PEDS Advance Access published online on October 24, 2007
Protein Engineering Design and Selection, doi:10.1093/protein/gzm053
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Enhancing the thermal stability of mitochondrial cytochrome b5 by introducing a structural motif characteristic of the less stable microsomal isoform
Department of Chemistry, University of Kansas, Lawrence, KS 66045, USA
1 To whom correspondence should be addressed. E-mail: drb{at}ku.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Outer mitochondrial membrane cytochrome b5 (OM b5) is the most thermostable cytochrome b5 isoform presently known. Herein, we show that OM b5 thermal stability is substantially enhanced by swapping an apparently invariant motif in its heme-independent folding core with the corresponding motif characteristic of its less stable evolutionary relative, microsomal cytochrome b5 (Mc b5). The motif swap involved replacing two residues, Arg15 with His and Glu20 with Ser, thereby introducing a Glu11-His15-Ser20 H-bonding triad on the protein surface along with a His15/Trp22
-stacking interaction. The ferric and ferrous forms of the OM b5 R15H/E20S double mutant have thermal denaturation midpoints (Tm values) of 
93°C and 104°C, respectively. A 15°C increase in apoprotein Tm plays a key role in the holoprotein thermal stability enhancement, and is achieved by one of the most common natural mechanisms for stabilization of thermophilic versus mesophilic proteins: raising the unfolding free energy along the entire stability curve.
Keywords: apoproteins/cytochrome b5/m values/residual structure/thermal stability
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Proteins that are stably folded and functional at high temperatures (thermophilic proteins) continue to attract attention, in part because they challenge our fundamental understanding of protein stability (Sterner and Liebl, 2001
; Li et al., 2005; Razvi and Scholtz, 2006), but also because of their potential industrial applications (Bruins et al., 2001; Vieille and Zeikus, 2001; de Miguel Bogas et al., 2006). Highly heat stable proteins may be obtained from thermophilic organisms (Vieille and Zeikus, 2001; van den Burg, 2003; Atomi, 2005). Alternatively, they can be generated via engineering approaches commencing with proteins of more modest stability (Renugopalakrishnan et al., 2005; DiTursi et al., 2006; Reetz et al., 2006; Sylvestre et al., 2006), efforts that are sometimes guided by studies comparing orthologous proteins from thermophilic and mesophilic organisms. We have recently initiated a similar rational approach toward thermal stability enhancement (Sun et al., 2005), but based on comparative studies (Silchenko et al., 2000; Altuve et al., 2001; Cowley et al., 2002, 2004, 2005; Altuve et al., 2004) of two paralogous heme proteins that arose via gene duplication in vertebrates (Lederer et al., 1983; Guzov et al., 1996; Wang et al., 2007): the outer mitochondrial membrane and microsomal isoforms of cytochrome b5 (OM b5 and Mc b5, respectively). OM and Mc b5s contain polar heme-binding domains that are separated from membrane anchoring domains via flexible medial regions (Lederer, 1994; Kuroda et al., 1998). Studies with recombinant proteins comprising only the heme-binding domains have revealed that mammalian OM b5s are considerably more thermostable than their Mc counterparts (Silchenko et al., 2000; Altuve et al., 2001, 2004). For example, thermal denaturation midpoints (Tm values) for the tryptic fragments of rat OM cytochrome b5 (rOM b5) and bovine Mc cytochrome b5 (bMc b5) are 85°C and 67°C, respectively. Figure 1 compares the amino acid sequences of bMc and rOM b5. An alignment of heme-binding domain sequences for all known mammalian OM and Mc b5 pairs is included in the Supporting Information (Figure S1).
|
The greater thermal stability exhibited by mammalian OM b5s in comparison with their Mc counterparts probably does not play a direct role in their specialized sub-cellular roles. It is more likely to be a source or consequence of two other notable biophysical differences, one or both of which may be related to their functional divergence. Specifically, OM b5s differ from Mc b5s in having (i) more negative reduction potentials (Altuve et al., 2004
), indicating a need for a stronger electron transfer driving force; and (ii) lower polypeptide conformational mobility (Altuve et al., 2001; Lee and Kuczera, 2003; Simeonov et al., 2005), suggesting greater specificity in interactions with redox partners. Our ideas about the possible relationship between OM and Mc b5 stability, redox potential and polypeptide conformational mobility have been detailed in several recent papers (Altuve et al., 2004; Cowley et al., 2004, 2005, 2006; Simeonov et al., 2005).
In light of the divergent stability properties of mammalian OM and Mc b5s, it is noteworthy that the corresponding apoproteins exhibit nearly identical unfolding free energies at pH 7 and 25°C [
GN
U 3 kcal/mol as determined in urea-mediated denaturation studies (Cowley et al., 2004)]. OM and Mc apo-b5s are also similar, in that their empty heme-binding pockets (core 1; see Figs 1 and 2A) are virtually devoid of secondary structure, whereas the remainder of each polypeptide (core 2) adopts a stable holoprotein-like fold (Falzone et al., 1996; Cowley et al., 2004). Studies in our laboratories have indicated that conformationally disordered core 1 is considerably more compact and less dynamic in OM apo-b5 than in Mc apo-b5, however, indicating the population of fewer non-native conformations (Cowley et al., 2004). The stronger heme-binding exhibited by OM apo-b5 relative to Mc apo-b5 may therefore reflect a less unfavorable entropy associated with the accompanying core 1 polypeptide conformational reorganization (Cowley et al., 2004, 2005).
