HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: ps.04817704v1 13/9/2316    most recent Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Cowley, A. B. Articles by Benson, D. R. Search for Related Content PubMed PubMed Citation Articles by Cowley, A. B. Articles by Benson, D. R. Social Bookmarking                   What's this?

Stabilizing roles of residual structure in the empty heme binding pockets and unfolded states of microsomal and mitochondrial apocytochrome b₅

Department of Chemistry, University of Kansas, Lawrence, Kansas 66045-7582, USA

Reprint requests to: David R. Benson, Department of Chemistry, University of Kansas, 1251 Wescoe Hall Drive, 2010 Malott Hall, Lawrence, KS 66045-7582, USA; e-mail: drb{at}ku.edu; fax: (785) 864-5396.

(RECEIVED April 18, 2004; FINAL REVISION June 3, 2004; ACCEPTED June 4, 2004)

	Abstract

TOP Abstract Introduction Results Discussion Materials and methods References

 
The microsomal (Mc) and mitochondrial (OM) isoforms of mammalian cytochrome b₅ are the products of different genes, which likely arose via duplication of a primordial gene and subsequent functional divergence. Despite sharing essentially identical folds, heme-polypeptide interactions are stronger in OM b₅s than in Mc b₅s due to the presence of two conserved patches of hydrophobic amino acid side chains in the OM heme binding pockets. This is of fundamental interest in terms of understanding heme protein structure–function relationships, because stronger heme–polypeptide interactions in OM b₅s in comparison to Mc b₅s may represent a key source of their more negative reduction potentials. Herein we provide evidence that interactions amongst the amino acid side chains contributing to the hydrophobic patches in rat OM (rOM) b₅ persist when heme is removed, rendering the empty heme binding pocket of rOM apo-b₅ more compact and less conformationally dynamic than that in bovine Mc (bMc) apo-b₅. This may contribute to the stronger heme binding by OM apo-b₅ by reducing the entropic penalty associated with polypeptide folding. We also show that when bMc apo-b₅ unfolds it adopts a structure that is more compact and contains greater nonrandom secondary structure content than unfolded rOM apo-b₅. We propose that a more robust -sheet in Mc apo-b₅s compensates for the absence of the hydrophobic packing interactions that stabilize the heme binding pocket in OM apo-b₅s.

Keywords: apocytochrome b₅; mitochondrion; endoplasmic reticulum; functional divergence; unfolding thermodynamics; m values; structure in the unfolded state; heme recognition

Article published online ahead of print. Article and publication date are at http://www.proteinscience.org/cgi/doi/10.1110/ps.04817704.

	Introduction

TOP Abstract Introduction Results Discussion Materials and methods References

 
Two membrane-anchored isoforms of the electron transfer heme protein cytochrome b₅ have been identified in mammals, one associated with the endoplasmic reticulum (ER) (microsomal, or Mc b₅), the other with the outer mitochondrial membrane (OM b₅) (Lederer et al. 1983; Vergeres and Waskell 1995; Kuroda et al. 1998; Soucy and Luu-The 2002; Ogishima et al. 2003). Mc and OM b₅ are the products of different genes, and phylogenetic analysis has revealed that they are more closely related to one another than to the lone b₅ gene product in plants, insects, and fungi (Guzov et al. 1996). On this basis, it was proposed that Mc and OM b₅ arose via duplication of an ancestral gene and subsequent functional divergence. Supporting this hypothesis is the recent discovery that plant b₅, long known to reside in the ER membrane, is also distributed to the outer mitochondrial membrane (Zhao et al. 2003). Studies in our laboratory with recombinant proteins representing the soluble heme-binding domains of rat OM (rOM) b₅ (Rivera et al. 1992, 1994; Silchenko et al. 2000) and human OM b₅ (Altuve et al. 2004) have revealed that functional divergence has led to substantial differences in the biophysical properties of the two mammalian b₅ isoforms. Nonetheless, the Mc and OM b₅ heme binding domains adopt essentially identical folds comprising two hydrophobic cores separated by a five-stranded -sheet as shown in Figure 1 (Durley and Mathews 1996; Rodriguez-Maranon et al. 1996).

View larger version (49K): [in this window] [in a new window]   Figure 1. (A) View of the X-ray crystal structure of rat OM b5 (PDB 1B5M [PDB] ; Rodriguez-Maranon et al. 1996) illustrating the packing interactions involved in one of the two conserved hydrophobic clusters described in the text. (B) The same structure rotated by ~180° to show the packing interactions involved in the second cluster. (C) The corresponding structure of bovine Mc b5 (PDB 1CYO [PDB] ; Durley and Mathews 1996), highlighting the two hydrophobic cores and the elements of secondary structure.

(1)

sequesters hemin (ferric heme) from bMc b₅ ~130-fold more rapidly (k_–H = 6.5 h^–1) than from rOM b₅ (k_–H ~0.05 h^–1) at 37°C (pH 5.0). The higher kinetic barrier for hemin release from OM b₅ relative to Mc b₅ also results in slower equilibration of the two heme orientational isomers that are formed in equal amounts upon initial binding of the prosthetic group, differing by a 180° rotation about the porphyrin ––meso axis. Equilibration to the thermodynamic ratio of isomers in Mc b₅s is complete within a few hours at 37°C (pH 7.0), but only proceeds at detectable rates above ~50°C in OM b₅s. Unfolding of OM and Mc b₅ polypeptides by heat is also accompanied by dissociation of hemin. Thus, OM b₅s have higher thermal denaturation midpoints (T_m = 86°C for rOM b₅) than Mc b₅s (T_m = 67°C for bMc b₅).

