Skip to main content
NCBI home page
Biophysical Journal logoLink to Biophysical Journal
. 2013 Sep 17;105(6):1491–1501. doi: 10.1016/j.bpj.2013.08.008

Proline Substitution of Dimer Interface β-strand Residues as a Strategy for the Design of Functional Monomeric Proteins

Prem Raj B Joseph †,‡, Krishna Mohan Poluri †,‡, Pavani Gangavarapu †, Lavanya Rajagopalan †,‡, Sandeep Raghuwanshi ‖, Ricardo M Richardson ‖, Roberto P Garofalo §,, Krishna Rajarathnam †,¶,‡,∗
PMCID: PMC3785864  PMID: 24048001

Abstract

Proteins that exist in monomer-dimer equilibrium can be found in all organisms ranging from bacteria to humans; this facilitates fine-tuning of activities from signaling to catalysis. However, studying the structural basis of monomer function that naturally exists in monomer-dimer equilibrium is challenging, and most studies to date on designing monomers have focused on disrupting packing or electrostatic interactions that stabilize the dimer interface. In this study, we show that disrupting backbone H-bonding interactions by substituting dimer interface β-strand residues with proline (Pro) results in fully folded and functional monomers, by exploiting proline’s unique feature, the lack of a backbone amide proton. In interleukin-8, we substituted Pro for each of the three residues that form H-bonds across the dimer interface β-strands. We characterized the structures, dynamics, stability, dimerization state, and activity using NMR, molecular dynamics simulations, fluorescence, and functional assays. Our studies show that a single Pro substitution at the middle of the dimer interface β-strand is sufficient to generate a fully functional monomer. Interestingly, double Pro substitutions, compared to single Pro substitution, resulted in higher stability without compromising native monomer fold or function. We propose that Pro substitution of interface β-strand residues is a viable strategy for generating functional monomers of dimeric, and potentially tetrameric and higher-order oligomeric proteins.

Introduction

Proteins that exist in monomer-dimer equilibrium are highly prevalent in biological systems from bacteria to humans, indicating positive selection and evolutionary advantages. There is substantial literature indicating that the ability to reversibly exist in two forms allows exquisite spatiotemporal regulation for a wide variety of functions, ranging from monomer/dimer-specific activity and allosteric regulation to differential activation of signaling pathways or simply increasing stability and protection from proteolysis (1–3). However, dissecting the structural basis and understanding the molecular mechanisms underlying the functions of the monomer are not trivial as the very phenomenon of the monomer-dimer equilibrium prevents the study of one species without interference from the other.

Protein dimer interfaces are stabilized via a combination of H-bonding, van der Waals, and electrostatic interactions, and furthermore, dimer association constants can vary by orders of magnitude from mM to pM affinities (3). In principle, monomers of weakly dimerizing proteins can be characterized by changing pH, protein concentration, or buffer conditions (4,5). The most common approach of introducing mutations is by destabilizing packing or electrostatic interactions of dimer interface residues. Such approaches have been successful in generating monomers (6–8). These studies, typically carried out with the specific goal of generating a monomer of a protein of interest, cannot ensure that the same strategy will work for other proteins. To our knowledge, there have been no systematic studies with the objective of establishing guidelines for the rational design of monomers. An ideal design strategy would be one that can be broadly applied and which minimally perturbs the native monomer structure. In this study, we show that disrupting β-strand dimer interface backbone H-bonding interactions via proline substitution is a viable approach for generating folded and fully functional monomers (Fig. 1 A).

Figure 1.

Figure 1

Open in a new tab

(A) Schematic showing Pro substitution strategy at the dimer interface for monomer design. The figure on top shows the H-bonding interactions between antiparallel β-strands of a dimer interface. On Pro substitution (bottom figure), the loss of H-bonding and disruption of the dimer interface is depicted using a blast and outward arrows. (B) Hydrogen-bonding network at the IL-8 dimer interface. Residues involved in H-bonding are labeled and ′ indicates the second monomer in the homodimer.

Proline is an unusual amino acid and its unique features contribute to the special roles that it plays in protein structure and function. Structure-function studies have shown that the prolines are usually conserved and/or are often essential, and play important roles in protein structure, stability, folding, splicing, cell signaling, and transmembrane active transport (9–14). The proline pyrrolidine ring side chain is covalently bonded to the backbone nitrogen, and therefore cannot form a backbone H-bond due to lack of the amide proton. The cyclic pyrrolidine side chain restricts its backbone conformation, in particular the ϕ angle to −63 ± 15°, resulting in proline adopting two distinct conformations tightly clustered around ψ ∼ −35° for the α-region and ψ ∼150° for the β-region. In addition, the conformational space of the preceding residue is influenced by the steric conflict from CδH2 attached to the imide nitrogen of proline (9). Not surprisingly, prolines are mostly observed in loops, turns, and at the beginning but not in the middle of helices or strands (9,10,14,15).

In this current study, we have explored proline substitution as a strategy for generating monomers by exploiting proline’s unique structural and conformational properties. We have used interleukin-8 (IL-8) as a model system, which exists reversibly as monomers and dimers (Kd ∼0.1 to 10 μM) (16,17). The structures and activities of the IL-8 monomers and dimers and of the importance of the monomer-dimer equilibrium have been well studied (18–21). The solution structures of the wild-type (WT) dimer and of a trapped nonassociating monomer exist (18–20). The trapped monomer was chemically synthesized and contains a nonnatural N-methyl amino acid for the dimer interface Leu-25 residue, which disrupts dimer interface H-bonding interactions (referred as the L25NMe trapped monomer). The receptor affinity and activity of the trapped monomer and the WT monomer are identical (22). Mutagenesis studies of IL-8 dimer interface residues have also shown that disrupting dimer-packing interactions can result in monomers or impaired dimers (16,17,23).

