Dissecting key residues in folding and stability of the bacterial immunity protein 7

Abstract

The small four-helix immunity protein, Im7, has previously been shown to fold via a compact intermediate containing three of the four native helices. The short, six-residue helix III only docks onto the developing Im7 structure after the rate-limiting second transition state has been traversed. Previous work demonstrated that mutation of the helix III sequence can be used to trap the protein in the on-pathway intermediate ensemble at equilibrium. Here the role played by individual residues in the native helix III sequence in locking Im7 into a stable native structure is further examined. This work commenced with an Im7 sequence trapped in the partially folded state by substitution of the six residues in helix III with a polyglycine sequence. Biophysical analysis of variants in which individual residues from the native helix III sequence, and combinations of these residues, were introduced into this background demonstrated a critical requirement for three residues, Leu 53, Ile 54 and Tyr 55, to lock Im7 into its unique native structure. The results demonstrate a stringent constraint on the evolution of the Im7 helix III sequence rationalizing its high-sequence identity in the fold family. Thus, Leu 53 and Ile 54 provide crucial stabilizing interactions in the hydrophobic core of native Im7, while Tyr 55 is required for both stability and function. In contrast, Tyr 56 is critical for colicin binding and has no role in maintaining a stable native fold.

Introduction

The small 87-residue immunity protein 7 (Im7) (Fig. 1A) folds to its native structure via a three-state mechanism involving an on-pathway intermediate (Capaldi et al., 2001), in which the rate-limiting transition state (TS) occurs late on the reaction coordinate. This is indicated by the beta-Tanford value measured for the TS (β_ts), which provides a measure of the relative compactness of the TS relative to that of the native state (β_ts = 0.89) (Fig. 1B). ϕ-Value analysis of Im7 folding (Capaldi et al., 2002; Friel et al., 2009) indicates that helix III only binds onto the developing structure after the rate-limiting TS (TS2) has been traversed, with residues Leu 53 and Ile 54 in helix III (Fig. 1A) preferentially stabilizing the native structure over the intermediate ensemble. We have previously shown that folding via an intermediate is an integral part of Im7's folding mechanism, as increasing the helix propensity of helix III by its extension with an eight-residue polyalanine sequence was not able to alter the folding mechanism (Knowling et al., 2009).

Fig. 1.

(A) Structure of wild-type Im7 highlighting the side chains of residues in helix III (PDB ID 1AYI (Dennis et al., 1998)). Aromatic residues making non-native contacts with Tyr 55 during folding (Phe 41 and Trp 75) are also shown. Residues mutated in this study are highlighted Leu 52 (green), Ile 53 (blue), Tyr 55 (gray) and Tyr 56 (pink). The diagram was drawn using Chimera (Pettersen et al., 2004). (B) Cartoon representation of the folding mechanism of Im7. The on-pathway intermediate forms on the submillisecond time-scale. How helices I, II and IV dock in the intermediate and TS ensembles are shown schematically based on restrained molecular dynamic simulations of these species using ϕ-values or hydrogen exchange protection factors as restraints (Gsponer et al., 2006; Friel et al., 2009). (C) Summary of the variants designed to determine the role of individual side chains in helix III in the folding and stability of Im7. The colors shown here are used to represent the different variants throughout.

Previous work (Capaldi et al., 2002) has demonstrated that it is possible to destabilize the native structure of Im7 relative to the intermediate ensemble such that the on-pathway intermediate becomes the most populated species at equilibrium (∼95%) (Spence et al., 2004). This can be achieved in a number of ways (Spence et al., 2004; Bartlett and Radford, 2010) which include mutating hydrophobic residues in helix III to alanine (L53A and/or I54A), or by replacement of the entire helix III sequence with a six-residue glycine linker (Spence et al., 2004). Several of these ‘trapped' variants have fluorescence signals similar to that of the hyperfluorescent folding intermediate (I) of wild-type Im7 (Spence et al., 2004). Interestingly, the variant H3G6, in which the six native residues of helix III (TDLIYY) are substituted with glycine residues (GGGGGG) (Fig. 1C), is partially folded but does not exhibit the hyperfluorescence characteristic of the kinetic folding intermediate of Im7, whereas a variant containing three Gly residues succeeded by Tyr 55 and Tyr 56 (GGGYY) exhibits a fluorescence signal comparable with that of the kinetic intermediate (Whittaker et al., 2007). These data suggest that one, or both, of the two tyrosine residues found in helix III play an important role in orienting the conformational properties of the single Trp 75 in the kinetic intermediate, such that it displays its characteristic hyperfluorescence signal.

