The Renaturation of Reduced Polyalanyl ... - Wiley Online Library

Eur. J . Biochrm. 59,433-440 (1975)

The Renaturation of Reduced Polyalanyl-chymotrypsinogen and Chymotrypsinogen Gilbert ORSINI. C e d e SKRZYNIA, and Michel E. GOLDBERG Service de Biochimie Cellulaire. Institut Pasteur. Paris (Received Junc 9. 1975)

Chymotrypsinogen has been successfully renatured in solution, after reduction of its 5 disulfide bonds in 6 M guanidine-HC1. This has been made possible by the study of the renaturation of a model derivative, polyalanyl-chymotrypsinogen. The reduced derivative is shown to refold and reoxidize spontaneously, with a 30-40% yield, into molecules which are monomeric and fully susceptible to activation by trypsin. Chymotrypsinogen can also be renatured but only in the presence of reagents allowing disulfide interchange and of moderate concentrations of guanidine-HC1 or urea. These results illustrate how the kinetic trapping of incorrectly folded molecules by wrong S-S bonds and aggregation can be overcome, thus allowing the correct refolding of the protein.

Following the pioneering work of Sela et al. on ribonuclease [l], several disulfide-cross-linked proteins have been successfully renatured after reduction of their S-S bonds in denaturing conditions (see [2] for review). These results have decisively substantiated the view that proteins fold so as to reach a low conformational free energy state among the kinetically accessible structures. The dynamic aspects of protein folding are, however, still debated and increasing evidence is accumulating against the equilibrium twostate approximation (see [3] for a recent review). During the renaturation of S-S-reduced proteins, the situation is further complicated by the covalent aspect of folding, the formation of disulfide bridges. In addition, it is a common observation in protein chemistry that denatured proteins in general and reduced-denatured S-S proteins particularly tend to be insoluble in aqueous buffers; in the latter case aggregation of disordered structures and formation of intermolecular S-S bonds may limit or prevent the renaturation of such proteins by a straightforward procedure. Reduced chymotrypsinogen provides a good example of this behaviour. Even with S-S bonds intact, many denatured forms of this zymogen are markedly ; with the 5 S-S bonds prone to aggregation [4,5] reduced, even at very low protein concentration, it is completely insoluble in aqueous media in the pH range required for the reoxidation of -SH groups [6]. A similar situation is found in attempts to reactivate -

E n x r n e ~Chymotrypsin . (EC 3.4.21.1); trypsin (EC 3.4.21.4).

reduced trypsin; this enzyme has, however, been renatured through coupling to carboxymethyl-cellulose [7]. An analogous procedure, involving immobilization of native chymotrypsinogen on a succinyl glass matrix, allowed renaturation of the reduced zymogen [8]. These results indicate the necessity of preventing protein-protein interactions, but the method precludes direct observation of the molecular species in the course of the refolding. In addition, little is known of the constraints exerted by an insoluble support upon a polypeptide chain to which it is covalently bound. For these reasons, a way other than immobilization of the protein has been looked for to avoid aggregation: acylation of proteins' lysyl amino groups by N-carboxyanhydrides of D and L-amino acids leads to the formation of poly(DL-amino acid)-proteins [9] ; this chemical modification is known to increase the solubility of proteins. Indeed, polyalanylation of trypsin has allowed the renaturation of this enzyme after reduction of its disulfide bonds [lo]. We therefore have investigated the possibility of renaturing reduced polyalanyl-chymotrypsinogen with the hope that the results might lead to a better understanding of the renaturation of the unmodified protein. The results reported in the present paper show that reduced polyalanyl-chymotrypsinogen can be successfully renatured. It is further shown that, starting from the conditions found for the renaturation of polyalanyl-chymotrypsinogen, a procedure could be set up for renaturing the reduced unmodified zymogen with a substantial yield. Among the various parameters

434

involved in the renaturation process it is shown that the presence of mixtures of reduced and oxidized thiol compounds is of primary importance. The role of disulfide interchange on the renaturation of reduced chymotrypsinogen is discussed and the process of reoxidation of reduced chymotrypsinogen is compared to that of lysozyme, ribonuclease and of the bovine pancreatic trypsin inhibitor [I 1 - 131. MATERIALS AND METHODS

Renaturation of Reduced Chymotrypsinogen

Active-Site Titration

The active-site concentration of chymotrypsin was determined by measuring at 400 nm the 'burst' of p-nitrophenol upon hydrolysis of p-nitrophenylacetate at pH 8.5 and 25°C [17]. Ultracentrijiigation

Sedimentation velocity measurements of native and renatured proteins were performed with an MSE Centriscan 75 analytical ultracentrifuge equipped with an ultraviolet-absorption scanning system.

