Skip to main content
NCBI home page
Proceedings of the National Academy of Sciences of the United States of America logoLink to Proceedings of the National Academy of Sciences of the United States of America
. 2003 Jan 27;100(7):3760–3765. doi: 10.1073/pnas.0336055100

Hg(II) binding to a weakly associated coiled coil nucleates an encoded metalloprotein fold: A kinetic analysis

Brian T Farrer 1, Vincent L Pecoraro 1,*
PMCID: PMC152995  PMID: 12552128

Abstract

A detailed kinetic analysis of metal encapsulation by a de novo-designed protein is described. The kinetic mechanism of Hg(II) encapsulation in the three-stranded coiled coil formed by the peptide CH3CO-G LKALEEK CKALEEK LKALEEK G-NH2 (Baby L9C) is derived by global analysis. The mechanism involves rapid initial collapse of two peptides by Hg(II) forming Hg(Baby L9C-H)2 with a linear thiolato Hg(II) bound to the cysteine sulfur atoms. Here, Baby L9C-H denotes Baby L9C with the cysteine thiol deprotonated. Addition of the third peptide, forming the three-stranded coiled coil, is the rate-determining step and results in an intermediate state involving two separate species. One of the species, termed the properly folded intermediate, undergoes rapid deprotonation of the third cysteine thiol, yielding the desired three-stranded coiled coil with an encapsulated trigonal thiolato Hg(II). The other species, termed the misfolded intermediate, rearranges in an experimentally distinguishable step to the properly folded intermediate. The order of the reaction involving the addition of the third peptide with respect to the concentration of Baby L9C indicates that addition of the third helix only occurs through reaction of Hg(Baby L9C-H)2 and Baby L9C that is unassociated with a coiled coil. Temperature dependence of the reaction afforded activation parameters for both the addition of the third helix (ΔH‡ = 20(2) kcal/mol; ΔS‡ = 40(5) cal/mol K) and the rearrangement of the misfolded intermediate steps (ΔH‡ = 23(2) kcal/mol; ΔS‡ = 27(5) cal/mol K). The mechanism is discussed with regard to metalloprotein folding and metalloprotein design.

Metalloproteins are ubiquitous throughout nature. In metalloproteins, the metal site is necessary for proper function through either stabilization of the overall structure or direct involvement in the active center of the protein. Design of metalloproteins is concerned with the structure of the polypeptide backbone, the coordination environment around the metal, and how the forces governing both interact to generate the overall structure of the metalloprotein.

An integral part of understanding the overall structure of a metalloprotein is to understand how, in time, the metalloprotein achieves its overall structure. Current research on natural protein systems has yielded information that both the metal and the structure of the polypeptide backbone play key roles in the folding processes for many proteins. Demonstrating the importance of metal coordination, the folding of ferricytochrome c from its chemically denatured form occurs through a multistep process where the metal coordination environment plays a pivotal role in determining the kinetic profile (1, 2). Highlighting the role of the polypeptide backbone, the addition of Cu(II) to azurin has been shown to occur faster in the unfolded apoenzyme than to the structured apo enzyme (3–6).

We have been investigating designed peptides that bind heavy metals. This family of peptides is based on the parent peptide CH3CO-G(LKALEEK)4G-NH2 (TRI) (7, 8). We intend to elucidate the structures and properties of heavy metal–protein interactions and to reveal the protein forces involved in controlling coordination environments around metal centers. At high pH (pH > 7.0), TRI assembles into a three-stranded coiled coil with a leucine core. By replacing one of the leucines with cysteine on each peptide strand, a metal-binding site presenting a trigonal thiol/thiolate environment to a metal is created. We reported trigonal thiolate complexation of As(III), Hg(II), and Cd(II) by these peptides (9, 10). In addition, we developed a detailed thermodynamic model for Hg(II) binding to CH3CO-GLKALEEK CKALEEK(LKALEEK)2G-NH2 (TRI L9C) and CH3CO-GLKALEEKCKALEEKLKALEEKG-NH2 (Baby L9C), a 23-aa peptide that forms a weakly associated three-stranded coiled coil in the absence of metal (Table 1) (11). This thermodynamic model led to the proposal of a three-step mechanism as a minimal kinetic model for Hg(II) encapsulation by Baby L9C.

Table 1.

Peptide sequences

Peptide Sequence
TRI CH3CO- G LKALEEK LKALEEK LKALEEK LKALEEK G-NH2
TRI L9C CH3CO- G LKALEEK CKALEEK LKALEEK LKALEEK G-NH2
Baby L9C    CH3CO- G LKALEEK CKALEEK LKALEEK G-NH2
Open in a new tab

Here, we describe a kinetic investigation of Hg(II) insertion into Baby L9C. We propose a mechanism that retains the basic outline of the minimal mechanism proposed from the thermodynamic data and thoroughly describes the kinetic data (Fig. 1). This mechanism includes a misfolded intermediate likely to be an antiparallel three-stranded coiled coil with encapsulated linear thiolato Hg(II).