The higher activation barrier for hemin dissociation from rOM b₅ in comparison to bMc b₅ is consistent with stronger hemin binding (G_H = k_H/k_–H; equation 1), which could reflect energetically more favorable interactions between hemin and the rOM apo-b₅ binding pocket, greater stability of the rOM apo-b₅ fold (G_NU in equation 2), or a combination of the two factors.

(2)

Evidence favoring the first possibility was obtained in guanidinium chloride (GdmCl)-mediated denaturation studies, which showed that bMc and rOM apo-b₅ are folded at 25°C (pH 7.0), and exhibit nearly identical unfolding free energies under those conditions (G_NU ~3 kcal/mole) (Silchenko et al. 2000; Cowley et al. 2002). In addition, comparison of the bMc and rOM b₅ amino acid sequences (see Fig. 2) in the context of their X-ray crystal structures allowed us to identify two conserved hydrophobic clusters in OM b₅ heme binding pockets (core 1) that seemed likely to favor stronger interactions with the prosthetic group than the corresponding packing motifs in their Mc counterparts (Altuve et al. 2001, 2004; Cowley et al. 2002). One cluster (Fig. 1A) brings together the side chains of Leu-47 in 3 and Leu-36 in adjacent helix 2, with additional interactions between the Leu-36 side chain and that of Ile-32 in 3, and between the Leu-47 side chain and that of Ala-18 in a loop between 1 and 4. The second cluster (Fig. 1B) involves interactions among the side chains of Phe-58 in 4, Leu-71 in adjacent helix 5 and Ile-25 in 4, all of which also make contact with heme. The corresponding residues in all known Mc b₅s (Ser-18, Leu-25, Leu-32, Leu-36, Arg-47, Phe-58, and Ser-71) have but two hydrophobic side-chain contacts, between Arg-47 (3) and Leu-36 (2), and between Phe-58 (4) and Leu-25 (4).

View larger version (19K): [in this window] [in a new window]   Figure 2. Amino acid sequences of the recombinant bMc (Funk et al. 1990) and rOM b5 (Rivera et al. 1992) polypeptides used in this study (63% identity/82% similarity based on residues 3–84). Numbering is based on the scheme introduced by Mathews for the lipase fragment of bovine Mc b5 (Mathews et al. 1979). Italicized residues in the rOM b5 sequence are disordered in the X-ray crystal structure.

	Results

TOP Abstract Introduction Results Discussion Materials and methods References

 
Expression and heme extraction
The synthetic gene coding for rOM b₅ (Rivera et al. 1992) is under control of the strong T7 promoter in pET11a, and when expressed in Escherichia coli BL21(DE3) cells at 37°C was found to yield 7.7 mg/L of holoprotein (Table 1). The analogous expression system recently developed (Cowley et al. 2002) for the synthetic bMc b₅ gene (Funk et al. 1990) produced 1.5 mg/mL of holoprotein at this temperature, and similarly low yields were obtained for rOM^5M holo-b₅ (Table 1). Holoprotein yields were improved in all cases when expressions were performed at 27°C, with the most substantial increase observed for bMc holo-b₅. The holoproteins were purified to homogeneity as assessed by native-PAGE (Fig. 3, lanes 1–3), and the corresponding apoproteins were obtained by extraction of heme using the acid-butanone method (Teale 1959). All three apoproteins migrate slowly in comparison to the corresponding holoproteins on native gels (Fig. 3, lanes 4–6), due in part to loss of the two negative charges from the heme propionates. Migration of rOM^5M apo-b₅ is further slowed by the fact that it is mostly unfolded at room temperature (vide supra), and therefore, has a larger average size than wild-type rOM apo-b₅.

View larger version (85K): [in this window] [in a new window]   Figure 3. Native PAGE (18%) data for bMc b5, rOM b5, and rOM5M b5. Lanes 1–3, pure holoproteins; lanes 4–6, pure apoproteins; lanes 7–9, supernatants from analytical expressions at 37°C.

Circular dichroism spectroscopy
Far-UV and near-UV circular dichroism (CD) spectra of the holo and apo forms of bMc and rOM b₅ recorded at 25°C (pH 7.0), are compared in Figure 4. Mean residue ellipticity values in all far-UV spectra have been calculated based on the length of the bMc b₅ polypeptide (82 residues), even though the rOM b₅ polypeptide contains 92 residues. This was done to facilitate comparison of secondary structure content based on signal intensity at 220 nm, as a comparison of the bMc and rOM b₅ X-ray crystal structures revealed the same number of residues in regions of well-defined secondary structure in the two proteins (beginning with invariant Val-4 and ending with residue 84). The seven N-terminal and three additional C-terminal residues that render rOM b₅ longer than bMc b₅ are in extended conformations or could not be detected in the diffraction data (Rodriguez-Maranon et al. 1996), and therefore are expected to make little if any contribution to ellipticity at 220 nm (far-UV CD spectra of random coil peptides are dominated by a strong signal of negative ellipticity centered below 200 nm; Compton and Johnson 1986).

View larger version (31K): [in this window] [in a new window]   Figure 4. Far-UV CD spectra of bMc b5 (A) and of rOM b5 (B). Holo, 25°C (dashed line); apo, 25°C (solid line); apo, 75°C (bold line). Corresponding near-UV CD spectra are shown in C (bMc b5) and D (rOM b5). All data were recorded at pH 7.0 in 50 mM potassium phosphate.

