutte, B., Lin, M. C., Caldi, D. G., and Merrifield, R. B. (1972). JBC 247, 4763. 255. Lin, M. C., Gutte, B., Caldi, D. G., Moore, S., and Merrifield, R. B. (1972). JBC 247, 4768. 256. Hayashi, R., Moore, S . , and Merrifield, R. B. (1973). JBC 248, 3889.

358

PETER BLACKBURN A N D STANFORD MOORE

ment of the residues at the active site and not from changes in the binding affinity for the substrate. Hodges and Merrifield (257) synthesized the 14-residue peptide with Tyr or Ala in position 120. In combination with RNase-(1-118), the substitution of Tyr at this position gave a product that had the same activity as the Phe-containing peptide toward RNA and 2',3'-cyclic CMP, but twice the activity toward 2',3'-cyclic UMP. With the Ala substitution, the relative activity toward 2',3'-cyclic CMP was less than 1%. These results led the authors to predict that giraffe RNase, which has a Tyr residue at position 120, would have a higher relative activity toward the UMP substrate than toward cyclic CMP. The prediction was verified; giraffe RNase showed a 2.6-fold greater activity toward the former substrate, taking the k,/K,,, values with bovine RNase as 1 [cf., Ronda et al. (258)l. The X-ray data ( I ) indicate a hydrogen bond between the hydroxyl of Ser-123 and the C-4 carbonyl oxygen of uridine. Potts et al. (259) showed that removal of Val-124 and Ser-123 by carboxypeptidase leaves 45% activity toward RNA; thus it is known that Ser-123 is not crucial. Hodges and Merrifield (260) synthesized RNase-( 111- 124) with Ala in the place of Ser at position 123. When the synthetic peptide was mixed with RNase(1-1 18), the substitution caused no change in activity toward 2',3'-cyclic CMP or in the transphosphorylation step with poly(C); but with 2',3'cyclic UMP the analog was 4 times less active, and with poly(U) two times less active. The results indicate that a hydrogen bond between Ser123 and the C-4 oxygen of uridine may contribute to substrate binding and catalytic activity with the uridine-containing substrate. The Pro residue at position 117 prevents carboxypeptidase A from carrying the degradation from the carboxyl end beyond residue 118. Hayashi et al. (256) tried carboxypeptidase Y, an enzyme capable of releasing proline. Carboxypeptidase Y removed Val-118, Pro-1 17, and Val-116. Tyr- 115 could then be removed by carboxypeptidase A. With the RNase A chain minus the last ten residues RNase-( 1- 114), reconstitution experiments with the nonapeptide RNase-( 116-124), in a combination thus missing Tyr-115, gave a maximum regenerable activity of 54%, which is essentially the same as that observed when the protein moiety was RNase-(1- 118). Tyr-115 is thus not required for activity. The result correlates with the finding of Jackson and Hirs (261) that in porcine RNase a 257. 258. 259. 3781. 260. 261.

Hodges, R. S. , and Merriiield, R. B. (1974). Int. J . Peptide Protein Res. 6 , 397. Ronda, G. J . , Gaastra, W., and Beintema, J. J. (1976). BBA 429, 853. Potts, J. T., Jr., Young, D. M., Anfinsen, C. B., and Sandoval, A. (1%4). JBC 239, Hodges, R. S., and Merrifield, R. B. (1975). JBC 250, 1231. Jackson, R. L., and Hirs, C. H. W. (1970). JBC 245, 637.

12. PANCREATIC RIBONUCLEASES

359

proline residue occupies position 115. The hydrogen bond between Tyr115 and Tyr-73 suggested by X-ray data is not crucial. Consideration of the associations that could hold the nonapeptide RNase-( 116- 124) t o the main chain, by reference to Wyckoff rf rrl. (262), indicates that six hydrogen bonds are possible with RNase-( 1- 114). The stronger binding ( K d = 3.0 x M ) than that obtained with the same peptide and RNase-(1- 118) (256) indicates that when there are overlapping residues in the 115 and 116 positions, there is competition for binding sites between the added peptide and the residual tail of the main chain. The transition temperature curves for RNase-( 1- 114) and the protein moiety plus RNase-(116-124) (see Ref. 285) show a change from a melted state at 25" to a more structured complex with a T,,, of 38". Andria and Taniuchi (263) mixed a tryptic hydrolysate of performic acid-oxidized RNase with RNase-( 1- 118) and gel filtered the mixture. The 105- 124 segment, which represents the residues following the last trypsin-susceptible bond in the sequence (Lys- 104), adsorbed to the RNase moiety that lacked the normal COOH terminus. The combination had about 509% of the activity of RNase A. The presence of RNase-(105bonds from reduced RNase-(I124) during the re-formation of -S-S118) increased the proportion of the species that exhibited the properties of RNase-( 1- 118) from zero to about 30%. C. CHEMICAL SYNTHESIS The Richards and Wyckoff review on RNase ( I ) recorded as dramatic events of 1969 the synthesis of RNase A (by Gutte and Merrifield) by the solid phase method, and the synthesis of RNase S-protein (by Denkewalter and Hirschmann and their associates) by solution methods. The details of the solid phase synthesis (264) document the isolation of 0.41 mg of purified synthetic enzyme with the substrate specificity of RNase A, and showing 78% of the activity of the native protein toward yeast RNA. A decade later, the synthesis of RNase A by solution methods was accomplished by Yajima and Fugii (265, 266). By including the use of affinity chromatography on pup-Sepharose (14% of the crude synthetic product was retained) they obtained 3 mg of a synthetic protein, which after fur262. Wyckoff, H. W., Tsernoglou, D., Hanson, A. W., Knox, J. R., Lee, B., and Richards, F. M. (1970). JRC 245, 305. 263. Andria, G., and Taniuchi, H. (1978). JBC 253, 2262. 264. Gutte. B., and Merrifield, R. B. (1971). JBC 246, 1922. 265. Yajirna, H., and Fujii, N. (1980). J . Clroni. S o c . Clrern. Couirn., p. 115. 266. Fujii, N . , and Yajima, H. (1981).J. Clwrn. Soc.. Perkin Trans. 1,789,797, 804, 811, 819. 831.

360

PETER BLACKBURN AND STANFORD MOORE

ther ion exchange chromatography was indistinguishable from the native enzyme by chemical and physical criteria and had 100% of the activities of RNase toward RNA and 2’,3’-cyclic cytidylic acid. The chemical synthesis of a 124-residue polypeptide chain represents atorrr de force by present methods, in which yields are in the 1% range. The challenge is stimulating innovations in some of the many steps of the process (267). Gutte (268) has undertaken the synthesis of shorter chains that might have ribonuclease action. From the study of the three-dimensional model of RNase A ( I ) , he aimed at the synthesis of a 70-residue chain that might have sufficient structure to form the central portion of the molecule that comprises the active site. The sequence selected was one in which the 104-residue S-protein structure was shortened by 34 residues; some surface loops and the 8 half-cystine residues were omitted, with Gly, Ala, or Leu inserted to fill seven of the gaps caused by the Cys deletions. The yield of a 70-residue analog of S-protein was, after ion exchange chromatography, about 1% of a product that in the monomeric form (with or without S-peptide) had 0.1% of the activity of RNase toward RNA. For reasons not growing from the original premise, a dimer of the product had 4% activity toward RNA, and the monomeric or dimeric analogs when added to S-protein gave 75 to 150% of the activity of RNase S in the hydrolysis of 2‘ ,3’-cyclic cytidylic acid. In a second approach, Gutte (269, 270) aimed for a 63-residue analog beginning with residue 26 and containing two of the four cystine bridges. The product was first purified by gel filtration. When affinity chromatography (265) was used to select molecules that had an affinity for a substrate analog, 98% of the preparation went straight through; the retarded 2%, isolated in pg quantities, represented an overall yield of about 0.25% of a product that had 7.8% of the activity of RNase A toward poly(C), without the addition of S-peptide. Refolding of the reduced 63residue preparation (or of RNase A) in the presence of mononucleotides (271) produced changes in activities toward synthetic substrates such as poly(C) and poly(U); the conformations thus obtained were thermolabile at 40”. Starting from theoretical considerations based upon secondary structure prediction rules and model building, Gutte et al. (272) designed a 267. Barany, G . , and Merrifieid, R. B. (1979). In “The Peptides” (E. Gross and J . Meienhofer, eds.), Vol. 2, p. 1. Academic Press, New York. 268. Gutte, B. (1975). JEC 250, 889. 269. Gutte, B . (1977). JEC 252, 663. 270. Gutte, B. (1978). JEC 253, 3837. 271. Gutte, B . (1978). EJE 92, 403. 272. Gutte, B., Daumigen, M., and Wittschieber, E. (1979). Nnture (London) 281, 650.

12. PANCREATIC RIBONUCLEASES

36 1

34-residue peptide structure with potential binding power for the anticodon of yeast tRNAPhe.The synthetic peptide (with two cysteine residues) cross-linked dimer that binds tRNA and 2',3'-cyclic forms an - S - S nucleotides. It also cleaves tRNAPheand poly(C); the activity toward yeast tRNA was 2.5% of that of RNase A. The product of the transphosphorylation step, the 2',3'-cycIic nucleotide, was not hydrolyzed. These experiments were designed to ascertain the feasibility of formulating polypeptides that would be artificial enzymes. The stimulating results demonstrate that structures can be built that fulfill predictions for ability to bind nucleotides, and that such a binding can labilize a phosphodiester bond to transphosphorylation. The data on the 70-residue and the 63-residue RNase A analogs have limited bearing on the specific features of pancreatic RNase per se. The 34-residue study shows most successfully that the ability to induce transphosphorylation in an RNA can be built into a polypeptide chain in a number of ways. When the yields of active product are as low as they are in the syntheses of the 70- and 63-residue peptides, it is difficult to assign definitive structures to the active fractions isolated in microgram amounts. Nature's design of a pancreatic enzyme also incorporates properties that are essential to the protein's function in vivo. The simplified analogs are readily hydrolyzed by proteolytic enzymes such as trypsin; the native structure of RNase A is uniquely constructed and cross-linked to render the catalyst resistant to proteolysis. The entire structure of RNase A is an entity that can function in the gastrointestinal tract. Semisynthetic studies to examine the roles of residues in the NHz- and COOH-terminal sections of RNase are summarized in Section II1,B.

D. BIOSYNTHESIS The signal hypothesis (273) for secreted proteins would predict that bovine pancreatic RNase should be synthesized as a presecretory protein. Demonstration of this fact has awaited a method for preparing intact mRNA from bovine pancreas by the employment of conditions that inactivate the RNase that otherwise destroys the message in a homogenate of the tissue. The method of Chirgwin ef crl. (274), using 5 M guanidinium thiocyanate plus 0.1 M mercaptoethanol in the extraction medium, opened the way to the in vi?ro translation experiment. Haugen and Heath (275) prepared mRNA by that procedure and conducted the translation with the 273. Blobel, G . , and Dobberstein, B. (1975). J . Cell B i d . 67, 852. 274. Chirgwin, J. M . , Rzybyla, A. E., MacDonald, R. J., and Rutter, W. J. (1979). Biochemi.str.v 18, 5294. 275. Haugen, T. H . , and Heath, E. C . (1979). P N A S 76, 2689.

362

PETER BLACKBURN AND STANFORD MOORE

rabbit reticulocyte lysate system. The pre-RNase was isolated by precipitation with antibody to bovine pancreatic RNase. The product showed a MW of 16,500 by SDS gel electrophoresis. Sequence analysis of the [14C]Ala-labeledprotein showed the presence of a 25-residue segment ahead of the normal NH2 terminus. When incubated with dog pancreas microsomal membranes, the product yielded a protein that comigrated with RNase in gel electrophoresis; a band corresponding to RNase B was shown to be the glycosylated enzyme. Although assays for enzymatic activity on the picogram quantities of protein synthesized are difficult, the results indicate that pre-RNase A can fold into an enzymatically active form. Since RNase is a pancreatic enzyme that does not have a zymogen form, it is understandable that the presecretory (the signaled form) could have an active conformation. The authors reflect on the potentially damaging effect of any RNase activity that might reach the cytoplasm, and suggest that the cytoplasmic RNase inhibitor may have a function in this regard (275). E . IMMu N oc HE M ISTRY

Experiments on antibody-mediated modification of RNase activity have been reviewed by Cinader (276). Rabbit sera drawn on the 255th day after the start of immunization were chromatographed on DEAE-Sephadex A-50 (277). One fraction contained a small quantity of antibody that increased the activity (up to twofold) of the enzyme toward 2’,3‘-cyclic CMP (the activity toward RNA remained unchanged), other fractions contained antibodies that inhibited the enzyme’s actions up to 80% with the cyclic substrate and 98% with RNA. The effects leveled off at high antibody concentrations. The activating antibody gave a similar response with RNase S (278). In Section IV,D, on the refolding of reduced RNase, we refer to the ways in which Chavez and Scheraga used their findings on the antigenic sites of the enzyme to follow the folding process. They reviewed (279) the previous literature on the location of antigenic determinants of the native enzyme and applied a variety of immunochemical techniques to the definition of four antigenic segments of the chain. The coupling of peptides from RNase to Sephadex for the isolation of antibodies to given segments was a 276. Cinader, B. (1977). I n “Methods in Immunology and Immunochemistry” (C. A. Williams and M. W. Chase, eds.), Vol. IV, p. 313. Academic Press, New York. 277. Suzuki, T., Pelichova, H . , and Cinader, B. (1969). J . Immunol. 103, 1366. 278. Cinader, B., Suzuki, T., and Pelichova, H. (1971). J . Immunol. 106, 1381. 279. Chavez, L. G . , Jr., and Scheraga, H. A. (1979). Biochemistry 18, 4386.

12. PANCREATIC RIBONUCLEASES

363

key step in their analysis. Working with late hyperimmune antisera, four sites in bovine RNase A were localized within residues 1-10, 40-61, 63-75, and 87-104; the data do not provide information on the possible antigenicity of the region 30-39. Welling et a / . (280) conducted an immunological comparison of pancreatic RNases from nine species with antisera toward four of the species. Cross-reactivities by Ouchterlony double-immunodiffusion tests ranged from identity, through partial identity, to no identity. The complement fixation titers varied over a 200-fold range. Lee ef al. (281) found that sheep antibodies to purified rabbit spleen RNase gave only one precipitin band with rabbit pancreatic RNase, thus providing immunological evidence on the similarity of the enzymes from the two rabbit tissues. However, the antibody to the rabbit enzyme did not react with bovine o r rat pancreatic RNases. Working with antisera to bovine RNase A, Welling and Groen (282) focused on the relative antigenicities of six pancreatic RNases rather closely related in sequence; the tests were made by competition experiments employing the modified phage technique of Haimovich el al. (283). Through consideration of both the immunological results and the known sequence substitutions, they concluded that residues 34, 35, 103, and 50 and/or 99 are parts of antigenically reactive regions in bovine RNase A. These results are consistent with those of Chavez and Scheraga (279). Bovine RNase B, which has a carbohydrate side chain on Asn-34, was measurably less competitive than RNase A in the tests by Welling and Groen (282), confirming that this region of the enzyme has an input into the reaction with antiserum. A synthetic peptide comprising residues 1-14, at a molar excess over RNase up to lo6, did not inhibit phage inactivation (282). Chavez and Scheraga (279) suggest that their positive result for this region may depend upon the use of late hyperimmune sera. The experiments of Brown and associates on the antigenic regions of performic acid-oxidized RNase have been summarized (284). There is minimal immuno-cross-reactivity between the oxidized chain and the active enzyme; one section, residues 38 (or 40)-61, is antigenic in both the native and oxidized molecules. 280. Welling, G . W., Groen, G . , Beintema, J. J . , Emmens, M., and Schroder, F. P. (1976). Imrnrrnach~mistrv 13, 653. 281. Lee, W. Y., Gyenes L.. and Sehon, A. H. (1971). ~ ~ ~ u n a ~ h ~8,m751. i . ~ t ~ y 282. Welling, G. W., and Groen, G. (1976). BBA 446, 331. 283. Haimovich, J., Hurwitz, E., Novik, N., and Sela, M. (1970). BBA 207, 115; ;bid, 125.

284. Liu, S., Johnsen, E . , and Brown, R. K. (1974). lmmurzochemistry 11, 55.

364

PETER BLACKBURN AND STANFORD MOORE

Reports on the immunosuppressive action of polyribonucleases prepared by cross-linkage with bisdiazobenzidine ( 2 8 5 , 2 8 5 ~have ) not been confirmed (286). IV.

Physical Properties

The following sections cover studies subsequent to the definition of the physical param2ters of RNase summarized in a chapter in this series (1). A. X-RAYDIFFRACTION Richards, Wyckoff and colleagues (1, 16, 286n) have thoroughly reported and interpreted the extensive data on the three-dimensional structure of RNase S . Subsequent results include those of Wodak et al. (21) on the binding of 2',5'-CpA, a substrate analog that binds to the enzyme but is not cleaved. Their results reinforce the evidence obtained with other dinucleotides for a hydrogen bond between the N-3 of cytosine and the 7-OH of Thr-45. The orientation of the cytosine ribose in 2',5'-CpA in the binding site ;s such that the 3'-OH is away from the active center and not in a position to be rendered more nucleophilic to facilitate a transphosphorylation. The results of White et a/. (287) indicated little interaction of 2',5'-UpA or 2',5'-4-thio-UpA with RNase A, in solution in 0.2 M imidazole buffer at pH 7.0, both in terms of inhibition of activity and direct binding by CD measurements. The combination of RNase S and the analog of 3'3'-UpA with a fluorine atom replacing the 2'-OH has been studied crystallographically by Pavlovsky et (11. (73). The data indicate that the distance from F-2' to N-1 of His-12 is 3 A, a result that is consistent with the positioning of a normal substrate with a hydrogen bond between the 2'-OH and His-12 (1). Allewell et al. (40) have examined the crystallographic structure of edinitrophenyl-Lys-41-RNase S. The derivative and the crystals were prepared by Fung and Doscher (199). They first prepared eDNP-Lys-41RNase A in a chromatographically purified form that had less than 0.001% of the activity of RNase A. That derivative was cleaved by subtilopepti285. Mowbray, J . F., and Scholand, J . (1966). Imrnrrnolngv 11, 421. 285a. Mowbray, J. F. (1967).Svmp. Tissue Org. Trnnsplnnt ( S u p p l . , J . Clin. Paihol.) u), 499. 286. Chakrabarty, A . K., and Friedman, H . (1970). Clin. Exp. Immunol. 6, 619. 286a. Richards, F. M., Wyckoff, H. W., Carlson, W. D., Allewell, N. M., Lee, B . , and Mitsui, Y. (1971). Cold Spring Harbor Svmp. Qunnt. B i d . 36, 35. 287. White, M. D., Keren-Zur, M., and Lapidot, Y. (1977). N i d e i c Acids Res. 4, 843.

12. PANCREATIC RIBONUCLEASES

365

dase A to give the RNase S analog. In solution, the inactive DNPsubstituted protein binds nucleotides such as 3'-CMP about one-tenth as strongly as does the parent enzyme; crystals of the RNase A and S derivatives show the same affinity for 3'-CMP. The presence of the DNP group displaces the E-N of Lys-41 by about 3 A; His-12 and His-119 are not moved. The loss of activity appears to result primarily from perturbation of the molecule in the region of Lys-41. H. C. Taylor has advised us of results (2870) on the crystal structure of the semisynthetic RNase S in which 4-fluoro-His replaces His- 12 in the synthetic segment of residues 1-15 (288). The overall structure, including the positioning of His-12, does not differ detectably from that of normal RNase S; the data support the view that the 4-F-His analog is inactive as a result of the lowered pK, of His-12 rather than distortion of the active site. The inhibition of RNase S by Cu'+ has been studied by Allewell and Wyckoff (289). They identified 7 binding sites, 3 of which were intermolecular. Two of the intramolecular sites are close to the active-site histidine residues 12 and 119 and could be expected to cause the observed inhibition. The effect of the pH of crystallization on the conformation of RNase has been examined by Martin et cd. (290). Crystals of RNase A and of RNase S formed at pH 9 have been found to be isomorphous with those of RNase S crystallized at pH 6.6. X-Ray diffraction shows small but significant intensity differences comparable in magnitude to those observed between RNase A and RNase S. The results of the crystallographic studies on RNase A by Carlisle and associates (119) are in general agreement with the data previously available for RNase A and RNase S ( I ) in terms of the relationships of His-12, His-119, and Lys-41 at the active center of the catalyst. However, the peptide bond between residues 20 and 21 restricts the conformational freedom of the S-peptide segment compared to that found with RNase S . Timchenko et a / . (291) have studied RNase A in solution by large-angle X-ray scattering, which is sensitive to the internal structure of globular proteins in solution. The observed scattering curve was compared with the curve calculated from the atomic coordinates available from X-ray 287a. Taylor, H . C., Richardson, D. C., Richardson, J. S., Wlodawer, A , , Komoriya, A,, and Chaiken, I . M . (1981). J M B . 149, 313. 288. Dunn, B . M . , DiBello, C . , Kirk, K . L., Cohen, L. A . , and Chaiken, I . M . (1974). JBC 249, 6295. 289. Allewell, N . M . , and Wyckoff, H. W. (1971). JBC 246, 4657. 290. Martin, P. D . , Petsko, G . A . , and Tsernoglou, D . (1976). J M B 108, 265. 291. Timchenko, A. A . , Ptitsyn, 0. B., Dolgikh, D. A , , and Fedorov, B. A. (1978). FEBS Lett. 88, 105.

366

PETER BLACKBURN AND STANFORD MOORE

analysis. The conclusion was that the structure in solution does not differ significantly from that in the crystal. Wlodawer (292) has undertaken a computational refinement of the available crystal structure coordinates af RNase A for comparison with results of neutron diffraction on deuterated crystals. Neutron diffraction can provide detailed information on the orientation of the imidazole rings of histidine residues. Large crystals (1 x 5 x 6 mm), successfully prepared in about 50% rerr-butyl alcohol at pH 5.3, were used. The results (Wlodawer and Sjolin, 292a) give evidence for a hydrogen bond between His-119 and Asp-121 in crystals suspended in deuterated tert-butyl alcohol-D20. B. NUCLEAR MAGNETIC RESONANCE In the early NMR work on RNase A, Meadows et al. (237,293,294)and Ruterjans and Witzel (77) differentiated between the C-2 proton resonances (HI, H2, H3, and H4) of the four histidine residues based upon their titration curves before and after carboxymethylation and upon cytidine monophosphate binding. Meadows et al. (237) differentiated between the His-12 and His-119 C-2 proton resonances by a comparison of the titration curves of RNase S and RNase S’ reconstituted with S-peptide deuterated at the C-2 proton of His-12, and assigned the resonance with the lower p K, value to His- 119. They were supported in their assignment by results from the deuterium exchange studies of Bradbury and Chapman (295). However, subsequent studies have shown that these original assignments of the active site histidine proton resonances were incorrect; current assignments are included in Table VII (see also Refs. 296-303). 292. Wlodawer, A. (1980). Acto Crystallogr. 836. 1826. 292a. Wlodawer, A., and Sjolin, L. (1981). PNAS 78, 2853. 293. Meadows, D . H., Markley, J. L., Cohen, J. S., and Jardetzky, 0. (1967). PNAS 58, 1307. 294. Meadows, D. H., and Jardetzky, 0. (1968). PNAS 61, 406. 295. Bradbury, J. H . , and Chapman, B. E. (1972). BBRC 49, 891. 296. Markley, J. L. (1975). Biochemisfn’ 14, 3546. 297. Shindo, H . , Hayes, M. B . , and Cohen, J . S. (1976). JBC 251, 2644. 298. Bradbury, J. H., Crompton, M. W., and Teh, J. S. (1977). EJB 81, 411. 299. King, N . L. R., and Bradbury, J. H. (1971). Nature (London) 229, 404. 300. Meadows, D. H., Roberts, G. C. K., and Jardetzky, 0. (1969). JMB 45, 491. 301. Kaptein, R . , Dijkstra, K . , Muller, F., van Schagen, C. G . , and Visser, A . J . W. G. (1978). J . Mugneric Resonanr‘e 31, 171. 302. Kaptein, R., Dijkstra, K., and Nicolay, K. (1978). Nature (London) 274, 293. 303. Kaptein, R. (1978). In “Nuclear Magnetic Resonance Spectroscopy in Molecular Biology” (B. Pullman, ed.), p. 21 1 . Reidel, Dordrecht, Netherlands.