|
The similar OM and Mc apo-b5 unfolding free energies have arisen by evolution of distinctly different conserved packing motifs, as initially revealed in mutagenesis studies involving residue 71 (Cowley et al., 2002
, 2005). Residue 71, located near the C-terminal end of 
5 in core 1 (Figs 1 and 2A), is leucine in all known mammalian OM b5s and serine in their Mc counterparts. Replacing Leu71 in rOM b5 with Ser was found (i) to destabilize the holoprotein, due in large measure to a >2 kcal/mol decrease in apoprotein stability (Cowley et al., 2002); and (ii) to extend conformational disorder in the apoprotein beyond core 1 and into core 2, including disruption of tertiary structure in the vicinity of Trp22 (Cowley et al., 2005). Mc b5s are therefore better able than OM b5s to accommodate Ser at position 71, which we concluded was a manifestation of a more stable core 2. A difference in packing that we predicted might contribute to this greater Mc b5 core 2 stability was identified by comparing X-ray crystal structures of bMc b5 (PDB 1CYO
[PDB]
) and rOM b5 (PDB 1B5M
[PDB]
), in the context of amino acid sequences of all known pairs of mammalian Mc and OM b5s (Wang et al., 2006). In the bMc b5 structure (Fig. 2B), the solvent-exposed side chain of His15 engages in a -stacking interaction with the side chain of Trp22, and also forms hydrogen bonds with the side chains of Glu11 (N
2–O 2.9 Å) and Ser20 (N
1–O
2.8 Å). As illustrated in Figure S1, Trp22 and Glu11 are present in all known mammalian Mc and OM b5s, but His15 and Ser20 are invariant only among the Mc proteins. The corresponding residues in the OM proteins are an invariant Arg15 and a highly conserved Glu20 (only bovine OM b5, with Asp at position 20, is known to differ). As shown in Fig. 2C for rOM b5, the Arg15 side chain is packed parallel to that of Trp22 and engages in electrostatic interactions with the side chains of Glu11 (NH–O 5.5 Å) and Glu20 (N– O 4.4 Å).
As a test of our prediction (Wang et al., 2006), we replaced Arg15 with His and Glu20 with Ser in a previously reported rOM b5 quintuple mutant that contains Ser at position 71, hereafter referred to as OM5M b5 (Cowley et al., 2002). The resultant septuple mutant (OM7M b5) was found to be considerably more thermostable (Tm = 76°C) than OM5M b5 (Tm = 69°C), but still less thermostable than rOM b5 (Tm = 85°C). We obtained a 2.1 Å X-ray crystal structure of OM7M b5 containing four protein molecules in the asymmetric unit cell, (PDB 2I89), all of which exhibit His15/Trp22 -stacking and His15/Ser20 H-bonding interactions analogous to those observed in the bMc b5 crystal structure (Wang et al., 2006). A bMc b5-like Glu11/His15 H-bond is present in only two of the four molecules in the OM7M b5 unit cell, however, whereas in a third molecule, the Glu11 side chain extends into solvent. The fourth molecule has Glu11 resolved in the two orientations just described. Considered in the light of a survey of Mc b5 solution structures, this led us to conclude that strength of interactions involving His15 in OM7M b5 and in Mc b5s decreases in the order His15/Trp22 -stack>His15/Ser20 H-bond>His15/Glu11 H-bond.
The increase in OM5M b5 thermal stability resulting from the R15H/E20S double mutation had a major contribution from enhancement of apoprotein thermodynamic stability (Wang et al., 2006). The R15H and E20S mutations introduced into OM5M b5 to generate OM7M b5 also (i) returned significant structural stability to core 2 of the apoprotein; and (ii) caused a change in apoprotein unfolding, from a reversible two-state reaction to a partly reversible one that occurs in two distinct stages. In the work described herein we tested a prediction that thermal stability of rOM b5 could be increased by introducing the R15H/E20S double mutation, thereby producing the most stable b5 variant reported to date. We were also interested in elucidating the role of apoprotein in governing the anticipated stabilization.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Proteins
The recombinant plasmid pET11a-OM b5, harboring the gene coding for the heme-binding domain of rOM b5, was used as template to construct a gene for the corresponding R15H/E20S double mutant (hereafter rOMR15H/E20S b5). The primers designed for this purpose were 5'-GAAGTTGCGAAACATAACACCGCGGAATCAACCTGGATG G-3' and 5'-CCATCCAGGTTGATTCCGCGGTGTTATGTTTCGCAACTTC-3'. Underlined codons represent mismatches introduced to generate the mutations. The recombinant constructs were transformed into Escherichia coli DH5 competent cells for amplification. Once the mutations had been confirmed by sequencing, the recombinant pET-11a plasmids carrying the mutant genes were transformed into E.coli BL21(DE3) cells for subsequent protein expression. Expression of rOMR15H/E20S b5 and subsequent steps preceding purification were accomplished using the protocol developed for rOM b5 (Rivera et al., 1992), and purification was achieved using a procedure recently reported for house fly cytochrome b5 (Wang et al., 2007). Purity was assessed by native and SDS–PAGE. The mutant protein was analyzed by electrospray ionization mass spectrometry (ESI-MS; KU mass spectrometry laboratory), using conditions that caused essentially complete dissociation of hemin. The experimental mass closely matched the calculated mass of the polypeptide, which includes the initiator methionine and residues (–)5 to 87: 10 447.7 Da (experimental); 10 447.5 (calculated average MW).