Chemical denaturation studies
In our previously reported GdmCl-mediated denaturation studies of bMc and rOM apo-b₅, changes in fluorescence intensity of the Trp-22 side chain were monitored (Cowley et al. 2002). The evidence presented above suggesting that Trp-22 in bMc and rOM apo-b₅ is located in a well-ordered region of tertiary structure (core 2) confirms its utility as a spectroscopic probe for monitoring unfolding reactions. bMc and rOM apo-b₅ begin to unfold at rather low GdmCl concentrations, which resulted in poorly defined baselines in the native (folded) region of the reported curves. As a result, we cautioned that unfolding free energies obtained from fits of those data should be considered as estimates. As an approach to broadening the native baselines and thereby obtaining more reliable data, we performed analogous unfolding reactions using the milder denaturant urea (Fig. 5). We also included in these studies the quintuple mutant rOM^5M apo-b₅, finding once again that it is only about 25% folded in aqueous solution at 25°C (the fraction of mutant apoprotein folded at each temperature was estimated based on the assumption that the fluorescence intensity of the Trp-22 side chain in rOM^5M apo-b₅ will undergo a change similar to that of wild-type rOM apo-b₅ upon denaturation).

View larger version (18K): [in this window] [in a new window]   Figure 5. Urea-mediated denaturation curves for bMc apo-b5 (filled squares), rOM apo-b5 (filled circles), and rOM5M apo-b5 (open circles). The experiments were performed at 25°C (pH 7.0) (50 mM potassium phosphate). Solid lines represent fits of the data to equation 2.

(3)

(4)

where [D] is the urea concentration, G_NU is the free energy of unfolding in the absence of denaturant, and m is the sensitivity of the unfolding free energy to denaturant concentration (Pace and Scholtz 1997). The concentration of denaturant at which each protein is 50% unfolded (C_m value) was determined using equation 4. The results of these fits support our previously reported conclusion (Cowley et al. 2002) based on GdmCl data that rOM and bMc apo-b₅ exhibit essentially identical unfolding free energies at 25°C (pH 7.0) (Table 2). The fact that the neutral denaturant urea and the ionic GdmCl yield essentially identical G_NU values for each apoprotein further indicates that electrostatic interactions do not make an unduly large contribution to stability in either case (Monera et al. 1994). A given increment in urea or GdmCl concentration causes a larger increase in the ratio of unfolded to folded rOM apo-b₅ than of bMc apo-b₅ (see Fig. 5), as reflected by the larger negative slopes (m values) of the rOM apo-b₅ unfolding curves. As a consequence, a much higher fraction of bMc apo-b₅ than of bMc apo-b₅ remains in the folded state at each denaturant concentration, resulting in much higher C_m values (see equation 4). The larger m values of rOM apo-b₅ indicate a larger change in accessible surface area (ASA) upon unfolding in comparison to bMc apo-b₅ (Myers et al. 1995). The larger m value of rOM apo-b₅ is not likely simply due to its greater length, as the "extra" residues at its termini are disordered and not involved in packing interactions, and therefore not likely to experience much change in exposed surface area upon unfolding. Furthermore, the m value obtained for bMc apo-b₅ in our GdmCl-mediated denaturation studies is remarkably similar to the corresponding value reported for a 103-residue fragment of the same protein (Table 2; Manyusa and Whitford 1999).

Thermal denaturation of bMc and rOM apo-b₅ is reversible, as demonstrated by the fact that far-UV CD spectra recorded at 25°C prior to and subsequent to each experiment are identical. Further evidence for reversibility was obtained in DSC experiments (Fig. 6; Table 3), which yielded identical results in the unfolding and refolding directions, as well as upon subsequent unfolding of the same sample. In addition, temperature-dependent changes in far-UV CD spectra occurred with an isodichroic point near 201 nm, also suggestive of a cooperative two-state unfolding equilibrium. Additional evidence in support of cooperativity is provided by the absence of significant temperature-dependent changes in far-UV CD spectra at any wavelength, between 5°C and the beginning of each unfolding transition region, and from the end of the transition region to 80°C, the highest temperature examined in each case. Figure 7 shows thermal denaturation curves for bMc, rOM, and rOM^5M apo-b₅ using data from a second set of experiments in which the time-averaged value of ₂₂₀ was recorded. Table 3 reports enthalpies of unfolding for bMc and rOM apo-b₅ referenced to the thermal denaturation midpoint [H(T_m)], obtained by fitting the thermal denaturation data to a two state model (Constans et al. 1998) that combines linear equations relating the fraction of protein in the folded and unfolded state at each temperature with an expanded form of the Gibbs-Helmholtz relation (equation 5), and accounts for nonzero slopes in the native and denaturing regions. Heat capacity changes derived from DSC data (C°_p values; Table 3) were used in all fits. Table 3 also reports entropies of unfolding referenced to the T_m value [S(T_m)], determined using equation 6. Absence of a plateau in the native region precluded fitting of the rOM^5M apo-b₅ denaturation curve in Figure 7.

View larger version (14K): [in this window] [in a new window]   Figure 6. Differential scanning calorimetry (DSC) data for rOM apo-b5 (1.87 mg/mL) (A) and bMc apo-b5 (2.04 mg/mL) (B) at pH 7.0 (50 mM potassium phosphate). The solid line in each panel indicates the observed data, and the dashed line is the fit of the data to an equation representing a two-state transition.