Materials

Bovine chymotrypsinogen A, 5 x crystallized, lot 1 EA and trypsin were purchased from Worthington. The commercial sources of the following products were : N-acetyl-L-tyrosine ethyl ester, p-nitrophenylacetate and diisopropylphosphorofluoridate from Koch-Light ; N-benzoyl-L-arginine ethyl ester and DL-alanine from Cyclo ; reduced dithiothreitol from Sigma ; oxidized dithiothreitol from Calbiochem; reduced glutathione from Fluka ; oxidized glutathione from Boehringer ; 2-mercaptoethanol and urea from Merck; guanidine-HC1 from Carlo Erba ; phosgene from Seppic-Labo. All other chemicals were from Prolabo or Merck.

Synthesis of N - Carboxy-DL-alanine Anhydride

N-Carboxy-m-alanine anhydride was synthesized according to the general method of Katchalski and Berger [ 181 except for the recrystallization of the anhydride, which was performed in cold pentane. Elemental composition: C = 42.16% (calcd = 41.74); H = 4.63 (calcd = 4.35); N = 12.12 (calcd = 12.17). Melting point = 52 "C (reported 46 "C [19]; 60 "C[20]). The dried crystals were stored at - 40°C in vamo in the presence of phosphorus pentoxide and could be used in the polymerisation reaction during more than twelve months. Polyalanylation of Chymotrypsinogen

Protein Concentration

Protein concentrations of chymotrypsinogen and polyalanyl-chymotrypsinogen were determined spectrophotometrically at 282 nm on a Cary 17 or a Zeiss PMQ I1 spectrophotometer, using a molar absorption coefficient E 5 . lo4 M-' cm-' [14]. For polyalanylchymotrypsinogen it was checked that the polyalanyl side-chains had no effect on this coefficient, by absorbance measurements coupled with amino acid analysis used as a protein concentration determination. Expressed in mg/ml, concentrations of polyalanylchymotrypsinogen refer to mg of chymotrypsinogen, not considering the polyalanyl additional weight. Activation of the Zymogen. Activity Assay

Chymotrypsinogen and the polyalanylated derivative were activated, at concentrations ranging from 20 pg/ml to 200 pg/ml, with trypsin-to-zymogen ratios of l j 5 to 1/10, at pH 8 , 2 5 "C.During these activations, chymotrypsin activity, toward N-acetyl-L-tyrosine ethyl ester, was measured according to Schwert et al. [15], until a constant specific activity level was reached. 6-Chymotrypsin hydrolyses this substrate with a rate constant (k) of about 230 s-' [16]; however, after activation the commercial preparation of chymotrypsinogen used in this work showed a hydrolysis rate constant of 180-200 s-'.

Prior to any chemical modification or denaturation procedure care was taken to treat chymotrypsinogen by diisopropylphosphorofluoridate in order to block the 0.4 chymotrypsin present in the commercial preparation. Following this treatment, this residual chymotrypsitic activity was abolished and polyacrylamide gel electrophoresis in dodecylsulfate of denatured and reduced chymotrypsinogen showed one single band migrating like the unreduced zymogen instead of the many proteolysed fragments found when this treatment was omitted (see [21]). Chymotrypsinogen was polyalanylated at a protein concentration of 5 mg/ml by polymerisation of N-carboxy-DL-alanine anhydride according to Epstein et al. [22]. The protein derivative was then dialysed extensively at 4 "C and precipitated by ammonium sulfate. The precipitate was redissolved in buffer, dialysed and passed through a Sephadex G50 column. All these steps were carried out in a 50 mM sodium phosphate buffer, pH 7. The amino acid composition of the fractions containing polyalanyl-chymotrypsinogen was determined, to measure the degree of the polyalanylation and to check the homogeneity of the labelling. Denaturation and Reduction of the Proteins

Chymotrypsinogen and the polyalanylated derivative were denatured at protein concentrations of 2-

435

G. Orsini, C. Skrzynia. and M. E. Goldberg

6 mg/ml in 8 M urea or 6 M guanidine-HCl in 200 mM Tris-HC1 buffer at pH 8.5. Reduction was carried out in these denaturants either with 100-400 mM 2-mercaptoethanol or with 10- 100 mM reduced dithiothreitol in stoppered tubes during 4 h at room temperature. After removal of the reducing agent by gel filtration at pH 2 - 4, the protein - SH groups were titrated at pH 8 with 5,5’-dithio-bis(2-nitrobenzoic acid) [23] in the presence of 6 M guanidine-HC1 or of 2 7; dodecylsulfate (w/v). Kinetics of reduction of chymotrypsinogen in 6 M guanidine-HC1 with reduced dithiothreitol could be followed by complexing the reducing agent with sodium arsenite [24] : 80 % of the protein - SH groups appear within the 30 min following the addition of dithiothreitol. Under most of the conditions used, 9 - 10 - SH groups/mol protein were titrated at the end of the reduction. Renaturation of the Reduced Proteins Different conditions ensuring renaturation of reduced chymotrypsinogen and polyalanyl-chymotrypsinogen are described in the Results. For both proteins the recovery of chymotrypsin activity upon activation by trypsin has been used in this study as a convenient measure of the extent of renaturation, assuming that the protein carries potential activity only in its native conformation. Renaturation yields are expressed as the percentage potential specific activity of the reoxidized zymogens relative to that of the control. These control renaturations were effected on zymogens denatured with S-S bonds intact and were routinely carried out in parallel with the renaturation of the reduced proteins.

