PEDS Advance Access originally published online on August 28, 2008
Protein Engineering Design and Selection 2008 21(11):645-652; doi:10.1093/protein/gzn043
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Making a single-chain four-helix bundle for redox chemistry studies
1Department of Biochemistry and Biophysics, University of Pennsylvania, 905 Stellar-Chance Laboratories, Philadelphia, PA 19104-6059, USA 2Department of Biochemistry and Biophysics, Arrhenius Laboratories for Natural Sciences, Stockholm University SE-106 91, Stockholm, Sweden 3Department of Chemistry, Columbia University, New York, NY 10027, USA
4 To whom correspondence should be addressed.E-mail: tommos{at}mail.med.upenn.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResults and discussionFundingReferences |
|---|
The construction and characteristics of the stable and well-structured
4W protein are described. The 117-residue, single-chain protein has a molecular weight of 13.1 kDa and is designed to fold into a four-helix bundle. Experimental characterization of the expressed and purified protein shows a 69.8 ± 0.8% helical content over a 5.5–10.0 pH range. The protein is thermostable with a TM > 355 K and has a free energy of unfolding as measured by chemical denaturation of –4.7 kcal mol–1 at 25°C and neutral pH. One-dimensional (1D) proton and 2D 15N-HSQC spectra show narrow, well-dispersed spectral lines consistent with a uniquely structured -helical protein. Analytical ultracentrifugation and NMR data show that the protein is monomeric over a broad protein concentration range. The 324 nm emission maximum of the unique Trp-106 is consistent with a sequestered position of the aromatic residue. Additionally, differential pulse voltammetry characterization indicates an elevated peak potential for Trp-106 when the protein is folded (pH range 7.0–8.5) relative to partly unfolded (pH range 11.4–13.2). The oxidation of Trp-106 is coupled to proton release as shown by a 53 ± 3 mV/pH unit dependence of the peak potential over the 7.0–8.5 pH range.
Keywords: amino-acid radicals/four-helix bundle/NMR/protein design/Rop
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingReferences |
|---|
The repressor of primer (Rop) protein from Escherichia coli (E.coli) has been used extensively as a model system to study protein folding and sequence/structure relationships (Magliery and Regan, 2004
; Levy et al., 2005; Glykos et al., 2006; Schug et al., 2007). Here we use Rop as a template to create a stable and well-structured model protein for the purpose of studying protein redox chemistry and metal binding (Reedy and Gibney, 2004; Gibney and Tommos, 2005; Westerlund et al., 2005). The redesign of Rop was guided by the properties of the structurally characterized 3W model protein (Dai et al., 2002). 3W is a de novo designed, single-stranded 67-residue three-helix bundle originally made to study radical reactions associated with tryptophan and tyrosine residues. Three single-site variants have been made of this protein that contains either a tryptophan (3W), a tyrosine (3Y), or a cysteine (3C) at core position 32. 3W, 3Y and 3C are stable and uniquely structured proteins that have been used to study amino-acid radicals and quinone redox chemistry (Tommos et al., 1999; Dai et al., 2002; Hay et al., 2005; Hay et al., 2007). Here we aim to compliment and extend the three-helix bundle studies by creating a single-chain four-helix bundle with a well-structured hydrophobic core. The larger protein scaffold provides the opportunity to study various effects, such as an increase in solvent screening, on the redox properties of endogenous and exogenous redox cofactors residing within the core of the protein. The sequence of the Rop-based model protein described here is designed to form a single-chain 117-residue protein containing a unique buried tryptophan. Wild-type Rop is a homodimeric (2)2 four-helix bundle with an anti-parallel topology (Banner et al., 1987; Eberle et al., 1991). Creating a single-stranded 4 protein from Rop involved a significant redesign of the hydrophobic core and inter-helical interactions to promote the switch from an anti-parallel to a parallel topology as well as the addition of a loop that connects the two subunits. The design process was iterative and the protein secondary and tertiary structures, global stability, solution aggregation state and spectroscopic properties were evaluated for each step. The redox properties of the unique tryptophan were studied by differential pulse voltammetry (DPV) for the final stable and well-structured protein. |
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingReferences |
|---|
Amino-acid sequences and nomenclature used for the
4 proteins
For consistency we denote the initially designed single-chain protein 4 and then name the redesigned sequences relative to the 4 sequence. The 117-residue sequence of 4 reads as follows: GSKQEKTALNMADEVRSQ(T)KTVLEKLNKLDADEQADIFKSLADAADELERSVKARFGGGGGETKQEKTALNKAREIRSQ(T)KTLLEKVNELDADEIARVAESLADAWDE(L)ERSIKARF. The protein named 4(+F) in the main text represents a L109F single-site variant of 4; 4(+2F,–T) a T80F and L109F double-site variant of 4; 4(+F,–2T) a T19I and T80F double-site variant of 4 and finally 4(+2F,–2T) is a T19I, T80F and L109F triple-site variant of 4. The positions that were changed in the four variants are in parentheses in the 4 sequence shown earlier. The 2 (helix–loop–helix monomer), (2)2 (helix–loop–helix dimer) and 4 (single-chain four-helix bundle) notation used here follows the nomenclature by Ho and DeGrado (1987).