Holoprotein thermal denaturation
Thermal denaturation experiments with the holo form of rOMR15H/E20S b5 were performed on a Varian Carey 100 Bio UV/Visible spectrophotometer equipped with a Peltier- thermostated multiple cell holder and a dedicated temperature probe accessory (±0.1°C). Solutions were buffered to pH 7.0 using 50 mM sodium phosphate. Protein concentration ranged from 3–5 µM, and was estimated via the absorbance at the Soret band [
max = 412 nm; max = 130 000 M–1 cm–1 in the ferric state (Beck von Bodman et al., 1986)]. Experiments with the ferric proteins were performed in quartz cuvettes of 1 cm path length and 1 ml sample volume, equipped with tight-fitting PTFE lids. For experiments with the ferrous proteins, we utilized a specially designed side-arm quartz cuvette of 1 cm path length and 3 ml sample volume featuring an opening that could be sealed with a rubber septum. Reduction of the ferric proteins to the ferrous state was accomplished by adding an aliquot of degassed aqueous sodium dithionite, after bubbling of the sample solutions for at least 30 min with N2 that had been passed through a chromous chloride solution to remove adventitious oxygen. A positive N2 pressure was maintained throughout the experiment. In all experiments, the temperature was increased in increments of 2°C, and samples were equilibrated for 5 min after reaching each desired temperature. For proteins having thermal unfolding curves that reached a plateau in the denaturing region, thermal denaturation midpoints (Tm values) were obtained by fitting plots of absorbance at the Soret band max (412 nm for the ferric forms; 423 nm for the ferrous forms) versus temperature to a previously described equation describing a two-state equilibrium (Constans et al., 1998). For proteins having denaturation curves with no plateau in the denaturing region, Tm values were estimated by visual inspection of the data.
Apoprotein chemical denaturation
Urea-mediated apoprotein denaturation studies were performed on a PTI QuantaMaster luminescence spectrometer (protein concentrations 0.5–1 µM), monitoring changes in emission (340 nm) of Trp-22 (exc 295 nm). Samples were buffered to pH 7.0 using 30 mM MOPS, and were incubated at 25°C for 1 h before spectra were recorded at the same temperature. Denaturation data were fit (Kaleidagraph v. 3.5; Synergy Software) to two-state Eq. (1), where [D] is the urea concentration, GNU the free energy of unfolding at 25°C when [urea ] = 0 and m the sensitivity of the unfolding free energy to urea concentration (Pace, 1986). Urea denaturation midpoints (Cm values) were then determined from the relationship in Eq. (2).
|
| 1 |
|
| 2 |
Differential scanning calorimetry
Scanning calorimetry experiments with rOMR15H/E20S apo-b5 (0.80 mg/ml) were performed on a VP-DSC microcalorimeter (Microcal Inc.). The sample was extensively dialyzed before use (four changes of buffer every 6 h) against 50 mM potassium phosphate, pH 7.0. Immediately prior to each experiment, insoluble matter was removed by centrifugation at 12 000g for 5 min and the sample was degassed at 0.5 atm for 15 min. Prior to making measurements, baselines were established via repeated scans in which the sample cell contained buffer solution from the final dialysis step. Scans were performed from low to high temperatures at 1 K/min. Data were analyzed via a statistical mechanics-based deconvolution as implemented in the CSC 5100 software package to obtain the calorimetric enthalpy change (Hcal), and via a non-linear least-squares fit to a two-state model to obtain the van't Hoff enthalpy change (HvH).
Dynamic light scattering (DLS) measurements 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. Samples (100 µM) were passed through 100 nm filters (Whatman, UK) immediately before use. Experiments were performed at 25°C, and samples were buffered to pH 7 with 50 mM potassium phosphate. Each experiment consists of six 30-s runs, and the instrument software reports the results of each experiment as the average of the six runs. Three independent experiments yielded nearly identical results, and therefore we only report the results from one experiment in the appropriate data table. 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 (RH) are related by the Stokes–Einstein equation [Eq. (3)].
|
| 3 |
the solution viscosity. Values of for urea solutions were determined by the method of Kawahara and Tanford (1966). |
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Holoprotein thermal denaturation
Thermal denaturation of the polypeptide in ferric and ferrous cytochrome b5 is accompanied by dissociation of the bonds between heme iron and its axial ligands His39 and His63, converting iron from the low-spin to the high-spin state (Sugiyama et al., 1980; Hewson et al., 1993). This allows the reactions to be conveniently studied by UV/vis spectroscopy, monitoring the decrease in intensity of the Soret band max (412 nm for the ferric proteins; 423 nm for the ferrous proteins). Figure 3A and B shows thermal denaturation curves for the ferric and ferrous forms, respectively, of bMc, rOM and rOMR15H/E20S b5. As previously observed for ferric rOM and bMc b5, thermal denaturation of ferric rOMR15H/E20S b5 and of ferrous rOM and bMc b5 is not fully reversible. For this and other reasons (Cowley et al., 2002), the data for those proteins cannot be extrapolated reliably through fits to two-state equations, and Table I therefore only reports their Tm values. Notably, ferric rOMR15H/E20S b5 (Tm = 93°C) exhibits substantially enhanced thermostability relative to ferric rOM b5 (Tm = 85.5°C).
|
|
Heme iron reduction enhances cytochrome b5 thermal stability (Hewson et al., 1993
), likely because it removes a formal +1 charge on iron and thereby strengthens heme–polypeptide interactions (McLachlan et al., 1986; Banci et al., 1998). Yao et al. (1997) have reported that, in 100 mM sodium phosphate buffer at pH 7, the Tm value of bMc b5 increases from 66°C in the ferric state to 77°C in the ferrous state (Tm = 11°C). We obtained a nearly identical result with bMc b5 in 50 mM potassium phosphate buffer at pH 7 (Fig. 3A and B; Table I). Reducing rOM b5 from the ferric to the ferrous state produced a somewhat smaller Tm increase, from 86°C to 93°C (Fig. 3B; Table I). Reduction was also found to enhance rOMR15H/E20S b5 thermal stability. In fact, ferrous rOMR15H/E20S b5 appears to be 80% folded at 99°C, the highest temperature we could reach in our experiments (Figs 3B and 4). We estimate that reduction increases the Tm of the double mutant from 92°C to 104°C, substantially greater than the effect observed in the ferric state. Returning the ferrous rOMR15H/E20S b5 samples to 25°C at the end of the thermal denaturation experiments reproduced the original spectra, demonstrating (i) absence of heme iron oxidation during the experiments; and (ii) that no solvent was lost to evaporation.
