- Email: [email protected]
Preparation and characterization of bovine albumin isoforms
International Journal of Biological Macromolecules 30 (2002) 259 /267 www.elsevier.com/locate/ijbiomac
Preparation and characterization of bovine albumin isoforms Marı´a Gabaldo´n * Unidad de Histoquimia, Centro de Investigacio´n, Hospital La Fe, Avenida Campanar 21, 46009 Valencia, Spain Received 4 January 2002; received in revised form 13 May 2002; accepted 20 May 2002
Abstract Albumin undergoes changes in conformation and isomerizations by disulfide interchange of unknown biological significance. The aim of this study was to prepare and characterize albumin isoforms, which were stable under near physiological conditions. Modified albumins were obtained by urea denaturation and renaturation, and by aging at low ionic strength and alkaline pH in the presence of cysteine. We describe a cathodic electrophoresis technique, which allows the separation of albumin isoforms with greater positive charge. Differences between native and modified albumins were analyzed by new criteria based on the reactivity of the thiol and histidyl residues and on the susceptibility of the disulfide bonds to sulfitolysis. Modified albumins had, (i) a more cationic component which disappears by sulfitolysis of the disulfide bonds or by incubation with a glutathione redox system; (ii) higher reactivities of the free thiol group and of the histidyl residues, and; (iii) decreased fluorescence. These differences were not observed when processes were carried out on albumin with the thiol group blocked by iodacetic acid, but reappeared with the addition of cysteine. Renatured and aged albumins differed in the nature of the cationic component. Generation of albumin isoforms is dependent on the presence of a free thiol group and seems to involve thiol disulfide interchanges. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Bovine albumin; Cathodic electrophoresis; Thiol-disulfide exchange
1. Introduction Albumin is a single chain multidomain protein with 17 disulfide bonds and one free thiol at Cys34, which partially forms mixed disulfides with cysteine or glutathione. Albumin undergoes pH dependent conformational changes in slightly alkaline solutions, the so called the neutral to basic transition, and isomerizes to the aged form at alkaline pH and low ionic strength. These changes in conformation and isomerizations determine a complex grade of microheterogeneity [1]. In contrast to the abundant literature concerning albumin heterogeneity, relatively little work has been done to assess the biological behavior of these albumin isoforms. A question to be answered is whether these different albumin molecules could form in vivo and, if so, whether they possess special functions. The basic transition has been proposed to have physiological significance in ligand transport [2] and in the regulation of the thiol-dependent * Tel.: /34-6-3862700x50870; fax: /34-6-1973018 E-mail address: [email protected] (M. Gabaldo´n).
antioxidant activity of albumin [3]. The aging reaction has been implicated in albumin catabolism, being the signal for selecting an albumin molecule for degradation [4]. There is sparse literature showing unexpected properties of albumin in biological systems. An albumin has been described with procoagulant activity which induces the expression of tissue factor in endothelial cells and monocytes [5], and regulates endothelial cell protein C activation, fibrinolysis [6] and prostacyclin secretion [7]. In previous works, we found that albumins obtained by heat shock produced an activation of the circulating leukocytes and an increase in leukocyte adhesion to the rat carotid endothelium [8,9]*/these proinflammatory activities not being related to differences in chemical composition. These results prompted us to study the effect of more subtle differences in the molecule by using isomeric forms of albumin. For this purpose we needed, first, to obtain albumin with changes in the conformation which were stable under physiological conditions and second, to establish criteria for distinguishing the native from the modified forms.
0141-8130/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 4 1 - 8 1 3 0 ( 0 2 ) 0 0 0 4 0 - 5
260
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
In this paper, we describe the preparation of modified albumins by aging in the presence of cysteine and by renaturation after treatment with urea in non-reducing conditions. Both procedures produced more cationic albumins, which were separated from the native albumin by a new technique of cathodic electrophoresis. To distinguish native from modified albumins, we applied new criteria based on the reactivity of the thiol and histidyl residues and on the susceptibility of disulfide bonds to sulfitolysis. The modified albumins differed with respect to each other and the native form according to all applied criteria, and exhibited their properties under near physiological pH and ionic strength conditions, far different to those in which they were obtained. In order to explore the role of the free thiol group and of the thiol disulfide interchanges in the formation of albumin isoforms, we utilized albumin with the thiol group blocked with iodoacetic acid, and performed experiments in the presence of protein disulfide isomerase (PDI) and a glutathione redox system. Results indicate that generation of albumin isoforms is dependent on the presence of a free thiol group and seems to involve thiol disulfide interchanges.
2. Materials and methods
2.3. Urea denaturation and renaturation An 8 mg/ml albumin solution in 20 mM sodium phosphate buffer, final pH 7, and 8 M urea in 20 mM sodium phosphate buffer pH 7 were mixed in 1:3 proportion, and denaturation was carried out for 24 h at 4 8C. The final urea concentration was kept at 6 M because this concentration produced a more cationic albumin in experiments of isoelectric focusing in the presence of urea [10]. Although a 6 M concentration is not sufficient for complete denaturation, it is proximal to the midpoint of the second transition in which all the three domains of albumin undergo conformational changes [11,12]. Another reason for using 6 M urea was to facilitate its removal by dialysis, so avoiding a prolonged time of renaturation, which would increase the formation of albumin polymers. Renaturation was performed by dialysis against 20 mM sodium phosphate buffer pH 7 for 24 h at 4 8C with stirring. During this time three changes each with a buffer volume 50 times that of the albumin were carried out. Under these conditions urea remaining in the dialysate was less than 1 mg/100 ml. Renatured albumin was concentrated to 8 mg/ml by ultrafiltration through a Filtron membrane 30 K Omega type and filtered through a 0.45 mm filter. 2.4. Preparation of aged albumins
2.1. Chemicals Bovine serum albumin fraction V obtained by cold alcohol precipitation (A-4503), S -carboxymethyl albumin (A-6285), PDI (lot 48H4055), glutathione reduced form (G-6529), glutathione oxidized form disodium salt, sodium tetrathionate, sodium sulfite (S-4672), L-cysteine free base (C-7755), 5,5?dithio-bis (2-nitrobenzoic acid) (DTNB), diethylpyrocarbonate (DEPC) and 8-anilino1-naphthalenesulfonic acid (ANS) hemimagnesium salt were provided by Sigma Chemical Co. Fuchsin basic (47860) was provided by Fluka Chemika-Biochemika (Switzerland). 2.2. Preparation of buffered albumin solutions Depending on concentration, albumin solutions in buffers may not reach the desired pH due to the buffering capacity of albumin. As the pH affects the reaction rate in all the techniques described below, pH adjustment is critical for comparative purposes. When starting from buffered albumin solutions of different pHs, samples must be adjusted individually with one of the components of the buffer until the desired pH is reached, before completing volume with the adequate buffer. Accordingly, the pHs indicated in the text always refer to final pH of the albumin solutions.