RESULTS

The average degree of polyalanylation of the different preparations of polyalanylchymotrypsinogen obtained varied from 90 to 140 alanyl residues/mol of protein. Gel filtration of a given preparation showed a slight heterogeneity of the labelling (Fig. 1). However, in sedimentation velocity measurements, polyalanyl-chymotrypsinogen (140 alanyl residues/mol) appeared monomeric and showed a single boundary with a sedimentation coefficient szo,w z 3 S, which compares favourably with szo,wz 2.6 S of chymotrypsinogen [25]. The different preparations of polyalanyl-chymotrypsinogen used in this work retained 80- 90 % potential chymotryptic activity. In drastic contrast with reduced chymotrypsinogen, the reduced derivative was freely soluble at mildly alkaline pH in the absence of denaturant even at a protein concentration of 1 mg/ml. The above features proved polyalanyl-chymotrypsinogen to be a convenient model

Fraction number

Fig. 1. Heterogeneity of ilie labelling o f p o l ~ a l a n v l - c h ~ n l o l r ~ ~ ~ i t l o g e n . The polyalanylated derivative, prepared as described in Methods, was passed through a Sephadex G-50 column at pH 7. The proteincontaining fractions were andlysed for amino acid compositions. (w) Absorbance of the fractions at 282 nm. (+-+) Number of alanyl residues/mol of polyalanylchymotrypsinogen in excess of the 22 existing in the unmodified zymogen [25]

derivative with which to investigate the conditions of the renaturation of non-modified chymotrypsinogen. Reactivation of’Reduced Pol-valanyl-Cliymotrypsinogen The excess reagents (the denaturant and the reducing agent) could be removed from the reduced protein by two main procedure: gel filtration or dialysis. Both were used to investigate the conditions required to obtain the optimal renaturation yields. When gel filtration was used the renaturation protocol involved two steps: the reduced protein was first separated from guanidine-HC1 and 2-mercaptoethanol on a SephadexG-50 column and then incubated at alkaline pH to allow renaturation. The effect of the pH at which these steps were performed was investigated. Table 1 shows that, after gel filtration of the protein either at acid pH (where it remains essentially reduced) or at pH 8.5 (where it has lost half of its - SH groups), incubations at pH 8.5 result in poor renaturation yields. Gel filtration followed by incubation at pH 10 brings about a significant renaturation of the reduced zymogen. However, as the highly alkaline pH 10 could influence both the reactivity of the - SH groups and the conformation of the protein, no clear assessment of this favourable effect of high pH on the renaturation process was possible at this stage. Renaturation by dialysis was studied to compare its effectiveness with the results obtained by the gel filtration-incubation procedure : preliminary experiments were undertaken to investigate the effect of a slow escape of denaturant on the renaturation yield.

436


Table 1. Renaturation of reduced polyalanyl-chymotrypsinagenufter gel filrratian Polyalanyl-chymotrypsinogen was reduced at 2.5 mg/ml with 0.15 M 2-mercaptoethanol in 6 M guanidine-HCI for 4 h. The protein was then diluted in 6 M guanidine-HCI to 1 mg/ml before gel filtration at acid pH or to 0.4 mg/ml before gel filtration at alkaline pH. Removal of guanidine-HC1 and 2-mercaptoethanol was then performed by gel filtration at 4 ’ C on a Sephadex (3-50 column equilibrated and eluted with 50 mM buffers at the indicated pH values. The collected protein was incubated overnight at 4‘C at 150 Fg/ml in 0.1 M alkaline buffers and subsequently activated and assayed for activity pH of gel filtration

Protein -SH content after gcl filtration

pH of incubation

%

- SH/mol protcin

2-4 4 8.5

10

8.5-10 8.5-10 5.5 4

Renaturation yields

8.5

3-

10

31-37

8.5 10

25

6

3

pH of first dialysis

Fig. 2. Activahility of reduced-reoxidized pol~alan?,l-chymotr?7)sinoKen as n functiou of’ the p H of refolding dialysis. Reduced polyalanyl-chymotrypsinogen, prepared as described in Table 1, was diluted to 200 pg/ml into 6 M guanidine-HCI solutions prepared in 0.1 M buffers of various pH values and dialysed overnight at 4 ° C against two changes of 50mM b u l k at the same pH (abscissa scale). This was followed by a second dialysis at pH 8 for 20 h. All dialyses were performed at 4 ’ C. No free - SH could bc detected in the proteins after this sequence of dialysis throughout the pH range investigated. The proteins were then assayed for activity