|
Apoprotein chemical denaturation
Urea-mediated denaturation of apo-b5s can be conveniently probed by fluorescence spectroscopy, monitoring changes in fluorescence emission of the lone Trp residue, Trp22. Previously published urea-mediated denaturation data for rOM and bMc apo-b5 at 25°C and pH 7.0 (Cowley et al., 2004) are shown in Fig. 5 and Table II, along with corresponding data for rOMR15H/E20S apo-b5 obtained in the present study. The curves are reported in terms of the fraction of protein that is folded at each urea concentration, and the lines through the curves represent fits to two-state Eq. (1) (see experimental). The results of this study show that rOMR15H/E20S apo-b5 is 1.1 kcal/mol more stable than rOM apo-b5 at 25°C.
|
|
We have reported that Trp22 fluorescence emission intensity decreases substantially upon unfolding of rOM and OM5M apo-b5 (Wang et al., 2006
), consistent with the Trp22 side chain moving from a well-ordered region in the folded state to a disordered region in the unfolded state. Trp22 fluorescence emission intensity was observed to be much weaker in the folded forms of bMc and OM7M ap-b5 than in the folded forms of rOM and OM5M apo-b5. Moreover, Trp22 emission intensity was found to increase upon unfolding of OM7M apo-b5, as had previously been demonstrated for bMc apo-b5 (Huntley and Strittmatter, 1972). On the basis of that literature precedent, we concluded that Trp22 fluorescence is quenched by -stacking with His15 in folded OM7M apo-b5 (Wang et al., 2006), an effect that becomes less efficient as the protein unfolds. The present study showed that Trp22 fluorescence emission is much less intense in rOMR15H/E20S apo-b5 than in rOM apo-b5, and that it increases substantially upon rOMR15H/E20S apo-b5 unfolding (data not shown). We can therefore conclude that Mc b5-like His15/Trp22 interactions are present in rOMR15H/E20S apo-b5. Apoprotein thermal denaturation
Differential scanning calorimetry (DSC) data for rOMR15H/E20S apo-b5 revealed a single transition that is fit well by a two-state model, as previously observed for rOM and bMc apo-b5 (Fig. 6). The fitting routine provides the heat capacity change associated with the unfolding transition (Cp), the Tm, and the enthalpy of unfolding referenced to the Tm (Hm; also known as the van't Hoff enthalpy, HvH). The DSC data for rOMR15H/E20S apo-b5 were also deconvoluted to obtain the calorimetric enthalpy (Hcal), which is independent of the unfolding pathway. The Tm, Cp, HvH and Hcal values obtained for rOMR15H/E20S apo-b5 in the present work, and for rOM and bMc apo-b5 in previous studies, are reported in Table IV. As with rOM and bMc apo-b5, Hcal and HvH values determined for rOMR15H/E20S apo-b5 differ by <5%. This provides convincing evidence (Zhou et al., 1999) that the R15H/E20S double mutation has not affected the two-state thermal unfolding equilibrium exhibited by rOM apo-b5. The data in Table IV show that the Tm value of rOMR15H/E20S apo-b5 (64.9°C) is 15°C higher than that of rOM apo-b5 (50.1°C), which in turn is 5.4°C higher than the Tm of bMc apo-b5 (44.7°C). The Tm values for the holoproteins and apoproteins in this study therefore both decrease in the order rOMR15H/E20S>rOM>bMc.
|
|
Although
Cp of unfolding varies to some extent with temperature (Privalov et al., 1986), it is commonly treated as a constant in order to estimate free energies of protein unfolding at temperatures other than the Tm (Freire, 1995). We calculated values of G from 0°C to 100°C for bMc, rOM and rOMR15H/E20S apo-b5 using Eq. (4), an expanded form of the Gibbs–Helmholtz equation.
|
| 4 |
The resultant stability curves (Becktel and Schellman, 1987) are presented in Fig. 7. In light of our assumption of a constant Cp, it is worth noting that the relative GNU values for bMc, rOM and rOMR15H/E20S apo-b5 determined by extrapolation of DSC data to 25°C (Table IV) are very similar to those obtained in urea denaturation experiments performed at 25°C (Table II). This gives us confidence that relative stabilities of the apoproteins calculated at both higher and lower temperatures are also reliable.
|
Dynamic light scattering
As noted above, urea-mediated denaturation data show that rOMR15H/E20S apo-b5 is 1.1 kcal/mol more stable than rOM apo-b5 at 25°C. Nonetheless, the data in Fig. 5 and Table II show that the R15H/E20S double mutation (i) greatly increased the concentration of urea required to achieve 50% unfolding of rOM apo-b5 (the Cm value); and (ii) substantially decreased the slope of the unfolding transition as reflected in a diminished m value. The data in Fig. 5 and Table II illustrate our previously reported finding that bMc apo-b5 likewise has a larger Cm value and a smaller m value than does rOM apo-b5 (Cowley et al., 2004). Mutations are known to lower a protein's denaturant m value if they diminish the change in solvent accessible surface area (ASA) that accompanies unfolding (Myers et al., 1995). Previously published DLS data shown in Fig. 8 and Table III have indicated that a smaller ASA is a source of the smaller m value of bMc apo-b5 relative to rOM apo-b5 (Cowley et al., 2004). Specifically, bMc apo-b5 unfolding involves a much smaller increase in hydrodynamic radius (RH value) than does rOM apo-b5 unfolding, largely because urea-denatured bMc apo-b5 is much more compact than urea-denatured rOM apo-b5. Figure 8 and Table III show that the same explanation holds for the smaller m value of rOMR15H/E20S apo-b5 relative to that of rOM apo-b5: urea-denatured rOMR15H/E20S apo-b5 is more compact than urea-denatured rOM apo-b5, and in fact is even more compact than urea-denatured bMc apo-b5. Notably, the double mutation exerted little, if any effect on the RH value of folded rOM apo-b5, consistent with the fact that both mutated residues are located on the surface of core 2.
|
|
We also performed additional DLS studies on OM7M apo-b5 which, as noted in the introduction, unfolds in two distinct stages. Table III shows that near the end of the first denaturation stage, in 3 M urea, the RH value of OM7M apo-b5 is somewhat smaller than those of urea-denatured bMc and rOMR15H apo-b5. On the other hand, the RH value of OM7M apo-b5 at the end of its second unfolding stage is similar to that of urea-denatured rOM apo-b5.
