Ca2+, K+-Regulated Intramolecular Crosslinking of S-100 Protein via Disulfide Bond Formation

Eur. J. Biochem. 71, 45-52 (1976)

Ca2 + , K +-Regulated Intramolecular Crosslinking of S-100 Protein via Disulfide Bond Formation Pietro CALISSANO, Delio MERCANTI, and Andrea LEV1

Laboratory of Cell Biology of the Consiglio Nazionale delle Ricerche, Rome

(Received June 14/September 4, 1976)

Reaction of the thiol reagent 5,5’-dithio-bis(2-nitrobenzoic acid) (Nbsz) with the brain-specific protein S-100 favours stabilization of the quaternary structure of the protein via disulfide bond forma- tion. This process is modulated by those cations (Ca2+ and K’) which are known to affect the con- formation of the protein. Ca2’ markedly favours the reaction of s-100 with Nbsz but inhibits subsequent disulfide bond formation; K’, on the contrary, is much less effective in promoting interaction with Nbs2 but strongly stimulates disulfide bond formation. These findings are inter- preted assuming that in presence of Ca2+ the three subunits forming the native s-100 protein have two cysteine residues exposed to the solvent but mismatched to form disulfides while in presence of K + the sulphydryl groups are in a less accessible position to Nbsz but suitable for S-S bond formation.

Crosslinking of S-100 subunits is characterized by the appearance in dodecylsulphate electro- phoresis of two very close protein bands having a molecular weight almost identical to that of the native, undenatured protein but not of higher or lower-molecular-weight components. This finding, and the demonstration that both the crosslinked and native S-100 proteins have identical profiles when analyzed by sucrose density centrifugation or gel chromatography indicate that disulfide bond formation occurs among subunits of the same molecule.

The contribution of the state of sulphydryl groups of the brain-specific protein S-100 to its electro- phoretic heterogeneity, conformational state and re- activity with specific antibodies has been reported already [l - 31. For instance, long-term incubation in the presence of CaC12 results in oxidation of two-SH groups and the appearance of two new bands in electrophoresis [l]. The protein elutes as two distinct molecular species in gel chromatography, unless previously treated with mercaptoethanol, which also changes the pattern of reaction with specific anti- bodies and exerts a protective action against thermal denaturation [2,4]. Thus, although the involvement of cysteine residues in the tertiary and possibly the quaternary structure of S-100 protein has been doc- umented, attempts to correlate their precise contri- bution to the structure(s) of the protein are still lack- ing. We have investigated this problem taking advdnt- age of the documented interplay between CaC12, monovalent cations and conformational state of this protein. It has been shown [5,6] that binding of Ca2+ to this protein induces a conformational change in- volving, beside other amino acids, also two cysteine residues which become rapidly titratable with the thiol -~

Ahhreviations. Nbsz, 5,5’-dithio-bis(2-nitrobenzoic acid); EGTA, [(ethylene-bis(oxoethylenenitrilo)]tetraacetic acid.

reagent Nbsz. This report will show that oxidation of these - SH groups with this reagent results in stabiliza- tion of S-100 quaternary structure to denaturing agents. Furthermore it will be demonstrated that reaction with Nbsz and disulfide formation are distinct, consecutive processes markedly dependent upon the cation present, Ca2+ favouring the former and in- hibiting the latter while K + acts in the opposite way. These findings show unequivocally that the reciprocal, sterical arrangement of S-1 00 subunits is finely depend- ent upon the cation present. Moreover, since cross- linking of subunits via disulfide formation requires a close proximity of the -SH groups, the occurrence of this process with a protein like $100, which appear to be composed of three non-identical subunits [7], indicates that the monomers are not randomly assem- bled but rather precisely organized, a consideration of importance for a protein of unknown function.