Figure 1.

Figure 1

Open in a new tab

Proposed kinetic mechanism for the encapsulation of Hg(II) by the three-stranded coiled coil (Baby L9C)3. The parameters for experiments done at 10°C (pH 8.5) are Kassoc = 4(1) × 1010 M−2, pKa = 7.6(2), k0 is fast on the time scale of the experiment, k1 = 1.4(2) × 105 M−1⋅sec−1, k−1 = 0.20(3) sec−1, k2 = 2.4(8) × 104 M−1⋅sec−1, k−2 = 0.9(3) sec−1. Kinetic data are derived from mechanism E in Scheme S1. The rate constants governing Kassoc and pKa are fast on the time scale of the experiment.

Methods

Synthesis of Peptides.

Baby L9C was synthesized on an Applied Biosystems 433A peptide synthesizer by using standard protocols (12) and purified and characterized as described (11). The stock solution concentrations were determined by using the Elmann test (13).

Circular Dichroism (CD) Titrations.

CD titrations were performed on an AVIV 14D spectrophotometer attached to a temperature-control bath. Samples containing 0, 6, 30, 60, 116, and 213 μM Baby L9C in 50 mM phosphate containing 100 mM KCl and 1 mM Tris(caboxyethyl)phosphine hydrochloride were titrated to pH 8.5. Spectra were taken at 8, 13, 18, 23, 28, 33, and 38°C from 210 to 300 nm. The path lengths of the quartz cells were 1.0 mm for [Baby L9C] ≥ 30 μM and 10 mm for [Baby L9C] < 30 μM.

The CD titration data were evaluated by using the association equilibrium for a three-stranded coiled coil defined by the equilibrium constant Kassoc:

graphic file with name M1.gif 1

Kassoc was determined by fitting [peptide]folded/[peptide]total vs. [peptide]total. MAPLE 7 (Waterloo Maple, Ontario, Canada) was used to solve for the fraction of folded peptide with respect to [Baby L9C]total by using Kassoc = [(Baby L9C)3]/[Baby L9C]3 and [Baby L9C]total = [Baby L9C]unfolded + 3[(Baby L9C)3] (see Supporting Text, which is published as supporting information on the PNAS web site, www.pnas.org).

Stopped-Flow Spectroscopy.

Stopped-flow spectroscopy was performed on an Olis-RSM stopped-flow spectrometer monitoring 250–400 nm. The spectrum was observed after mixing (mixing time ≈ 4 msec) for 1–3 sec at 1,000 scans per sec or for 10 sec at 62 scans per sec. Reactions of 5 μM HgCl2 with 50 μM Baby L9C were performed by mixing equal volumes of 10 μM HgCl2, 100 mM KCl, and 50 mM K+ phosphate (desired pH) with a solution of 100 μM Baby L9C, 100 mM KCl, and 50 mM K+ phosphate (desired pH). The temperature was controlled to within ±1°C by a Neslab Instruments (Portsmouth, NH) RTE-111 refrigerated circulating bath and set to 10°C unless otherwise stated.

Concentration-dependent studies were performed with 0–395 μM Baby L9C. pH-dependent studies were done at pH 7.0, 7.5, 8.0, 8.5, and 9.0 by making phosphate solutions of the desired pH before the addition of HgCl2 and peptide. All the studies were done in the presence of 0.10 M KCl and 50 mM phosphate buffer. The concentration of KCl did not affect the rate of the reactions noticeably. The addition of peptide acidified the solution slightly, and the pH was readjusted by using a small aliquot of 1 M KOH.

Hg(Baby L9C-H)2 (Baby L9C with a deprotonated cysteine thiol) was generated by adding a one-half equivalent of HgCl2 to a Baby L9C stock solution of known concentration (≈5 mM). The stock solution then was diluted to make a 10 μM Hg(Baby L9C-H)2 solution that was used in the experiment.

Data obtained from the stopped-flow experiments were analyzed by using global analysis by singular value decomposition with SPECFIT (Spectrum Software, Chapel Hill, NC) as described in Results.

Supporting Information.

Plots of ln(k/T) vs. 1/T for k1 and k−2, an overlay of reaction profiles using HgCl2 and Hg(Baby L9C-H)2 as starting materials, and the rate expression for mechanism E shown in Scheme S1 at high peptide concentration are provided in Supporting Text and Figs. 6–8, which are published as supporting information on the PNAS web site. A comparison of the results using this expression and those obtained by using SPECFIT are provided.

Scheme 1.

Scheme 1

Open in a new tab

Kinetic mechanisms used to fit stopped-flow data.

Results

Concentration Dependence of Coiled-Coil Association.