Table 2. Renaruration of reduced polyuIun~~l-c~iymoIrypsinu~en bv dia1ysi.s at p H 8.5 Reduced polyalanyl-chymotrypsinogen as described in Tablc 1 was diluted to 200 pg/ml either in pH 8.5 buffcr or in 6 M guanidine-HCI, pH 8.5 (the residual 2-mercaptoethanol concentration was 12 mM). The proteins were then dialysed overnight at 4°C against 50 mM Tris-HCI buffer, pH 8.5. Active sites were titrated as described in Methods and corrected for the presence of the trypsin used for activation. n.d., not determined Denatured state

Dilution buffer

Activability after dialysis specific activity toward N-acetyl-L-tyrosine ethyl ester

fraction of active site titrated

S-’

With S-S bonds reduced With S-S bonds rcduccd With S-S bonds intact (control)

50 mM Tris-HCI 6 M guanidine-HCI in 0.1 M Tris-HCI 6 M guanidine-HC1 in 0.1 M Tris-HC1

The experiments reported in Table 2 indeed show that a rapid decrease of the guanidine-HCI concentration from 6 M to 0.5 M results in a renaturation yield lower than does a slow elimination of the denaturant by dialysis. This dialysis procedure was, therefore, used to study the dependence of the renaturation yield of reduced polyalanyl-chymotrypsinogen upon the pH at which the 6 M guanidine-HC1 and the residual 2-mercaptoethanol are eliminated. Fig.2 shows that refolding by dialysis in the pH 8-8.5 range gave renaturation yields close to the plateau values reached at higher pH. This result clearly contrasts with the very low yields reproducibly obtained by incubation of the reduced protein at p H 8.5 after gel separation at acid pH (see Table 1). In this latter case the lack of experimental reversibility was undoubtfully attributable to aggregation of the reduced protein, either at acid pH or at pH 8.5, where S-S bonds are likely to reform. Trying to avoid

25 30 (yield = 15- 19 7,) 55 - 60 (yield = 34- 37 160 (yield = lOO>J ~

x)

n.d. 0.25 0.80

this aggregation effect the following experiment was performed : polyalan yl-chymotrypsinogen was fully reduced by 0.1 M dithiothreitol in 6 M guanidine-HCI and separated from these reagents by gel filtration on a Sephadex G-50 column equilibrated with 10 mM HCl at 4 T . The protein was collected and total reduction of the S-S bonds was checked. It was then diluted to 300 pg/ml in aliquots of 0.1 M Tris-HC1 buffer, pH 8.5, containing various concentrations of guanidine-HC1 (from 0 - 3 M) or urea (from 0 - 4 M). These aliquots were immediately dialysed overnight at 4°C against 50 mM pH 8.5 Tris-HC1 buffer. The protein diluted in moderate concentrations of denaturant, approximately 3 M urea and 2 M guanidineHC1, showed 25 - 30 % renaturation, whereas it was totally devoid of potential activity after dilution in the absence of the perturbant. These ‘uncooking’ experiments provide some understanding of the discrepancy between the results obtained at pH 8.5 reported in Tables 1 and 2. During

G. Orsini, C . Skrzynia. and M. E. Goldberg

437

which are likely to be in the native state, since upon activation by trypsin they exhibit one active site per molecule and full activity. It should be added that the distribution between the aggregated and monomeric species is not due to a heterogeneity in the degree of polyalanylation; the monomeric as well as the aggregated species, when submitted to a new cycle of reduction and reoxidation, yielded both renatured and aggregated polyalanylchymotrypsinogen in the same proportions as described above.

I

1 .c

m N

p"

0.5

1

+++

0

t d L

20

fi


'

25

30

35

40

45

Fraction number

Fig. 3 . Srp1iacie.u G- 100 cliwimtogrupliy of reduced-reoxidized p l y t r l n r ~ ~ ~ l - c h s r ~ r o t r ~ p s Reoxidizcd i r i o ~ ~ ~ n . polyalanyl-chymotrypsinogen was obtained by dialysis at pH 10 as described in Fig.2. The protein was then concentrated to 2 mg/ml by ultrafiltration, checked for activity and chromatographed. The Sephadex column (65 x 2 cm) was equilibrated, and the proteins eluted, at 4°C with 50 m M TrisHCI, 0.1 M KCI, pH 8.5, buffer. Arrows 1 and 2 indicate the elution volumes of /I-galactosidase and polyakdnyl-chymotrypsinogcn rcspcctively, both proteins in the native state. (M Ab) sorbance of the fractions at 282 nm. (+) Specific activity of the the retarded fractions, upon activation