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Effects of the motif swap on holoprotein stability and structure
We have substantially enhanced the thermal stability of rOM b5 in both the ferric and ferrous oxidation states by replacing a conserved mammalian OM b5 core 2 packing motif (Fig. 2C) with the corresponding motif characteristic of the less stable Mc isoform (Fig. 2B). To the best of our knowledge, the resultant mutant rOMR15H/E20S b5 is the most thermostable cytochrome b5 variant reported to date, with Tm values of 93°C (FeIII) and 104°C (FeII) at pH 7.
On the basis of evidence summarized in the introduction, we can assume that the motif swap introduced the following structural features into rOM b5: (i) a Glu11-His15-Ser20 H-bonding triad that appears likely to be quite stabilizing when fully intact, as judged by the 2.8–2.9 Å N(His)–O(Glu) distances in the bMc b5 crystal structure; and (ii) a His15/Trp22 -stacking interaction in which the His15 imidazole ring is located directly above and parallel to the six-membered ring of the Trp22 indole side chain. This is one of the most commonly observed His/Trp side chain interactions in proteins (Samanta et al., 1999). Figure 9 shows two distinct patterns that are possible for the fully H-bonded Glu11-His15-Ser20 triad. NMR studies have shown that His15 has a pKa of 8.5 in bMc holo-b5 (Altman et al., 1989), as well as in rat Mc holo- and apo-b5 (Moore et al., 1991). It is therefore safe to assume that His15 in bMc apo-b5 also has a pKa of 8.5, and is protonated at the pH of 7.0 utilized in our studies (as in Fig. 9B). The Glu11/His15 H-bond is therefore also a salt bridge, and the His15/Trp22 interaction has cation- character (Loewenthal et al., 1991; Fernandez-Recio et al., 1997; Berry et al., 2007).
|
The Glu11-His15-Ser20 triad in rOMR15H/E20S b5 replaced a Glu11-Arg15-Glu20 triad, which appears likely to stabilize OM b5s via ion-pair interactions. Although the relevant N–O distances in the rOM b5 crystal structure are rather long, they can be much shorter (and the electrostatic interactions correspondingly stronger) as demonstrated by the crystal structure of OM5M b5 (Cowley et al., 2002
). In all four molecules in the OM5M b5 unit cell, Arg15 NH to Glu11 O distances are 2.6–2.7 Å (versus 4.4 Å in rOM b5), and the side chains are much better oriented for H-bond formation. Arg15N to Glu20 O distances are also shorter in all four OM5M b5 molecules (3.1–3.6 Å) than in the rOM b5 structure (5.5 Å), but are not well oriented for H-bonding. The shorter N–O distances in the OM5M b5 structures than in the rOM b5 structure are primarily a consequence of different Glu11 and Glu20 side chain conformations. In contrast, the Arg15 side chains in rOM b5 and the four OM5M b5 molecules have virtually identical conformations. Analogous to the His15 imidazolyl side chain in the Mc b5 motif described earlier, the Arg15 guanidino moiety in rOM and OM5M b5 engages in parallel -stacking/cation- interactions with the six-membered ring of the Trp22 indole group, a particularly common and stabilizing arrangement for Arg and Trp side chains in proteins (Flocco and Mowbray, 1994; Mitchell et al., 1994; Gallivan and Dougherty, 1999; Tatko and Waters, 2003). Role of the apoprotein in the stability increase
As noted in Results, we are unable to accurately determine thermodynamic parameters for unfolding of the holoproteins in this study. This reflects two factors: (i) most of the denaturation reactions are not fully reversible; and (ii) not all of the unfolding reactions are complete at the highest temperature achievable in our studies (99°C). In principle, however, thermodynamic stability of a given holo-b5 variant can also be estimated from the sum of the free energies of apoprotein folding (GUN = –GNU) and of heme binding by the folded apoprotein (Gbind). Given that none of the residues in the Mc and OM b5 motifs compared in Fig. 2B and C make van der Waals contact with the heme, we consider it unlikely that the R15H/E20S double mutation enhanced rOM holo-b5 thermostability via strengthened heme–polypeptide interactions. Moreover, the R15H/E20S double mutation did not detectably alter the RH value of folded rOM apo-b5, suggesting little if any effect on empty core 1. It is therefore reasonable to conclude that the R15H/E20S double mutation had little if any effect on Gbind, another parameter we are unable to measure directly. The holoprotein thermal stabilization instead appears to arise primarily from enhancement of apoprotein thermodynamic stability (GUN), most notably an 15°C increase in apoprotein Tm.
Three mechanisms have been described which, alone or in combination, can increase Tm of a protein and thereby make it more thermophilic (Nojima et al., 1977; Razvi and Scholtz 2006): (I) raising G along the entire stability curve; (II) flattening the stability curve by decreasing Cp and (III) shifting the entire stability curve to higher temperatures, for example, by increasing the temperature at which the protein is maximally stable (TS; where S = 0). Figure 7 and Table IV show that (i) rOM and rOMR15H/E20S apo-b5 have nearly identical TS and Cp values; and (ii) rOMR15H/E20S apo-b5 is more stable than rOM apo-b5 at all temperatures between 0°C and 100°C. The 15°C increase in rOM apo-b5 Tm resulting from the R15H/E20S double mutation therefore arises almost exclusively via mechanism (I), the most common one utilized by nature for stabilizing proteins from thermophilic organisms relative to their mesophilic counterparts (Razvi and Scholtz, 2006). It is also worth noting that rOMR15H/E20S apo-b5 becomes progressively more stable in comparison with rOM apo-b5 as the temperature increases (G 1.2 kcal/mol at 25°C, 2.0 kcal/mol at 60°C and 3.1 kcal/mol at 100°C).