Gene synthesis of 4 and variants
The amino-acid sequence of 4 was back translated into a DNA sequence using E.coli preferred codons. A BamHI restriction site represents the two first residues (GS) in the 4 sequence. Three stop codons were added at the end of the 4 gene followed by an EcoRI restriction site and GC rich overhangs flanked both restriction sites. The resulting 378 base-pair sequence was divided into six 88 or 89 base-pair overlapping oligonucleotides, which were synthesized chemically (Cybergene, Sweden). In a stepwise version of the nested PCR strategy described by Dillon and Rosen (1993), the half-products from oligonuclotides number 1 + 2 + 3 + 4 and 3 + 4 + 5 + 6 were synthesized in separate PCR runs to prevent the incorrect association of oligos 2 and 5. The half-products were amplified separately and then combined in a final PCR experiment. The sequence of the resulting gene product was confirmed by DNA sequencing and three single base-pair insertion/deletion errors were corrected by site-directed mutagenesis. The 4(+F) and 4(+2F,–T) variants were made from the in-house generated 4 gene. The gene of the 4(+2F,–2T) variant was purchased from GenScript (www.genscript.com) and the 4(+F,–2T) variant was made from the commercially obtained gene. All mutations were made using the Stratagene QuikChange kit.
Protein expression of 4 and variants
The genes of the five proteins were digested with BamHI/EcoRI and ligated into a modified pET32b expression vector (Novagen) and co-expressed with thioredoxin in BL21(DE3)pLysS cells (Stratagene). The purification of the 4 proteins followed standard protocols for His6-tagged proteins. Cells were cultivated on rich or minimal media, harvested after a 3–6 h induction period and stored at –80°C. Frozen cells were thawed and resuspended in a nickel column binding buffer and lysed by sonication or in a French press. The lysate was clarified by centrifugation and the resulting supernatant was passed over a nickel column to isolate the thioredoxin-4 fusion protein. Thrombin (Sigma) was added to the eluate and the resulting mixture was dialyzed against thrombin cleavage buffer (20 mM Tris–HCl, 500 mM NaCl, 2.5 mM CaCl2, pH 8.0) overnight at room temperature. Following a buffer exchange, the digestion mixture was passed over a second nickel column to remove the His6-tagged thioredoxin and any remaining undigested fusion protein. Finally, the target 4 protein was purified by C18 reverse phase HPLC using a acetonitrile/water gradient containing 0.1% (v/v) trifluoroacetic acid. The molecular weights of 4 and variants were verified by matrix-assisted laser desportion-ionization (MALDI) mass spectrometry. The experimentally determined and (calculated) molecular weights of the five proteins are as follows: 4, 12 980 (12 980); 4(+F), 13 008 (13 014); 4(+2F,–T), 13 058 (13 060); 4(+F,–2T), 13 038 (13 038); 4(+2F,–2T), 13 072 (13 072) g mol–1.