The coiled-coil association constants at various temperatures were determined at pH 8.5 via titration of Baby L9C into a solution of Hg(II) by monitoring the α-helical signature in the UV CD spectrum. A peptide within a coiled coil has greater α-helical character than a peptide that is not associated with a coiled coil, leading to a more negative CD signal at 222 nm for those peptides within a coiled coil. The titration showed a decrease in the molar absorptivity of the solutions with increasing [Baby L9C], consistent with the equilibrium expression of three-stranded coiled-coil formation (Eq. 1). The fit to the data at 10°C indicated an association constant (Kassoc) of 3.5 × 1010 M−2 (ΔG° = −14 kcal/mol).

Temperature Dependence of Coiled-Coil Association.

Temperature affects two properties of the CD titrations (Fig. 2). First, the overall curvature of the titrations decreases with increasing temperature. Second, the endpoint (the signal due to the assembled form) decreases with increasing temperature. These results are consistent with a thermodynamic model involving two steps:

graphic file with name M2.gif 2

and

graphic file with name M3.gif 3

The partially folded state of the (Baby L9C) can be understood by considering N and C termini fraying. With this interpretation, the fraying increases with increasing temperature, an observation supported by 2D NMR experiments showing a loss of nuclear Overhauser effect cross peaks for residues at the ends of the helix (K. Clarke-Baldwin and V.L.P., unpublished work). The fits to the data show that the unassociated peptide has a mean residue ellipticity of ≈−7,000 deg/(mol cm residue) at 222 nm at infinite dilution, much different from that for purely random coil (≈0). A plot of ln(K) vs. 1/T (Fig. 2 Inset) indicates that the ΔG° of bundle formation (Eq. 1) at pH 8.5 is governed by a ΔH° of −18(2) kcal/mol and a ΔS° of −16(8) cal/mol K. The thermodynamic parameters showing an enthalpic driving force and unfavorable entropy are consistent with results in the literature measuring the association constant in coiled-coil formation (14) and indicative of hydrophobic collapse driving the folding and an unfavorable entropic term due to the decrease in species number from reactants to products.

Figure 2.

Figure 2

Open in a new tab

Temperature-dependent CD titrations monitored at 222 nm with fits to the data. Temperatures are 8 (filled circles), 13 (open circles), 18 (filled squares), 23 (open squares), 28 (filled triangles), 33 (open triangles), and 28°C (filled diamonds). Fitting assumes equilibrium 3 Baby L9C ⇌ (Baby L9C)3.

General Kinetic Features of Hg(II) Encapsulation by Baby L9C.

Mixing of equivalent volumes of a solution of 50 μM Baby L9C and a solution of 10 μM HgCl2 at pH 8.5 leads to a rapid increase in the UV absorbance spectrum at wavelengths lower than 300 nm (Fig. 3). These bands are characteristic of the charge-transfer band arising from trigonal thiolato Hg(II) formation (15). Analysis of the time-dependent absorbance spectra is consistent with Hg(SR)Inline graphic being the only species contributing to changes in the UV spectrum during the time of the experiment. The reaction is essentially (>99%) complete at 10 sec and is best fit to a biexponential function (Fig. 3 Inset), indicating that there are two distinct phases of the reaction. The biexponential behavior is more pronounced at higher concentrations of peptide, and the amplitude of the faster phase contributes ≈80% of the total change in the UV spectrum. Global analysis by singular value decomposition using SPECFIT (16, 17) allows evaluation of kinetic parameters for both phases.

Figure 3.

Figure 3

Open in a new tab

Stopped-flow data taken from mixing 10 μM HgCl2 and 790 μM Baby L9C in 50 mM potassium phosphate buffer (pH 8.5) containing 200 mM KCl. Difference spectra taken at 100-msec intervals using the spectrum at t = 0 sec as a background are shown. (Inset) Kinetic trace at 270 nm, fit, and residuals derived from mechanism E in Scheme S1.

The faster phase shows simple first-order kinetics with respect to HgCl2 concentration but has a more complicated order with respect to Baby L9C concentration as discussed below. These observations give an overall rate expression of (d[Hg(SR)Inline graphic]/dt)fast = k1 [HgCl2] [Baby L9C]n, where n depends on Baby L9C concentration. The second phase of the reaction is first order with respect to HgCl2 concentration and zero order with respect to Baby L9C concentration, giving an overall rate expression (d[Hg(SR)Inline graphic]/dt)slow = k−2 [HgCl2].

Dependence of the Reaction on Baby L9C Concentration.

The initial rate of Hg(SR)Inline graphic formation shows a nonlinear dependence on Baby L9C concentration (Fig. 4). At low concentration of Baby L9C, the reaction rate appears to have a first-order dependence; however, saturation-like behavior develops at higher concentrations of peptide. One explanation for the nonlinear behavior is that the rate-determining step in the formation of Hg(SR)Inline graphic involves Baby L9C that is unassociated with coiled coils and does not involve preformed unmetallated coiled coils. The rate constants are linearly dependent on the concentration of peptide unassociated with a coiled coil calculated by using the coiled-coil association constants derived from the titration data (Fig. 4 Inset). This result supports the model that in the formation of Hg(SR)Inline graphic the rate-determining step involves Baby L9C that is unassociated with coiled coils.