METHODS AND MATERIALS

Purijkation of S-I00 Protein

This protein, called S-100 because it is soluble in 100% ammonium sulfate, was purified from beef, horse and calf brain as described by Moore [8]. All the buffers employed during the purifications steps con-

46 Conformation-Dependent Crosslinking of S-I 00 Subunits via Disulfide Bonds

tained 1.0 mM EDTA and 5 mM mercaptoethanol. The chelating agent was omitted in the final Sephadex G-100 column. The concentration of beef S-100 pro- tein was calculated assuming that a 1.0 mg/ml solu- tion of the dried protein in 60 mM KC1,20 mM Tris-C1 pH 8.3 gives an absorbance of 0.344 at 280 nm in a l-cm cell [5] .

Chemicals

All common salts were reagent grade. KC1 and CaC12 'suprapur' were from Merck. Glass-twice- distilled water was used to prepare all solutions. Nbsz from Sigma was dissolved at a final 10 mM concentra- tion in 50 mM Tris-C1 buffer pH 8.3 and stored at 5 - 10 "C for up to one month after preparation.

Reaction with Nbsz and Dodecylsulphate Acrylumide Gel Electrophoresis

Reaction of S-100 protein with Nbsz was followed at 412 nm with a Zeiss PM QII spectrophotometer. The general procedure of -SH titration followed by electrophoresis was the following. To 100- 300 pg of the protein in 0.3 ml of 50 mM Tris-C1 pH 8.4, also containing CaC12, monovalent cations or other sub- stances at the concentration reported in figures, 10 p1 of 10mM Nbs2 dissolved in the same buffer were added. After following the rate and extent of reaction for the time desired, 50 pl of 100 mM iodoacetic acid in 100 mM Tris-HC1 pH 8.4 were added and the solutions allowed to stand for 2 h at room temperature. A sufficient amount of sodium dodecylsulfate to reach a final concentration of 1 % was then added and the solution put to dialyze in acetylated tubes for at least 12 h against 50 mM Tris-C1 pH 8.4 also containing 0.1 % sodium dodecylsulfate. When KC1 was present, an almost instantaneous turbidity occurred after addition of the detergent, but the solution rapidly clarified during dialysis. The samples dialyzed were kept at 4 "C until used for electrophoresis which was accomplished in the absence of reducing agents. The buffer present in the reservoir and in the gels was 100 mM Tris/P04 pH 6.6. Gels were prepared at a concentration of 12.5 acrylamide, 0.31 % bisacryl- amide 0.05 % sodium dodecylsulfate in the absence of reducing agents. A prerun lasting 12 h was performed at 3.0 mA/tube to remove the ammonium persulfate used for polymerization. The buffer was then sub- stituted with a fresh solution and samples of 50- 150 pl were applied and run for 10- 15 h at 1 mA/tube at the constant temperature of 15 "C. Staining was accom- plished with 0.25 Coomassie blue in 44 % methanol/ 8 % acetic acid and destaining in 15% methanol/8% acetic acid. Densitometric tracings of the stained gels were performed with a Joyce-Loebl chromoscan and the fraction of the bands present was calculated by

cutting out and weighing the areas of the densitometric tracings. The protein standards employed to calculate the molecular weight of the S-100 protein bands were : egg albumin (43 500), u-chymotrypsinogen from bo- vine pancreas (23 650), myoglobin from horse heart (16 890), cytochrome c from horse heart (1 3400) and epidermal growth factor (6045).

Separation and Renaturation of S-100 Monomers and Polymer

Separation by ultrogel chromatography was per- formed as follows. The sample of S-100 protein treated with urea as described in Fig. 5 was layered on a 1.5 x 180-cm AcA 54 ultrogel column equilibrated with 50mM Tris/P04 pH 8.4, 100mM NaCl and 6 M urea. Fraction of 1.0 ml were collected at a flow rate or 7.0 ml/h at room temperature and absorbance at 220 nm was measured.