Circular dichroism spectroscopy
Circular dichroism (CD) studies were performed on an Aviv 215 CD spectrometer equipped with a Pelletier thermoelectric controller. Protein unfolding/refolding transitions were monitored by measuring the ellipticity at 222 nm, 
222, as a function of pH, guanidine hydrochloride (Gdn:HCl) concentration, or temperature. The unfolding/refolding kinetics were determined to be sufficiently fast that the denaturations by both heat and temperature were complete within 2 min. Equilibration times were adjusting accordingly to ensure equilibrium conditions. Samples were prepared by dissolving lyophilized protein in a 20 mM potassium phosphate (KPi), 100 mM KCl, buffer. Final protein concentrations were in the 0.9–1.4 µM range as calculated by assuming an 
280 of 5600 M–1 cm–1. pH dependencies were investigated by titrating either HCl or KOH into the protein samples in steps of 0.05 pH units. Chemical denaturation of the 4 proteins was conducted by constant volume titration of 8 M Gdn:HCl, adjusted to pH 7.2, to a sample containing 20 mM KPi, 100 mM KCl, pH 7.2. The Gdn:HCl and buffer samples had the same protein concentration. The pH and chemical denaturation experiments were performed at 25°C and using a rectangular quartz cell with a 1.0 cm pathlength. Thermal denaturation studies were performed at pH 7.2 and using a capped 1.0 cm pathlength rectangular fluorescence cell. Temperature was increased in one degree increments between 5 and 110°C, with a 5 min equilibration time after each increment. After a thermal denaturation experiment, the temperature was held at the final point (95 or 105°C) for 5 min, and then returned to 25°C to measure the signal recovery. The proteins regained at least 80% of the initial signal even after exposure to high temperatures of up to 4 h. The helical content of the proteins were estimated from [
]222/–39.5, where []222 is the mean residue ellipticity at 222 nm in degrees cm2 dmol–1 residue–1 and the denominator represents the mean residue ellipticity of an infinitely long -helix. [] is given by /10lCn, where is the measured ellipticity in degrees, l the pathlength in cm, C the molar protein concentration, and n is the number of residues. The chemical and thermal denaturation curves were fit to two-state monomer folded to monomer unfolded equilibrium models using a nonlinear least-squares routine in Kaleidagraph (www.synergy.com). The Gdn:HCl denaturation curves were fitted by the method of Santoro and Bolen (1988), in which the baselines representing the folded and unfolded protein, the stability of the protein in the absence of denaturant 
G (kcal mol–1), and the m value (kcal mol–1 M–1) were allowed to vary. The thermal denaturation curves were fit to the Gibbs–Helmholtz equation (Greenfield, 2004):
| 1 |
H (kcal mol–1) is the change in enthalpy at TM, and CP (kcal mol–1 K–1) is the change in heat capacity between the folded and unfolded protein. The slopes of the baselines representing the folded and unfolded protein were included as fixed parameters in the fit.
Fluorescence spectra were collected on a Horiba Jobin Yvon Spex Fluorolog spectroflurometer equipped with a Neslab RTE-111 refrigerated bath/circulator. Samples were prepared by dissolving lyophilized protein in a 50 mM KPi, 30 mM KCl, pH 8.0 buffer to an A290 of 0.15–0.20 (1.0 cm path). The protein samples were then exchanged into a 5 mM KPi, 5 mM KCl pH 6.5 buffer and the A290 adjusted to 0.10 (1.0 cm path, protein concentration 
20 µM assuming an 290 of 5000 M–1 cm–1). Spectra were recorded at 25°C using 
ex = 290 nm and em = 295–445 nm. The slit width for the excitation and emission light was 0.8 and 2.0 nm, respectively, and the data averaging time for each 0.05 nm step was 0.2 s.