dialysis at pH 8.5 the slowly decreasing denaturant concentration, be it initially 2- 3 M or 6 M (Table 2 and Fig. 2) interferes with the refolding of the protein, in that it prevents aggregation of incorrectly folded polypeptide chains, otherwise cross-linked by incorrect s-S bonds. Some molecular properties of the renaturedreoxidized species were characterized. Table 2 shows that the active-site titer of renatured and activated polyalanyl-chymotrypsinogen parallels its specific activity. In sedimentation velocity experiments, renatured-reoxidized polyalanyl-chymotrypsinogen sedimented with two distinguishable boundaries: the first corresponded to 30-40'j/, of the protein and sedimcnted with a szo,w z 3 S like the non-denatured derivative; the second showed considerable heterogeneity and was composed of fast-sedimenting species. These high-molecular-weight aggregates were separated by gel chromatography on a Sephadex (3-100 column as shown in Fig. 3. The retarded protein fraction emerged from the column with the same elution volume as non-denatured polyalanyl-chymotrypsinogen. This fraction could be activated into 90'j/, active polyalanyl-chymotrypsin, whereas the excluded material was completely devoid of potential activity. Hence, renaturation of reduced polyalanyl-chymotrypsinogen results in 30 - 40 % monomeric molecules,

Polyalanylation of chymotrypsinogen has been an experimental approach to study the renaturation of the reduced zymogen without the added polyalanyl side-chains; in this latter case the major difficulty to overcome was that in the absence of denaturant, at pH values suitable for reoxidation of the cysteines and even at very low protein concentration, intermolecular interactions leading to insolubility were massively predominant over the proper refolding of the polypeptide chain. However. in the absence of urea or guanidine-HCl the reduced protein is soluble in the acid pH range up to pH 6 . Reduced chymotrypsinogen was thus freed from guanidine-HC1 by gel filtration or dilution at acid pH and some of its properties were examined. At pH 4 and 4 "Cthe - SH titer of the reduced protein (10 - SH/mol) remained unchanged for several hours. Ultracentrifugation of fully reduced chymotrypsinogen at pH 4 and 4 "C showed that the protein at 0.2 mg/ml formed heavy aggregates. The accessibility to the solvent of the aromatic chromophores of the reduced and aggregated protein was investigated by difference spectroscopy in the ultraviolet region. from 250 nm to 310 nm. It was observed that the difference spectrum between reduced chymotrypsinogen in pH 4 buffer and reduced chymotrypsinogen in 8 M urea at the same pH was identical to the 8 M urea denaturation spectrum of this zymogen with S-S bonds intact reported by Chervenka [26]. This latter spectrum is obtained between the native and 8-M-urea-denatured zymogen. This observation shows that 'spectral' refolding of the reduced protein occurs in denaturant-free buffer. This indicates only that the aromatic chromophores of reduced and aggregated chymotrypsinogen are buried, as compared to their complete exposure to the solvent in 8 M urea, whether or not the protein S-S bond are reduced [27]. The formation of these incorrectly folded aggregates could be restricted by moderate concentrations of guanidine-HCI as described for the refolding of reduced polyalanyl-chymotrypsinogen : it was observed that when reduced chymotrypsinogen, after


438

gel filtration at pH 4, was brought to pH 8.5 in the presence of 2 M guanidine-HC1 and then free from the denaturant by dialysis at pH 8.5, it did not precipitate, Furthermore, the protein showed a low (2 but significant potential activity. The same result was observed when the pH 4 gel filtration was omitted, i.e. when the denatured-reduced protein was directly diluted in 2 M guanidine-HC1 at pH 8.5 and dialysed. This favourable effect of guanidine-HC1 made it possible to study the renaturation of the reduced zymogen. As air oxidation of reduced polyalanyl-chymotrypsinogen occurred very slowly even at pH 10, reoxidation of reduced chymotrypsinogen was investigated in the presence of several agents known to favour the reoxidation of -SH groups in proteins: copper ions [28,29]; mixtures of reduced and oxidized glutathione [l 1,121; oxidized dithiothreitol [131 and mixtures of reduced and oxidized dithiothreitol [13]. The effectiveness of these different systems on the renaturation process is summarized in Table 3 :whereas simple oxidation reagents such as copper ions and oxidized dithiothreitol have no effect, mixtures of reduced and oxidized thiols lead to substantial renaturation yields. The same was observed by keeping constant, in the dialysis buffer during the escape of guanidine-HC1, the concentrations of the following reagents : copper ions, oxidized dithiothreitol and a mixture of reduced and oxidized glutathione. Various ratios of reduced to oxidized glutathione were tested but no additional information was gained over the careful investigation reported by Saxena et al. [ll]. Varying the initial concentration of guanidine-HC1 from 2 M to 6 M, had no significant effect on the final yield of renatured chymotrypsinogen. The presence of a redox system during the escape of guanidine by dialysis thus ensured a 45-50% renaturation yield of the reduced zymogen. This effect is further demonstrated by the experiments described in Table 4: native chymotrypsinogen, the 6 M-guanidine-HC1-denatured protein with S-S bonds intact and the reduced protein in 6 M guanidine-HC1 were tested for disulfide bond interchanges in the presence and during the escape of 2 M guanidine-HC1. It can be seen in Table 4 that denatured chymotrypsinogen with intact S-S bonds behaves like the fully reduced zymogen once in contact with the redox system. On the contrary, the native zymogen is insensitive to S-S interchange (i.e. 100% activability). This result shows that denatured chymotrypsinogen is at least partially reduced in these conditions and therefore has to follow the same general renaturation route as the fully reduced protein: in 2 M guanidineHCl, and during the dialysis the correct refolding of reduced chymotrypsinogen proceeds through disulfide bond interchanges which require the presence of a reducing agent.