The two peaks containing S-100 polymer (I) and $100 monomer (11) were pooled and regeneration of their sulphydryl groups achieved by addition of 4.0 mM dithiothreitol followed by incubation at 21 "C for 4 h before dialysis against 1 1 of 50 mM Tris/P04 pH 8.4 containing 0.01 mM dithiothreitol and 50 mM KCl. After overnight dialysis, the two peaks were dialyzed against the same buffer without dithiothreitol and KCl followed by another dialysis against bidistilled water.

After lyophilization, the two samples were redis- solved in 2 ml of 50 mM Tris-C1 pH 8.4 and centri- fuged at 20000 x g for 10 min to remove the portion of S-100 protein not completely soluble (generally 30 - 15% of the total) and the supernatant was used for another cycle of Nbsz-catalyzed disulfide formation, performed with the procedure described in Fig. 5 .

RESULTS

Fig. 1 reports a correlation between rate, extent of Nbsz reaction with S-100 protein in different condi- tions, and appearance of a band having a molecular weight of 21 500 - 22 500 in dodecylsulphate acryl- amide gel electrophoresis. In the same gels, another portion of S-100 protein run as a single, monomeric band having an M , of 7500 -t 500. These two molec- ular species will be referred to as respectively S-100 polymer and S-100 monomer since the latter is the monomeric component, while the former appears to be the native polymer stabilized through disulfide bond formation. It is worth mentioning that S-100 polymer can be resolved into two very close com- ponents having M , of 21500 and 22500 (Fig.3B). It can be seen that in the presence of 3 mM CaClz the rate of reaction with the thiol reagent is extremely fast as compared to that measurable in the presence of K + ;

P. Calissano, D. Mercanti, and A. Levi 41

E c N

a T

0

/ I

0.4

0.2

0

1.4 ? cn

I I I A 40 50 V 120 0 10 2 0 30

0

ov.

a c h . P-s100

CY

rn-S100

eg f

c 1 2 3 4 5 6 7 8 St.

Fig. 1. ( A ) Rate of reaction of sulphydryl groups o j S-100 protein with N b s ~ in the presence of various substances and ( B ) dodecylsulpliate acrj~larnide gel electrophoresis of sumplei reacted with Nhsz. (A) The rate of reaction of the protein with Nbs2 was measured as described under Methods; 160 pg S-100 protein was in a final volume of 0.3 ml. Curves 1, 2, 3 = CaC12 at 0.03, 0.3, 3.0 mM. Curve 4 = S-100 protein alone. Curves 5 and 6 = 150 m M NaCl or KCI respectively. Curves 7 and 8 = 6 M urea and 1 sodium dodecylsulphate respectively added to the protein sample 60 min before the sulphydryl reagent. Reaction was initiated by addition of 100 nmol of N b s ~ to each sample. After 120 min of reaction 50 pl of 100 mM iodoacetic acid were added and the samples processed as described under Methods for electrophoresis (B). Samples 5 and 6 contained beside the monovalent cation, also 0.1 mM EDTA. (B) Each number corresponds to a sample of 50 pg of S-100 protein after incubation with Nbsz and dialysis as reported under Mcthods. C refers to a sample of the protein not incubated with N b s ~ and directly mixed with iodoacetic acid. Gel 9 contains the following proteins: epidermal growth factor (ego, cytochrome c (cy), myoglobin (my), r-chymotrypsinogen (cxch) and ovalbumin (ov). Gel 10 contains the same standard proteins plus 40 pg S-300 of sample 6 reacted with Nbsz and KCI. The bars indicate the position of the tracking dye. p-S-100 = S-100 polymer; m-S100 = S-100 monomer

intermediate concentrations of CaClz give intermediate rates of reaction confirming previous findings [ 5 ] , 0.03 mM CaClz being only slightly more effective than K’ in favouring -SH titration. Analysis by dodecylsulphate acrylamide gel electrophoresis of the S-100 samples reacted with Nbsz reveals that there is a type of inverse relationship between rate of reaction and appearance of the S-100 polymer which has an M ,

almost identical to that reported by Dannies and Levine for the native, non-denatured protein [7,9]. While in the presence of KCl the rate and the extent of reaction, as mentioned above, is much slower than with 3.0 mM CaC12, the amount of S-100 polymer is greater with the monovalent than with the divalent cation. Decreasing concentrations of CaClz (0.3 - 0.03 mM) although allowing a slower titration, favour the