Sedimentation equilibrium analytical ultracentrifugation
Analytical ultracentrifugation analyses were performed using a Beckman XL-I analytical ultracentrifuge operating at 20 000, 25 000 and 30 000 r.p.m. and 4°C. Samples were prepared by dissolving lyophilized protein in a 20 mM KPi, 100 mM KCl, pH 8.0 buffer to an A280 of 0.35 (1.0 cm path, protein concentration 60 µM). Partial specific volumes 
, (ml g–1) for the proteins were calculated from the residue-weighted average of the amino-acid sequence using the method of Cohn and Edsall (Laue et al., 1992). The density, 
(g ml–1), of the solvent buffer was calculated using the Sedenterp software (Laue et al., 1992). The radial distribution absorbance scan data at 280 nm were fit to a single exponential using the WinNonlin software (www.pharsight.com). The buoyant molecular weight, Mb (g mol–1), was converted to the average molecular weight of the molecular species in solution, Mr (g mol–1), using the following relationship:
| 2 |
NMR experiments were performed on a Varian Inova 500 MHz spectrometer. Samples were prepared by dissolving lyophilized protein in 2.5 ml of a 50 mM KPi, 30 mM KCl, pH 8.0 buffer to an A290 of 0.5 (1.0 cm path). The proteins were then exchanged into 3.5 ml of a 5 mM KPi, 5 mM KCl, pH 6.5 buffer, freeze-dried, and finally dissolved in 650 µl D2O/H2O containing 250 µM 2,2-dimethyl-2-silapentane-5-sulfonate sodium salt. The protein concentration was 0.5 mM in all samples. Data processing was performed with the Felix95 software (Molecular Simulations Inc., San Diego, CA, USA). NMR diffusion measurements were performed with the pulsed field gradient spin-echo experiment (PFG-LED) (Gibbs et al., 1991) in 5 mm NMR tubes. A 600 MHz Varian Inova spectrometer equipped with a HCX probe with z-axis gradient was used for these experiments. To correct for systematic errors due to nonlinearity of the gradient field, a distribution function for the gradient strength was used when calculating the diffusion coefficient (Damberg et al., 2001). This function was calibrated using a standard sample of 99.95% D2O. The diffusion experiments were performed using 30 different linearly spaced strengths of gradients with 150 ms longitudinal storage time. Diffusion experiments were done at 25°C.
Measurements were performed with an Autolab PGSTAT12 potentiostat equipped with a conventional three-electrode glass cell containing a glassy carbon working electrode, a platinum wire counter electrode, and a Ag/AgCl (3.0 M NaCl) reference electrode as described earlier (Hay et al., 2005). The potentials given here are relative to the normal hydrogen electrode. All solutions for electrochemistry were prepared with milli-Q water and all measurements were made at room temperature under an argon atmosphere. The peak positions and peak heights of each voltammogram were determined with the Autolab GPES software.
|
Results and discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingReferences |
|---|
Design of
4
The aim of this study was to make a stable and well-structured model protein that subsequently will be used as a platform to study tryptophan and tyrosine radicals, quinone redox chemistry, and metal binding. The model protein should be single stranded and, initially, contain a single tryptophan but no tyrosine, cysteine or histidine residues. These residues will be introduced at a later stage to change the amino-acid radical species (Trp to Tyr) or to direct the binding of quinones (Trp to Cys) or metal ions (addition of His or Cys residues). The sequence of the model protein was derived from E.coli Rop as described earlier (Westerlund et al., 2005). Rop is a homodimer containing two 63-residue helix–loop–helix subunits (Banner et al., 1987; Eberle et al., 1991). The two subunits have an anti-parallel orientation, which places the C- and N-termini of each monomer on opposite sides of the protein. To make a single-chain 4 protein by connecting the C-terminus of subunit 1 to the N-terminus of subunit 2 requires a parallel orientation of the two 2 subunits (see Supplementary Fig. S1 available at PEDS online for a schematic illustration of the Rop redesign). About 60% of the Rop heptad a and d core residues were changed following a ‘small versus large residue’ packing pattern (Fig. 1) with the aim to induce and stabilize an 180° reorientation of one of the subunits. The unique tryptophan was placed at heptad a position 106 and a five-residue glycine loop was introduced to connect the two subunits. Multiple alterations were also made at inter-helical positions to exchange potentially unfavorable electrostatic interactions to favorable ones (for more details see Westerlund et al., 2005).
|
The approach used here differs significantly from earlier studies on the design and synthesis of single-chain Rop variants. As described earlier, our main emphasis is on redesigning hydrophobic and electrostatic inter-helical interactions to change the topology of the two
2 subunits from an anti-parallel to a parallel orientation while maintaining the LDA hairpin bends connecting the individual helices. In contrast, Predki and Regan (1995) and Kresse et al. (2001) maintained the helical interactions and the orientation of the helices but redesigned the non-helical loop regions to change the connections between the helices and form single-chain Rop proteins.