x)

Table 3. Effectiveness of rapid regeneration systems in the renaruration of reduced chymotrypsinogen Chymotrypsinogen was reduced with 10 mM or 100 mM reduced dithiothreitol in 6 M guanidine-HCI at pH 8.5. The resulting protein showed a titer of 9.9- 10.2 -SH/mol. The reduced protein was diluted to 20- 50 pg/ml in 2 M guanidine-HC1 solutions, pH 8.5, in the presence of the indicated concentrations of reagent and dialysed overnight against 20 mM Tris-HCI, 0.1 M KCI, pH 8.5 Renaturation yield

Reagent concentration

C U S O0.1 ~ - 10 PM Oxidized dithiothreitol 1- 10 mM Oxidized dithiothreitol 5 mM and reduced dithiothreitolO.5 mM Oxidized dithiothreitol 10 mM and reduced dithiothreitolO.5 mM Oxidized glutathione 0.5 mM and reduced glutathione 5 mM Oxidized glutathione 5 mM and reduced glutathione 50 mM

2 2 25 35 45 - 50

6

Table 4. Susceptibility of various forms of chymotrypsinogen to disurfide interchange in 2 M guanidine-HCI Three solutions of chymotrypsinogen at the same concentration, native, denatured in 6 M guanidine-HCI with disulfide bonds intact and reduced in 6 M guanidine-HCl, were diluted in 2 M guanidineHCI, pH 8.5, with or without 5 mM reduced glutathione + 0.5 mM oxidized glutathione, dialysed as described in Table 3 and tested for renaturation ~

Chymotrypsinogen before dilution

~

~~

Renaturation yield after dilution into 2M guanidineHCI

2 M guanidineHCI + reduced and oxidized glutathione

100 90- 100 0

100 50 45 - 50

-

Native Denatured with S-S bonds intact Reduced and denatured

DISCUSSION To our knowledge the present work is the first report of the renaturation of reduced and denatured chymotrypsinogen in solution. Although complete renaturation was not achieved, the significant return of 35-50% of the protein to the native state was obtained. The conditions required for the correct refolding of both the zymogen and the polyalanylated derivative share some common features and somehow explain the processes occurring in this renaturation. Indeed, two distinct phenomena interfere with the renaturation of these proteins : first, the formation of protein aggregates, which can be limited by moderate concentrations of urea or guanidine-HC1.The presence

G. Orsini, C. Skrzynia, and M. E. Goldberg

of such a ‘perturbant’ is an absolute requirement during the refolding of reduced chymotrypsinogen. The second effect is the formation of incorrect disulfide bonds followed by disulfide interchange. In the case of the polyalanylated derivative, this interchange very likely occurs spontaneously, through hydroxyl-ion catalysis [30]. On the contrary, renaturation of the unmodified zymogen requires the presence of significant levels of thiol reagents. This indicates that breakage of non-native S-S bonds prevails in this case (see Tables 3 and 4). The reason for this different behaviour of the two proteins lies in the ‘solubilizing’ effect of the polyalanyl side-chains : this chemical modification, by preventing proteinprotein interactions, presumably allows the protein to ‘search’ the correct pathways of folding without being trapped in abortive aggregates. Oxidized dithiothreitol alone or copper ions do not ensure higher renaturation yields than simple air oxidation. In this respect chymotrypsinogen differs from the other disulfide-containing proteins, for which similar investigations have been recently reported. Oxidized dithiothreitol is effective in the renaturation of reduced bovine pancreatic trypsin inhibitor [13] : this 58-residue protein refolds through the highly restricted formation of a single correct S-S bond at the early stages of its reoxidation [31]. Reduced lysozyme has been shown to refold, in the presence of copper ions, through a moderately restricted pathway of S-S bond reformation [29]; this was also observed in the presence of a disulfide interchange system [32]. Reduced ribonuclease renatures after the general reshuffling of incorrect S-S bonds at the early stages of the reoxidation [12] ; however, the initial observations of Sela et al. indicated that ribonuclease is able to reach its native reoxidized conformation in the absence of added thiols favouring disulfide interchange [l]. Reduced chymotrypsinogen thus requires for its renaturation more exacting conditions than the above-mentioned proteins : the correct pairing of S-S bonds occurs efficiently only in the presence of disulfide-reducing agents. Some structural features of the native and denatured zymogen may account for the difficulty to renature the reduced protein. The disulfide bonds greatly restrict the flexibility of denatured chymotrypsinogen: in 8 M urea or in 6 M guanidine-HC1, there is a more than two-fold increase of the intrinsic viscosity of the denatured zymogen upon reduction of the S-S bonds [30,33]. As a consequence, the zymogen with S-S bonds intact renatures easily after 8 M urea denaturation [34] and the same is shown in the present work to occur after 6 M guanidine-HC1 denaturation of the oxidized protein. The origin of the abortive aggregation, which interferes with the proper refolding of the reduced zymogen, may be understood by examining the con-