Because IL-8 dimerizes via β1-strand residues (Fig. 1 B), we generated a panel of single and double mutants by substituting prolines for each of the dimer interface H-bond-forming residues. We characterized the structures, dynamics, stability, dimerization state, and activities of the Pro mutants using NMR, molecular dynamics (MD) simulations, fluorescence, cellular assays, and animal models. Our data convincingly show that a single Pro substitution in the middle of the dimer interface is sufficient to generate a properly folded functional monomer, and most interestingly, double Pro substitutions, compared to single Pro substitutions, result in higher stability without compromising native monomer fold or function.

Materials and Methods

Cloning, expression, and purification of IL-8 Pro mutants

Clones of the single and double Pro mutants (L25P, V27P, E29P, L25P/E29P, V27P/E29P, and L25P/V27P) were generated using the Stratagene QuikChange site-directed mutagenesis protocol (24), and recombinantly expressed and purified as discussed earlier (25). 15N- and 15N/13C-labeled Pro mutants were produced by growing cells in minimal media containing 15NH4Cl and 13C glucose as the sole nitrogen and carbon sources, respectively. The purity and molecular weight of the proteins were confirmed using matrix-assisted laser desorption/ionization mass spectrometry.

NMR spectroscopy

15N- and 15N/13C-labeled Pro mutants were prepared in 50 mM sodium phosphate pH 6.0 buffer containing 1 mM DSS (2,2-dimethyl-2-silapentanesulfonic acid), 1 mM sodium azide, and 10% 2H2O (v/v). The protein concentrations were ∼0.2 to 0.3 mM. The chemical shifts of the Pro mutants were assigned using 1H-15N HSQC, 15N-edited NOESY, 15N-edited TOCSY, CBCA(CO)NH, and HNCACB NMR data (26). The NMR spectra were acquired at 30°C using a Bruker Avance III 800 MHz (equipped with a TXI cryoprobe) or 600 MHz (equipped with a QCI cryoprobe) spectrometers. The mixing time for the 15N-edited NOESY and TOCSY experiments were 150 and 80 msec, respectively. Chemical shifts were referenced to DSS. The spectra were processed with NMRPipe (27), and analyzed using NMRView (28) or Bruker Topspin 3.2 software.

NMR self-diffusion coefficients were measured using a stimulated echo and longitudinal-eddy-current delay incorporating bipolar gradient pulses for diffusion (29). The diffusion delay (Δ) was 80 msec and the gradient pulse length (δ) was 5 msec. Sixteen 1D spectra were collected at equal intervals by varying the gradient strength from 10% to 95%, with 54.4 G/cm as the maximum (100%) gradient strength. A 1% (w/v) solution of β-cyclodextrin in 90% H2O and 10% D2O, with a diffusion coefficient 3.239 × 10−6 cm2/sec at 25°C, was used as a standard for gradient strength calibration (30). The self-diffusion coefficients (DS) were calculated by nonlinear least-square fitting of the intensities from the individual one-dimensional spectra using the analysis tool in Bruker Topspin 3.2 software.

Native-state hydrogen exchange (NHX) studies of the Pro mutants were initiated by dissolving the lyophilized protein samples in D2O. The samples were loaded onto a tuned and shimmed NMR spectrometer at 25°C, and a series of 1H-15N HSQC spectra were recorded for 6–48 h time period depending on the rate of decay. The time from the addition of D2O to the start of the first HSQC spectrum was 8 min (dead time). Each spectrum was recorded for 22 min and consisted of 64 complex increments in the indirect 15N-dimension. The NHX decay rates (kobs) of the individual NH protons were calculated using nonlinear least-square fitting of the peak intensities, and the stabilization free energies (ΔGNHX) were calculated as discussed previously (31).

Stability of the Pro mutants

The stability of the Pro mutants was determined from changes in tryptophan fluorescence. The fluorescence spectra were collected on a Spex Fluoromax fluorimeter (Horiba Jobin Yvon, Edison, NJ), with excitation at 295 nm and emission measured from 300 to 400 nm at 25°C. The blank spectrum of the buffer solution was subtracted from each sample spectrum. The data were collected using 5 μM protein samples in 50 mM sodium phosphate buffer, pH 6.0, and the GdnHCl concentration was varied from 0 to 8 M at intervals of 0.3 M. The measurements were repeated twice with freshly prepared protein samples, and the free energy of unfolding (ΔGF-U) was obtained using the linear extrapolation method (32). As the unfolding of WT and E29P dimer is concentration dependent and those of the Pro monomer mutants are not, we have calculated ΔGF-U of the WT and E29P using the 1M standard state as described previously (33).

MD simulations

The L25P, V27P, E29P, and V27P/E29P mutants were modeled in both the monomeric and dimeric forms from the NMR IL-8 structure using Pymol (18,34). The WT dimer, WT monomer, and the eight Pro variants (four each in the monomer and dimer forms, respectively) were subjected to multiple cycles of constrained and free energy minimizations using the AMBER 12 suite software to remove steric hindrances introduced by the mutations (35,36). The energy-minimized structures were subjected to an equilibration protocol in explicit solvent (36), and 500 ns of MD production runs were carried out using the PMEMD (particle mesh Ewald molecular dynamics) module of the AMBER 12 software suite on the Lonestar Dell Linux Cluster at the Texas Advanced Computing Center (TACC, UT Austin, TX). Analysis of the trajectory was carried out using the Ambertools12 and the VMD molecular visualization software (35,37). All molecular plots were prepared using Pymol (34).