The greater thermal stability exhibited by rOM holo-b5 in comparison with bMc holo-b5 also has a contribution from a higher apoprotein Tm (Tm = 5.4°C). rOM and bMc apo-b5 have very similar Cp values and are equal in stability at 34°C, but the TS value of rOM apo-b5 is 13°C higher than that of bMc apo-b5. The higher Tm value of rOM apo-b5 relative to bMc apo-b5 therefore occurs mainly via mechanism (III). rOM apo-b5 becomes progressively less stable in comparison with bMc apo-b5 below 34°C (G approximately –0.4 kcal/mol at 25°C; approximately –1.6 kcal/mol at 0°C), but progressively more stable relative to bMc apo-b5 above 34°C (G 1.0 kcal/mol at 60°C; 2.2 kcal/mol at 100°C).
Because the R15H/E20S double mutation appears to have affected GUN much more than Gbind, we consider it likely that relative stabilities of rOMR15H/E20S and rOM holo-b5 as a function of temperature closely parallel those of rOMR15H/E20S and rOM apo-b5. In other words, rOMR15H/E20S holo-b5 is almost certainly more stable than rOM holo-b5 at all temperatures between 0°C and 100°C, and to a greater extent at high temperatures than at more modest temperatures. It can also be argued that rOM and bMc holo-b5 exhibit greater stability differences at high temperatures than at more modest temperatures, although the relative stabilities of these proteins may actually reverse at a sufficiently low temperature. rOM holo-b5 is almost certainly more stable in comparison with bMc holo-b5 at any given temperature than the relative apoprotein unfolding free energies at that temperature would indicate, however. Indeed, all available evidence indicates that rOM holo-b5 is considerably more stable than bMc holo-b5 at 25°C, in spite of our calculations indicating that rOM apo-b5 is 0.4 kcal/mol less stable than bMc apo-b5 at that temperature. As noted in the introduction, this may reflect the fact that conformationally disordered core 1 is more dynamic and expanded in folded bMc apo-b5 than in folded rOM apo-b5, which increases the entropic penalty associated with heme binding and thereby makes Gbind less favorable (Cowley et al., 2004).
The motif swap appears to stabilize both the folded and unfolded forms of rOM apo-b5
After demonstrating in our previous work that converting OM5M b5 to OM7M b5 via the H15R/E20S double mutation resulted in substantial increases in both holoprotein and apoprotein stability, we performed the reverse motif swap experiment with bMc b5 (Wang et al., 2006). Introducing the H15R/S20E double mutation into bMc b5 led to a modest 3.4°C decrease in holoprotein Tm, consistent with our expectations. However, bMcH15R/S20E apo-b5 was found to be 1.8 kcal/mol more stable than bMc apo-b5 at 25°C. We suggested that this increase in apoprotein unfolding free energy arose because the H15R/S20E double mutation destabilized unfolded bMc apo-b5 to a greater extent than it destabilized the folded apoprotein. Supporting this conclusion was the observation that the H15R/S20E double mutation substantially diminished residual structure present in urea-denatured bMc apo-b5 (as evidenced, for example, by an increase in RH), strongly suggesting disruption of energetically favorable interactions. This is consistent with results by Pace et al. (2000), who showed that stabilizing interactions among ionic residues can lead to unusually compact states for some unfolded proteins, and thereby affect the unfolding free energies of those proteins.
The residues involved in the bMc b5 motif highlighted in Fig. 2B clearly favor a compact denatured state, as further demonstrated by the following observations: (i) urea-denatured bMc apo-b5 has a considerably smaller RH value than urea-denatured rOM apo-b5 (Cowley et al., 2004); (ii) urea-denatured rOM and OM5M apo-b5 exhibit nearly identical RH values (Cowley et al., 2005); (iii) converting OM5M b5 to OM7M b5 via the R15H/E20S double mutation caused a change in apoprotein unfolding from a cooperative two-state reaction to a process that occurs in two sequential steps (Wang et al., 2006) and (iv) near the end of its first stage of urea-mediated denaturation OM7M apo-b5 is more compact than urea-denatured bMc and rOMR15H/E20S apo-b5, whereas fully urea-denatured OM7M apo-b5 has an RH value nearly identical to that of urea-denatured rOM apo-b5 (Table III).
The present study has revealed that urea-denatured rOMR15H/E20S apo-b5 is considerably more compact than urea-denatured rOM apo-b5, providing strong evidence that the former is more stable. The fact that the R15H/E20S double mutation significantly increased rOM apo-b5 unfolding free energy therefore allows us to conclude that it stabilized the folded apoprotein to a greater extent than the unfolded apoprotein. In the absence of this effect on the unfolded apoprotein, the stabilization achieved via the motif swap described herein would almost certainly have been considerably greater than observed.
Possible relevance to stabilization of natural thermophilic proteins
The factors stabilizing natural thermophilic proteins relative to their mesophilic homologues are often subtle, and vary from case to case. Surface electrostatic interactions, most notably salt bridges, have been shown to play a particularly important role, however (Xiao and Honig 1999; Kumar et al., 2000; Szilagyi and Zavodszky 2000). Moreover, recent studies have indicated that the number of salt bridges is generally less important for thermophilicity than their optimization on the protein surface (Greaves and Warwicker, 2007). Indeed, a number of researchers have made mesophilic proteins much more thermophilic by introducing favorable surface ion pairs, often by changing a single residue (Torrez et al., 2003; Gribenko and Makhatadze, 2007). Recall that the marked thermal stability enhancement obtained in the present work for rOM holo- and apo-b5 resulted from a motif swap in which two spatially close surface residues were mutated. The Arg15 to His mutation almost certainly maintained a positive charge, whereas the Glu11 to Ser mutation resulted in loss of a negative charge. The mutations altered the nature of salt bridging and hydrogen-bonding interactions among surface exposed side chains of residues 11, 15 and 20, and of stacking interactions between the side chains of residues 15 and 22. The R15H/E20S double mutation also entailed loss of four conformationally mobile side chain bonds.