View larger version (19K): [in this window] [in a new window]   Figure 7. Thermal denaturation curves for bMc apo-b5 (filled squares), rOM apo-b5 (filled circles), and rOM5M apo-b5 (open circles). The experiments were performed at 25°C (pH 7.0)(50 mM potassium phosphate).

(5)

(6)

The corresponding DSC data were analyzed via a statistical mechanics-based deconvolution to obtain the calorimetric enthalpy change at T_m (H_cal), and via a nonlinear least-squares fit to a two-state model to obtain van’t Hoff enthalpies of unfolding at T_m (H_vH) (Table 3). H_cal and H_vH differed by 2% or less in each case, and by 3% or less from H (T_m) determined by CD, lending particularly strong support to a hypothesis that thermal unfolding of each apoprotein occurs via a cooperative two-state process (Zhou et al. 1999).

The thermal denaturation and DSC data in Table 3 reveal that bMc apo-b₅ is more enthalpically stable than rOM apo-b₅, but that it unfolds with a significantly larger entropy increase. Extrapolation of the bMc and rOM apo-b₅ denaturation data using equation 5 reveals that the two proteins consequently share almost identical thermodynamic stability at 25°C (Table 3), consistent with the results of chemical denaturation studies. Importantly, the CD and DSC data are consistent in showing that neither bMc apo-b₅ nor rOM apo-b₅ undergoes cold denaturation. In this context, it is worth noting that stock solutions of bMc apo-b₅ maintained at 4°C (pH 7.0), form noticeable amounts of precipitate even after a few days, whereas little if any precipitate is formed under these conditions in solutions containing similar rOM apo-b₅ concentrations. The bMc apo-b₅ remaining in solution is compromised over time, as well, as evidenced by changes in its CD spectrum and thermal denaturation curve. In contrast, CD spectra and thermal denaturation curves recorded for rOM apo-b₅ immediately following its preparation and after the solution has been stored at 4°C for up to three weeks are identical.

Dynamic light scattering studies
DLS data were recorded for the holo and apo forms of bMc and rOM b₅ at 25°C (pH 7.0), as well as of the apoproteins denatured by 8 M urea. Size distribution plots for folded and urea-denatured bMc and rOM apo-b₅ are compared in Figure 8, and hydrodynamic radii (R_H values), polydispersity values (a measure of size distribution) and diffusion constants obtained from all of the experiments are compiled in Table 4. The R_H values obtained for bMc and ROM apo-b₅ are consistent with expectations based on the lengths of the proteins as determined by crystallography (~37 Å in the case of bMc b₅; Durley and Mathews 1996). Removal of hemin from bMc b₅ is accompanied by a 9.8% increase in R_H, similar to a previously reported value obtained for the 93-residue lipase fragment (9%) in sedimentation studies (Huntley and Strittmatter 1972), as well as by a 36% increase in polydispersity. A considerably smaller increase in R_H (2.8%) and polydispersity (13%) results when heme is removed from rOM b₅. Interestingly, the data reveal a much smaller increase in R_H and polydispersity when bMc apo-b₅ unfolds (2.2-fold and 2.0-fold, respectively) than when rOM apo-b₅ unfolds (3.0-fold and 3.1-fold, respectively).

View larger version (25K): [in this window] [in a new window]   Figure 8. DLS size distribution plots for rOM apo-b5 (black bars) and bMc apo-b5 (white bars). (A) 0 M urea; (B) 8M urea. Data were recorded at 25°C (pH 7.0) (50 mM potassium phosphate), with protein concentrations of 100 µM.

	Discussion

TOP Abstract Introduction Results Discussion Materials and methods References

 
Consequences of evolutionary divergence of OM and Mc apo-b₅ on structure in the folded state
The heme prosthetic group occupies a sizeable volume in the interior of cytochrome b₅, engaging in a variety of specific (His-Fe bonds; hydrogen bonds to heme propionates) and nonspecific (hydrophobic) interactions with the polypeptide. As initially demonstrated in studies with the 93-residue lipase fragment of bMc b₅ (Huntley and Strittmatter 1972), and shown to be the case for the 82-residue bMc b₅ tryptic fragment in the studies presented herein, disruption of these interactions by heme extraction results in a substantial loss of secondary and tertiary structure and an expansion of the polypeptide fold. Greater insight into the structure of bMc apo-b₅ can be gleaned from solution NMR studies of rat Mc apo-b₅ (Falzone et al. 1996). Core 2 of rat Mc apo-b₅ adopts a mostly native-like fold, while the only residual native-like structure in core 1 is the irregular helix 2 (residues 33–38) and the final turn of helix 5. Helices 3 and 4 are completely disordered, exhibiting motional parameters similar to those of unfolded or partly folded proteins (Bhattacharya et al. 1999). As a consequence, 5 (intervening between 3 and 4) only makes occasional contact with 4 (Falzone et al. 2001).

The CD and DLS data presented herein indicate that heme removal from rOM b₅ disrupts core 1 polypeptide conformation to an even greater extent than in bMc b₅, but is accompanied by a much smaller expansion of the local fold and a smaller increase in dynamic motion. Greater compactness of rOM apo-b₅ in comparison to bMc apo-b₅ is further highlighted by its faster migration on native-PAGE, which contrasts with the nearly identical electrophoretic mobilities of the corresponding holoproteins (see Fig. 3). We interpret the greater compactness of rOM apo-b₅ to indicate that the conserved hydrophobic patches that stabilize rOM holo-b₅ in comparison to bMc holo-b₅ (Fig. 1A,B) exert a similar stabilizing effect when heme is absent. The apparently lower secondary structure content in core 1 of rOM apo-b₅, suggested by the CD data, may reflect that the most stable arrangement among the participating side chains is less compatible with native conformations for 2–5 than the corresponding, more loosely organized packing in bMc apo-b₅ that allows 2 and one turn of 5 to remain folded.