Figure 4.

Figure 4

Open in a new tab

Peptide concentration dependence of initial rate kini vs. [Baby L9C]total. (Inset) kini vs. [Baby L9C]unassociated. The rate constant derived from the initial rate data in Inset is 1.8(1) × 105 M−1⋅sec−1.

The Reaction (Hg + Baby L9C) vs. the Reaction [Hg(Baby L9C-H)2 + Baby L9C].

According to the minimal model proposed from the thermodynamic data, the rate-determining step could be one of three steps: (i) the formation of a two-stranded coiled coil with linear mercury, (ii) the addition of the third strand to the coiled coil, or (iii) the deprotonation of the third thiol to form the ultimate trigonal thiolato Hg(II) site. To probe the contribution of the first step in the kinetic mechanism to the overall rate, the reaction between HgCl2 and excess Baby L9C was compared with that of Hg(Baby L9C-H)2 and excess Baby L9C at pH 8.5.

Stopped-flow UV spectroscopy was performed by mixing a 20 μM Hg(Baby L9C-H)2 solution with solutions of 0–790 μM Baby L9C, all containing 50 mM phosphate and 100 mM KCl. For comparison, stopped-flow experiments were also performed with 20 μM HgCl2 and 0–790 μM Baby L9C. The kinetic profiles of both systems were identical. The temperature dependences of both reactions were also investigated by using initial rate analysis. This analysis yielded a similar activation energy (Ea) for the rate-determining step of both reactions; the Ea values are 17 ± 2 and 18 ± 2 kcal/mol for the reactions with the starting material Hg(Baby L9C-H)2 and HgCl2, respectively. These results are strong evidence that the rate-determining step is identical in both cases and that the formation of Hg(Baby L9C-H)2 is fast in comparison to the overall rate of Hg(Baby L9C-H)Inline graphic formation from HgCl2 and excess Baby L9C.

pH Dependence.

The pH dependence of the reaction was investigated to probe the role of the cysteine thiol deprotonation in the overall rate of mercury encapsulation. Different phosphate buffer solutions were made at pH values ranging from 7.0 to 9.0 for use in stopped-flow experiments (traces shown in Fig. 5). The analysis of these data is complicated by the pKa for the third cysteine thiol in a mercurated three-stranded coiled coil (pKa = 7.6). The reactions therefore result in incomplete formation of the trigonal thiolato Hg(II) complex at lower pH.

Figure 5.

Figure 5

Open in a new tab

Kinetic traces at 270 nm of the reaction between 10 μM HgCl2 and 100 mM Baby L9C at various pH values.

Although the pH does have an effect on the extent of formation of the Hg(SR)Inline graphic site, the half-life of the reaction remains relatively constant throughout the pH range, suggesting that the rate of three-stranded coiled-coil formation and the rate of the formation of the product state [the mixture between the Hg(SR)Inline graphic and Hg(SR)2(HSR) sites] does not seem to be affected significantly by pH. Consequently, the effect of pH on the reaction is not due to slow deprotonation of the third cysteine thiol affecting the rate-determining step and seems to be isolated to perturbation of a rapid equilibrium between the protonated and deprotonated forms of the third cysteine thiol changing the product state. The slight change in the rate (factor of 2 decrease from pH 7.0 to 9.0) may be due to lowering the transition-state energy due to a transient protonation of the peptide during the reaction.

The minimal mechanism proposed before this investigation seemed to be supported by the initial analysis. However, the rate-determining step involving collapse of the third helix onto a two-stranded coiled coil was more complicated than the original model suggested.

General Fitting.

All the kinetic data were analyzed by using the initial spectrum at 0 sec as the background. The variables for global analysis not only included rate constants but also the spectra of the absorbing species denoted with bold type in Scheme S1. The initial chemical mechanism used in the fitting was the simple first-order mechanism Hg(SR)2 → Hg(SR)Inline graphic (mechanism A). Subsequent mechanisms became more complex to describe the data satisfactorily. Notice that the process of finding an overall mechanism to fit the data with SPECFIT is a numerical process, and thus the mechanisms shown in Scheme S1 only include mechanistic steps that affect the rate of reaction. A more complete mechanism is proposed in Fig. 1.

