Logo of prosciprotein sciencecshl presssubscriptionsetoc alertsthe protein societyjournal home
Protein Sci. 2016 May; 25(5): 999–1009.
Published online 2016 Mar 16. doi: 10.1002/pro.2908
PMCID: PMC4838642
PMID: 26916981

Salt bridge as a gatekeeper against partial unfolding

Mark W. Hinzman, 1 , 2 Morgan E. Essex, 1 , 3 and Chiwook Parkcorresponding author 1
Author information Article notes Copyright and License information Disclaimer

Associated Data

Supplementary Materials

Abstract

Salt bridges are frequently observed in protein structures. Because the energetic contribution of salt bridges is strongly dependent on the environmental context, salt bridges are believed to contribute to the structural specificity rather than the stability. To test the role of salt bridges in enhancing structural specificity, we investigated the contribution of a salt bridge to the energetics of native‐state partial unfolding in a cysteine‐free version of Escherichia coli ribonuclease H (RNase H*). Thermolysin cleaves a protruding loop of RNase H* through transient partial unfolding under native conditions. Lys86 and Asp108 in RNase H* form a partially buried salt bridge that tethers the protruding loop. Investigation of the global stability of K86Q/D108N RNase H* showed that the salt bridge does not significantly contribute to the global stability. However, K86Q/D108N RNase H* is greatly more susceptible to proteolysis by thermolysin than wild‐type RNase H* is. The free energy for partial unfolding determined by native‐state proteolysis indicates that the salt bridge significantly increases the energy for partial unfolding by destabilizing the partially unfolded form. Double mutant cycles with single and double mutations of the salt bridge suggest that the partially unfolded form is destabilized due to a significant decrease in the interaction energy between Lys86 and Asp108 upon partial unfolding. This study demonstrates that, even in the case that a salt bridge does not contribute to the global stability, the salt bridge may function as a gatekeeper against partial unfolding that disturbs the optimal geometry of the salt bridge.

Keywords: protein stability, protein folding, salt bridge, partial unfolding, proteolysis

Abbreviations

CD
circular dichroism
RNase H*
cysteine‐free E. coli ribonuclease HI
SDS‐PAGE
sodium dodecyl sulfate polyacrylamide gel electrophoresis.

Introduction

Pairs of opposing charges in proximity (salt bridges) are frequently observed in protein structures.1, 2, 3 The higher occurrence of salt bridges in proteins from thermophilic organisms than their analogous proteins from mesophilic organisms has suggested that salt bridges contribute to protein stability in a favorable manner.1, 4, 5, 6 Experimental assessments of the energetic contribution of salt bridges have shown salt bridges may stabilize protein structures significantly.7, 8, 9 However, salt bridges observed in natural proteins are not always stabilizing. The contribution of surface salt bridges to protein stability seems minimal in many cases.10, 11 Moreover, it has been shown that a proper substitution of a buried salt bridge with hydrophobic residues may stabilize protein more effectively than the salt bridge itself.12 Computational surveys of salt bridges in natural proteins have also discovered a quite broad spectrum of stabilizing to destabilizing salt bridges.3, 13 The energetic contribution of a salt bridge depends on multiple factors, such as the degree of the desolvation penalty, the geometry of the bridge, and the interaction of the charges in the salt bridge with the rest of the protein.3, 13, 14, 15, 16 Apparently, whether a salt bridge is stabilizing or destabilizing depends on how much the favorable energetic factors compensate the unfavorable energetic factors. Stabilization of protein structure by an engineered salt bridge also requires careful considerations of these multiple factors by which the newly added salt bridge affects the stability of the protein.15

From the frequent observations of salt bridges that contribute to protein stability only marginally or even unfavorably, it has been suggested that salt bridges may contribute to the specificity in protein structure rather than the stability itself.13 Because the energetic contribution of a salt bridge is quite sensitive to the geometry and the structural context, a salt bridge may selectively stabilize only the conformations in which the salt bridge can achieve the proper geometry and context. This argument is also supported by the observation that salt bridges are somewhat rare between domains with flexible hinges.3 While flexibility requires the presence of multiple conformations with similar conformational energies, a salt bridge between the domains may increase the selectivity for a subset of conformations and reduce flexibility in the domain motion.

This proposed role of salt bridges in enhancing structural selectivity suggests that salt bridges may have a role in suppressing transient partial unfolding in proteins. Even under native conditions, folded proteins exist in dynamic equilibrium with multiple non‐native conformations, which include partially or even globally unfolded proteins. Selective stabilization of the native form of protein will maximize the population of the protein in the active conformation. Using a cysteine‐free version of Escherichia coli ribonuclease H (RNase H*)17 as a model system, we have tested the role of a salt bridge in suppression of transient partial unfolding. RNase H* has been investigated extensively as a model system for protein folding and native‐state partial unfolding.17, 18, 19, 20, 21 The protein unfolds in a two‐state manner without any equilibrium intermediate.17 However, hydrogen/deuterium exchange revealed that partially unfolded forms exist under native conditions.18 We have also probed transient partial unfolding of RNase H* by native‐state proteolysis.22 Because binding to the active site of a protease requires an extended conformation of the cleavage site,23 proteolysis of compactly folded proteins under native conditions occurs through transient unfolding.24, 25, 26 The kinetic analysis of native‐state proteolysis allows us to determine the free energy difference between the native form and the most accessible form out of proteolytically susceptible conformations (the cleavable form).22, 27, 28 The initial cleavage of RNase H* by thermolysin, a nonspecific bacterial protease, occurs at the peptide bond between Thr92 and Ala93 in a protruding loop [Fig. ​[Fig.11(A)].22 The analysis of the proteolysis kinetics revealed that the free energy difference between the native form and the cleavable form (ΔG C‐N°) is about 6 kcal/mol. The negligible effect of urea on ΔG C‐N°22 and also a negligible ΔC p° determined from the temperature effect on ΔG C‐N°29 showed that the solvent accessible surface area of the cleavable form is not so different from that of the native form, indicating that the scale of the unfolding is small and localized. Considering that 8–12 residues in an extended conformation are the minimum requirement for proteolysis,26 it is likely that the protruding loop with the cleavage site is unfolded and exists in an extended conformation in the cleavable form. Interestingly, a salt bridge between Lys86 and Asp108 connects two helices that are linked by the protruding loop with the cleavage site [Fig. ​[Fig.1(A)].1(A)]. The proximity and the location of the salt bridge suggest that unfolding of the protruding loop for proteolysis may result in disruption of the salt bridge, which may increase the energetic cost of partial unfolding. To assess the energetic contribution of the salt bridge to partial unfolding, we investigated the energetics of partial unfolding of K86Q, D108N, and K86Q/D108N variants of RNase H* by native‐state proteolysis as well as the global stability of each variant.

