Influence of the Carbohydrate Moiety on the Proteolytic ... - Springer Link

Journal of Protein Chemistry, Vol. 17, No. 5, 1998

Influence of the Carbohydrate Moiety on the Proteolytic Cleavage Sites in Ribonuclease B Ulrich Arnold,1,2 Angelika Schierhorn,1 and Renate Ulbrich-Hofmann1 Received December 5, 1997

The influence of glycosylation on proteolytic degradation was studied by comparing cleavage sites in ribonuclease A (RNase A) and ribonuclease B (RNase B), which only differ by a carbohydrate chain attached to Asn34 in RNase B. Primary cleavage sites in RNase B were determined by identifying complementary fragments using matrix-assisted laser desorption/ionization mass spectrometry and compared with those in RNase A [Arnold et al. (1996), Eur. J. Biochem. 237, 862869]. RNase B was cleaved by subtilisin even at 25°C at Ala20-Ser21 as known for RNase A. Under thermal unfolding, the peptide bonds Asn34-Leu35 and Thr45-Phe46 were identified as primary cleavage sites for thermolysin and Lys3 l-Ser32 for trypsin. These sites are widely identical with those in RNase A. Treatment of reduced and carbamidomethylated RNase A and RNase B with trypsin led to a fast degradation and revealed new primary cleavage sites. Therefore, the state of unfolding seems to determine the sequence of degradation steps more than steric hindrance by the carbohydrate moiety does. KEY WORDS: Ribonuclease B; proteolysis; carbohydrate chain; trypsin; thermolysin.

of the couple ribonuclease A (RNase A)3 and ribonuclease B (RNase B). Both enzymes occurring in bovine pancreas have identical protein sequences consisting of 124 amino acid residues (Smyth et al., 1963; Plummer et al, 1968), which form a nearly identical tertiary structure (Berman et al, 1981; Williams et al, 1987). They only differ by an oligosaccharide chain N-linked to Asn34 in RNase B (Plummer et al, 1968). This carbohydrate chain is not uniform, but contains 5-9 mannose units (GlcNAc2Man5_9), resulting in an increase of the molecular mass from 13,683 Da (RNase A) to 14,899-

1. INTRODUCTION Glycosylation is the most extensively occurring natural modification of proteins. Its biological role, however, is not sufficiently understood. Such cellular functions as traffic marker, signal modifier, or regulator of biological activity are attributed to the carbohydrate content in glycoproteins (Rademacher et al., 1988; Lis and Sharon, 1993). The comparison of the molecular properties of glycosylated and nonglycosylated proteins revealed differences in their solubility (Jaenicke, 1991), thermal stability (Mer et al., 1996), and susceptibility toward proteolytic attack (Rudd et al, 1995), which might be of biological significance, too. An ideal model system for studying effects induced by glycosylation has been provided by nature in the form

3

1

Department of Biochemistry/Biotechnology, Martin-Luther University Halle-Wittenberg, D-06120 Halle, Germany 2 To whom correspondence should be addressed.

Abbreviations: CAM-RNase, reduced and carbamidomethylated RNase; CD, circular dichroism; DTE, 1,4-dithioerythritol; EDTA, ethylenediaminetetraacetic acid disodium salt; GdnHCl, guanidine hydrochloride; MALDI-MS, matrix-assisted laser desorption/ ionization mass spectrometry; NMR, nuclear magnetic resonance; PMSF, phenylmethanesulfonyl fluoride; RNase, ribonuclease; RPHPLC, reversed-phase high-performance liquid chromatography; SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; TFA, trifluoroacetic acid; Tris, tris(hydroxymethyl)aminomethane; UV, ultraviolet.

