Effect of redox state on unfolding energetics of heme proteins

Biochimica et Biophysica Acta 1432 (1999) 401^405 www.elsevier.com/locate/bba

Rapid report

E¡ect of redox state on unfolding energetics of heme proteins Pernilla Wittung-Stafshede * Chemistry Department, Tulane University, 6823 St. Charles Ave., New Orleans, LA 70118-5698, USA Received 8 April 1999; received in revised form 6 May 1999; accepted 6 May 1999

Abstract Both the enthalpic and entropic contributions to unfolding of three heme proteins, cytochrome b562 , cytochrome c and myoglobin, are larger for the reduced than for the oxidized form. Thus, the higher thermodynamic stability of a reduced, as compared to an oxidized, heme protein is the net result of a large increase of favorable enthalpy and a small increase in unfavorable entropy. Upon comparing the unfolding energetics of the heme proteins to those of other single-domain proteins I find that protein length is the primary determinant of the thermodynamics. ß 1999 Elsevier Science B.V. All rights reserved. Keywords: Protein folding; Spectroscopy; Energetics; Heme protein

Our understanding of the molecular details that underlay the stabilities of native proteins remains incomplete [1^4]. The thermodynamic stabilization of a folded protein can be attributed to many intramolecular forces (such as van der Waals interactions, interactions between charged groups, between polar groups and between non-polar groups) as well as to interactions with surrounding water. The main driving force for folding is generally believed to be formation of hydrophobic structures [5,6]. In addition, a cofactor may add to the stability of a protein; in case of a redox-active cofactor, the extent of stabilization depends on the reduction state [7,8]. For example, heme proteins are more stable towards unfolding when reduced than when oxidized [9]. Here I report thermodynamic characterization of unfolding of both reduction states of three heme proteins, cytochrome

* Fax: +1-504-865-5596; E-mail: [email protected]

b562 (cyt b562 ), with a four-helix-bundle structure [10,11], cytochrome c and myoglobin. Comparisons of the data to a spectrum of other single-domain proteins reveal that chain length is the major determinant for protein unfolding energetics, also in the case of heme proteins. Both denaturant- and heat-induced unfolding of oxidized cyt b562 are reversible processes [12,13]. I performed temperature-induced unfolding experiments at various GuHCl concentrations (inset Fig. 1A); at each GuHCl concentration a di¡erent midpoint temperature for the unfolding transition was detected (Fig. 1A). The transitions were analyzed in terms of two-state processes and the thermodynamic parameters corrected for their temperature dependencies [14] using a change in heat capacity (vCp ) of 3.8 kJ/mol K as determined by calorimetry for cyt b562 (see [13]: this paper reports a vH = 100 kcal/mol at Tm = 67³C; 33% higher than the vH we estimated at the same T (see Fig. 1B). The cyt b562 unfolding enthalpies and entropies increased linearly with the transition temperature [5] (Fig. 1B); by extrapola-

0167-4838 / 99 / $ ^ see front matter ß 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 7 - 4 8 3 8 ( 9 9 ) 0 0 1 1 6 - 8

BBAPRO 30441 5-7-99

402

P. Wittung-Stafshede / Biochimica et Biophysica Acta 1432 (1999) 401^405

Fig. 1. (A) The temperature of the transition midpoint (Tm ) for unfolding of oxidized cyt b562 is shown as a function of GuHCl concentration. (inset: CD signal (mdeg at 220 nm) versus temperature (³C)). Plasmid pNS207 containing the cyt b562 gene (S.G. Sligar, University of Illinois) was expressed in a BL21 strain of Escherichia coli and puri¢ed by anion exchange chromatograpy [10,12]. Thermal-unfolding experiments on oxidized cyt b562 were performed at various GuHCl concentrations (0^1.5 M GuHCl) by following the far-UV circular dichroism (CD) signal at 220 nm upon increasing the temperature. Steps of 2³C (20^90³C), 2 min equilibrium time at each temperature, 10 WM protein in 1-mm cell, Aviv 62A Spectropolarimeter. (B) vH(Tm ) for unfolding of oxidized cyt b562 is shown as a function of Tm . From each thermal unfolding curve, the value of the enthalpy change, vH(Tm ), at Tm was calculated from a modi¢ed van 't Ho¡ equation, correcting for temperature dependencies of vH and vS [14]. Plots of ln Kcorr versus 1/T gave straight lines with slopes of 3vH(Tm )/R, from which vH(Tm ) at each Tm was determined. At each Tm , ln K = 0 so that vS(Tm ) also could be calculated at each temperature.