An external file that holds a picture, illustration, etc. Object name is PRO-25-0999-g001.jpg

Lys86–Asp108 salt bridge in RNase H*. (A) The location of the Lys86–Asp108 salt bridge of RNase H* (PDB CODE: 1F21). Both residues of the salt bridge are shown as spheres with Lys86 in blue and Asp108 in red. The image on the left is rotated by 90° to show the image on the right. The amide bond between Thr92 and Ala93, which is the cleavage site by thermolysin, is indicated by an arrow. (B) Hydrogen bonds between Lys86 and Asp108 in wild‐type RNase H* and Gln86 and Asn108 in K86Q/D108N RNase H*. Hydrogen bonds are indicated by yellow dashed line. The distances between the donor and acceptor atoms are shown in Å. The structure of K86Q/D108N RNase H* was modeled with PyMOL (Schrödinger).

Results

Choice of the reference state

The choice of a reference state is critical in evaluating the energetic contribution of a salt bridge to a protein structure.15 In this study, we chose K86Q/D108N RNase H* as the reference state. In wild‐type RNase H*, the distances between the nitrogen atom of the amine group of Lys86 and the two oxygen atoms of the carboxyl group of Asp108 are 2.9 Å and 3.1 Å [Fig. ​[Fig.1(B)],1(B)], indicating that the salt bridge between Lys86 and Asp108 forms hydrogen bonds as well. When Lys86 and Asp108 are mutated to Gln and Asn, respectively, the residues lose their charges but still maintain a large portion of the side chain and may also form a weak hydrogen bond. A modeling of the double mutant showed that the distance between the oxygen atom of the amide group of Gln86 and the nitrogen atom of the amide group of Asn108 would be about 3.3 Å [Fig. ​[Fig.1(B)].1(B)]. The use of K86Q/D108N RNase H* as a reference state, therefore, focuses our investigation of the mutational effect mostly to the loss of charges and the decrease in the strength of the hydrogen bonds. For consistency, we also use K86Q/D108N RNase H* as a reference state in describing the energetic contribution of the salt bridge; Instead of discussing the effect of removal of Lys86 or Asp108, we discuss the effect of addition of Lys86 and/or Asp108 to this reference state.

Computational analysis of the electrostatic environment of Lys86 and Asp108

To evaluate the electrostatic environment that Lys86 and Asp108 experience in the structure of RNase H*, we calculated the Coulombic potential energy of each charge at the location of Lys86 and Asp108 using the Tanford–Kirkwood formalism30 with solvent accessibility correction.31 This formalism calculates the work required to put a charge at locations based on the three dimensional structure of a protein. Though using a simple treatment of Coulobmic interactions, the approach has been successfully applied to protein engineering.32, 33 The calculation showed a handful of charges in RNase H* that experience favorable or unfavorable Coulombic potential (Fig. S1 in Supporting Information). The strongly unfavorable Coulombic potential energy of Asp10 is consistent with the previous finding that D10A mutation stabilizes RNase H* by ∼3 kcal/mol under low‐salt conditions.20, 34 Both Lys86 and Asp108 have a favorable Coulombic potential energy and are within top 10 residues with the most favorable potential energies. When the charge of Lys86 or Asp108 is set to 0, the Coulombic potential energy of the partner residue in the pair becomes negligible. The effects of the mutations on the Coulombic potential energy of other charged residues in RNase H* are negligible, suggesting that the Coulombic interaction between Lys86 and Asp108 are somewhat isolated from the rest of the charges in the protein and the favorable Coulombic interaction originates mostly from the interactions between the residues in the pair.

Contribution of the Lys86‐Asp108 salt bridge to global stability

We tested the effect of the mutations on the overall structure of the protein by far‐UV circular dichroism (CD). The spectra of the three mutant proteins are virtually identical to that of wild‐type protein, indicating that the mutations do not have a significant effect on the overall structure of RNase H* (Fig. S2 in Supporting Information). To measure the effect of mutations on the global stability (ΔG U·N°), we monitored urea‐induced equilibrium unfolding of each variant by CD (Fig. ​(Fig.2).2). We determined global stability of the proteins at two different salt concentrations (0.050M and 1.0M NaCl) to assess the effect of ions in the solvent on the energetic contribution of the salt bridge. At 0.050M NaCl, the addition of the negative charge (K86Q RNase H*) or the positive charge (D108N RNase H*) in the background of K86Q/D108N RNase H* destabilizes the protein by 2.8 ± 0.3 kcal/mol or 1.0 ± 0.3 kcal/mol, respectively (Table 1). Interestingly, the addition of both charges (wild‐type RNase H*) does not stabilize the protein significantly (ΔΔG U·N° of 0.1 ± 0.5 kcal/mol). Increasing the salt concentration affected the stability of the proteins favorably. The increase in the concentration of NaCl from 0.050 to 1.0M resulted in the increase in the stability of wild‐type RNase H* by 0.9 kcal/mol (Table 1), which suggests that NaCl may alleviate some unfavorable Coulombic interactions in the protein by shielding charges. The effect of adding the negative charge (K86Q RNase H) on stability is less severe at 1.0M NaCl (1.1 ± 0.5 kcal/mol) than at 0.050M NaCl (2.8 ± 0.5 kcal/mol). At 1.0M NaCl, the effects of adding the positive charge (D108N RNase H*) or both charges (wild‐type RNase H*) are similar to the results at 0.050M NaCl. This result shows that addition of one charge in the salt bridge is detrimental to the stability of the protein, but addition of both charges does not stabilize the protein either, whether the salt concentration is high or low. Apparently, the energetic contribution of the Lys86–Asp108 salt bridge to the global stability of RNase H* is not significantly greater than the energetic contribution of the Gln86–Asn108 pair in the double mutant (the reference state).