A 50 mg/ml albumin solution was filtered through a 0.45 mm filter and 10 ml were desalted by passing through a Hi Prep† 26/10 desalting column (Amersham Pharmacia Biotech) equilibrated with H2O. The albumin was eluted with H2O, adjusted to pH 8.9 with 0.3 M Tris or 40 mM NaOH, and diluted to 20 mg/ml. Aging in the presence of cysteine was performed by adding Lcysteine free base to yield a final concentration of 1.5 mM (cysteine/albumin molar ratio, 5). Solutions were gassed with nitrogen and stored in the dark for 24 h at room temperature. 2.5. Cathodic PAGE Polyacrylamide gel electrophoresis was performed in native conditions at acid pH. In this type of electrophoresis, the lower electrode is the cathode and the upper is the anode, since proteins migrate as cations. Slab gels (100 /100/1 mm) were obtained by standard procedures. The buffer of the resolving gel was prepared as a stock solution (4 /) containing 120 mM KOH adjusted to pH 4 with acetic acid The composition of the resolving gel was as follows: acrylamide, 7.5%; bisacrylamide, 0.27% (T /7.8% and C /3.5%); buffer of the resolving gel (4 /), 25%; glycerol, 25% (v/v); TEMED, 3.7 ml/ml of the gel; 1.5% ammonium persulfate, 50 ml/ml of the gel. The buffer of the stacking gel was prepared as
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
a stock solution (2 /) containing 30 mM KOH adjusted to pH 4.4 with acetic acid. The composition of the stacking gel was as follows: acrylamide, 3.75%; bisacrylamide, 0.14% (T /3.9% and C /3.5%); buffer of the stacking gel (2 /), 50%; glycerol, 10% (v/v); TEMED, 4 ml/ml of the gel; 1.5% ammonium persulfate, 50 ml/ml of the gel. The electrophoresis buffer was: 100 mM balanine adjusted to pH 4.2 with acetic acid. The final concentrations in the sample buffer were: glycerol 20% (v/v); buffer of the stacking gel (2 /), 50%; tracking dye, 0.4% fuchsin basic, 25 ml/ml. Equal volumes of 1 mg/ml of albumin and sample buffer were mixed and 10 ml of this solution were loaded into the well. Gels were run by applying 50 V until the dye passed through the stacking gel. The voltage was then increased to 150 V and electrophoresis was continued until disappearance of the tracking dye. Gels were stained with colloidal Coomassie (ICN Biomedicals Inc.) for 2 h to overnight, rinsed with water and placed into the image enhancing solution (8% acetic acid (v/v), 25% ethanol (v/v)) for 10 min with gentle swirling. Gels were destained in 10% acetic acid (v/v). 2.6. Reactivity of the histidyl groups DEPC reacts with histidyl residues in albumin to yield a N -carbethoxyhistidyl derivative which absorbs at 240 nm. Stock 91 mM DEPC solutions were prepared in acetonitrile and stored at /20 8C. Since the purity of the commercial DEPC may be variable owing to hydrolysis, its concentration must be determined by reaction with imidazole [13]. Albumin solutions (1 mg/ml) were prepared in 0.1 M sodium phosphate buffer, final pH 6. Into a semimicrocuvette 1.18 ml of 1 mg/ml albumin were placed and autozeroed at 240 nm; 20 ml of 91 mM DEPC in acetonitrile were added and thoroughly mixed. Readings were started at 15 s and monitored for 7 min at 20 s intervals. A blank without albumin was processed in the same way. As DEPC is partially hydrolyzed during reaction, its absorption at t/0 was calculated from the reading of a blank, zeroed with 1.18 ml of acetonitrile, to which 20 ml of DEPC were added. The concentration of the N -carbethoxyhistidyl derivative was calculated from the absoption at 240 nm after subtracting the absorbance of the blank by using an extinction coefficient of 3.2 per mM/cm [13]. Determinations were performed in duplicate. As Tris accelerates the decomposition of DEPC, when reactivity of the histidyl groups is determined in albumin aged in the presence of Tris, this must be added to the controls in the same proportion as in the aged albumins. DEPC reacts with thiol groups and their presence must be taken into account when determining the reactivity of the histidyl groups. The interference produced by the cysteine used in the aging process is discrete and may be
261
corrected by using a blank with a final concentration of cysteine similar to that of the reaction mixture. The technique described cannot be applied in the presence of a glutathione redox system, since the higher concentration of thiol groups causes a notorious loss of DEPC.
2.7. Reactivity of the /SH group DTNB reacts with the thiol group in albumin to produce an equivalent of 5-thio-2-nitrobenzoate (TNB) which absorbs at 412 nm [14]. Albumin solutions (5 mg/ ml) were prepared in 40 mM sodium phosphate buffer, final pH 7; DTNB (0.7 mM) was prepared in 0.1 M Tris /HCl buffer-4.5 mM disodium EDTA, final pH 7. The spectrophotometer was autozeroed at 412 nm with a blank containing 0.8 ml of the DTNB solvent and 0.4 ml of the albumin solution. The reaction was started by adding 0.8 ml of DTNB to 0.4 ml of albumin. A blank without albumin was processed in the same way. The concentration of TNB was calculated from the absorption at 412 nm after subtracting the absorbance of the blank, assuming an extinction coefficient of 13.6 per mM/cm [14]. Determinations were performed in duplicate. Pseudo-first-order rate constants (kobs) were calculated by fitting the data to a single exponential.
2.8. Reactivity of the disulfide bonds Sulfitolysis is a reaction in which sulfite ions cleave disulfide bonds to S -sulfonates. The composition of the sulfitolysis reagent was as follows: Tris, 0.1M; Na2EDTA, 2.5 mM; sodium sulfite, 0.2 M; sodium tetrathionate, 32 mM; pH adjusted to 8 with HCl [15]. A blank was prepared without sodium sulfite and sodium tetrathionate. Equal volumes of albumin (2 mg/ml in 0.1 M sodium phosphate buffer, final pH 8) and sulfitolysis reagent or blank were mixed and incubated for 30 min. After reaction, aliquots were mixed with the same volume of sample buffer and subjected to cathodic PAGE. Sulfitolysis of accessible disulfide bonds increases the hydrodynamic volume and the negative charge of albumin, which produces a lesser mobility in cathodic electrophoresis.
2.9. Fluorescence measurements The fluorescent probe ANS was used to monitor conformational changes in albumin as it fluoresces strongly in non-polar environments such as the hydrophobic binding sites of albumin [16]. Intrinsic and ANS fluorescence were determined as described previously [9].
262
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
Fig. 1. Fluorescence spectra of aged and renatured albumins. (A) Intrinsic fluorescence A-4503, (1) control; (2) aged; (3) renatured; (4) aged-Cys. (B) ANS fluorescence A-4503, (1) control; (2) aged; (3) aged-Cys; (4) renatured. (C) Intrinsic fluorescence A-6285 (S -carboxymethylated), (1) control; (2) aged; (3) renatured; (4) aged-Cys. (D) ANS fluorescence A-6285 (S -carboxymethylated), (1) renatured; (2) aged; (3) control; (4) aged-Cys. The emission spectra were recorded in albumin dissolved in 60 mM sodium phosphate buffer pH 7, upon excitation at 280 nm (intrinsic fluorescence) and at 374 nm (ANS fluorescence). The ANS/albumin molar ratio was 6. Fluorescence is expressed in arbitrary units and the values are the mean of two independent determinations.
2.10. Analytical procedures Determination of the free thiol group in albumin was performed with the technique described for the study of the /SH group reactivity, except that albumin concentration was 6 mg/ml and pH for both albumin and reagents was 7.5. Absorbances were measured at 15/25 min. Fractionation of albumin was performed by gel filtration using Superdex 200 prep grade on a HiLoad 16/60 column (Amersham Pharmacia Biotech) as described previously [9]. The column was loaded with 1 ml of 20 mg/ml of albumin in the eluent buffer. A pool of dimeric and polymeric fractions from albumin A-4503 were rechromatographied, concentrated and used as electrophoretic and biochemical standards. 2.11. Reactions with PDI and glutathione PDI is a multifunctional enzyme which catalyzes the reduction and isomerization of misfolded proteins [17]. When reaction was carried out in the presence of PDI, the enzyme was first conditioned for 10 min in a
glutathione redox system by incubating 28 U of PDI in 0.5 ml of a solution containing 0.1 M sodium phosphate buffer, 2 mM reduced glutathione (GSH), 1.5 mM oxidized glutathione (GSSG) and 1 mM Na2EDTA, final pH 7.5. The final concentrations in the reaction mixture were: albumin, 5 mg/ml; GSH, 5 mM; GSSG, 3.75 mM; Na2EDTA, 1 mM; sodium phosphate buffer, 0.1 M; PDI, 19 U/ml; total volume, 1.5 ml; final pH, 7.5. Incubation was carried out for 4 h.
3. Results 3.1. Renatured albumin After 24 h in 6 M urea at 4 8C and pH 7, albumin A4503 was renatured during 24 h by dialysis against 20 mM sodium phosphate buffer pH 7. When albumin concentration was 2 mg/ml, this process determines a discrete increase in the dimeric and polymeric fractions and a concomitant decrease in the content of the free / SH group. Control: 78% monomer, 17% dimer, 5% polymer, and /SH/albumin molar ratio, 0.45. Rena-
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
tured albumin: 63% monomer, 24% dimer, 13% polymer and /SH/albumin molar ratio, 0.37. With an albumin concentration of 25 mg/ml, intensive polymerization occurs and the monomeric fraction decreases to 24%. Albumin A-6285 with the thiol group blocked by iodoacetic acid had 81% monomer, 15% dimer and 4% polymer, and these fractions were not modified after renaturation. Native and renatured albumin A-4503 differed in, (i) the fluorescence profile. The spectrum of renatured albumin after exciting at 280 nm (intrinsic fluorescence) showed a shift in the emission maximum from 340 to 335 nm (Fig. 1(A)). The decreases in intrinsic and ANS fluorescence observed in renatured albumin are shown in Fig. 1(A and B) and in Table 1; (ii) the electrophoretic pattern. Cathodic PAGE of renatured albumin showed new bands with greater mobility than the native monomer. These more cationic albumin isoforms were poorly resolved, producing a significant band broadening. The dimer also presented a new band with greater mobility (Fig. 2); (iii) the response to the sulfitolysis reagent. Native albumin was not affected after 30 min incubation. Renatured albumin lost the more cationic monomeric and dimeric bands and practically all the polymeric bands. These losses were not produced when sulfitolysis was performed at t /0, nor when the reagent was diluted 1/5. Renatured albumin showed a new narrow band at the beginning of the resolving gel which was absent in the native albumin (Fig. 2(A)); (iv) the reactivity of the histidyl groups. Histidyl groups in
Fig. 2. Cathodic PAGE of aged and renatured albumins. (A) Sulfitolysis of the disulfide bonds (A-4503), (1) control; (2) aged; (3) aged-Cys; (4) renatured; (5 /8), the same albumins after sulfitolysis for 30 min. (B) Effect of a glutathione redox system, (1) A-4503; (2) agedCys A-4503; (3) aged-Cys A-4503/glutathione; (4) renatured A-4503; (5) renatured A-4503/glutathione; (6) aged A-6285 (S -carboxymethylated); (7) aged-Cys A-6285; (8) aged-Cys A-6285/glutathione.
renatured abumin were more reactive than in native albumin (Table 2). Results were not subjected to a rigorous kinetic analysis because of two difficulties.