Intracellular Ca2+ mobilization

CXCR2-transfected RBL-2H3 cells (5 × 106) were washed with HEPES-buffered saline and loaded with 1 μM Indo-I/AM in the presence of 1 μM pluronic acid for 30 min at room temperature. The cells were then washed with HEPES and resuspended in 1.5 ml of Siriganian buffer, and the intracellular Ca2+ mobilization activity of the Pro mutants were measured as described previously (38).

Heparin binding

The WT and Pro mutants (∼0.3 mg/ml) were prepared in 5 mM sodium phosphate pH 7.0 buffer, and loaded onto a HiTrap Heparin affinity column (GE LifeSciences). The proteins were eluted using a step gradient (50 mM) of 0–1 M sodium chloride in the same buffer, and their concentrations were determined using ultraviolet absorption spectroscopy at 280 nm.

Neutrophil recruitment in a mouse peritoneum model

8- to 10-week-old female BALB/c mice were purchased from Harlan (Houston, TX) and housed under pathogen-free conditions in the animal research facility, in accordance with the National Institutes of Health and University guidelines for animal care. Under light anesthesia, mice were inoculated intraperitoneally with 1 μg of IL-8 WT and Pro mutants in Dulbecco’s phosphate-buffered saline (D-PBS). Mice were sacrificed, and peritoneal neutrophil levels were determined as described (39).

Results

Design and oligomeric state of the Pro mutants

The dimer interface β1-strand L25, V27, and E29 amides are involved in a total of six backbone H-bond interactions across the dimer interface with the β1′-strand (Fig. 1 B; residues of the second monomer across the dimer interface are represented by ′). We substituted Pro for each of these three residues (L25, V27, and E29), and generated three single (L25P, V27P, and E29P) and three double (L25P/E29P, L25P/V27P, and V27P/E29P) mutants. In the case of L25P/V27P double mutant, despite multiple attempts, we could not isolate the protein during the purification process due to possible precipitation, and so was not pursued further. We determined the oligomeric state of the Pro mutants using NMR self-diffusion coefficient (DS) measurements. Nonlinear least-squares fitting of the intensity data gave a DS value of 0.95 × 10 −6 cm2 sec−1 for the E29P, which was identical as for the WT dimer. For the V27P, L25P, V27P/E29P, and L25P/E29P mutants, a DS value of ∼1.14 × 10−6 cm2 sec−1 was obtained that is consistent with that expected for the monomer (Table 1 and Fig. S1 in the Supporting Material). DS(monomer)/DS(dimer) ratio of ∼1.21 obtained for IL-8 is consistent with those expected for a monomeric protein of spherical shape (40).

Table 1.

NMR translation self-diffusion coefficients calculated for the IL-8 Pro mutants

IL-8 variants Diffusion coefficient (cm2/s)
WT 0.96 ± 0.02
E29P 0.95 ± 0.03
V27P 1.14 ± 0.02
V27P/E29P 1.14± 0.02
L25P 1.07± 0.03
L25P/E29P 1.15± 0.03
Open in a new tab

NMR structural characterization of the Pro mutants

We characterized the structural features of the single and the double mutants using solution NMR spectroscopy. 1H-15N HSQC spectra of the mutants showed the characteristic upfield and downfield amide proton chemical shifts (Q8, K15, F17, K20, C34, and V58) indicating that all the mutants have an IL-8-like fold (Fig. S2). The E29P mutant showed chemical shift profiles comparable to the WT dimer, whereas the remaining mutants showed characteristics of the monomer. Based on the quality of the HSQC data, we chose the V27P, V27P/E29P, and E29P mutants for further detailed structural characterization.

Comparison of the backbone HN and Hα chemical shifts between the V27P and V27P/E29P mutants and the trapped L25NMe monomer showed that the perturbations are localized in and around the substitution sites in the β1- and β2-strands (Fig.S3, A–C) (20). Analysis of the secondary structural propensities of the V27P and V27P/E29P mutants using the program CSI based on Hα, Cα, and Cβ chemical shifts also suggests that the monomer structural scaffold is intact (Fig. S3 E) (41). Furthermore, the last six residues in the C-terminal helix are unstructured, a feature that is also observed in the structure of the trapped L25NMe monomer (20). Analysis of the three-dimensional 15N-edited NOESY-HSQC spectra of V27P and V27P/E29P mutants showed no evidence of intermolecular NOEs from β1-strand backbone amides, indicating that the mutants are monomeric (Fig. 2 A).

Figure 2.

Figure 2

Open in a new tab

NMR NOESY spectra of IL-8 Pro mutants. (A) Strip plots of the 15N-edited NOESY spectra from L25 and E29 amides of the V27P mutant and L25 amide of the V27P/E29P mutant show no evidence of intermolecular NOEs. The missing NOEs are shown as open gray circles. The E29P mutant shows intermolecular NOEs at the dimer interface. (B and C) Strip plots of the 15N-edited V27P/E29P and V27P NOESY spectra of the β1- and β2-strand interface residues. The interstrand NOEs between the β1- and β2-strands are labeled in bold italics.