439

formation of the protein. The overall structure of native chymotrypsinogen is similar to that of a-chymotrypsin [35,36]. Like the enzyme, the zymogen is essentially composed of two folded units or domains closely packed together through an interface which runs across the molecule [37]. The 5 disulfide bonds are distributed within each domain and there is no such bond bridging the two domains. Each of these units is formed by twisted anti-parallel /?-pleated sheets and contains a core of non-polar buried residues [35,38]. Many, if not all of these iesidues, are found clustered in the close vicinity of the cysteine residues forming the S-S bonds. The limited a-helical portion of the protein is made of the 16 C-terminal residues, the side-chains of which are essentially in contact with the solvent. One may speculate that once formed, this a-helical C-terminus would thus provide a poor stabilizing contribution to the refolding of the rest of the molecule. This makes reasonable the concept that refolding of the molecule would have to occur through some formation of secondary structures within each of the domains prior to their final close packing, according to currently developped views on the folding of protein structures [39,40]. In this line, aggregation of the reduced protein in the absence of denaturant would be a consequence of the fact that the B-structures, which have to build up within each of the domains, involve mainly “non-local interactions”, between “residues that are far apart along the polypeptide chain” [41]. The correct non-local secondary interactions would be in unfavourable competition with incorrect interactions between elements belonging to different domains or to different polypeptide chains. Wrong disulfide bonds would only later cross-link these incorrect structures and lead to irreversibility. This tentative interpretation of the origin of the abortive aggregation of the reduced protein may account for the effects of guanidine-HC1 and thiol reagents during the renaturation of reduced chymotrypsinogen. In this respect the use of moderate concentrations of guanidine-HC1 (or urea) and of thiol reagents favouring disulfide-bond interchange may prove of general applicability for the renaturation of proteins hitherto reluctant to undergo proper refolding. This work was supported by funds from the Centre Naiional de la Recherche Scienri$yue, the Delegation GinnPrale a la Recherche Scientifique et Technique, the CJniversifP Paris VII and the U.S. National Institutes of Health. We thank 1. Emod for the synthesis of the N-carboxyanhydride of DL-ahine.

REFERENCES 1 . Sela, M., White, F. H. & Anfinsen, C. B. (1957) Science (Wash. D . C . ) 125,691 -692. 2 . Epstein, C . J. (1972) in Aspects of Protein Synthesis (Anfinsen, C. B., ed.) pp. 367-431, New York, Academic Press.

440

G. Orsini. C. Skrzynia. and M. E. Goldberg: Renaturdtion of Reduced Chymotrypsinogen

3. Wetlaufer, D. B. & Ristow, S. (1973) Annu. Rev. Biochern. 42, 135 - 158. 4. Eisenberg, M. A. & Schwert, G. W. (1951) J . Gen. Physiol. 34,583 - 606. 5. Brandts, J. F. (1964) J . A m . Chem. Soc. 86, 4291-4301. 6. Brown, J. C. & Horton, H. R. (1972) Proc. Soc. Exp. B i d . Med. 140, 1451- 1455. 7. Epstein, C. J . & Anfinsen, C. 8. (1962) J . Biul. Chem. 237, 2175 - 2179. R. Brown, J . C., Swdisgood, H. E. & Horton, H. R. (1972) Biochem. Biophys. Res. Commun. 48, 1068 - 1073. 9. Sela, M. & Arnon, R. (1972) Merhods Enzymol. 25B, 553-558. 10. Epstein, C. J . & Anfinsen, C. B. (1962) J . Biul. Chem. 237, 3464 - 3467. 11, Saxena. V . P. & Wetlaufer, D. B. (1970) Biochemistry. 9, 5015 - 5023. 12. Hantgan. R. R., Hammes, G. G. & Scheraga, H. A. (1974) Biochemixfry, 13, 3421 - 3431. 13. Creighton, T. E. (1974) J . Mol. Biol. 87, 563-577. 14. Dixon, G . H. & Neurath, H. (1957) J . B i d . Chem. 225, 1049 - 1059. 15. Schwert, G. W. & Takenaka, Y. (1955) Biochim. Biophys. Acra, 16. 570- 575. 16. Ghelis, C., Garel, J. R. & Labouesse, J. (1970) Biochemisfry, 9, 3902-3913. 17. Bender, M. L.. Begue-Canton, M. L., Blakeley, R. L., Brubacher, M. J., Feder, J., Gunter, C. R., Kezdy, F. J., Killheffer. J. V., Jr, Marshall, T. €1.. Miller, C. G., Koeske. R. W. & Stoops,J. K. (1966) J . Am. Chem. Soc. 88, 58905913. 18. Katchalski, E. & Berger, A. (1957) Merhods Enzymol. 3, 546 - 549. 19. Bailey, J . L. (1950) J . Chem. SOC.3, 3461 -3466. 30. Sela, M. & Berger, A. (1955) J . A m . Chem. Soc. 77. 18931898. 21. Egan, K. & Gross, E. M. (1957) Fed. Proc. 16, 176.