Mechanism A in Scheme S1 did not adequately describe the shape of the data, which appeared to be biexponential. Therefore, mechanism B was proposed and satisfactorily described the shape of all data yielding a residual that resembled noise. Mechanism B, however, did not provide consistent rate constants when the Baby L9C concentration was varied. Mechanism C takes into account the dependence on Baby L9C concentration that adequately describes the nonlinear relationship between kobs and [Baby L9C]total. The intermediate in mechanism C was shown to have a spectrum proportional to the product complex [ɛintermediate state ≈ 0.8 × ɛHg(Baby L9C-H)3 for all wavelengths observed] (see Fig. 6). It was proposed, therefore, that the intermediate was a mixture of the product complex and a misfolded state that has the same spectrum as the initial state and not an intermediate with a different Hg(II)-trigonal thiolato site, which would be expected to have a spectrum with a different amplitude and shape. The misfolded state then can rearrange to yield the properly folded product state. Mechanism D accounts for the misfolded intermediate; however, it does not account for substoichiometric binding of Hg(II) to the three-stranded coiled coil formed by Baby L9C (11). Mechanism E includes dissociation rates that both accounts for substoichiometric binding of Hg(II) to the three-stranded coiled coil and provides a dissociative mechanism for refolding the misfolded state. Notice that this mechanism does not take into account the deprotonation of the third cysteine thiol (Fig. 1), which would be necessary to fit pH-dependent data. However, this mechanism is sufficient to fit experimental data taken at constant pH. To verify the results obtained from SPECFIT analysis, the rate equation for mechanism E was solved for the condition of constant peptide concentration and used to fit a kinetic profile (see Supporting Text). The rate constants obtained from fits to the rate expression agreed with the results obtained from SPECFIT.

Temperature Dependence.

The temperature dependence of the reaction between HgCl2 (10 μM) and excess Baby L9C was investigated to determine activation parameters for both phases of the reaction. The kinetic mechanism used for the fitting was mechanism E in Scheme S1, with the preequilibrium being defined by the association constants derived from the temperature-dependent investigation (above). This analysis was done at both low (50 μM) and high (830 μM) concentration of Baby L9C in the presence of 50 mM potassium phosphate buffer and 200 mM KCl from 5 to 35°C to verify that activation parameters for k1 at both extremes are equivalent. Activation parameters for k−2 can be adequately obtained only at high Baby L9C concentration where the first and second phases are sufficiently separated. Plots of ln(k) vs. 1/T indicate that the Ea for k1 is 21 ± 2 kcal/mol, which agrees within error with the value obtained from initial rate analysis (18 ± 2 kcal/mol), and that for k−2 is 23 ± 3 kcal/mol. The enthalpy and entropy of activation (Table 2) were obtained from plots of ln(k/T) vs. 1/T (Fig. 7).

Table 2.

Kinetic and thermodynamic parameters obtained from fits to the data

ΔH, kcal/mol ΔS, cal/mol K ΔH‡, kcal/mol ΔS‡, cal/mol K
Kassoc −18 (2) −8 (4) — —
k1 — — +20 (2) +40 (5)
k−2 — — +23 (2) +27 (5)
Open in a new tab

Discussion

We have developed a program to design small, untethered metallopeptide assemblies that provides variable coordination environments to metals. It is our hope that these systems will provide useful, well defined, rapidly modified, miniature metalloproteins capable of revealing general information about metalloprotein folding. The systems investigated thus far are coiled coils derived from the peptide TRI (8). TRI is acetylated on the N terminus and amidated on the C terminus to stabilize coiled-coil structure. TRI at pH > 7.0 is a well defined three-stranded coiled coil.

Substitution of a leucine for cysteine within the interior of the coiled coil results in a soft-metal binding site containing sulfur atoms. Although Hg(II) prefers to bind to thiolates with a linear geometry in aqueous solution at micromolar concentration, the peptides TRI L16C and TRI L9C provide a rigid trigonally symmetric framework before addition of metal capable of binding Hg(II) with trigonal coordination (7). We have derived a thermodynamic model for the encapsulation of Hg(II) by TRI L9C (11), which indicated that a rigid peptide backbone was not a prerequisite for trigonal thiolato coordination of Hg(II). This prediction was verified by the observation that Baby L9C, a 23-aa peptide that does not form a coiled coil before addition of Hg(II), stabilizes a Hg(SR)Inline graphic site. The Baby L9C system demonstrated that the information required for a trigonal Hg(II) site was embedded in the primary amino acid sequence, and the final configuration could be triggered by metal addition. A minimal kinetic mechanism for the Hg(II) coiled-coil construct was proposed based on the thermodynamic mechanism.

The initial reaction between HgCl2 and two equivalents of Baby L9C was proposed to be fast on the time scale of the overall reaction. This contention is supported by the observation that the formation of the Hg(II)(Baby L9C-H)Inline graphic proceeded with the same rate constant and activation enthalpy independent of mercury starting material [HgCl2 or Hg(Baby L9C-H)2]. In addition, the proposed rapid deprotonation of the third cysteine thiol was supported by the observation that the half-life of the reaction was relatively unaffected by pH (decreasing only by a factor of 2 from pH 7.0 to 9.0).