397 0277-8033/98/0700-0397$15.00/0 © 1998 Plenum Publishing Corporation

398 15,547 Da (RNase B). Despite the great similarities in the conformation of the protein component of RNase A and RNase B, differences were described for the thermodynamic and kinetic stabilities. Puett (1973) found an increase of the thermodynamic stability by several hundreds of calories for an RNase B preparation from pancreatic juice. H-D exchange nuclear magnetic resonance (NMR) and circular dichroism (CD) spectroscopy revealed a slight stabilization of RNase B involving not only the region near the glycosylation site, but even remote regions (Joao et al, 1992; Joao and Dwek, 1993). H-D exchange NMR experiments by Rudd et al. (1994) showed an increased dynamic stability of RNase B connected with a decreased functional activity. These properties were attributed to an increased rigidity and a shielding of substrate-binding sites of RNase B by the carbohydrate moiety. Comparing the thermodynamic and kinetic thermal stabilities of RNase A and RNase B, Arnold and Ulbrich-Hofmann (1997) found that RNase B is more stable than RNase A by 2.5 kJ mol-1 for AG° and 2.2 kJ mol-1 for AG#. Yamaguchi and Uchida (1996) studied the oxidative refolding of RNase A and RNase B and observed that the oligosaccharide not only prevents the protein from aggregation, but also enhances the rate of reactivation of RNase B in a chaperone-like fashion. The proteolytic susceptibility of RNase A and RNase B, however, has been scarely compared, although limited proteolysis is a sensitive tool to reveal delicate differences in local conformation, and increasing resistance toward proteolytic attack by shielding potential cleavage sites due to glycosylation may also be important in vivo. By following the activity decrease, Birkeland and Christensen (1975) showed that RNase B is more resistant toward proteolytic attack by trypsin, chymotrypsin, elastase, and pepsin than RNase A. Rudd et al. (1994) observed a decreased susceptibility of RNase B toward proteolytic attack by Pronase. Arnold and Ulbrich-Hofmann (1997) used trypsin and thermolysin to determine the thermal unfolding constants of RNase A and RNase B. In contrast to the restricted information on the proteolytic degradation of RNase B, there has been a large number of studies on the proteolytic degradation of RNase A, which were mainly focused on probing the conformation and unfolding or refolding mechanisms of RNase A. Results by Ooi et al. (1963), Rupley and Scheraga (1963), and Klee (1967) gave some information on the thermally induced unfolding pathway of RNase A. The role of proline isomerization in the unfolding and refolding of RNase A was investigated by isomerspecific proteolysis (Lin and Brandts, 1983, 1984), while

Arnold, Schierhorn, and Ulbrich-Hofmann the refolding of GdnHCl-denatured RNase A was followed by a trypsin-pulse method (Lang and Schmid, 1986). Recently, Arnold et al. (1996) localized the structural region first unfolded under thermal denaturation by limited proteolysis with trypsin and thermolysin. This region is modified by the carbohydrate chain in RNase B, so that dramatic changes in the unfolding pathway and the position of primary proteolytic cleavage sites were to be expected. In the present paper, the primary cleavage sites in native and thermally denatured RNase B are located for subtilisin, trypsin, and thermolysin and are compared to those of RNase A. The studies have been completed by the analysis of proteolytic fragments in the degradation of the reduced and carbamidomethylated RNase A and RNase B (CAM-RNases). The results allow us to compare the structural flexibility and steric accessibility of the two enzymes in their native, thermally denatured, and reduced/carbamidomethylated states. 2. MATERIALS AND METHODS RNase A from Serva and RNase B from Sigma were purified on an FPLC-column MONO S (Pharmacia) resulting in single bands in the SDS-PAGE. Thermolysin, subtilisin Carlsberg, angiotensin II (human), insulin (bovine), cytochrome c (horse heart), soybean trypsin inhibitor, bovine pancreas trypsin inhibitor from Sigma, and trypsin (treated with N-a-tosyl-L-phenylalanine chloromethyl ketone) from Serva were used without further purification. Acrylamide, N,N'-methylenebisacrylamide, N,N,N', N'-tetramethylethylenediamine, and ammonium persulfate were purchased from Pharmacia, tris(hydroxymethyl)aminomethane (Tris), N-tris(hydroxymethyl) methylglycine, and Coomassie brillant blue G250 from Serva, calcium chloride and phenylmethanesulfonyl fluoride (PMSF) from Merck, acetonitrile (ultravioletgrade) from Roth, ultrapure GdnHCl from SchwarzMann Biotech, and trifluoroacetic acid (TFA), 1,4-dithioerythritol (DTE), ethylenediaminetetraacetic acid disodium salt (EDTA), iodoacetamide, and SDS from Sigma. Sinapinic acid from Aldrich and a-cyano-4-hydroxycinnamic acid from Sigma were twice recrystallized from methanol. All other reagents were the purest ones commercially available. 2.1. Proteolysis In a typical experiment 80 (0,1 of 50 mM Tris-HCl buffer, pH 8.0, was preincubated in a thermostat RM 6

Proteolytic Cleavage Sites in Ribonuclease B (LAUDA, accuracy ±0.1°C). At zero time 10 ul of protease solution of trypsin or subtilisin (0.005 mg ml -1 ) in 50 mM HCl buffer, pH 8.0, or thermolysin (0.0125 mg ml-1) in 50 mM Tris-HCl buffer, 10 mM CaCl2, pH 8.0, and 10 u1 of RNase (2.5 mg ml -1 ) was added. After distinct time intervals, samples of 15 ul were removed and mixed with 5 ul of 50 mM EDTA in the case of thermolysin or 5 ul of 50 mM PMSF (dissolved in 2propanol) in the case of trypsin and subtilisin.