367

12. PANCREATIC RlBONUCLEASES TABLE VII SUMMARY O F

Resonanceb) H1 H2 H3 H4' HI' H2' H3'

Yl" y2d Y3"

THE

ASSIGNMENTS O F AROMATIC PROTON RESONANCES OF BOVINERIBONUCLEASE A"

Chemical shiftb (ppm)

8.08 7.91 7.77 7.18 6.70 6.43

( ( 7.19 6.76 { ;:I

Assignment His-105, C-2H His-119, C-2H His-12, C-2H His-48, C-2H His-105, C-4H His-1 19, C-4H His-12. C-4H

I

Tyr-76* C"H Tyr-76, Cf-H Tyr-115, Cd-H Tyr-115, Cf-H

Y4

Y5

6.90 6.70

Tyr-25, C'-H Tyr-92, Cs-H + Cf-H Tyr-73, CS-H or Cf-H

F3 F4

6.83 6.61

Phe-120, ring protons Phe-46, ring protons

I

6.3-6.5

Phe-120, ring protons near active-site inhibitor

Reference or evidence (78. 237, 296-298) (237, 296-298) (237, 296-298) (237, 296. 298) (237, 296, 299) (299); CIDNP spectrum' (299); Comparison with RNase S ; pH midpoint; effect of inhibitors Comparison with nitrated RNase Comparison with nitrated RNase Comparison with RNase S; pH titration Alkaline titration pH dependence in nitrated RNases Effect of active-site inhibitors Comparison with RNase S; pH titration curve (300)

From Lenstra et r i l . (80). The last five assignments are tentative. "The chemical shifts were measured at pH 7.0, 38", downfield from sodium 2,2dimethyl-2-silapentane 5-sulfonate. ' Not visible at pH 7.0. The upfield doublets are assigned to tyrosine C' protons on the basis of the polarizations in the CIDNP (chemically induced dynamic nuclear polarization) spectra. In the CIDNP technique (301-303), the appearance of resonances from histidine, tyrosine, and tryptophan residues in these spectra depends upon the accessibility of the aromatic ring systems to photo-excited flavin, resulting in nuclear spin polarization through reversible hydrogen or electron transfer (302. 303 ).

"

1. Histidine Assignments

A resonance, designated n , downfield of the C-2 proton resonances of histidine residues was observed by Patel et NI. (304) and was assigned to a 304. Patel, D. J., Woodward, C. K., and Bovey, F. A. (1972). PNAS 69, 599.

368

PETERBLACKBURN ANDSTANFORDMOORE

slowly exchangeable, possibly buried or hydrogen bonded imidazole N-H whose chemical shift and linewidth are governed by ionization of the other imidazole ring nitrogen. In the presence of 3'-CMP the p K , values of the active-site histidine residues are increased by > 1 pH unit (see Table VIII) (see also refs. 305-307), whereas the pK, values of His-48 and His-I05 were found to increase by only 0.1 pH unit (214). The p K , of resonance a increases from 6.3 to 7.2 in the presence of 3'-CMP (214), and this rules out its assignment to His-48 or His-105. In N-l-carboxymethyl-His119-RNase A, des-(121- 124)-RNase A, and DNP-Lys-41-RNase A, resonance (i titrates with constant band area and a pK, -6.8, which rules out its assignment to N-3 of His-119 and the e N H 2 group of Lys-41. In N-3-carboxymethyl-His-12-RNase A, resonance CI is observed with the chemical shift of a protonated histidine residue; it does not tirate, but disappears at alkaline pH as the proton exchanges with solvent. Resonance n was thus assigned to the N-1 proton of His-12 (214), which is proposed to hydrogen bond to the peptidyl carbonyl of Thr-45 ( 1 , 203). Based upon the titration behavior of resonance (I in the presence and absence of phosphopyrimidine nucleotides, by correlation with the previous histidine C-2 proton assignments (237), resonance a had been incorrectly assigned to the N-H of His-119 (308). The ionization characteristics of resonance fi were associated with the active-site histidine C-2 proton resonance that exhibits the lower p K, (H3) on the basis of Gun, Ag', and spin labeling data (309) and active-site inhibitor binding (214,308), suggesting that the histidine C-2 proton resonance with the lower pK, (H3) is from His-12 (78),and not from His-119, as was previously indicated (237, 294). The assignments of the C-2 proton resonances of His-12 and His-119 were thus shown to require reinvestigation. Patel et a/. (78) demonstrated in RNase S and RNase S', reconstituted with S-peptide deuterated at the C-2 proton of His-12, and by Cu" and 3'-CMP binding, that the resonance with the lower pK, (H3) should be assigned to His-12. Markley (296) from a comparison of the deuterium exchange kinetics of the histidine C-2 protons of RNase A, measured by NMR, with the order of tritium exchange rates into the individual his305. Griffin, J . H., Schechter, A. N., and Cohen, J. S . (1973).Awi. N. Y. Actrtl. Sci. 222, 693. 306. Haar, W., Maurer, W., and Riiterjans, H. H. (1974). EJB 44, 281. 307. Haffner, P. H., and Wang, J. H. (1973). Biochenrisfry 12, 1608. 308. Griffin, J. H., Cohen, J. S., and Schechter, A. N. (1973). Biochemistry 12, 2096. 309. Patel, D.J . , Woodward, C., Canuel, L. L., and Bovey, F. A. (1975). Biopdymers 14, 975.

12. PANCREATIC RIBONUCLEASES

369

tidine residues ( 3 / U ) , confirmed that resonances H4 and H1 could be assigned to His-48 and His-105 (237, 294) but that H3 and H2 should be reassigned to His-12 and His-1 19, respectively. Cohen et a / . (238) demonstrated that the C-2 proton resonances of the active-site histidine residues are essentially identical in RNase A and RNase S. Meadowset a/. (237) had concluded that these resonances were different. These differences were shown to be most probably a result of phosphate in the sample of RNase S (238). Moreover, all four C-2 proton resonances are resolved at pH >5.5 in RNase S. By a comparison of the rate of tritium exchange of the C-2 proton of His-12 of S-peptide with the rates of deuterium exchange of individual C-2 protons of the histidine residues of RNase S, Shindo et 01. (297) confirmed the reassignment of resonance H3 to His-12. A comparison of the titration of the imidazole C-2 protons of N- 1-carboxymethyl-His-119-RNase A with those of the model compounds N-1-carboxymethyl- and N-3-carboxymethylhistidine (p K, values of 6.5 and 5.9, respectively) and N- l,N-3-dicarboxymethylhistidine, permitted reassignment of both resonances H3 and H2 to His-12 and His-1 19 respectively. A, the p K, of His-12 is elevated In N-1-carboxymethyl-His-119-RNase by 1.1 pH units to 7.17 as a result of interaction with the carboxyl group on the modified His-119 side chain. The C-2 proton resonance of the imidazole moiety of carboxymethyl-His-119 was undetected; this possibly indicates that its motion is restricted upon carboxymethylation (297). From the data of Meadows et a/. (237), in the light of the new assignments, it can be concluded that His-12 has a pK, of 6.9 inN-1-carboxymethyl-His119-RNase A, and His-1 19 has a pK, of 6.7 inN-3-carboxymethyl-His-12RNase A (a change of 0.5 units upon derivatization). The pK, values and chemical shifts of the resonances of His-48 and His-105 were not affected in these derivatives. The increased pK, values of the active-site histidine residues in the carboxymethylated derivatives arise from local electrostatic interactions. The reassignment of the H2 and H3 C-2 proton resonances of the active-site histidine residues requires that work (18, 42, 71, 77, 300, 307, 311,312,313) based on the assignments of Meadows l’f d . (237) should be interpreted in this light. Accordingly, where referred to in this review, the appropriate reassignments of the H2 and H3 resonances have been made to the C-2 protons of His-I19 and His-12, respectively. 310. Ohe, M . , Matsuo, H . , Sakiyama, F., and Narita, K . (1974). J . Biochewz. ( T o k y o ) , 75, 1197. 311. Westmoreland, D. G . , and Matthews, C. R. (1973). P N A S 70, 914. 312. Benz, F. M . , and Roberts, G. C. K . (1975). J M B 91, 345. 313. Benz, F. M., and Roberts, G . C. K . (1975). J M B 91, 367.

370

PETER BLACKBURN AND STANFORD MOORE

“1

t

RNase A.

I

9

I

7-0

I

I

1

8

7

6

6 (Ppm) FIG.3. Aromatic region of the 360 MHz proton NMR spectrum of RNase A. The prefixes H, Y, and F denote assignments to histidine, tyrosine, and phenylalanine residues, respectively. From Lenstra ef c d . (80).

The high resolution, 360 MHz proton NMR spectra of RNase A have been reported by Lenstraer a/. (80). Most of the resonances in the aromatic region have been assigned to specific residues (Fig. 3). The assignments, where identified, and their chemical shifts are presented in Table VII. The two spectra shown in Fig. 3, recorded at pH 7.0 and pH 3.4, illustrate resonances that undergo a pH-dependent chemical shift, presented in more detail in Fig. 4. Four singlets H1, H2, H3, and H1’ are identified (80) as histidine resonances on the basis of their titration curves (78, 237, 296-299, 314). The singlet H4 is not observed between pH 5.2 and pH 7.9 (237, 315, 316) except in acetate buffer (79,315,316) and is assigned to the C-2 proton of His-48 (237, 296, 298). H2’ was assigned to the C-4 proton of His-119 based on results from laser-induced photo-CIDNP spectra (see footnote to Table VII) and the effect of the active-site inhibitor 2’-CMP, associating it with the H2 resonance of the C-2 proton of His-119 assigned previously (214, 2%-298). 314. Meadows, D. H. (1972). “Methods in Enzymology” Vol. 26, p. 638. 315. Roberts, G . C. K., Meadows, D. H., and Jardetzky, 0 . (1969). Biochc~niisrry8,2053. 316. Markley, J. L. (1975). Biorhemi.sfry 14, 3554.

12.

37 1

PANCREATIC RIBONUCLEASES 90

85

\'2

-44

.\

80

7s 72

71

g 70 Q

1

og

i

\

69

68

61

66

65

64

3

4

S

PH

6

7

e

FIG.4. The pH dependencies between pH 3.4 and 8.0 of the chemical shifts of aromatic proton resonances of RNase A. The prefixes are those defined for Fig. 3. Crosses, triangular crosses, and triangles denote singlets representing one proton; squares denote doublets, and circles denote nonresolved multiplets. From Lenstra et rrl. (80).

372

PETERBLACKBURN ANDSTANFORDMOORE

Unlike RNase A at neutral pH, RNase S exhibits the H4 resonance of the C-2 proton of His-48 (77, 294, 315). Comparison of the NMR spectra of RNase A and RNase S permitted assignment of resonance H4’ to the C-4 proton of His-48 and thus the remaining resonance, H3’, was assigned to the C-4 proton of His-12 (80). The assignments of H2’ and H3’, as shown here, to the C-4 protons of His-119 and His-12, respectively, as reported by Lenstraet al. (80),agree with the assignments of King and Bradbury (299) and the assignment of H2‘ by Markley (296). However, Markley had assigned the resonance reported here as H3’ (80) to the C-4 proton of His-48. 2. Tvrosine Assigrimerits The proton resonances of tyrosine rings that rotate rapidly about their CB- Cybonds are represented by two mutually coupled doublets, both . such pairs were revealed by corresponding to two protons ( 3 1 6 ~ )Three double resonance experiments (80); they are Y 1, Y2, and Y3. Two other resonances that do not appear as mutually coupled doublets are Y4 and Y5. Y1, Y2, Y4, and Y5 exhibit alkaline shifts between pH 9 and 11, and therefore titrate as normal tyrosine residues. Earlier, Egan et ul. (135) had also found four normally titrating tyrosine residues in 13C NMR studies. This conclusion is contrary to the data obtained by spectrophotometric titration of the tyrosine residues of RNase A [reviewed by Richards and Wyckoff (1) and in Section III,A,7], which indicate three tyrosine residues with normal pK, values and three with abnormally high pK, values. The resonances Y1 and Y2 are identified as being due to exposed tyrosine residues from photo-CIDNP spectra of RNase A (80).They were assigned to Tyr-76 and Tyr-115, respectively, after their elimination from the NMR spectrum following selective nitration of Tyr-115 and Tyr-115 plus Tyr-76 of RNase A with tetranitromethane according to van der Zee et a / . (155).The titration behavior of resonance Y5 between pH 6 and pH 7 in RNase A nitrated at either Tyr-115 or both Tyr-115 and Tyr-76 indicated that Y5 is due to a normally titrating tyrosine residue close to the 3-nitro-Tyr-115 phenolic hydroxyl, which has a pK, of 6.3, and thus Y5 was assigned to Tyr-73 (80). The photo-CIDNP spectra of RNase A and RNase S are essentially identical except for an emission at the position of the upfield Y3 doublet; thus, the tyrosine residue corresponding to Y3 is accessible in RNase S but not in RNase A, and has been assigned to Tyr-25 (80). The fourth normally titrating resonance Y2 is assigned to Tyr-92 (80). 316a. Campbell, I . D . , Dobson, C. M., and Williams, R. J. P. (1975). Proc. Roy. S O C . London A345, 23.

12. PANCREATIC RIBONUCLEASES

373

The two sharp doublets of Y3 in the spectrum of RNase A, seen at low pH are exhibited by RNase S at both neutral and low pH, and indicate rotation about the Cp-Ccy bond of Tyr-25. At neutral pH these Y3 signals are less sharp in RNase A, indicating this movement is more hindered at neutral pH in RNase A. Y3 exhibits a pH-dependent chemical shift with a midpoint at pH 6.2; the other tyrosine resonances are unaffected in this pH range. Lenstra et al. (80) were unable to locate the resonances assigned by Markley (316) to Tyr-25. 3. Conformational Trcinsition Involving H i s 4 The different titration behavior of His-48 in the proton NMR spectra of RNase A in the presence and absence of acetate (294, 315-317), and its different behavior in the proton NMR spectra of RNase A and RNase S, (77, 238) led to the suggestion that His-48 is involved in a local conformational transition of the protein (237, 238, 294, 315-317). The C-2 proton resonance of His-48 of RNase A, observed in acetate buffer, exhibits an inflection in its titration curve near pH 4.2 (77-79, 238); this inflection is not observed with RNase S-protein (239). Moreover, the active-site histidine residues of RNase A (77, 79, 238, 317, 318) and RNase S (79) also exhibit an inflection near pH 4.5 in the titration curves of their C-2 proton resonances. Markley and Finkenstadt (318)have proposed a model based on mutual interaction between His-12 and His-119 to describe the pH titration curves of the respective C-2 proton resonances. They suggest that the low pH inflections of the titration curves arise from a conformational transition with a midpoint at pH 3.7. Cohen and Shindo (79) proposed that the inflections at pH -4.2 in the NMR titration curves of histidine residues 12, 119, and 48 of RNase A are derived from a common event, namely a conformational change involving the side chains of Asp-14 and His-48. In the crystal structure of RNase S ( I )and RNase A (119), the p-COOH of Asp-14 is hydrogen-bonded to the phenolic hydroxyl of Tyr-25. The proton NMR data of Lenstra et 01. (80) for the Y3 doublet assigned to Tyr-25 indicate that in solution at acidic pH, Tyr-25 is not involved in a hydrogen bond. They suggest that the behavior of the Tyr-25 (Y3) resonances may be explained by the breaking of the Tyr-25-Asp-14 hydrogen bond as His-48 is protonated, with formation of a His-48-Asp-14 hydrogen bond. Riiterjans and Witzel (77), and later Santoro et a/. ( 7 4 , based 317. Schechter, A. N . , Sachs, D. H . , Heller, S. R., Shrager, R. I . , and Cohen, J. S. (1972). JMB 71, 39. 318. Markley, J. L., and Finkenstadt, W. R. (1975). Biochemistry 14, 3562.

374

PETER BLACKBURN AND STANFORD MOORE

upon the titration behavior of resonances assigned to Tyr-25 and the /3-COOS of Asp-14 in the proton-decoupled I3C NMR spectra of RNase A, have also proposed a hydrogen bond between Asp-14 and His-48 at acidic pH. They propose that as His-48 is deprotonated, a hydrogen bond forms between the Tyr-25 phenolic group and the Asp- 14 p-COOH or with one of the ring nitrogens of His-48. I3C NMR studies by Cohen and his associates (203,203a),working with the RNase S' complex in which an S-peptide of residues 1 through 15 was and synthesized with 13C-enrichedamino acids at His-12 (C'), Met-13 ((3, Asp-14 (Cy), obtained a pK, of 2.4 for the p-COOH of Asp-14 in the complex by curve fitting $0 the data points; a p Ka of 3.8 was obtained in the free peptide. The titrabion curve of Asp-14 exhibits an inflection with a pK, of 6.1, thought to originate from His-48. On the other hand, the PKa of 2.4 of Asp-14 in this RNase S' complex is too low to account for the inflection seen at pH 4.2 in the proton NMR titration curve of His-48 and those of His-12 and His-119 of RNase A and RNase S. The C' resonance of His-12 of this RNase S' complex exhibited a single pH transition with a pK, of 5.7; an inflection at pH 4.2 was not observed. Santoro et a/. (74) identified several titrating resonances in the carbonyl and carboxyl region of the I3C NMR spectra of RNase A. They tentatively assigned two of these, one to the a-carboxyl of Val-124 (pKa 3.5) and the other to the /3-carboxyl of Asp-14 (pKa 4.33), which demonstrated an inflection and peak splitting with a midpoint in the pH range 6.5 to 7.0, thought to be a result of interaction with His-48. In view of the results of Niu et ol. (203, 203a), this assignment to Asp-14 may require reexamination, but should not be ruled out. 4. His-I 19-Asp121 Interaction

An interaction between these residues was proposed by Sacharovsky et

d.(71). Such an interaction could perhaps explain the results obtained

by Santoro et ul. (74) previously described. However, they consider this unlikely since they observed a broadening and splitting of the carboxyl resonance around the imidazole p K,, which they suggest is from His-48. Also, based on curve fitting to their data on the low pH inflection for the C-2 proton resonance of His-119, Cohen and Shindo (79) consider this His-1 19-Asp-121 interaction unlikely. Deprotonation of the imidazolium form of a histidine residue can yield either the Nc2-H(N-3-H)tautomer or the N"-H (N-1-H) tautomer of the imidazole form of the residue. From the titration behavior of the resonances of the nonprotonated aromatic carbons assigned to the four histidine residues in the I3C NMR spectrum of RNase A in acetate buffer, it was found by Walters and Allerhand (75)that only His- 119 exists predom-

12. PANCREATIC RIBONUCLEASES

375

inantly as the N"-H tautomer. The imidazole forms of His-12 and His-48 are predominantly in the N6'-H tautomeric form; His- I05 appears to be between 50 to 90% as the N"-H tautomer. The NC2-Htautomer is predominant for 1-histidine and a number of L-histidyl peptides (319-321). Stabilization of the N6'-H tautomeric forms in the protein is thought to result from hydrogen bonds between residues His-48 and Asp-14, His-I2 and Thr-45, and between NS'-H of His-105 and the a-carboxyl group of Val-124 ( 7 3 . Thus, evidence for a hydrogen bond between the p-COOS of Asp-121 and NC2of His-119 was not found by this technique. 5 . Lysine Amino Groirps

Direct observation of the lysine e N H 2 groups of proteins by NMR spectroscopy is not possible due to the rapid rates of exchange of the amine protons. Lysine residues may be studied, after reductive methylation with formaldehyde and borohydride, by the proton resonances of the resultant N-methyl groups (18, 322). Titration of the N-methyl proton resonances of fully reductively methylated RNase A permitted assignments for the derivatives of the a-NHz group of Lys-1 and the e N H 2 group of Lys-41, and enabled pK, values, corrected for the effects of methylation (323), of 6.6 and 8.8 to be determined, respectively, for their amino groups (18). The pK, values obtained for the other nine lysine e N H 2 groups were between 10.2 and 10.8 (18). The pK, of 8.8 obtained for the e N H 2groups of Lys-41 agrees closely with previously determined values (41, 324, 325). Lys-41 is the last lysine residue to react upon guanidination with l-guanyl-3,5-dimethylpyrazole( I 7); reaction of Lys-41 is accompanied by greater than 95% loss of enzymatic activity (19). The proton NMR spectra of RNase guanidinated on 9 or 10 lysine residues and of RNase A reductively methylated are essentially identical with that of the native enzyme and undergo similar changes on thermal unfolding (18, 19). Following reductive methylation of the e N H Zgroup of Lys-41 of RNase guanidinated on 9 lysine NH2groups, titration of the N-methyl proton resonances indicated a pK, of 8.8 for the e N H 2 group of this residue. The major loss of activity upon modification of Lys-41 does not appear to result from any 319. Reynolds, W. F., Peat, I . R., Freedman, M . H . , and Lyerla, J. R., Jr. (1973). JACS 95, 328. 320. Wasylishen, R. E., and Tomlinson, G. (1977). Cnn. J . Biorhcm. 55, 579. 321. Blomberg, F., Maurer, W., Riiterjans, H. (1977). JACS 99, 8149. 322. Bradbury, J. H . , and Brown, L. R . (1973). EJB 40, 565. 323. Perrin, D. D. (1964). Airst. J . Cl7m. 17, 484. 324. Murdock, A. L . , Grist, K. L., and Hirs, C . H . W. (1966). A B B . 114, 375. 325. Carty, R . P., and Hirs, C . H. W. (1968). JBC 243, 5244.

376

PETER BLACKBURN AND STANFORD MOORE

major conformational change in the protein arising from the chemical modifications. Moreover, similar chemical shifts and pK, values for the active-site histidine residues were observed for the fully methylated and native RNase (18). Upon binding phosphate, the pK, of EN-methyl-Lys-41 was increased by 0.3 pH units. Similar pK, values for the respective lysine residues of RNase were obtained by Jentoft et al. (20, 34) in I3C NMR studies after reductive methylation of the enzyme with [‘3C]formaldehyde and cyanoborohydride. Binding of 3’-CMP by the fully methylated RNase caused a small but significant shift in the resonance assigned to eN.N‘-dimethyl-Lys-41 but was without effect on the other N-methyllysine resonances. The spin lattice relaxation time and nuclear Overhauser enhancement effect for theN, N’-dimethyl-Lys-41resonance indicate that the e N H 2 group of this residue is in a relatively restricted environment (37). The pH titration curve of this resonance has two inflections, with pK1 at 9.0 and pK2 at 5.7. Based upon a comparison of the effects of various substrate analogues on pK2 and the pK, values of the active-site histidine residues it was suggested that pK2 arises by a perturbation from His- 12, possibly via a conformational transition of the protein. Markley (316) postulated two slow conformational transitions of the enzyme, one involving His-48, as described previously, and a second with a midpoint at pH 5.6, which might be due to titration of His-12. Also, Riiterjans and Witzel (77) postulated that an electrostatic interaction between the e N H 2 group of Lys-41 and the active-site histidine residues could explain the strong dependence of their pK, values on ionic strength. 6 . Inntemctions wirli Substrate Analogs a. Mononucleotide Phosphates. The effects of nucleotide monophosphates on the histidine C-2 proton resonances have been described by a number of authors (42, 72, 77, 294, 300, 306, 307). The C-2 proton resonances of the active-site histidine residues 12 and 119 shift downfield on binding the pyrimidine mononucleotides 2’- and 3’-CMP; the major shift is that of H3 of His-12. The shift seen on binding of 2’-CMP or 3’-CMP by the C-2 proton resonance of His-12 is beyond the position of a fully protonated histidine, resembles that seen on formation of an imidazoliumphosphate complex (300, 326), and suggests direct contact between the imidazolium group of His-12 and the phosphate of 2’-CMP or 3’-CMP. When 5’-CMP or phosphate binds to RNase, this extensive shift is not observed. The effects on the C-2 proton resonance of His-119 are similar for all three mononucleotides. 326. Cohen, J . S. (1968). BBRC 33, 476.

12. PANCREATIC RIBONUCLEASES

377

The C-2 proton resonance of His-48 is shifted slightly downfield by the binding of nucleotide monophosphates, but not by phosphate (300) or pyrophosphate (3061,whereas that of His-105 is not shifted at all. Binding of pyrmidine nucleotide monophosphates to RNase A also produces an upfield shift of an aromatic resonance (294, 300, 306). Four resonances designated F1 through F4, which correspond to the three phenylalanine residues of RNase A, are shown in the high resolution proton NMR spectrum reported by Lenstra et al. (80) and presented in Fig. 3. All four of these resonances are affected upon binding of the active-site inhibitors 2'-UMP or 2'-CMP (Fig. 5). Resonance F3 is most strongly affected and is tentatively assigned to Phe-120 (80,300)since this residue is most directly involved in binding active-site inhibitors (21, 73). A similar shift in the I3C NMR resonance assigned to Phe-120 was also reported by Santoro et al. (74).A resonance, designated I (Fig. 5 ) , which originates from the enzyme-nucleotide complex, titrates with a small shift between pH 4 and 5 and has been assigned to Phe-120 (80,294,300).There appears to be no shift in resonances due to Tyr-25 [contrary to the suggestion by Markley (316) and Antonov et af. (72)1, since the Y3 doublet of Tyr-25 was not affected (80.).The resonances assigned to His-105, and tyrosine residues 25, 73, 76, 92, and 115 do not shift on the binding of either 2'-UMP or 2'-CMP (80). By contrast, the purine mononucleotides 2'- and 3'-AMP have no effect on the C-2 proton resonances of either His-48 or His-105 (306). No downfield shift of the C-2 proton resonance of His-12 is seen on binding of 5'-AMP. Moreover, 2'-, 3', and 5'-AMP are all without effect on the Phe-120 resonance (306, 307). 6. Histidine p K , Changes. The effects of various active-site inhibitors on the pK, values of the histidine residues of RNase A and some of its derivatives are summarized in Table VIII. The various mononucleotide phosphates are not equivalent in their effects, especially with regard to the pK, of His-12. Meadows et al. (300) suggest that the magnitude of the effect on the pK, of His-12 reflects the order of the binding constants, 2'-CMP > 3'-CMP > 5'-CMP. The effect on the pK, of His-12 cannot depend exclusively on the geometry of the phosphate esters, since 2'-UMP and 3'-UMP have much less influence on the pK, of His-12 (305, 306) than do 2'-CMP and 3'-CMP, and do not cause the major downfield shift of its C-2 proton resonances seen with these cytidine mononucleotides (42, 306). Also 2'-UMP and 2'-CMP do not equally affect the resonances F1, F2, and F3 (80). Only small effects on the pK, of either active-site histidine are seen with 2'- and 3'-AMP (306). By contrast, 5'-AMP significantly raises the pK, values of both active-site histidine residues (306,307).