The rate of the reaction depended on the concentration of Baby L9C as predicted by the mechanism proposed from the thermodynamic data. The reaction profile, however, was not adequately fit by using a single exponential, indicating that this step was more complicated than simple first-order addition of the third helix. The data were described adequately by a biexponential function, indicating that the reaction occurred in two phases: a fast and a slow phase. Although the fast phase was first order with respect to Hg(Baby L9C-H)2 concentration, the reaction order with respect to total Baby L9C concentration was more complex, showing saturation-like behavior at high concentration of peptide (Fig. 4). Global analysis was used to fit the data, which were described adequately by the kinetic mechanism E in Scheme S1. The main differences between this mechanism and the minimal mechanism derived from the thermodynamic data are (i) the preequilibrium, involving the formation of an apo coiled coil, (ii) the formation of the three-stranded coiled coil with encapsulated Hg only occurs through reaction between Hg(Baby L9C)2 and free Baby L9C that is explicitly unassociated with a coiled coil, and (iii) the formation of a misfolded intermediate that is consistent with an antiparallel three-stranded coiled coil with encapsulated linear thiolato Hg(II) (see Fig. 1).

Mechanism E describes the refolding of the misfolded state as a dissociative process. Another option is intracomplex reorganization where the misfolded peptide simply rearranges without dissociating completely from the bundle. Both of these processes can be zero order in [peptide]total, an observed property of the second phase of the reaction. The bundle dissociation enthalpy [(Baby L9C)3 ⇌ 3 Baby L9C] is +18(2) kcal/mol, and the activation enthalpy for the refolding of the misfolded intermediate is +23(2) kcal/mol, showing that complete dissociation of the third strand is energetically feasible. A dissociative mechanism is also supported by a large entropic contribution to the transition-state energy for the refolding step (ΔS‡ = +27(5) cal/mol K). In addition, a process where dissociation is taken as the mechanism of refolding adequately describes the kinetic data, accounting for substoichiometric binding of Baby L9C to Hg(Baby L9C-H)2. Furthermore, the reorganization rate is comparable to the rate reported for the dissociation of the third strand of a three-stranded coiled coil with similar stability (18).

The assignment of the misfolded state as an antiparallel coiled coil is consistent with the data presented here as well as literature data for coiled-coil folding. Several literature studies investigating the folding of nonmetallated coiled coils indicate a single, monoexponential transition from the unfolded to folded states (19–21). These investigations monitor fluorescence or UV CD spectra characteristic of α-helical structure, techniques that cannot clearly distinguish between parallel and antiparallel structures. Therefore, an intermediate that contained a mixture of antiparallel and parallel coiled coils would look similar to an intermediate containing all parallel coiled coils. The reaction would appear complete after attainment of the intermediate mixture. However, because the formation of trigonal thiolato Hg(II) in the three-stranded coiled coil formed by Baby L9C depends on the parallel conformation, a difference in the UV spectra is observed for a mixture of parallel and antiparallel coiled coils as compared with a solution of all parallel coiled coils, allowing for observation of the intermediate state.

Implications for Metalloprotein Folding.

The area of protein design has provided insight into the formation of metallocenters by using protein redesign to generate novel metal centers in existing protein scaffolds and by using de novo design to make novel scaffolds to investigate the role of the protein structure on metal binding. Initial studies focused on using the strength of the metal–ligand bonds to impart structure on the polypeptide backbone (22, 23). More recently, polypeptide scaffolds have been designed that bind metals in thermodynamically favored environments (24–27). Examples where the metals are bound in ligation environments that are not thermodynamically favored in the absence of the polypeptide backbone are uncommon. However, recent designs of de novo proteins containing Hg(SR)Inline graphic centers (7, 28), dimeric metallocenters (29, 30), iron-sulfur clusters (31), Ni-X-Fe4S4 centers (32, 33), and hemes with open coordination sites (34) demonstrate that the design of complex metal centers where the ultimate structure of the metalloprotein is determined by an interplay of polypeptide backbone and metal coordination stabilities is feasible. A fundamental understanding of the formation of these metal centers is essential to facilitate the successful design of functional metalloproteins.

A common conception among the design community has been that a highly stable peptide scaffold is optimal for the formation of this type of metal center. Our results here and in our previous report (11) indicate that this is not necessarily the case. The present report suggests that highly stable apoprotein structures may not be kinetically optimal for two reasons. First, in the rate-determining step the formation of the three-stranded coiled coil with the encapsulated Hg(II) site occurs by reaction between Hg(Baby L9C-H)2 and free unassociated peptide. Preformed three-stranded coiled coil is not observed to compete kinetically with its unassociated counterpart. This result predicts that apoproteins with flexible metal binding sites may be more apt to bind metal quickly than apoproteins with very stable, rigid binding sites. Second, the presence of a misfolded intermediate on the pathway to Hg(Baby L9C-H)3 warns that stability in the apo structure may also serve to stabilize local minima during the folding process. Less-stable peptide backbones will allow facile conformational interchange to reach the global minimum more quickly.

Studies of natural protein systems provide examples to support both of these statements. First, the insertion of Cu(II) into the unmetallated form of the blue copper protein azurin has been studied thoroughly (3–6). Azurin has a very stable backbone structure. In fact, crystal structures of apo- and holoazurin show that the backbone structures are nearly identical. The addition of Cu(II) to the folded apoprotein has a very slow rate, which is decreased further when the insertion is performed in complexing media. The insertion rate can be increased 20,000-fold when the apo form is denatured before addition of the copper, highlighting the importance of having an accessible metal-binding site for rapid incorporation of the cofactor.