Table 2 Reactivity of the histidyl residues in aged and renatured albumins Albumin
Table 1 Effect of a glutathione redox system on intrinsic and ANS fluorescence Albumin
Intrinsic fluorescence (% control)
ANS fluorescence (% control)
A-4503 Dimer A-4503 Renatured A-4503 IdemPDI IdemGLUT IdemPDIGLUTa Aged A-4503 Aged-Cys A-4503 IdemGLUT A-6285b Renatured A-6285 Aged A-6285 Aged-Cys A-6285 IdemGLUT
100 88 80 85 88 90 86 78 93 100 100 100 82 89
100 79 52 57 65 68 82 72 99 100 104 100 77 90
Values were taken at the emission maximum upon excitation at 280 nm (intrinsic fluorescence) and at 374 nm (ANS fluorescence) and represent the mean of two independent determinations. a protein disulfide isomerase (PDI); glutathione redox system (GLUT). b A-6285 has the SH group blocked with iodoacetic acid.
263
N -carbethoxyhistidyl derivativea
Histidine/albumin molar ratiob
pH 6 A-4503 Dimer A-4503 Renatured A-4503 A-4503 Aged A-4503 Aged-Cys A-4503 A-6285c Aged A-6285 Aged-Cys A-6285 Renatured A-6285
13.7 16.8 30.0 15.9 25.4 36.5 16.6 16.9 26.9 18.8
8.3 9.0 11.0 8.2 9.4 10.5 8.1 8.1 9.0 8.4
pH 7 A-4503 Renatured A-4503 A-4503 Aged A-4503 Aged-Cys A-4503
34.5 60.4 33.9 46.2 64.2
12.1 13.4 12.0 12.5 13.1
Values represent the mean of two independent determinations. Concentration (mM) of the N -carbethoxyhistidyl derivative in the incubation mixture at 15 s, formed in the reaction of DEPC (1517 mM) with histidyl residues in albumin (15 mM). b Histidine/albumin (molar ratio) at 7 min. Bovine albumin has 17 histidyl residues. c A-6285 has the SH group blocked with iodoacetic acid. a
264
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
First, DEPC hydrolyzes partially during reaction and second, the alternative of increasing its concentration must be discarded because excess DEPC can react to form a disubstituted histidyl derivative with a higher absorbance at 240 nm [18]. At pH 6, the renatured albumin reacted more rapidly with DEPC, and had accessible for reaction 2.7 histidyl residues more than the native albumin. Since only the unprotonated form of histidine reacts with DEPC, the reaction is affected by external pH and by the particular pK of the each of the histidyl groups. When reaction is performed at pH 7, the responses of native and renatured albumin were different. Both increase the reactivity of the histidyl groups as a consequence of the greater amount of unprotonated histidine, but the increase in reaction rate was lower in renatured albumin, having accessible for reaction only 1.3 histidyl residues more than the native albumin; and (v) the reactivity of the free thiol group. The pseudofirst-order rate constants for the reaction of DTNB with the free /SH group of albumin at pH 7 were 64 and 996 (kobs /104, (per s)), for the native and renatured albumin, respectively. Since only the thiolate anion reacts with DTNB, the reaction is affected by external pH and by the pK of the thiol group. At pH 7.5, both albumins increased the reactivity of the /SH group as a consequence of the greater concentration of the thiolate anion, but their response was quite different as native albumin multiplied its rate constant by 5.7 while renatured albumin did so by only 1.3. In order to see whether differences observed between native and renatured albumin A-4503 were partly due to the excess of albumin dimer, we studied the fluorescence spectra and the reactivity of the histidyl residues in the pure dimer. We found that the dimer had 88% of the intrinsic and 79% of the ANS fluorescence of A-4503 (Table 1), without changes in the l of the emission maximum. The reactivity of histidyl residues in the pure dimer was discretely higher than in A-4503 (Table 2). As differences in the dimer content between native and renatured albumin were only of 7%, the contribution of the albumin dimer to the changes observed in renatured albumin may be considered negligible. When albumin had the thiol group blocked by iodoacetic acid (A-6285), the differences between native and renatured albumin disappeared. Neither the l of the emission maximum nor the intrinsic and ANS fluorescence were modified (Fig. 1 (C and D) and Table 1); the more cationic component was not detected by electrophoresis (data not shown), and the reactivity of histidyl residues was only discretely increased (Table 2). 3.2. Aged albumins The aging process of albumin A-4503 in the presence or absence of cysteine did not modify the content of dimers and polymers, or the free thiol content.
Native and aged albumins differed in: (i) the fluorescence profile, the greater changes being present in the albumins aged in the presence of cysteine. Spectra of aged albumins after exciting at 280 nm showed a blue shift from 340 to 337 nm (aged) or 335 nm (aged-cys) (Fig. 1(A)). The decreases in intrinsic and ANS fluorescence observed in aged albumins are shown in Fig. 1 (A and B) and Table 1. S -carboxymethyl albumin with the free thiol group blocked with iodoacetic acid (A-6285) showed changes in the fluorescence spectrum only when aged in the presence of cysteine (Fig. 1 (C and D) and Table 1; (ii) the electrophoretic pattern. Cathodic PAGE of aged albumins showed two new bands with greater mobility than the native monomer. These more cationic albumin isoforms were well resolved and were present in a greater amount in aged-cys albumins. The dimer also presented a new band with greater mobility (Fig. 2). S carboxymethyl albumin produced these forms only when aged in the presence of cysteine (Fig. 2(B)). Anodic native and SDS-PAGE failed to show any difference between native and aged albumins; (iii) the response to the sulfitolysis reagent. Upon sulfitolysis aged albumins lost the more cationic monomeric and dimeric bands and showed a new narrow band at the beginning of the resolving gel, which was absent in the native albumin (Fig. 2(A)); (iv) the reactivity of the histidyl groups. Histidyl groups in aged albumins were more reactive than in native albumin (Table 2). At pH 6, aged albumins had accessible for reaction 1.2 histidyl residues more than native albumin and 2.3 histidyl residues more when aging was performed in the presence of cysteine. At pH 7, the responses of native and aged albumins were different. Both increase the reactivity of the histidyl groups but the increase in reaction rate was lower in aged albumins. S -carboxymethyl albumin showed an increase in the reactivity of the histidyl groups only when aged in the presence of cysteine (Table 2); and (v) the reactivity of the free thiol group. The pseudo-first-order rate constants for the reaction of DTNB with the free /SH group of albumin at pH 7 were 64 and 144 (kobs /104, (per s)), for the native and aged albumin. At pH 7.5, both albumins increased the reactivity in the same proportion being the the kobs / 104, 365 and 740 (per s) for the native and aged albumin, respectively. 3.3. Effect of PDI and glutathione The incubation of the modified albumins with a glutathione redox system partially reverted the effects produced by aging and renaturation. The increase in intrinsic and ANS fluorescence of renatured albumin was higher with glutathione than with PDI (Table 1). Addition of glutathione produced a red shift in the emission maximum (intrinsic fluorescence), reaching a value of 338 nm, intermediate between that of the native
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
and modified albumins. The effect of PDI was discrete and may be due to the residual glutathione utilized for the enzyme activation. The reversal effect on the fluorescence intensities was more notorious in agedCys albumin A-4503 than in renatured albumin. The same occurred when they were analyzed by cathodic PAGE. In aged albumins the more cationic bands disappeared completely with glutathione, whereas in renatured albumin the bands remain and the possible variations were not apparent due to the broadening of the bands (Fig. 2(B)).