tions I determined the unfolding enthalpy and entropy for oxidized cyt b562 to be vH = 155 þ 5 kJ/mol and vS = 475 þ 25 J/mol K at 20³C. From these values, an unfolding-free energy of 16 þ 7 kJ/mol for oxidized cyt b562 was calculated at 20³C, in excellent agreement with the value of 18 þ 2 kJ/mol determined from GuHCl-titrations at this temperature [12].

The entropy for reduction of folded cyt b562 has been determined to be 390 J/mol K at pH 7.0 [15]. Moreover, electrochemical studies (M.G. Hill, S. Zittin, personal communication) of the temperature dependence of the reduction potential of microperoxidase-8, a peptide fragment of cytochrome c, have suggested the entropy for reduction of an unfolded heme protein to be zero. Upon incorporating these

Table 1 Thermodynamic unfolding data for three heme proteins at 20³C Protein (length)

Redox state

vG (kJ/mol)

vH (kJ/mol)

vS (J/mol K)

Cytochrome b562 (106)

Ox Red Ox Red Ox Red

18a 43a 40 74 24 30

155 208 193c 245 223c 265

475 565 522d 584 697d 789

Cytochrome cb (104) Myoglobinb (153) a

Data from [12]. Cytochrome c, horse; myoglobin, whale. c Calculated from vH(Tm ) and vCp [23,24]. d Calculated from vH and vG [17,18] at 20³C. b

BBAPRO 30441 5-7-99

P. Wittung-Stafshede / Biochimica et Biophysica Acta 1432 (1999) 401^405

403

Fig. 2. Correlation between thermodynamics of unfolding (at 20³C) and protein length (number of amino acids) for 12 simple singledomain proteins (listed in Tables 1 and 2). vH as function of length (A), vS as function of length (B). For heme proteins, data points for unfolding of oxidized proteins are shown as ¢lled circles, reduced proteins are shown as open circles (directly above each other since length does not change). The correlation coe¤cients (r) and P-values (for linear ¢ts) were calculated using S-plus (MathSoft). The P-value measures the probability that a normally distributed data set, as large as that reported here would, by random chance, produce a slope with a magnitude as great or greater than that which we actually observe.

data in a thermodynamic cycle [7] that connects entropies for reduction of folded (390 J/mol K) and unfolded (0 J/mol K) proteins and entropies for unfolding of oxidized (475 J/mol K) and reduced forms, we could determine the entropic contribution to unfolding of reduced cyt b562 (vS = 565 J/mol K at 20³C). From the unfolding-free energy (vG = 43 kJ/ mol [12]) and the unfolding entropy, the unfolding enthalpy for reduced cyt b562 was then calculated to be vH = 208 kJ/mol. In Table 1, I present the thermodynamic parameters for unfolding of three heme proteins, cyt b562 , cytochrome c and myoglobin. I calculated the thermodynamic parameters for unfolding of oxidized cytochrome c and myoglobin at 20³C using vH(Tm ), vCp , and vG, as described in the footnotes of Table 1; the corresponding numbers for the reduced forms were calculated as for reduced cyt b562 , using vS = 362 J/mol K for reduction of folded cyt c and vS = 392 J/mol K for reduction of folded myoglobin [16]. Since the unfolding-free energies of reduced cytochrome c and myoglobin at 20³C have been reported [17,18], I could determine both entropies and enthalpies of unfolding of the reduced forms (Table 1). The heme cofactor stabilizes both folded

forms of cyt b562 , myoglobin, and cytochrome c; in addition, the reduced proteins are signi¢cantly more stable than the oxidized forms [12,17,18]. Only myoglobin can exist as a fully folded apoprotein; the other two proteins do not fold completely without heme. I ¢nd both the enthalpy and entropy of unfolding to be larger for the reduced proteins (Table 1); thus, the folded, reduced state of each of these heme proteins appears less dynamic and more hydrophobic around the heme cavity than the folded, oxidized protein (assuming similar unfolded structures). In accord, the reduced heme is charge-neutral and hydrophobic, whereas the oxidized form is cationic [8,9]. The two oxidation states of these heme proteins have essentially identical folded structures [9] and thermodynamic di¡erences should relate only to the local environment near the heme. The increased enthalpic stabilization of the charge-neutral heme in the proteins is signi¢cantly larger than the accompanying entropic penalty from decreased dynamics and, therefore, the overall stability (free energy) of the reduced proteins is larger than that of the oxidized proteins. Relationships between protein stability and protein