48 Conformation-Dependent Crosslinking of S-100 Subunits via Disulfide Bonds

appearance of the new band whose concentration, however, never reaches the amount noticeable with KCI of NaCl. Fig. 1 B shows also that, in contrast to the results obtained with Ca2 + at low concentrations, or with monovalent cations, the protein denatured with a detergent or urea prior to Nbsz addition, gives rise only to monomer bands despite the fact that a maximum number of -SH groups have reacted.

Formation of S-100 polymer appears to be con- sequence of Nbsz-catalyzed disulfide bonds, stabiliz- ing the molecule to the action of denaturing agents. The apparent discrepancy between extent of -SH titration and stabilization of quaternary structure of S-100 protein can be explained if we assume that the reciprocai position of the subunits of S-100 in the presence of KC1 is more suitable to form Nbsz- catalyzed S-S bonds than in the presence of CaCI2. In the presence of this divalent cation the cysteine residues are more available for reaction with the thiol reagent but mismatched for disulfide formation. The validity of this hypothesis will be discussed in detail elsewhere.

Indication that interchain cross-linking mediated by Nbs2 occurs within subunits of the same molecule rather than between different S-100 protein molecules comes from the finding that no higher-M, species are detectabie except those having the same size as the native protein. This conclusion is a!so supported by experiments of sedimentation velocity in sucrose gra- dient and gel chromatography of S-100 protein previously reacted with Nbsz in the presence of K + , in conditions identical to those described in Fig. 1. When this sample is layered on a 5-20% sucrose gradient and run at 45000 rev./min for 24 h at 21 "C on a 50-L Spinco rotor, it sediments as a single, homogenous peak in the same position as a control sample of S-300 protein. An analogous situation occurs when the elution profile of this sample is analyzed and compared to unreacted S-1 00 protein on a Sephadex G-75 superfine column equilibrated with 50 mM Tris/P04 pH 8.3, 100 mM NaCI.

Fig. 2 shows that the S-100 polymer progressively disappears when incubated with a reducing agent, con- firming that stabilization of the quaternary structure occurs via disulfide bond formation. It can be seen that already after a 30-min incubation with dithio- threitol the polymer band obtained by incubation with Nbs2 and KCl or CaC12 has disappeared.

Notice that soon after addition of the reducing agent, a release of the bound portion of Nbsz occurs, as indicated by an increased obsorbance at 412 nm. The amount of the nitrobenzoate anion released is twice as high if the protein has previously been incu- bated with CaC12 than with KCI.

It is well known that reaction of this reagent with thiol groups may proceed through two distinct steps as depicted in the reaction sequence reported below. The aromatic disulfide Nbsz (R-S-S-R) reacts with

,-" 50 c J 0.5

Time a f t e r di th iothrei to l ( rn in )

Fig.2. Rate oj 5-100 protein regenerution ajter dithiothreitof treat- ment. 1.6 mg of the protein were dissolved in 1.2 ml of 50 mM Tris-CI pH 8.4 and divided in two 0.6-ml portions to which were added respectively 3.0 mM CaCll or 150 mM KCI and 200 nmol of Nbsz. After reaction for 120 min, the two samples were chromato- graphed on Sephadex G-25 column equilibrated with 50 mM Tris-C1 pH 8.4. Fractions containing protein were pooled and 1 sodium dodecylsulphate was added. After 30 min at room temperature di- thiothreitol was added to reach a concentration of 10 mM and the increase in absorbance at 412 nm recorded. Before addition of dithiothreitol (zero time) and after 5,10,30,60 and 120 min, samples were taken and mixed with equal volumes of 0.1 M iodoacetic acid. After standing overnight in the cold room the samples were used for electrophoresis. The open symbols report the release of the nitro- benzoate anion following addition of dithiothreitol to samples previously reacted in the presence of KCI (A) or CaClz (0) and the filled symbols refer to the percentage of the polymeric S-100 still detectable by electrophoresis after dithiothreitol addition

aliphatic thiols by an exchange reaction [10,11] to form a mixed disulfide of the protein and 1 mol of 2-nitro-5-thiobenzoate is released :