4. Discussion The renaturation of albumin denatured by urea depends on the experimental conditions utilized, the main determinants in this process being, (i) the existence of an oxidative environment if the starting molecule is a reduced and unfolded protein; (ii) the procedure of urea removal, by dilution (rapid) or dialysis (gradual); (iii) the albumin concentration and pH; and (iv) the presence of a free or blocked thiol group. The existence of multiple protocols, together with the different criteria applied to characterize the renatured albumin, explain the discrepancies found in the literature concerning the reversibility of denaturation upon urea removal [10,11,19,20]. Urea weakens hydrogen bonds and hydrophobic forces of the albumin mainly in domain III and to a lesser extent in domain I [12,19,21]. Albumin alkalinization produces the rupture of salt bridges between positively charged histidyl residues in domain I and anionic residues in domain III [2]. This loosening of the forces which maintain the tertiary structure in native albumin produces more open conformations with unmasking of ionizable groups whose normalization determines changes in the isoelectric point, which are higher in the urea denatured albumin (/1 pH unit) [10] than in aged albumin (/0.2 /0.4 pH units) [22,23]. The reversibility of these changes upon stimulus removal has not been well established and may not to be complete if a fraction of the molecules covalently stabilize their more open configuration with a thiol disulfide interchange. This possibility has been considered in aged albumin, suggesting a shuffling of one or two disulfide bonds in domain I [24]. In this work, we describe an electrophoretic technique which allows separation at a pH lower than the isoelectric point of albumin isoforms with a greater positive charge which are not resolved from the native form by anodic conventional electrophoresis. The presence of these more cationic isoforms in aged albumins is concordant with two of their characteristics, the greater proton binding below the isoionic point [24], and their higher isoelectric point [22,23]. The one or two more
265
cationic bands present in aged albumins may correspond to isomers formed by a restricted thiol disulfide interchange since, (i) the bands were not produced when the free thiol group of albumin was blocked with iodoacetic acid but reappeared on addition of cysteine, an external /SH donor acting as a catalyst; (ii) the bands disappeared on sulfitolysis, indicating an abnormal accessibility of the disulfide bonds to the reagent as the buried location of all 17 disulfide bonds in albumin makes them resistant to sulfitolysis in the absence of denaturing agents [15]; and (iii) the bands disappeared after incubation with a glutathione redox system which can produce a rearrangement of erroneous disulfide pairings following a pathway of sequential reactions involving mixed disulfides with glutathione, as has been described for the refolding of ribonuclease [25]. It has been proposed that prolonged exposure of albumin to urea results in intramolecular disulfide interchange [10]. The more profound conformational change produced by urea would allow a certain randomization of the disulfide interchange in contrast to the restricted shuffling of one or two disulfide bonds proposed in the aging process [24]. This would explain the heterogeneous pattern found by isoelectric focusing techniques, with 10/20 bands focused over the pH range between the isoelectric point of native and denatured albumin [10]. Our results are in accordance with this interpretation, since we found albumin isoforms with higher mobility in cathodic PAGE, consistent with a higher pI, and a poorly resolved banding pattern indicative of multiple intermediate states. The more cationic bands in renatured albumin disappeared upon sulfitolysis, which indicates an abnormal accessibility of the disulfide bonds. The occurrence of thiol disulfide interchanges in renatured albumin is also supported by the results obtained with S -carboxtmethylated albumin. In the absence of a free thiol group acting as a catalyst, the more cationic components were not produced and the fluorescence spectra were indistinguishable from those of the native form. The latter can mean that the more open conformations produced by urea are not stabilized covalently by a thiol disulfide interchange and that upon urea removal a complete renaturation allows achievement of the more folded conformation of the native form. In this work, we applied new criteria to distinguish native from modified albumins, based on the reactivity of the thiol and histidyl residues. The reactivity of both groups depends in part on steric factors, and an indirect approximation of the accessibility to the target groups is given by the fluorescence measurements [26]. The decreases in both intrinsic and ANS fluorescence observed in aged and renatured albumins point to a more open conformation than in native albumin, as these changes are produced by a higher exposure of fluorophors to solvent and by a loss of hydrophobicity,
266
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
respectively. The reactivity depends also on the pK of the thiol and histidyl residues. A decrease in the thiol pK by a more cationic environment will produce a higher concentration of thiolate anion, the reactive form with DTNB, and a greater reactivity of the /SH group. The severance of salt bridges by alkalinization results in a deprotonation and in a lowering in pK of the histidyl residues [2]. This downward pK shift determines a higher concentration of the unprotonated form of histidine and, as a consequence, a higher reactivity */ since only the unprotonated form reacts with DEPC. The contribution of the pK changes to the reactivity of both groups may be explored by performing reactions at different pHs. Differences in reactivity which are pH dependent suggest pK changes in the groups of modified albumins. This is the case for renatured albumins whose thiol and histidyl residues seem to have a lower pK than the native albumin. The mechanism of ANS binding to albumin is still far from being completely understood despite its common use as a fluorescent hydrophobic probe. Neither the location of the five sites on albumin within which ANS is fluorescent [27] nor the microenvironment surrounding the binding sites are known. Binding of ANS to albumin comes from both the nonpolar anilinonaphthalene group (hydrophobic binding where intrinsic ANS fluorescence is not quenched by water) and the negatively charged sulfonate group (ionic binding which depends on the protein cationic charge and pH). Difficulties in interpreting results obtained by ANS are due to the fact that electrostatic as well as hydrophobic mechanisms contribute to the ANS /albumin interaction [28,29]. The modifications in tertiary structure produced by aging and renaturation may affect hydrophobicity in different ways. An open conformation is compatible with decreased ANS fluorescence when hydrophobic pockets are widened or change in position. In our study, we have found that modified albumins have decreased ANS fluorescence and greater reactivity of the thiol group, possibly due to a less steric restriction produced by the opening of the hydrophobic crevice which contains Cys34 [30]. The greater accessibility of the sulfitolysis reagent to the disulfide bonds in modified albumins may indicate that certain pockets have a greater facility for water penetration, which would quench the ANS fluorescence. Apart from a hydrophobic environment, ANS binding requires proximal cationic charges that may attract negatively charged ANS molecules. In our study, we have found that modified albumins had more reactive histidyl residues, which indicates a loss of positive charges since only the unprotonated form of histidine reacts with DEPC. In the event histidyl residues were in the microenvironment of any of the ANS binding sites, the decrease in ANS fluorescence in modified albumins
could also be explained by a reduction of the histidine positive charges capable of retaining negatively charged ANS in the surroundings of the hydrophobic binding site. Albumin carries out the transport and distribution of exogenous and endogenous ligands. The affinity of these ligands is dependent on the state of the neutral to basic equilibrium, which is in turn affected by pH and calcium concentration [2]. As these conditions vary in different tissues and organs, it is clear that albumin conformational changes may play a significant role in the pharmacokinetics of drugs. The preparation and characterization of albumin isoforms may help to better understand albumin functionality in vivo, and may be a useful tool for studying the anomalous behavior of certain albumins with procoagulant and proinflammatory activities [5,8,9].
Acknowledgements The author thanks Elena Carrascosa and Nuria Martin for technical assistance. This work was supported by a grant from Plan Nacional I/D No. SAF 96-0010.
References [1] Peters T, Jr. All about albumin. Biochemistry, genetics and medical applications. San Diego, CA: Academic Press, 1996. [2] Bos OJM, Labro JFA, Fischer MJE, Wilting J, Janssen LHM. J Biol Chem 1989;264:953. [3] Lee H, Cha MK, Kim IH. Arch Biochem Biophys 2000;380:309. [4] Wallevik KJ. Clin Invest 1976;57:398. [5] Faucette KJ, Parker CJ, McCluskey T, Bernshaw NJ, Rodgers GM. Blood 1992;79:2888. [6] Gubler DB, Wilson BD, Parker CJ, Rodgers GM. Thromb Res 1993;70:459. [7] Gubler DB, Ahlstrom CR, Liu L, Zhou JF, Parker CJ, Rodgers GM. Thromb Haemost 1995;74:1573. [8] Gabaldo´n M, Zu´n˜iga A, Capdevila C, Marque´s A. Atherosclerosis 1994;110:209. [9] Zu´n˜iga A, Marque´s A, Gabaldo´n M. Thromb Res 1997;86:243. [10] Salaman MR, Williamson AR. Biochem J 1971;122:93. [11] Muzammil S, Kumar Y, Tayyab S. Proteins Struct Funct Genet 2000;40:29. [12] Tayyab S, Sharma N, Khan MM. Biochem Biophys Res Commun 2000;277:83. [13] Miles EW. Methods Enzymol 1977;47:431. [14] Janatova J, Fuller JK, Hunter MJ. J Biol Chem 1968;243:3612. [15] Malhotra M, Sahal D. Int J Pept Protein Res 1996;48:240. [16] Damodaran S. Int J Pept Protein Res 1986;27:589. [17] Zheng J, Gilbert HF. J Biol Chem 2001;276:15747. [18] Roosemont JL. Anal Biochem 1978;88:314. [19] Khan MY, Agarwal SK, Hangloo S. J Biochemistry 1987;102:313. [20] Burton SJ, Quirk AV, Wood PC. Eur J Biochem 1989;179:379. [21] Tanaka N, Nishizawa H, Kunugi S. Biochim Biophys Acta 1997;1338:13. [22] Wallevik K. Biochim Biophys Acta 1976;420:42.
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267 [23] Nikkel HJ, Foster JF. Biochemistry 1971;10:4479. [24] Stroupe SD, Foster JF. Biochemistry 1973;12:3824. [25] Ruoppolo M, Freedman RB, Pucci P, Marino G. Biochemistry 1996;35:13636. [26] Dockal M, Carter DC, Ru¨ker F. J Biol Chem 2000;275:3042.
267
[27] Daniel E, Weber G. Biochemistry 1966;5:1893. [28] Matulis D, Lovrien R. Biophys J 1998;74:422. [29] Matulis D, Baumann CG, Bloomfield VA, Lovrien RE. Biopolymers 1999;49:451. [30] Cornell CN, Kaplan LJ. Biochemistry 1978;17:1755.