The results of the present study comparing bMc and rOM aqo-b₅, considered in the context of previously reported computational studies of the corresponding holoproteins, provide improved insight into the markedly divergent stability and heme-binding properties exhibited by the two mammalian b₅ isoforms. Substantial polypeptide conformational mobility was observed near the interface of core 1 and core 2 of bMc holo-b₅ in two independent molecular dynamics (MD) simulations, including the formation of a large cleft on the protein surface that provided access of water to the protein interior (Storch and Daggett 1995; Altuve et al. 2001; Lee and Kuczera 2003). Lower conformational flexibility in the corresponding region of rOM holo-b₅, a consequence of its more extensive hydrophobic packing interactions, prevented formation of an analogous cleft (Altuve et al. 2001; Lee and Kuczera 2003). The higher kinetic barriers for hemin transfer and hemin isomer equilibration exhibited by rOM b₅ in comparison to bMc b₅ can therefore be attributed to a combination of interrelated factors arising from more extensive hydrophobic packing, including: (1) less frequent dissociation of one or both axial His ligands in OM b₅s because conformational changes such as cleft formation that can trigger the process will occur less frequently; (2) less effective protonation of a dissociated His ligand, and less effective competition with it for coordination to heme iron due to a lower internal water content; and (3) more rapid and more efficient religation following dissociation of one or both His ligands due to less extensive melting of local secondary structure. In other words, there may not only be fewer opportunities, and less time during any given opportunity for heme to depart from core 1 of rOM b₅ than of bMc b₅, there may also be less room for the prosthetic group to turn around and reseat in the opposite orientation, or to be captured by apoMb following a given His dissociation event. Stated from a more thermodynamic standpoint (on the as yet untested assumption that hemin will bind to the two apoproteins with similar rate constants), the results of this study support a hypothesis that the hydrophobic clusters in OM b₅ render binding of heme more entropically favorable in comparison to Mc b₅ by diminishing the extent of core 1 conformational flexibility when heme is absent. It is worth noting in this context that increased binding pocket preorganization (Zartler et al. 2001) and the presence of more extensive hydrophobic packing interactions (Yamamoto et al. 2002) have also been shown to stabilize metalloproteins from thermophilic organisms in comparison to their analogues from mesophiles.

It is interesting to consider the implications of the above considerations for the specialized roles played by Mc and OM b₅ in the cell. For example, it has been proposed that conformational flexibility in Mc b₅, including formation of a surface cleft, may be important for function because it increases the diversity of potential substrate recognition motifs (Storch and Daggett 1995; Dangi et al. 1998). Lower conformational mobility might thus facilitate OM b₅ function by enabling more specific substrate recognition (Altuve et al. 2004), with stronger heme–polypeptide interactions an interesting but functionally unimportant consequence. Evolution of stronger heme–polypeptide interactions in OM b₅ in comparison to Mc b₅, with lower conformational mobility as the consequence, could alternatively represent part of nature’s solution to rendering OM b₅ reduction potentials more negative without need for substantial changes in exposure of heme to solvent or in its electrostatic environment (Altuve et al. 2004). This interpretation was suggested by studies of two helical heme protein models, one less conformationally dynamic than the other as a consequence of enhanced side chain–heme packing interactions (Liu et al. 1999). For reasons that are not yet clear, the decrease in conformational flexibility resulting from these interactions stabilized His-Fe(III) coordination to a greater extent than His-Fe(II) coordination, leading to a negative shift in reduction potential (~60 mV; Kennedy et al. 2001) similar in magnitude to what is observed for OM b₅ versus Mc b₅ (40 mV; Altuve et al. 2004).

Consequences of evolutionary divergence of OM and Mc apo-b₅ on structure in the unfolded state
As discussed in the previous section, bMc apo-b₅ differs from rOM apo-b₅ in exhibiting a more extended and more conformationally mobile empty heme binding pocket. As illustrated by the DLS data in Figure 8, however, unfolding of bMc apo-b₅ is accompanied by a considerably smaller increase in hydrodynamic radius (2.2-fold) than is unfolding of rOM apo-b₅ (threefold), and by a correspondingly smaller increase in polydispersity (disorder). This is clearly attributable to the fact that bMc apo-b₅ retains greater non-random secondary structure content in the "unfolded" state than does rOM apo-b₅, which is further highlighted by the observation that bMc apo-b₅ migrates more slowly than rOM apo-b₅ on SDS-PAGE (data not shown) even though it is shorter by 10 residues. Residual structure in the unfolded state may also be a factor in the propensity of bMc apo-b₅ to form inclusion bodies during expression in E. coli at 37°C but not at 27°C, as thermal denaturation data suggest that the decrease in expression temperature will diminish the fraction of apoprotein in the "unfolded" state from ~16% to <2%. Furthermore, neither rOM apo-b₅ nor rOM^5M apo-b₅ forms inclusion bodies during expression at 37°C, despite having a significant unfolded population (~9% and >99%, respectively).