399 stirring in a solution containing 40% ethanol, 10% methanol, 10% acetic acid, and 20% water) for at least 2 hr followed by destaining with methanol solutions of increasing concentrations (20-40% methanol). The gels were scanned at 595 nm using a densitometer CD 60 (Desaga). Peak areas of the obtained chromatograms were estimated by use of the software supplied with the densitometer. 2.4. RP-HPLC of the Proteolytic Fragments

2.2. Preparation of CAM-RNases and Tryptic Digestion Bisulfide bonds of the intact RNases were reduced by 20 mM DTE, 5 M GdnHCl in 0.2 M Tris-HCl buffer, pH 8.0, in the dark for 2 hr. The resulting cysteine residues were carbamidomethylated by 100 mM iodoacetamide for 30 min. Both operations were performed under nitrogen at room temperature. Desalting of the samples was performed on an inert HPLC system (Merck-Hitachi) using an octyl reversed-phase column (250 X 4 mm) from Vydac. The solvent gradient was produced from 0.05% TFA in degassed HPLC-grade water (solvent A) and 0.045% TFA in degassed acetonitrile (solvent B). The flow rate was 1.0 ml min -1 . Absorbance was followed at 214 nm. The fractions of CAM-RNases were collected manually, dried under nitrogen, and resolved in 50 mM Tris-HCl buffer, pH 8.0. Trypsin in 50 mM Tris-HCl buffer, pH 8.0, was added to give a final RNase to protease ratio of 5000:1 (by mass). The reaction was stopped after 1 min at 25°C by adding PMSF (dissolved in 2-propanol) to give a final concentration of 15 mM.

2.3. Electrophoresis and Densitometric Evaluation Electrophoresis was carried out on a Midget electrophoresis unit (Hoefer) according to Schagger and von Jagow (1987), but using 10%, 14%, and 18% acrylamide for the sampling, spacer, and separating gels, respectively. Reducing conditions were applied in order to obtain linear peptide fragments. Silver staining was performed according to Blum et al. (1987). The molecular masses of the fragments were estimated from the semilogarithmic plot of the molecular masses against the distance of migration by use of soybean trypsin inhibitor (21.0 kDa), cytochrome c (12.4 kDa), and bovine pancreas trypsin inhibitor (6.5 kDa) as standard proteins. Coomassie staining was performed by incubation of the SDS-PAGE gels in a Coomassie brillant blue G250 solution (2.5 g of Coomassie G250 was dissolved under

Disulfide bonds of the proteolytic fragments were reduced and carbamidomethylated as described for CAM-RNases. The resulting peptides were separated by the same equipment, solvents, and technical parameters as used for the purification of the CAM-RNases. The fractions applied to protein sequencing and MALDI-MS were collected manually. 2.5. MALDI-MS Determination of the molecular masses was performed on the reflectron-type time-of-flight mass spectrometer Reflex™ (Bruker-Franzen). Ions formed by laser desorption at 337 nm (N2 laser, 4 nsec pulse width) were recorded at an acceleration voltage of 28.5 kV in the linear mode. In general, 10-50 single spectra were accumulated for improving the signal-noise ratio and analyzed by use of the software XMASS supplied with the spectrometer. Sinapinic acid and a-cyano-4-hydroxycinnamic acid were used as UV-absorbing matrices. One ul of a saturated solution of the matrix compounds in 0.1% TFA/acetonitrile (2:1) was mixed with 1 ul of the analyte solution (5-10 pmol ul -1 ). For MALDI-MS 1 ul of this mixture was spotted on a stainless steel probe tip and dried at room temperature. The mass spectra were calibrated with angiotensin II (human), insulin (bovine), and cytochrome c (horse heart) as external standards. The mass accuracy was in the range of 0.05%. 2.6. Protein Sequencing Amino acid sequences were determined using the protein sequencer 476A (Applied Biosystems) according to the instructions of the manufacturer. 3. RESULTS 3.1. Proteolysis by Subtilisin RNase A is known to be degraded by subtilisin even in the native state, where the Ala20-Ser21 peptide

Arnold, Schierhorn, and Ulbrich-Hofmann

400

Fig. 1. SDS-PAGE of proteolytic degradation of RNase A (a) and RNase B (b) by subtilisin at 25°C. RNases were incubated in the presence of subtilisin at a 500:1 ratio (by mass) for 0, 0.5, 1, 2, 3, and 5.5 hr (lanes 1-6). (c) The reference proteins (see Materials and Methods).