Preparation and characterization of bovine albumin isoforms Marı´a Gabaldo´n * Unidad de Histoquimia, Centro de Investigacio´n, Hospital La Fe, Avenida Campanar 21, 46009 Valencia, Spain Received 4 January 2002; received in revised form 13 May 2002; accepted 20 May 2002
Abstract Albumin undergoes changes in conformation and isomerizations by disulfide interchange of unknown biological significance. The aim of this study was to prepare and characterize albumin isoforms, which were stable under near physiological conditions. Modified albumins were obtained by urea denaturation and renaturation, and by aging at low ionic strength and alkaline pH in the presence of cysteine. We describe a cathodic electrophoresis technique, which allows the separation of albumin isoforms with greater positive charge. Differences between native and modified albumins were analyzed by new criteria based on the reactivity of the thiol and histidyl residues and on the susceptibility of the disulfide bonds to sulfitolysis. Modified albumins had, (i) a more cationic component which disappears by sulfitolysis of the disulfide bonds or by incubation with a glutathione redox system; (ii) higher reactivities of the free thiol group and of the histidyl residues, and; (iii) decreased fluorescence. These differences were not observed when processes were carried out on albumin with the thiol group blocked by iodacetic acid, but reappeared with the addition of cysteine. Renatured and aged albumins differed in the nature of the cationic component. Generation of albumin isoforms is dependent on the presence of a free thiol group and seems to involve thiol disulfide interchanges. # 2002 Elsevier Science B.V. All rights reserved. Keywords: Bovine albumin; Cathodic electrophoresis; Thiol-disulfide exchange
1. Introduction Albumin is a single chain multidomain protein with 17 disulfide bonds and one free thiol at Cys34, which partially forms mixed disulfides with cysteine or glutathione. Albumin undergoes pH dependent conformational changes in slightly alkaline solutions, the so called the neutral to basic transition, and isomerizes to the aged form at alkaline pH and low ionic strength. These changes in conformation and isomerizations determine a complex grade of microheterogeneity [1]. In contrast to the abundant literature concerning albumin heterogeneity, relatively little work has been done to assess the biological behavior of these albumin isoforms. A question to be answered is whether these different albumin molecules could form in vivo and, if so, whether they possess special functions. The basic transition has been proposed to have physiological significance in ligand transport [2] and in the regulation of the thiol-dependent * Tel.: /34-6-3862700x50870; fax: /34-6-1973018 E-mail address: [email protected] (M. Gabaldo´n).
antioxidant activity of albumin [3]. The aging reaction has been implicated in albumin catabolism, being the signal for selecting an albumin molecule for degradation [4]. There is sparse literature showing unexpected properties of albumin in biological systems. An albumin has been described with procoagulant activity which induces the expression of tissue factor in endothelial cells and monocytes [5], and regulates endothelial cell protein C activation, fibrinolysis [6] and prostacyclin secretion [7]. In previous works, we found that albumins obtained by heat shock produced an activation of the circulating leukocytes and an increase in leukocyte adhesion to the rat carotid endothelium [8,9]*/these proinflammatory activities not being related to differences in chemical composition. These results prompted us to study the effect of more subtle differences in the molecule by using isomeric forms of albumin. For this purpose we needed, first, to obtain albumin with changes in the conformation which were stable under physiological conditions and second, to establish criteria for distinguishing the native from the modified forms.
0141-8130/02/$ - see front matter # 2002 Elsevier Science B.V. All rights reserved. PII: S 0 1 4 1 - 8 1 3 0 ( 0 2 ) 0 0 0 4 0 - 5
260
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
In this paper, we describe the preparation of modified albumins by aging in the presence of cysteine and by renaturation after treatment with urea in non-reducing conditions. Both procedures produced more cationic albumins, which were separated from the native albumin by a new technique of cathodic electrophoresis. To distinguish native from modified albumins, we applied new criteria based on the reactivity of the thiol and histidyl residues and on the susceptibility of disulfide bonds to sulfitolysis. The modified albumins differed with respect to each other and the native form according to all applied criteria, and exhibited their properties under near physiological pH and ionic strength conditions, far different to those in which they were obtained. In order to explore the role of the free thiol group and of the thiol disulfide interchanges in the formation of albumin isoforms, we utilized albumin with the thiol group blocked with iodoacetic acid, and performed experiments in the presence of protein disulfide isomerase (PDI) and a glutathione redox system. Results indicate that generation of albumin isoforms is dependent on the presence of a free thiol group and seems to involve thiol disulfide interchanges.
2. Materials and methods
2.3. Urea denaturation and renaturation An 8 mg/ml albumin solution in 20 mM sodium phosphate buffer, final pH 7, and 8 M urea in 20 mM sodium phosphate buffer pH 7 were mixed in 1:3 proportion, and denaturation was carried out for 24 h at 4 8C. The final urea concentration was kept at 6 M because this concentration produced a more cationic albumin in experiments of isoelectric focusing in the presence of urea [10]. Although a 6 M concentration is not sufficient for complete denaturation, it is proximal to the midpoint of the second transition in which all the three domains of albumin undergo conformational changes [11,12]. Another reason for using 6 M urea was to facilitate its removal by dialysis, so avoiding a prolonged time of renaturation, which would increase the formation of albumin polymers. Renaturation was performed by dialysis against 20 mM sodium phosphate buffer pH 7 for 24 h at 4 8C with stirring. During this time three changes each with a buffer volume 50 times that of the albumin were carried out. Under these conditions urea remaining in the dialysate was less than 1 mg/100 ml. Renatured albumin was concentrated to 8 mg/ml by ultrafiltration through a Filtron membrane 30 K Omega type and filtered through a 0.45 mm filter. 2.4. Preparation of aged albumins
2.1. Chemicals Bovine serum albumin fraction V obtained by cold alcohol precipitation (A-4503), S -carboxymethyl albumin (A-6285), PDI (lot 48H4055), glutathione reduced form (G-6529), glutathione oxidized form disodium salt, sodium tetrathionate, sodium sulfite (S-4672), L-cysteine free base (C-7755), 5,5?dithio-bis (2-nitrobenzoic acid) (DTNB), diethylpyrocarbonate (DEPC) and 8-anilino1-naphthalenesulfonic acid (ANS) hemimagnesium salt were provided by Sigma Chemical Co. Fuchsin basic (47860) was provided by Fluka Chemika-Biochemika (Switzerland). 2.2. Preparation of buffered albumin solutions Depending on concentration, albumin solutions in buffers may not reach the desired pH due to the buffering capacity of albumin. As the pH affects the reaction rate in all the techniques described below, pH adjustment is critical for comparative purposes. When starting from buffered albumin solutions of different pHs, samples must be adjusted individually with one of the components of the buffer until the desired pH is reached, before completing volume with the adequate buffer. Accordingly, the pHs indicated in the text always refer to final pH of the albumin solutions.
A 50 mg/ml albumin solution was filtered through a 0.45 mm filter and 10 ml were desalted by passing through a Hi Prep† 26/10 desalting column (Amersham Pharmacia Biotech) equilibrated with H2O. The albumin was eluted with H2O, adjusted to pH 8.9 with 0.3 M Tris or 40 mM NaOH, and diluted to 20 mg/ml. Aging in the presence of cysteine was performed by adding Lcysteine free base to yield a final concentration of 1.5 mM (cysteine/albumin molar ratio, 5). Solutions were gassed with nitrogen and stored in the dark for 24 h at room temperature. 2.5. Cathodic PAGE Polyacrylamide gel electrophoresis was performed in native conditions at acid pH. In this type of electrophoresis, the lower electrode is the cathode and the upper is the anode, since proteins migrate as cations. Slab gels (100 /100/1 mm) were obtained by standard procedures. The buffer of the resolving gel was prepared as a stock solution (4 /) containing 120 mM KOH adjusted to pH 4 with acetic acid The composition of the resolving gel was as follows: acrylamide, 7.5%; bisacrylamide, 0.27% (T /7.8% and C /3.5%); buffer of the resolving gel (4 /), 25%; glycerol, 25% (v/v); TEMED, 3.7 ml/ml of the gel; 1.5% ammonium persulfate, 50 ml/ml of the gel. The buffer of the stacking gel was prepared as
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
a stock solution (2 /) containing 30 mM KOH adjusted to pH 4.4 with acetic acid. The composition of the stacking gel was as follows: acrylamide, 3.75%; bisacrylamide, 0.14% (T /3.9% and C /3.5%); buffer of the stacking gel (2 /), 50%; glycerol, 10% (v/v); TEMED, 4 ml/ml of the gel; 1.5% ammonium persulfate, 50 ml/ml of the gel. The electrophoresis buffer was: 100 mM balanine adjusted to pH 4.2 with acetic acid. The final concentrations in the sample buffer were: glycerol 20% (v/v); buffer of the stacking gel (2 /), 50%; tracking dye, 0.4% fuchsin basic, 25 ml/ml. Equal volumes of 1 mg/ml of albumin and sample buffer were mixed and 10 ml of this solution were loaded into the well. Gels were run by applying 50 V until the dye passed through the stacking gel. The voltage was then increased to 150 V and electrophoresis was continued until disappearance of the tracking dye. Gels were stained with colloidal Coomassie (ICN Biomedicals Inc.) for 2 h to overnight, rinsed with water and placed into the image enhancing solution (8% acetic acid (v/v), 25% ethanol (v/v)) for 10 min with gentle swirling. Gels were destained in 10% acetic acid (v/v). 2.6. Reactivity of the histidyl groups DEPC reacts with histidyl residues in albumin to yield a N -carbethoxyhistidyl derivative which absorbs at 240 nm. Stock 91 mM DEPC solutions were prepared in acetonitrile and stored at /20 8C. Since the purity of the commercial DEPC may be variable owing to hydrolysis, its concentration must be determined by reaction with imidazole [13]. Albumin solutions (1 mg/ml) were prepared in 0.1 M sodium phosphate buffer, final pH 6. Into a semimicrocuvette 1.18 ml of 1 mg/ml albumin were placed and autozeroed at 240 nm; 20 ml of 91 mM DEPC in acetonitrile were added and thoroughly mixed. Readings were started at 15 s and monitored for 7 min at 20 s intervals. A blank without albumin was processed in the same way. As DEPC is partially hydrolyzed during reaction, its absorption at t/0 was calculated from the reading of a blank, zeroed with 1.18 ml of acetonitrile, to which 20 ml of DEPC were added. The concentration of the N -carbethoxyhistidyl derivative was calculated from the absoption at 240 nm after subtracting the absorbance of the blank by using an extinction coefficient of 3.2 per mM/cm [13]. Determinations were performed in duplicate. As Tris accelerates the decomposition of DEPC, when reactivity of the histidyl groups is determined in albumin aged in the presence of Tris, this must be added to the controls in the same proportion as in the aged albumins. DEPC reacts with thiol groups and their presence must be taken into account when determining the reactivity of the histidyl groups. The interference produced by the cysteine used in the aging process is discrete and may be
261
corrected by using a blank with a final concentration of cysteine similar to that of the reaction mixture. The technique described cannot be applied in the presence of a glutathione redox system, since the higher concentration of thiol groups causes a notorious loss of DEPC.
