How strong are side chain interactions in the folding intermediate?

Abstract

Influence of 12 nonpolar amino acids residues from the hydrophobic core of apomyoglobin on stability of its native state and folding intermediate was studied. Six of the selected residues are from the A, G and H helices; these are conserved in structure of the globin family, although nonfunctional, that is, not involved in heme binding. The rest are nonconserved hydrophobic residues that belong to the B, C, D, and E helices. Each residue was substituted by alanine, and equilibrium pH-induced transitions in apomyoglobin and its mutants were studied by circular dichroism and fluorescent spectroscopy. The obtained results allowed estimating changes in their free energy during formation of the intermediate state. It was first shown that the strength of side chain interactions in the apomyoglobin intermediate state amounts to 15–50% of that in its native state for conserved residues, and practically to 0% for nonconserved residues. These results allow a better understanding of interactions occurring in the intermediate state and shed light on involvement of certain residues in protein folding at different stages.

Introduction

Average-sized proteins (approximately 100–300 residues) usually fold in vitro via an intermediate state, as it was first reported by Tanford.¹ The later studies in this field are summarized in a number of reviews.^2–6 The formation of intermediate states in the course of globular protein renaturation often takes less than 10⁻³ s (see, e.g., reviews^3,7). These intermediate states are quite compact and possess a pronounced secondary structure, but their side chains lack tight packing.^8–11 The refolding of a protein may depend on the arrangement of its intermediate state. Therefore, studies of the intermediate state structure and its stability are important to shed light on relationship between the protein amino acid sequence and its spatial structure.

Apomyoglobin (apoMb) ranks classical in studies of this kind. It is a single-domain protein of 17 kDa formed by seven helices (by eight ones for holomyoglobin). Its structure differs from that of holomyoglobin by the position of its loops and helices, with one of the latter, F-helix, being completely destroyed,^12–14 and the total helicity being reduced from 78% to 65%.¹⁵ Nevertheless, apoMb holds its hydrophobic core¹⁶ and a peculiar holomyoglobin-like tertiary structure.¹⁷ According to the thermodynamic criteria, when heated, both holomyoglobin and its apo-form undergo cooperative denaturing transitions accompanied by notable changes in their enthalpy and heat capacity. Of the two, the apo-form, displays a lower stability.¹⁸ At pH 4.2, it acquires the “molten globule” intermediate state characterized by a pronounced helical structure content, compactness, and the hydrophobic core which can be unfolded by a strong denaturant.¹⁹ Further acidification to pH 2.0 also provides almost complete unfolding of the intermediate.²⁰

A number of articles report on the structures of both kinetic and equilibrium apoMb intermediates with an emphasis on the presence of intrinsic native contacts and on their contribution to stability of the intermediate state. For example, NMR studies demonstrated that the equilibrium and early (<10⁻³ s, formation time) kinetic intermediates have a virtually complete secondary structure, and the side chains of the A, G and H helices, and in part, of the B and E helices too, are somewhat screened from the solvent.^14,21,22 Some authors^23–27 used point mutagenesis to estimate contributions of various amino acids to intermediate stability. In particular, it has been shown that, with A130, F123, and S108 (responsible for A-H and G-H interactions) substituted by either hydrophobic or charged amino acids, all apoMb mutants demonstrated a change in stability of the native state relative to the intermediate state and a surprisingly small (though measured with a large error) alteration of the intermediate state stability relative to the unfolded state.²³ The aforementioned articles and Bertagna and Barrick²⁸ offer evidence both for and against involvement of the apoMb side chains in stabilization of the intermediate state.

To resolve this uncertainty, here we measure contribution of side chains of the 12 residues into interactions occurring within apoMb folding intermediate. For each mutant, we determined thermodynamic parameters of the conformational transitions U↔I and I↔N and the resulting Fersht's parameter²⁹ φ_I that reflects the comparative contribution of the substituted side chain to stabilization of the intermediate state. Analysis of all these experimental data allowed concluding that the strength of side chain interactions in the apomyoglobin intermediate state does not exceed 50% of that in its native state for conserved residues, and is practically equal to 0% for nonconserved residues.