and RNase B migrate in the SDS-PAGE as proteins of 15.4 and 18.0 kDa, respectively, the molecular masses of the fragments can be estimated to be 11.5 and 12.8 kDa. These values correspond to the molecular mass of the Ser21-Val124 fragment without (S-proteinA) and with the carbohydrate moiety (S-proteinB). As seen in Fig. 1, the S-proteins were further degraded as time of proteolysis proceeded, yielding small fragments which are not detectable in the SDS-PAGE. The estimation of the molecular masses of the large fragments was confirmed by mass spectrometry (Fig. 2). In the m/z range from 8,000 to 16,000 in both cases only signals for the intact protein (RNase A, 13,683 Da, or RNase B, 14,899 Da) and one large fragment (SproteinA, 11,536 Da, or S-proteinB, 12,752 Da) were detected. In the case of RNase B the signal at 12,752 shows clearly a series of additional peaks differing by 162 Da, which corresponds to the mannose units of the five RNase B types. The signals in the m/z range from 5,000 to 8,000 represent the doubly charged RNases and S-proteins, labeled by the mannose peak series in the case of RNase B. No other large fragments were found. The signals at 2,169 indicate the complementary fragment Lys1-Ala20 (S-peptideA and S-peptideB). Fig. 2. MALDI mass spectra of RNase A (A) and RNase B (B) after incubation in the presence of subtilisin at a 50:1 ratio (by mass) at 25°C for 15 min.

bond is cleaved to give the Lys1-Ala20 S-peptide and the Ser21-Val124 S-protein (Richards and Vithayathil, 1959). In Fig. 1, this degradation is compared with that of RNase B. After treatment of RNase A and RNase B with subtilisin at 25°C, pH 8.0, SDS-PAGE reveals fragment bands of about 13 kDa for RNase A and 15 kDa for RNase B. Under consideration that intact RNase A

3.2. Proteolysis by Thermolysin The susceptibility of RNase A toward proteolytic attack by thermolysin has been analyzed recently (Arnold et al, 1996). The enzyme proved to be resistant to proteolysis in the native state, but was degraded at higher temperatures. Asn34-Leu35 and Thr45-Phe46 were identified to be primary cleavage sites. Starting from these results, the proteolytic degradation of RNase B by thermolysin was analyzed by SDS-PAGE and mass spectrometry after RP-HPLC. Like RNase A, na-

Proteolytic Cleavage Sites in Ribonuclease B

401

Fig. 3. SDS-PAGE of the proteolytic degradation of RNase A (a) and RNase B (b) by thermolysin at 65°C. RNases were incubated in the presence of thermolysin at a 200:1 ratio (by mass) for (a) 0, 15, 45, 90, and 180 sec (lanes 1-5) and (b) 0, 0.75, 2, 5, and 8 min (lanes 1-5). (c) The reference proteins (see Materials and Methods).

Fig. 4. RP-HPLC of RNase B after proteolysis by thermolysin. RNase B was incubated in the presence of thermolysin at a 200:1 ratio (by mass) at 65°C for 1.5 min. Chromatographic separation was carried out as described in Materials and Methods. The gradient of the mobile phase was obtained by increasing concentrations of solvent B: 0-5 min, 3%; 5-10 min, 3-5%; 10-49 min, 5-22%; 49-74 min, 22-32%; 74-90 min, 32-45%.

tive RNase B also was resistant toward proteolytic attack by thermolysin, as proved by densitometric evaluation of Coomassie-stained SDS-PAGE gels (not shown). With increasing temperature above 45°C, the band of intact RNase B as well as of intact RNase A disappeared with progressing time of incubation in the presence of thermolysin. Fragments, however, became visible only at temperatures above 55°C. Figure 3 compares the time course of the proteolytic degradation of RNase A and RNase B by thermolysin at 65°C by SDS-PAGE. The band of intact RNase B disappears more slowly than that of RNase A, but the band patterns are qualitatively similar, with one exception: the band of the largest fragment in the RNase A degradation seems to be absent in RNase B degradation. According to Arnold et al. (1996), the

largest fragments of RNase A in Fig. 3a represent the fragments Leu35-Val124 (I), Phe46-Va1l24 (II), Lsy1lThr45 (III) and Lys1-Asn34 (IV), forming two complementary pairs: I + IV and II + III. The identification of the RNase B fragments was gained by MALDI-MS. A sample of RNase B was incubated in the presence of thermolysin at a 200:1 ratio (by mass) at 65°C for 1.5 min and separated by RPHPLC after reducing and carbamidomethylating the cysteine residues (Fig. 4). More than 40 fractions of this digestion were analyzed by MALDI-MS. The high accuracy of the MALDI-MS measurements, the limited number of large fragments being possible to occur, and the mannose signal series labeling the Asn34-containing fragments enabled the unambiguous assignment of all molecular masses to the RNase sequence. As a result, two pairs of complementary N- and C-terminal fragments (peak I + IV and peak II + III in Fig. 4) were identified (Table I). All the other fractions represented products of the further degradation of the primary fragments or intact RNase B (peak V).