378

PETER BLACKBURN AND STANFORD MOORE 3 mM RNae A, pH 7 1

I

7

65

C-l'H

C-5H

6

6 (ppm) FIG. 5 . Aromatic regions of the proton NMR spectra of RNase A in the presence of active site inhibitors. The dashed lines indicate the positions of the F resonances at pH 7.1 without inhibitor (see Fig. 3). The C resonances are from the inhibitor. From Lenstra e l d. (80).

379

12. PANCREATIC RIBONUCLEASES TABLE VIII T H EEFFECTO F N U C L E O T I D B EI N D I NOGN T H E pK, RIRONUCLEASE A

OF

H I S T I D I NRESIDUES E OF

Histidine residue p K , Nucleotide

His-I2

His-119

His-48

His-105

Native enzyme + 2'-CMP + 3'-CMP + 5'-CMP + Z'deoxy-3'CMP + 2'-UMP + 3'-UMP + 2'-AMP + 3'-AMP + 5'-AMP + Phosphate + UpcA + 2'-FdUpA + 2'-FdUp(Me) + 2'-FdUp Des-( 12I - 124) + 3'-CMP S-protein + Phosphate

5.8

6.2 8.0 7.9-8.0 8.0 8.0 7.8 7.8-8.0

6.3-6.6

6.7 6.7 6.7 6.7 6.7 6.7 6.7 6.7 6.7 6.7 6.7

6.4 6.4 6.3 6.6

6.7 6.7 6.7 6.6 6.7 6.7 6.9

8.0 7.4-7.5 7.0 6.5 6.3 6.2-6.4 6.3 6.0 6.3 6.6 5.8

6.6 6.5 7.3 7.1 7.5

-

6.0 7.6 6.9 6.1 6.7 6.7 7.8 6.8 7.4 6.98 7.4

5.6 6.0

Ref

c. Efects on Sirbstrate Resonances. Two new resonances appearing at high field (Fig. 5 ) are identified as the proton resonances of C-1' of the ribose and C-5 of the pyrimidine base (300, 316, 327). The proton resonances C-5-H and C-6-H of the cytidine bases of 2'-, 3'-, and 5'-CMP, all shift downfield similarly on binding to the enzyme. The proton resonances C-1'-H of the ribose rings are also downfield shifted similarly €or 3'- and 5'-CMP, but slightly differently for 2'-CMP (300). Thus all three mononucleotide bases bind similarly to the enzyme. The shift to lower field of the C-6 and C-5 proton resonances is attributed to anisotropy caused by the base-stacking with Phe-120 (77,300,306);a similar shift in the C-6 proton resonance of 2'- and 3'-UMP is also observed (306). The C-2 and C-8 proton resonances of the adenine base of 2'-, 3'-, and 5'-AMP are all shifted slightly upfield on binding to the enzyme and may arise in part from a stacking interaction with His-1 19 (306, 307). 327. Gorenstein, D. G., and Wyrwicz, A. (1974). Biorlwmisrr~13, 3828.

380

PETER BLACKBURN A N D STANFORD MOORE

d. Interactions with the Phosphate Group. While 31Pchemical shifts are a function of the geometry of the phosphate ester (328, 329), they are relatively insensitive to the nature of the group bonded to the phosphate oxygen (328). Gorenstein et a/. (328. 330) have reported the pH dependence of the 31Presonance of pyrimidine nucleotides, 2'(3')- and 5'-CMP, and 3'-UMP, both free in solution and when bound to RNase A. Two ionizations are observed for the complex, found also by Haar et ul. (331), with pK, of 4.0-5.5 and pK2 of 5.9-6.7. The pK1 is associated with ionization of the monoanionic inhibitor, and pK2 with ionization of the protonated His-12, which hydrogen bonds to the phosphate. Gorenstein et a/. (328) calculated the microscopic pK, values that describe the pHdependent equilibria between the monoanionic and dianionic forms of the phosphonucleotides with the enzyme protonated and unprotonated at His-12. For the unprotonated form of His-12, the microscopic pK, for the ionization of the monoanionic phosphonucleotide is nearly the same as the ionization constant for the phosphonucleotide free in solution, which suggests little interaction between the phosphate and the other protonated groups at the active site, His-119, and Lys-41. The pK, values of the 2'and 3'- (but not the 5'-) monoanionic phosphonucleotides bound to the protonated His-12 enzyme are perturbed by 1.5 to 2.0 pH units, a result consistent with the stabilization afforded by hydrogen bonding between the 2'- and 3'-phosphates and His-12 (328). Binding of 3'-CMP to RNase A decreases the spin-lattice relaxation time, T I , of the C-2 proton resonances of His-I2 by 25%, and that of H i s 4 by lo%, but is without effect on those of His-119 or His-105 (332), and further demonstrates the close association between His-12 and this mononucleotide phosphate. The decreased relaxation times, chemical shift changes, and titration behavior of the C-2 proton resonances of His- 12 and His-48 suggest local conformational changes upon binding of active-site inhibitors. e . Ititercrctiuns witlz Dinucleotide Substrate Analogs. The mononucleotide phosphates, around neutral pH, bind in the dianionic state, whereas the phosphodiester substrate is monoanionic and would be expected to have less effect on the pK, values of the active-site histidine residues as a result of charge interactions. Griffin et a / . (305)studied the binding of UpcA, an analog of UpA in which a methylene group re328. 329. 330. 331. 332.

Gorenstein, D. G., Wyrwicz, A. M., and Bode, J. (1976). JACS 98, 2308. Blackburn, G. M . , Cohen, J. S., and Weatherall, I . (1971). Tetrahedron 27, 2903. Gorenstein, D. G., and Wyrwicz, A. (1973). BBRC 54, 976. Haar, W., Thompson, J. C., Maurer, W., and Riiterjans, H. (1973). EJB 40, 259. Benz, F. W., Roberts, G . C. K., Feeny, J., and Ison, R . R. (1972). BBA 278, 233.

12. PANCREATIC RIBONUCLEASES

38 1

places the 5’-oxygen of the adenosine nucleoside (334, and is not cleaved by the enzyme. For the binding of this inhibitor, Griffin et al. (305) found that the pK, values of His-12 and His-119 were not significantly altered. Moreover, they found, by NMR studies (305, 329), little effect on the ionization or conformation of the phosphonate group, which these authors suggest indicates little direct interaction between the phosphonate group and the active-site histidine residues in the RNase A-UpcA complex (305). Antonov rt a/. (72) studied the binding of a nonhydrolyzable analog of UpA to RNase A, 2’-deoxy-2’-fluorouridyl-3‘-p-5’-adenosine (2’-FdUpA). The binding of 2‘-FdUpA, 2’-FdUp methyl ester and 2’-FdUp were studied by proton NMR and 31PNMR. The position of the purine and pyrimidine bases in complex with RNase A were identified from the chemical shifts of the C-6 and C-5 proton resonances of the pyrimidine base, the C-1’ proton resonance of the uridylribose, the C-2 and C-8 proton resonances of the purine base, and the C-1‘ proton resonance of the adenosylribose. The results indicated that the positions of the pyrimidine bases of 2’-FdUpA and 2’-FdUp methyl ester and 2’-FdUp (72) were identical to those previously observed for pyrimidine mononucleotides (300, 306). Binding of the purine base of 2’-FdUpA (72) was somewhat different from that observed for 5’-AMP (306). The chemical shifts of the proton resonances of 2’-FdUpA in complex with RNase A and when free in solution were compared. Measurements were made of the nuclear Overhauser effect for the C-I’ proton resonances of the ribose moieties and those of the C-6 (pyrimidine) and C-8 (purine) proton resonances to determine the glycoside torsion angles in both fragments of the dinucleoside monophosphate when bound to RNase A. The results indicated that the dinucleotide monophospate is bound at the active site in an extended conformation with both nucleotides in the anti conformation. The effects of these fluorine nucleotide analogs on the pK, values of the histidine residues of RNase A are shown in Table VIII. As expected, binding of the monoanionic dinucleotide phosphate and the methyl ester have less of an influence on the pK, values of the active-site histidine residues. A comparison of the effect of 3‘-UMP on the pK, of His-12 with that of 2’-FdUp suggests that the fluorine atom at the 2‘ position of the ribose influences the pK, of His-12, increasing it by about 1 pH unit, and demonstrates the close proximity expected between the 2’-OH of the ribose in UpA and His-12. If the effect of the 2’-fluorine is taken into account, binding of the 333. Jones, G . H . , Albrecht, H. P., Damodoran, N . P., and Moffatt, J . G. (1970). JACS 92, 5510.

382

PETER BLACKBURN AND STANFORD MOORE

substrate UpA would not be predicted to affect the pKa of His-12 (72), which agrees well with the results for UpcA reported by Griffin et al. (305). C. OPTICAL PROPERTIES 1. UV Absorption Spectra

Spectrophotometric titrations to study the PKa values of tyrosine residues in RNase are facilitated by the fact that the bovine enzyme and all of the other mammalian pancreatic RNases studied to date (see Section V) contain no tryptophan residues. The absorption near 280 nm is almost entirely due to tyrosine residues. UV spectra have been used to study the degree of exposure of tyrosine side chains in RNase A [reviewed in Ref. (I); also see Ref. (334) for the application of derivative spectra] and the effects of derivatization of available phenolic groups by iodination or nitration (see Section III,A,7). NMR spectroscopy has provided independent data on the pK, values of tyrosine residues (Section IV,B).

2. Circular Dichroism The characteristics of the CD spectra of RNase A in the far- and nearUV regions have been reviewed by Timasheff (335) and by Richards and Wyckoff (I). The near-UV CD spectrum obtained at 25” is characterized by a positive band with a maximum near 240 nm, and a negative band near 275 nm. Strickland and his colleagues (336-338) have shown that the resolution between 250 and 320 nm can be improved by conducting the experiments at 77” K in 1 : 1 water: glycerol solutions (Fig. 6). The nearUV CD spectrum of RNase A shows changes in inflection between 276 and 283 nm that arise from exposed tyrosine residues. The shoulder at 289 nm observed with RNase A (and missing with RNase S, where it may be shifted to about 286 nm) is attributed to a tyrosine residue (probably Tyr-25) that becomes more accessible in RNase S. The shoulders at 268, 261, and 255 nm are attributed to phenylalanine residues. Studies with derivatives of tyrosine have suggested that the major source of tyrosine circular dichroism bands in proteins arises from dipole-dipole coupling between the near-UV transition of tyrosyl side 334. 335. 336. 337. 338.

Brandts, J. F., and Kaplan, L. J. (1973). Biochemistry 12, 2011. Timasheff, S . (1970). “The Enzymes,” 3rd ed., Vol. 2, p. 371. Horwitz, J., Strickland, E. €I and .,Billups, C. (1970). JACS 92, 2119. Honvitz, J., and Strickland, E. H. (1971). JBC 246, 3749. Strickland, E. H. (1974). CRC Crit. Rev. Biochern. 2, 113.

12. PANCREATIC RIBONUCLEASES

383

260 270 280 290 300 Wavelength Inm)

FIG.6. Comparison of the CD spectra of RNase A and RNase S at 77°K. From Horwitz and Strickland (337).

chains with strong far-UV transitions of other nearby moieties. Strong coupling interactions are expected primarily with aromatic amino acids, peptide bonds, and other groups having T orbitals (339-341). Strickland (342) has examined the near-UV CD bands of RNase S on the basis of atomic coordinates for the crystalline enzyme ( 1 ) and theoretical predictions of dipole-dipole interactions within the RNase molecule in order to calculate the rotatory strengths of the individual residues in the near-UV. The interactions between Tyr-73 and Tyr-115 are predicted to provide a major contribution to the negative bands at 276 and 283 nm. Interactions between tyrosine and phenylalanine transitions are considered to contribute little to the optical rotatory properties of the molecule. Despite the numerous peptide groups surrounding some tyrosine side chains, the total rotatory strengths from tyrosyl coupling with peptide bonds are limited by cancellation of contributions having opposite signs. Most of the difference between the calculated and experimental CD spectra comes from the -S-Sbonds of cystine, which contribute to the broad valley (Fig. 6) upon which the finer structure is superimposed. At low temperature (77" K), the bands between 250 and 300 nm in the CD spectra of RNase A (336) and RNase S (337) exhibit similar intensities to those observed at ambient temperature; this result suggests that the 339. 340. 341. 342.

Hooker, T. M . , Jr., and Schellrnan, J. A. (1970). Biopolymers 9, 1319. Chen, A. K . , and Woody, R. W. (1971). JACS 93, 29. Hsu, M.-C., and Woody, R. W. (1971). JACS 93, 3515. Strickland, E. H. (1972). Biochemistry 11, 3465.

384

PETER BLACKBURN AND STANFORD MOORE

contributing tyrosine residues have relatively restricted motions (342). Upon cooling solutions of N-acetyl-0-methyl-L-tyrosine ethyl ester from 297" K to 140" K, the negative rotational strength exhibited by this model compound was intensified 10-fold (343). However, Pflumm and Beychok (344) showed that the positive elipticity observed at 240 nm with RNase is temperature-sensitive; a change from 25" to 5" causes an increase in intensity of two- to threefold. Goux and Hooker (345)have extended the calculations of mean residue rotatory strengths for side chain transitions to the positive elipticity at 240 nm. The calculations are based upon the three-dimensional structure ( 1 ) and are fitted to CD data obtained at 27". Their predictions are in general agreement with those of Strickland (342). The contributions from disulfides and the interaction between Tyr-73 and Tyr-115 are considered to be significant for both the positive and the negative bands in the spectrum. Several applications of CD measurements in the course of researches covered in this chapter are included in Sections 111,A; IV,D; V; VI; and VII. 3. Fluorescence Seagle and Cowgill (156, 158) have extended the studies by Cowgill [cf. Ref. (Z)] on the use of fluorescence to study the tyrosine residues in RNase. In the native enzyme, the fluorescence exhibited by RNase A arises from the normally titrating surface tyrosine residues of the molecule. The low quantum yield results from quenching by disulfide bonds and through hydrogen bonding that involves tyrosine residues 25,92, and 97. Seagle and Cowgill (156) reported on the fluorescence characteristics of RNase A modified with tetraintromethane (147)and reduced with bisulfite to produce 3-amino-Tyr residues (148). For one tyrosine residue that was unusually susceptible to such modification, the fluorescence emission maximum was near 395 nm, a high wavelength for 3-amino-Tyrin peptides and on the surface of fibrous proteins; generally A,, is at 350 to 370 nm (158) (see Section III,A,7). Churchich (346) has reported on the fluorescence properties of 4-pyridoxic-5'-phosphate,a competitive inhibitor of the enzyme, with a K1 of M , that is bound firmly to the active site as demonstrated by nanosecond emission anisotropy measurements. Churchich and Wampler (347)reported on the luminescent properties of RNase A at 77" K. The low 343. 344. 345. 346. 347.

Strickland, E. H., Wdchek, M., Horwitz, J., and Billups, C. (1972). JEC 247, 572. Mumm, M. N., and Beychok, S. (1969). JBC 244, 3973. Goux, W. J . , and Hooker, T. M., Jr. (1980). JACS 102, 7080. Churchich, J. E. (1976). Modern Fluorescence Spectroscopy 2, 217. Churchich, J . E., and Wampler, J. (1971). BBA 243, 304.

12. PANCREATIC RIBONUCLEASES

385

phosphorescence yield of RNase A is related to the quenching effect exerted by disulfides on the triplet-excited state of tyrosyl residues. Grandi er al. (348), in studies on the dimeric RNase of bovine seminal plasma, used fluorescence-quenching experiments to observe that the tyrosine residues of the seminal dimer are less exposed to solvent than those of the monomeric species. Iodide and cesium ions and acrylamide [cf., Refs. 349, 3501 have minimal quenching effect with the dimer.

D. THEFOLDING PATHWAY The folding of ribonuclease to its native conformation proceeds in a directed pathway to a final structure of overall minimum free energy (351); the nature of the process has been the subject of several reviews (123,352,353). Interactions within the polypeptide chain must dictate the pathway and the final conformation. Short-range interactions between amino acid residues dominate the determining factors that contribute to the native conformation of a globular protein such as RNase A (354-358). The precise orientation of an amino acid residue in the h a 1 conformation of the protein is further refined via medium-range (359) and long-range interactions with other regions of the polypeptide chain (360). The unfolded protein molecule has been described by Anfinsen and Scheraga (361) as “a fluctuating ensemble of conformations.” By a random mechanism, the folding of a protein molecule would take an inordinately long time (362,363) and can be ruled out (353). The formation of a nucleus, which would then direct subsequent protein folding, has been postulated as the rate-determining step. Both hydrogen-bonded sec348. Grandi,,G., D’Alessio, G., and Fontana, A. (1979). Biochemistry 18, 3413. 349. Eftink, M . R., and Ghiron, C. A. (1976). Biochemistry 15, 672. 350. Eftink, M. R., and Ghiron, C. A. (1977). Biochemistry 16, 5546. 351. Anfinsen, C. B . (1973). Science 181, 223. 352. Baldwin, R. L., and Creighton, T.E . (1980). In “Protein Folding” (R. Jaenicke, ed.), p. 217. Elsevier, Amsterdam. 353. Baldwin, R. L. (1980). In “Protein Folding” (R. Jaenicke, ed.), p. 369. Elsevier, Amsterdam. 354. Kotelchuck, D., and Scheraga, H. A. (1968). PNAS 61, 1163. 355. Kotelchuck, D., and Scheraga, H. A. (1%9). PNAS 62, 14. 356. Finkelstein, A. V., and Ptitsyn, 0. B. (1971). J M E 62, 613. 357. Scheraga, H. A. (1973). Pure Appl. Chem. 36, 1 . 358. Scheraga, H. A. (1978). Pure Appl. Chem. 50, 315. 359. Ponnuswamy, P. K., Warme, P. K . , and Scheraga, H. A. (1973). PNAS 70, 830. 360. Burgess, A. W., and Scheraga, H. A. (1975). PNAS 72, 1221. 361. Anfinsen, C. B., and Scheraga, H. A. (1975). Advan. Protein Chem. 29, 205. 362. Levinthal, C. (1968). J . Chem. Phys. 65, 44. 363. Karplus, M., and Weaver, D. L. (1976). Nature (London) 260, 404.

386

PETER BLACKBURN AND STANFORD MOORE

ondary structures (364, 365) and a cluster of hydrophobic residues (366, 367) have been proposed to form such a nucleus. An alternative view is that folding proceeds via discrete intermediates that determine the pathway and the rate of folding. These alternatives have been discussed by Baldwin (353). 1. Equilibrium Studies

a. Stability of the Native Conformation. The unfolding transition of RNase A was observed originally by spectrophotometric measurements of tyrosine absorption (368-370). Since then, a number of techniques have been applied to the study of this problem. Calorimetric measurements give direct thermodynamic data on the stability of the native conformation, and have demonstrated that the folding transition for RNase A is a highly cooperative process (371). The native protein has been shown to exhibit substantial heat uptake prior to entrance into the cooperative transition zone (371, 372). A pretransition zone preceding unfolding by guanidinium chloride has also been demonstrated (373). These results suggest a loosening of the native structure prior to the cooperative unfolding of the molecule. Laser Raman spectroscopy, which measures the vibrational frequencies of a given class of groups, has been used to study the thermal unfolding of RNase A (374). The results provided evidence for stepwise unfolding of the protein between 32 and 70". NMR spectroscopy has been used to follow the folding transition of RNase A (80, 237, 307, 312). The NMR spectrum of the random coil conformation of a protein represents the sum of the superposition of the spectra of the constitutent amino acids (375,376).The area of a particular resonance is a direct measure of the quantity of the species responsible for that resonance. Thermal and acid unfolding of RNase A produces shifts in 364. Ptitsyn, 0. B., Finkelstein, A. V., and Falk, P.(1979). FEBS Lett. 101, 1. 365. Lim, V. I. (1978). FEES Lett. 89, 10. 366. Wuthrich, K., and Wagner, G . (1978). Trends Biochem. Sci. 3, 227. 367. Matheson, R. R., Jr., and Scheraga, H. A. (1978). Macromolecules 11, 819. 368. Harrington, W. F., and Schellman, J. A. (1956). CR Truv. Lab. Curlsberg, Ser. Chim. 30, 21. 369. Hermans, J . , and Scheraga, H. A. (1961). JACS 83, 3283. 370. Scott, R. A., and Scheraga, H. A. (1%3). JACS 85, 3866. 371. Privalov, P. L., and Khechinashvili, N. N. (1974). J M B 86, 665. 372. Matheson, R. R., Jr., Dugas, H . , and Scheraga, H. A. (1977). BBRC 74, 869. 373. Miller, J. F., and Bolen, D. W. (1978). BBRC 81, 610. 374. Chen, M. C., and Lord, R. C. (1976). Biochemistry 15, 1889. 375. Cohen, J. S . , and Jardetzky, 0. (1968). PNAS 60, 92. 376. McDonald, C. C., and Phillips, W. D. (1969). JACS 91, 1513.

12. PANCREATIC RIBONUCLEASES

387

the C-2 proton resonances of His-12, His-119, and His-105 that can be accounted for by the changes in the pK, values and protonation of these residues. However, the C-2 proton resonance of His-48 is an exception; the shift in its C-2 proton resonance indicates that a conformational change occurs in its locale in the pretransition zone. Urea and guanidinium chloride both produce large chemical shifts in the C-2 proton resonances of the histidine residues of RNase A (313). No effect is observed on the resonance of His-105. A small downfield shift of His-119, and a small upfield shift of His-48, are observed throughout changes in concentrations of guanidinium chloride that are insufficient to effect the major unfolding transition of the molecule (up to 1.3 M guanidinium chloride). A large downfield shift for the C-2 proton resonance of His-12 is observed throughout a wide concentration range of the denaturants. Except for that of His-12, the shifts are completed below 1.3 M guanidinium chloride. The areas of the histidine C-2 proton resonances, which give a direct measure of the relative populations of those molecules with an individual histidine in a folded (not necessarily native) state, decrease with increased concentrations of denaturant above 2.0 M guanidinium chloride. The sum of the areas of the individual folded resonances and the resonance that corresponds to the unfolded state remains constant throughout the unfolding transition. The resonance of each histidine residue exhibits a different dependency on denaturant concentration, suggesting that partially folded species must exist during the unfoldirig transition. Chemical trapping experiments have been performed to follow stages in the unfolding of RNase A. One procedure has been to use proteases, which generally hydrolyze the peptide bonds of unfolded proteins much more rapidly than those of the native molecule. RNase A is a particularly good protein for this approach, since its native structure is very resistant to most proteases; as the protein unfolds with increasing temperature, specific peptide bonds are cleaved (377-381). The order in which particular peptide bonds hydrolyze has been taken to indicate the order in which corresponding sections of the molecule unfold. Since the time spans of unfolding steps are commonly of the order of milliseconds (3731, chemical reactions used to trap intermediate stages of unfolding should be fast; photochemical labeling has been used to this 377. Rupley, J . A., and Scheraga, H. A. (1963). Biochemistry 2, 421. 378. Ooi, T., Rupley, J. A., and Scheraga, H. A. (1963). Biochemistry 2, 432. 379. Ooi, T., and Scheraga, H. A. (1964). Biochemistry 3, 648. 380. Klee, W. A. (1967). Biochemistry 6, 3736. 381. Burgess, A. W., Weinstein, L. I., Gabel, D., and Scheraga, H. A. (1975). Biochemistry 14, 197.

388

PETER BLACKBURN AND STANFORD MOORE

end. Information on the thermal unfolding of RNase has been gained using an aryl nitrene generated by flash photolysis (184-186, 188) (see Section III,A,&b); amino acid analysis indicates which types of residues undergo modification. 6 . Contributions of Terminal Segments to the Conformational Stability. The S-protein is folded significantly like RNase A as evidenced by the chemical shifts and titration behavior of the C-2 proton resonances of His-48 and His-105, which are similar to those of RNase A (239). However, the C-2 proton resonance corresponding to histidine residues that are fully exposed to solvent is present to varying degrees in the proton NMR spectra of S-protein at all pH values; there is a pH-dependent equilibrium between folded and unfolded forms of the S-protein. The C-2 proton resonance of His-1 19 is affected by removal of S-peptide, yet this resonance is also shifted significantly by phosphate, suggesting some residual phosphate binding capacity of S-protein (239). Furthermore, it has been shown that S-protein undergoes cooperative thermal unfolding (197, 382), although the melting temperature and thermodynamic parameters (383, 384) indicate that the protein is more labile and exhibits a broader thermal transition than does RNase A. The importance of the six carboxyl terminal residues of RNase to the proper folding of the chain is discussed in Section 111,B,2. c. Properties of the Unfolded Molecule. RNase A, reduced and under nondenaturing conditions, is largely unfolded with all six tyrosine residues exposed to solvent; however, far-UV circular dichroism measurements differ from those typical of a random coil and indicate the presence of some ordered structure (127, 384a). Moreover, fully reduced RNase A retains 0.04% of the activity of native RNase A toward cyclic 2',3'-CMP, with the same substrate concentration-dependence and characteristics of inhibition by 2'-CMP as those of native RNase A (385).This activity is lost after treatment with the sulfhydryl reactant N-ethylmaleimide, indicating that the species responsible for this activity are at most only partially reoxidized (385). NMR studies have shown that, above pH 5.0, thermally unfolded 382. Shenvood, L. M., and Potts, J. T., Jr., (1965). JBC 240, 3799. 383. Tsong, T. Y., H e m , R. P., Wrathall, D. P., and Sturtevant, J. M. (1970). Biochemistry 9, 2666. 384. Rocchi, R., Borin, B., Marchion, F., Moroder, L., Peggion, E., Scoffone, E., Crescenzi, V., and Quadrifoglio, F. (1972). Biochemistry 11, 50. 384a. Takahashi, S., Kontani, T., Yoneda, M., Ooi, T. (1977). J . Biochern. (Tokyo) 82, 1127. 385. Garel, J.-R. (1978).IMB 118, 331.