An external file that holds a picture, illustration, etc. Object name is PRO-25-0999-g002.jpg

Equilibrium unfolding of wild‐type RNase H* and its variants. Urea‐induced unfolding of wild‐type (●), K86Q (○), D108N (▼), and K86Q/D108N (▵) RNase H* at 0.050M NaCl (A) and 1.0M NaCl (B) was monitored by CD. The CD signal is normalized to the fraction of native protein (f N) for comparison. The results from nonlinear curve‐fitting with a two‐state model are shown as solid lines.

Table 1

Effect of Mutations on Global Stability of RNase H*

RNase H Cm (M) m‐Value (kcal mol−1 M−1)ΔG U–N° (kcal/mol)ΔΔG U–N°a (kcal/mol)
0.050M NaCl
K86Q/D108N4.21 ± 0.012.50 ± 0.0510.5 ± 0.2—
K86Q4.05 ± 0.011.90 ± 0.057.7 ± 0.2−2.8 ± 0.3
D108N4.13 ± 0.012.30 ± 0.069.5 ± 0.2−1.0 ± 0.3
WT* 4.60 ± 0.012.30 ± 0.1010.6 ± 0.50.1 ± 0.5
1.0M NaCl
K86Q/D108N5.31 ± 0.022.10 ± 0.1011.2 ± 0.5—
K86Q5.20 ± 0.011.94 ± 0.0610.1 ± 0.3−1.1 ± 0.6
D108N5.26 ± 0.011.97 ± 0.0510.4 ± 0.3−0.8 ± 0.6
WT* 5.68 ± 0.012.03 ± 0.0711.5 ± 0.40.3 ± 0.6

The free energies for global unfolding (ΔG U‐N°) of wild‐type (WT*), K86A, D108N, and K86Q/D108N RNase H*'s were determined from equilibrium unfolding at 0.050M NaCl and 1.0M NaCl. The standard errors are from curve‐fitting and error propagation.

aΔΔG U‐N° = ΔG U‐N° − ΔG U‐N°(K86Q/D108N).

One caveat of the stability measurements is that the mutations affected the m‐values (the dependence of ΔG U·N° on the concentration of urea) of the protein at 0.050M NaCl. The m‐value of K86Q RNase H* (1.90 ± 0.05 kcal mol−1 M−1) was 0.4 kcal mol−1 M−1 less than that of wild‐type RNase H* (2.30 ± 0.10 kcal mol−1 M−1), while the m‐value of K86Q/D108N RNase H* (2.50 ± 0.05 kcal mol−1 M−1) was 0.2 kcal mol−1 M−1 greater than that of wild‐type RNase H* (Table 1). Interestingly, the effect of mutations on m‐values becomes negligible at 1.0M NaCl (Table 1). Empirically, m‐values are known to correlate with the change in the solvent‐accessible surface area upon unfolding.35 When a point mutation does not affect the structures of a protein in the native and unfolded forms, the mutation is not likely to affect the m‐value associated with global unfolding. As the CD spectra of the mutants are close to that of wild‐type RNase H* (Fig. S2 in Supporting Information), the effect of the mutation on the native form seems minimal. Therefore, the change in the m‐value seems to result from the mutational effects on the unfolded form. The effect of salt on m‐values suggests that the distribution of conformations in the unfolded state ensemble is influenced by some charge‐charge interactions, which can be screened effectively by salt at 1.0M NaCl. In spite of this caveat, the insensitivity of the global stability to the double mutation at both 0.050 and 1.0M NaCl clearly demonstrates that the energetic contribution of the Lys96–Asp108 salt bridge to the global stability of RNase H* is not significant. This finding is consistent with the suggestion that stabilizing salt bridges tend to be part of a network of charge‐charge interactions, and isolated salt bridges are likely to be neutral or destabilizing.36, 37

Contribution of the salt bridge to the energetics of transient partial unfolding

Native‐state proteolysis offers a simple way to assess the effect of a mutation on transient partial unfolding.22, 27, 28, 38 Under our experimental condition, proteolysis of RNase H* by thermolysin occurs in the following kinetic scheme:

Native ⇄Kop Cleavable →kintCleaved,

where K op is the equilibrium constant between the native and the cleavable form, and k int is the pseudo‐first‐order rate constant for the proteolysis of the cleavable form. When the proteolysis step is rate‐limiting in this kinetic scheme (EX2‐like kinetics), a pre‐equilibrium approximation is valid.22 In this kinetic regime, the observed proteolysis rate (k p) is simply expressed as:

kp = Kopkint.
(1)

When k int is approximated as the product of the protease concentration ([E]) and k cat/K m for proteolysis of an unstructured peptide substrate,22 Eq. (1) is recast as:

kp = Kop(kcat/Km)[E].
(2)

By measuring k p at varying concentration of the protease, we determine K op(k cat/K m) from the slope of the plot of k p versus [E]. Using k cat/K m determined with a peptide substrate, we determine K op from which we calculate ΔG C‐N° (=−RTlnK op), the free energy difference between the native form (N) and the cleavable form (C). In case of RNase H*, the initial cleavage by thermolysin occurs at the peptide bond between Thr92 and Ala93 in the protruding loop (Fig. ​(Fig.1),1), and we use the k cat/K m values determined with a tetrapeptide substrate that contains the sequence of the cleavage site (Lys‐Thr‐Ala‐Asp).