Despite exhibiting substantial differences in structure and dynamics in the folded and unfolded states, chemical and thermal denaturation data are consistent in showing that bMc and rOM apo-b₅ unfold with essentially identical free energy changes in aqueous solution at 25°C (pH 7.0). The stability of bMc apo-b₅ becomes progressively lower in comparison to that of rOM apo-b₅ as temperature is increased (lower T_m), however, but progressively higher in comparison to that of rOM apo-b₅ as the concentration of GdmCl or urea increases (higher C_m). The thermal denaturation and DSC data (Table 3) provide an explanation for the different temperature dependencies of bMc and rOM apo-b₅ stability: bMc apo-b₅ has the more positive H of unfolding, but also the more positive S of unfolding. A more positive S for bMc apo-b₅ at first seemed counterintuitive, in light of the CD, DLS, and electrophoresis data, indicating that it undergoes a considerably smaller increase in polypeptide conformational disorder during unfolding than does rOM apo-b₅. Differences in extent of solvation of the folded and unfolded forms of the two apoproteins is likely to be a major factor: Folded bMc apo-b₅ should be more extensively solvated than folded rOM apo-b₅ due to its more open heme binding pocket, but denatured bMc apo-b₅ should be less extensively solvated than denatured rOM apo-b₅ because of its more compact structure. Unfolding of bMc apo-b₅ will therefore be accompanied by a smaller increase in the number of water molecules involved in entropically unfavorable solvation of exposed nonpolar (Dill 1990) and polar (Makhatadze and Privalov 1996) moieties.

The higher stability of bMc apo-b₅ in comparison to rOM apo-b₅ toward denaturants is also likely related to differences in solvation in the unfolded and denatured states, perhaps better discussed in terms of the extent of polypep-tide surface area (side chains and backbone) that becomes exposed to solvent during unfolding. Myers et al. calculated the increase in solvent exposed surface area (ASA) expected to accompany the complete unfolding of 45 proteins of known structure, and perhaps not surprisingly found a very strong correlation with polypeptide length (R >0.98; equation 7; Myers et al. 1995). A reasonably strong correlation was also found between the calculated ASA values and experimentally determined GdmCl and urea m values for the same group of proteins (R = 0.84–0.87), which improved significantly (R = 0.9) when corrections were made for proteins that contain disulfide bridges and therefore cannot unfold completely (equations 8, 9). Using equation 7, it can be assumed that complete unfolding of a typical 82-residue protein such as bMc and rOM apo-b₅ (counting only the residues involved in distinct packing interactions in the corresponding holoproteins) will be accompanied by ASA = 6719 Ų (Table 5). Similar ASA values are predicted for unfolding of rOM apo-b₅ from equations 8 and 9, using our experimentally determined GdmCl and urea m values, respectively (Table 5). The corresponding ASA values determined for bMc apo-b₅ are smaller by at least half, consistent with our experimental observations. Incorporating the ASA values predicted for bMc apo-b₅ in this manner into equation 7 suggests that unfolding of bMc apo-b₅ exposes the equivalent of 42–45 residues to solvent.

View this table: [in this window] [in a new window]   Table 5. Predicted ASA values for bMc and rOM apo-b5

(7)

(8)

(9)

(10)

It is noteworthy that the C_p of unfolding determined for rOM apo-b₅ by DSC is only modestly larger than that determined for bMc apo-b₅, as it is known that C_p also correlates with ASA (Myers et al. 1995; Robertson and Murphy 1997). Indeed, Myers et al. found a much stronger correlation between experimental C_p values for the proteins in their study and ASA calculated on the assumption that they unfold completely, than between m values and calculated ASA values, even without correcting for the effects of disulfide bridges (R = 0.994; equation 10). Using our experimentally determined C_p values, equation 10 predicts a ASA for bMc apo-b₅ significantly larger than predicted based on m values, and a ASA for rOM apo-b₅ somewhat smaller than predicted based on m values (Table 5). In fact, ASA predicted for bMc and rOM apo-b₅ on the basis of their C_p values are strikingly similar to one another, in discord with our experimental observations. A possible explanation for this apparent contradiction may lie in the fact that solvation of nonpolar amino acid side chains contributes positively to C_p of unfolding and solvation of polar side chains and of the polypeptide backbone contributes negatively (Murphy and Freire 1992; Privalov and Makhatadze 1992), whereas polar and nonpolar surfaces make positive contributions to m (Schellman 1978; Alonso and Dill 1991; Myers et al. 1995; Shortle 1995). A protein that unfolds to a greater extent than another beginning from a similar structure and a similar distribution of nonpolar and polar residues, as in the case of rOM apo-b₅ versus bMc apo-b₅, is likely to expose more nonpolar and polar surface area to solvent, which will tend to cancel in terms of con-tributions to C_p, but will be additive in terms of contributions to m. This is consistent with the results of a recent study linking lower C_p values for thermophilic proteins relative to their mesophilic counterparts with the presence of more extensive salt-bridges and other polar interactions (Zhou 2002).

There is growing recognition that unfolded proteins do not adopt extended random coil conformations (Plaxco and Gross 2001; Klein-Seetharaman et al. 2002), and in some cases the structure of an unfolded polypeptide can even resemble the native fold (Shortle and Ackerman 2001; Ohnishi et al. 2004). It has been suggested that a compact denatured state has the advantage of reducing the unfavorable entropy change associated with folding, thereby simplifying the folding landscape (Pappu et al. 2000). It has recently been proposed that differences in C_p resulting from differences in the unfolded state structure can also endow two evolutionarily related proteins with similar thermodynamic stability at dramatically different temperatures, as in the case of ribonuclease H from Thermus thermophilus (T_m = 86°C; C_p = 1.8 kcal • mole^–1 • K^–1), which has a much more compact unfolded state than the analogous protein from E. coli (T_m = 66°C; C_p = 2.7 kcal • mole^–1 • K^–1) (Robic et al. 2003).