Second, the folding of cytochrome c underscores the importance of having conformation flexibility for rapid refolding of misfolded states (1, 2). Cytochrome c has a final structure containing a heme axially ligated by a methionine and a histidine. During folding, an intermediate is formed that has bis-histidine ligation. The polypeptide backbone then rearranges to displace the histidine and replace it with the methionine. This process is aided by the conformational flexibility of the polypeptide backbone before coordination of the heme by methionine.

Conclusions.

This article describes a detailed kinetic study of metal binding by a de novo-designed peptide. Specifically, the kinetics of Hg(II) encapsulation by the three-stranded coiled coil formed by Baby L9C are detailed. The mechanism of encapsulation involves the nucleation of the polypeptide backbone fold by mercury to form Hg(Baby L9C-H)2, a rate-determining addition of the third peptide by reaction between Hg(Baby L9C)2 and free unassociated Baby L9C to form a three-stranded coiled coil with an encapsulated linear thiolato Hg(II) intermediate state and fast deprotonation of the third cysteine thiol. The intermediate state contains a significant amount (≈20%) of misfolded intermediates that can rearrange to form the properly folded intermediate, which continues on to the final Hg(Baby L9C-H)Inline graphic structure. This mechanism indicates that less-structured peptides can bind metals more quickly and allow for fast reorganization from misfolded intermediates in comparison to their more-structured counterparts.

Supplementary Material

Supporting Information
pnas_0336055100_index.html (1.3KB, html)

Acknowledgments

B.T.F. thanks the National Institutes of Health for supporting this research through a National Research Service Award postdoctoral fellowship.

Abbreviations

TRI

CH3CO-G(LKALEEK)4G-NH2

TRI L9C

CH3CO-GLKALEEK CKALEEK- (LKALEEK)2G-NH2

Baby L9C

CH3CO-GLKALEEKCKALEEKLKALEEKG-NH2

Baby L9C-H

Baby L9C with a deprotonated cysteine thiol

CD

circular dichroism

Footnotes

This paper was submitted directly (Track II) to the PNAS office.