The addition of the salt bridge decreases the rate of proteolysis significantly [Fig. ​[Fig.3(A),3(A), Table 2]. At 0.050M NaCl, wild‐type RNase H* showed 14‐fold slower proteolysis kinetics than K86Q/D108N RNase H*. We determined the k cat/K m value with the tetrapeptide substrate to be 3.3 × 105 M−1 s−1 under this experimental condition. Using the k cat/K m value, we determined ΔG C‐N° (Table 2). ΔΔG C‐N°, the effect of addition of the charges on ΔG C‐N°, was determined to be −0.11 ± 0.05 kcal/mol, 0.58 ± 0.04 kcal/mol, and 1.56 ± 0.04 kcal/mol for K86Q, D108N, and wild‐type RNase H*, respectively. Therefore, the addition of the salt bridge increase ΔG C‐N° significantly (1.56 ± 0.04 kcal/mol), though the effect of the addition of the salt bridge on the global stability is negligible (0.1 ± 0.5 kcal/mol).

An external file that holds a picture, illustration, etc. Object name is PRO-25-0999-g003.jpg

Proteolysis kinetics of wild‐type RNase H* and its variants. The rate constants for proteolysis (k p) of wild‐type (●), K86Q (○), D108N (▼), and K86Q/D108N (▵) RNase H* were determined at varying concentrations of thermolysin at 0.050M NaCl (A) and 1.0M NaCl (B). The results from fitting with Eq. (2) are shown as solid lines.

Table 2

Effect of mutations on partial unfolding of RNase H*

RNase H K op (k cat/K m) (M−1 s−1)ΔG C‐N° (kcal/mol)ΔΔG C‐N° (kcal/mol)a
0.050M NaCl
K86Q/D108N390 (± 20)3.98 ± 0.03—
K86Q470 (± 30)3.87 ± 0.04−0.11 ± 0.05
D108N146 (± 3)4.56 ± 0.020.58 ± 0.04
WT* 27.6 (± 0.3)5.54 ± 0.021.56 ± 0.04
1.0M NaCl
K86Q/D108N25 (± 1)5.63 ± 0.03
K86Q54 (± 5)5.18 ± 0.05−0.45 ± 0.06
D108N37 (± 3)5.40 ± 0.05−0.23 ± 0.06
WT*16 (± 1)5.90 ± 0.040.27 ± 0.05

The free energies for partial unfolding (ΔG C‐N°) of wild‐type (WT*), K86Q, D108N, and K86Q/D108N RNase H*'s were determined using native‐state proteolysis at 0.050M and 1.0M NaCl. The standard errors are from curve‐fitting and error propagation

aΔΔG C‐N° = ΔG C‐N° − ΔG C‐N°(K86Q/D108N).

We also determined ΔG C‐N° in 1.0M NaCl (Table 2), using the k cat/K m value determined with the tetrapeptide substrate (3.5 × 105 M−1 s−1) under this experimental condition. 1.0M NaCl increases ΔG C‐N° for all four proteins (Table 2), suggesting that the stabilizing effects of the salt is somewhat greater in the native form than in the cleavable form. 1.0M NaCl also mitigates the effect of the salt bridge on ΔG C‐N° [Fig. ​[Fig.3(B),3(B), Table 2]. The addition of both charges to K86Q/D108N RNase H* decreases the proteolysis rate by less than twofold, which correspond to increase in ΔG C‐N° by ∼0.3 kcal/mol. It is noteworthy that ΔG C‐N° values from native‐state proteolysis have some uncertainty due to the choice of k int and the effect of product inhibition on determination of k p.39 However, the uncertainty in ΔG C‐N° does not affect ΔΔG C‐N° values because the systematic errors in ΔG C‐N° determination are cancelled out in calculation of ΔΔG C‐N°.28

The energy diagram of the free energy of the native and cleavable forms relative to that of the unfolded form shows the effect of the salt bridge and the salt concentration on each form (Fig. ​(Fig.4).4). We calculated the stability of the cleavable form (ΔG U‐C°) by subtracting the energy for partial unfolding (ΔG C‐N°) from the global stability (ΔG U‐N°). The addition of charges affect the energy of the native and cleavable forms in a similar trend but to different extents. Addition of a single charge has somewhat less prominent effects on the cleavable form than on the native form. The notable difference is the effect of the salt bridge at 0.050 mM NaCl. The salt bridge does not affect the energy of the native form (−0.1 ± 0.5 kcal/mol) but significantly destabilizes the cleavable form (1.4 ± 0.5 kcal/mol). 1.0M NaCl clearly stabilizes both native and cleavable forms of wild‐type RNase H* (Fig. ​(Fig.4),4), though the effects of 1.0 NaCl on the proteolysis rates (Table 2) indicate that NaCl stabilizes the native form greater than the cleavable form. The destabilizing effect of the salt bridge in the cleavable form diminishes at 1.0M NaCl, suggesting that 1.0M NaCl effectively mitigates unfavorable energetic contributions of the salt bridge in the cleavable form. In generating the energy diagrams in Figure ​Figure4,4, we assume that the mutations do not affect the energy of the unfolded form and use the unfolded form as the reference state. If mutations affect the energy of the unfolded form, the effect of the mutations on the energy of the cleavable form and the native form can be interpreted somewhat differently. However, regardless of the effect of mutations on the energy of the unfolded form, the presence of the Lys86–Asp108 pair clearly increases the energy gap between the cleavable form and the native form.

An external file that holds a picture, illustration, etc. Object name is PRO-25-0999-g004.jpg

Energy diagrams of global unfolding and partial unfolding of wild‐type RNase H* and its variants. The energies of the cleavable (C) and native (N) forms of K86Q/D108N (0/0), K86Q (0/−), D108N (+/0), and wild‐type (+/−) RNase H* at 0.050M and 1.0M NaCl are shown with the unfolded form (U) as the reference. The standard errors of the free energies are less than 0.5 kcal/mol (Table I).