On the other hand, a number of characteristic intermolecular NOEs from the amide protons of L25 and V27 were observed for the E29P mutant confirming that it is dimeric (Fig. 2 A). The CSI plot indicates that the secondary structural features of E29P are very similar to WT dimer (Fig. S3 E). Comparison of the HN and Hα chemical shifts between E29P and WT dimer also showed that the structural changes are localized to the β1-strand and C-terminal end of the α-helix (Fig. S3 D) (18).

The network of intramolecular interactions between the adjacent β1- and β2-strands were intact in the V27P, V27P/E29P, and E29P mutants except for local perturbations at the site of substitution, indicating that the proline can be accommodated with no major structural perturbation of the β1-strand conformation. Strip plots of three-dimensional 15N-edited NOESY-HSQC planes of amino acids in the β1- and β2-strands illustrating the intramolecular interactions for V27P and V27P/E29P are shown in Fig. 2, B and C. Analysis of the backbone conformations of the V27P/E29P mutant based on NMR chemical shift data using TALOS+ also showed only local perturbations in the β1-strand conformation (42). The ϕ/ψ angles were −68°/145° for P27 and −61°/145° for P29, which fall into the allowed proline conformational space. Except for the ϕ angle of R26 (∼ −107°) that deviates from the WT values (ϕ ∼ −122°), the remaining residues retain the WT β-sheet conformation. Most interestingly, the V27P/E29P double mutant showed stronger intramolecular NOEs between the β1- and β2-strands, suggesting that the double mutant is more stable compared to the single V27P mutant.

Equilibrium unfolding measurements

The IL-8 structure reveals a single tryptophan (Trp) at the beginning of the α-helix, and changes in its fluorescence emission wavelength can be used as a probe to study the global unfolding of the protein (43). The stability (ΔGF-U) of the Pro mutants was measured using fluorescence spectroscopy and guanidine hydrochloride (GdnHCl) as the denaturant (Fig. 3 and Table 2). Analysis of the unfolding data showed that E29P was thermodynamically slightly less stable than the WT, which can be attributed to the loss of E29-K23 H-bonding and E29 side-chain packing interactions with the overlying helix. Under the experimental conditions (5 μM protein in 50 mM phosphate buffer pH 6.0), the monomeric mutants were less stable, and can be directly attributed to loss of dimer stabilizing interactions. In addition to the six H-bonding interactions, the IL-8 structures also reveal that the dimer interface is stabilized by packing interactions involving a number of residues including L25, V27, and E29 side chains (18,19). Most interestingly, the double mutants, L25P/E29P and V27P/E29P, were more stable by ∼0.5 to 1 kcal/mol compared to their single mutant counterparts (Table 2). The higher stability for V27P/E29P is also consistent with stronger interstrand NOEs compared to the V27P mutant, and from hydrogen/deuterium exchange studies (see below).

Figure 3.

Figure 3

Open in a new tab

Stability of the Pro mutants. The unfolding isotherms of IL-8 WT and Pro mutants. The plot shows the λmax of tryptophan fluorescence against GdnHCl concentration.

Table 2.

Fluorescence equilibrium unfolding data for the IL-8 Pro mutants

IL-8 variants ΔGunfolding (kcal/mol) C1/2 (M) m-value (cal/[mol M])
WTa −8.2 ± 0.5 4.1 1992 ± 109
E29Pa −7.7 ± 0.5 4.6 1660 ± 97
V27P −2.5 ± 0.2 1.8 1398 ± 78
V27P/E29P −3.2 ± 0.6 2.1 1521 ± 124
L25P −2.2 ± 0.6 1.5 1462 ± 231
L25P/E29P −3.2 ± 0.4 2.1 1559 ± 160
Open in a new tab
a

The ΔGunfolding for the WT and E29P are given in kcal/mol monomer and are twice if given per mol dimer.

Stability differences among the Pro mutants were also evident from the Trp λmax showing a redshift, compared to the WT (Fig. 3). A redshift even in the case of the E29P mutant indicates that the Trp side chain is more solvent-exposed suggesting subtle structural changes caused by altered packing interactions and breathing motions or dynamics. The larger redshift for the monomers indicates higher solvent exposure due to loss of quaternary structure, which is also evident from the hydrogen/deuterium exchange studies as described below.

Residue-level stabilities of the Pro mutants

NHX monitored by NMR spectroscopy is a powerful tool for studying residue-wise free energy stabilities and folding energy landscapes of proteins (44–47). The exchange rates of backbone amide protons depend on the accessibility to the solvent deuterons, which in turn correlate to the secondary, tertiary, and quaternary structures, and thermodynamic stability.

The first spectra of the WT and Pro mutants, after exchange was initiated, are shown in Fig. 4 A. NHX data of the WT dimer indicate that about half of the resonances disappear within the dead time (8 min), and these solvent accessible residues mostly reside in the N-terminal, N-loop, C-terminal, and interconnecting loops of the protein. The slow-exchanging residues located in the β1-, β2-, β3-strands and the α-helix are predominantly hydrophobic and form the protein core (Fig. 4 and Table S1). Sequence analyses of nine related chemokines from mouse and human reveal that most of these hydrophobic residues are highly conserved, indicating their fundamental roles in defining the chemokine fold and stability. The slow exchanging charged residues (R26, E38, and K42) are involved in interstrand H-bonding interactions.

Figure 4.