Results and Discussion

Selection of amino acid residues for point substitutions

To estimate the side chain contribution into stabilization of the apoMb intermediate state, the residues intended for point substitutions were selected on the basis of two theoretical approaches proposed previously for prediction of the protein folding nucleus. One of these was developed by Ptitsyn's team.^30,31 An analysis of 728 globin sequences revealed only six residues that are conserved in globin structure but nonfunctional (i.e., not involved in heme binding): Val10 and Trp14 (A helix), Ile111 and Leu115 (G helix), Met131 and Leu135 (H helix).³⁰ These are the first six residues selected for substitution. The other six, from the B helix (I28, F33), C helix (L40), D helix (M55) and E helix (L61, L76), are nonconserved but deeply buried in the protein structure residues that have been predicted by the “landscape” approach in modeling of protein unfolding^32,33 as residues having significant influence upon the rate of apoMb folding. For short, the former will be referred to as “conserved residues” and the latter as “nonconserved residues.” All these residues are hydrophobic, and their side chains have numerous contacts with their surroundings in the native structure (Fig. 1).

Figure 1.

(A) Number of interatomic contacts for each residue side chain in the native myoglobin structure (contact radius is 6 Å). Filled circles denote the conserved nonfunctional residues, nonfilled circles – the nonconserved residues. α-Helices of the myoglobin structure³⁴ are shown under the plot. (B) Predicted probability φ_nucl of implication in the apomyoglobin folding nucleus. Adapted from Ref. 33.

Each selected residue was substituted by Ala. These mutations met two common requirements: (i) to keep the type of amino acids unchanged (i.e., a hydrophobic amino acid was replaced by another hydrophobic one); (ii) to diminish the number of contacts of their side chains as compared with the wild type (WT) native structure.

To learn the effect of these mutations on stability of the native and intermediate states (a thorough analysis of which for apoMb was made earlier by Baldwin and coworkers),^23,35 we studied pH-induced equilibrium transitions of both the mutant and WT forms using far UV CD and tryptophan (Trp) fluorescence.

Strength of side chain interactions in the intermediate and native states: estimation and comparison

Stability of the native and intermediate states of apoMb with point mutations was estimated on the basis of studying pH-induced equilibrium transitions by far UV CD and Trp fluorescence (Figs. 2 and 3). The equilibrium urea-induced denaturing transitions provide a poor evidence for existence of an intermediate state in this process,³⁶ thus being inadequate for measuring a relative stability of the native and intermediate states of apoMb and its mutants.

Figure 2.

Equilibrium pH-induced denaturing transitions in WT apoMb and in its mutants as recorded by changing molar ellipticity at 222 nm, [θ]₂₂₂. (A, B), proteins with substituted conserved residues within the A, G, and H helices; (C, D), proteins with substituted nonconserved residues within the B, C, D, and E helices. Insets show the mutant proteins presented. Experimental errors do not exceed the size of corresponding experimental points.

Figure 3.

Equilibrium pH-induced denaturing transitions in WT apoMb and in its mutants as recorded by changing intensity of Trp fluorescence at 335 nm, I₃₃₅. (A, B), proteins with substituted conserved residues within the A, G, and H helices; (C, D), proteins with substituted nonconserved residues within the B, C, D, and E helices. Insets show the mutant proteins presented. Experimental errors do not exceed the size of corresponding experimental points.

In WT apoMb, two pH-induced equilibrium transitions were clearly revealed by changing molar ellipticity at 222 nm ([θ]₂₂₂) and by Trp fluorescence intensity at 335nm (I₃₃₅).³⁷ These were the transition from the native to intermediate state with its middle at pH 4.8 and the transition from the intermediate to unfolded state with its middle at pH 3.3. It is around pH 4.2 where accumulation of the equilibrium “molten globule” state was observed. As seen from Figures 2 and 3, for mutants, the accumulated intermediate was observed over a wider pH range than for WT apoMb. However, the U-I mid-transition of both the mutant and WT forms occurred within virtually the same pH range (pH = 3.3–3.5), unlike the I-N transition that, for mutants, was shifted towards higher pH values (from pH = 4.8 to pH = 5.2–5.8). Hence, these mutations caused a strong destabilization of the apoMb native structure and only a slight destabilization of its folding intermediate structure.