The interaction energy between Lys86 and Asp108

A double mutant cycle reports the interaction energy (ΔG int°) between two mutated residues. As the stability of the cleavable form (ΔG U‐C°) is available for each variant, we can determine the interaction energy (ΔG int°) between Lys86 and Asp108 in the cleavable form as well as in the native form (Fig. ​(Fig.5).5). The interaction energy (ΔG int°) corresponds to the difference between the effect of the double mutation and the sum of the effects of single mutations. Because we calculated ΔG int° in the double mutant cycle with K86Q/D108N RNase H* as the reference state, ΔG int° actually indicates the gain in the interaction energy by replacing the Gln–Asn pair to the Lys–Asp pair.

An external file that holds a picture, illustration, etc. Object name is PRO-25-0999-g005.jpg

Double‐mutant cycles of Lys86 and Asp108 in RNase H*. Double‐mutant cycles of the effect of mutations on the native (A) and cleavable (B) forms at 0.050M and the native (C) and cleavable (D) forms at 1.0M NaCl were constructed to determine ΔΔG int.° between Lys86 and Asp108.

The ΔG int° value in the native form at 0.050M NaCl was determined to be 3.9 ± 0.6 kcal/mol [Fig. ​[Fig.5(A)],5(A)], which indicates the presence of a strong favorable interaction between Lys86 and Asp108. This interaction energy is much greater than the interaction energy observed in salt bridges in protein surfaces10, 15 and somewhat comparable to those of buried salt bridges.40 According to GetArea,41 Lys86 and Asp108 are 48% and 33% solvent‐exposed, respectively. The partial burial of the salt bridge in an optimal geometry apparently confers the salt bridge the considerable interaction energy. The interaction energy seems mostly result from the favorable Coulombic interaction and stronger hydrogen bonds between Lys86 and Asp108. In spite of the interaction energy, the salt bridge still does not stabilize the native form because the unfavorable contributions from desolvation and possibly unfavorable interactions with the neighboring charges or dipoles cancel out the favorable interaction. The ΔG int° value in the cleavable form at 0.050M NaCl was determined to be 2.9 ± 0.6 kcal/mol [Fig. ​[Fig.5(B)],5(B)], which is lower than that in the native form. Though more steps are involved in the calculation of the interaction energies in the cleavable form, the error in ΔG int° for the cleavable form is not significantly greater than that for the native form because ΔΔG C‐N° has much smaller errors than ΔΔG U‐N° (Tables I and II). The decrease in the interaction energy indicates that partial unfolding induces a less favorable geometry for the salt bridge. Unfolding of the protruding loop in the cleavable form seems to be coupled with some increase in the distance between the two residues in the salt bridge, which would weaken the favorable Coulombic interaction and the hydrogen bonds. This change in the interaction energy explains largely the increase in the free energy for partial unfolding (ΔG C‐N°) by the salt bridge (1.56 kcal/mol) (Table 2, Fig. ​Fig.44).

When the concentration of NaCl is increased from 0.050 to 1.0M, ΔG int° decreases by 1.7 kcal/mol and 1.6 kcal/mol in the native and cleavable forms, respectively [Fig. ​[Fig.5(C,D)].5(C,D)]. Whether the protein is in the native form or in the cleavable form, the increase in the salt concentration apparently exhibits similar effects on the interaction between the two charges in the salt bridge. Also, even at 1.0M NaCl, partial unfolding decreases ΔG int° by 0.9 kcal/mol, which is similar to the change in ΔG int° upon partial unfolding at 0.050M NaCl, suggesting that the salt bridge experiences a similar structural change at both 0.050 and 1.0M NaCl.

Discussion

By using native‐state proteolysis and urea‐induced equilibrium unfolding, we determined the contributions of the Lys86‐Asp108 salt bridge to the energetics of transient partial unfolding as well as the global stability of RNase H*. When the Gln–Asn pair is replaced by the Lys–Asp pair, the change in the global stability is negligible, but the energetics of partial unfolding is significantly altered, especially at low salt concentrations (Fig. ​(Fig.4).4). The cleavage site (Thr92 and Ala93) is located between Lys86 and Asp108 in the primary structure. In the tertiary structure, the salt bridge between Lys86 and Asp108 tethers the loop containing the cleavage site (Fig. ​(Fig.1).1). For proteolysis, the loop containing the cleavage site needs to be extended to fit within the active site of the protease. The tethering of the loop with the salt bridge apparently increases the energetic cost of unfolding the loop for proteolysis. A similar tethering mechanism has been reported in α‐defensin family. The microbicidal peptides in α‐defensin family have a conserved salt bridge on a protruding loop, mutation of which greatly increases the rate of proteolysis without affecting the activity.42, 43, 44 Our observation suggests that a salt bridge can act as a “gatekeeper”45 against partial unfolding even when the salt bridge does not contribute to the global stability.

The sensitivity of the interaction energy to geometry makes salt bridges a useful device for gatekeeping against partial unfolding. The total energetic contribution of a salt bridge to protein stability with respect to that of an equivalent pair of uncharged isosters (ΔΔG tot°) can be expressed as the sum of three components:3, 13

ΔΔGtot°  =  ΔΔGdslv°  +  ΔΔGbrd°  +  ΔΔGprt°,
(3)

where ΔΔG dslv° is the desolvation penalty of the charged residue upon folding, ΔΔG brd° is the favorable interaction energy between the charged residues, and ΔΔG prt° is the interaction energy with the rest of the charges in the protein. The Lys86–Asp108 salt bridge is neither stabilizing nor destabilizing to RNase H* due to the sum of these favorable and unfavorable terms is close to 0. The negligible Coulombic potential energies of Lys86 in D108N variant and Asp108 in K86Q variant (Fig. S1 in Supporting Information) suggests that the contribution of ΔΔG prt° may be insignificant, and ΔΔG dslv° and ΔΔG brd° are major determinants of ΔΔG tot° for the Lys86–Asp108 salt bridge. Partial unfolding in the protruding loop is likely to result in the change in the geometry of the salt bridge in the cleavable form. When this change is subtle, solvation of the salt bridge may not change significantly. However, the interaction energy of the salt bridge is strongly dependent on the geometry, and even a subtle change in the geometry may decrease the interaction energy significantly. Therefore, the neutral salt bridge in the native form becomes unfavorable in the cleavable form, because the considerable desolvation penalty remains mostly while the interaction energy is significantly weakened.