References

  • 1.Tezcan F A, Findley W M, Crane B R, Ross S A, Lyubovitsky J G, Gray H B, Winkler J R. Proc Natl Acad Sci USA. 2002;99:8626–8630. doi: 10.1073/pnas.132254499. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Yeh S-R, Han S, Rousseau D L. Acc Chem Res. 1998;31:727–736. [Google Scholar]
  • 3.Tennent D L, McMillin D R. J Am Chem Soc. 1979;101:2307–2311. [Google Scholar]
  • 4.Blaszak J A, McMillin D R, Thornton A T, Tennent D L. J Biol Chem. 1983;258:9886–9892. [PubMed] [Google Scholar]
  • 5.Pozdnyakova I, Wittung-Stafshede P. Biochemistry. 2001;40:13728–13733. doi: 10.1021/bi011591o. [DOI] [PubMed] [Google Scholar]
  • 6.Wittung-Stafshede P. Acc Chem Res. 2002;35:201–208. doi: 10.1021/ar010106e. [DOI] [PubMed] [Google Scholar]
  • 7.Dieckmann G R, McRorie D K, Tierney D L, Utschig L M, Singer C P, O'Halloran T V, Penner-Hahn J E, DeGrado W F, Pecoraro V L. J Am Chem Soc. 1997;119:6195–6196. [Google Scholar]
  • 8.Dieckmann G R, McRorie D K, Lear J D, Sharp K A, DeGrado W F, Pecoraro V L. J Mol Biol. 1998;280:897–912. doi: 10.1006/jmbi.1998.1891. [DOI] [PubMed] [Google Scholar]
  • 9.Matzapetakis M, Farrer B T, Weng T-C, Hemmingsen L, Penner-Hahn J E, Pecoraro V L. J Am Chem Soc. 2002;122:8042–8054. doi: 10.1021/ja017520u. [DOI] [PubMed] [Google Scholar]
  • 10.Farrer B T, McClure C P, Penner-Hahn J E, Pecoraro V L. Inorg Chem. 2000;39:5422–5423. doi: 10.1021/ic0010149. [DOI] [PubMed] [Google Scholar]
  • 11.Farrer B T, Harris N P, Balchus K E, Pecoraro V L. Biochemistry. 2001;40:14696–14705. doi: 10.1021/bi015649a. [DOI] [PubMed] [Google Scholar]
  • 12.Chan W C, White P D. In: Fmoc Solid Phase Peptide Synthesis: A Practical Approach. Hames B D, editor. New York: Oxford Univ. Press; 2000. p. 346. [Google Scholar]
  • 13.Ellman G M. Arch Biochem Biophys. 1959;82:70. doi: 10.1016/0003-9861(59)90090-6. [DOI] [PubMed] [Google Scholar]
  • 14.Bosshard H R, Durr E, Hitz T, Jelesarov I. Biochemistry. 2001;40:3544–3552. doi: 10.1021/bi002161l. [DOI] [PubMed] [Google Scholar]
  • 15.Wright J G, Natan M J, MacDonnell F M, Ralston D M, O'Halloran T V. Prog Inorg Chem Bioinorg Chem. 1990;38:323–412. [Google Scholar]
  • 16.Binstead R A, Chronister C W, Ni J, Hartshorn C M, Meyer T J. J Am Chem Soc. 2000;122:8464–8473. [Google Scholar]
  • 17.Stultz L K, Binstead R A, Reynolds M S, Meyer T J. J Am Chem Soc. 1995;117:2520–2532. [Google Scholar]
  • 18.Durr E, Bosshard H R. Protein Sci. 2000;9:1410–1415. doi: 10.1110/ps.9.7.1410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Durr E, Jelesarov I, Bosshard H R. Biochemistry. 1999;38:870–880. doi: 10.1021/bi981891e. [DOI] [PubMed] [Google Scholar]
  • 20.Wendt H, Berger C, Baici A, Thomas R M, Bosshard H R. Biochemistry. 1995;34:4097–4107. doi: 10.1021/bi00012a028. [DOI] [PubMed] [Google Scholar]
  • 21.Sosnick T R, Jackson S, Wilk R R, Englander S W, DeGrado W F. Proteins Struct Funct Genet. 1996;24:427–432. doi: 10.1002/(SICI)1097-0134(199604)24:4<427::AID-PROT2>3.0.CO;2-B. [DOI] [PubMed] [Google Scholar]
  • 22.Ghadiri M R, Choi C. J Am Chem Soc. 1990;112:1630–1632. [Google Scholar]
  • 23.Ghadiri M R, Soares C, Choi C. J Am Chem Soc. 1992;114:825–831. [Google Scholar]
  • 24.Choma C T, Lear J D, Nelson M J, Dutton P L, Robertson D E, DeGrado W F. J Am Chem Soc. 1994;116:856–865. [Google Scholar]
  • 25.Robertson D E, Farid R S, Moser C C, Urbauer J L, Mulholland S E, Pidikiti R, Lear J D, Wand A J, DeGrado W F, Dutton P L. Nature. 1994;368:425–432. doi: 10.1038/368425a0. [DOI] [PubMed] [Google Scholar]
  • 26.Sharp R E, Diers J R, Bocian D F, Dutton P L. J Am Chem Soc. 1998;120:7103–7104. [Google Scholar]
  • 27.Suzuki K, Hiroaki H, Kohda D, Nakamura H, Tanaka T. J Am Chem Soc. 1998;120:13008–13015. [Google Scholar]
  • 28.Li X, Suzuki K, Kanaori K, Tajima K, Kashiwada A, Hiroaki H, Kohda D, Tanaka T. Protein Sci. 2000;9:1327–1333. doi: 10.1110/ps.9.7.1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Summa C M, Rosenblatt M M, Hong J-K, Lear J D, DeGrado W F. J Mol Biol. 2002;321:923–938. doi: 10.1016/s0022-2836(02)00589-2. [DOI] [PubMed] [Google Scholar]
  • 30.Lombardi A, Summa C M, Geremia S, Randaccio L, Pavone V, DeGrado W F. Proc Natl Acad Sci USA. 2000;97:6298–6305. doi: 10.1073/pnas.97.12.6298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Gibney B R, Rabanal F, Skalicky J J, Wand A J, Dutton P L. J Am Chem Soc. 1999;121:4952–4960. [Google Scholar]
  • 32.Laplaza C E, Holm R H. J Biol Inorg Chem. 2002;7:451–460. doi: 10.1007/s00775-001-0320-4. [DOI] [PubMed] [Google Scholar]
  • 33.Laplaza C E, Holm R H. J Am Chem Soc. 2001;123:10255–10264. doi: 10.1021/ja010851m. [DOI] [PubMed] [Google Scholar]
  • 34.Moffet D A, Case M A, House J C, Vogel K, Williams R D, Spiro T G, McLendon G L, Hecht M H. J Am Chem Soc. 2001;123:2109–2115. doi: 10.1021/ja0036007. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information
pnas_0336055100_index.html (1.3KB, html)
pnas_0336055100_1.pdf (91.8KB, pdf)
pnas_0336055100_2.html (950B, html)
pnas_0336055100_6055Fig_6.jpg (49.6KB, jpg)
pnas_0336055100_3.html (1,014B, html)
pnas_0336055100_6055Fig_7.jpg (47.7KB, jpg)
pnas_0336055100_4.html (869B, html)
pnas_0336055100_6055Fig_8.jpg (47.7KB, jpg)

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences

RESOURCES