Figure 4

Open in a new tab

NHX data of the Pro mutants. (A) The first HSQC spectra of the WT dimer and Pro mutants, after initiating exchange with D2O at pH 6.0 and 25°C (dead time ∼8 min). The dimer-interface residues not observed in the E29P mutant are labeled in bold italics in the WT dimer spectrum. The β1-β2 strand interface and helical residues that are more protected in the V27P/E29P, compared to the V27P mutant, are also labeled in bold italics. (B) The decay profiles for Leu-43 of WT and Pro mutants. (C) Plot of the stabilization free energy, ΔGHX of individual amino acids in the WT (open bars) and the Pro mutants (shown using the same symbol scheme as in panel B). Core residues that are significantly protected from exchange in the experimental time frame have been given an upper ΔGHX value of 8.0 kcal/mol.

The residue-specific stabilities of the Pro mutants were determined from the exchange rates (kex) (31). NHX can provide information about the thermodynamics as well as kinetics depending on the exchange regime (EX2/EX1), and it is well established that the amide protons exist in the EX2 regime at acidic and neutral pH. Residues whose intensities remain high in the experimental time (48 h) due to high protection are given an upper value of 8 kcal/mol. A representative decay profile is shown in Fig. 4 B, and the values of the residue-specific free energy of stabilization of the individual mutants are plotted in Fig. 4 C and tabulated in Table S1.

In the case of the E29P mutant, the overall decay profile and the slow exchange of the dimer interface residues indicate that it is a dimer. However, the free energy stabilities of some of the dimer interface β1-strand, and a large stretch of C-terminal helical residues (V58 to L66) are also lower (Fig. 4 C). The IL-8 dimer structure reveals that E29 is packed against A69′ and F65′ across the dimer interface, suggesting that the loss of these favorable intermolecular packing interactions results in local structural changes and increased solvent accessibility.

In the case of the L25P, V27P, V27P/E29P, and L25P/E29P mutants, only a few residues from the β1-, β2-, β3-strands and α-helix were observed, and none of the dimer interface residues could be observed confirming that these mutants are monomeric (Fig. 4 A and Fig. S4). Similarities between these spectra with those reported for the trapped L25NMe monomer also indicate that these mutants are monomers (22). Furthermore, the differences in the number of resonances and peak intensity decay in the first HSQC spectra also highlight the stability differences. The data indicate that V27P is more stable compared to the L25P monomer, and the V27P/E29P and L25P/E29P double mutants are more stable than the V27P and L25P mutants (Fig. 4 and Fig. S4), which are also consistent with the unfolding studies and the NMR structural data (Figs. 2 and 3, and Table 2). The higher stability of the double mutants can be directly attributed to the optimal packing of the β1- and β2-strand residues (Fig. 2 B).

Structures and dynamics of the Pro mutants

To assess the consequence of Pro substitutions on the structural integrity, we carried out a 500 ns MD simulations of the L25P, V27P, and V27P/E29P monomers and E29P dimer, using the Amber 12 Suite (35). In addition, we also carried out simulations for the L25P, V27P, and V27P/E29P mutants in the dimeric background to provide a structural rationale as to why these mutations could not be accommodated in the dimer and what could be the primary trigger for the disruption of the dimer interface.

MD simulations in the dimeric background

Analysis of the MD trajectory of the V27P mutant showed that the β1-strands forming the dimer interface show large rocking and sliding motions within 100 ns, resulting in a loss of interface H-bonds (Fig. 5, A–C). During the course of the simulation, only two out of the four dimer-interface H-bonds remain (Fig. 5, A–C). This is a consequence of two fewer intermolecular backbone H-bonds in the middle of the dimer interface due to the V27P mutation, and also could be due to the disruption of intermolecular packing interactions involving V27 with F65′ and A69′ in the C-terminal helix and realignment of the helix packing with the β-strands. In the case of V27P/E29P simulations, a similar but more facile disruption of the dimer interface was observed even during the initial stages of the MD simulation, which can be directly attributed to the absence of two additional H-bonds involving E29 at the interface. This MD exercise provided a clear structural and dynamics picture as to why the V27P and V27P/E29P mutants exist as monomers and not dimers.

Figure 5.

Figure 5

Open in a new tab

MD simulations of V27P and L25P mutants in the dimeric background. (A and B) Molecular plots of the V27P mutant showing disruption of the dimer interface over the course of the simulation through rocking and sliding motions due to loss of the central H-bonds and unfavorable intermolecular packing of P27 with the C-terminal helix. (C) Distance plots for the V27P mutant showing breaking of the dimer interface. Bars in orange and blue show the regions of the trajectory that represent the structures shown in panels A and B, respectively. (D) Overlay of molecular snapshots of the L25P mutant at 100 ns (gray) and 300 ns (green) of the MD trajectory showing disruption of the dimer interface due to flipping of the E24 side chain into dimer interface. (E) Distance plots for the L25P mutant showing breaking of the H-bonds of the dimer interface residues.

For the L25P dimer, MD simulations indicate that the effect of the mutation is more drastic and more pronounced at the dimer interface compared to V27P. The intermolecular H-bonds are lost even during the equilibration step of the MD simulation (Fig. 5, D and E). The altered ϕ angle for P25 and the unfavorable intramolecular packing of P25 with F65 and V62 leads to flipping of the E24 side chain into the dimer interface, resulting in steric hindrance causing disruption of dimer formation.

In the case of E29P mutant, the central four H-bonds at the dimer interface are intact and the structure retains the WT dimer fold (Fig. S5, A and B). Analysis of the E29(N)…K23′(O) H-bond in the WT simulation showed that it is transient (Fig. S5 C), suggesting that the H-bond involving E29 is not essential for retaining the dimer interface.