On the basis of the data presented herein, we surmise that Mc apo-b₅s have evolved folded states enthalpically more stable than those of OM apo-b₅s to compensate for the absence of the hydrophobic packing interactions stabilizing core 1 in the latter, and that this has rendered them more resistant to complete unfolding. Given the fact that core 1 of bMc apo-b₅ exhibits characteristics of a partly folded protein, it seems reasonable to expect that any compensating stabilizing interactions are located in core 2, and involve the antiparallel -sheet (we anticipate that the very short helices near the N and C termini in core 2 will be much less likely than the -sheet to resist unfolding). It is worth noting in this context that the number of amino acids apparently exposed to solvent upon unfolding of bMc apo-b₅ (42–45; vide supra) is only slightly greater than the number of residues comprising its empty heme binding pocket, which represents a nearly contiguous stretch of amino acids commencing near residue Thr-33 and ending near residue Thr-73. It is thus conceivable that much of the "unfolding" that occurs during denaturation of bMc apo-b₅ involves the already partly unfolded core 1. More direct evidence for stabilizing interactions involving core 2 and the -sheet in bMc that are absent from OM b₅s is provided by observations recorded for rOM^5M apo-b₅. This Mc/OM b₅ "hybrid" mutant contains Mc b₅-like packing interactions in core 1 and OM b₅-like packing interactions in core 2 (including a mostly OM b₅ -sheet), but is dramatically less stable than the two wild-type proteins and exhibits a CD spectrum when unfolded that is essentially identical to that of unfolded rOM apo-b₅.

	Materials and methods

TOP Abstract Introduction Results Discussion Materials and methods References

 
Proteins
Proteins were expressed and purified as previously described (Rivera et al. 1992; Cowley et al. 2002), with the exception that a temperature of 27°C was used. Analytical expression experiments were performed using the same protocol but with 250 mL of LB medium, and IPTG was added when the OD₆₀₀ of each sample had reached a value of 1.0 (approximately 3 h at 37°C and 5 h at 27°C). Identical solution volumes were maintained throughout the lysis and centrifugation steps. The cell pellets were analyzed for inclusion bodies using an established method (De Bernardez Clark 1998). Protein concentrations were determined on a Varian Carey 100 Bio UV/visible spectrometer, using the absorbance at 412 nm (Soret band _max) for the holoproteins, and at 280 nm for the apoproteins. For rOM holo-b₅ and bMc holo-b₅ an extinction coefficient of 130,000 M^–1 cm^–1 was used (Beck von Bodman et al. 1986). Extinction coefficients at 280 nm for rOM apo-b₅ ( = 12,950 M^–1 cm^–1) and bMc apo-b₅ ( = 11,460 M^–1 cm^–1) were calculated based on the number of Trp and Tyr residues in each protein (Gill and von Hippel 1989).

Circular dichroism spectroscopy
CD spectra were recorded on a Jasco J-710 spectropolarimeter equipped with a Peltier thermostated cell holder. Temperature within the cell was monitored with an Omega model HH200 thermometer with T thermocouple (±0.2°C). Spectra were acquired at 1.0 nm intervals with a response time of 4 sec and a scan rate of 50 nm/min, and represent the average of at least five scans. Background correction was accomplished by subtraction of a spectrum recorded at the same temperature and containing only buffer. To test for isosbestic behavior during apoprotein unfolding, far-UV spectra were recorded at 6°C intervals between 5°C and 75°C. Increments of 3°C were used in the thermal denaturation experiments, for which the mean residue ellipticity at 222 nm (₂₂₂) over a 5-min period was recorded and averaged. An identical set of measurements was performed using a blank sample, and subtracted to correct for background changes. Samples were equilibrated for 15 min at each a temperature prior to recording data. Data were fit to the two-state equation noted in the text (Constans et al. 1998), generously made available to us by Prof. Juliette Lecomte, using the program Igor Pro v. 4.0 (Wavemetrics, Inc.). The mean values and average deviations from three independent runs are reported in Table 3.

Chemical denaturation studies
Stock solutions of urea (Fischer) were prepared in 50 mM potassium phosphate, buffered to pH 7.0. To eliminate potential errors due to weighing of the hygroscopic denaturant, stock solution concentrations were verified from measurements of the solution refractive index (Pace 1986). Sample solutions were equilibrated at 25°C for 1 h prior to recording fluorescence emission spectra on a PTI Quanta Master luminescence spectrometer, with excitation of Trp-22 at 295 nm. Data fitting was performed using the program Kaleidagraph, version 3.6. The mean values and average deviations from two independent runs are reported in Table 2.