Protein -SH + R-S-S-R

(SH)fl Protein-S-S-R + R -SH .

I (1) ( S W n This reaction can stop with one nitrobenzoate

anion released and one bound to the protein or it may proceed, depending upon availability of other - SH groups and steric factors, according to this second reaction :

According to the hypothesis postulated above (see also Discussion) in the presence of CaC12 reaction of

P. Calissano, D. Mercanti, and A. Levi

I -

1.8 -

0.4 E ,2

0.3 Y

49

IT A

-

- a

1.0

0.2

C 1 1’ 1“ 2 3 St.

Fig. 3. ( A ) Reaction of Nbsz with sulphydryl groups of S-I00 protein in the presence ofCcrCl2 andsuhsequent disuljideformation in dqferent ionic conditions; ( B ) dodecylsulphate electrophoresis of the samples. (A) 2.1 mg of beef S-100 protein in 1.0 ml 50 mM Tris-C1 pH 8.4 and 1.5 mM CaCh were allowed to react at room temperature with 1.0 mM Nbsz for 5 min (I). After reaction the mixture was chromatographed on a Sephadex G-25 column equilibrated with the buffer and CaClz at the same concentration used for incubation. The fractions of the column con- taining the protein were pooled and divided into three aliquots. One portion was allowed to stand at room temperature without further additions (0) while the other two aliquots were incubated at room temperature after addition of 150 mM KCI (A) or 6.0 mM EGTA (H). The increase in absorbance at 412 nm was followed in each sample as a measure of release of nitrobenzoate anion still bound to S-100 protein after chro- matography (11). Aliquots for electrophoretic analysis were taken immediately before and after chromatography and at the end of incubation, which lasted 6 h. (B) Electrophoresis was performed as described under Methods. C refers to a sample of S-100 protein before reaction with Nbs2. Gels 1, l’, 1” correspond respectively to S-100 samples after reaction with Nbs2 and before (1) or immediately after (1’) chromatog- raphy on Sephadex G-25 and after 6 h incubation with CaClz (I”), Gels 2 and 3 correspond to S-100 protein samples after 6 h incubation with 150 mM KCl (2) or 6 mM EGTA (3). Standard proteins (St) as in Fig. 1

0 ’ I h ,

Nbsz with S-100 protein occurs very rapidly but stops at the first step. Accordingly, little or no disulfide formation occurs, although an average of two moles

of -SH have reacted per mole of protein. This con- clusion is supported by the finding reported in Fig.2 that following addition of the reducing agent dithio-

50 Conformation-Dependent Crosslinking of S-100 Subunits via Disulfide Bonds

threitol a large portion of nitrobenzoate anion is released and that prior to its addition few disulfide bonds had formed as indicated by the smali amount of S-100 polymer.