389

12. PANCREATIC RIBONUCLEASES

RNase, although similar to a random coil (80, 96,31 f -313, 386, 387), still retains some residual structure which may be further unfolded by or by thermal unfolding below pH guanidinium chloride or urea (80,313), 2.0 (3ff , 312). Two of the resonances that shift upon addition of denaturant arise from incompletely exposed His-48 and His-105 (96), and others arise from tyrosine and phenylalanine residues (80). d. Irnrnunochernical Estimation of Protein Conformation. Using an immunochemical approach, Chavez and Scheraga (388) have determined the equilibrium constant, Kconffor the equilibrium between unfolded and folded forms of RNase A and some of its derivatives. Thus, for the equilibrium

where P, and P, are the protein in its random and native forms. The association with anti-native RNase antibodies is defined by Ab

+ P,

Z

AbP,, where K,,,

=

[AbP,I/[Abl[P,I

Thus,

and may be calculated from measurable parameters. At 4" and pH 8.3, for native RNase A the value of Kconf is very large. For reduced RNase A the value of Kconfis 0.06; thus, the antigenic deterof the minants of fully reduced RNase A have about 6% (Kconf/l + &,) native conformation of RNase A at equilibrium. This represents a high degree of structure in the fully reduced molecule, since Kconffor a single unstructured determinant is estimated to be to low6(388). Removal of residues 121-124 reduces Kconffor each of the antigenic regions [average value = 0.29, T,,, = 44.5" (252)l; when Phe-120 and His-119 are also removed, Kconffor each region is decreased markedly [average value = 6.8 x lo+, T,,, = 32.5" (252)]. Deletion of the S-peptide to give the Sprotein also reduces Kconf[average value = 0.07, T, = 37.5" (38311. The data clearly indicate the importance of residues in the COOH- and NH2-terminal regions for stability of the native conformation of RNase A. 386. Howarth, 0. W. (1979). BBA 576, 163. 387. Sadler, P. J . , Benz, F. W., and Roberts, G. C. K. (1974). BBA 359, 13. 388. Chavez, L. G . , Jr., and Scheraga, H. A. (1980). Biochemistry 19, 996; ibid., 1005.

390

PETER BLACKBURN AND STANFORD MOORE

2, Kinetic Studies a . Refolding of RNase with Intact Disuljides. RNase, with its native disulfides intact, that has been unfolded by heat or guanidinium chloride, refolds rapidly according to biphasic kinetics when followed by tyrosine absorption or fluorescence emission, and by difference spectra upon 2'-CMP binding. The process has been described by Baldwin and associates (154, 389-392) as one in which Us

k, kz * UF * N k-, k-,

where Us (80%) and UF (20%) represent slow- and fast-folding forms of the unfolded protein, and N represents the native molecule. b. Role of Proline Isomerization. The molecular basis for the difference between Us and UF, first suggested by Brandts et af. (393, appears to result from the cis-trans isomerization of at least two X-Propeptide bonds. RNase A has two cis isomers, at Pro-93 and Pro-114, and two trans isomers, at Pro-42 and Pro-117 (I, 292). Isomerization at the bond at Pro-42 has been considered to be one possibility [see Ref. (399)l. Isomerization at Pro-114 has been observed spectrophotometricaly in RNase nitrated at tyrosine residues 73, 76, and 115, as measured by the effect of isomerization on the ionization of the adjacent 3-nitro-Tyr-115(394,395). The conversion of UF to Us under unfolding conditions and cis-trans proline isomerization are both acid-catalyzed at similar rates and have similar activation enthalpies of the order of 20 kcaUmol(392, 393). The distribution of Us and UF at equilibrium is temperature-dependent (396), and corresponds to the enthalpic difference between cis and trans proline isomers (397) of - 1.4 kcaYmol (396). To account for the fast-folding species UF, Brandts et al. (393) suggested that the four proline residues in this species adopted their native peptide configurations. However, measured under refolding conditions, the activation enthalpy for the Us UF reaction is less than 5 kcdmol (398), as opposed to the 18 to 20 kcaUmol for cis-trans isomerization at

*

389. Garel, J.-R., and Baldwin, R. L. (1973). PNAS 70, 3347. 390. Garel, J.-R., and Baldwin, R. L. (1975). JMB 94, 611. 391. Garel, J.-R., Nall, B. T., and Baldwin, R. L. (1976). PNAS 73, 1853. 392. Hagerman, P. J., and Baldwin, R. L. (1976). Biochemistry 15, 1462. 393. Brandts, J. F., Halvorson, H. R., and Brennan, M. (1975). Biochemistry 14, 4953. 394. Garel, J.-R. (1980). PNAS 77, 795. 395. Garel, J.-R., and SiEert, 0. (1979). BBRC 89, 591. 396. Henkens, R. W., Gerber, A. D., Cooper,M. R . , and Herzog, W. R., Jr. (1980). JBC 255, 7075. 397. Cheng, H. N., and Bovey, F. A. (1977). Biopolymers 16, 1465. 398. Nall, B. T., Garel, J.-R., and Baldwin, R. L. (1978). J M B 118, 317.

12. PANCREATIC RIBONUCLEASES

391

proline (394, 399). Moreover, the rates of refolding do not correlate with the temperature-dependence of the distribution of Us and UF during prefolding conditions (3%); thus isomerization at proline in the unfolded molecule cannot be rate limiting in the folding of Us to N. The rate of isomerization at proline is significantly accelerated during folding, 30-fold at 100 (400).During the refolding of Us to N, the existence of intermediates in the folding pathway prevent the otherwise rapid exchange of amide protons, probably as a result of their engagement in hydrogen-bonded structures (244). Guanidinium chloride (2 M) destabilizes these intermediates and allows the amide protons to exchange. Guanidinium chloride does not affect the rate of conversion of UF to Us (394, 401) and it has no effect on cis-trans isomerization in Lalanyl-L-proline (398);however, it does decrease the rate of isomerization at proline during refolding. Apparently isomerization at proline is accelerated by formation of an intermediate during refolding (400). c . Role of S-Peptide During Refolding. Measurements of protection of amide protons against exchange have indicated that the S-peptide is necessary for the stabilization of early folding intermediates (244). RNase S-protein with its native disulfides intact exists in two unfolded forms, as does RNase A, with a similar distribution of the species, Us (80%) U, (20%) (245). At low temperatures, dissociation of RNase S by pH <2 (226) yields S-protein that remains partially folded (70% at 100) and rapidly recombines with S-peptide (246). RNase A refolds 60 times more rapidly than S-protein; in the presence of S-peptide, RNase S refolds much more rapidly than S-protein. The refolding kinetics of RNase S are concentration-dependent , which suggests that the S-peptide combines with and stabilizes a folding intermediate of the S-protein (245). The refolding kinetics of thermally unfolded RNase A, studied by 360 MH, proton NMR resonances of the C-2 protons of histidine residues, suggest that formation of the a-helical structure in residues 3 to 12 constitutes an early event during refolding of the molecule (243). d . Refolding of RNase with Reduced Disuljides. The refolding of reduced RNase is a complex and slow process. Recovery of the native enzyme by air oxidation requires hours (402, 403). In the presence of a redox mixture of reduced and oxidized glutathione (404&06), reoxidation 399. 400. 401. 402. 403.

Schmid, F. X . , and Baldwin, R. L. (1978). PNAS 75, 4764. Cook, K. H., Schmid, F. X., and Baldwin, R. L. (1979). PNAS 76, 6157. Schmid, F. X., and Baldwin, R. L. (1979). JMB 133, 285. Anfinsen, C. B . , Haber, E., Sela, M., and White, F. H . , Jr. (1961). PNAS 47, 1309. Epstein, C. J . , Goldberger, R . F., Young, D. M., and Anfinsen, C. B. (1962). ABB

Suppl. 1, 223.

404. Saxena, V. P., and Wetlaufer, D. B. (1970). Biochemistry 9, 5015. 405. Ahmed, A . K., Schaffer, S. W., and Wetlaufer, D. B. (1975). JBC 250, 8477.

392

PETER BLACKBURN AND STANFORD MOORE

of protein sulfhydryl groups to intramolecular disulfides occurs within minutes near pH 8.0, but with initially incorrect pairing of the protein disulfides (407,408). In a much slower reaction (tllz 90 min by recovery of activity), these bonds then rearrange to the native pairings via glutathione-catalyzed mixed-disulfide exchange reactions (407, 409). In 6 M guanidinium chloride, although the rate of thiol-disulfide exchange is unaltered (410), species with only one or two randomly paired disulfides accumulate along with a significant amount of insoluble material, presumably interchain disulfide cross-linked products (129). e. Trapping of SulJhydryl Intermediates. After trapping sulfhydryl groups by S-alkylation with iodoacetate, during glutathione-facilitated refolding of reduced RNase A, and separation of the S-carboxymethylated products by ion exchange chromatography, Creighton (129,411) identified intermediates with 1,2, 3, and 4 non-native disulfides as well as the native molecule, As the number of disulfide bonds formed increased, the number of non-native disulfide pairings decreased from those statistically possible. Creighton (129) found that cysteine residues 65 and 72 were uniformly more engaged in disulfides throughout refolding than were the other six cysteine residues. Similarly, during Cun-catalyzed air oxidation of reduced RNase, Takahashi and Ooi (412) have found that the native disulfide between cysteine residues 65 and 72 was generated most rapidly. Creighton (413) has also isolated and identified a refolding intermediate (IIIn) with three correctly paired disulfides; the last disulfide between residues Cys-40 and Cys-95 had not formed. This last disulfide is near two proline residues at positions 42 (trans) and 93 (cis); the slower reformation of this disulfide may be related to the isomerization of one or both of these proline residues. The conformation of intermediate IIIn is substantially similar to that of native RNase A; IIIn demonstrates enzymatic activity, and spectral, hydrodynamic, and immunochemical properties similar to those of the native molecule, but its conformation is less stable to denaturation by urea. With the exception of IIIn (asjudged by tyrosine absorption and fluorescence emission, electrophoretic analysis of unfolding in urea gradients (414), and immunochemical cross-reactivity), other intermediates resem-

-

406. 407. 408. 409. 410. 411. 412. 413. 414.

SchaBer, S. W., Ahmed, A. K., and Wetlaufer, D. B. (1975). JBC 250, 8483. Hantgan, R. R., Hammes, G. G . , and Scheraga, H. A. (1974).Biochemistry 13,3421. Creighton, T. E. (1977). J M B 113, 329. Schaffer. S. (1975). Int. J . Peptide Protein Res. 7, 179. Creighton, T. E. (1977). JMB 113, 313. Creighton, T. E. (1978). Progr. Biophys. Mol. B i d . 33, 231. Takahashi, S., and Ooi, T. (1976). Bull. Inst. Chem. Res. Kyoto Univ. 54, 141. Creighton, T. E. (1980). FEES Lett. 118, 283. Creighton, T. E. (1979). J M B 129, 235.

12. PANCREATIC RIBONUCLEASES

393

bled the unfolded species. However, conformational analyses of intermediates trapped by S-carboxymethylation, or with other charged or bulky reagents, may be misleading as a result of charge repulsion and steric effects on the conformations of the intermediates (388,425). Intermediates with two or three correctly paired disulfides, which might be expected to possess native-like structure, would be affected most because the charged groups would be located at precisely the regions predicted to interact with one another. The reshuffling of disulfide bonds in partially regenerated RNase can be arrested without disturbing the disulfide bonds or the conformation of the protein by lowering the pH to 4.0. By this approach, Konishi and Scheraga (416) have studied the temperature-dependence of the initial velocity of cyclic 2',3'-CMP hydrolysis as a probe of the thermodynamic properties of the active site during refolding. Temperature-dependence of tyrosine absorption at 287 nm and changes in optical rotation at 436 nm were also studied as indicators of the polypeptide chain conformation during glutathione-facilitated refolding of reduced RNase A. The thermodynamic parameters AHo (T,) and ASo (T,) and the T, (melting transition temperature) of folding intermediates with enzymatic activity over the range from 0.6 to 100% were identical to those of native RNase. The thermodynamic parameters determined by activity measurements and spectral measurements were essentially identical. Thus, the intermediates formed during refolding of reduced RNase are inactive, and by these techniques their conformations are apparently disordered (416). Proton NMR spectra indicate that the reappearance of the C-2 proton resonances of histidine residues 48, 105, and 119, and the corresponding decrease of the C-2 proton resonance of histidine residues in disordered conformations also parallels the recovery of enzymatic activity during refolding of RNase A (417), again suggesting that the dominant conformations of the intermediates formed during refolding are disordered. f. Immunochemical Approach to RNase Refolding. An alternative approach to seek out intermediates in the folding of reduced RNase has been to use immunochemical methods, based upon the ability of antibodies (toward the native protein) to recognize specifically only native conformation in their antigenic determinants (418). A significant advantage of this approach is its greater sensitivity over physicochemical techniques, permitting low concentrations of folding intermediates to be detected. A 415. Goto, Y.,and Hamaguchi, K. (1979). J . Biochem. (Tokyo) 86, 1433. 416. Konishi, Y., and Scheraga, H. A. (1980). Biochemistry 19, 1308. 417. Konishi, Y., and Scheraga, H. A. (1980).Biochemistry 19, 1316. 418. Sachs, D. H . , Schechter, A. N., Eastlake, A., and Anfinsen, C. B. (1972).P N A S 69, 3790.

394

PETER BLACKBURN AND STANFORD MOORE

disadvantage is that only events on a relatively long time scale can be observed. RNase A has at least four antigenic sites, possibly more, only three of which may be occupied at any one time as a result of competition between two of the sites (279). Subfractions of anti-native RNase A were purified b> affinity chromatography on RNase A-CNBr fragments comprised of residues 1-13, 31-79, and 80-124 that had been coupled to Sepharose (419). The ability of refolded RNase A to compete 100% with lZ5I-labeled native RNase A for binding to anti-native RNase A antibodies and their specific subfractions, and the absence of appreciable crossreactivity with reduced S-carboxymethylated RNase A, made it possible to follow the kinetics of refolding of reduced RNase A (419). Antigenic activity begins to return immediately during air oxidation of reduced RNase A, and returns to 100% after 240 min. Antigenic activity appears first in residues 80-124, followed by residues 1-13, and, finally, residues 31-79. Enzymatic activity reappears after 30 min and correlates with the folding of residues 3 1-79, which therefore probably represents a late event during folding. The thermal stability of these antigenic regions follows the order 1-13 < 31-79 < 80-124 (419). Matheson and Scheraga (367) have proposed that a segment of a polypeptide chain that can fold into a pocket that maximizes hydrophobic interactions forms the primary nucleation site for folding of a protein. They suggest that residues 106 to 118 form the primary nucleation site for RNase A(367). Upon air oxidation of reduced des-( 121-124)-RNase A, randomly paired disulfides are formed (142). RNase S-protein, on the other hand, possesses a limited capacity (20 to 30%) to fold to the native conformation by air oxidation of the reduced S-protein (420,421).Chavez and Scheraga (3881, by the irnmunochemical approach, followed the kinetics of refolding of reduced RNase A, S-protein, and des-( 121- 124)-RNase A, during glutathione-catalyzed thiol-disulfide exchange. It was found that the antigenic site in residues 87- 104 of all three proteins folded before the others. They suggest that nucleation at residues 106- 118 induces early folding of residues 87-104. The fact that S-protein readily folds eliminates residues 1-20 as the primary nucleation site. The partial folding of des(121-124)-RNase A indicates that residues 121-124 are not in the nucleation site; otherwise segment 87-104 would not have folded. Residues 121-124 are required to stabilize later stages of the folding process (388). 419. Chavez, L. G., Jr., and Scheraga, H. A. (1977). Biochemistry 16, 1849. 420. Haber, E., and Anfinsen, C. B. (l%l). JBC 236, 422. 421. Kato, I., and Anfinsen, C. B. (1969). JBC 244, 1004.

12. PANCREATIC RIBONUCLEASES

395

Burgess and Scheraga (422) suggested that segments of RNase that unfold only at high temperatures should provide nucleation sites for the folding of the molecule. Six overlapping stages in the thermal unfolding/ folding process were identified (422), modified by Matheson and Scheraga (186), and later modified by Chavez and Scheraga (388). Some of the observations that have helped to elucidate the thermal unfolding of RNase A, are summarized in Table IX (including Refs. 423-426). The process is first localized to the outermost residues and continues in a series of overlapping stages (186,388,422). In stage I (15-3Y), the side chain of Tyr-92 unfolds, and other localized changes involving Met-13 and Ala-19 and/or Ala-20 occur; in stage I1 (30-4Y), residues 13-25 unfold, and changes occur in the exposure of residues 1-12 to solvent; in stage 111 (40-50"), residues 27-34 and 75-80 unfold; in stage IV (50-60°), residues 51-60 and 121-124 unfold; in stage V (55-65"), residues 1-12, 35-50, and 62-74 unfold; and in stage VI (60-70"), residues 81-102 and 104-120 unfold (186, 388). The outer shell of amino acids, mostly polar residues, unfolds in a broad transition starting at 35-40" and is complete by 70"; the hydrophobic core unfolds in a sharper transition between 60 and 70". The overall transition temperature at pH 5.0 is about 56 to 60". The unfolded form of RNase at 78" still has some residual structure (80, 186, 312, 386, 427).

The folding pathway of reduced RNase is determined primarily by short-range interactions in the protein chain. The formation of disulfide bonds is not considered by Scheraga (123) to influence the folding pathway, but is thought to stabilize the final folded form (361) by reducing the entropy of the unfolded molecule [cf. Refs. 428, 4291. The folding pathway, as proposed by Chavez and Scheraga (388), is as follows: Nucleation by the hydrophobic core of residues 106-1 18, which may include cis-trans proline isomerization, and includes folding of the antigenic determinant in residues 87- 104. Folding continues as residues 63-75 reduce the exposure of the primary nucleation site to solvent. The ,&bend at residues 66-68 brings the chain around as residues 40-48 fold into place, along with residues 63-75 and probably residues 1-12, possibly with stabilizing interactions with the COOH-terminal residues 121422. 423. 424. 425. 426. 427. 428. 429.

Burgess, A. W., and Scheraga, H. A. (1975). J . Theoret. Biol. 53, 403. Li, L.-K., Riehm, J. P., and Scheraga, H. A. (1966). Biochemistry 5, 2043. Bigelow, C. C . , and Sonenberg, M. (1962). Biochemistry 1, 197. Bigelow, C. C. (l%l). JBC 236, 1706. Zaborsky, 0. R., and Mfiman, G. E. (1973). BBA 271, 274. Roberts, G . C. K., and Benz, E W. (1973). Ann. N . Y. Acad. Sci. 221, 130. Lapanje, S ., and Rupley, J. A. (1973). Biochemistry 12, 2370. Johnson, R. E., Adams, P., and Rupley, J. A. (1978). Biochemistry 17, 1479.

396

PETER BLACKBURN AND STANFORD MOORE TABLE IX EVENTS IN

THE

THERMAL UNFOLDING OF RIBONUCLEASE A

Events Stage I 15-35" Tyr-92 side chain unfolds and normalizes Localized changes involve Met-13, Ala-19/20 Stage I1 30-45" Tyr-25-Cys-26 bond becomes accessible Normalization of Tyr-25 Conformational change involves His-48 and Tyr-25 with Asp-14 Disruption of Met-13, Val-47 hydrophobic contact His-12 environment begins to change Stage 111 40-50' Lys-31-Ser-32, Arg-33-Asn-34bonds become accessible Met-79-Ser-80, Tyr-76-Ser-77 bonds become accessible NHpterminal and COOH-terminal sections remain intact Stage IV 50-60" Phe-120-Asp-121accessible at low pH as Val-124 is protonated Ser-123-Val-124 bond becomes accessible a-Helix of residues 51-60 unwinds Stage V 55-65' a-Helix of residues 3-12 unwinds Lys-1 exposed Lys-37-ASP-38, Arg-39-CYS-40, Lys-66-Asn-67 Tyr-73-Gln-74 Stage VI 60-70" Tyr-97 still not normalized L ys-9 1-Or-92

Ile-106-Val-118 nucleation site retained

Evidence

Ref.

Spectrophotornetric titration Photochemical labeling

(370. 423, 425)

a-Chymotrypsin Spectrophotometric titration NMR

(377, 379, 380) (370,423, 425) (312, 426, 427)

Photochemical labeling

(186)

NMR

(312)

Trypsin

(378,.380)

a-Chymotrypsin

(377)

Resistance to exopeptidases

(380,381)

Pepsin

(251

Immobilized carboxypeptidase A Optical rotation

(381 )

NMR Spin labeling; aminopeptidase Stable to trypsin

(375) (372, 380)

(186)

(422

(378)

Stable to a-chymotrypsin

(378)

Spectrophotometric titration Stable to trypsin Theory; antibodies

(425 1

(378) (367, 388)

12. PANCREATIC RIBONUCLEASES

397

124. These events bury the hydrophobic core of the molecule; the remaining residues then pack together in order to decrease the overall free energy of the protein molecule. 3 . Folding in Vivo

Bovine pancreatic RNase is synthesized in vivo with a 25-amino acid extension of its amino terminus (275), the signal peptide (273). This presecretory form of the protein is reported to fold via thiol-disulfide exchange to an enzymically active species. However,this is not always the case; folding of the presecretory proinsulin molecule is inhibited by its signal peptide (430). The folding of presecretory RNase A to a native-like structure is in keeping with observations that the S-peptide segment of the protein is less important for stability and folding in v i m . The observation that the last four residues of the molecule, residues 121-124, are important for maintaining the overall conformational stability of the protein and the inability of reduced des-(121-124)-RNase A to refold, indicates that, at least for RNase A, folding of the nascent protein chain during biosynthesis (431 433) cannot direct the final native conformation. A microsomal enzyme has been described (434, 435) that will catalyze the rearrangement of the disulfide bonds of a protein and facilitate refolding in vitro (436-438). The role of this enzyme in vivo is not clear. The kinetics of refolding for a small protein like RNase A via glutathionecatalyzed mixed-disulfide exchange may be adequate to account for its folding in vivo. Some of the aspects of protein folding in vivo that are pertinent to RNase A have been discussed by Baldwin and Creighton (352). V.

Species Variations

In the course of studies on pancreatic RNases from a wide variety of mammalian species, Beintema and his colleagues measured the concen430. 797 1. 431. 432. 433. 434. 435. 436. 437. 438.

Lomedico, P. T., Chan, J. S., Steiner, D. F., and Saunders, G . F. (1977). JBC 252, Chantrenne, H.(l%l). “Biosynthesis of Proteins,” p. 122. Pergamon, Oxford. Phillips, D. C. (1967). P N A S 57, 484. DeCoen, J. L. (1970). J M B 49, 405. Goldberger, R. F., Epstein, C. J., and Anfinsen, C. B. (1963). JEC 238, 628. Venetianer, P., and Straub, F. B. (1963). EEA 67, 166. Givol, D., Goldberger, R. F., and Anfinsen, C. B . (1%4). JBC 239, PC3114. Venetianer, P., and Straub, F. B. (1964). BBA 89, 189. Fuchs, S., De Lorenzo, F., and Anfinsen, C. B. (1967). JEC 242, 398.