Since the mutants constructed for the current study had the unchanged number of ionized amino acids, that is, since pH dependence of the N and I (relative to U) free energies was preserved, it is possible²⁵ to calculate the change in mutant free energy^1,38 (counted off the U state one), as compared to that of the WT form, resting on the three-state model N↔I↔U and the pH-dependent equilibrium transitions in the mutants. The thermodynamic parameters were calculated from the protein unfolding curves using equations based on the three-state model N↔I↔U described earlier.^1,38 The experimentally observed fluorescence intensity or molar ellipticity at a current pH value is presented as . Here A_i is the pH-independent fluorescence intensity or molar ellipticity of state i (i = N, I, U); is the fraction of state i (R being the gas constant and T the temperature); ΔG_iU(pH) = ΔG_iU(pH₀)+β_iU(pH − pH₀) is free energy of this state relatively to that of the state U [thus, ΔG_UU(pH) ≡ 0, ΔG_IU(pH) = ΔG_IU(pH₀) + β_IU(pH − pH₀), and [cf. Eq. (4)] ΔG_NU(pH) = (ΔG_NI(pH₀) + ΔG_IU(pH₀)) + (β_NI + β_IU) (pH − pH₀)]; pH₀ = 4.2 where the most pronounced “molten globule” state accumulation was observed, while β_NI and β_IU are pH-independent multipliers for the N-I and I-U transitions (since it has been shown by Baldwin and coworkers^35,39 that the pH dependence of ΔG_NI is close to linear in the transition region, we also assume a linear dependence of ΔG_IU on pH in the corresponding region; this assumption is justified by calculations made in⁴⁰). The values of , , and so forth, are fitted to the experimental graphs (Figs. 2 and 3) using the least-square method (see “Methods” section).

Table I gives values of ΔΔG_IU, ΔΔG_NU, and φ_I at pH₀ = 4.2. They are calculated from CD and FL data using Eqs. (1)– (4) (see later). As seen, all point mutations resulted in a considerable destabilization of the native state, whereas the mutant intermediates showed a markedly less destabilization, in particular those with substitutions in the B, C, D, and E helices.

Table I.

Changes in Stability of N and I States of apoMb Mutants Upon Equilibrium Unfolding, and φ_I Values Estimated from Results by Different Techniques

	Circular dichroism			Fluorescence
	ΔΔGIU (kcal/mol)	ΔΔGNU (kcal/mol)	φI	ΔΔGIU (kcal/mol)	ΔΔGNU (kcal/mol)	φI
Conserved
V10A	0.5 ± 0.5	0.9 ± 0.6	0.6 ± 0.7	0.6 ± 0.3	1.3 ± 0.7	0.5 ± 0.3	0.5 ± 0.3
W14A	1.0 ±0.4	3.0 ± 1.3	0.3 ± 0.2	0.6 ± 0.3	6.0 ± 2.2	0.1± 0.1	0.16 ± 0.06
I111A	1.2 ± 0.4	3.0 ± 0.6	0.4 ± 0.1	0.6 ± 0.4	1.9 ± 0.8	0.3 ± 0.2	0.37 ± 0.12
L115A	0.8 ± 0.4	2.9 ± 0.8	0.3 ± 0.2	0.7 ± 0.2	2.7 ± 0.7	0.3 ± 0.1	0.26 ± 0.08
M131A	1.0 ± 0.4	2.6 ± 0.8	0.4 ± 0.2	0.3 ± 0.2	1.0 ± 0.3	0.2 ± 0.2	0.32 ± 0.15
L135A	0.9 ± 0.4	2.1 ± 0.7	0.4 ± 0.2	0.2 ± 0.2	1.7 ± 0.8	0.1 ± 0.1	0.18 ± 0.12
Nonconserved
I28A	0.1 ± 0.7	0.7 ± 0.8	0.1 ± 1.0	0.0 ± 0.2	1.5 ± 0.4	−0.0 ± 0.1	−0.0 ± 0.1
F33A	0.5 ± 0.4	2.1 ± 0.6	0.2 ± 0.2	−0.1 ± 0.2	1.9 ± 0.7	−0.1 ± 0.1	0.0 ± 0.1
L40A	−0.4 ± 0.5	0.2 ± 0.5	−1.8 ± 4.8	−0.4 ± 0.2	1.2 ± 0.6	−0.4 ± 0.3	−0.4 ± 0.3
M55A	−0.3 ± 0.5	0.1 ± 0.6	−2.0 ± 8.9	−0.1 ± 0.2	2.2 ± 1.1	−0.1 ± 0.1	−0.1 ± 0.1
L61A	0.4 ± 0.5	1.2 ± 0.6	0.3 ± 0.4	−0.4 ± 0.3	1.4 ± 0.4	−0.3 ± 0.2	−0.2 ± 0.2
L76A	0.7 ± 0.4	5.0 ± 1.6	0.1 ± 0.1	0.2 ± 0.2	10.1 ± 4.7	0.02 ± 0.03	0.03 ± 0.03

All values correspond to pH₀ = 4.2 and 11°C in a buffer system with 0.01 M sodium acetate.