Molecular dynamics simulations, restrained by experimental constraints (hydrogen exchange protection factors or ϕ-values), of the ensembles populated during Im7 folding suggest that non-native interactions not only stabilize the intermediate ensemble, but also continue to be of major importance in folding beyond the intermediate ensemble (Gsponer et al., 2006; Friel et al., 2009). A subtle rearrangement of the hydrophobic core takes place in the folding step from I to TS2, resulting in a native-like orientation of helices I, II and IV and a native-like radius of gyration for core hydrophobic residues (Friel et al., 2009). These simulations highlighted Tyr 55, a partially solvent-exposed residue in native Im7, as one residue that forms significant and substantial non-native contacts with other aromatic residues throughout the folding process (particularly with Phe 41 and Trp 75). Im7 has also been shown to have an unusually frustrated native structure, in part due to the requirement to fold efficiently and also to bind to its cognate colicin-binding partner (Kuhlmann et al., 2000; Sutto et al., 2007). Redesign of the Im7 native structure to reduce frustration focused on frustrated interactions involving residues in the helix III sequence. Using this approach the least frustrated variant identified was the double variant Y55R, Y56N (Sutto et al., 2007). Subsequent coarse-grained simulations predicated that this double variant will exhibit two-state folding behavior with no populated intermediate species. Similar coarse-grained simulations of wild-type Im7 folding included in this study also pointed to the importance of non-native contacts between residues in the region surrounding helix III in stabilizing the folding intermediate (Sutto et al., 2007).

Here we describe a series of experiments that aim to further define the role of the helix III sequence in enabling Im7 to fold to a unique, stable native structure. Using the previously designed variant (H3G6) (Spence et al., 2004) (Fig. 1C and Table I) as a starting point, residues from the sequence of helix III in wild-type Im7 (TDLIYY) were reintroduced either individually or in combination (Fig. 1A and C), with the aim of identifying the minimal sequence in this region required to enable Im7 to fold to its stable native structure. Additional single aromatic substitution variants were also created in the background of the wild-type Im7 sequence to further examine the role of Tyr 55 and Tyr 56 in stabilization of the native structure. The variants were analyzed for changes in structural properties when compared with either native wild-type Im7 or partially folded H3G6. The results reveal a critical requirement for the presence of three highly conserved residues in helix III, Leu 53, Ile 54 and Tyr 55 to stabilize the native Im7 structure. This finding rationalizes the conservation of these residues in the Im7 sequence and demonstrates how folding, stability and functional requirements have led to the frustration characteristic of the Im7 folding landscape.

Table I.

Spectroscopic and thermodynamic parameters determined for Im7 and its helix III variants (pH 7.0, 10°C and 0.4 M Na₂SO₄)

	Im7	H3G6a	GGLGGG	GGGIGG	GGLIGG	GGLIYG	Y55A	Y56A
Expected mass (Da)	10 846	10 421	10 476	10 476	10 532	10 638	10 754	10 754
Actual mass (Da)b	10 846	10 420	10 474	10 476	10 532	10 637	10 754	10 754
ΔG°UF (kJ mol−1)c	−24.9 ± 1.3	−10.2 ± 1.3	−11.7 ± 2.4	−11.4 ± 1.5	−13.5 ± 2.1	−17.9 ± 2.9	−17.2 ± 1.3	−23.6 ± 1.1
MUF (kJ mol−1 M−1)c	5.05 ± 0.3	3.1 ± 0.34	3.5 ± 0.6	3.5 ± 0.4	4.1 ± 0.5	4.7 ± 0.8	4.8 ± 0.4	4.9 ± 0.3
εU (M−1 cm−1)	9530	–	6970	6970	6970	8250	8250	8250
εN (M−1 cm−1)	9370	6250	6970	6960	6970	8650	8140	8150
Helical content (%)d	50	45	35	35	43	43	40	49
λmax (ANS)e	473 ± 1	477 ± 0.7	482 ± 1	484 ± 1	478 ± 2	473 ± 2	475 ± 2	474 ± 2
Intensity (ANS)e,f	1	6.4 ± 0.2	2.47 ± 0.2	1.99 ± 0.2	1.8 ± 0.1	1.2 ± 0.1	2.50 ± 0.2	1 ± 0.1
λmax (Trp)e	311 ± 1	340 ± 0.6	340 ± 1	338 ± 1	336 ± 1	334 ± 1	334 ± 1	312 ± 1
Intensity (Trp)e,g	0.26 ± 0.1	0.96 ± 0.01	1.0 ± 0.2	0.97 ± 0.2	0.72 ± 0.1	0.44 ± 0.2	0.90 ± 0.2	0.20 ± 0.1