As noted earlier, the substantial thermal stabilization in rOM holo- and apo-b5 resulting from the R15H/E20S double mutation arose by the most common general mechanism employed by nature to achieve high stability in thermophilic versus mesophilic proteins: raising the unfolding free energy across the entire unfolding curve. It is interesting to consider the possibility that it may also have been achieved by one of the most common natural mechanisms at the molecular level, optimization of salt bridges and other surface electrostatic interactions, and that a more favorable entropy change associated with formation of those interactions plays a key role.
|
Future directions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
In future studies, we intend to ascertain the extent to which the various interactions involving His15 (
-stack with Trp22; H-bonds with Glu11 and Ser20) contribute to the greater stability and more compact unfolded states exhibited by bMc versus rOM apo-b5 and OMR15H/E20S versus rOM apo-b5. We will also continue our protein engineering studies with rOM b5, with the ultimate goal of rendering both the ferric and ferrous forms of the protein stable in boiling water and under other extreme conditions. |
Supplementary Material |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Supplementary data are available at PEDS online
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
National Science Foundation (MCB-0446326).
|
Footnotes |
|---|
Edited by Susan Marqusee
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
We thank Prof. Russell Middaugh for use of his DSC and DLS equipment.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFuture directionsSupplementary MaterialFundingAcknowledgementsReferences |
|---|
Altman J., Lipka J.J., Kuntz I., Waskell L. Biochemistry (1989) 28:7516–7523.[CrossRef][Medline]
Altuve A., Silchenko S., Lee K.-H., Kuczera K., Terzan S., Zhang X., Benson D.R., Rivera M. Biochemistry (2001) 40:9469–9483.[CrossRef][Medline]
Altuve A., Wang L., Benson D.R., Rivera M. Biochem. Biophys. Res. Commun. (2004) 314:602–609.[CrossRef][Web of Science][Medline]
Atomi H. Curr. Opin. Chem. Biol. (2005) 9:166–173.[CrossRef][Web of Science][Medline]
Banci L., Bertini I., Cavazza C., Felli I.C., Koulougliotis D. Biochemistry (1998) 37:12320–12330.[CrossRef][Medline]
Beck von Bodman S., Schuler M.A., Jollie D.R., Sligar S.G. Proc. Natl Acad. Sci. USA (1986) 83:9443–9447.
Becktel W.J., Schellman J.A. Biopolymers (1987) 26:1859–1877.[CrossRef][Web of Science][Medline]
Berry B.W., Elvekrog M.M., Tommos C. J. Am. Chem. Soc. (2007) 129:5308–5309.[CrossRef][Web of Science][Medline]
Bruins M.E., Janssen A.E.M., Boom R.M. Appl. Biochem. Biotechnol. (2001) 90:155–186.[CrossRef][Web of Science][Medline]
Constans A.J., Mayer M.R., Sukits S.F., Lecomte J.T.J. Protein Sci. (1998) 7:1983–1993.[Web of Science][Medline]
Cowley A.B., Altuve A., Kuchment O., Terzyan S., Zhang X., Rivera M., Benson D.R. Biochemistry (2002) 41:11566–11581.[CrossRef][Medline]
Cowley A.B., Rivera M., Benson D.R. Protein Sci. (2004) 13:2316–2329.[CrossRef][Web of Science][Medline]
Cowley A.B., Sun N., Rivera M., Benson D.R. Biochemistry (2005) 44:14606–14615.[CrossRef][Medline]
Cowley A.B., Kennedy M.L., Silchenko S., Lukat-Rodgers G.S., Rodgers K.R., Benson D.R. Inorg. Chem. (2006) 45:9985–10001.[CrossRef][Web of Science][Medline]
de Miguel Bogas T., Barros-Velazquez J., Villa T.G. Protein Pept. Lett. (2006) 13:645–651.[CrossRef][Medline]
DeLano W.L. PyMol. (2002) San Carlos, CA: DeLano Scientific.
DiTursi M.K., Kwon S.J., Reeder P.J., Dordick J.S. Protein Eng. Des. Select. (2006) 19:517–524.[CrossRef]
Durley R.C.E., Mathews F.S. Acta Cryst. (1996) D52:65–76.[Web of Science]
Falzone C.J., Mayer M.R., Whiteman E.L., Moore C.D., Lecomte J.T.J. Biochemistry (1996) 35:6519–6526.[CrossRef][Medline]
Fernandez-Recio J., Vazquez A., Civera C., Sevilla P., Sancho J. J. Mol. Biol. (1997) 267:184–197.[CrossRef][Web of Science][Medline]
Flocco M.M., Mowbray S.L. J. Mol. Biol. (1994) 235:709–717.[CrossRef][Web of Science][Medline]
Freire E. Protein Stability and Folding, Theory and Practice—Shirley B.A., ed. (1995) Totowa, NJ: Humana Press. 191–218.
Gallivan J.P., Dougherty D.A. Proc. Natl Acad. Sci. USA (1999) 96:9459–9464.
Greaves R.B., Warwicker J. BMC Struct. Biol. (2007) 7.
Gribenko A.V., Makhatadze G.I. J. Mol. Biol. (2007) 366:842–856.[CrossRef][Web of Science][Medline]
Guzov V.M., Houston H.L., Murataliev M.B., Walker F.A., Feyereisen R. J. Biol. Chem. (1996) 271:26637–26645.
Hewson R., Newbold R.J., Whitford D. Protein Eng. (1993) 6:953–964.
Huntley T.E., Strittmatter P. J. Biol. Chem. (1972) 247:4641–4647.
Kawahara K., Tanford C. J. Biol. Chem. (1966) 241:3228–3232.