MD simulations in the monomer background

Throughout the MD simulation of the V27P mutant, the interstrand H-bond network between the β1- and β2-strands is retained. The altered backbone conformation at the site of Pro mutation (ϕ(Pro-27) = −80 ± 25°), which is slightly outside the allowed conformational range for β-strand (−125 ± 15°) only minimally perturbs the local structure. It is also interesting to note that H-bonding interactions of the neighboring R26 and I28 residues are still intact [I40(N)…R26(O) and E38(N)…I28(O)], except for some transient loss over the course of the simulation (Figs. 6, A and B, and Fig. S6). The proline most likely induces local strain, which results in increased dynamics as there are overall increased breathing motions at the β1-β2 strand interface, which are also consistent with the weaker NOEs and the lower measured stabilities from the NHX experiments. In addition, we observe local unwinding of the last 6 helical residues (K67–S72), which is also consistent with the NMR data (Fig. S3 E). The overlay of a snapshot of the V27P mutant and the trapped monomer structures shows only local differences around the site of mutation (Fig. 6 A). MD simulations data of the V27P/E29P mutant showed very similar features as observed for the V27P mutant (Fig. 6 C, Fig. S7, and Fig. S8). P29 ϕ angle (−70 ± 25°), as in the case of P27, falls slightly outside the preferred range, but nevertheless is well accommodated in the E29 position of β1-strand (Fig. S8). The intactness of the β-strands in this double mutant is also evident from the slow exchange rates of the R26, I28, E38, and I40 residues (Fig. 4 and Table S1). Though our 500 ns MD simulations studies are fairly long to understand structural dynamics, it is not sensitive enough to capture the subtle structural/dynamics changes responsible for the increased stability of V27P/E29P compared to V27P, and larger timescale sampling may be necessary to capture these stability differences.

Figure 6.

Figure 6

Open in a new tab

MD simulations of V27P, V27P/E29P, and L25P monomers. (A) Overlay of the β1- and β2-strand interface of the L25NMe monomer (gray) and a snapshot of V27P monomer (orange) from the MD simulations. Local perturbation introduced by the proline is minimal and localized around the site of mutation. (B) Distance plots showing the changes in the H-bond network at the β1-β2 strand interface of V27P monomer. Throughout the simulation, the H-bonds are intact except for transient loss of I40(N)…R26(O) and E38(N)…I28(O) hydrogen bonds adjacent to the proline mutation. (C) Overlay of snapshots of V27P (orange) and V27P/E29P (yellow) monomers from the MD simulations. (D) Overlay of the β1-β2 strand interface of a snapshot (100 ns) of L27P monomer (green) from the MD simulation with the L25NMe monomer (gray). Perturbation introduced by the P25 results in flipping of the E24 side chain into the dimer interface, resulting in partial opening of the β1-β2 strand interface. (E) Distance plots showing changes in the H-bond network at the β1-β2 strand interface in the L25P monomer MD simulations. Note the loss of K42(N)…K24(O) H-bond around 30 ns of the MD trajectory.

Analysis of the L25P monomer trajectory showed the flipping of the E24 side chain into the dimer interface as observed in the L25P dimer background, but the extent of flipping was more pronounced, providing the structural basis for the inability of L25P to form a stable dimer. Moreover, the flipping of E24 results in the loss of the K42(N)…E24(O) H-bond and local opening of the β1-β2 strand interface (Figs. 6, D and E). These structural details are in line with the observed lower free energy values for the adjacent β2-strand residues I40 and K42 (missing after the first few data points) in the NHX experiments (Fig. 4 C and Table S1).

Functional activities of the Pro mutants

The chemokine IL-8 level is upregulated in response to infection, and so its concentration could vary by orders of magnitude as a function of space and time. Therefore, IL-8 could exist as monomers, dimers, or both, and animal model studies have shown that monomer-dimer equilibrium regulates neutrophil recruitment (21,38). IL-8 function involves activation of the CXCR2 receptor, and also binding to cell surface negatively charged glycosaminoglycans (GAG) that regulate receptor functions. We characterized Ca2+ release activity in a cellular assay that captures receptor activation (Fig. S9 A), binding to heparin affinity column that captures GAG interactions (Fig. S9 B), and neutrophil recruitment in a mouse model that captures the sum of receptor and GAG interactions (Fig. 7). We have previously shown that at the 1 μg dose, WT recruits neutrophils like a monomer in the mouse peritoneum model (21), and observe that all of the mutants have WT-like activity. The WT functions like a monomer also in cellular assays, and we observe that all the Pro mutants have WT-like activity. In the GAG binding experiments, WT exists as a dimer, and the E29P mutant elutes like the WT from the heparin column, and all other mutants eluted with lower salt concentration similar to what was observed for the trapped L25NMe monomer indicating that the V27P, L25P, V27P/E29P, and L25P/E29P mutants function as monomers.

Figure 7.

Figure 7

Open in a new tab

Neutrophil recruitment in a mouse peritoneum model. The recruitment of the Pro mutants are scaled to WT. Data are presented as mean ± SE, represent 3 to 5 mice/group, and are representative of one of two to three independent experiments.