However, as seen in many cases, especially for the I-U transition, an error in ΔΔG determination by either technique exceeded ΔΔG itself. This leads to a large error in φ_I value calculated by each single method. To reduce this error and to obtain more accurate φ_I values, we averaged φ_I obtained by the two techniques used in such a way, that φ_I values obtained with smaller errors have larger weights in the averaging (see “Materials and Methods” section).

These averaged values and their errors for each of the 12 mutants are shown in Figure 4. Inset on Figure 4 presents, as an example, a decrease of error in the φ_I value for the I111A mutant as a result of joint usage of the two techniques.

Figure 4.

Estimated values (with errors) for the studied side chains, averaged over the two techniques used, for the conserved (filled triangles) and nonconserved (nonfilled triangles) residues. The Inset shows φ_I values for I111A substitution separately calculated from CD and FL equilibrium data with their errors for each technique.

As mentioned above, the parameter φ_I shows^29,41 the relative strength (as compared with the native state) of residue side chain contacts in the intermediate state: φ_I = 1 means that stabilization of the intermediate state by the given side chain is comparable with its stabilizing effect on the native state, while φ_I = 0 means that no stabilization of the intermediate state is produced by this side chain. As seen from Figure 4, all side chains had values below 0.5, with some of these (specifically, for nonconserved residues) being close to zero, which means that mutations of the studied residues, nonconserved in particular, only slightly (as compared with the native state) affected the strength of contacts formed by their side chains in the intermediate state. Then, the fact that substitution of large hydrophobic groups by small ones (Ala) does not result in a considerable destabilization of the intermediate state may be explained by relatively weak interactions of these side chains and in particular, by their weak hydrophobic interactions in the I state. The latter conclusion is justified by investigation²⁸ of substitutions of Met131 of apoMb for various nonpolar residues of different hydrophobicity.

However, it is noteworthy that values were reliably higher ( approximately 0.15–0.5) for mutations in the A, G, and H helices, as compared with those in the B, C, D, and E helices ( approximately 0.0 and even negative). This occurs not only because the B – E helices are less structured in the intermediate state (as shown by H-D exchange data¹³), but may also depend on the side chain of the structured region. Indeed, Val17 is a part of the A helix structured in the I state, but is nonconserved in the globin family, and behavior of its side chain (as shown by³⁷) is different from that of conserved residues and closer to nonconserved ones. This demonstrates heterogeneity in strength of side chain contacts in the intermediate state. It is noteworthy that the A, G, and H helices are not only structured but also show a native-like topology²² in the I state, which correlates with a higher contact density between these helices.

Conclusions

The results of the current study are summarized in Figures 4 and 5. In Figure 5, the color intensity is indicative of the extent to which the studied side chains are involved in stabilizing the apoMb intermediate state. It is evident that side chains of the conserved residues are involved in stabilizing of the intermediate state to a larger extent than those of the nonconserved conserved residues.

Figure 5.

The conserved nonfunctional residues in the A (V10, W14), G (I111, L115), and H (M131, L135) helices and the nonconserved residues in the B, C, D, and E helices are shown on the background of the sperm whale myoglobin backbone fold;³⁴ residue numbers are indicated. In the studied residues, Cα atoms are shown as circles; the color intensity (see the scale under the figure) increases with increasing parameter for the side chain. Labels of the conserved residues of the A, G, and H helices are underlined.

Figures 4 and 5 show that the interaction strength of the side chains in the intermediate state, although sufficient for maintaining of the intermediate state structure (as follows from NMR results by¹³), is lower than that in the native state even for the conserved residues: it amounts to 15–50% (of the native state strength) for the conserved residues forming the A, G, H-helical complex and is equal practically to 0% for the nonconserved residues from the B, C, D, and E helices. However, it is shown (unpublished results) that the situation in the folding nucleus for the rate-limiting I↔N transition is quite the opposite to what we found in the folding intermediate: contribution of the conserved residues from the A, G, H helices to stabilization of the folding nucleus in this transition is minor, while contribution of the residues from the B, D, E helices is significant. This means that the intermediate state structure acts as a template for formation of other helices and then the native state.

Qualitatively, our results supplement theoretical^31,42 and experimental^13,14 studies on protein denaturation as to estimation of the strength of side chain interactions in the molten globule. Besides, they supplement the supposition by Ptitsyn⁴³ that conserved residues are of particular importance for the intermediate of apoMb folding.