3.3. Proteolysis by Trypsin The resistance of native RNase A toward proteolytic attack by trypsin was shown previously (Ooi et al., 1963; Klee, 1967; Arnold et al, 1996). In our experiments, also RNase B proved to be resistant toward proteolytic attack by trypsin below 45°C at pH 8.0 as studied by SDS-PAGE. At temperatures above 45°C, degradation of RNase A and RNase B started as detected by densitometric evaluation of Coomassie-stained SDSPAGE gels (not shown). However, fragments could be detected only from 55°C (RNase A) and 60°C (RNase B). Figure 5 compares the proteolytic degradation of


402 Table I. Molecular Masses of the Complementary RNase B Fragments Obtained by Limited Proteolysis by Thermolysina RNase B fragment

Peak

I II III IV

Molecular massb (Da)

Mannose seriesc

Sequence

Molecular massd (Da)

5,071e 6,386" 8,996 10,310

+ + -

1-34 1-45 46-124 35-124

5,071 6,385 8,996 10,309

Fractions of RP-HPLC separation (Fig. 4) were analyzed by MALDIMS as described in Materials and Methods. Peak numbers correspond to those in Fig. 4. * Molecular mass estimated by MALDI-MS. " Presence of signal series in the MALDI mass spectra for RNase B types possessing 5-9 mannose units. d Molecular mass of the Man5 type calculated from the amino acid sequence (Smyth et al, 1963). e Lowest molecular mass in signal series. a

RNase A and RNase B by trypsin at 65°C and reveals distinct differences in the band patterns. Band I in Fig. 5a, with an estimated molecular mass of 10 kDa, represents the fragments Ser32-Val124 and Asn34-Va1l24 of RNase A (Arnold et al, 1996). In the case of RNase B, these fragments would possess the carbohydrate moiety, increasing the molecular mass to about 13 kDa, so that band I in Fig. 5b might represent the corresponding fragments of RNase B. This assumption was confirmed by the results of the MALDI-MS measurements described subsequently. The fragment band II in Fig. 5a represents the fragment Cys40-Vall24 (Arnold et al., 1996). Presumably, band II in Fig. 5b shows the same fragment of RNase B. In order to localize primary cleavage sites, trypsintreated RNase B was separated by RP-HPLC (not shown) and analyzed by MALDI-MS. While two primary cleavage sites yielding the complementary fragments Lys1-Lys31 + Ser32-Va124andLysl-Arg33 + Asn34-Vall24 were identified for RNase A (Arnold et al., 1996), in case of RNase B only one pair of complementary N- and C-terminal fragments (Lysl-Lys31 + Ser32-Vall24) could be identified by MALDI-MS (Table II). Another large fragment with a molecular mass of 11,645 Da possessing the mannose-peak series (Asn34-Vall24, Table II) was detected, but no complementary fragment could be found. Although the estimation of the molecular masses of the tryptic fragments allows an unambiguous assignment to the RNase B sequence, N-terminal sequences of the two larger fragments were determined, which confirmed the results of MALDI-MS analysis (Table II).

3.4. Tryptic Degradation of CAM-RNase A and CAM-RNase B Thermally unfolded RNase A is reported to contain still a high degree of secondary and possibly tertiary structures (Sosnick and Trewhella, 1992). In order to evaluate proteolytic degradation of both RNases with minimum remaining structure, disulfide bonds were reduced and the resulting thiol groups were carbamidomethylated. Under these conditions, a-helix and (3-sheet content decreases by about 50% each, as found by Takeda et al. (1988) using CD measurements and Wilson et al. (1996) using dynamic light scattering measurements. In contrast to the RNases with intact disulfide bridges, both the reduced proteins were degraded by trypsin extremely fast even at 25°C. In order to identify primary cleavage sites, samples of RNase A and RNase B were incubated in the presence of trypsin at a 5000:1 ratio (by mass) at 25°C for 1 min and analyzed by MALDI-MS (Table III). In both cases the peptide bonds Lys7-Phe8 and Arg10-Gln11 could be identified as primary cleavage sites. Additionally, for RNase A the known fragments Lysl-Lys31 and Asn34-Vall24 were identified, whereas for RNase B only Lysl-Lys31 was found. For both RNases the fragment Cys40-Vall24 accumulated to a larger extent. Hence, Arg39-Cys40 was expected to be a primary cleavage site. However, the complementary N-terminal fragment Lysl-Arg39 was absent, whereas fragments with either N-terminal Ser32 or Asn34 and Cterminal Arg39 were found. Therefore, Cys40-Vall24 has to be regarded as a secondary fragment. The fragment Ser32-Arg39 was found only in the CAM-RNase B degradation. Obviously, this fragment is accumulated because the carbohydrate moiety delays the degradation by shielding the Arg33-Asn34 peptide bond, whereas in CAM-RNase A this peptide bond is not protected, which results in a fast degradation of Ser32-Arg39.