BBAPRO 30441 5-7-99

404

P. Wittung-Stafshede / Biochimica et Biophysica Acta 1432 (1999) 401^405

Table 2 Thermodynamic unfolding data for nine single-domain proteins at 20³C Protein (reference)

Length

vH (kJ/mol)

vS (J/mol K)

Protein L [25] Csp B [26] CI-2 [27] N-terminal domain L9 [28] Spectrin SH3 [29] Lysozyme [22] Ribonuclease A [22] Protein G [30] Paralbumin [31]

63 67 83 56 62 129 124 56 100

84 64 113 41 57 220 280 60 136

210 186 251 87 148a 530 790 119b 280

a b

Calculated from vH and vG at 25³C. Derived from Fig. 4 in [30].

structural features have been examined [21], though only at high temperatures. Reasonable correlations between protein length and vS and vH of unfolding were found at the convergence temperature (around 100³C) where hydrophobic contributions are thought to approach zero, as had been suggested earlier [22]. The data set in this study [21] included an impressive number of proteins, with lengths up to 365 residues and many containing disul¢des. In contrast, I here compare unfolding thermodynamics at 20³C for proteins that are all small single-domain proteins exhibiting two-state equilibrium unfolding curves, to place the heme-protein unfolding data in a wider perspective. In Table 2, the enthalpy and entropy of unfolding of nine small single-domain proteins are listed. Examination of the data in Tables 1 and 2 clearly suggests that there are correlations between the protein length (number of amino acids) and the thermodynamic parameters; in Fig. 2, I show linear correlations between protein chain length and the enthalpy (A) and entropy (B) of unfolding. The correlations are statistically highly signi¢cant, correlation coe¤cients (r) of 0.925, and P-values (P) of 0.0001 in both cases. Thus, among this set of single-domain proteins (including the heme proteins), the total number of amino acids is the primary determinant of the magnitude of the enthalpic and entropic contributions to unfolding. The di¡erences in enthalpies and entropies of the two redox states of the heme proteins appear small in these plots; both oxidized and reduced data ¢t equally well to the correlations (see Fig. 2; connected open and ¢lled symbols). Thus, it appears

that the heme moiety makes the heme protein behave `normal'; presence of the heme does not change the thermodynamics as compared to non-heme proteins. A linear correlation between unfolding-free energy (vG) and protein length is also observed (not shown, r = 0.780, P = 0.0046); this must occur since vG is a linear function of vH and vS. The correlation between vG and protein length suggests that approximately 40 residues is the minimum length required to form a stable native protein state. This is roughly consistent with both theoretical predictions (a 60 residues minimum size was derived from a statisticalenergy function approach [19]) and experimental observations; only a single stable protein domain of 6 40 residues has been reported (the 36-residue villin headpiece subdomain [20]). I also attempted to correlate the unfolding energetics with the size of the hydrophobic core across these data set of single-domain proteins. No statistically signi¢cant correlations were found upon plotting the number of hydrophobic amino acids (ranging from including only Ile, Leu, Ala, to including all of Ile, Leu, Ala, Val, Met, Phe in the proteins) against vH or vS of unfolding. To summarize, the higher stability of the reduced form of three heme proteins compared to the oxidized form is the net result of a more favorable enthalpy term (increased hydrophobic nature of a charge-neutral heme) and an increased negative entropy (decreased dynamics in a more hydrophobic heme cavity). In addition, I demonstrate that for 12 single-domain proteins (including the three proteins containing heme cofactors) the primary determinant