^aData for H3G6 have been taken from Spence et al. (2004) and are included for comparison.

^bDetermined by electrospray ionization mass spectrometry.

^cErrors quoted are the fit errors.

^dDetermined from the mean residue molar ellipticity relative to the known helical content of Im7 from its X-ray structure (Dennis et al., 1998).

^eError quoted is the replicate error from three experiments.

^fNormalized relative to the native state of Im7 in the presence of ANS.

^gNormalized relative to the unfolded state of Im7 in the presence of 8 M urea.

Materials and methods

Expression, purification and mutagenesis

All variants were created as described by Spence et al. (2004). In the initial round of mutagenesis, the wild-type sequence of helix III, TDLIYY, was mutated to GGLIYY. Tyrosine 55 and Tyrosine 56 were then changed to glycine in a second round of mutagenesis. All mutants of H3G6 were created in one round of polymerase chain reaction apart from the mutant GGLIGG, which was created in two rounds. Production and purification of all Im7 variants were carried out as described by Whittaker et al. (2007).

Equilibrium denaturation

Equilibrium denaturation was performed using circular dichroism (CD), observing the signal at 222 nm. A stock solution containing 2 mg ml⁻¹ of protein in 50 mM sodium phosphate, 0.4 M Na₂SO₄ and 1 mM Ethylenediaminetetraacetic acid (EDTA), pH 7.0 was prepared. This solution was then used to make further substock solutions of 0, 2, 4, 6 and 8 M urea by mixing with 50 mM sodium phosphate, 0.4 M Na₂SO₄, 1 mM EDTA and 8 M urea, pH 7.0 (final protein concentration 0.2 mg ml⁻¹). These substock solutions were used to make 1 ml samples ranging in urea concentration from 0 to 8 M in 0.2 M increments. These solutions were thoroughly mixed and then incubated at 10°C overnight prior to measurement. CD measurements were performed using a 1 mm cuvette (Hellma, UK) and the signal at 222 nm was averaged over 1 min. The average signal as a function of denaturant concentration was determined and fitted to a two-state equilibrium model (Equation (1)) using non-linear least-squares regression function in SigmaPlot (Systat Software, UK):

(1)

where is the difference in free energy between the denatured and natively folded states in the absence of urea, M_UN is the dependence of the free energy on the concentration of urea, a is the denaturant dependence of the signal of the native state, b is the signal of the native state in water, c is the denaturant dependence of the signal of the unfolded state, d is the signal of the unfolded state in water, R is the molar gas constant, T is the temperature in Kelvin and [D] is the concentration of urea. The data for all variants were then converted into the fraction folded (f_F) according to Equation (2):

(2)

where a, b, c, d and [D] are as specified above for Equation (1). The data were then normalized to the 222 nm value obtained for wild-type Im7 in 50 mM sodium phosphate, 0.4 M Na₂SO₄, pH 7.0 to reflect the differences in native helical content for each of the variants.

Fluorescence emission spectra

The fluorescence of samples containing ∼20 μM of protein in 50 mM sodium phosphate, 0.4 M Na₂SO₄ and 1 mM EDTA, pH 7.0 was measured using a QM-1 spectrofluorimeter (Photon Technology International, UK). Prior to steady-state fluorescence analysis, the samples were incubated at 10°C overnight. Fluorescence was measured in a 1 cm pathlength quartz cuvette (Hellma). Fluorescence was excited at 280 nm with a 3 nm bandwidth. Fluorescence emission spectra were recorded from 270 to 500 nm (3 nm emission bandwidth), with a 1 nm step resolution and 1 s integration time. Spectra containing only buffer were subtracted from the observed signal and the spectra of the folded state were normalized to the emission spectra of the unfolded state at an identical concentration in 50 mM sodium phosphate, 0.4 M Na₂SO₄, 1 mM EDTA and 8 M urea, pH 7.0.