4. DISCUSSION For characterizing the proteolytic susceptibility of proteins, two main strategies can be applied: first, the determination of the rate of proteolysis by measuring the decrease of intact protein molecules, yielding information on the global proteolytic resistance of the protein, and second, the determination of the primary cleavage sites, i.e., cleavages in the intact protein molecule, resulting in conclusions on local accessibilities of different structural regions in the molecule. While the results of Arnold and Ulbrich-Hofmann (1997) contained information on the proteolytic degradation rates of intact RN-

Proteolytic Cleavage Sites in Ribonuclease B

403

Fig. 5. SDS-PAGE of the proteolytic degradation of RNase A (a) and RNase B (b) by trypsin at 65°C. RNases were incubated in the presence of trypsin at a 500:1 ratio (by mass) for (a) 0, 0.5, 1, 2.5, and 5 min (lanes 1-5) and (b) 0, 1, 3, 8, and 15 min (lanes 1—5). Because of the progressive inactivation of trypsin, trypsin addition was repeated after 8 min. (c) The reference proteins (see Materials and Methods).

Table II. Molecular Masses and N-Terminal Sequences of the RNase B Fragments Obtained by Limited Proteolysis by Trypsina RNase B fragment

Peak

I II III

Molecular mass* (Da)

Mannose seriesc

N-terminal sequence

Assigned RNase B sequence

3,498 11,888e 11,645e

+ +

Not determined Ser-Arg-Xaaf-Leu-Thr-Lys Xaaf-Leu-Thr-Lys-Asp-Arg

32-37 34-39

—

Sequence


1-31 32-124 34-124

3,498 11,883 11,640

RNase B was incubated in the presence of trypsin at a 500: 1 ratio (by mass) for 8 min at 65°C. The fragments were separated by RP-HPLC and analyzed by MALDI-MS and N-terminal protein sequencing as described in Materials and Methods. * Molecular mass estimated by MALDI-MS. c Presence of signal series in the MALDI mass spectra for RNase B types possessing 5-9 mannose units. d Molecular mass of the Man, type calculated from the amino acid sequence (Smyth et al., 1963). * Lowest molecular mass in signal series. fThe nonidentified amino acid residue Xaa is assumed to be glycosylated Asn34. a

ase A and RNase B, in this paper the primary cleavage sites in the RNase B molecule for subtilisin, mermolysin, and trypsin have been determined and compared with those in RNase A (Arnold et al, 1996). The results allow conclusions on the influence of the glycosylation on the course of proteolytic degradation and on the interplay of conformational unfolding and proteolytic susceptibility. In our studies subtilisin, thermolysin, and trypsin were selected as proteases since they markedly differ in their substrate specificities. Subtilisin is very unspecific and is able to cleave nearly all peptide bonds, with a preference for bonds at hydrophobic amino acid residues (Bond, 1990), while thermolysin selects peptide bonds at the N-terminal side of Leu, Ile, Phe, Val, Met, and Ala for cleavage (Aitken et al., 1989). Trypsin as the

most specific protease of these three enzymes cleaves only peptide bonds at the carboxylic side of the basic amino acids Lys and Arg (Aitken et al., 1989). In addition to these amino acid specificities, the positions of primary cleavage sites depend on the steric accessibility of the corresponding peptide bonds and the conformation of the amino acid residues at the cleavage site and in their vicinity (Hubbard et al, 1994). Both RNase A and RNase B are attacked by subtilisin even under native conditions (25°C, pH 8.0), but RNase B was degraded more slowly than RNase A (Fig. 1). The results of MALDI-MS (Fig. 2) show that RNase B is cleaved between Ala20 and Ser21, which is the same site as in RNase A (Richards and Vithayathil, 1959). Therefore, the decelerated rate of proteolysis is not connected with a change of the degradation pathway,


404

Table III. Main Fragments of CAM-KNase A and CAM-RNase B after Limited Proteolysis by Trypsin RNase A

RNase B

Assigned fragment Molecular mass a (Da)