Discussion

In this current study, we have explored the strategy of Pro substitution for generating monomers by disrupting the H-bond network of dimer interface antiparallel β-strands. NMR structural characterization of the IL-8 mutants showed that prolines could be accommodated in the middle of the dimer interface β-strand with minimal perturbation resulting in stable functional monomers. The edge mutation E29P in the β1-strand only resulted in local perturbation of the dimer interface. MD simulation studies of the Pro mutants provided structural snapshots of how prolines can be accommodated in a β-strand, effect of mutations on the integrity of the β-sheet, changes in the flexibility and breathing motions of the β1- and β2-strand interfaces, and why some mutations resulted in stable monomers compared to others. MD studies of the monomeric mutants in a dimeric background provided a structural basis as to why these mutants cannot be accommodated in a dimer. The NHX data provided a residue-specific description of the cause and effect of the observed stability differences among Pro mutants. The NHX data also allow us to propose a model describing the sequential events in the folding of the IL-8 dimer: the hydrophobic collapse is initiated by the side-chain residues of β1-, β2-, β3-strands and α-helix, followed by the formation of the rest of the structural elements, and finally the monomers associate via β1-strand interactions to form a homodimer.

Interestingly, the structural, unfolding, and NHX data together show the double Pro mutants are more stable than the single Pro mutants. NHX data also indicated that the higher stability of the double mutants is due to the higher integrity of the β1-β2 inter-strand H-bonds. Furthermore, the greater stabilization of the second Pro substitution at E29 in L25P compared to V27P mutant, and because the L25P/V27P mutant was only marginally folded, indicates that the stabilities of the double Pro mutants are strictly context dependent. We propose that stabilization is most likely because the individual proline substitutions counteract and dampen the breathing motions at the β1- and β2-strand interface to maintain native fold and interactions. Future studies on other proteins are necessary to better understand how double Pro substitutions result in higher stability of the monomer.

Considering that the Pro backbone conformation is restricted, it has been argued that Pro substitution minimizes the loss of conformational entropy on folding, and hence could be used as a strategy to increase protein stability. Indeed, Pro mutations in a wide variety of proteins have shown that such a strategy can result in higher stability (48–55). The objective of our proline substitution study, in contrast, was to disrupt dimer interactions, and our data indicate that the inability to form H-bonds is the major factor driving monomer formation, and that other factors including conformational strain and packing could also play a role.

Our data show that the H-bonds in the middle, compared to the edge, of the dimer interface play a more prominent role in stabilizing the dimer. Kelly’s group has systematically studied the role of individual H-bonds, by substituting amide esters for amide bonds in chemically synthesized model proteins such as WW, and have shown that the energetics of the individual H-bonds can vary significantly and that the local structure and hydrophobicity play important roles in determining the H-bond energetics in the folded protein (56,57). Our study cannot provide such insights of the individual H-bonds as Pro substitutions also influence packing interactions. However, NHX data of WT indicate that the V27 H-bond, compared to L25 H-bond, is more stable by as much as 1 kcal/mol (Table S1) suggesting that the NHX data and the preference for middle over edge residues could be used to rank order the potential dimer interface residues for the Pro substitution strategy.

Conclusions

In summary, our studies have shown that a single Pro mutation at the middle of the dimer interface β-strand is sufficient to generate a fully folded and functional monomer, and that double Pro substitutions can actually result in higher stability of the monomer without compromising native monomer fold or function. Considering dimers and higher-order oligomers are highly pervasive in nature and is involved in a wide spectrum of functions, we propose Pro substitution as a viable strategy for generating functional monomers of dimers, and potentially higher-order oligomers, stabilized by β-strand interactions. Our current systematic approach has also provided useful guidelines for identifying residues and their locations in the dimer interface for the efficient disruption of dimers resulting in stable and functional monomers.

Acknowledgments

We thank Ms. Meena Shanmugasundaram and Dr. Tianzhi Wang (UTMB) for technical assistance. We are also thankful for the high performance computing resource provided by the Texas Advanced Computing Center (TACC) at UT Austin.

This work was supported by grants PO1HL107152 and R01AI069152 from the National Institutes of Health.

Supporting Material

Document S1. One table and nine figures
mmc1.pdf (2.4MB, pdf)
Document S2. Article plus Supporting Material
mmc2.pdf (4.2MB, pdf)