The 4 far-UV CD spectrum has been shown in a preliminary report on the properties of 4 (Westerlund et al., 2005; We note that the protein originally introduced as 4W by Westerlund et al. is the 4 protein described and more fully characterized here.). The spectrum represents the protein at neutral pH and displays the characteristic features of an -helical coiled coil with an ellipticity () maximum 195 nm, minima at 208 and 222 nm, and a 222/208 ratio of 0.96. Consistent with the CD data, one- and two-dimensional (2D) NMR spectra of 4 display spectral features suggesting a predominantly helical structure (data not shown). Resonances assigned to main chain -protons in the 1D proton spectrum of 4 are found in the 3.6–4.4 p.p.m. spectral region, which is consistent with the chemical shifts of protons involved in -helical hydrogen bonding (Wang and Jardetzky, 2002). The pH sensitivity of the helical content was examined by monitoring 222 as a function of solution pH. As shown in Supplementary Fig. S2 available at PEDS online, the -helical content of 4 is pH independent between pH 5.5 and 10.0. From the mean residue ellipticity value at 222 nm, []222, we estimate that 4 is 71.0 ± 0.6% helical in this pH range (Table I). As the pH is lowered below 5.5, the helical content drops by 10–15%. At pH > 10, the helical content drops to reach a value of 50% at pH 12. We conclude that the secondary structure of 4 is fairly robust toward changes in the solution pH.
|
The global stability of
4 was investigated by chemical and thermal denaturation. The data from these studies are displayed in Fig. 2A and Supplementary Fig. S3A available at PEDS online and the results from the data analyses are summarized in Table I. Figure 2A displays the reversible denaturation of 4 by Gdn:HCl. The solid gray line in Fig. 2A represents a least-squares fit of the experimental data (open circles) in which the baselines of the folded and unfolded state of the protein, the protein stability at zero molar denaturant, G, and the m value of the folding/unfolded transition were allowed to vary (Santoro and Bolen, 1988). The data fit yields a stability of –2.2 kcal mol–1 and a m value of 2.7 kcal mol–1 M–1. The relatively low stability of 4 as shown by chemical denaturation was confirmed by thermal denaturation of the protein (Supplementary Fig. S3A is available at PEDS online). The solid gray line represents a least-square fit of the experimental data (open circles) in which the midpoint temperature of the unfolding transition TM, and the change in enthalpy, H, and heat capacity, CP, between the folded and unfolded state of 4, were allowed to vary. The thermal melt estimates the 4 stability to –3.2 kcal mol–1 at 25°C.
|
Fluorescence spectroscopy and NMR spectroscopy were used to probe the local environment of Trp-106 and the overall tertiary structure of
4. The emission maximum (Emmax) of the Trp-106 fluorescence spectrum is at 330 nm (Table I), which suggests that the aromatic residue is partly shielded from the solvent. As a comparison, Emmax of N-acetyl-L-tryptophanamide solvated in phosphate buffer is 352 nm (Tommos et al., 1999) while Emmax of Trp-32 in 3W (accessible surface area <3%; Dai et al., 2002) is 324 nm. Analytical sedimentation equilibrium ultracentrifugation was used to investigate the solution aggregation state of 4. A sample containing 60 µM protein dissolved in a 20 mM KPi, 100 mM KCl, pH 8.0 buffer was run at 4°C at a speed of 20, 25 and 30 k r.p.m. The data were collected by measuring the absorbance at 280 nm as a function of rotor distance. The three data sets from the same sample were fit as a set to a single species with a molecular weight of 14 479 g mol–1 (data not shown). This value is very close to the 12 980 g mol–1 value we calculate from the protein sequence and obtained from mass spectrometry (Table I). This conclusion was confirmed by NMR diffusion measurements, which showed that the protein is monomeric in the 100 µM concentration range (data not shown). Thus the protein is monomeric in the concentration range used for all of the optical measurements described earlier. The 1D proton NMR spectrum of 4 shows a rather limited dispersion of 1.8 p.p.m. in the amide region (Fig. 3A) and the poor dispersion and broad spectral lines observed of the 2D 13C (Fig. 4A) and 15N (data not shown) HSQC spectra suggest a dynamic ensemble of structures rather than a uniquely structured protein. Protein aggregation at the higher protein concentrations (0.5 mM) used for NMR measurements may also contribute to the broad NMR lines.