Materials and Methods

Expression and isolation

Plasmids with mutant apomyoglobin genes were constructed by PCR using pET17a plasmid (a kind gift of P.E. Wright), containing the sperm whale apomyoglobin gene, and appropriate primers, with the use of a QuikChange (“Stratagene”) kit.⁴⁴ WT apoMb and its mutants were isolated and purified as described previously³⁶ after expression of appropriate plasmids in E. coli BL21 (DE3) cells.

The protein purity was checked by SDS/PAGE electrophoresis. Protein concentration was determined spectrophotometrically from absorbance at 280 nm using the extinction coefficient for all samples,^45,46 except for one Trp-lacking mutant with substitution W14A; for this one, the used extinction coefficient was determined from nitrogen-based measurements according to⁴⁶ and appeared .^37,47

Spectroscopy techniques

Equilibrium measurements of pH-induced denaturing transitions

Equilibrium protein fluorescence (FL) measurements were taken on a Shimadzu RF-5301 PC spectrofluorimeter using standard 1-cm path length quartz cuvettes. The excitation wavelength was 293 nm. The emission spectra were recorded between 300 and 500 nm [with a special attention to the emission intensity at 335 nm (I₃₃₅)] at a protein concentration of 0.03 mg/mL.

A JASCO J-600 spectropolarimeter (Japan) was used to perform circular dichroism (CD) experiments. Far UV CD spectra were measured between 200 and 250 nm using 0.1 mm path length cells. The protein concentration was 1.0 mg/mL. Molar ellipticity [θ] was calculated as previously:³⁶ where θ is ellipticity measured (deg, 10⁻³); MRW is the average molecular weight of a residue calculated from the sequence; l is the path length (mm); c is protein concentration (mg/mL).

All experiments were carried out at 11°C in a buffer system with 0.01 M sodium acetate (NaAc) under conditions described previously.³⁶

For every mutant protein, two independent experimental unfolding curves were obtained. Each such titration curve was approximated, as a whole, by a continuous bisigmoid curve using the “SigmaPlot” program (SPSS, www.sigmaplot.com) that allows finding , and auxiliary β_NI, β_IU values and their standard deviations. Then the averaged , values and their standard deviations , were estimated for every mutant from the corresponding values found for the pair of titration curves obtained for this mutant.

Calculation of thermodynamic parameters

The thermodynamic parameters^1,38 were calculated from the protein unfolding curves using equations based on the three-state model of N↔I↔U transition.

Fersht's parameter φ_I for the equilibrium intermediate state (I) was calculated for pH₀ = 4.2. It was calculated as in,²⁹ that is, as the ratio between mutation-induced change in stability of the protein intermediate state relative to its unfolded state (ΔΔG_IU) and the change occurring in protein stability during the N↔I↔U transition (ΔΔG_NU):

(1)

where

(2)

(3)

with the values of of the total transition N↔I↔U calculated from the equation

(4)

since for each separate transition (N↔I and I↔U) its thermodynamic parameters were calculated individually.

Estimation of errors

To calculate the values of φ_I with all due care, we used a combination of equilibrium data on apoMb unfolding obtained as (i) equilibrium CD measurements, and (ii) equilibrium fluorescence (FL) measurements. Each gave a value of φ_I with a substantial error. An accurate calculation of the final (average) value of requires not only values calculated by Eq. (1) from results by the above techniques (CD, i = 1; FL, i = 2) but also an optimal choice of weights p_i from estimates of errors [see Eq. (5) below]. To have the most accurate estimate of , the contribution of each technique used to the averaging has to be taken, according to,^48,49 as

(5)

where is the error of obtained by the i-technique, n is the number of techniques (here, n = 2).

Since obtained by each technique (i = 1, 2) was calculated by Eq. (1), its error is

(6)

The value is calculated from Eq. (2) for the technique i, and its error is

(7)

while is calculated, for the same technique, from Eqs. (3) and (4), and thus its error is

(8)

For each technique, and were calculated with allowance made for (a) the instrument error, (b) errors of approximation of the experimental points by the continuous bisigmoid curve.

The final, averaged from the two techniques used, parameter was obtained as the weighted sum of values (with i = 1, 2; see above)

(9)

where the weight p_i was calculated from Eq. (5); the resultant error of is given (see Refs. ⁴⁸ and⁴⁹) by

(10)