Hydrogen bonds seem to contribute significantly to the interaction energy of the salt bridge between Lys86 and Asp108. According to our model [Fig. ​[Fig.1(B)],1(B)], when the two residues are mutated to Gln and Asn, the two hydrogen bonds with distance of 2.9 Å and 3.1 Å are replaced with one hydrogen bond with distance of 3.3 Å. Because we determined the interaction energy with respect to the double mutant with Gln and Asn (Fig. ​(Fig.5),5), the weakening of the hydrogen bond interaction as well as the loss of the Coulombic interaction contributes to the reduction in ΔΔG brd°. Due to the chemical nature of the functional groups of charged side chains, it is quite common that salt bridges in proteins contain hydrogen bonds.3 Because hydrogen bonds are actually more sensitive to alterations in geometry (distance and angle) than Coulombic interactions, a subtle structural change upon partial unfolding might be more detrimental to the energy of the hydrogen bonds than that of the Coulombic interaction.

The gatekeeping role of salt bridges against partial unfolding has an important implication in protein engineering. Transient partial unfolding under native conditions has a strong relevance with the maintenance of the integrity of protein structure.28 A protein may lose its activity through irreversible aggregation or modification in its partially unfolded forms as well as the globally unfolded form. Under native conditions, however, the population of partially unfolded forms is much greater than that of the globally unfolded form due to the lower free energy. How resistant a protein is against these inactivation events is likely to be dependent more strongly on the energetics of partial unfolding rather than the energetics of global unfolding. To suppress the chance of detrimental inactivation under native conditions, the protein structure may have evolved to increase the energy gap between the native and partially unfolded forms rather than that between the native and globally unfolded forms. Therefore, to engineer robust proteins with longevity, one needs to consider maximizing the free energy of partial unfolding by stabilizing the native form selectively. Stabilizing the native and partially unfolded forms together would not protect proteins from inactivation through partial unfolding. This study suggests that a salt bridge in a proper geometry can be an effective means to suppress partial unfolding. Placing a salt bridge in a region that is disrupted upon partial unfolding will increase the energy gap between the native and partially unfolded forms. Regardless of the effect of the newly introduced salt bridge on the global stability of the protein, the increase in the energy required for partial unfolding will suppress partial unfolding and may confer the protein the desired robustness against inactivation.

Materials and Methods

Expression and purification RNase H* variants

RNase H* variants were constructed with site‐directed mutagenesis using QuikChange (Agilent, Santa Clara, CA), and the mutation was confirmed by DNA sequencing. RNase H* used in this study is a cysteine‐free variant of E. coli ribonuclease HI in which all free cysteine residues are substituted with alanine residues.17 The cysteine‐free variant of RNase H has been used extensively as a model system for folding studies.18, 19, 20, 46 For convenience, we refer to this cysteine‐free variant as wild‐type RNase H* here. Over‐expression and purification were performed as previously described.20 An overnight dialysis was performed after heparin affinity chromatography and Source 15S ion exchange chromatography (GE Healthcare, Piscataway, NJ) against 0.020M sodium acetate buffer (pH 5.5) containing 0.020M NaCl to remove salt from the elution. Purity was confirmed to be greater than 95% for each mutant by SDS‐PAGE.

Calculation of Coulombic potential energy

We calculated the Coulombic potential energy of each charged residue in RNase H* using Tanford–Kirkwood formalism30 with solvent accessibility correction.31 The Coulombic potential energy of charged residue i (E i) is calculated with Eq. (4):

Ei=q2(∑j≠iAij−Bij2b−∑jCij2a)(1−SAi),
(4)

where q is the unit charge (1.602 × 10−19 C), b is the radius of a sphere representing the protein (27.6 Å), a is the radius of the sphere from which solvent ions are excluded (29.5 Å), and SA is the solvent accessibility of the residue. A ij, B ij, and C ij are the three terms to evaluate charge–charge interactions in the protein.30 The solvent accessibility was determined using ASAView.47 We used the dielectric constant of 4 for the protein interior and 78.5 for the protein exterior.

Circular dichroism spectra

Protein samples were allowed to equilibrate overnight in 0.020M sodium acetate buffer (pH 5.5) containing 0.050M NaCl, 1.0 mM CaCl2, and 0.030 mg/mL protein. Spectra were collected for 400–200 nm using a Jasco J‐815 Circular Dichroism spectrophotometer (Easton, MA) in a 1‐cm quartz cuvette.

Determination of global stability

The effect of the mutation on global stability of each RNase H* variant was determined by monitoring equilibrium unfolding in 0.020M sodium acetate buffer (pH 5.5) containing 0.050M or 1.0M NaCl, 1.0 mM CaCl2, 0.030 mg/mL protein, and varying concentrations of urea (0–9M). As urea modifies pK a of acetate,48 pH was adjusted to 5.5 with 1.0M HCl after the solutions were prepared. We monitored the degree of unfolding by the change in ellipticity at 222 nm using a Jasco J‐815 circular dichroism spectrophotometer (Easton, MA) in a 1‐cm quartz cuvette. The resulting equilibrium unfolding curves were fit with OriginPro (Northampton, MA) by the linear extrapolation method with a two‐state assumption.49

Native‐state proteolysis

Native‐state proteolysis was conducted as described previously.22 Briefly, proteolysis was initiated by adding a small volume of a concentrated thermolysin stock solution to a 0.50 mg/mL protein solution in 0.020M sodium acetate buffer (pH 5.5) containing 0.050 or 1.0M NaCl and 1 mM CaCl2. At designated time points, 15 μL aliquots of the reaction were quenched by 5 μL 0.050M EDTA (pH 8.0). The amounts of remaining intact protein were determined by quantifying the band intensities of intact RNase H* on SDS‐PAGE gels. The plots of the band intensity of intact protein versus time were fit with a first‐order rate equation to determine k obs. K op(k cat/K m) values were determined from the slope of the plot of k obs versus the concentrations of thermolysin. The k cat/K m values were determined with a fluorogenic substrate with the sequence of the initial cleavage site in RNase H* (ABS‐Lys‐Thr‐Ala‐Asp‐NBA)22 in the identical conditions for native‐state proteolysis as described above.