398

PETER BLACKBURN AND STANFORD MOORE

tration of the enzyme in about one hundred species of mammals (4394 4 0 ~ )The . data extend those of Barnard (441, 442), and show that the concentration in pg/g of tissue ranges from about 0.5 in dog and 5 in man to 1000 in cow and 2000 in eland. The generally high concentrations found in ruminants led Barnard to conclude that the enzyme’s special role in that instance is in the digestion of bacterial RNA. A. VARIATIONS I N AMINO ACIDSEQUENCE Beintema and associates have conducted sequence determinations on the purified RNases from a wide variety of species. Taken together with the earlier sequence data, the results summarized in Table X (see Refs. 4 4 3 4 6 5 ) demonstrate the variable and conserved regions of the sequences for thirty-eight homologous RNases. The bovine seminal enzyme (see 439. Zendzian, E. N., and Bamard, E. A. (1967). ABB 122, 699. 440. Gaastra, W. (1975). Ph.D. Thesis, Univ. of Groningen, The Netherlands. 440a. Beintema, J. J., Scheffer, A. J., van Dijk, H., Welling, G. W., and Zwiers, H. (1973). Nature New Biol. 241, 76. 441. Barnard, E. A. (1969). Nature (London) 221, 340. 442. Bamard, E. A. (1969). Annu. Rev. Eiochem. 38,677. 443. Beintema, J. J., and Gruber, M. (1973). BBA 310, 161. 444. Beintema, J. J., and Gruber, M. (1967). BEA 147, 612. 445. Lenstra, J. A., and Beintema, J. J. (1979). EJB 98, 399. 446. Jekel, P. A , , Sips, H. J., Lenstra, J. A., and Beintema, J. J. (1979). Biochirnie 61, 827. 447. van Dijk, H., Sloots, B., van den Berg, A., Gaastra, W., and Beintema, J. J. (1976). t i i t . J . Peptide Protein Res. 8, 305. 448. van den Berg, A., and Beintema, J. J. (1975). Nature (London) 253, 207. 449. van den Berg, A., van den Hende-Timmer, L., Hofsteenge, J., Gaastra, W., and Beintema, J. J. (1977). EJB 75, 91. 450. van den Berg, A., van den Hende-Timmer, L., and Beintema, J. J. (1976).BBA 453, 400. 451. Emmens, M., Welling, G. W., and Beintema, J. J. (1976). EJ 157, 317. 452. Jackson, R. L., and Hirs, C. H. W. (1970). JBC 245, 637. 453. Phelan, J. J., and Hirs, C. H. W. (1970). JEC 245, 654. 454. Welling, G. W., Groen, G., and Beintema, J. J. (1975). BJ 147, 505. 455. Welling, G. W., Mulder, H., and Beintema, J. J. (1976). Biochem. Gener. 14, 309. 456. Beintema, J. J. (1980). BEA 621, 89. 457. Suzuki, H., Greco, L., Parente, A., Farina, B., La Montagna, R., and Leone, E. (1976).In “Atlas of Protein Sequence and Structure” (M. 0. Dayhoff, ed.), Vol. 5, Suppl. 2, p. 93. National Biomedical Research Foundation, Washington, D. C. 458. Russchen, F., De Vrieze, G., Gaastra, W., and Beintema, J. J. (1976).BBA 427,719. 459. Groen, G., Welling, G. W., and Beintema, J. J. (1975). FEBS Lett. 60, 300. 459a. Kuper, H., and Beintema, J. J. (1976). BEA 446, 337. 460. Welling, G. W., Scheffer, A. J., and Beintema, J. J. (1974). FEBS Lett. 41, 58. 461. Kobayashi, R., and Hirs, C. H. W. (1973). JBC 248, 7833. 461a. Beintema, J. J., Gaastra, W., and Munniksma, J. (1979). J. Mol. Evol. 13, 305.

12. PANCREATIC RIBONUCLEASES

399

Section VI) is included in the comparison. Residues essential for maintaining the correct secondary and tertiary structure of the protein (including bridges) and residues with essential roles for binding of the four -S-Ssubstrate and for catalytic function should be expected to remain constant throughout evolution; only nonessential residues should occupy variable domains. Among the pancreatic RNases studied, a number have been found to exhibit sequence heterogeneity (466). A gene duplication has occurred in the guinea pig where two forms of the enzyme RNase A and B are found. They differ from each other at 3 1 positions in the sequence, and one of the forms, RNase B, exhibits heterogeneity at position 64 where the major species has Leu and a minor species has Pro (448, 449). The enzyme in bovine seminal plasma is considered to arise through gene duplication (466); the seminal enzyme is a dimer of identical subunits cross-linked by two adjacent disulfide bonds at positions 31 and 32 (Section VI). The sequence of the subunit is homologous with that of pancreatic RNase A (457).

Allelic polymorphism has been demonstrated in the dromedary pancreatic RNase, where Gln or Lys is found in the ratio 3 : 1 at position 103; the pancreatic RNase of the bactrian species has the same sequence with Gln at this position (454, 455). In roe deer, either Ile or Ala is found at position 64 in the ratio 7 : 3 (464,465). Sequence heterogeneities have also been found in the pancreatic RNase from chinchilla, where either Gly or Asp is found at position 32 (448), in porcupine RNase, where either Gly or Arg is found at position 98 (466), and in hippopotamus RNase where either Gln or Lys is found at position 37 (10). Mutations that have affected the amino terminus are seen in the pancreatic RNase from rat (443, 444), which has three extra residues, and those from kangaroo (9) and wallaby, both of which have a single amino acid deletion at the amino terminus. These events probably represent mutations that have affected cleavage of :he signal peptide from the polypeptide chain during biosynthesis (see Section 111,D). Mutations of the termination codon that have led to chain extensions of the carboxyl-terminus have been found in guinea pig RNase B (448, 449) and in the pancreatic RNase from horse (139), coypu (448), casiragua, 462. Gaastra, W., Groen, G ., Welling, G. W., and Beintema, J. J. (1974). FEBS L e f t . 41, 227. 463. Leijenaar-van den Berg, G., and Beintema, J. J. (1975). FEBS Lett. 56, 101. 464. Zwiers, H., Scheffer, A. J., and Beintema, J. J. (1973). EJB 36, 569. 465. Oosterhuis, S., Welling, G. W., Gaastra, W., and Beintema, J. J. (1977). BBA 490, 523. 466. Beintema, J. J . (1980). Proc. 28th Colloq. Profides B i d . Fluids, p. 139. Pergamon, Oxford.

TABLE X AMINOACIDSEQUENCES Order

Infraorder Superfamily or suborder or family

OF MAMMALIAN PANCREATIC RIBONUCLEASES'

Species (Ref.) 1

10

20

1. ox fBoo DIu11u61. bison IBiJon bi6onl ( 1 5 1 K E T A A A K F E R Q H M D S S T S A A 2. water buffalo swamp type IBubaeu bubaEbI (4561 K E T A A A K F Q R Q H U D S S T S S A 3. river type K E T A A A K F Q R Q H M D S S T S S A 4. eland ITauno-tnagub omizl 14561 K E T A A A K F E R Q H U D S S T S S A 5. nilgai lBo4ePaphud t h a g o c a n ~ u 6 1 (4561 K E T A A A K F E R Q H M D S S T S S A Hippo6. gnu ICclnnwltaetcla tnuninubl (4591 K E S A A A K F E R Q H M D S S T S S A thaqinae 7. topi [ W b C t M kotnigiunl I 4 m l K E S A A A K F E R Q H U D S S T S S A AMbpinae 8. impala [Aepycmob mr'an)wdI K E S A A A K F E R Q H M D S S T S S A 9. Thomson's gazelle (CflzclCn thoMorul (4561 K E S A A A K F Z R Z H M B S S T S S A K E S A A A K F E R Q H U D S S T S S A 10. goat [ C a w hihCUA1, sheep [OUid a h i e ~ l T 6 0 . 4 6 1 1 Caphinae AnZiPocap%idae 11. pronghorn IAtdLPocapta anphicanal (461aI - K E T A A A K F E R Q H I D S N P S S V Gihad6idae 12. giraffe IGina66a c a n c ( q n t d a l i s 1 i 4 T K E S A A A K F E R Q H I D S S T S S V CPnuidae Odoicoiee- 13. reindeer [Rangi@t &zhnnrlw) 14631K E S A A A K F E R Q H U D P S P S S A inae 14. roe deer [ C a p e o h c a ~ . v 1 e o t u l ~ 6 4 , 4 6 5 I K E S A A A K F E R Q H U D P S P S S A 15. moose IACceS dceal 14631 K E S A A A K F E R Q H U D P S A S S I Cehvinae 16. red deer ICehvw &p7LTl 1464,4651 K E S A A A K F E R Q H U D P S T S S A K E S A A A K F E R Q H U D P S M S S A 17. fallow deer (Oana d m l 1 4 6 ' j r Tylopoda Camdidae 18. dromedary [Caneew d f i o ~ d Z & 1 4 ] 1454) S E T A A E K F E R Q H M D S Y S S S S 19. bactrian camel ( C a m d m Iiacthinnud-[4551 S E T A A E K F E R Q H M D S Y S S S S Ancodonta 20. hippopotamus [H4ppOtJOh-u3 mphibiub P I K E T A A E K F Q R Q H M D T S S S L S SUitlL7 21. pig (Sub 4 c h u 6 a l 17,452,5531 K E S P A K K F Q R Q H M D P D S S S S R E s P A u K F Q R Q H M D s G N s P G CetaCU 22. lesser rorqual I&a?aenop& acdoaostnatn) P t A i hdo d a c t y t a 23. horse I E q u cahLepu6I 11391 K E S P A U K F E R Q H M D S C S T S S Rod& Myomqha Mcuidae 24. rat (Rattub n o 4 u e g i c u I m 3 . 4 4 4 1 G E S R E S S A D K F K R Q H M D T E G P S K R E S A A Q K F Q R Q H U D P D G S S I 25. m u s e (ku m u n c u p ~ ~(~4] 4 5 r Chicetidae 26. hamster I#eooc%icetlo nrulafub) (4461 K E S A A H K F E R Q H U D S T V A T S K E T S A Q K F E R Q H U D S T C S S S 27. muskrat IUndatna z i h u t l u c a ) HybthiHybtnicoidea 28. porcupine I H I ( 4 f h i X c h i 4 t a & z l K E S S k M K F E R Q H M D S S C S P S 29. guinea pig 1Cavi.a p o t c e U u 4 I (446,449) A A E S S A M K F E R Q H V D S G C S S S cononpha Cauoidea 30. guinea pig ( C a v i a pozcePPuaI ( I J Z , ~ B] A E S S A M K F Q R Q H M D P E G S P S Ckinckieeoidea 31. chinchilla (CkinckiPPa baovicalZ€iibT(446,450) K E S S A M K F Q R Q H M O S S C S P S K E S S A K K F Q R Q H I D S S C S P S Dotodontoidea 32. casiragua IP.toeckimip guaihnelb S E S S A K K F E R Q 11 M D S R C S P S 33. coypu IMyoroMtua co!it~u51 (446,4501 K E T A A M K F Q R Q H U D S C S S L S 34. two-toed 510th [ C ~ O ~ O P ~h I ~U m & Z l I I01 EdenK E S S A A K F Q R Q H M D S D S S(X X 35. three-toed 510th iB4adypu5 g'Li6eu1 - E T P A E K F Q R Q H M D T E H S T A bllmupiaLia 36. red kangaroo [ M a c t o t w i u d u l I 9 1 - E T A A E K F Q R Z(H U B T)E H S T(A 37. wallaby [lhcrlopw zul(oqGbeu6J - / 9 l b i R i b o n d u e om 4eminaP pPaMma:l 38. ox (804 tarnu61 (4571 K E S A A A K F E R Q H M O S G D S P S Wodac&~fa

Pecona

Bouidne

Bovinae

ml

-

[s]

cr

im

30 40 Y) 60 70 1 . S S S N Y C N Q M M K S R N L T K D R C K P V N T F V H E S L A D V Q A V C S Q K N V A C K N C Q T N C Y Q S Y

2 3 4 5

. . . .

S S S S

S S S S

S S S S

N N N N

Y Y Y Y

C C C C

N N N N

Q Q Q Q

M n M M

M M M M

K K K K

S S S S

R R R R

S N D S

M M M M

T T T T

S S K Q

D D D N

R R R R

C C C C

K K K K

P P P P

V V V V

N N N N

T T T T

F F F F

V V V V

H H H H

E E E E

S S S S

L L L L

A A A A

D D D D

V V V V

Q Q Q Q

A A A A

V V V V

C C C C

S S S S

q Q Q Q

K E K K

N N N N

V V V V

A A A A

C C C C

K K K K

N N N N

G G G G

Q Q Q Q

T T T T

N N N N

C C C C

Y Y Y Y

Q Q Q Q

S S S S

Y Y Y Y

6 . S S S N Y C N Q M M K S R N L T Q D R C K P V N T F V H E P L A D V Q A V C S Q K N V A C K N G Q T N C Y Q S Y 7 . S S S N Y C N Q M M K S R N L T Q D R C K P V N T F V H E S L A D V Q A V C S Q K N V A C K N G Q T N C Y ~ S Y

6 . S S S N Y C N Q n M K S R N L T Q S R C K P V N T F V H E S L A D V Q A V C S Q K N V A C K N G Q T N C Y Q S Y 9 . S S S N Y C N Q M M K S R N L T Q D R C K P V N T F V H E S L A D V Q A V C S Q K N V Q C K N G Q T N C Y Q S Y 1 O . S S S N Y C N Q M M K S R N L T Q D R C K P V N T F V H E S L A D V Q A V C S Q K N V A C K N G Q T N C Y Q S Y

1 1 1 2 1 3 l ~ 1 5

. . . . .

S S S S S

S S S S S

S S S S S

N N N N N

Y C N Q M M Y S B N L T Q C R C K P V N T F V H E S L A D V ~ A V C S Q K N V A C K N G Q T N C Y Q S Y Y C N Q M M T S R N L T Q D R C K P V N T F V H E S L A D V Q A V C S Q K N V A C K N C Q T N C Y Q S Y Y C N Q M M Q S R D L T Q D R C K P V N T F V H E S L A D V Q A V C F Q K N V A C K N G Q S N C Y Q S N Y C N Q M M Q S R N L T Q D R C K P V N T F V H E S L A D V Q A V C F Q K N V I C K N G Q S N C Y Q S N Y C N Q M M Q S R N L T Q D R C K P V N T F V H E S L A D V Q A V C F Q K N V A C K N G Q S N C Y Q S N

1 6 . S S S N Y C N Q M M Q S R K M T Q D R C K P V N T F V H E S L A D V Q A V C F Q K N V A G K N G Q S N C Y Q s N 1 7 . S S S N Y C N Q M M Q S R K M T Q D R C K P V N T F V H E S L A D V Q A V C F Q K N V A C K N G Q S N C Y O S N 1 6 . S N S N Y C N Q M M K R R E M T N G - C K P V N T F I H E S L E D V Q A V C S Q K S V T C K N G Q T N C H Q S S

1 9 . S N S N Y C N Q M M K R R E M T N G - C K P V N T F I H E S L E D V Q A V C S Q K S V T C K N G Q T N C H Q S S Z 2 2 2

O . N D S N Y C N Q M M V R R N M T Q D R C K P V N T F V H ~ S E A D V K A V C S Q K N V T C K N G Q T N C Y Q ( S N 1 . N S S N Y C N L M M S R R N M T Q G R C K P V N T F V H ~ S L A D V Q A V C S Q I N V N C K N G Q T N C Y Q S N 2 . N N P N Y C N Q M M M R R K M T Q G R C K P V N T F V H E S L E D V K A V C S Q K N V L C K N G R T N C Y E S N 3 . N P T N Y C N Q M M K R R N M T Q G - C K P V N T F V H E P L A D V Q A I C L Q K N I T C K N C Q S N C Y Q S S

2 2 2 2 2 3 3 3 3

5 C 7 8 9 0 1 2 3

2 4 . S S P T Y C N Q M M K R Q C M T K G S C K P V N T F V H E P L E D V Q A I C S Q G Q V T C K N G R D N C H K S S

. N S P T Y C N Q M M K R R D M T N C S C K P V N T F V H E P L A D V Q A V C S Q E N V T C K N R K S N C Y K S S S S P T Y C N Q M M K R R N M T Q G Y C K P V N T F V H E S L A D V H A V C S Q E N V A C K N G K S N C Y K S H . S S ( P T Y ) C N Q M M K R R E M T Q G Y C K P V N T F V H E P L A D V Q A V C S ~ E N V T C K N G N S N C Y K S R . S N S N Y C N E M M R R R N M T Q D R C K P V N T F V H E P L A D V R A V C F Q K N V A C K N G Q T N C Y Q S N , S N A N Y C N E M M K K R E M T K D R C K P V N T F V H E P L A E V Q A V C S Q R N V S C K N G Q T N C Y Q S Y . N S S N Y C N V M M I R R N M T Q G R C K P V N T F V H E S L A D V Q A V C F Q K N V L C K N G Q T N C Y Q S Y . T N A N Y C N E M M K G R N M T Q G Y C K P V N T F V H ~ P L A D V Q A V C F Q K N V P C K N G Q S N C Y Q S N . T N P N Y C N A M M K S R N M T Q ~ R C K P V N T F V H E P L A D V Q A V C F Q K N V P C K N G Q S N C Y ~ S T . T N P N Y C N E M M K S R N M T Q C R C K P V N T F V H E P L A D V Q A V C F Q K N V L C K N G Q T N C Y Q S N

3 4 , S S S D Y C N K H M K V R N M T Q E S C K P V N T F V H E S L Q D V Q A V C F Q E N V T C K N G Q Q N C H Q S R 3 S . X X X X X X X ) K M M K S R N M T Q E S C K A V N T F V H ~ P L T D V ~ A V C L Q E N V T C K B G Q Q B C H X X X

3 6 . S S S N Y C N L M M K A R D M T S C R C K P L N T F I H E P K S V V D A V C H Q E N V T C K N G R T N C Y K S N 37. S S S B Y C B L M M)K A R E M T S D R C K P V N T F I H E P K S V V B A V(C 2 Z ) E B ( V T C)K N G Q T N ( C Y)K S ( N

3 6 . S S S N Y C N L M M C C R K M T Q C K C K P V N T F V H E S L A D V K A V C S Q K K V T C K N C Q T N C Y Q S K

(continued)

TABLE X (Coritinuetl) 80 90 M 110 1a0 1 . S T M S I T D C R E T G S S K Y P N C A Y K T T Q A N K H l I V A C E G N P Y V P V H F D A S V

2 . S T M S I T D C R E T G S S K Y P N C A Y K T T Q A N K H l I V A C E G N P Y V P V H F D A S V 3 . S T U S I T D C R E T C S S K Y P N C A Y K T T Q A N K H l l V A C E C N P Y V P V ~ Y D A S V

~ 5 6 7

. S . S T . S T . S T

T M M M

~ S I S I T D S I T D S I T D

T D C C R E T C R E T C R E T

R E T G S S K G S S K G S S K

G Y Y Y

S S P N C P N C P N C

K A A A

Y P N ~ A Y K T T Q A ~ K H I I ~ A ~ E ~ N P Y ~ P ~ H F D A ~ ~ Y T T T Q A K K H I I V A C E G N P Y V P V H F D A S V Y K A T Q A K K H I I V A C E G N P Y V P V H F D A S V Y K T T Q A K K H I I V A C E G N P Y V P V H F D A S V

6 . S T M S I T D C R E T G S S K Y P N C A Y K T T Q A K K H I I V A C E G N P Y V P V H F D A S V

1 1 1 1 ~ 1 1 1 1 1 2 2 2

9 . S T M S I T D C R E T C S S K Y P N C A Y K T T Q A Q K H I I V A C E G N P Y V P V H F D A S V D 1 2 3

. S T M S I T D C R E T C S S K Y P N C A Y K T T Q A E K H I I V A C E G N P Y V P V H F D A S V . S T M S I T D C R E T C S S K Y P N C A Y K T T Q A K K H I I V A C E C N P Y V P V H Y D A S V . S A M S I T D C R E T C N S K Y P N C A Y Q T T Q A E K H I I V A C E G N P Y V P V H Y D A S V . S A M H I T D C R E T G S S K Y P N C V Y K T T Q A E K H I I V A C E G N P Y V P V H F D A S V . S A M H I T D C R E S G N S K Y P N C V Y K T T Q A E K H I I V A C E C N P Y V P V H F D A S V S . S A M H I T D C R E S C N S D Y P N C V Y K T T Q A E K H I I V A C E G N P Y V P V H F D A S V 6 . S A M H I T D C R E S G N S K Y P N C V Y K A T Q A E K H I I V A C E G N P Y V P V H F D A S V 7 . S A M H I T D C R E S C N S K Y P N C V Y K A T Q A E K H l I V A C E C N P Y V P V H F D A S V 6 . T S M H I T D C R E T G S S K Y P N C A Y K A S N L K K H I I I A C E C N e Y V P V H F D A S V 9 . T S M H I T D C R E T G S S K Y P N C A Y K A S N L Q K H I I I A C E C N P Y V P V H F D A S V D . S T ) M H I T D C R E T C S S K Y P N C A Y K T S Q L Q K H I I V A C E C D P Y V P V H Y D A S V l . S T M H I T D C R Q T C S S K Y P N C A Y K A S Q E Q K H I I V A C E G N P P V P V H F D A SV 2 . S T M H I T D C R Q T G S S K Y P N C A Y K T S Q K E K H I I V A C E G N P Y V P V H F D N S V

2 3 . S S M H I T D C R L T S G S K Y P N C A Y Q T S Q K E R H l I V A C E G N P Y V P V H F D A S V Q T

2 2 2 2

~ 5 . 6 . 7 .

. S S S

S T L R I T D C R L K G S S K Y P N C T Y N T T " P Y V P A L H I T D C H L K G N S K Y P N C D Y K T T Q Y Q K H I I V A C E C N P A L H I T D C R L K G N A K Y P N C D Y Q T S O L Q K Q I I V A C ~ G N A L H I T D C R L K G N S K Y P N C D Y Q ( T S Q L ) Q K Q V I V A C E G S P

V H F D A S V Y V P V H F D A T V P F V P V H F D A S V F V P V H F D A S V

2 8 . S L M H I T D C R V T G S S K Y P D C S Y R M S Q L E R S I V V A C E C S P Y V P V H F D A S V G P S T 2 9 . S S M H I T E C R L T S G S K F P N C S Y R T S Q A Q K S I l V A C E G K P Y V P V H F D N S V 3 D . S R M R I T D C R V T S S S K F P N C S Y R M S Q A Q K S I I V A C E G D P Y V P V H F D A S V E P S T 3 3 3 ~ 3 3

1 2 3 . 5 6

. S N M H I T D C R L T S N S K Y P N C S Y R T S R E N K G I I V A C ~ : G N P Y V P V H F D A S V . S N M H I T D C R L T S N S K F P D C L Y R T S Q E E K S I I V A C E G N P Y V P V H F D A S V A A S A . S N M H I T D C R V T S N S D Y P N C S Y R T S Q E E K S I V V A C E G N P Y V P V H F D A S V A A S A S N M H I T D C R Q T S C S K Y P N C L Y Q T S N M N R H I I I A C E G N P Y V P V H F D A S V E D S T . X X M H I T D C R Q ( T S G S T Y P N C L Y ) K T T N K X X X X X X X X X X X X X V P V H F D A T V . S R L S I T N C R Q T C A S K Y P N C Q Y E T S N L N K Q l l V A C E C Q - Y V P V H F D A Y V 37. S)R L(S 1 T N C)R Q(T G A S B Y P B C 7. Y 2 T S B ) L Q K Q ( I I V A C ) E G Q Y(V P V H F ) D A Y V 3 6 . S T M R I T D C R E T G S S K Y P N C A Y K T T Q V E K H I I V A C C G K P S V P V H F D A S V

-

Classified according to Beintema and Lenstra (468), with sequence data from the cited references. Many residues in peptides have been positioned for the RNases of bovidae and pronghorn by homology with the ox enzyme; a similar procedure has been used with the RNases of deer species (with reference to red deer) and bactrian camel (with reference to dromedary). J . J. Beintema et al., unpublished. Single letter code: A = Ala, B = Asx, C = Cys, D = Asp, E = Glu, F = Phe, G = Gly, H = His, I = Ile, K = Lys, L = Leu, M = Met, N = Asn, P = Pro, Q = Gln, R = A r g , S = Ser, T = Thr, V = Val, W = Trp. X = ?, Y = Tyr, Z = Glx, - = Gap, ( ) = By analogy, from composition, / = )(, back-to-back parentheses.

'

403

12. PANCREATIC RIBONUCLEASES True ruminants Came 1s

ALA (GCX) G U L‘

pig

LYS (AApu)

Horse

MET (AUG)

Lesser rorqual

MET (AUG)

Rat Muskrat

ASP (GAPY) GLU ‘ GLN (CApu)/

Guinea pig

MET (AUG)

Chinchilla

MET (AUG)-

COYPU

LYS (AApu)

Kangaroo

GLU (GApu)

(GAG) \ G L U ( a G )

(GAG)-LYS

(AAG)

GLU (GAG)

F I G .7. Evolutionary history of mutations at position 6 in RNase. Of the ancestral codons, only those that require minimal base changes are given. From Welling et cil. (472).

porcupine (Table X), and two-toed sloth (10). Examples of deletions have not been found at the carboxyl terminus. Havinga and Beintema (10) pointed out that carboxyl terminal additions do not occur frequently, and when they are present the sequences of the extensions are not random. The molecular evolution of mammalian pancreatic RNases has been discussed in detail by Beintema and his colleagues (467, 468). Pancreatic RNase is a rapidly evolving protein by comparison with cytochrome c and several other proteins (469471). The evolution rate of bovid RNases is 2-3 times slower than the average rate observed among the other mammalian RNases (456).On the other hand, rat RNase has evolved at an unusually high rate, 2.5 to 4.3 times as high as the enzyme in related rodent species, and 23 times as high as the average rate in the bovid RNases (445). An example of the evolutionary changes that affect a particular residue is shown in Fig. 7. The residue at position 6 is rather variable, although not all amino acids are permitted at this position. Probably one reason for this restricted variability is the importance of the a-helix in this region of the 467. Beintema, J . J . , Gaastra, W., Lenstra, J. A , , Welling, G . W., and Fitch, W. M. (1977). J . Mol. E t d . 10, 49. 468. Beintema, J. J., and Lenstra, J. A. (1982). / / I “Macromolecular Sequences in Systematics and Evolutionary Biology” (M. Goodman, e d . ) , in press. Plenum, New York. 469. Dickerson, R. E. (1971). J . Mol. Evol. 1, 26. 470. McLaughlin, P. J . , and Dayhoff, M. 0 . (1973). J . Mol. Evol. 2, 99. 471. Smith, E. L. (1970). “The Enzymes,” Vol. 1 , p. 267.