BBAPRO 30441 5-7-99

P. Wittung-Stafshede / Biochimica et Biophysica Acta 1432 (1999) 401^405

of their unfolding energetics is simply the size (length) of the protein. I am grateful to K. Plaxco, H.B. Gray and J.R. Winkler, for helpful discussions, and L. Plaxco for statistical analyses. References [1] K.A. Dill, H.S. Chan, Nat. Struct. Biol. 4 (1997) 10^19. [2] T.R. Sosnick, L. Mayne, S.W. Englander, Proteins Struct. Funct. Genet. 24 (1996) 413^426. [3] R.L. Baldwin, Fold. Design 1 (1996) R1^R8. [4] C.M. Dobson, A. Sali, M. Karplus, Angew. Chem. Int. Ed. 37 (1998) 868^893. [5] W. Kauzmann, Adv. Protein Chem. 14 (1959) 1^63. [6] G.I. Makhatadze, P.L. Privalov, Adv. Protein Chem. 47 (1995) 307^425. [7] J. Bixler, G. Bakker, G. McLendon, J. Am. Chem. Soc. 114 (1992) 6938^6939. [8] J.R. Winkler, P. Wittung-Stafshede, J. Leckner, B.G. Malmstro«m, H.B. Gray, Proc. Natl. Acad. Sci. USA 94 (1997) 4246^4249. [9] A. Schejter, in: R.A. Scott, A.G. Mauk (Eds.), Cytochrome c: A Multidisciplinary Approach, University Science Books, 1996, pp. 335^345. [10] H. Nikkila, R. Gennis, S.G. Sligar, Eur. J. Biochem. 202 (1991) 309^313. [11] K. Hamada, P.H. Bethge, F.S. Mathews, J. Mol. Biol. 247 (1995) 947^962. [12] P. Wittung-Stafshede, H.B. Gray, J.R. Winkler, J. Am. Chem. Soc. 119 (1997) 9562^9563. [13] C.R. Robinson, Y. Liu, R. O'Brian, S.G. Sligar, J.M. Sturtevant, Protein Sci. 7 (1998) 961^9650.

405

[14] P. Wittung-Stafshede, B.G. Malmstrom, D. Sanders, J.A. Fee, J.R. Winkler, H.B. Gray, Biochemistry 37 (1998) 3172^3177. [15] P.D. Barker, J.L. Butler, P. Olivera, A.O. Hill, N.I. Hunt, Inorg. Chim. Acta 252 (1996) 71^77. [16] V.T. Taniguchi, N. Saliasuta-Scott, F.C. Anson, H.B. Gray, Pure Appl. Chem. 52 (1980) 2275^2281. [17] P. Wittung-Stafshede, B.G. Malmstrom, J.R. Winkler, H.B. Gray, J. Phys. Chem. A 102 (1998) 5599^5601. [18] G.A. Mines, T. Pacher, S.C. Lee, J.R. Winkler, H.B. Gray, Chem. Biol. 3 (1996) 491^497. [19] D. Xu, R. Nussinov, Fold. Design 3 (1997) 11^17. [20] C.J. McKnight, D.S. Doeing, P.T. Matsudaira, P.S. Kim, J. Mol. Biol. 260 (1996) 126^134. [21] A.D. Robertson, K.P. Murphy, Chem. Rev. 97 (1997) 1251^ 1267. [22] P.L. Privalov, N.N. Khechinashvili, J. Mol. Biol. 86 (1974) 665^684. [23] S. Potekhin, W. Pfeil, Biophys. Chem. 34 (1989) 55^62. [24] P.L. Privalov, Y.V. Griko, S.Y. Venyaminov, V.P. Kutyshenko, J. Mol. Biol. 190 (1986) 487^498. [25] M.L. Scalley, D. Baker, Proc. Natl. Acad. Sci. USA 94 (1997) 10636^10640. [26] T. Schindler, F.X. Schmid, Biochemistry 35 (1996) 16833^ 16842. [27] Y.-J. Tan, M. Oliveberg, A.R. Fersht, J. Mol. Biol. 264 (1996) 377^389. [28] B. Kuhlman, D.P. Raleigh, Protein Sci. 7 (1998) 2405^2412. [29] A.R. Viguera, J.C. Martinez, V.V. Filimonov, P.L. Mateo, L. Serrano, Biochemistry 33 (1994) 2142^2150. [30] P. Alexander, S. Fahnestock, T. Lee, J. Ovban, P. Bryan, Biochemistry 31 (1992) 3597^3603. [31] P.L. Privalov, S.J. Gill, Adv. Protein Chem. 39 (1988) 191^ 234.

BBAPRO 30441 5-7-99