|
|
To conclude,
4 is 71.0 ± 0.6% helical between pH 5.5 and 10.0. The helical content drops at pH values outside this range. The stability of the protein is fairly low, with chemical and thermal denaturation data showing a stability of –2.2 and –3.2 kcal mol–1, respectively. An Emmax of 330 nm for Trp-106 suggests partial solvent screening. 4 is monomeric at low protein concentrations (
100 µM), although aggregation may occur at higher concentrations.
The characterization of 4 resulted in two clear conclusions: the global stability and the chemical shift dispersion of the single-chain model protein must be improved. The latter is important in order to reduce spectral degeneracy and make the final protein suitable for NMR structural characterization. This point is crucial because of the all-helical nature of the model protein, which will limit the overall spectral dispersion and make resonance and NOE assignments more problematic. The predicted -helical segments of 4 follow a classic coiled-coil pattern with largely apolar residues in the heptad a and d positions. However, the fifth core packing layer (T19, L41, T80 and L102; Fig. 1) does contain two threonine residues. In an attempt to boost the stability of the protein by increasing the overall buried hydrophobic surface area, T19 and T80 residues were replaced by apolar residues. Visual inspection of an 4 working model based on the Rop crystal structure (1rop.pdb; Banner et al., 1987) suggested that the side chains of I19, L41, F80 and L102 could come together and form a packing layer that roughly fits with the layers above and below. The T80F change will also increase the total number of aromatic residues in the protein sequence. To facilitate a full NMR structural characterization it is critical to reduce spectral overlap, particularly of those resonances that are associated with the atoms that reside in the protein interior. Aromatic side chains will provide ring current shifts that are likely to significantly influence the chemical shift of nearby atoms. By placing a second phenylalanine at position 109, five out of the eight predicted packing layers now contain aromatic residues (Fig. 1). To reflect the two main redesign criteria, i.e. the removal of hydrophilic residues from predicted core positions (to increase protein stability) and the addition of phenylalanine residues (to increase NMR spectral resolution), the 4 variants are denoted as follows: 4(+F) (L109F, addition of one Phe), 4(+2F,–T) (T80F and L109F, addition of two Phe and removal of one Thr), 4(+F,–2T) (T19I and T80F, addition of one Phe and removal of two Thr), and 4(+2F,–2T) (T19I, T80F and L109F, addition of two Phe and removal of two Thr).
Characterization of 4 variants
Since none of the changes in the amino-acid sequence involve residues with aqueous pKa values in the pH 5.5–10.0 range, the pH stabilities of the helical contents in the 4 variants were expected to be similar to the structural pH stability of 4. Supplementary Fig. S2B available at PEDS online shows that this is indeed the case. The degree of secondary structure for the four 4 variants show little variance over the 5.5–10.0 pH range (Table I). In addition, the -helical content of the 4 variants are all close to the 71.0 ± 0.6% value of 4, with the exception of 4(+F,–2T) for which the helical content is 63.6 ± 0.4% (pH 5.5–10.0, Table I). The global stability of the model protein appears to be sensitive to the number of threonine residues at predicted core positions. The 4(+F,–2T) and 4(+2F,–2T) proteins in which both T19 and T80 were replaced by hydrophobic residues have a global stability at 25°C of –4.6 and –4.7 kcal mol–1, respectively, which is more stable than 4 (–2.2 kcal mol–1), 4(+F) (–3.0 kcal mol–1), and 4(+F,–T) (–2.6 kcal mol–1). The chemically induced unfolding transitions of the four 4 variants are shown in Fig. 2B and the corresponding free energy of unfolding values are listed in Table I. The thermal stability of the protein also appears to depend on the T19I and T80F changes. The TM of 4, 4(+F) and 4(+2F,–T) are 338, 339 and 346 K, respectively, while the TM values of 4(+F,–2T) and 4(+2F,–2T) are estimated at >355 K (Supplementary Fig. S3 is available at PEDS online; Table I).