In contrast, in the presence of K + or Naf , reaction with Nbs2 proceeds much more slowly (although faster than in its absence) but seems to go also through step 2 since much less nitrobenzoate anion is released by addition of the reducing agent dithiothreitol and a large portion of S-100 protein migrates as a 21 000- 22000-M, species. Thus, S-100 protein can exist in two different conformational states : a Ca2+ conformation, highly reactive with a thiol reagent but poorly suited to form disulfide bonds and a K + or Na+ conforma- tion with cysteine residues in close proximity to each other and suitable to form disulfides, although less accessible to the thiol reagent. Fig.3 reports experi- ments confirming this conciusion and showing how by switching from the Ca2+ to the K + conformation a maximum stabilization of S-100 quaternary structure is achieved. Titration of the protein with Nbs2 in the presence of Ca2+ leads to 2.1 -SH groups reacted with the thiol reagent within a 5-min incubation (Fig.3A) but only 15% transformed into the S-100 polymer species (Fig. 3 B, gel 1). Chromatography of this protein sample on a Sephadex G-25 column con- taining CaC12 separates all the free Nbsz from the pro- tein which elutes from the column with an equimolar portion of the aromatic anion bound to its -SH groups. Addition of K + to this S-100 sample, or re- moval of Ca2+ with a specific chelating agent, induces a progressive and marked release of the bound nitro- benzoate anion corresponding in the presence of EGTA, to 0.9 - 1.1 mol/mol S-100 protein (Fig. 3 A) with concomitant appearance in dodecylsulfate elec- trophoresis of S-100 polymer which, after 150 min of incubation, accounts for 40% in presence of K + and 67% in presence of EGTA of the total S-100 protein as measured by densiometric scanning (Fig. 3 B, gel 2 and 3).

In the absence of K + and EGTA, almost no release of bound anion, nor concomitant stabilization of S-100 protein in its native conformation occurs (Fig. 3B, gel 1’ and 1”) showing that in the presence of Ca2+ the conformation of the protein is not suitable for S - S bond formation even during prolonged incubation.

We found that the same amount of S-100 protein is transformed in conditions in which the molar con- centration of Nbs2 is varied between an excess of 3- 30-fold over the protein or when S-100 concentration is varied between 0.2-2.0 mg/ml in the presence of a constant amount of Nbs2 (data not shown). In addition, prolonged incubation of the protein with KCl up to 6 h does not increase the yield of the polymeric S-100 indicating that only part of the sulphydryl groups are available for disulfide formation, unless previously reacted in the presence of Ca2+.

+ -

Beef Horse Cal f

Fig. 4. Dodecylsulpliare electrophoresis of Nlisz treuted S-100 protein f r o m dijjerent species. Samples of S-100 protein purified from beef, horse and calf were incubated for 2 h in the presence of Nbsz and KCI in the same conditions as those described in Fig. 1 and processed for dodecylsulphate electrophoresis as described under Methods. For each type of S-100 protein the gel on the left was loaded with the protein after reaction with Nbsz while that on the right is a control sample. The control of calf S-100 protein is not present in the figure

Finally, the formation of intersubunit S-S bonds is not restricted to beef S-100 protein since horse and calf S-100 proteins show almost identical transforma- tion in the presence of K + and Nbsz (Fig.4).

Attempts to separate on a preparative scale the S-100 polymer and S-100 monomer were successful by the use of gel chromatography on an Ultrogel column equilibrated with 6 M urea as shown in Fig.S. These two populations of S-100 molecules, renatured by ad- dition of a reducing agent, followed by removal of urea, may give rise again independently to poiymer and S-100 monomer species as shown in Fig.6. This finding suggests that stabilization of S-100 protein is a process dependent upon equilibrium conditions rather than the expression of the intrinsic hetero- geneity the protein.

DISCUSSION

The experiments reported allow one to draw some conclusions on the number and reciprocal position