2.7. Reactivity of the /SH group DTNB reacts with the thiol group in albumin to produce an equivalent of 5-thio-2-nitrobenzoate (TNB) which absorbs at 412 nm [14]. Albumin solutions (5 mg/ ml) were prepared in 40 mM sodium phosphate buffer, final pH 7; DTNB (0.7 mM) was prepared in 0.1 M Tris /HCl buffer-4.5 mM disodium EDTA, final pH 7. The spectrophotometer was autozeroed at 412 nm with a blank containing 0.8 ml of the DTNB solvent and 0.4 ml of the albumin solution. The reaction was started by adding 0.8 ml of DTNB to 0.4 ml of albumin. A blank without albumin was processed in the same way. The concentration of TNB was calculated from the absorption at 412 nm after subtracting the absorbance of the blank, assuming an extinction coefficient of 13.6 per mM/cm [14]. Determinations were performed in duplicate. Pseudo-first-order rate constants (kobs) were calculated by fitting the data to a single exponential.
2.8. Reactivity of the disulfide bonds Sulfitolysis is a reaction in which sulfite ions cleave disulfide bonds to S -sulfonates. The composition of the sulfitolysis reagent was as follows: Tris, 0.1M; Na2EDTA, 2.5 mM; sodium sulfite, 0.2 M; sodium tetrathionate, 32 mM; pH adjusted to 8 with HCl [15]. A blank was prepared without sodium sulfite and sodium tetrathionate. Equal volumes of albumin (2 mg/ml in 0.1 M sodium phosphate buffer, final pH 8) and sulfitolysis reagent or blank were mixed and incubated for 30 min. After reaction, aliquots were mixed with the same volume of sample buffer and subjected to cathodic PAGE. Sulfitolysis of accessible disulfide bonds increases the hydrodynamic volume and the negative charge of albumin, which produces a lesser mobility in cathodic electrophoresis.
2.9. Fluorescence measurements The fluorescent probe ANS was used to monitor conformational changes in albumin as it fluoresces strongly in non-polar environments such as the hydrophobic binding sites of albumin [16]. Intrinsic and ANS fluorescence were determined as described previously [9].
262
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
Fig. 1. Fluorescence spectra of aged and renatured albumins. (A) Intrinsic fluorescence A-4503, (1) control; (2) aged; (3) renatured; (4) aged-Cys. (B) ANS fluorescence A-4503, (1) control; (2) aged; (3) aged-Cys; (4) renatured. (C) Intrinsic fluorescence A-6285 (S -carboxymethylated), (1) control; (2) aged; (3) renatured; (4) aged-Cys. (D) ANS fluorescence A-6285 (S -carboxymethylated), (1) renatured; (2) aged; (3) control; (4) aged-Cys. The emission spectra were recorded in albumin dissolved in 60 mM sodium phosphate buffer pH 7, upon excitation at 280 nm (intrinsic fluorescence) and at 374 nm (ANS fluorescence). The ANS/albumin molar ratio was 6. Fluorescence is expressed in arbitrary units and the values are the mean of two independent determinations.
2.10. Analytical procedures Determination of the free thiol group in albumin was performed with the technique described for the study of the /SH group reactivity, except that albumin concentration was 6 mg/ml and pH for both albumin and reagents was 7.5. Absorbances were measured at 15/25 min. Fractionation of albumin was performed by gel filtration using Superdex 200 prep grade on a HiLoad 16/60 column (Amersham Pharmacia Biotech) as described previously [9]. The column was loaded with 1 ml of 20 mg/ml of albumin in the eluent buffer. A pool of dimeric and polymeric fractions from albumin A-4503 were rechromatographied, concentrated and used as electrophoretic and biochemical standards. 2.11. Reactions with PDI and glutathione PDI is a multifunctional enzyme which catalyzes the reduction and isomerization of misfolded proteins [17]. When reaction was carried out in the presence of PDI, the enzyme was first conditioned for 10 min in a
glutathione redox system by incubating 28 U of PDI in 0.5 ml of a solution containing 0.1 M sodium phosphate buffer, 2 mM reduced glutathione (GSH), 1.5 mM oxidized glutathione (GSSG) and 1 mM Na2EDTA, final pH 7.5. The final concentrations in the reaction mixture were: albumin, 5 mg/ml; GSH, 5 mM; GSSG, 3.75 mM; Na2EDTA, 1 mM; sodium phosphate buffer, 0.1 M; PDI, 19 U/ml; total volume, 1.5 ml; final pH, 7.5. Incubation was carried out for 4 h.
3. Results 3.1. Renatured albumin After 24 h in 6 M urea at 4 8C and pH 7, albumin A4503 was renatured during 24 h by dialysis against 20 mM sodium phosphate buffer pH 7. When albumin concentration was 2 mg/ml, this process determines a discrete increase in the dimeric and polymeric fractions and a concomitant decrease in the content of the free / SH group. Control: 78% monomer, 17% dimer, 5% polymer, and /SH/albumin molar ratio, 0.45. Rena-
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
tured albumin: 63% monomer, 24% dimer, 13% polymer and /SH/albumin molar ratio, 0.37. With an albumin concentration of 25 mg/ml, intensive polymerization occurs and the monomeric fraction decreases to 24%. Albumin A-6285 with the thiol group blocked by iodoacetic acid had 81% monomer, 15% dimer and 4% polymer, and these fractions were not modified after renaturation. Native and renatured albumin A-4503 differed in, (i) the fluorescence profile. The spectrum of renatured albumin after exciting at 280 nm (intrinsic fluorescence) showed a shift in the emission maximum from 340 to 335 nm (Fig. 1(A)). The decreases in intrinsic and ANS fluorescence observed in renatured albumin are shown in Fig. 1(A and B) and in Table 1; (ii) the electrophoretic pattern. Cathodic PAGE of renatured albumin showed new bands with greater mobility than the native monomer. These more cationic albumin isoforms were poorly resolved, producing a significant band broadening. The dimer also presented a new band with greater mobility (Fig. 2); (iii) the response to the sulfitolysis reagent. Native albumin was not affected after 30 min incubation. Renatured albumin lost the more cationic monomeric and dimeric bands and practically all the polymeric bands. These losses were not produced when sulfitolysis was performed at t /0, nor when the reagent was diluted 1/5. Renatured albumin showed a new narrow band at the beginning of the resolving gel which was absent in the native albumin (Fig. 2(A)); (iv) the reactivity of the histidyl groups. Histidyl groups in
Fig. 2. Cathodic PAGE of aged and renatured albumins. (A) Sulfitolysis of the disulfide bonds (A-4503), (1) control; (2) aged; (3) aged-Cys; (4) renatured; (5 /8), the same albumins after sulfitolysis for 30 min. (B) Effect of a glutathione redox system, (1) A-4503; (2) agedCys A-4503; (3) aged-Cys A-4503/glutathione; (4) renatured A-4503; (5) renatured A-4503/glutathione; (6) aged A-6285 (S -carboxymethylated); (7) aged-Cys A-6285; (8) aged-Cys A-6285/glutathione.
renatured abumin were more reactive than in native albumin (Table 2). Results were not subjected to a rigorous kinetic analysis because of two difficulties.