The emission spectra of Trp-106 in the 4 variants are blue shifted by various extents relative to the emission spectrum of Trp-106 in 4 (Table I). The tryptophan emission of 4(+F,–2T) protein is shifted by a modest 2 nm relative to the 4 emission spectrum (Emmax 328 versus 330 nm), which suggests that Trp-106 is partly solvent exposed in these two proteins. In contrast, 4(+F), 4(+2F,–T) and 4(+2F,–2T) all display an emission maximum close to 324 nm indicating a very limited or no solvent exposure of the aromatic residue in these proteins.
Analytical sedimentation equilibrium ultracentrifugation was used to determine the aggregation state of the 4 variants. All four variants were found to be monomeric at a protein concentration 60 µM (Table I). NMR characterization showed that 4(+2F,–2T) is also monomeric at higher protein concentrations. The broadened line shapes observed in the HSQC data of 4 (Fig. 4A), 4(+F), 4(+2F,–T) (data not shown), and to lesser extent in the data from 4(+F,–2T) (Fig. 4B), are absent in the data from 4(+2F,–2T) (Fig. 4C). The 15N-HSQC spectrum of 4(+2F,–2T) displays narrow spectral lines consistent with a non-aggregated protein. The overall chemical shift dispersion and spectral resolution is moreover consistent with a uniquely structured protein. The expected 117 amide cross-peaks can easily be identified in the 15N-HSQC spectrum, which suggests that the protein is suitable for NMR structural studies. Spectral differences are also observed in the 1D proton NMR spectra of the five proteins (Fig. 3), although they are less dramatic relative to the 2D data.
Electrochemical characterization of 4(+2F,–2T)
DPV was used to investigate the redox characteristics of 4(+2F,–2T) in aqueous solution. DPV peak potentials as the function of pH are shown in Fig. 5. The data points obtained in the pH 7.0–8.5 range represents the fully folded protein (closed circles) while the data in the pH 11.4–13.2 range represents a partly unfolded protein (Fig. 1B). The DPV peak potential is pH dependent and a linear fit to the data yields a slope of 53 ± 3 mV/pH unit in the pH 7.0–8.5 range and 52 ± 6 mV in the 11.4–13.2 range. These numbers are close to the theoretical value of 59 mV/pH unit dependence of a one electron/one proton redox reaction at room temperature. Thus, the electrochemical process is charge neutral. The magnitude of the apparent 100 mV increase in DPV peak potential in the fully folded protein (relative to the alkaline partly unfolded protein) is similar to the increase in DPV peak potential observed upon burying a phenol hydroxyl group within the core of the 3C protein (Hay et al., 2005).
|
Concluding remarks
This paper describes the design, construction and properties of a stable and well-structured four-helix bundle. The single-chain protein contains 117 amino acids and has a molecular weight of 13.1 kDa. The -helical content is 69.8 ± 0.8% over a pH range of 5.5–10.0. Chemical denaturation yields a global stability of –4.7 kcal mol–1 at 25°C and pH 7.2 and thermal denaturation predicts a TM> 355 K. Analytical ultracentrifugation and NMR data show that the protein is monomeric over a broad protein concentration range. The 1D proton and 2D 15N-HSQC spectra show narrow spectral lines consistent with a uniquely structured protein. The protein contains a single tryptophan whose fluorescence properties suggest that this residue is located in a sequestered position. DPV characterization indicates an elevated peak potential for this tryptophan in the folded form of the protein (pH range 7.0–8.5) relative to a partly unfolded form (pH range 11.4–13.2). Oxidation of the tryptophan is coupled to proton release as indicated by a 53 ± 3 mV/pH unit dependence of the peak potential of the 7.0–8.5 pH range.
Throughout the text the preliminary name of 4(+2F,–2T) is used for the protein showing the characteristics summarized in the previous paragraph. We end by renaming 4(+2F,–2T) as 4W to reflect that it is a single-stranded four-helix bundle containing a unique tryptophan residue and to follow the nomenclature used for our 3W-based proteins. The 4W protein will be used as a platform for future studies involving protein radicals, quinone cofactors and metal ions.