718 13,453 1,151 13,018 3,499 10,421

746 9,695

Sequence Lys1-Lys7 Phe8-Val124 Lys1-Arg10 Glnll-Val124 Lys1-Lys31 Asn34-Val124 Asn34-Arg39 Cys40-Vall24

Assigned fragment

Molecular massb (Da) Molecular massa (Da) Mannose seriesc

Sequence


718

718

13,447 1,150 13,014 3,498 10,424

14,662e 1,151 14,234e 3,499

+ + -

Lys1-Lys7 Phe8-Vall24 Lys1-Arg10 Gln11-Vall24 Lysl-Lys31

14,663 1,150 14,230 3,498

2,206" 1,964e 9,697

+ + -

Ser32-Arg39 Asn34-Arg39 Cys40-Val124

2,205 1,962 9,695

746 9,695

718

" Molecular mass estimate by MALDI-MS. * Molecular mass calculated from the amino acid sequence (Smyth et al, 1963). c Presence of signal series in the MALDI mass spectra for RNase B types possessing 5-9 mannose units. dMolecular mass of the Man5 type calculated from the amino acid sequence (Smyth et al,, 1963). " Lowest molecular mass in signal series.

but it is presumably caused by a decreased flexibility of the loop region around Ala20. As derived from recent results (Arnold et al., 1996; Arnold and Ulbrich-Hofmann, 1997), thermolysin and trypsin require conformational unfolding of RNase A or RNase B before they can come into action. In thermally unfolded RNase A, two primary cleavage sites were found for thermolysin (Asn34-Leu35 and Thr45-Phe46) and two for trypsin (Lys31-Ser32 and Arg33-Asn34). Interestingly, these cleavage sites are immediately at or in close vicinity to the carbohydrate attachment site in RNase B (Asn34), so that dramatic changes in the proteolysis of RNase B with respect to the velocity of degradation and the position of primary cleavage sites were expected. In fact, however, no basically different fragment spectra were obtained in the proteolysis of thermally unfolded RNase B by thermolysin or trypsin even if they reflected an evident influence of the carbohydrate moiety on the accessibility of the cleavage sites. Thus, also in the degradation of RNase B with thermolysin, the two pairs of complementary fragments indicating the cleavage of the peptide bonds Asn34-Leu35 and Thr45Phe46 could be detected in mass spectra (Table I). The fragments proving the primary cleavage of Asn34Leu35, however, were found only in traces and were not detectable in the SDS-PAGE (Fig. 3), whereas no striking differences between RNase A and RNase B were observed with respect to the cleavage of Thr45-Phe46. Also in the proteolysis of RNase B by trypsin, no new primary cleavage sites beyond the carbohydrate-containing region were found. Unambiguously, again the peptide bond Lys31-Ser32 was identified as primary

cleavage site (Table II). The cleavage of Arg33-Asn34 was detected as well, but only one of the complementary fragments (Asn34-Val124) was found. Hence, it cannot be decided whether this cleavage is a primary one or the fragment results from a secondary cleavage in the fragment Ser32-Val124. The comparison of the progress of proteolysis, however, reveals considerable differences between RNase A and RNase B, which are stronger for trypsin than for thermolysin. Obviously, the degradation of RNase B is much more hindered for trypsin, where both the primary cleavage sites are in closest vicinity of the carbohydrate moiety (at Asn34), than for thermolysin, where one of the primary cleavage sites (Thr45Phe46) is more remote. Surprisingly, despite remarkable shielding effects of the preferred cleavage sites for thermolysin and trypsin by the voluminous carbohydrate chain, the RNase B molecule does not become cleavable at any site (from 33 potential cleavage sites for thermolysin and 14 for trypsin) in another region of the molecule. These findings suggest that both the thermally unfolded RNases still contain a rather compact structure which shields large regions of the sequence from proteolytic attack. This interpretation was confirmed by the experiments with reduced RNases. After breaking the four disulfide bridges in both RNases, trypsin as the most specific protease degraded RNase A as well as RNase B very fast and under occurrence of two new primary cleavage sites at the peptide bonds Lys7-Phe8 and Arg10-Gln11 (Table III). This result proves that protein conformation and not substrate specificity is responsible for the high selectivity of the proteases in the nonreduced RNases.

Proteolytic Cleavage Sites in Ribonuclease B Since trypsin and thermolysin cleave the nonreduced RNases only at denaturing temperatures, the results of this paper also allow comparative conclusions on the thermal unfolding of RNase A and RNase B. In RNase A, the structural region around Asn34 with the neighboring sequence around Thr45, which is first attacked by these proteases, has been assumed to become most exposed in thermal unfolding (Arnold et al, 1996). The present results suggest that this unfolding region is not changed in RNase B, although the carbohydrate moiety increases both thermodynamic and kinetic stability (Puett, 1973; Rudd et al., 1994; Arnold and UlbrichHofmann, 1997) and the modification site is situated immediately in this region. Obviously, the voluminous carbohydrate moiety decreases local movements of flexible regions of the molecule without altering the unfolding pathway or the accessibility of the unfolded peptide chain significantly.