404

PETER BLACKBURN AND STANFORD MOORE

molecule (472). In general, the various amino acid substitutions can be accomodated into the three-dimensional model of RNase S (1, 16) without altering the folding of the backbone. However, mutations that result in the deletion of an amino acid have occurred at residue 39 in horse (472), and dromedary and bactrian camel RNases (454,455). In order to accommodate the deletion of residue 39, the loop that comprises residues 36-39 must be shortened in such a way that the hydrogen bond between the phenolic hydroxyl of Tyr-92 and the carbonyl oxygen of residue 37 is broken, causing Tyr-92 to turn away into the solvent (139). With the exception of Tyr-92, which is conservatively substituted by phenylalanine in guinea pig RNases A and B and in the RNase from casiragua, the surface tyrosine residues have been variously substituted nonconservatively; the two buried tyrosine residues at positions 25 and 97 are invariant. Proline substitution into a polypeptide chain restricts the number of possible chain conformations available at that site because the dihedral angle imposed by proline must be around -70". Among the various pancreatic RNases, proline substitutions occur at ten different sites in addition to the four positions found in bovine RNase A. Oosterhuis et al. (465) have listed the dihedral angle at these ten sites in the X-ray structure of RNase S (16) and find that proline can be accommodated at all of these sites without main chain distortion. Of the four proline residues in bovine RNase, the two at positions 93 and 114 are in the cis configuration; no changes at these positions are found, with the exception of the kangaroo RNase in which residue 114 is deleted (9). No substitutions have been found for the two proline residues at positions 42 and 117, with the exception of Ala at position 42 in the three-toed sloth RNase (10). Further evidence that amino acid substitutions, comprising the variable regions of the pancreatic RNases, have not appreciably affected the three-dimensional structure of the molecule comes from studies on S-peptide and S-protein interactions. Although, the first 24 residues of the sequence contain some of the most variable positions of the RNase molecule, Glu-2, Ala-5, Phe-8, Arg-10, Gln-11, His-12, and Asp-I4 are invariant. Hybrids composed of the S-peptides derived from cow, dromedary, and kangaroo RNases, and a synthetic 17-residue S-peptide corresponding to that sequence of rat RNase, with the S-proteins derived from cow and dromedary camel had very similar properties (190). The dissociation constants for the S-peptides, K , values for cyclic 2',3'-CMP, and maximal velocities showed no significant differences. Immunological 472. Welling, G. W., Leijenaar-vanden Berg, G., van Dijk, B., van den Berg, A., Groen, G., Gaastra, W., Emmens, M., and Beintema, J. J. (1975). Bio Systems 6, 239.

12. PANCREATIC RIBONUCLEASES

405

cross-reactivities with antibodies toward bovine RNase S failed to reveal any difference between the various hybrids. By NMR, the conformation of a hybrid RNase S' composed of bovine S-protein and the synthetic S-peptide corresponding to residues 3-20 of the rat enzyme, was essentially identical to that of bovine RNase S' (473). However, not all pancreatic RNases are cleaved by subtilisin (474,475). The region of the RNase molecule recognized by the protease, residues 16 to 22, is a highly variable part of the sequence; minor changes in the threeLdimensiona1structure of this region due to slightly different conformational preferences of the different amino acids might account for the resistance of some RNases to subtilisin cleavage (475). In particular, the presence of proline residues in this loop is associated with resistance to subtilisin (465). Despite the many amino acid substitutions in the S-peptide segments of the pancreatic RNases, residues that have been shown to be important from binding studies with synthetic S-peptides (see Section II1,B) are invariant, with the exception of the conservative substitutions of Met-13 by Ile in giraffe (462), pronghorn, and casiragua (Table X), and by Val in guinea pig RNase B (448). Most of the variable residues of the pancreatic RNases are hydrophilic and are located in less structured, looped regions on the surface of the molecule. Of the hydrophilic residues that remain constant, most can be ascribed roles. His-12, His-119, and Lys-41 are important for the catalytic process; Lys-7, Lys-66, and Arg-85 are close to the active site and their positively charged side chains may be important for enzymic function. However, Arg-85 has been substituted by histidine in mouse RNase (445). Substitutions of other residues thought to have important roles include Arg-39, which may be responsible for the lower pK, of the e N H 2group of Lys-41 (see Section III,A,3), and has been substituted by tyrosine in chinchilla (448), muskrat (447), and hamster RNase (446),by serine in rat (443,444)and mouse RNase (35),and by lysine in bovine seminal plasma RNase (457). Ser-123 has been substituted by threonine in mouse (445) and three-toed sloth RNase (lo), and by tyrosine in kangaroo (9) and wallaby RNase (Table X); Ser-123 is engaged in a hydrogen bond with uridine-containing substrates (260) (see Section 111,B,2). Other residues on the surface of the molecule that are invariant, or relatively so, include those in the adjacent sequences that involve residues 42-45 and 90-95, both of which contain nonpolar residues and appear to 473. Beintema, J. J . , and Lenstra, J . A . (1980). I n t . J . Pepfide Protein Res. 15, 455. 474. Klee, W. A , , and Streaty, R. A. (1970). JBC 245, 1227. 475. Welling, G . W., Groen, G . , Gabel, D., Gaastra, W., and Beintema, J. J. (1974). FEBS L e f t . 40, 134.

406

PETER BLACKBURN AND STANFORD MOORE

form part of the binding domain for the naturally occurring RNase inhibitor (45) (see Section VII). Residues 65-72 form part of the second basebinding site (21, 73). These relatively nonpolar domains are also among the more thermostable regions of the surface of bovine RNase A (see Table IX, Section IV,D). The eight half-cystine residues are conserved among all of the mammalian pancreatic species examined in detail to date. Weickman et al. ( / I ) have reported on the human pancreatic RNase; amino acid analyses for S- carboxymethylcysteine of the reduced, carboxymethylated protein have indicated 6 half-cystine residues per molecule. This result merits further study. Residues involved in hydrophobic contacts, and residues that shield such contacts, are in general invariant. They appear to be the most important features of the amino acid sequence for the formation of secondary and tertiary structure. Lenstra et a/. (476) applied the predictive methods of Burgess et (11. (477),Chou and Fasman (478,479),and Lim (480, 481 ) to the sequences of some 24 different mammalian pancreatic RNases. The predictions by the method of Lim, based on the relative positions of hydrophobic residues in a-helices and P-sheet structures, gave the best agreement with the X-ray structure of the bovine RNase. All of the residues that according to Lim’s theory are essential for the formation of secondary structure are invariant or conservatively substituted in the ribonucleases tested (476). Moreover, residues 106- 118, proposed by Matheson and Scheraga (367) (see Table IX, Section IV,D) as the nucleation site for the folding of bovine RNase A, form one of the more conserved regions of the sequence. Proton NMR studies have been performed on a number of the mammalian pancreatic RNases (42, 482, 4 8 4 , and pK, values have been determined for their histidine residues (Table XI). His-48 has a higher pK,, in rat (42) and pig (482) RNase than in bovine RNase A. It was suggested for the rat enzyme that this arises because of the influence of the negatively charged glutamate residue at position 16 of the rat enzyme (42); the pig RNase has aspartate at position 16 ( 7 , 452, 453). The high pK, values of 476. Lenstra, J. A . , Hofsteenge, J . , and Beintema, J. J. (1977). J M B 109, 185. 477. Burgess, A . W., Ponnuswamy, P. K . , and Scheraga, H. A. (1974). I s r . J . Biochem. 12, 239. 478. Chou, P. Y., and Fasman, G. D. (1974). Biochemistry 13, 211. 479. Chou, P. Y., and Fasman, G . D. (1974). Biochemistry 13, 222. 480. Lim, V. I. (1974). J M B 88, 857. 481. Lim, V. I. (1974). J M B 88, 873. 482. Wang, F.-F. C., and Hirs, C. H. W. (1979). JBC 254, 1090. 483. Leijenaar-van den Berg, G., Migchelsen, C., and Beintema, J . J. (1974). FEBS L e f t . 48, 218.

407

12. PANCREATlC RIBONUCLEASES TABLE XI pK, VALUES OF H I S T I D I NRESIDUES E IN

THE

PANCREATIC RNASES

p K , values Position

Reindeer"

Ratb

Chinchilla"

Coypu"

12 119 48 73 80 105

6.1 6.5 6.3 >7.0 6.5

6'2 6.6 7.6 6.1 6.3

6.0-6.1"

6.3d

4.9 7.2

5.8 8.0 -

-

Pig' 6.4 6.3 >7.5 >7.0 6.6

From van den Berg and Migchelsen (48.1). From Migchelsen and Beintema (42). ' From Wang and Hirs (482). Unassigned.

His-80 in coypu, chinchilla (42),reindeer (483),and pig RNase (482) arise most probably as the result of a salt bridge with Glu-49 (450). From NMR data, the active-site conformations of the bovine and rat RNAses resemble one another more closely in the presence of pyrimidine mononucleotides. This result suggests evolutionary constraints on the preservation of the active-site structure in the enzyme-substrate complex rather than in the substrate free state (42). Myer et a/. (484) arrived at a similar conclusion, based on the near-UV circular dichroism difference spectra of bovine RNase A and turtle RNase (442,485)in the presence of 2'- and 3'-CMP. B.

VARIATIONS IN

CARBOHYDRATE MOIETIES

Post-translational modifications that attach carbohydrate side chains to pancreatic RNases occur in many species. Beintema rt a/. (486) have tabulated (Table XII) (see refs. 487-490) the points of attachment and the approximate compositions of the carbohydrate side chains found with the enzymes from different species. Four carbohydrate attachment sites have 484. Myer, Y. P., Barnard, E. A., and Pal, P. K. (1979). JBC 254, 137. 485. Barnard, E. A . , Cohen, M. S., Gold, M. H., and Kim, J. (1972). Nrrrrrre (Londun) 246, 395. 486. Beintema, J. J., Gaastra, W., Scheffer, A. J., Welling, G. W. (1976). EJB 63, 441. 487. Plummer, T. H., Jr. (1968).JBC 243, 5961. 488. Becker, R. R., Halbrook, J. L., and Hirs, C. H. W. (1973). JBC 248, 7826. 489. Kabasawa, I . , and Hirs, C. H. W. (1972). JBC 247, 1610. 490. Tsuruo, T., Yamashita, S., Terao, T., and Ukita, T. (1970). BBA 200, 544.

TABLE XI1 APPROXIMATE COMPOSITION

OF T H E

CARBOHYDRATE MOIETIESOF GLYCOSIDATED RIBONUCLEASES" ~~

Ratio of monosaccharide residuesb Species Cow pancreas Sheep Pig

Giraffe Okapi

Component B C D B C

Attachment site 34 34 34 34 34 21 34 76 all 34a 34b 34c 34a 34b 34c

G~UCOSamine 2 4 4 2 5 7- 10 2 8-11

11

4 7 3 2 5

4

Mannose 6 4 3 6 6 3 6 3

7 5 6 5 4 4 19?

Galactose 2 2 2 2-4

Fucose 1 1 -

2 1

3-4 2 3 3

2 3

1

1

2 2 -

1 1 1

1 1

Sialic acid

21

4

Refs.

Moose Horse Lesser rorqual B COYPU Chinchilla Guinea pig Roe deer Hippopotamus Two-toed sloth Hamster River-type water buffalo Porcupine Casiragua

I I1

B

B

34 21 34 62 76 34 34 34 21a 21b 34 34 34 34 34 34 34 34

3

5 6 5 3 2 4 3 2 3 2 3

+

4

+

4 3

" Adapted from Beintema et a / . (486);galactosamine not present. not found; + , number of residues not determined.

* - Residue

Approximate values, J. J. Beintema et d..unpublished.

4 2 Not determined 5

1

1

(463 )

5

5

4-6 10- IS 6 7 3 9 4 9

+

5

(450)

(449)

2 -

4 6

1

1

I1

Trace

410

PETER BLACKBURN AND STANFORD MOORE

been found at positions 2 1,34,62, and 76. All occur at asparagine residues in an Asn-X-Thr/Ser sequence, coupled via N- acetylglucosamine in an N-glycosidic linkage (491, 492). The four carbohydrate attachment sites occur at exposed regions of the RNase molecule, removed from the active site, in variant parts of the sequence. The enzymes in even closely related species differ by the presence or absence of carbohydrate (486). Moreover, not all potential glycosylation sites are coupled to carbohydrate and there are notable differences in the complexity of the carbohydrate chains. The carbohydrate side chain of bovine RNase B (66, 493) has been sequenced by Liang ei al. (494) and is Mand (Mancvl

+ 2)0--3

Manal

I f

!Mancvl Mancvl

I f

!Man01

+

4GlcNAcpl .+ 4GlcNAc

The side chain is variously elaborated through the addition of one to three mannose residues joined by an a(1 + 2) linkage. Most of the pancreatic RNases studied have experienced acid conditions at some point during their isolation, with the exceptions of some of the preparations of the proteins from cow (66,493), pig (489,495,496), and sheep (488). Carbohydrate side chains, especially at sialic acid residues, undergo gradual degradation when exposed to acid (496). Sialic acid residues have been found in gl ycopeptides isolated from various pancreatic RNases that have been exposed to acidic conditions; thus their carbohydrate side chains cannot have undergone extensive acid degradation. The function of the carbohydrate chains attached to some of the pancreatic RNases is unknown. Beintema et nl. (486) have suggested that since species with cecal digestion (like pig, horse, guinea pig, chinchilla, and coypu) produce RNases with large carbohydrate chains attached to one or several sites on the surface of the molecule; the carbohydrate perhaps protects the RNase molecule from absorption by the gut. This 491. Eylar, E. H. (1965). J . Tlieorrt. 861. 10, 89. 492. Neuberger, A., and Marshall, R. D. (1%9). “Symposium on Foods-Carbohydrates and Their Roles” (H. W. Schultze, R. F. Chain, and R. W. Wrotstad, eds.), p. 115. Avi, Westport, Connecticut. 493. Plummer, T. H., Jr., and Hirs, C. H. W. (1963). JBC 238, 1396. 494. Liang, C.-J., Yamashita, K., Kobata, A. (1980). J. Biochem. (Tokyo) 88, 51. 495. Reinhold, V. N., Dunne, F. T., Wriston, J. C., Schwartz, M., Sarda, L., and Hirs, C. H. W. (1968). JEC 243,6482. 496. Jackson, R. L., and Hirs, C. H. W. (1970). JBC 245, 624.

12. PANCREATIC RIBONUCLEASES

41 1

woutd then facilitate its transport to the large intestine where it should hydrolyze the RNA from the cecal microflora, analogous to the postulated function of pancreatic RNases in ruminants (441). The effect of the carbohydrate side chains on the properties of the porcine pancreatic RNase molecule has been studied by Wang and Hirs (497). They compared a number of physicochemical properties of the native molecule and of molecules in which the carbohydrate was substantially reduced after digestion with a mixture of exoglycosidases. The size of the carbohydrate side chains had no influence on the rate at which the fully reduced, denatured protein reassumed the native folded structure on reoxidation, nor on the overall conformational stability of the molecule. The results of spectrophotometric titrations at high and low pH indicated that the carbohydrate chains increased the conformational stability of local surface regions associated with tyrosine residues. Circular dichroism measurements and UV absorption-difference spectra suggested that the carbohydrate chains affected the local tertiary structure around at least one tyrosine residue. Based upon the results of Puett (498), who found by circular dichroism measurements that the carbohydrate side chain at Asn-34 of bovine RNase B has no effect on the tertiary structure of the molecule and no influence on the nearby residue Tyr-92, Wang and Hirs (497) tentatively identified Tyr-25 of the porcine RNase as the residue most likely affected, possibly through interactions with the carbohydrate attached to nearby residue Asn-21. The conclusion from the results with pig pancreatic RNase was that no special mechanisms are necessary for correct polypeptide chain folding when glycoproteins are synthesized on membrane-associated ribosomes; this may not be true for all glycoproteins. VI.

Bovine Seminal Plasma RNase

The ribonuclease of bovine seminal plasma was first described by D’Alessio and Leone (499); the enzyme represents more than 2% of the total protein of the fluid (500).Two components were identified, BS-1 and BS-2, and the major component, BS-1, was purified and shown to have a MW of 29,000 and an isoionic point at pH 10.3 (500, 501). Similar proper497. 498. 499. 500. 153. 501.

Wang, F.-F. C., and Hirs, C. H. W. (1977). JBC 252, 8358. Puett, D. (1973). JBC 248, 3566. D’Alessio, G . , and Leone, E. (1963). BJ 89, 7P. D’Alessio. G . , Floridi, A , , De Prisco, R . , Pignero, A , , and Leone, E. (1972). EJB 26, Floridi, A., and De Prisco, R . (1973). I d . J . Biochern. 22, 1.

412

PETER BLACKBURN AND STANFORD MOORE

ties were reported by Hosakawa and Irie (502)for ribonucleases isolated from bovine seminal vesicles. The early studies demonstrated that while the seminal plasma enzyme had some properties in common with the bovine pancreatic enzyme, the differences in molecular size and charge were prominant (503-506). RNase BS- 1 is a pyrimidine-preferring endoribonuclease that yields 3’-phosphonucleotides via cyclic 2’ ,3‘intermediates (507), as is bovine RNase A (508), but differs from the pancreatic enzyme in its ability to hydrolyze double-stranded RNA (509). Seminal RNase BS-1 has been found to be comprised of two identical subunits (510) with an amino acid sequence (457, 511-513) that is homologous with that of the mammalian pancreatic RNases (see Table XI. There are 23 substitutions in nonessential positions of the chain; all key residues (such as His-12, His-119, and Lys-41) are present, and the four intrachain disulfides (514) are at positions identical with those found in the pancreatic RNases. The two subunits are covalently cross-linked by two readily reduced, adjacent disulfides at positions 31 and 32 of the polypeptide chain (515-517). Selective reduction of the interchain disulfides by dithiothreitol occurs rapidly with 30% dissociation of RNase BS-1 into monomers, and 70% into noncovalently associated dimers (517); complete dissociation to monomers requires mildly denaturing conditions. The refolding of fully reduced, denatured RNase BS-1 is facilitated optimally at pH 8.0 and 30” with a mixture of 3.0 mM reduced glutathione and 0.6 mM oxidized 502. Hosakawa, S. , and hie, M. (1971). J . Biochem. (Tokyo) 69, 683. 503. Forlani, L., Chiancone, E., Vecchini, P., Floridi, A., D’Alessio, G., and Leone, E. (1967). BBA 104, 170. 504. Floridi, A., and DAlessio, G. (1%7). Boll. Soc. Itul. Biol. Sper. 43, 32, 505. Floridi, A. (1968). BBRC 32, 179. 506. D’Alessio, G., Demma, G., Farina, B., Leone, E., and Parente, A. (1970).Boll. Soc. Ital. B i d . Sper. 46, 96. 507. Floridi, A., D’Alessio, G., and Leone, E. (1972). EJB 26, 162. 508. Volkin, E., and Cohn, W. E. (1953). JBC 205, 767. 509. Libonati, M., and Floridi, A. (1969). EJB 8, 81. 510. D’Alessio, G., Parente, A., Guida, C., and Leone, E. (1972). FEBS Lett. 27, 285. 511. D’Alessio, G., Parente, A., Farina, B., La Montagna, R., De Prisco, R., Demma, G. B., and Leone, E. (1972). BBRC 47, 293. 512. Leone, E., Suzuki, H., Greco, L., Parente, A , , Farina, B., and La Montagna, R. (1972). Proc. 8th FEBS. Meet. Amsterdrim, Abstr., p. 359. 513. Suzuki, H., and Greco, L . (1972). Boll. Soc. Itnl. B i d . Sper. 48, 1124. 514. Di Donato, A., and D’Alessio, G. (1979). BBA 579, 303. 515. Malorni, M. C., Di Donato, A., and D’Alessio, G. (1972). Boll. Soc. Ifnl. B i d . Sper. 48, 606. 516. Di Donato, A., and D’Alessio, G. (1973). BBRC 55, 919. 517. D’Alessio, G., Malorni, M. C., and Parente, A. (1975). Biochemistry 14, 1116.

12. PANCREATIC RIBONUCLEASES

413

glutathione (518), in much the same way as for bovine RNase A (405). The major product is the monomeric species with two moles of glutathione per mole of subunit as mixed disulfides at positions 31 and 32 (518).Selective reduction of the interchain disulfides with reduced glutathione yieids bis-S-glutathione-RNase-BS- 1 monomers (5f9 ) that are quite stable and show no tendency to re-form disulfide cross-linked dimers. Selective reduction of the interchain disulfides of RNase BS-1 followed by bis-Salkylation with iodoacetate (to give MCM-BS- 1) ( 5 / 7 ) , iodoacetamide (MCAM-BS-I), and ethylenimine (MAE-BS-l), yields catalytically active subunits (520). Smith and Schaffer (521) showed that these derivatives, after full reduction, will refold to the active species. Bovine RNase A (522-524) and seminal RNase BS-1 (525) aggregate when lyophilized from 30 to 50% acetic acid. The alkylated monomeric species, MCM-, MCAM-, and MAE-BS-1 showed less tendency to aggregate ( 5 , 8, and 11%, respectively) than did RNase A (24%) when similarly lyophilized (520).A structure has been proposed (522-524) for the aggregated dimer of RNase A in which the NH2-terminal segment (the S-peptide region) of one molecule adsorbs on the main chain (the S-protein region) of another, and vice versa. Low resolution X-ray diffraction studies on crystalline RNase BS-1 (526) provide data that indicate that a similar structure is possible for the seminal dimer, but the resolution is not sufficient to eliminate alternative orientations. The two monomeric subunits are disposed in the anti-parallel configuration, with half-Cys-3 I of one linked to residue 32 of the other, and vice versa. The dissociated monomeric subunits of RNase BS-1 adopt a different conformation from that which they have in disulfide cross-linked dimers. The circular dichroism spectra of bis-S-glutathione-BS-1 (518, 519) and MCM-BS- 1 monomers (348) closely resemble that of native bovine pancreatic RNase A, but the monomeric subunits appear to contain more a-helical content and less @-structure than the covalently cross-linked dimer (348). Near-UV circular dichroism measurements, fluorescence and fluorescence-quenching studies, and UV-absorption difference spectra all 518, Smith, G. K . , D'Alessio, G . , and Schaffer, S. W. (1978). Biochemistry 17, 2633. 519. Smith, G. K . , and Schaffer, S. W. (1979). ABB 196, 102. 520. Parente, A . , Albanesi, D . , Garzillo, A . M., and D'Alessio, G . (1977). I r a / . J . Biorhem. 26, 451. 521. Smith, G. K . , and Schaffer, S. W. (1980). ABB 203, 282. 522. Crestfield, A. M., Stein, W. H., and Moore, S . (1962). ABB Suppl. 1, 217. 523. Fruchter, R. G., and Crestfield, A. M. (1965). JBC 240, 3868. 524. Fruchter, R . G., and Crestfield, A . M . (1965). JBC 240, 3875. 525. Libonati, M. (1969). Itol. J . Eiochem. 18, 407. 526. Capasso, S . , Giordano, F., Mattia, C. A , , Mazzarella, L., and Zagari, A. (1979). Go::. Chim. Itci(. 109, 55.

414

PETER BLACKBURN AND STANFORD MOORE

suggest an increased exposure of tyrosine residues upon reduction of the interchain disulfides and dissociation to monomers (348). Of the four tyrosine residues, at positions 25, 73, 92, and 97, only one, presumably Tyr-73, titrates spectrophotometrically with a normal pK, (527). Other evidence of conformational differences between monomeric and dimeric forms of RNase BS-1 subunits is suggested from the differences in the extinction coefficients, E!Fm, of native RNase BS-1, 4.65 (500); MCMBS-1, 4.82 (517); MCAM-BS-1, 5.37; and MAE-BS-1, 5.49 (520). Also, the monomeric species MCM-BS- 1 exhibits susceptibility to digestion with trypsin, whereas RNase BS-1 does not (528), and MCM-BS-1 is less stable to heat, acid, and urea denaturation than pancreatic RNase A and dimeric RNase BS-1 (348). Immunologically, RNase BS-1 dimer demonstrates less reactivity with anti-RNase A serum than does RNase A, while the monomeric species is very similar to RNase A in its interaction with this antiserum (529). Bovine pancreatic RNase A, at physiological ionic strength and pH, has essentially no activity toward double-stranded RNA (530) or poly(A); activity toward the latter is observed at high concentrations of both enzyme and poly(A) (53f). On the other hand,-RNase BS-1 is active under physiological conditions toward double-stranded RNA, and will hydrolyze the polypyrimidine strand of a poly(A). poly(U) complex (509) and the RNA strand of a DNA-RNA hybrid (532). Similarly, aggregates of bovine pancreatic RNase A and its chemically cross-linked dimers (55, 533) act on double-stranded RNA (509,525), the RNA strand of a DNARNA hybrid (534), and measurably on poly(A) (530). Libonati and Floridi (509) considered it unlikely that the two active sites in the dimeric enzymes were important in the activity toward doublestranded substrates and suggested that the determining factor was probably the positive charge density of the enzyme. This hypothesis received support from studies with monomeric derivatives of RNase BS-1 (520, 535) and with species variants of the enzyme that exhibited different 527. Irie, M., and Suito, F. (1975). J . Eiochetn. (Tokyo) 77, 1075. 528. Parente, A., Branno, M., Malorni, M. C.. Welling, G. W., Libonati, M., and D’Alessio, G. (1976). EEA 445, 377. 529. Floridi, A., and Fini, C. (1972). f m l . J . Biochem. 21, 72. 530. Libonati, M. (1971). BEA 228, 440. 531. Beers, R. F., Jr. (1960). JEC 235, 2393. 532. Taniguchi, T., and Libonati, M. (1974). BERC 58, 280. 533. Bartholeyns, J . , and Moore, S. (1974). Science 186, 444. 534. Libonati, M., Sorrentino, S. , Galli, R., La Montagna, R., and Di Donato, A . (1975). BBA 407, 292. 535. Libonati, M . , Malorni, C., Parente, A., and D’Alessio, G. (1975). EBA 402, 83.