Table 2 Reactivity of the histidyl residues in aged and renatured albumins Albumin
Table 1 Effect of a glutathione redox system on intrinsic and ANS fluorescence Albumin
Intrinsic fluorescence (% control)
ANS fluorescence (% control)
A-4503 Dimer A-4503 Renatured A-4503 IdemPDI IdemGLUT IdemPDIGLUTa Aged A-4503 Aged-Cys A-4503 IdemGLUT A-6285b Renatured A-6285 Aged A-6285 Aged-Cys A-6285 IdemGLUT
100 88 80 85 88 90 86 78 93 100 100 100 82 89
100 79 52 57 65 68 82 72 99 100 104 100 77 90
Values were taken at the emission maximum upon excitation at 280 nm (intrinsic fluorescence) and at 374 nm (ANS fluorescence) and represent the mean of two independent determinations. a protein disulfide isomerase (PDI); glutathione redox system (GLUT). b A-6285 has the SH group blocked with iodoacetic acid.
263
N -carbethoxyhistidyl derivativea
Histidine/albumin molar ratiob
pH 6 A-4503 Dimer A-4503 Renatured A-4503 A-4503 Aged A-4503 Aged-Cys A-4503 A-6285c Aged A-6285 Aged-Cys A-6285 Renatured A-6285
13.7 16.8 30.0 15.9 25.4 36.5 16.6 16.9 26.9 18.8
8.3 9.0 11.0 8.2 9.4 10.5 8.1 8.1 9.0 8.4
pH 7 A-4503 Renatured A-4503 A-4503 Aged A-4503 Aged-Cys A-4503
34.5 60.4 33.9 46.2 64.2
12.1 13.4 12.0 12.5 13.1
Values represent the mean of two independent determinations. Concentration (mM) of the N -carbethoxyhistidyl derivative in the incubation mixture at 15 s, formed in the reaction of DEPC (1517 mM) with histidyl residues in albumin (15 mM). b Histidine/albumin (molar ratio) at 7 min. Bovine albumin has 17 histidyl residues. c A-6285 has the SH group blocked with iodoacetic acid. a
264
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
First, DEPC hydrolyzes partially during reaction and second, the alternative of increasing its concentration must be discarded because excess DEPC can react to form a disubstituted histidyl derivative with a higher absorbance at 240 nm [18]. At pH 6, the renatured albumin reacted more rapidly with DEPC, and had accessible for reaction 2.7 histidyl residues more than the native albumin. Since only the unprotonated form of histidine reacts with DEPC, the reaction is affected by external pH and by the particular pK of the each of the histidyl groups. When reaction is performed at pH 7, the responses of native and renatured albumin were different. Both increase the reactivity of the histidyl groups as a consequence of the greater amount of unprotonated histidine, but the increase in reaction rate was lower in renatured albumin, having accessible for reaction only 1.3 histidyl residues more than the native albumin; and (v) the reactivity of the free thiol group. The pseudofirst-order rate constants for the reaction of DTNB with the free /SH group of albumin at pH 7 were 64 and 996 (kobs /104, (per s)), for the native and renatured albumin, respectively. Since only the thiolate anion reacts with DTNB, the reaction is affected by external pH and by the pK of the thiol group. At pH 7.5, both albumins increased the reactivity of the /SH group as a consequence of the greater concentration of the thiolate anion, but their response was quite different as native albumin multiplied its rate constant by 5.7 while renatured albumin did so by only 1.3. In order to see whether differences observed between native and renatured albumin A-4503 were partly due to the excess of albumin dimer, we studied the fluorescence spectra and the reactivity of the histidyl residues in the pure dimer. We found that the dimer had 88% of the intrinsic and 79% of the ANS fluorescence of A-4503 (Table 1), without changes in the l of the emission maximum. The reactivity of histidyl residues in the pure dimer was discretely higher than in A-4503 (Table 2). As differences in the dimer content between native and renatured albumin were only of 7%, the contribution of the albumin dimer to the changes observed in renatured albumin may be considered negligible. When albumin had the thiol group blocked by iodoacetic acid (A-6285), the differences between native and renatured albumin disappeared. Neither the l of the emission maximum nor the intrinsic and ANS fluorescence were modified (Fig. 1 (C and D) and Table 1); the more cationic component was not detected by electrophoresis (data not shown), and the reactivity of histidyl residues was only discretely increased (Table 2). 3.2. Aged albumins The aging process of albumin A-4503 in the presence or absence of cysteine did not modify the content of dimers and polymers, or the free thiol content.
Native and aged albumins differed in: (i) the fluorescence profile, the greater changes being present in the albumins aged in the presence of cysteine. Spectra of aged albumins after exciting at 280 nm showed a blue shift from 340 to 337 nm (aged) or 335 nm (aged-cys) (Fig. 1(A)). The decreases in intrinsic and ANS fluorescence observed in aged albumins are shown in Fig. 1 (A and B) and Table 1. S -carboxymethyl albumin with the free thiol group blocked with iodoacetic acid (A-6285) showed changes in the fluorescence spectrum only when aged in the presence of cysteine (Fig. 1 (C and D) and Table 1; (ii) the electrophoretic pattern. Cathodic PAGE of aged albumins showed two new bands with greater mobility than the native monomer. These more cationic albumin isoforms were well resolved and were present in a greater amount in aged-cys albumins. The dimer also presented a new band with greater mobility (Fig. 2). S carboxymethyl albumin produced these forms only when aged in the presence of cysteine (Fig. 2(B)). Anodic native and SDS-PAGE failed to show any difference between native and aged albumins; (iii) the response to the sulfitolysis reagent. Upon sulfitolysis aged albumins lost the more cationic monomeric and dimeric bands and showed a new narrow band at the beginning of the resolving gel, which was absent in the native albumin (Fig. 2(A)); (iv) the reactivity of the histidyl groups. Histidyl groups in aged albumins were more reactive than in native albumin (Table 2). At pH 6, aged albumins had accessible for reaction 1.2 histidyl residues more than native albumin and 2.3 histidyl residues more when aging was performed in the presence of cysteine. At pH 7, the responses of native and aged albumins were different. Both increase the reactivity of the histidyl groups but the increase in reaction rate was lower in aged albumins. S -carboxymethyl albumin showed an increase in the reactivity of the histidyl groups only when aged in the presence of cysteine (Table 2); and (v) the reactivity of the free thiol group. The pseudo-first-order rate constants for the reaction of DTNB with the free /SH group of albumin at pH 7 were 64 and 144 (kobs /104, (per s)), for the native and aged albumin. At pH 7.5, both albumins increased the reactivity in the same proportion being the the kobs / 104, 365 and 740 (per s) for the native and aged albumin, respectively. 3.3. Effect of PDI and glutathione The incubation of the modified albumins with a glutathione redox system partially reverted the effects produced by aging and renaturation. The increase in intrinsic and ANS fluorescence of renatured albumin was higher with glutathione than with PDI (Table 1). Addition of glutathione produced a red shift in the emission maximum (intrinsic fluorescence), reaching a value of 338 nm, intermediate between that of the native
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
and modified albumins. The effect of PDI was discrete and may be due to the residual glutathione utilized for the enzyme activation. The reversal effect on the fluorescence intensities was more notorious in agedCys albumin A-4503 than in renatured albumin. The same occurred when they were analyzed by cathodic PAGE. In aged albumins the more cationic bands disappeared completely with glutathione, whereas in renatured albumin the bands remain and the possible variations were not apparent due to the broadening of the bands (Fig. 2(B)).