1-Anilinonapthalene-8-sulphonic acid binding

The fluorescence emission spectra of 1-anilinonapthalene-8-sulphonic acid (ANS) were measured in the presence or absence of Im7 and its variants using a QM-1 spectrofluorimeter (Photon Technology International). Protein samples were dissolved into 50 mM sodium phosphate, 0.4 M Na₂SO₄ and 1 mM EDTA, pH 7.0 containing 250 μM ANS to give a final protein concentration of 1 μM. The samples were then incubated overnight at 10°C prior to measurement. ANS fluorescence was measured by excitation at 389 nm and the emission spectra recorded between 380 and 600 nm, with a 1 nm step resolution and 1 s integration time. Six spectra were acquired and the average spectra determined. All spectra were normalized to the fluorescence emission of ANS in the presence of 1 μM native Im7 at its emission maximum.

Results

Reintroduction of Leu 53 or Ile 54 into H3G6

As a starting point for this study, the residues Leu 53 or Ile 54 were introduced individually into the sequence of H3G6 to create the variants GGLGGG and GGGIGG, respectively (Fig. 1C). Studies of these variants using equilibrium denaturation and far-ultraviolet (UV) CD (Figs 2A, B and 3A) revealed that these proteins have a helical content less than that of the native protein as judged by far-UV CD and show similar stability (ΔG°_UF) and compactness (M_UF) to the kinetic folding intermediate of Im7 (ΔG°_UF −11.7 ± 2.4 kJ mol⁻¹ and M_UF 3.5 ± 0.6 kJ mol⁻¹ M⁻¹ for GGLGGG and ΔG°_UF−11.4 ± 1.5 kJ mol⁻¹ and M_UF 3.5 ± 0.4 kJ mol⁻¹ M⁻¹ for GGGIGG, respectively) (Table I). For comparison, wild-type Im7 has a ΔG°_UN of −24.0 ± 1.5 kJ mol⁻¹ and an M_UF of 4.6 ± 0.3 kJ mol⁻¹ M⁻¹ (Spence et al., 2004) and the variant L53AI54A, which has been shown to populate the intermediate ensemble at equilibrium, has a ΔG°_UF of−11.1 ± 1.3 kJ mol⁻¹ and an M_UF of 3.5 ± 0.35 kJ mol⁻¹ M⁻¹ (Spence et al., 2004). The variants described in this study all denature with apparent two-state equilibrium behavior. However, it is possible that these variants may populate a mixture of the native, intermediate and unfolded ensembles at equilibrium. The free energies and M-values obtained from fitting the data to a two-state model thus represent the apparent values for the equilibrium between the unfolded ensemble and all folded species as a function of denaturant concentration and are defined as ΔG°_UF and M_UF. Comparison of the ΔG°_UF and M_UF obtained from fitting the equilibrium unfolding data for these variants to a two-state model with the values expected for native wild-type Im7 and the intermediate species provides a fingerprint of whether these variants predominantly populate the intermediate or native species at equilibrium or a mixture of the two ensembles.

Fig. 2.

Equilibrium denaturation of the Im7 variants (A) GGLGGG (B) GGGIGG (C) GGLIGG and (D) GGLIYG measured at pH 7.0, 10°C in 0.4 M Na₂SO₄. Filled circles represent data obtained and the solid line is the best fit of the data to a two-state transition. To aid comparison the data for wild-type Im7 (black) and H3G6 Im7 (pale blue) are shown in all panels.

Fig. 3.

(A) Far-UV CD spectra of Im7 (black) and variants GGLGGG (green), GGGIGG (blue), GGLIGG (red), GGLIYG (yellow) and H3G6 (pale blue). (B) Near-UV CD spectra of Im7 and helix III variants. (C) Fluorescence emission spectra of the folded states of Im7 and the helix III variants. For comparison, the spectrum of the denatured state of Im7 in 8 M urea is also shown as a dashed black line. (D) ANS fluorescence emission spectra of 250 μM ANS (dashed black line) in the presence of 1 μM helix III variants. The data are normalized to the intensity of ANS plus native Im7 at 476 nm. Spectra for Y55A and Y56A are shown by gray and pink lines, respectively.