P. Calissano, D. Mercanti, and A. Levi 51

2.4

E & 1.6 N N

T a

0.8

n

n

U 50 60 7 0 80 90 100 110 120

Fraction number

Fig. 5. Ultrogel chromatography in urea of crosslinked S-100 protein. 9.0 mg of beef S-100 protein were treated as described in Fig. 3A by incubation in 3.0 ml of 50 mM Tris-CI pH 8.4 containing 1.5 mM CaClz and 1 .O mM Nbsz. After 15 min incubation, the mixture was chromatographed on a Sephadex (3-25 column equilibrated with the same buffer + CaZ+ used for incubation. Fractions containing the protein with a portion of the nitrobenzoate anion bound were pool- ed, 3.0 mM EGTA was added and the mixture alllowed to stand at room temperature for 4 h followed by dialysis against 2 1 of 50 mM TrisIPOL, pH 8.4. After dialysis, solid urea was added to reach a final 6 M concentration and the sample allowed to stand for 4 h at room temperature before chromatography on AcA Ultrogel (1.5 x 180 cm). Reduction and renaturation of the two peaks was per- formed as described under Methods. The bars indicate the fractions (each of 1 ml) collected to prepare the two pools. I = S-100 polymer; I1 = S-100 monomer

of the subunits forming the native S-100 protein. The portion of this molecule stabilized through di- sulfide bond formation runs in dodecylsulphate gel electrophoresis as two distinct, very close bands hav- ing M , of 21 500 and 22 500 and accounting, altogether, for a maximum of 67 % of the total protein. Depending upon the experimental conditions, e.g. the concentra- tion of Ca2+, the extent of these two bands relative to each other varies, probably in connection with the con- formational change induced by the divalent cation. The remaining monomeric portion of S-100 protein runs as a single, relatively broad band having an M , of 7500 i 500. The possibility of comparing, in denatur- ing conditions and within the same system, the molec- ular weight of both the native and monomeric form of S-100 protein gives an excellent opportunity to extrapolate the number of subunits forming the pro- tein. This is particularly important for molecules with exceptionally high net charge and small size such as S-1 00 protein, which cou!d have abnormal dodecyl- sulphate binding and consequent mobility. It is worth

0

c 1 2 3 4 5

Fig. 6. Interconversion of' the S-100 motzomer and S-100 polymer species after separation by Ultrogel chromatograpliy. The two peaks from theUltrogel column were pooled and treated as described under Methods to regenerate the sulphydryl groups. Samples were then used for another cycle of reaction with Nbsz performed in con- ditions identical to those employed for the first reaction (Fig. 5). C = control sample of S-100 protein. Gel 1 = after Nbsz reaction, Sephadex G-25 chromatography and incubation for 4 h with EGTA; gel 2 and gel 3 = samples from peak I and I1 of the Ultrogel column. Gel 4 and gel 5 = peak I and 11 after regeneration with dithiothrei- tol, reassembly of the subunits by removal of urea and reincubattion with Nbsz. Notice the reappearance of both forms of S-100 protein species in previously homogeneous I and I I peaks

considering, however, that the presence of disulfide bonds, as in the case of S-100 polymer, may diminish the usual dodecylsulphate binding [12]. A comparison of the mobility of S-100 monomers with the stabilized polymer would indicate that the protein is composed of three subunits, a value in agreement with that proposed by Dannies and Levine [9] who reached this conclusion by measuring the M , of native and denatured protein with two independent methods, i.e. ultracentrifugation and dodecylsulphate gel electrophoresis. According to these authors there arc four cysteine residues per S-100 protein molecule ( M , 21 SOO), while previous analysis gave a value of three residues [7,5]. The results presented in this paper indicate that at least 67 of the protein is composed of three subunits with four - SH groups since this is the minimum number neces- sary to form the two disulfide bridges stabilizing the trimer in its native, quaternary structure. The existence of two polymeric species of the protein suggests the possibility that the covaient link occurring through the disulfide bridges stabilizes different combinations of non-identical subunits. In this connection it is worth mentioning that Dannies and Levine obtained a partial separation of one subunit which has the single trypto- phan present in S-100 protein. It is unclear, as yet, why only 67 of the total protein can be transformed, un-

52 P. Calissano, D. Mercanti, and A. Levi : Conformation-Dependent Crosslinking of S-100 Subunits via Disulfide Bonds

less we assume that (a) this process is under equilibrium conditions, or (b) that 30 % of S-100 protein is made of subunits structurally unsuitable to form stable trimers. The data reported in Fig. 5 and 6 are in favour of the former hypothesis since the portion of S-100 protein separated by column chromatography as the 7 000-M, species may in fact be transformed into further stable trimers after reduction of its - SH groups and renatur- ation. This finding also indicates that subunits can be separated in denaturing conditions and reassociated to form trimers whose steric and reciprocal arrange- ment is identicai to that of native, untreated protein.