12. PANCREATIC RIBONUCLEASES

415

basicities (536);both the net positive charge and the positions of the basic residues are considered to contribute to the effectiveness of the enzyme toward double-stranded substrates. In that connection, Wang and Moore (59) have shown that the highly basic derivative of RNase A cross-linked to octaspermine with dimethyl suberimidate is 115 and 380 times as active as RNase A toward poly(A) 'poly(U) and the hybrid poly(rU) 'poly(dA), respectively, the increased activity being primarily a result of a 100-fold decrease in the K m for these substrates (537). The resistance of double-stranded RNA and poly(A) to hydrolysis by bovine RNase A is only observed at physiological ionic strength. At onetenth the physiological ionic strength, the activity of bovine RNase A is equal to or greater than the activities of RNase BS-1, the more basic whale pancreatic RNase, and chemically cross-linked dimers of bovine RNase A (538, 539). Although the mechanisms are not fully established (540), the effects of ionic strength on the relative activities of the various RNases toward single- and double-stranded substrates reflect changes in the K , of the substrates for each enzyme. For example, the K ifor poly(A) inhibition of bovine RNase A activity toward yeast ribosomal RNA decreases 10-fold with respect to the substrate K , as the ionic strength is decreased from 0.1 M to 0.005 M in Tris-HC1 buffer at pH 7.5 (P. Blackburn and R. Jacoby, unpublished results), making it a relatively stronger competitive inhibitor, whereas at physiological ionic strength the Ki for poly(A) and the K , for RNA are almost identical. The K, of RNase BS-I for yeast RNA increases significantly as the ionic strength is increased above 0.2 M salt (541). Bovine RNase A dimers and RNase BS-1 destabilize the structure of double-stranded DNA (542) much more efficiently than does bovine RNase A (543). When the known interactions of nucleotides with bovine pancreatic RNase are considered with respect to the X-ray structure of the molecule ( I , I 6 , 2 / , 7 3 ) it is clear that at least local strand separation of a double-stranded substrate must occur to accommodate the polynucleotide into the enzyme active site. 536. Libonati, M . , Furia, A , , and Beintema, J . J . (1976). EJB 69, 445. 537. Wang, D. (1979). BBA 568, 488. 538. Palmieri, M., and Libonati, M. (1977). BBA 474, 456. 539. Libonati, M., and Palmieri, M . (1978). BBA 518, 277. 540. Sorrentino, S . , Carsana, A., Furia, A , , Doskofil, J., and Libonati, M. (1980). BBA 609, 40. 541. Floridi, A . , and Fini, C. (1973). ftu1. J . Biocliern. 22, 7 . 542. Libonati, M . , and Beintema, J . J. (1977). Biochem. Soc. TrcrnA. 5, 470. 543. Felsenfeld, G., Sandeen, G., and von Hippel, P. H . (1963). P N A S 50, 644.

416 VII.

A.

PETER BLACKBURN AND STANFORD MOORE

Cytoplasmic RNase Inhibitor

PURIFICATION A N D CHEMICAL PROPERTIES

Nonpancreatic tissues contain small amounts of RNases that in many respects resemble the pancreatic enzyme. These RNases have not been thoroughly characterized, primarily as a result of the minute quantities available for study. Levy and Karpetsky (544) and Maor and Mardiney (545) have reviewed studies on these RNases from human tissue, serum, plasma, and urine in relation to neoplastic and nonneoplastic diseases. Investigations of tissue RNases of the pancreatic type require consideration of the presence of an endogenous RNase inhibitor. Normally, more than 95% of the available RNase activity measured near neutral pH in the postmitochondrial supernatant fraction of mammalian tissues is in a latent form. The RNase, bound by an inhibitor, forms an inactive complex maintained by a 6- to 8-fold molar excess of free inhibitor over the enzyme. The presence in mammalian tissues of this inhibitor of neutral RNase activity was first described in 1952 by Pirotte and Desreux (546). Since that time, most if not all mammalian tissues have been found to contain small amounts (e.g., 1 part in 10,000 on a protein basis) of this RNase inhibitor in the cytoplasm. Many of the early studies on the inhibitor were performed by Roth and his colleagues (547-550) and have been reviewed by Roth (551). The inhibitor is not restricted to mammalian species and has been found in marsupial (552,553),amphibian (553n), and avian livers (552, 554, 55.5) where it exhibits species specificity toward its respective neutral RNase. 544. Levy, C. C., and Karpetsky, T. P. (1981).//1"Enzymes as Drugs'' ( J . S. Holcenberg, ed.), p. 103. Wiley, New York. 545. Maor, D . , and Mardiney, M. R., Jr. (1979). CRC Crit. Rev. Clin. Lab. Sci. 10, 89. 546. Pirotte, M., and Desreux, V. (1952). Bull. Soc. Chim. Belges 61, 167. 547. Roth, J. S. (1956). BBA 21, 34. 548. Roth, J. S. (1958).JBC 231, 1085. 549. Roth, J. S . (1958). JBC 231, 1097. 550. Roth, J. S . (1962). BBA 61, 903. 551. Roth, J. S. (1967). Merhods Cancer Res. 3, 153. 552. Kraft, N., and Shortman, K. (1970). Austr. J . B i d . Sci. 23, 175. 553. Meyer, D. H . , Meyer, W. L . , and Kakulas, B. A. (1976). I n "Recent Advances in Myology" (W. G. Bradley, D. Gardner-Medwin, and J. N. Walton, eds.), p. 277. Excerpta Medica, Amsterdam. 553a. Malicka-Blaskiewicz, C. (1978). Proc. 12th FEES Meet. Dresdeti. Absrr. 0122. 554. Kraus, A. A., and Scholtissek, C. (1974). EJB 48, 345. 555. Dijkstra, J., Touw, J., Halsema, I., Gruber, M., and AB, G. (1978). BBA 521, 363.

12. PANCREATIC RIBONUCLEASES

417

The researches by Roth (547, 548) and Shortman (556, 557) on the RNase inhibitor of rat liver demonstrated that the inhibitor was a heatand acid-labile, sulfhydryl-dependent protein readily inactivated by p-hydroxymercuribenzoate with concomitant activation of the latent RNase. The protein does not inhibit acid lysosomal RNase (557) and is specific for neutral RNases of the pancreatic type. Generally, most purification procedures have used combinations of salt fractionation of the postmicrosomal or postmitochondrial supernatant fractions of isotonic extracts of mammalian tissues, followed by ion exchange chromatography and gel filtration. Using this approach Gribnau et a/. (558) determined that the molecular weight of the rat liver RNase inhibitor was near 50,000; although a 3000-fold purification was achieved, the protein was still impure by SDS gel electrophoresis. This purification was further extended by Gribnau et af. (559) using affinity chromotography on RNase A coupled to carboxymethyl-cellulose. A similar approach was later used by Gagnon and de Lamirande (560). Studies on the properties of the rat liver inhibitor purified by Gribnau el al. (5.58) demonstrated the importance of EDTA and free thiol, especially dithiothreitol, for maintaining the protein in its active form (561). However, the best preparations (559) were contaminated by a potent leucine aminopeptidase that was detrimental to studies on cell-free protein synthesis (562). A number of the properties of the partially purified RNase inhibitor have been described (563-571 ). The inhibitor was observed to be present 556. Shortman, K. (1961). BBA 51, 37. 557. Shortman, K . (1962). BBA 55, 88. 558. Gribnau, A. A. M., Schoenmakers, J. G. G . , and Bloemendal, H. (1969). ABB 130, 48. 559. Gribnau, A. A. M., Schoenmakers, J. G . G., van Kraaikamp, M., and Bloernendal, H. (1970). BBRC 38, 1064. 560. Gagnon, C., and de Lamirande, G. (1973). BBRC 51, 580. 561. Gribnau, A. A. M., Schoenmakers, J . G . G . , van Kraaikamp, M., Hilak, M . , and Bloemendal, H. (1970). BBA 224, 55. 562. Berns, A. J. M., Zweers, A., Gribnau, A. A. M., and Bloemendal, H. (1971). BBA 247, 62. 563. Ortwerth, B. J . , and Byrnes, R . J. (1972). Expt. Eye R e s . 14, 114. 564. Greif, R . L . , and Eich, E. F. (1977). Metabolism 26, 851. 565. Takahashi, Y., Mase, K., and Suzuki, Y. (1967). Experientia 23, 525. 566. Suzuki, Y., and Takahashi, Y. (1970). J . Neiirochem. 17, 1521. 567. Takahashi, Y., Mase, K., and Suzuki, Y. (1970). J . Neirrochem. 47, 1433. 568. Goto, S., and Mizuno, D. (1971). ABB 145, 64. 569. Goto, S . , and Mizuno, D. (1971). ABB 145, 71. 570. Bishay, E. S., and Nicholls, D. M. (1973). ABB 158, 185. 571. Nicholls, D. M., and Markle, H . V. (1974). C h e m . B i d . Inrercictions 8, 225.

418

PETER BLACKBURN AND STANFORD MOORE

in human placenta (572, 573), a tissue with considerable biosynthetic activity. From this readily available human tissue, the homogeneous inhibitor was isolated through the use of ion exchange chromatography and affinity chromatography on RNase A Sepharose (574). A simplified procedure employs only (NH&SOa fractionation and affinity chromatography (575). In common with the inhibitors of pancreatic RNase from other tissues, the placental inhibitor is an acidic protein with an isoionic point at pH 4.6 to 4.8. It has a mean average molecular weight, determined by SDS-polyacrylamide gel electrophoresis and gel filtration, of about 50,000, and forms a I : 1 complex with bovine pancreatic RNase .4. Ortwerth and Byrnes (563) reported a molecular weight of 32,000 for bovine lens RNase inhibitor. Bloemendal et (11. (576) obtained a molecular weight by gel filtration of near 55,000 for a preparation of calf-lens RNase inhibitor; an inactive preparation obtained by electroelution of the protein after polyacrylamide gel electrophoresis at pH 8.9 was found to give two bands on subsequent polyacrylamide electrophoresis in the presence of 6 M urea or sodium dodecyl sulfate. RNase inhibitor activity is generally assayed against bovine pancreatic RNase A by its ability to inhibit the enzymatic activity toward RNA according to the procedures of Roth (547)and Shortman (556). One unit of RNase inhibitor activity is defined as the amount of inhibitor required to inhibit the activity of 5 ng of RNase A by 50% (556). The spectrophotometric assay for bovine pancreatic RNase A toward 2' ,3'-cyclic CMP (252. 577, 578) has been adapted to the assay of purified RNase inhibitor (575). While this assay is not sensitive enough to be used with tissue extracts it provides a convenient method, and is the assay of choice for measuring the activities of purified preparations of the inhibitor. As is the case with the RNase inhibitors from bovine lens (563), rat kidney (570), and rat liver (570, 579) the placental inhibitor is a strong noncompetitive inhibitor of bovine pancreatic RNase A, with a K i of 3 x lO-'OM (574). Like the RNase inhibitor from rat liver (547),the placental RNase inhibitor is rapidly inactivated by agents that react with sulfhydryl groups, especially by p - hydroxymercuribenzoate. The inactivation is not reversed by the presence of excess free thiol, and results in dissociation of 572. Brody, S. (1957). BBA 24, 502. 573. Bardon, A., Pamula, Z , and Hillar, M.(1969). A m Biochim. Polon. 16, 119. 574. Blackburn, P., Wilson, G., and Moore, S. (1977). JBC 252, 5904. 575. Blackburn, P. (1979). JBC 254, 12484. 576. Bloernendal, H., Zweers, A., Koopmans, M., and van den Broek, W. (1977). BBRC 17, 416. 577. Richards, F. M. (1966). CR Troi.. Lob. Corlshrrg Ser. Chim. 29, 315. 578. Crook, E. M., Mathias, A. P., and Rabin, B. R. (1960). BJ 74, 234. 579. Bartholeyns, J . , and Baudhuin, P. (1977). BJ 163, 675.

419

12. PANCREATIC RIBONUCLEASES TABLE XI11

A M I N OACIDCOMPOSlTlONS

OF

MAMMALIA RNASE N INHIBITORS Residues/Molecule"

Amino acid

Human placentab

Bovine brain'

Asx Thr Ser Glx pro Gly Ala Val Met I le Leu TY Phe His LYS Arg t cys + cys Trp Total residues

47 16 45 64 17 36 34 24 2 12

43 20 40 65 18 53 38 22 2 9 88 5 3 5 15 20 30 -5 48 1

85

4 6 6 17 23 30 5 473

' I Calculated to the closest integer fit based on a molecular weight of 51,000. From Blackburn et rrl. (574). ' From Burton ef d.(581).

active enzyme from the RNase-inhibitor complex. In the absence of free thiol, the free RNase inhibitor is rapidly inactivated; the depletion of free thiol from extracts of human placenta can lead to inactivation of the inhibitor during extraction (574. 575). Electrophoretic studies on the inhibitors from rat, ovine, and bovine tissues indicated that their RNase inhibitors were similar (580). The RNase inhibitor has been purified to apparent homogeneity from bovine brain (581) and mammalian liver (582)by procedures based upon those of Blackburn ef NI. (574, 575). For comparison, the amino acid composition of the inhibitors from placenta and brain are shown in Table XIII. Of the 580. van den Broek, W. G. M., Koopmans, M. A. G., and Bloemendal, H . (1974). W o l . B i d . RtJp. 1, 295. 581. Burton, L. E . , Blackburn, P., and Moore, S. (1980). 1u/.J . P ~ p / i d r/ndPro/ein t~ Res. 16, 359. 582. Burton, L . E., and Fucci, N . P. (1982). f i i / . J . Prpritlc Proreill Res.. in press.

420

PETER BLACKBURN AND STANFORD MOORE TABLE XIV

INTERACTION OF MODIFIED RNASEA

Reagent or derivative Reduced and carboxamido methylated BJtanedione C-clohexanedione Iodoacetate Iodoacetate Methylacetimidate Methyl-p -hydroxybenzimidate Cyanate Cyanate Bromoacetate Des-( 121- 124)-RNase Des-(l19-124)-RNase RNase S-protein RNase S-peptide RNase S

WITH

RNASEINHIBITOR

Type and number of residues modified

4 Cystine 8 Arginine 4.0 Arginine 3.34 Histidine (residue-12) Histidine (residue-1 19) Lysine 9.1 Lysine 10.0 Lysine 3 Lysine 6.6 Lysine (residue-41)

FROM

HUMAN PLACENTA"

Enzymatic activity (% of RNase A)

Strength of interaction with inhibitor (l/R&

0


5.1 3.2 7 <1 0.9 0 0 0 0 1.2 0 0.4 0 99

1.0 0.45 1.3 3.6 0.27 0.25 0.1 <0.1 0.1 1.o 1.o 1.o 10.1 1.or

" From

Blackburn and Jailkhani (47) and Blackburn and Gavilanes ( 4 5 ) . of the inhibition of RNase A by the inhibitor. ' Determined from percentage inhibition of RNase S by inhibitor.

* Rm = Molar ratio of derivative to RNase A that gives 50% reversal

30 half-cystine plus cysteine residues per molecule of placental RNase inhibitor, at least 8 are present as free sulfhydryl (574). Studies on the functional groups involved in the enzyme-inhibitor interaction to date have been performed primarily with bovine pancreatic RNase A used as a model. Knowledge of the sequence of the enzyme (15) and the details of its three-dimensional structure (1, 16), along with the wealth of data concerning the importance of specific functional groups of the enzyme, have permitted conclusions to be drawn as to which regions of this enzyme are involved in binding to the inhibitor. Specific proteolytic and chemical modifications of RNase A were made. Competition-binding experiments were performed using the 2' ,3'-cyclic CMP assay described by Blackburn (575) to examine the effect of these modifications on the ability of RNase to interact with the inhibitor (45,47). The results, some of which are shown in Table XIV, were consistent with the noncompetitive mode of inhibition. The active-site residues of

12. PANCREATIC RIBONUCLEASES

42 1

the enzyme, His-12 and His-119, and the auxiliary residues Phe-120, Asp-121, and Ser-123 are not essential for the interaction with the inhibitor. Also, it was demonstrated that the binding site for the inhibitor resides within the S-protein part of the molecule; residues 1 through 20 are not essential for the interaction. The results suggest that one or more lysine residues might be involved. When bound to the enzyme, the inhibitor protected the enzyme from inactivation by reagents that react at Lys-41 of RNase A, indicating that this residue was essential for the interaction (47). Subsequently, specific carboxymethylation of Lys-41 of RNase A with bromoacetate (81) was shown to reduce significantly the strength of the interaction between the enzyme and its inhibitor (45). Modification of the four arginine residues of the enzyme had little effect on the interaction ( 4 7 ) . Unexpectedly, carboxymethylation of the active-site residue His- 119 of RNase A (583)resulted in a 3.5-fold increase in the strength of the interaction between the enzyme and the placental inhibitor (45, 47). To assess whether the enzyme modifications that altered the interaction with the inhibitor were specific, circular dichroism measurements were performed on a number of the derivatives listed in Table XIV (45). The results, coupled to the CD data that had been reported previously by others for des-(121-124)-RNase A (142, 143), des-(119- 124)-RNase A (263), and RNase S-protein (584), suggested that one or more tyrosine residues of the enzyme were important for its interaction with the inhibitor. Based on studies of the binding of the inhibitor with different pancreatic RNases of known sequence (26/, 443, 445, 446, 454), and with the bovine seminal plasma RNase BS-1 (457)and its monomeric component, it was concluded that Tyr-92 was important for the interaction. It is hypothesized that Tyr-92 is rendered more accessible for interaction with the inhibitor as a result of carboxymethylation of His- 119 of the enzyme (45). To more clearly define the regions of the RNase molecule that are in contact with the inhibitor, the available lysine residues of the enzyme inhibitor complex have been amidinated with methyl acetimidate (585589) under conditions that preserve the complex functionally intact (47). 583. 584. 585. 586. 587. 595. 588. C89.

Crestfield, A. M . , Stein, W. H . , and Moore, S. (1963). JBC 238, 2413; ibid.. 2421. Pflumrn, M. N . , and Beychok, S . (1969). JBC 244, 3973. Lambert, J . M . , and Perham, R. N . (1977). BJ 161, 49. Hunter, M . J . , and Ludwig, M. L. (1962). JACS 84, 3491. Ludwig, M. L . , and Hunter, M . J. (1967). “Methods in Enzymology,” Vol. 1 1 , p. Browne, D. T., and Kent, S. B. H. (1975). BBRC 67, 126. Browne, D. T., and Kent, S. B. H . (1975). BBRC 67, 133.

422

PETER BLACKBURN AND STANFORD MOORE TABLE XV THEPROTECTIVE EFFECTO F RNASEINHIBITORO N THE A M I D I N A T IOF ON LYSINE RESIDUESI N RNASEA

Lysine residue Protection afforded (56)”

1 <5

7

31

37

100

100

71

41 100

61 100

66 <5

91 100

98

<5

104 <5

Calculated from relative recovery of individual tryptic peptides and their overlap peptides, the sum of which equals 100%. From Blackburn and Gavilanes (44a). If

The resistance of eacetimidyllysine residues to hydrolysis by trypsin (590) has allowed (44n), after peptide mapping, the identification of the lysine residues of the enzyme that are fully protected by the inhibitor from amidination (Table XV). Lysine residues in the enzyme that are protected by the substrate analogue polyadenylic acid from amidination are identified in Section III,A, Table I. The results indicate that the contact regions for substrate and inhibitor are not identical, as to be expected for a noncompetitive inhibitor, but there is some overlap. From the data presented in Table XV, and those reported earlier (45. 4 7 ) , studied in reference to the three-dimensional structure of the enzyme (Fig. 2), the known contact points with the inhibitor can be placed into three groups, A, B, and C, with respect to their locations in the molecule. Group A includes (a) Lys-7, (b) Lys-41, Pro-42, and Val-43, and (c) Lys91, Tyr-92, and Pro-93. Group B includes Lys-31 and Lys-37. Group C is represented by Lys-61 and adjacent residues. Groups A and B are adjacently located, whereas C lies distal to both A and B. The current hypothesis to describe the interaction involves an extensive contact between the inhibitor and RNase A that spans from A through B and on around through C. Other contact areas must lie between A , B, and C; these cannot involve the regions of the RNase A molecule where the e N H 2 groups of lysine residues 1, 66, 98, and 104 are located, since these residues are not protected from amidination. The intervening contacts between B and C possibly involve the groove formed by residues Ser-77 to Thr-82, which are in a P-structure with residues Thr-100 to His-105. The interaction between the enzyme and the inhibitor involves both polar and nonpolar residues. A key ionic interaction involves the positively charged eNH2-group of Lys-41, probably through an interaction with a negatively charged group of the inhibitor. This interaction of Lys-41 may account for the inactivation of the enzyme, since the activity of the enzyme is sensi590. Hunter, M. J., and Ludwig, M. L. (1972). “Methods in Enzymology,” Vol. 25, p. 585.

12. PANCREATIC RIBONUCLEASES

423

tive to modification of the e N H , group of this residue (19, 23, 81, 591, 592). B.

STUDIES O N IN V I T R O PROTEIN SYNTHESIS

Kraft and Shortman (552, 593) first noted that the ratio of the inhibitor to neutral RNase activity tends to increase in tissues characterized by increased rates of RNA synthesis and accumulation (e.g., 594-598); conversely, tissues in which protein synthesis decreases and catabolic activity increases usually demonstrate lower levels of the inhibitor and elevated neutral RNase activity (e .g., 596-600). A detailed discussion of the extensive literature on this subject is beyond the scope of this review. Evidence has been obtained in iGtvo (601, 602) that the inhibitor serves to preserve fully functional messenger RNA in the course of protein biosynthesis. Inclusion of the inhibitor purified from the human placenta into such systems significantly increases the incorporation of radioactive amino acids into the larger molecular weight translation products (Fig. 8). Also, it has been shown by de Martynoff crl. (603) that the synthesis of complementary DNA by reverse transcriptase is significantly improved in the presence of a preparation of placental RNase inhibitor. This finding has wide significance for genetic engineering experiments as an aid in the synthesis of full-size reverse transcripts for insertion into plasmid DNA and also for mRNA sequence determination studies. Further evidence for the role of the RNase inhibitor comes from studies on the isolation of polysomes, the stability of which is greater in tissues characterized by high levels of inhibitor activity, as seen in regenerating rat liver (604, 605) and the liver of estrogenized roosters (555). Moreover, the endogenous RNase inhibitor in rat liver high-speed supernatants (558, 591. Hirs, C. H. W. (1962). Brookhaven Syrnp. Biol. 15, 154. 592. Carty, R . P., and Hirs, C. H. W. (1968). JBC 243, 5244; ibid.. 5254. 593. Kraft, N . , and Shortman, K. (1970). BBA 217, 164. 594. Liu, D. K., Williams, G. H., and Fritz, P. J. (1975). BJ 148, 67. 595. Kyner, D., Christman, J. K., and Acs, G. (1979). EJB 99, 395. 596. Greif, R. L., and Eich, E. F. (1972). BBA 286, 350. 597. Murthy, P. V. N., and McKenzie, J. M. (1974). Endocrinology 94, 74. 598. Brewer, E. N., Foster, L. B., and Sells, B. H. (1969). JBC 244, 1389. 599. Liu, D. K., and Matrisian, P. E. (1977). BJ 164, 371. 600. Karplus, M.,and Weaver, D. L . (1976). Nrrt/rrr (London) 260, 404. 601. Scheele, G., and Blackburn, P. (1979). PNAS 76, 4898. 602. Robbi, M., and Lazarow, P. B. (1978). PNAS 75, 4344. 603. de Martynoff, G., Pays, E., and Vassart, G. (1980). BBRC 93, 645. 604. Moriyama, T., Umeda, T., Nakashirna, S., Oura, H., and Tsukada, K. (1969). J. Biochrm. (Tokyo) 66, 151. 6 0 5 . Bont, W. S., Rezelman, G., Meisner, I., and Bloemendal, H. (1967). ABB 119, 36.