Despite possessing only a single Tyr (Tyr 10), the near-UV spectra of GGLGGG and GGGIGG show much larger ellipticities than that of native wild-type Im7, indicating that these variants have failed to adopt a native-like structure (Fig. 3B). The single tryptophan residue (Trp 75) of native wild-type Im7 gives rise to a small negative signal in the near-UV CD at ∼295 nm. This signal is absent in the spectra of GGLGGG and GGGIGG, suggesting that this residue is not well ordered in these variants (Fig. 3B). The altered positioning of Trp 75 is reflected in the fluorescence emission spectra of GGLGGG and GGGIGG: both variants give rise to fluorescence emission intensities similar to that of Im7 denatured in 8 M urea, although the emission maxima are blue shifted by 10–12 nm relative to urea-denatured Im7, consistent with burial of Trp 75 in these partially folded species (Fig. 3C). These findings echo those previously found for H3G6 (Spence et al., 2004) (Table I). Binding of ANS to the GGLGGG and GGGIGG variants results in a blue shift in the λ_max of ANS fluorescence compared with binding to native wild-type Im7, although the intensity is lower than expected for binding to a hydrophobic region in a molten globule state (Semisotnov et al., 1991) (Fig. 3D and Table I). The data indicate, therefore, that GGLGGG and GGGIGG bury a substantial part of their hydrophobic core, despite being partially folded. This observation is consistent with previous data for H3G6 (Spence et al., 2004) and the kinetic folding intermediate (Capaldi et al., 2002).

Taken as a whole, these experiments show that substitution of the individual side chains Leu 53 and Ile 54 into the background of H3G6 is insufficient to stabilize the native state of Im7 such that it is significantly populated at equilibrium.

Tandem reintroduction of Leu 53 and Ile 54

To determine whether the simultaneous reintroduction of Leu 53 and Ile 54 into the sequence of H3G6 is sufficient to permit the native structure of Im7 to be stably populated at equilibrium, the variant GGLIGG was created (Fig. 1C). The ΔG°_UF and M_UF for GGLIGG (ΔG°_UF−13.5 ± 2.1 kJ mol⁻¹ and M_UF 4.1 ± 0.5 kJ mol⁻¹ M⁻¹) are within error of the values obtained for GGLGGG and GGGIGG, indicating that the simultaneous presence of these two side chains in the H3G6 background is not sufficient to allow Im7 to fold to a stable native state (Table I). Although equilibrium denaturation (Fig. 2C) indicates little difference in the thermodynamic stability of GGLIGG, GGLGGG and GGGIGG, the far-UV CD spectrum of GGLIGG suggests that this variant has a higher helical content than the variants containing single substitutions (Fig. 3A). The near-UV CD spectrum of GGLIGG is similar to that observed for native wild-type Im7 except for the lack of negative ellipticity arising from Trp 75 ∼295 nm (Fig. 3B). The ANS binding data indicate that the hydrophobic surface area exposed in GGLIGG is similar to that in GGLGGG and GGGIGG, and smaller than that of H3G6 (Fig. 3D). In contrast, both the λ_max and the intensity of the Trp fluorescence emission spectrum for GGLIGG are reduced compared with the trapped intermediate variants discussed thus far (Table I and Fig. 3C). These data are consistent with GGLIGG remaining trapped in an intermediate ensemble but adopting a distinctly different ensemble of conformations to the variants GGLGGG and GGGIGG. Nonetheless, this variant still remains predominantly non-native in structure in the region of helix III. The inclusion of either Tyr 55 or Tyr 56 within the sequence of helix III, therefore, must be critical for the stabilization of the Im7 native state.