Supporting information

Supporting Information

Acknowledgments

The authors thank Joseph R. Kasper, Nathan Gardner, and Chen Chen for helpful comments on this manuscript and the Topp laboratory for sharing their CD instrument.

References

1. Perutz MF (1978) Electrostatic effects in proteins. Science 201:1187–1191. [PubMed] [Google Scholar]
2. Barlow DJ, Thornton JM (1983) Ion‐pairs in proteins. J Mol Biol 168:867–885. [PubMed] [Google Scholar]
3. Kumar S, Nussinov R (1999) Salt bridge stability in monomeric proteins. J Mol Biol 293:1241–1255. [PubMed] [Google Scholar]
4. Vogt G, Woell S, Argos P (1997) Protein thermal stability, hydrogen bonds, and ion pairs. J Mol Biol 269:631–643. [PubMed] [Google Scholar]
5. Kumar S, Tsai CJ, Nussinov R (2000) Factors enhancing protein thermostability. Protein Eng 13:179–191. [PubMed] [Google Scholar]
6. Yip KS, Britton KL, Stillman TJ, Lebbink J, de Vos WM, Robb FT, Vetriani C, Maeder D, Rice DW (1998) Insights into the molecular basis of thermal stability from the analysis of ion‐pair networks in the glutamate dehydrogenase family. Eur J Biochem 255:336–346. [PubMed] [Google Scholar]
7. Anderson DE, Becktel WJ, Dahlquist FW (1990) pH‐induced denaturation of proteins: a single salt bridge contributes 3‐5 kcal/mol to the free energy of folding of T4 lysozyme. Biochemistry 29:2403–2408. [PubMed] [Google Scholar]
8. Marqusee S, Sauer RT (1994) Contributions of a hydrogen bond/salt bridge network to the stability of secondary and tertiary structure in lambda repressor. Protein Sci 3:2217–2225. [PMC free article] [PubMed] [Google Scholar]
9. Luisi DL, Snow CD, Lin JJ, Hendsch ZS, Tidor B, Raleigh DP (2003) Surface salt bridges, double‐mutant cycles, and protein stability: an experimental and computational analysis of the interaction of the Asp 23 side chain with the N‐terminus of the N‐terminal domain of the ribosomal protein l9. Biochemistry 42:7050–7060. [PubMed] [Google Scholar]
10. Serrano L, Horovitz A, Avron B, Bycroft M, Fersht AR (1990) Estimating the contribution of engineered surface electrostatic interactions to protein stability by using double‐mutant cycles. Biochemistry 29:9343–9352. [PubMed] [Google Scholar]
11. Sun DP, Sauer U, Nicholson H, Matthews BW (1991) Contributions of engineered surface salt bridges to the stability of T4 lysozyme determined by directed mutagenesis. Biochemistry 30:7142–7153. [PubMed] [Google Scholar]
12. Waldburger CD, Schildbach JF, Sauer RT (1995) Are buried salt bridges important for protein stability and conformational specificity? Nat Struct Biol 2:122–128. [PubMed] [Google Scholar]
13. Hendsch ZS, Tidor B (1994) Do salt bridges stabilize proteins? A continuum electrostatic analysis. Protein Sci 3:211–226. [PMC free article] [PubMed] [Google Scholar]
14. Kumar S, Ma B, Tsai CJ, Sinha N, Nussinov R (2000) Folding and binding cascades: Dynamic landscapes and population shifts. Protein Sci 9:10–19. [PMC free article] [PubMed] [Google Scholar]
15. Makhatadze GI, Loladze VV, Ermolenko DN, Chen X, Thomas ST (2003) Contribution of surface salt bridges to protein stability: Guidelines for protein engineering. J Mol Biol 327:1135–1148. [PubMed] [Google Scholar]
16. Bosshard HR, Marti DN, Jelesarov I (2004) Protein stabilization by salt bridges: Concepts, experimental approaches and clarification of some misunderstandings. J Mol Recognit 17:1–16. [PubMed] [Google Scholar]
17. Dabora JM, Marqusee S (1994) Equilibrium unfolding of Escherichia coli ribonuclease H: Characterization of a partially folded state. Protein Sci 3:1401–1408. [PMC free article] [PubMed] [Google Scholar]
18. Chamberlain AK, Handel TM, Marqusee S (1996) Detection of rare partially folded molecules in equilibrium with the native conformation of RNase H. Nat Struct Biol 3:782–787. [PubMed] [Google Scholar]
19. Raschke TM, Marqusee S (1997) The kinetic folding intermediate of ribonuclease H resembles the acid molten globule and partially unfolded molecules detected under native conditions. Nat Struct Biol 4:298–304. [PubMed] [Google Scholar]
20. Raschke TM, Kho J, Marqusee S (1999) Confirmation of the hierarchical folding of RNase H: A protein engineering study. Nat Struct Biol 6:825–831. [PubMed] [Google Scholar]
21. Hu W, Walters BT, Kan ZY, Mayne L, Rosen LE, Marqusee S, Englander SW (2013) Stepwise protein folding at near amino acid resolution by hydrogen exchange and mass spectrometry. Proc Natl Acad Sci USA 110:7684–7689. [PMC free article] [PubMed] [Google Scholar]
22. Park C, Marqusee S (2004) Probing the high energy states in proteins by proteolysis. J Mol Biol 343:1467–1476. [PubMed] [Google Scholar]
23. Tyndall JD, Nall T, Fairlie DP (2005) Proteases universally recognize beta strands in their active sites. Chem Rev 105:973–999. [PubMed] [Google Scholar]
24. Linderstrøm‐Lang K, Hotchkiss RD, Johansen G (1938) Peptide bonds in globular proteins. Nature 142:996. [Google Scholar]
25. Rupley JA (1967) Susceptibility to attack by proteolytic enzymes. Methods Enzymol 11:905–917. [Google Scholar]