4. Discussion The renaturation of albumin denatured by urea depends on the experimental conditions utilized, the main determinants in this process being, (i) the existence of an oxidative environment if the starting molecule is a reduced and unfolded protein; (ii) the procedure of urea removal, by dilution (rapid) or dialysis (gradual); (iii) the albumin concentration and pH; and (iv) the presence of a free or blocked thiol group. The existence of multiple protocols, together with the different criteria applied to characterize the renatured albumin, explain the discrepancies found in the literature concerning the reversibility of denaturation upon urea removal [10,11,19,20]. Urea weakens hydrogen bonds and hydrophobic forces of the albumin mainly in domain III and to a lesser extent in domain I [12,19,21]. Albumin alkalinization produces the rupture of salt bridges between positively charged histidyl residues in domain I and anionic residues in domain III [2]. This loosening of the forces which maintain the tertiary structure in native albumin produces more open conformations with unmasking of ionizable groups whose normalization determines changes in the isoelectric point, which are higher in the urea denatured albumin (/1 pH unit) [10] than in aged albumin (/0.2 /0.4 pH units) [22,23]. The reversibility of these changes upon stimulus removal has not been well established and may not to be complete if a fraction of the molecules covalently stabilize their more open configuration with a thiol disulfide interchange. This possibility has been considered in aged albumin, suggesting a shuffling of one or two disulfide bonds in domain I [24]. In this work, we describe an electrophoretic technique which allows separation at a pH lower than the isoelectric point of albumin isoforms with a greater positive charge which are not resolved from the native form by anodic conventional electrophoresis. The presence of these more cationic isoforms in aged albumins is concordant with two of their characteristics, the greater proton binding below the isoionic point [24], and their higher isoelectric point [22,23]. The one or two more
265
cationic bands present in aged albumins may correspond to isomers formed by a restricted thiol disulfide interchange since, (i) the bands were not produced when the free thiol group of albumin was blocked with iodoacetic acid but reappeared on addition of cysteine, an external /SH donor acting as a catalyst; (ii) the bands disappeared on sulfitolysis, indicating an abnormal accessibility of the disulfide bonds to the reagent as the buried location of all 17 disulfide bonds in albumin makes them resistant to sulfitolysis in the absence of denaturing agents [15]; and (iii) the bands disappeared after incubation with a glutathione redox system which can produce a rearrangement of erroneous disulfide pairings following a pathway of sequential reactions involving mixed disulfides with glutathione, as has been described for the refolding of ribonuclease [25]. It has been proposed that prolonged exposure of albumin to urea results in intramolecular disulfide interchange [10]. The more profound conformational change produced by urea would allow a certain randomization of the disulfide interchange in contrast to the restricted shuffling of one or two disulfide bonds proposed in the aging process [24]. This would explain the heterogeneous pattern found by isoelectric focusing techniques, with 10/20 bands focused over the pH range between the isoelectric point of native and denatured albumin [10]. Our results are in accordance with this interpretation, since we found albumin isoforms with higher mobility in cathodic PAGE, consistent with a higher pI, and a poorly resolved banding pattern indicative of multiple intermediate states. The more cationic bands in renatured albumin disappeared upon sulfitolysis, which indicates an abnormal accessibility of the disulfide bonds. The occurrence of thiol disulfide interchanges in renatured albumin is also supported by the results obtained with S -carboxtmethylated albumin. In the absence of a free thiol group acting as a catalyst, the more cationic components were not produced and the fluorescence spectra were indistinguishable from those of the native form. The latter can mean that the more open conformations produced by urea are not stabilized covalently by a thiol disulfide interchange and that upon urea removal a complete renaturation allows achievement of the more folded conformation of the native form. In this work, we applied new criteria to distinguish native from modified albumins, based on the reactivity of the thiol and histidyl residues. The reactivity of both groups depends in part on steric factors, and an indirect approximation of the accessibility to the target groups is given by the fluorescence measurements [26]. The decreases in both intrinsic and ANS fluorescence observed in aged and renatured albumins point to a more open conformation than in native albumin, as these changes are produced by a higher exposure of fluorophors to solvent and by a loss of hydrophobicity,
266
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267
respectively. The reactivity depends also on the pK of the thiol and histidyl residues. A decrease in the thiol pK by a more cationic environment will produce a higher concentration of thiolate anion, the reactive form with DTNB, and a greater reactivity of the /SH group. The severance of salt bridges by alkalinization results in a deprotonation and in a lowering in pK of the histidyl residues [2]. This downward pK shift determines a higher concentration of the unprotonated form of histidine and, as a consequence, a higher reactivity */ since only the unprotonated form reacts with DEPC. The contribution of the pK changes to the reactivity of both groups may be explored by performing reactions at different pHs. Differences in reactivity which are pH dependent suggest pK changes in the groups of modified albumins. This is the case for renatured albumins whose thiol and histidyl residues seem to have a lower pK than the native albumin. The mechanism of ANS binding to albumin is still far from being completely understood despite its common use as a fluorescent hydrophobic probe. Neither the location of the five sites on albumin within which ANS is fluorescent [27] nor the microenvironment surrounding the binding sites are known. Binding of ANS to albumin comes from both the nonpolar anilinonaphthalene group (hydrophobic binding where intrinsic ANS fluorescence is not quenched by water) and the negatively charged sulfonate group (ionic binding which depends on the protein cationic charge and pH). Difficulties in interpreting results obtained by ANS are due to the fact that electrostatic as well as hydrophobic mechanisms contribute to the ANS /albumin interaction [28,29]. The modifications in tertiary structure produced by aging and renaturation may affect hydrophobicity in different ways. An open conformation is compatible with decreased ANS fluorescence when hydrophobic pockets are widened or change in position. In our study, we have found that modified albumins have decreased ANS fluorescence and greater reactivity of the thiol group, possibly due to a less steric restriction produced by the opening of the hydrophobic crevice which contains Cys34 [30]. The greater accessibility of the sulfitolysis reagent to the disulfide bonds in modified albumins may indicate that certain pockets have a greater facility for water penetration, which would quench the ANS fluorescence. Apart from a hydrophobic environment, ANS binding requires proximal cationic charges that may attract negatively charged ANS molecules. In our study, we have found that modified albumins had more reactive histidyl residues, which indicates a loss of positive charges since only the unprotonated form of histidine reacts with DEPC. In the event histidyl residues were in the microenvironment of any of the ANS binding sites, the decrease in ANS fluorescence in modified albumins
could also be explained by a reduction of the histidine positive charges capable of retaining negatively charged ANS in the surroundings of the hydrophobic binding site. Albumin carries out the transport and distribution of exogenous and endogenous ligands. The affinity of these ligands is dependent on the state of the neutral to basic equilibrium, which is in turn affected by pH and calcium concentration [2]. As these conditions vary in different tissues and organs, it is clear that albumin conformational changes may play a significant role in the pharmacokinetics of drugs. The preparation and characterization of albumin isoforms may help to better understand albumin functionality in vivo, and may be a useful tool for studying the anomalous behavior of certain albumins with procoagulant and proinflammatory activities [5,8,9].
Acknowledgements The author thanks Elena Carrascosa and Nuria Martin for technical assistance. This work was supported by a grant from Plan Nacional I/D No. SAF 96-0010.
References [1] Peters T, Jr. All about albumin. Biochemistry, genetics and medical applications. San Diego, CA: Academic Press, 1996. [2] Bos OJM, Labro JFA, Fischer MJE, Wilting J, Janssen LHM. J Biol Chem 1989;264:953. [3] Lee H, Cha MK, Kim IH. Arch Biochem Biophys 2000;380:309. [4] Wallevik KJ. Clin Invest 1976;57:398. [5] Faucette KJ, Parker CJ, McCluskey T, Bernshaw NJ, Rodgers GM. Blood 1992;79:2888. [6] Gubler DB, Wilson BD, Parker CJ, Rodgers GM. Thromb Res 1993;70:459. [7] Gubler DB, Ahlstrom CR, Liu L, Zhou JF, Parker CJ, Rodgers GM. Thromb Haemost 1995;74:1573. [8] Gabaldo´n M, Zu´n˜iga A, Capdevila C, Marque´s A. Atherosclerosis 1994;110:209. [9] Zu´n˜iga A, Marque´s A, Gabaldo´n M. Thromb Res 1997;86:243. [10] Salaman MR, Williamson AR. Biochem J 1971;122:93. [11] Muzammil S, Kumar Y, Tayyab S. Proteins Struct Funct Genet 2000;40:29. [12] Tayyab S, Sharma N, Khan MM. Biochem Biophys Res Commun 2000;277:83. [13] Miles EW. Methods Enzymol 1977;47:431. [14] Janatova J, Fuller JK, Hunter MJ. J Biol Chem 1968;243:3612. [15] Malhotra M, Sahal D. Int J Pept Protein Res 1996;48:240. [16] Damodaran S. Int J Pept Protein Res 1986;27:589. [17] Zheng J, Gilbert HF. J Biol Chem 2001;276:15747. [18] Roosemont JL. Anal Biochem 1978;88:314. [19] Khan MY, Agarwal SK, Hangloo S. J Biochemistry 1987;102:313. [20] Burton SJ, Quirk AV, Wood PC. Eur J Biochem 1989;179:379. [21] Tanaka N, Nishizawa H, Kunugi S. Biochim Biophys Acta 1997;1338:13. [22] Wallevik K. Biochim Biophys Acta 1976;420:42.
M. Gabaldo´n / International Journal of Biological Macromolecules 30 (2002) 259 /267 [23] Nikkel HJ, Foster JF. Biochemistry 1971;10:4479. [24] Stroupe SD, Foster JF. Biochemistry 1973;12:3824. [25] Ruoppolo M, Freedman RB, Pucci P, Marino G. Biochemistry 1996;35:13636. [26] Dockal M, Carter DC, Ru¨ker F. J Biol Chem 2000;275:3042.
267
[27] Daniel E, Weber G. Biochemistry 1966;5:1893. [28] Matulis D, Lovrien R. Biophys J 1998;74:422. [29] Matulis D, Baumann CG, Bloomfield VA, Lovrien RE. Biopolymers 1999;49:451. [30] Cornell CN, Kaplan LJ. Biochemistry 1978;17:1755.