Introduction of Tyr 55 into a GGLIGG background

To investigate the role of Tyr 55 in enabling Im7 to fold to a unique native structure the variant GGLIYG was created (Fig. 1C). As shown in Fig. 2D the inclusion of Tyr 55 in this variant stabilizes the protein by ∼6 kJ mol⁻¹ (ΔG°_UF = −17.9 ± 2.9 kJ mol⁻¹) relative to the kinetic intermediate of Im7 (Table I). The M_UF value for GGLIYG (4.7 ± 0.8 kJ mol⁻¹ M⁻¹) is also within error of that of native Im7 (M_UN= 5.05 ± 0.3 kJ mol⁻¹ M⁻¹) (Table I). The far-UV CD spectrum of GGLIYG indicates that this variant has reduced helical content compared with wild-type Im7 (Fig. 3A). The CD signal at 222 nm indicates that both GGLIGG and GGLIYG show a helical content decreased by approximately seven residues compared with wild-type Im7, a figure remarkably similar to the number of residues found in native helix III (six residues) (Table I). However, analysis of the conformation of GGLIYG by near-UV CD revealed a spectrum similar to that of native, wild-type Im7, including negative ellipticity for Trp 75 at ∼295 nm (Fig. 3B). This finding supports the view that Tyr 55 plays a key role in formation of and stabilization of the Im7 native fold.

The GGLIYG variant is less fluorescent than GGLGGG, GGGIGG and GGLIGG, consistent with this variant adopting a more native-like packing within its hydrophobic core (Fig. 3C). However, GGLIYG is also more fluorescent than wild-type Im7, consistent with this variant containing a structure somewhat reorganized compared with native wild-type Im7. ANS fluorescence data (Fig. 3D) show an exposed hydrophobic surface area very similar to that of native wild-type Im7. Together, these data suggest that GGLIYG either populates both the intermediate and the native states at equilibrium or, alternatively, adopts a unique structure that is partially folded, but distinct from that of GGLIGG.

The role of tyrosines 55 and 56 in Im7 stability

The data presented above suggest that Tyr 55 is critical for Im7 to adopt a stable native structure. To examine the role of this residue in stabilizing Im7 in more detail Tyr 55 was substituted with alanine, creating the variant Y55A (Fig. 1C). As a control the solvent-exposed Tyr 56 was also individually substituted with alanine, creating Y56A.

Urea equilibrium denaturation of Y55A shows that the native state of this variant is destabilized by ∼7 kJ mol⁻¹ compared with wild-type Im7 (ΔG°_UF = −17.2 ± 1.3 kJ mol⁻¹ and M_UF = 4.8 ± 0.4 kJ mol⁻¹ M⁻¹, for Y55A; Fig. 4A and Table I). In contrast, equilibrium denaturation of Y56A reveals that this variant has a free energy similar to that of wild-type Im7 (ΔG°_UN = −24.9 kJ mol⁻¹ and M_UN = 4.9 ± 0.3 kJ mol⁻¹ M⁻¹; Table I). The near-UV CD spectrum of Y55A (Fig. 4B) shows features that are typical of a protein with a fixed packing of aromatic residues around the protein's core (Kelly et al., 2005), but reveals that this variant has failed to adopt a native-like structure. Consistent with the equilibrium denaturation data, the near-UV CD spectrum of Y56A is similar to that of native wild-type Im7 (Fig. 4B). Far-UV CD spectra recorded for these Tyr substitution variants indicate that Y56A has a helical content identical to that of wild-type Im7, while Y55A shows a reduction in helical content (Fig. 4C, Table I).

Fig. 4.

(A) Equilibrium denaturation of Y55A (gray) and Y56A (pink). Filled circles represent data obtained and the solid lines are the best fit of the data to a two-state transition. To aid comparison the data for wild-type Im7 (black) and H3G6 Im7 (pale blue) are shown. (B) Near-UV CD spectra of Y55A (gray), Y56A (pink) and Im7 (black). (C) Far-UV CD spectra of Y55A (gray), Y56A (pink) and Im7 (black). (D) Fluorescence emission spectra of the folded states of Im7 (black), Y55A (gray) and Y56A (pink). For comparison, the spectrum of the denatured state of Im7 in 8 M urea is also shown as a dashed black line.

The fluorescence emission spectra obtained for the Y55A and Y56A variants demonstrate that Y56A has a fluorescence emission intensity and λ_max identical to that of wild-type Im7, consistent with this variant having a native-like fold (Fig. 4D). In contrast, Y55A shows dramatic changes in its fluorescence properties. The Trp fluorescence emission spectrum of Y55A shows an emission intensity similar to that of the denatured state of Im7, although, akin to the partially folded Im7 species described above, its λ_max is blue shifted to 334 nm (Fig. 4D and Table I). Together, the data reinforce the key role of Tyr 55 in stabilizing the native structure of Im7, its substitution with Ala resulting in a structure or an ensemble of structures that is substantially different from the native wild-type protein.