ACKNOWLEDGMENTS We thank Dr K.P. Riicknagel, Max-Planck-Forschungsstelle "Enzymologie der Proteinfaltung," Halle, Germany, for performing N-terminal protein sequencing. This work was supported by the Deutsche Forschungsgemeinschaft, Bonn, Germany (Ul 130/2-3).

REFERENCES Aitken, A., Geisow, M.J., Findley, J. B. C., Holmes, C., and Yarwood, A. (1989). In Protein Sequencing. A Practical Approach (Findley, J. B. C., and Geisow, M. J., eds.), IRL Press, Oxford, pp. 43-68. Arnold, U., and Ulbrich-Hofmann, R. (1997). Biochemistry 36, 21662172. Arnold, U., Rucknagel, K. P., Schierhorn, A., and Ulbrich-Hofmann, R. (1996). Eur. J. Biochem. 237, 862-869.

405 Berman, E., Walters, D. E., and Alerhand, A. (1981). J. Biol. Chem. 256, 3853-3857. Birkeland, A. J., and Christensen, T. B, (1975). J. Carbohydrates Nucleosides Nucleotides 2, 83-90. Blum, H., Beier, H., and Gross, J. H. (1987). Electrophoresis 8, 9399. Bond, J. S. (1990). In Proteolytic Enzymes. A Practical Approach (Beynon, R. J., and Bond, J. S., eds.), IRL Press, Oxford, pp. 239-240. Hubbard, S. J., Eisenmenger, F., and Thornton, J. M. (1994). Protein Sci. 3, 757-768. Jaenicke, R. (1991). Biochemistry 30, 3147-3161. Joao, H. C., and Dwek, R. A. (1993). Eur. J. Biochem. 218, 239-244. Joao, H. C., Scragg, I. G., and Dwek, R. A. (1992). FEBS Lett. 307, 343-346. Klee, W. A. (1967). Biochemistry 6, 3736-3742. Lang, K.., and Schmid, F. X. (1986). Eur. J. Biochem. 159, 275-281. Lin, L.-N., and Brandts, J. F. (1983). Biochemistry 22, 559-563. Lin, L.-N., and Brandts, J. F. (1984). Biochemistry 23, 5713-5723. Lis, H., and Sharon, A. (1993). Eur. J. Biochem. 218, 1-27. Mer, G., Hietter, H., and Lefevre, J.-F. (1996). Nature Struct. Biol. 3, 45-53. Ooi, T., Rupley, J. A., and Scheraga, H. A. (1963). Biochemistry 2, 432-437. Plummer, T. H., Tarentino, A., and Maley, F. (1968). J. Biol. Chem. 243, 5158-5164. Puett, D. (1973). J. Biol. Chem. 248, 3566-3572. Rademacher, T. W., Parekh, R. B., and Dwek, R. A. (1988). Annu. Rev. Biochem. 57, 785-838. Richards, F. M., and Vithayathil, P. J. (1959). J. Biol. Chem. 234, 1459-1465. Rudd, P. M., Joao, H. C., Coghill, E., Fiten, P., Saunders, M. R., Opdenakker, G., and Dwek, R. A. (1994). Biochemistry 33, 1722. Rudd, P. M., Woods, R. J., Wormald, M. R., Opdenakker, G., Downing, A. K., Campbell, I. D., and Dwek, R. A. (1995). Biochim. Biophys. Acta 1248, 1-10. Rupley, J. A., and Scheraga, H. A. (1963). Biochemistry 2, 421^31. Schagger, H., and von Jagow, G. (1987). Anal. Biochem. 166, 368379. Smyth, D. G.; Stein, W. H., and Moore, S. (1963). J. Biol. Chem. 238, 227-234. Sosnick, T. R., and Trewhella, J. (1992). Biochemistry 31, 8329-8335. Takeda, K., Sasa, K., Nagao, M., and Batra, P. P. (1988). Biochim. Biophys. Acta 957, 340-344. Williams, R. L., Greene, S. M., and McPherson, A. (1987). J. Biol. Chem. 262, 16020-16031. Wilson, G., Hecht, L., and Baron, L. D. (1996). Biochemistry 35, 12518-12525. Yamaguchi, H., and Uchida, M. (1996). J. Biochem. 120, 474-477.