Kumar S., Tsai C.-J., Nussinov R. Protein Eng. (2000) 13:179–191.
Kuroda R., Ikenoue T., Honsho M., Tsujimoto S., Mitoma J., Ito A. J. Biol. Chem. (1998) 273:31097–31102.
Lederer F. Biochimie (1994) 76:674–692.[Medline]
Lederer F., Ghrir R., Guiard B., Cortial S., Ito A. Eur. J. Biochem. (1983) 132:95–102.[Web of Science][Medline]
Lee K.-H., Kuczera K. Biopolymers (2003) 69:260–269.[CrossRef][Web of Science][Medline]
Li W.F., Zhou X.X., Lu P. Biotechnol. Adv. (2005) 23:271–281.[CrossRef][Web of Science][Medline]
Loewenthal R., Sancho J., Fersht A.R. Biochemistry (1991) 30:6775–6779.[CrossRef][Medline]
Mathews F.S., Gerwinsky E.W., Argos P. The Porphyrins—Dolphin D., ed. (1979) New York: Academic Press. 107–147.
McLachlan S.J., LaMar G.N., Burns P.D., Smith K.M., Langry K.C. Biochim. Biophys. Acta (1986) 874:274–284.[CrossRef][Medline]
Mitchell J.B.O., Nando C.L., McDonald I.K., Thornton J.M., Price S.L. J. Mol. Biol. (1994) 239:315–331.[CrossRef][Web of Science][Medline]
Moore C.D., Al-Misky O.N., Lecomte J.T.J. Biochemistry (1991) 30:8357–8365.[CrossRef][Medline]
Myers J.K., Pace C.N., Scholtz J.M. Protein Sci. (1995) 4:2138–2148.[Web of Science][Medline]
Nojima H., Ikai H., Oshima T., Noda H. J. Mol. Biol. (1977) 116:429–442.[CrossRef][Web of Science][Medline]
Pace N.C. Methods Enzymol. (1986) 131:267–280.
Pace C.N., Alston R.W., Shaw K.L. Protein Sci. (2000) 9:1395–1398.[Web of Science][Medline]
Privalov P.L., Griko Y.V., Venyaminov S.Y., Kutyshenko V.P. J. Mol. Biol. (1986) 190:487–498.[CrossRef][Web of Science][Medline]
Razvi A., Scholtz J.M. Protein Sci. (2006) 15:1569–1578.[CrossRef][Web of Science][Medline]
Reetz M.T., Carballeira J.D., Vogel A. Angew. Chem. Int. Ed. (2006) 45:7745–7751.[CrossRef]
Renugopalakrishnan V., Garduno-Juarez R., Narasimhan G., Verma C.S., Wei X., Li P. J. Nanosci. Nanotechnol. (2005) 5:1759–1767.[CrossRef][Web of Science][Medline]
Rivera M., Barillas-Mury C., Christensen K.A., Little J.W., Wells M.A., Walker F.A. Biochemistry (1992) 31:12233–12240.[CrossRef][Medline]
Rodriguez-Maranon M.J., Qiu F., Stark R.E., White S.P., Zhang X., Foundling S.I., Rodriguez V., Schilling III C.L., Bunce R.A., Rivera M. Biochemistry (1996) 35:16378–16390.[CrossRef][Medline]
Samanta U., Pal D., Chakrabarti P. Acta Crystallogr. D (1999) 55:1421–1427.[CrossRef][Medline]
Silchenko S., Sippel M.L., Kuchment O., Benson D.R., Mauk A.G., Altuve A., Rivera M. Biochem. Biophys. Res. Commun. (2000) 271:467–472.
Simeonov M., Altuve A., Massiah M.A., Wang A., Eastman M.A., Benson D.R., Rivera M. Biochemistry (2005) 44:9308–9319.[CrossRef][Medline]
Sterner R., Liebl W. Crit. Rev. Biochem. Mol. Biol. (2001) 36:39–106.[CrossRef][Web of Science][Medline]
Sugiyama T., Miura R., Yamano T., Shiga K., Watari H. Biochem. Biophys. Res. Commun. (1980) 97:22–27.[CrossRef][Web of Science][Medline]
Sun N., Wang A., Cowley A.B., Altuve A., Rivera M., Benson D.R. Protein Eng. Des. Sel. (2005) 18:571–579.
Sylvestre J., Chautard H., Cedrone F., Delcourt M. Org. Process Res. Dev. (2006) 10:562–571.[CrossRef]
Szilagyi A., Zavodszky P. Structure (2000) 8:493–504.[Medline]
Tatko C.D., Waters M.L. Protein Sci. (2003) 12:2443–2452.[CrossRef][Web of Science][Medline]
Torrez M., Schultehenrich M., Livesay D.R. Biophys. J. (2003) 85:2845–2853.[Web of Science][Medline]
van den Burg B. Curr. Opin. Microbiol. (2003) 6:213–218.[CrossRef][Web of Science][Medline]
Vieille C., Zeikus G.J. Microbiol. Mol. Biol. Rev. (2001) 65:1–43.
Wang L., Sun N., Terzyan S., Zhang X., Benson D.R. Biochemistry (2006) 45:13750–13759.[CrossRef][Medline]
Wang L., Cowley A.B., Terzyan S., Zhang X., Benson D.R. Proteins (2007) 67:293–304.[CrossRef][Web of Science][Medline]
Xiao L., Honig B. J. Mol. Biol. (1999) 289:1435–1444.[CrossRef][Web of Science][Medline]
Yao P., Xie Y., Wang Y.-H., Sun Y.-L., Huang Z.-X., Xiao G.-T., Wang D. Protein Eng. (1997) 10:575–581.
Zhou Y., Hall C.K., Karplus M. Protein Sci. (1999) 8:1064–1074.[Web of Science][Medline]
Received March 13, 2007; revised September 12, 2007; accepted September 14, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||