;:I 424

PETER BLACKBURN AND STANFORD MOORE

(B)

12

0

O O ’

/

-1

0

C .-

a

o0 0

0

C .-

E 0

‘0

30

60 90 Time (minl

120

FIG.8. (A) Effect of human placental RNase inhibitor on the in virro translation of dog pancrease mRNA in the wheat germ system. Radioactivity incorporated into protein was measured with (+I) and without (-1) inhibitor. (B) Fluorographic analysis of SDSpolyacrylamide gel patterns obtained with equal volumes of the translation mixtures prepared as described for A. The M, values correspond to those of the presecretory proteins preamylase (55K), preprocarboxypeptidases (46K), and the serine preproteases (27K). From Scheele and Blackburn (60/).

606-609) and the purified human placental RNase inhibitor (601) have been used to protect polysomes from degradation during their extraction. VIII.

Catalytic Properties

A. ASSAYS In the decade since a review of procedures for RNase assay was published in this treatise ( I ) ,there have been improvements in the sensitivity and modes of measurement of the transphosphorylation step and the hy606. 607. 608. 609.

Bont, W. S., Rezelman, G., and Bloemendal, H. (1965). BJ 95, 1%. Blobel, G., and Potter, V. R. (1966). PNAS 55, 1283. Takahashi, Y., Mase, K., and Sugano, H . (1966). BBA 119, 627. Burghouts, J. Th. M., Stols, A. L. H., and Bloemendal, H. (1970). BJ 119, 749.

12. PANCREATIC RIBONUCLEASES

425

drolysis step. The most commonly used procedures continue to be modifications of the spectrophotometric assay toward RNA as introduced by Kunitz (610), the measurement of acid-soluble nucleotides released from polynucleotides by Anfinsen et al. (61I), and the spectrophotometric assay of the course of hydrolysis of 2’,3’-cyclic CMP by Crooker al. (578) [cf., see Richards Ref. (189)l. The sensitivity of the precipitation assay has been increased by the use of a variety of radioactively labeled RNAs. With [3H]tRNAor [32P]tRNA, less than 1 part of RNase per 10 million in preparations of DNase can be measured (12). Mendelsohn and Young (612) determined the acid-soluble radioactivity from the action of RNase on leucyl-tRNA charged with [14C]leucine.For a direct assay of the transphosphorylation step, White et al. (30) used Up[’T]C as substrate; after thin-layer chromatography, the radioactivity was counted in the areas containing [“CIC and Up[ 14C]C. When the assay of Anfinsenet al. (611) was modified (613) to precipitate acid-soluble nucleotides from very dilute solution (0.02% RNA instead of the usual 0.25% RNA), anomalous results due to incomplete precipitation were obtained (55), particularly at acid pH in the presence of low concentrations of phosphate (0.005 M ) (614). Anfinsen et al. (611) diluted the sample after (not before) the precipitation step, and this is preferable. A spectrophotometric assay has been based upon measurement of color solubilized by RNase action on an insoluble RNA-acridine orange complex (615). Fluorometric assays have been described that measure the drop in fluorescence when RNase acts on the complex of RNA and ethidium bromide (616, 617). The drop in radioactivity of an insolubilized substrate, [1251]RNA-agarose,has been used to detect RNase down to a concentration of lov8pdml(618). The measurement of the hydrolysis step, as described by Crook et af. (578), requires substrate concentrations close to the K , value for RNase A and measures the spectrophotometric change at 286 nm, which is on the 610. Kunitz, M. (1946). JBC 164, 563. 611. Anfinsen, C. B., Redfield, R. R . , Choate, W. L., Page, J . , and Carroll (1954). JBC 207, 201. 612. Mendelsohn, S . L., and Young, D. A . (1978). BBA 519, 461. 613. Bartholeyns, J., Peeters-Joris, G . , Reychler, H . , and Baudhuin, P. (1975). EJB 57, 205. 614. Bartholeyns, J . , Wang, D., Blackburn, P., Wilson, G., Moore, S . , and Stein, W. H. (1977). I n t . J . Peptide Protein Res. 10, 172. 615. Chaplinski, T., and Webster, D. A. (1973). Anal. Biochem. 54, 395. 616. LePecq, J. B . , and Paoletti, C. (1966). Anal. Biocllern. 17, 100. 617. Kamm, R. C., Smith, A . G . , and Lyons, H. (1970). Anal. Biochem. 37, 333. 618. Egly, J. M., and Kempf, J . (1976). FEES Lett. 63, 250.

426

PETER BLACKBURN A N D STANFORD MOORE

descending arm of the absorption spectrum of the substrate. Precautions necessary for reproducible recordings have been discussed by Hugli et al. (619) and Blackburn (575). An assay has been described that measures phosphate liberated by the action of alkaline phosphatase on the products of RNase action on RNA (620).

The detection of nuclease activity in bands obtained electrophoretically in SDS-polyacrylamide gels that contain RNA has been studied by Rosenthal and Lacks (621). The SDS prevents adsorption of the enzyme to nucleic acid; after extensive washing with pH 7.6 buffer to remove SDS (to allow renaturation of RNase and its subsequent action on the substrate) the presence of RNase was evidenced by a dark band under UV light after staining the gel with ethidium bromide. Alternatively, Karpetsky et al. (622) conducted the electrophoresis at acid pH, without SDS, and then incubated the gels at neutral pH to allow the enzyme to act on the incorporated polynucleotide. Spermine was included to prevent binding to the polynucleotide during electrophoresis; staining was with pyronine Y. Both techniques were sensitive to 0.5 ng of RNase A. Sierakowska and Shugar ( 2 ) have reviewed the use of chromogenic substrates, such as uridine-3‘-(a-naphthyl phosphate) and 5’-O-benzyluridine-3’-(a-naphthyl phosphate) for assay of pancreatic-like RNases. B. INHIBITORS A N D ACTIVATORS Additions to the list of inhibitory nucleotides ( 1 ) include substrate analogs in which the carbohydrate moiety is arabinose; Pollard and Nagyvary (623) have found that Ara-3’-CMP (Ki= 0.1 mM) is bound five times as strongly as the ribose analog under the same conditions (pH 7.0, 25”, with 2’,3’-cyclic CMP as substrate). The anti conformation of arabinonucleosides (624) is considered to be a possible contributing factor to the stronger binding. White ef d.(30) have studied forty oligonucleotides as competitive inhibitors of the hydrolysis of RNA. ApUp ( K , = 0.5 mM) was the strongest of the series. Folk acid is an inhibitor of RNase when cyclic 2’,3‘-CMP is the substrate but not when RNA is the substrate (624 a ) . 619. Hugli, T. E . , Bustin, M., and Moore, S. (1973). Brtiin Res. 58, 191. 620. Stern, R . , and Wilczek, J. (1973). A n d . Biochem. 54, 419. 621. Rosenthal, A. L., and Lacks, S. A. (1977). AnuI. Biocliern. 80, 76. 622. Karpetsky, T. P., Davies, G . E . , Shriver, K . K . , and Levy, C. C. (1980). BJ 189, 277. 623. Pollard, D. R., and Nagyvary, J. (1973). Biochernisrry 12, 1063. 624. Emerson, T. R . , Swan, R. J . , and Ulbricht, T. L. V. (1967). Biochemistry 6, 843. 624a. Sawada, F., Kamesaka, Y.,and Irie, M. (1977). BBA 479, 188.

12. PANCREATIC RIBONUCLEASES

427

The inhibition of RNases is crucial in a variety of experiments that depend upon biologically active RNA. The role of the cytoplasmic inhibitor of RNases of the pancreatic type in the maintenance of functional RNA is reviewed in Section VII. Mendelsohn and Young (612) found a combination of sodium dodecyl sulfate and diethyl pyrocarbonate to be effective in the protection of RNA from degradation during its isolation. Chirgwin et ul. (274)demonstrated the efficiency of homogenization in 4 M guanidinium thiocyanate plus 0.1 M 2mercaptoethanol for this purpose. Oxovanadium ion (VO'+) forms stable complexes with nucleotide monophosphates, which are strong competitive inhibitors of RNase A (625).These complexes have been demonstrated to inhibit RNase activity very efficiently during extraction of polysomes from lymphocytes (626); they cannot be used during translation. Compounds that bind to nucleic acids can either decrease or increase the action of RNase. Chloroquine (627)at 0.13 mM can double the rate of hydrolysis of tRNA by RNase A in 0.02 M Tris-HC1 buffer at pH 7.3 and 37"; there was no effect with 2',3'-cyclic CMP as substrate. Spermine at 0.13 mM has been found to reduce the hydrolysis of tRNA by RNase A to about 50% of the control value under the above conditions (627). At lower concentrations (0.02 mM), spermine increased the activity of RNase twoto fourfold toward cyclic substrates and poly(C), but not toward poly(U) (628). The apparent K , of the substrate was not influenced; the effect was on the velocity of the reaction. In 0.01 M phosphate buffer at neutral pH, spermine has been observed to stimulate the cleavage of RNA by a human plasma RNase (629);the effect, being in part electrostatic in nature, can be expected to be less at physiological salt concentration. Wang (537) studied the kinetics of the action of RNase A cross-linked to polyspermine, a combination that shows increased ability to hydrolyze double-stranded substrates (59) at p H 7.5 in the presence of 0.125 M NaCI. The coupling of a single chain of octaspermine to the enzyme strengthens the binding to poly(A). poly(U) ( K , decreases from 270 to 2.7 pM in total U) and increases the V,,,, for hydrolysis of the susceptible poly(U) strand from 2.5 to 16.2 AAZ5*min-lmg-' of enzyme. There is evidence for inhibition by the RNase-resistant poly(A) tracts in the substrate; free poly(A) shows a K 1 of about 8 pM in total A (537). 625. Lienhard, G . E., Secemski, 1. I . . Koehler, I<. A , , and Lindquist, R . N . (1971). CSHSQB 36, 45. 18, 5143. 626. Berger, S. L . , and Birkenmeier, C. S. (1979). Eioclic~mistq~ 627. Holbrook, D. J . , Jr., Whichard, L . P., and Washington, M. E. (1975). EJB 60, 317. 628. Kumagi, H . , Igarashi, K . , Tsuji, I., Mori, C., and Hirose, S . (1980). Cliem. Phcirm. Blrll. (Jcipcrn) 28, 1189. 629. Schmukler, M., Jewett, P. B . , and Levy, C. C . (1975). JBC 250, 2206.

428

PETER BLACKBURN AND STANFORD MOORE

Jensen and von Hippel (630) examined the kinetics and the thermodynamic parameters that characterize the complex formation between RNase A and DNA; the system provides a model for studies of the destabilization of double-stranded DNA by a protein that binds to singlestranded sequences, which may be transiently exposed by fluctuations below the normal transition temperature for the double-stranded structure. C. KINETICS Rubsamen et t i / . ( 6 3 / ) observed sigmoidal kinetics in the action of RNase A at pH 7.6 on 2',3'-cyclic CMP and on esters of 3'-UMP in the substrate concentration range 1-15 mM; the results were explained by proposing a conformation equilibrium between two enzyme species in terms of the model for RNase A action proposed by Witzel and his colleagues (631).Working in a higher substrate concentration range, the initial rates of hydrolysis were observed by Walker rf t i / . (632) to change markedly with substrate concentration at pH 7.0 in Tris-HC1 buffer or in a pH-stat; there was a dip in the initial rate at 25 to 30 mM substrate, followed by a second rise and a gradual decline. Such data provided an explanation of the variations reported for K , values for 2' ,3'-cyclic CMP, which have ranged from near 1 mM to 7 mM with upper substrate concentrations of 10 to 250 mM. The authors propose an allosteric model for RNase A in which there is a substrate-dependent change in the equilibrium between the two enzyme conformations. They postulate the binding of six substrate molecules in the course of a cooperative substrateinduced transition; the binding is reminiscent of Crestfield and Allen's (633) early observation that the apparent isoionic point of RNase in phosphate buffer varied with phosphate concentration as a result of multiple binding sites for phosphate. The kinetic treatment was extended (634) to include inhibition by product or by a competitive inhibitor such as phosphate, and demonstrated (635) the decrease in the affinity for nucleotides with a decrease in the net positive charge of the protein, which can vary with the method of preparation of the enzyme, with the phosphate or sulfate content, or with the degree of deamidation upon storage of the phosphate-free protein (636). 630. Jensen, D. E . , and von Hippel, P. H. (1976). JBC 251, 7198. 631. Rubsamen, H . , Khandker, R., and Witzel, H. (1974). H o p p e Seyler's Z . Plrysiol. Cham. 355, 687. 632. Walker, E. J . , Ralston, G. B . , and Darvey, I. G. (1975). BJ 147, 425. 633. Crestfield, A. M., and Allen, F. W. (1954). JBC 211, 363. 634. Walker, E. J., Ralston, G . B . , and Darvey, 1. G. (1976). BJ 153, 329. 635. Walker, E. J . , Ralston, G. B., and Darvey, I. G. (1978). BJ 173, 1. 636. Walker, E. J., Ralston, G. B . , and Darvey, 1. G . (1978). BJ 173, 5 .

12. PANCREATIC RIBONUCLEASES

429

The temperature-dependence of the hydrolysis of 2‘ ,3‘-cyclic CMP by RNase A at pH 5 has been studied by Matheson and Scheraga (637); the data indicate that a small conformational change occurs in the enzyme near 32”, well below the temperature of the main thermal transition. Walz (638) measured the binding of deoxyuridine 3’-phosphate to assess the role of the 2’-OH group in the affinity for the enzyme. Kinetic and equilibrium binding studies showed that the deoxy derivative has an apparent Kd of 0.38 mM compared with 0.07 mM for 3‘-UMP and that the bound 2’-hydroxyl group of 3’-UMP interacts with RNase in a specific fashion that influences the interactions of the 3’-phosphate group with the enzyme, as well as an isomerization process associated with formation of the RNase-nucleotide complex. Li and Walz (639) studied the influence of a phosphate group on the 5‘-OH of 2‘,3‘-cyclicUMP; at pH 5.5 the derivative is bound more strongly ( K , is 0.03 mM, which is 23-fold lower than that for 2‘,3‘-cyclic UMP) but the turnover number was fivefold lower. The authors consider it likely that the 5’-phosphate group is subject to attraction by a group such as Lys-41. As noted in Section III,A,I, chemical and physical studies provide data for the presence [see Fig. 20 of Ref. ( f ) ] of three base- and phosphate-binding sites (B and R for base and ribose), PO,B1, R1, p1, R2, B2, pZ, R3, B3. A comparison of the kinetic parameters for the action of pancreatic RNases of known sequence from five mammalian species toward 2‘,3’cyclic CMP and UMP have been determined by Ronda et al. (2.58). The results for all five enzymes are similar, a finding consistent with the preservation of the overall features of the molecule deemed essential for activity; the main difference is in the turnover numbers rather than in the K , values. Avramova et crl. (640) used CD spectra to confirm that the very slow depolymerization of poly(A) at pH 6.5 by RNase A is a result of a low V,,,,,rather than to a change in K , ; the latter is similar to that for poly(U). Cozzone and Jardetsky (641) studied the transphosphorylation step with poly(A) by 3’P NMR. The reaction proceeds slowly at pH 7.9 in 0.1 M Tris-HC1 buffer at temperatures above 35”. The enzymatic process with poly(A) (1) is known to stop at the cyclic phosphate stage; at 160 hours the presence of a small amount of 2’- and 3’-AMP is attributed to nonenzymatic hydrolysis at pH 7.9. Libonati and associates (530, 540) studied the action 637. 638. 639. 640. Biol. 8, 641.

Matheson, R . R . , J r . , and Scheraga, H. A. (1979). Biorhemisrry 18, 2446. . ~ 2156. Walz, F. G . , Jr. (1971). B k x - h t ~ m b t r 10, Li, J. R.-T., and Walz, F. G . , Jr. (1974). A B B 161, 227. Avrarnova, 2. V., Dudkin, S . M . , and Karabashyan, L. V. (1974). Mo/eXir/yirn~r.wr 501. Cozzone, P. J . , and Jardetsky, 0. (1977). FEBS Lett. 73, 77.

430

PETER BLACKBURN AND STANFORD MOORE

RNase- A

O\/O

H,O"

"nu

s o

on

Q

s

(Et O), P 0 C I

2-

FIG.9. The method used to determine the geometry of the ring opening step. From Usher cf t i / . ( 6 4 3 ) .

of RNase A on poly(A) and on double-stranded substrates such as poly(A) . poly(U), with emphasis on the multiple effects of ionic environment on the substrate and/or on the enzyme-substrate interaction.

D. MECHANISM O F CATALYSIS Steps 1. Geometry of the TUY>

The synthesis of the two stereoisomers of uridine 2',3'-cyclic phosphorothioate by Eckstein (642) opened the way to the establishment of the geometry of the two steps in the action of RNase A [Usher et d.(643, fj44)]. The hydrolysis step, the opening of the ring of isomer N (Fig. 9) by the enzyme, was conducted in water enriched in lXO.The ring was then reclosed by a known in-line (SN2)cyclization. If the geometry of the enzymatic ring opening is also in line (as is shown in Fig. 9), then the isomer ( I that is produced should contain no excess IXO;the isotope should be in the h isomer. The results of measurement of the IXOincorporation in the two isomers were consistent with the in-line hypothesis. The transphosphorylation step was then studied (644). Advantage was taken of the reversibility of the transphosphorylation reaction (Fig. 10). The dinucleotide Up(S)C was synthesized from the CI isomer of the cyclic substrate and cytidine by the catalytic action of RNase A. The reformation of the cyclic phosphorothioate from the dinucleotide by base catalysis, which is known to follow an in-line mechanism, gave the origi642. Eckstein, F. (1970). JACS 92, 4718. 643. Usher, D. A., Richardson, D. I . , Jr., and Eckstein, F. (1970). Nutiire (London) 228, 663. 644. Usher, D. A . , Erenrich, E. S . , and Eckstein, F. (1972). P N A S 69, 115.

43 1

12. PANCREATIC RIBONUCLEASES

c +

RNase A UC(S)

d

Base Up(S)C ---+U^P(S)

In-Line

Up(S)C

+ c

F I G .10. The overall method used to test the geometry of the first step of RNase action. From Usher cf t r l . ( 6 4 4 ) .

nalu isomer (determined by NMR). The simplest explanation of these geometric results, and the one consistent with most evidence from organic chemistry, is that both steps in the action of RNase A proceed by in-line mechanisms, each of which results in the inversion of the absolute configuration around the phosphorus. Witzel and associates ( 6 3 / ) ,in their proposal for the mode of action of RNase A, postulate an adjacent mechanism; to accommodate the data of Usher et ti/. (643, 644), they invoke turnstile rotation to permit inversion of the absolute configuration as a consequence of an adjacent mechanism. 2. Confbmiirtioti of tlir Siibstmte

Several lines of investigation have been directed toward the question of the torsional angle for the glycosidic bond in the enzyme-substrate complex. Interpretation of nuclear Overhauser effects have led Karpeisky and Yakovlev (64s) to conclude that the pH-dependent conformation of 3’-CMP holds in the anti orientation up to near pH 7. Earlier NMR data (327) suggested that the syn conformer may prevail at neutral pH. Gorenstein rt (11. (646) applied molecular orbital calculations to the torsional activation of phosphodiester bonds, with reference to RNase A; they conclude that molecules in a gauche, transconformation are activated for cleavage, and consider that the data on the preferential hydrolysis of the intranucleotide linkage 33-34 in Phe-tRNA is evidence that the stereoelectronic consequences of a gauche, trans conformation are taken advantage of in this instance. 3. M e c h t i i s t i c , Models

Richards and Wyckoff ( I ) and Benz and Roberts (647)have discussed in detail the chemical and physical data upon which proposals for the mechanism of action of RNase A can be built. The two main lines of reasoning continue to yield two quite different hypotheses. The initial suggestion, 645. Karpeisky, M . Ya, and Yakolev, G . I . (1977). FEBS Lett. 75, 70. 646. Gorenstein, D. G., Findlay, J. B . , Luxon, B. A , , and Kar, D. (1977).JACS 99, 3473. 647. Benz, F. W., and Roberts, G. C. K. (1973). 1~ “Physico-Chemical Properties of Nucleic Acids” (J. Duchesne, e d . ) , Vol. 3, p . 77. Academic Press, London.

432

PETER BLACKBURN AND STANFORD MOORE

proposed on kinetic grounds by Mathias and Rabin and their colleagues, [see Ref. (I)] has over the years been fitted into the accumulated framework of structural and kinetic data on the enzyme. Deakyne and Allen (648) have reviewed this approach and applied molecular orbital theory calculations to some of the steps. The imidazole ring of His-119 is considered to activate the leaving group and to facilitate the in-line addition of 0-2’ to the phosphodiester group. Hydrogen bonds between the peptidy1 carbonyl of Thr-45 and N-1 of His-12, and N-3 of His-12 and the 2’-OH may increase the nucleophilicity of 0-2’. The backbone N-H from residue 120 can increase the electrophilicity of the phosphorus; Gln11 * HOH * phosphoryl hydrogen bonds are expected to have a similar function. Lys-41 increases the electrophilicity of the phosphorus and stabilizes a trigonal-bipyramidal intermediate. Holmes (649) discussed a square pyramidal model as an alternative to the trigonal intermediate. Rein et a / . (650)combined molecular orbital and perturbation theory to consider the configurational flexibility of the active site; their conclusions on the positioning of the two imidazole rings are consistent with those of earlier reviews (I, 647). In applying molecular orbital theory to the second step, Umeyama et al. (651) postulate that coincident with a change in the shape of the active site, a charge relay system involving Asp- 121 and His- 119 may be involved in the opening of the 2’ ,3’-cyclic substrate. Bellmann and Witzel’s (652) observation that fully purified carboxymethyl-His- 12-RNase A has no activity strengthens the view that in RNase A His-12 should have a well-defined role in whatever mechanism is proposed. The approach of Witzel and his associates (631) is entirely different. Their central premise is that the 2-0x0 group of the pyrimidine activates the 2’-OH group of the sugar in the first step, or a water molecule in the second step. A diimidazole system is proposed in which His-119 is hydrogen-bonded to His- 12. Lys-41 contributes to the electrostatic binding of the phosphate group. The ring of Phe-120 interacts with the pyrimidine ring. [Lin et a / . ’ s (255) observation that a substitution of Leu for Phe at position 120 gives a product with 13% activity indicates that an aromatic ring at this position is helpful but not essential.] Witzel and

-

648. Lkakyne, C. A . , and Allen, L. C. (1979). JACS 101, 3951. 649. Holmes, R. R. (1976). I n ? . J . Peptide Protein Res. 8, 445. 650. Rein, R., Renugopalakrishnan, V., and Barnard, E. A. (1971). In “Proceedings of the First European Biophysics Congress” (E. Broda, A. Locker, and H. Springer-Lederer, eds.), Vol. 6, p. 35. Verlag der Wiener Medizinischen Akademie, Vienna, Austria. 651. Umeyama, H . , Nakagawa, S. , and Fuji, T. (1979). Chem. Pharm. Bull. (Japan) 27, 974. 652. Bellmann, B . , and Witzel, H. (1980). Hoppe-Sryler’s 2. Physiol. Cheni. 361, 218.

12. PANCREATIC RIBONUCLEASES

433

colleagues presented kinetic analyses and NMR data in support of individual structural features of the formulation (631 ). IX.

Research Applications

Among the uses of RNase as a probe in the study of ribonucleic acids, the enzyme is one of the nucleases widely employed in the course of the sequencing of RNA; references to the current methodology are included in the article on RNase TI (653) in this volume. ThF usefulness of the cytoplasmic inhibitor of RNase to protect RNA in the course of the synthesis of complementary DNA by reverse transcriptase is cited in Section VII; that section also covers the use of the inhibitor to protect mRNA during the course of in vitro translation and during the preparation of rough microsomes and detached polysomes. Clinical applications, in general, are subject to further research; these include the possible value of plasma or serum levels of neutral RNase as indicators of neoplastic disease (544,545).Encouraging clinical trials have been reported on the efficacy of intramuscular administration of bovine RNase A for the treatment of infection by the RNA-containing virus of tick-borne encephalitis (654). The antitumor activities of bovine seminal plasma RNase BS-1 (e.g., 655, 656) and cross-linked dimers of bovine RNase A (e.g., 57, 58, 657) merit further study. Ac K NOW L EDGMENTS The literature survey for this review grew in part from the researches under NIH Grant GM 25323 on the Biochemistry of Nucleases. The library research, organization of the bibliography, and the assembly of the manuscript were completed with the skillful cooperation of Lorraine Ackerman. The authors are especially indebted t o J. J. Beintema and associates for making available unpublished sequence data on species variations and to readers of several sections of the manuscript who provided counsel and papers in press from their laboratories.

653. Takahashi, K . , and Moore, S. (1981). “The Enzymes,” 3rd Ed., this volume, Chap. 13. 654. Glukhov, B. N., Jerusalimsky, A. P., Canter, V. M., and Salganik, R. I. (1976). Arch. Nertrol. 33, 598. 655. MatouSek, J. (1973). Experienria 29, 858. 656. Vescia, S . , Tramontano, D., Augusti-Tocco, G., and D’Alessio, G. (1980). Cuncer Res. 40, 3740. 657. Bartholeyns, J . , and Baudhuin, P. (1976). P N A S 73, 573.