26. Hubbard SJ, Eisenmenger F, Thornton JM (1994) Modeling studies of the change in conformation required for cleavage of limited proteolytic sites. Protein Sci 3:757–768. [PMC free article] [PubMed] [Google Scholar]
27. Chang Y, Park C (2009) Mapping transient partial unfolding by protein engineering and native‐state proteolysis. J Mol Biol 393:543–556. [PubMed] [Google Scholar]
28. Park C (2014) Probing transient partial unfolding in proteins by native‐state proteolysis. Bio Des 3:117–128. [Google Scholar]
29. Youn K, Park C (2009) Investigating the effect of temperature on transient partial unfolding by proteolysis. Protein Pept Lett 16:1093–1097. [PubMed] [Google Scholar]
30. Tanford C, Kirkwood JG (1957) Theory of protein titration curves: I. General equations for impenetrable spheres. J Am Chem Soc 79:5333–5339. [Google Scholar]
31. Matthew JB, Gurd FR (1986) Calculation of electrostatic interactions in proteins. Methods Enzymol 130:413–436. [PubMed] [Google Scholar]
32. Gribenko AV, Patel MM, Liu J, McCallum SA, Wang C, Makhatadze GI (2009) Rational stabilization of enzymes by computational redesign of surface charge‐charge interactions. Proc Natl Acad Sci USA 106:2601–2606. [PMC free article] [PubMed] [Google Scholar]
33. Zhang L, Tang X, Cui D, Yao Z, Gao B, Jiang S, Yin B, Yuan YA, Wei D (2014) A method to rationally increase protein stability based on the charge‐charge interaction, with application to lipase LipK107. Protein Sci 23:110–116. [PMC free article] [PubMed] [Google Scholar]
34. Kanaya S, Oobatake M, Liu Y (1996) Thermal stability of Escherichia coli ribonuclease HI and its active site mutants in the presence and absence of the Mg2+ ion: Proposal of a novel catalytic role for Glu48. J Biol Chem 271:32729–32736. [PubMed] [Google Scholar]
35. Myers JK, Pace CN, Scholtz JM (1995) Denaturant m values and heat capacity changes: Relation to changes in accessible surface areas of protein unfolding. Protein Sci 4:2138–2148. [PMC free article] [PubMed] [Google Scholar]
36. Xiao L, Honig B (1999) Electrostatic contributions to the stability of hyperthermophilic proteins. J Mol Biol 289:1435–1444. [PubMed] [Google Scholar]
37. Kumar S, Ma B, Tsai CJ, Nussinov R (2000) Electrostatic strengths of salt bridges in thermophilic and mesophilic glutamate dehydrogenase monomers. Proteins 38:368–383. [PubMed] [Google Scholar]
38. Kasper JR, Liu PF, Park C (2014) Structure of a partially unfolded form of Escherichia coli dihydrofolate reductase provides insight into its folding pathway. Protein Sci 23:1728–1737. [PMC free article] [PubMed] [Google Scholar]
39. Kasper JR, Andrews EC, Park C (2014) Product inhibition in native‐state proteolysis. PLoS One 9:e111416. [PMC free article] [PubMed] [Google Scholar]
40. Robinson AC, Castaneda CA, Schlessman JL, Garcia‐Moreno EB (2014) Structural and thermodynamic consequences of burial of an artificial ion pair in the hydrophobic interior of a protein. Proc Natl Acad Sci USA 111:11685–11690. [PMC free article] [PubMed] [Google Scholar]
41. Fraczkiewicz R, Braun W (1998) Exact and efficient analytical calculation of the accessible surface areas and their gradients for macromolecules. J Comput Chem 19:319–333. [Google Scholar]
42. Wu Z, Li X, de Leeuw E, Ericksen B, Lu W (2005) Why is the Arg5‐Glu13 salt bridge conserved in mammalian alpha‐defensins? J Biol Chem 280:43039–43047. [PubMed] [Google Scholar]
43. Rajabi M, de Leeuw E, Pazgier M, Li J, Lubkowski J, Lu W (2008) The conserved salt bridge in human alpha‐defensin 5 is required for its precursor processing and proteolytic stability. J Biol Chem 283:21509–21518. [PMC free article] [PubMed] [Google Scholar]
44. Andersson HS, Figueredo SM, Haugaard‐Kedstrom LM, Bengtsson E, Daly NL, Qu X, Craik DJ, Ouellette AJ, Rosengren KJ (2012) The a‐defensin salt‐bridge induces backbone stability to facilitate folding and confer proteolytic resistance. Amino Acids 43:1471–1483. [PMC free article] [PubMed] [Google Scholar]
45. Miller KH, Karr JR, Marqusee S (2010) A hinge region cis‐proline in ribonuclease A acts as a conformational gatekeeper for C‐terminal domain swapping. J Mol Biol 400:567–578. [PMC free article] [PubMed] [Google Scholar]
46. Dabora JM, Pelton JG, Marqusee S (1996) Structure of the acid state of Escherichia coli ribonuclease HI. Biochemistry 35:11951–11958. [PubMed] [Google Scholar]
47. Ahmad S, Gromiha M, Fawareh H, Sarai A (2004) ASAView: Database and tool for solvent accessibility representation in proteins. BMC Bioinformatics 5:51. [PMC free article] [PubMed] [Google Scholar]
48. Das Gupta PK, Moulik SP (1987) Interaction of urea with weak acids and water. J Phys Chem 91:5826–5832. [Google Scholar]
49. Greene RF, Pace CN (1974) Urea and guanidine‐hydrochloride denaturation of ribonuclease, lysozyme, a‐chymotrypsin, and b‐lactoglobulin. J Biol Chem 249:5388–5393. [PubMed] [Google Scholar]
Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society