Differential scanning calorimetry
Scanning calorimetry experiments were performed on a Nano Differential Scanning Calorimeter CSC 5100 (Calorimetry Sciences Corp.), using samples containing 1.4–1.9 mg/mL of apoprotein that had been extensively dialyzed (four changes of buffer every 6 h) against 50 mM potassium phosphate (pH 7.0). Immediately prior to each experiment, insolubles were removed by centrifugation at 12,000 x g for 5 min and the sample was degassed at 0.5 atm for 15 min. Prior to introducing each sample, baselines were established via repeated scans in which the sample and reference cell contained buffer solution from the final dialysis step. Scans were performed from low to high temperatures ("upscans") at 1 K/min, followed by a downscan and a second upscan at the same rate to evaluate reversibility. Each experiment was repeated with a fresh apoprotein sample. Data from both upscans and the single downscan from each experiment were analyzed via a statistical mechanics-based deconvolution as implemented in the CSC 5100 software package to obtain the calorimetric enthalpy change (H_cal), and via a nonlinear least-squares fit to a two-state model to obtain the van’t Hoff enthalpy change (H_vH). The mean values and average deviations from the six fits are reported in Table 3.

Dynamic light scattering
DLS measurements (Schmitz 1990) were performed on a BI-200SM research goniometer and laser light scattering system, equipped with a BI-9000AT digital correlator (Brookhaven Instruments Corporation). Incident light of = 532 nm (0.3–1.0 W) was used, with scattered light detected at an angle of 90° via a photomultiplier tube. Sample temperature was controlled by means of a thermostated cell jacket and monitored with the thermocouple described above. Samples (100 µM) were passed through 100 nm filters (Whatman) immediately before use. Three independent measurements were performed for each protein, each consisting of six 30-sec runs. The instrument software reports the results of each experiment as the average of the six runs. Data from one measurement on each protein are reported in Table 4 (the duplicate run yielded essentially identical results). All data could be fitted multimodally, and essentially 100% of the scattering mass was attributed to a single low molecular mass component. The diffusion coefficient (D) and the hydrodynamic radius (R_h) are related by the Stokes-Einstein equation (equation 11).

(11)

where T is the temperature in Kelvin, k is the Boltzmann constant (1.38 x 10^–16 erg/K), and is the solution viscosity (set at 1.0 g/cm•sec and at 1.663 g/cm•sec for samples examined in buffered aqueous solution and in 8 M urea, respectively, at 25°C; Kawahara and Tanford 1966).

	Acknowledgments

 
This work was supported by a grant to D.R.B. and M.R. from the NSF (MCB-0110139). We thank Prof. Russell Middaugh, Haihong Fan, and Jason Rexroad for assistance in obtaining the DSC and DLS data, and Adriana Altuve for providing some of the holoprotein samples.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.

	References

TOP Abstract Introduction Results Discussion Materials and methods References

 
Alonso, D.O.V. and Dill, K.A. 1991. Solvent denaturation and stabilization of globular proteins. Biochemistry 30: 5974–5985.[CrossRef][Medline]

Altuve, A., Silchenko, S., Lee, K.-H., Kuczera, K., Terzyan, S., Zhang, X., Benson, D.R., and Rivera, M. 2001. Probing the differences between rat liver outer mitochondrial membrane cytochrome b₅ and microsomal cytochromes b₅. Biochemistry 40: 9469–9483.[CrossRef][Medline]

Altuve, A., Wang, L., Benson, D.R., and Rivera, M. 2004. Mammalian mitochondrial and microsomal cytochromes b₅ exhibit divergent structural and biophysical characteristics. Biochem. Biophys. Res. Commun. 314: 602–609.[CrossRef][Medline]

Beck von Bodman, S., Schuler, M.A., Jollie, D.R., and Sligar, S.G. 1986. Synthesis, bacterial expression, and mutagenesis of the gene coding for mammalian cytochrome b₅. Proc. Natl. Acad. Sci. 83: 9443–9447.[Abstract/Free Full Text]

Bhattacharya, S., Falzone, C.J., and Lecomte, J.T.J. 1999. Backbone dynamics of apocytochrome b₅ in its native, partially folded state. Biochemistry 38: 2577–2589.[CrossRef][Medline]

Compton, L.A. and Johnson Jr., W.C. 1986. Analysis of protein circular dichroism spectra for secondary structure using a simple matrix multiplication. Anal. Biochem. 155: 155–167.[CrossRef][Medline]

Constans, A.J., Mayer, M.R., Sukits, S.F., and Lecomte, J.T.J. 1998. A test of the relationship between sequence and structure in proteins: Excision of the heme binding site in apocytochrome b₅. Protein Sci. 7: 1983–1993.[Abstract]

Cowley, A.B., Altuve, A., Kuchment, O., Terzyan, S., Zhang, X., Rivera, M., and Benson, D.R. 2002. Toward engineering the stability and hemin binding properties of microsomal cytochromes b₅ into rat outer mitochondrial membrane cytochrome b₅: Examining the influence of residues 25 and 71. Biochemistry 41: 11566–11581.[CrossRef][Medline]

Dangi, B., Sarma, S., Yan, C., Banville, D.L., and Guiles, R.D. 1998. The origin of differences in the physical properties of the equilibrium forms of cytochrome b₅ revealed through high-resolution NMR structures and backbone dynamic analyses. Biochemistry 37: 8289–8302.[CrossRef][Medline]

De Bernardez Clark, E. 1998. Refolding of recombinant proteins. Curr. Opin. Biotechnol. 9: 157–163.[CrossRef][Medline]

Dill, K.A. 1990. Dominant forces in protein folding. Biochemistry 29: 7133–7155.[CrossRef][Medline]

Durley, R.C.E. and Mathews, F.S. 1996. Refinement and structural analysis of bovine cytochrome b₅ at 1.5 Å resolution. Acta Crystallogr. D 52: 65–76.