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingReferences |
|---|
National Science Foundation (DMR-0520020); Swedish Research Council (C.T.); American Heart Association Grant-in-Aid (0755879T to B.R.G); National Science Foundation GK-12 Graduate Fellowship (DGE-02-31875 to S.D.M); Fulbright Fellowship (H.P.); Wenner-Gren Foundation (Stiftelsen Wenner-Grenska Samfundet to S.H.).
|
Footnotes |
|---|
Edited by Professor Lars Baltzer
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResults and discussionFundingReferences |
|---|
Banner D.W., Kokkinidis M., Tsernoglou D. J. Mol. Biol. (1987) 196:657–675.[CrossRef][Web of Science][Medline]
Dai Q.-H., Tommos C., Fuentes E.J., Blomberg M.R.A., Dutton P.L., Wand A.J. J. Am. Chem. Soc. (2002) 124:10952–10953.[CrossRef][Web of Science][Medline]
Damberg P., Jarvet J., Gräslund A. J. Magn. Reson. (2001) 148:343–348.[CrossRef][Web of Science][Medline]
Dillon P.J., Rosen C.A. Methods in Molecular Biology—White B.A., ed. (1993) vol. 15. Totowa, NJ, USA: Human Press Inc. 263–268.
Eberle W., Pastore A., Sander C., Rosch P. J. Biomol. NMR (1991) 1:71–82.[CrossRef][Medline]
Gibbs S.J., Morris K.F., Johnson C.S. J. Magn. Reson. (1991) 94:165–169.[Web of Science]
Gibney B.R., Tommos C. Photosystem II: The Light-Driven Water:Plastoquinone Oxidoreductase.—Wydrzynski T.J., Satoh K., eds. (2005) The Netherlands: Springer. 729–751.
Glykos N.M., Papanikolau Y., Vlassi M., Kotsifaki D., Cesareni G., Kokkinidis M. Biochemistry (2006) 45:10905–10919.[CrossRef][Web of Science][Medline]
Greenfield N.J. Methods Enzymol. (2004) 383:282–317.[Web of Science][Medline]
Hay S., Westerlund K., Tommos C. Biochemistry (2005) 44:11891–11902.[CrossRef][Web of Science][Medline]
Hay S., Westerlund K., Tommos C. J. Phys. Chem. B (2007) 111:3488–3495.[Medline]
Ho S.P., DeGrado W.F. J. Am. Chem. Soc. (1987) 109:6751–6758.[CrossRef][Web of Science]
Kresse H.P., Czubayko M., Nyakatura G., Vriend G., Sander C., Bloecker H. Protein Eng. (2001) 14:897–901.
Laue T.M., Shah B.D., Ridgeway T.M., Pelletier S.L. Analytical Ultracentrifugation in Biochemistry and Polymer Science.—Harding S.E., Rowe A.J., Horton J.C., eds. (1992) London, UK: Royal Society of Chemistry. 90–125.
Levy Y., Cho S.S., Shen T., Onuchic J.N., Wolynes P.G. Proc. Natl Acad. Sci. USA (2005) 102:2373–2378.
Magliery T.J., Regan L. Protein Eng. Des. Sel. (2004) 17:77–83.
Predki P.F., Regan L. Biochemistry (1995) 34:9834–9839.[CrossRef][Web of Science][Medline]
Reedy C.J., Gibney B.R. Chem. Rev. (2004) 104:617–649.[CrossRef][Web of Science][Medline]
Santoro M.M., Bolen D.W. Biochemistry (1988) 27:8063–8068.[CrossRef][Web of Science][Medline]
Schug A., Whitford P.C., Levy Y., Onuchic J.N. Proc. Natl Acad. Sci. USA (2007) 104:17674–17679.
Tommos C., Skalicky J.J., Pilloud D.L., Wand A.J., Dutton P.L. Biochemistry (1999) 38:9495–9507.[CrossRef][Web of Science][Medline]
Wang Y.J., Jardetzky O. Protein Sci. (2002) 11:852–861.[CrossRef][Web of Science][Medline]
Westerlund K., Berry B.W., Privett H.K., Tommos C. Biochim. Biophys. Acta (2005) 1707:103–116.[Medline]
Received April 9, 2008; revised July 2, 2008; accepted July 29, 2008.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||