On the basis of the data reported we propose that the three subunits forming the native S-100 protein before incubation with thiol-reacting substances can exist in two main, opposite conformations : a Ca2 +

conformation, with some cysteine residues exposed to the solvent and thus highly reactive with Nbs2 but too far apart to form disulfide bonds, and a K + confor- mation with - SH groups facing each other, possibly into the interior of the trimer. A two-step incubation of the protein with Nbs2, first in a Ca2+-rich medium, followed by a K+-rich, Ca2+-depleted environment allows stabilization of the quaternary structure.

The fact that the subunits may vary their reciprocal position, but disulfide bond formation results in ap- pearance of two very similar polymeric species having the same size as the native protein, strongly suggests that this protein is precisely organized in its quaternary structure. An intriguing question is whether analogous oxido-reductive processes may occur and affect also the protein - SH groups in a nervous cell and, in the affirmative case, whether they are relevant to its func- tion. Several studies performed on crude or purified preparations of S-100 protein indicate that it is often isoiated in ‘aggregate’ forms having M , higher than those expected for the native protein [7,13] and that the presence of multiples S-100 species [14,15] in

crude extracts cannot always be attributed to a real structural heterogeneity. This report shows that indeed this protein may easily undergo - SH-mediated stab- ilization and/or polymerization, although it cannot be decided whether this process could also occur in the intact cell or after homogenization.

Tests on the possible existence of enzymic or non- enzymic activities oxidizing S-1 00 sulphydryl groups should eventually clarify this controversial and crucial matter.

We wish to thank Prof. P. Fasella for stimulating discussion.

REFERENCES 1. Rusca, G. & Calissano, P. (1970) Biochim. Biophys. Actu, 221,

2. Dannies, P. S. & Levine, L. (1971) J . B id . Chem. 246, 6276-

3. Ansorg, R. & Neuhoff, V. (1971) I n / . J . Neurosci. 2, 151 - 159. 4. Kessler, D., Levine, L. & Fasman, G. (1968) Biochemistry, 7,

5. Calissano, P., Moore, B. M. & Friesen, A. (1969) Biochcmistry,

6. Calissano, P., Alemi, S. & Fasella, P. (1974) Biochemistry, 13,

7. Dannies, P. S. & Levine, L. (1971) J . B id . Chem. 246, 6284-

8 . Moore, B. M. (1965) Biochem. Biophys. Res. Cornmiin. 19,

9. Dannies, P. S. & Levine, L. (1969) Biochem. B i o p h ~ ~ . Res.

10. Habeeb, A. F. S. A. (1972) Methods Enzymol. 25B, 457-464. 11. Leherer, S. S. (1975) Proc. Nut/ Acad. Sci. U.S.A. 72, 3377-

12. Weber, K., Pringle, J. R . & Osborn, M. (1972) Methods En-

13. Stewart, J . A. (1972) Biochem. Biophys. RPS. Commun. 46,

14. Filipowicz, W., Vincendon, G., Mandel, P. & Combos, G.

15. Gombos, G., Zanetta, J. P. & Vincendon, G. (1971) Biochimie

74-86.

6283.

758.

8, 4318-4326.

4553 -4560.

6287.

739 - 744.

Commun. 37,587- 592.

3381.

zymol. 26, 3 - 27.

1405-1410.

(1968) Lge Sci. 7, 1243.

(Puris), 53, 645-655.

P. Calissano, D. Mercanti, and A. Levi, Laboratorio di Biologia Cellulare del C.N.R., Via G. Romagnosi 18/A, 1-00196 Roma, Italy