Discussion

The data presented above indicate that the presence of a triumvirate of residues within the helix III sequence, Leu 53, Ile 54 and Tyr 55, is critical to allow Im7 to fold to a stable native-like structure. While mutation of both Leu 53 and Ile 54 to alanine destabilizes Im7 to the extent that the intermediate ensemble is populated at equilibrium (Spence et al., 2004), biophysical characterization of the GGLIGG variant demonstrates that the presence of these two aliphatic residues alone is not sufficient to enable folding to a stable, native Im7 structure. It is not clear from these data whether the GGLGGG, GGGIGG and GGLIGG variants could cross the rate-limiting TS barrier but, if this is the case, they provide insufficient energy to stabilize the native state once the barrier has been crossed.

The results presented for variants GGLIYG and Y55A demonstrate the essential role played by Tyr 55 in stabilizing the native structure of Im7. Previous experimentally restrained molecular dynamics simulations indicated that Tyr 55 makes non-native contacts with other aromatic groups, particularly Trp 75 and Phe 41 (Fig. 1A), throughout Im7 folding (Friel et al., 2009). The importance of non-native interactions involving aromatic residues in shaping the folding energy landscape of Im7 and other proteins has been demonstrated in a number of other studies (Radford et al., 1992; Rothwarf and Scheraga, 1996; Rea et al., 2008; Friel et al., 2009). Tryptophan residues in hen egg white lysozyme have been shown to form non-native interactions that slow folding and must be rearranged prior to adoption of the native structure (Radford et al., 1992; Rothwarf and Scheraga, 1996). Insertion of Phe or Trp residues into the N-terminal β-hairpin of ubiqutin results in the formation of a compact intermediate that is stabilized by non-native interactions with the C-terminus of the protein (Rea et al., 2008). The data presented here for Im7 demonstrate that Tyr 55 not only plays a role in directing folding of the polypeptide to the native structure (Friel et al., 2009), but is also a key determinant of native-state stability.

The Im7 native structure has been shown to contain an unusually high proportion of highly frustrated interactions, a cluster of which encompass residues found in helix III (Sutto et al., 2007). Sutto et al. (2007) highlighted the Tyr residues within helix III as candidates for mutation to reduce frustration in the Im7 native structure and predicted that the Im7 structure becomes less frustrated upon binding to its cognate binding partner, E7. Tyr 55 and Tyr 56, along with Asp 52, have previously been shown to provide a large component of the binding free energy of the Im7:E7 complex that is vital for protection from colicin intoxication in vivo (Wallis et al., 1998; Kuhlmann et al., 2000). The data presented here provide further experimental evidence supporting the importance of Tyr 55 in the Im7 folding landscape, not only for folding and function, but also in enabling a stable native structure to be adopted by this 87 residue protein.

The data presented here demonstrate that three residues, Leu 53, Ile 54 and Tyr 55, play a key role both in allowing helix III to dock onto the developing Im7 structure and in locking the protein into a stable native conformation. The residues Leu 53 and Ile 54, which are buried in the hydrophobic core of native Im7, are essential for the stability of the native structure and make contact with a large number of hydrophobic residues spread throughout the Im7 core. Tyr 55 (which is 100% conserved among the nuclease-specific immunity proteins (Dennis et al., 1998)) plays a key role in minimizing the hydrophobic surface area exposed in natively folded Im7 and, as demonstrated by the Y55A variant, its substitution results in failure to adopt a native structure. Overall, therefore, the triad of residues, Leu 53, Ile 54 and Tyr 55, holds the key to successful helix III docking during Im7 folding and subsequently locks Im7 into a stable native structure. Consideration of the folding kinetics, function and stability of Im7 together rationalize why the sequence of helix III is so highly conserved (85%) across all four nuclease-specific immunity proteins (Dennis et al., 1998; Friel et al., 2009) given the essential role played by every residue in this sequence in one or more of the key facets involved in folding to a functional, native Im7 structure.

Funding

The work was supported by the EPSRC (S02/B196), BBSRC (BB/F01614X/1), and the Wellcome Trust (045223 and 075099).