22. Epstein, C. J., Anfinsen, C . B. & Sela, M. (1962) J . Biul. Chem. 237,3458- 3463. 23. Ellman, G. L. (1959) Arch. Biochem. Biophjjs. 82. 70-77. 24. Zahler, W. L. & Cleland, W. W. (1968) J . Biol. Chem. 243, 716-719. 25. Wilcox, P. E., Krant, J., Wade, R. D. & Neurath, H. (1957) Biochim. Biophys. Acfu, 24, 72- 78. 26. Chervenka, C. H. (1959) Biochim. Biophys. Acra, 31, 85-95. 27. Williams. E. J.. Herskovits. T. T. & Laskowski, M., Jr (1965) J . Biol. Cham. 240, 3574- 3519. 28. Takagi, T. & Isemura, T. (1964) J . Biochem. ( T u k . ~ )56, 344- 350. 29. Ristow, S. & Wetlaufer, D. B. (1973) Biochem. Biophjx. Rex. Commun. 50,540 - 550. 30. Ryle, A. P. & Sanger, F. (1955) Biochem. J . 60. 535-540. 31. Creighton, T. E. (1974) J . Mu/. Biol. 87, 603-624. 32. Anderson, W. L. & Wetlaufer, D. B. (1974) Fed. Proc. 33. 1309. 33. Tanford, C., Kawahara, K. & Lapanje. S. (1967) J . Am. Chem. SOC.89, 729- 736. 34. Neurath, H., Rupley, J. A. & Dreyer. W. J. (1956) Arch. Biochem. Biuphys. 65, 243 - 259. 35. Freer,S.T., Krdnt,J., Robertus, J . D., Wright, H . T . & Xuong, Ng. H. (1970) Biochemistry, 9, 1997-2009. 36. Wright, H. T. (1970) ./. Mol. Biul. 79, 1-11, 37. Birktoft. J . .I..Blow, D. M., Henderson, R. & StcitL, T. A. (1970) Phil. Trans. R . Soc. Lond. S i r , B B i d . Sci. 275. 67- 76. 38. Birktort, J. J. & Blow, D. M. (1972)J. Mol. B i d . 68. 187-2240, 39. Wetlaufer, D. B. (1973) Proc. Natl Amd. Sci. U . S . A . 70. 697 - 701. 40. Chothia, C. (1975) Nature (Lond.) 254, 304-308. 41. Schulz. G . E., Barry, C. D., Friedman, J . , Chou, P. Y., Fasman, G. D., Finkelstein, A. V., Lim, V. I., Ptitsyn. 0. B., Kabat, E. A., Wu, T. T., Levitt, M., Robson, B. & Nagano, K. (1974) Nature (Lond.) 250, 140- 142.

G . Orsini and M. E. Goldberg, Service dc Biochimie Cellulaire, Institut Pasteur, 28 Rue du Docteur-Roux, F-75724 Paris-Cedex-15, France

C. Skrzynia, Dcpartment of Neuroscience, Children's Hospital Medical Center, 300 Longwood Avenue, Boston, Massachusetts, U.S.A. 021 15

Nore A d d d i n Proof(0ctober 1,1975). Since this paper was submitted, two reviews on protein folding have appeared [Anfinsen, C. B. and Sheraga, H. A. (1975) Adv. Prorein Chem. 29,205-3001 and (Baldwin, R. L. (1975) Annu. Rev. Biochem. 44, 453-4751. Both report that the successful renaturation of reduced chymotrypsinogen was demonstrated by Givol and co-workers [Givol, D., De Lorenzo, F., Goldberger, R. F. and Anfinsen, C . B. (1965) Proc. Natl Acad. Sci. U.S.A. 53, 676-6841, Contrary to this claim, the experiments described in the article cited, in no way indicate that the zymogen has been renatured after reduction of its disulfide bonds: the protein was denatured at pH 4 in 8 M urea and in the absence of any reducing agent. To our view (clearly also that of Givol er a/.) the experiments cited above indicate only the inaccessibility of the S-S bonds in renatured chymotrypsinogen to the disulfide interchange enzyme.