References

  • 1.Matthews J.M., Sunde M. Dimers, oligomers, everywhere. Adv. Exp. Med. Biol. 2012;747:1–18. doi: 10.1007/978-1-4614-3229-6_1. [DOI] [PubMed] [Google Scholar]
  • 2.Goodsell D.S., Olson A.J. Structural symmetry and protein function. Annu. Rev. Biophys. Biomol. Struct. 2000;29:105–153. doi: 10.1146/annurev.biophys.29.1.105. [DOI] [PubMed] [Google Scholar]
  • 3.Nooren I.M., Thornton J.M. Diversity of protein-protein interactions. EMBO J. 2003;22:3486–3492. doi: 10.1093/emboj/cdg359. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Veldkamp C.T., Peterson F.C., Volkman B.F. The monomer-dimer equilibrium of stromal cell-derived factor-1 (CXCL 12) is altered by pH, phosphate, sulfate, and heparin. Protein Sci. 2005;14:1071–1081. doi: 10.1110/ps.041219505. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Kang X., Zhong N., Xia B. Foldon unfolding mediates the interconversion between M(pro)-C monomer and 3D domain-swapped dimer. Proc. Natl. Acad. Sci. USA. 2012;109:14900–14905. doi: 10.1073/pnas.1205241109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Mishra V., Kumar A., Bhakuni V. Glu-108 is essential for subunit assembly and dimer stability of D-phosphoglycerate dehydrogenase from Entamoeba histolytica. Mol. Biochem. Parasitol. 2012;181:117–124. doi: 10.1016/j.molbiopara.2011.10.008. [DOI] [PubMed] [Google Scholar]
  • 7.Hor L., Dobson R.C., Perugini M.A. Dimerization of bacterial diaminopimelate epimerase is essential for catalysis. J. Biol. Chem. 2013;288:9238–9248. doi: 10.1074/jbc.M113.450148. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Tong Y., Hughes D., Buck M. When monomers are preferred: a strategy for the identification and disruption of weakly oligomerized proteins. Structure. 2005;13:7–15. doi: 10.1016/j.str.2004.10.018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.MacArthur M.W., Thornton J.M. Influence of proline residues on protein conformation. J. Mol. Biol. 1991;218:397–412. doi: 10.1016/0022-2836(91)90721-h. [DOI] [PubMed] [Google Scholar]
  • 10.Wilmot C.M., Thornton J.M. Analysis and prediction of the different types of beta-turn in proteins. J. Mol. Biol. 1988;203:221–232. doi: 10.1016/0022-2836(88)90103-9. [DOI] [PubMed] [Google Scholar]
  • 11.Brandl C.J., Deber C.M. Hypothesis about the function of membrane-buried proline residues in transport proteins. Proc. Natl. Acad. Sci. USA. 1986;83:917–921. doi: 10.1073/pnas.83.4.917. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Yutani K., Hayashi S., Ogasahara K. Role of conserved proline residues in stabilizing tryptophan synthase alpha subunit: analysis by mutants with alanine or glycine. Proteins. 1991;9:90–98. doi: 10.1002/prot.340090203. [DOI] [PubMed] [Google Scholar]
  • 13.Sarkar P., Reichman C., Kalodimos C.G. Proline cis-trans isomerization controls autoinhibition of a signaling protein. Mol. Cell. 2007;25:413–426. doi: 10.1016/j.molcel.2007.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Barlow D.J., Thornton J.M. Helix geometry in proteins. J. Mol. Biol. 1988;201:601–619. doi: 10.1016/0022-2836(88)90641-9. [DOI] [PubMed] [Google Scholar]
  • 15.Bhattacharyya R., Chakrabarti P. Stereospecific interactions of proline residues in protein structures and complexes. J. Mol. Biol. 2003;331:925–940. doi: 10.1016/s0022-2836(03)00759-9. [DOI] [PubMed] [Google Scholar]
  • 16.Lowman H.B., Fairbrother W.J., Hébert C.A. Monomeric variants of IL-8: effects of side chain substitutions and solution conditions upon dimer formation. Protein Sci. 1997;6:598–608. doi: 10.1002/pro.5560060309. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Kagiampakis I., Jin H., Tsai J. Conservation of unfavorable sequence motifs that contribute to the chemokine quaternary state. Biochemistry. 2008;47:10637–10648. doi: 10.1021/bi702288a. [DOI] [PubMed] [Google Scholar]
  • 18.Clore G.M., Appella E., Gronenborn A.M. Three-dimensional structure of interleukin 8 in solution. Biochemistry. 1990;29:1689–1696. doi: 10.1021/bi00459a004. [DOI] [PubMed] [Google Scholar]
  • 19.Baldwin E.T., Weber I.T., Wlodawer A. Crystal structure of interleukin 8: symbiosis of NMR and crystallography. Proc. Natl. Acad. Sci. USA. 1991;88:502–506. doi: 10.1073/pnas.88.2.502. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Rajarathnam K., Clark-Lewis I., Sykes B.D. 1H NMR solution structure of an active monomeric interleukin-8. Biochemistry. 1995;34:12983–12990. doi: 10.1021/bi00040a008. [DOI] [PubMed] [Google Scholar]
  • 21.Das S.T., Rajagopalan L., Rajarathnam K. Monomeric and dimeric CXCL8 are both essential for in vivo neutrophil recruitment. PLoS ONE. 2010;5:e11754. doi: 10.1371/journal.pone.0011754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Rajarathnam K., Sykes B.D., Clark-Lewis I. Neutrophil activation by monomeric interleukin-8. Science. 1994;264:90–92. doi: 10.1126/science.8140420. [DOI] [PubMed] [Google Scholar]
  • 23.Horcher M., Rot A., Besemer J. IL-8 derivatives with a reduced potential to form homodimers are fully active in vitro and in vivo. Cytokine. 1998;10:1–12. doi: 10.1006/cyto.1997.0251. [DOI] [PubMed] [Google Scholar]
  • 24.Ho S.N., Hunt H.D., Pease L.R. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene. 1989;77:51–59. doi: 10.1016/0378-1119(89)90358-2. [DOI] [PubMed] [Google Scholar]
  • 25.Rajagopalan L., Rajarathnam K. Ligand selectivity and affinity of chemokine receptor CXCR1. Role of N-terminal domain. J. Biol. Chem. 2004;279:30000–30008. doi: 10.1074/jbc.M313883200. [DOI] [PubMed] [Google Scholar]
  • 26.Permi P., Annila A. Coherence transfer in proteins. Prog. Nucl. Magn. Reson. Spectrosc. 2004;44:97–137. [Google Scholar]
  • 27.Delaglio F., Grzesiek S., Bax A. NMRPipe: a multidimensional spectral processing system based on UNIX pipes. J. Biomol. NMR. 1995;6:277–293. doi: 10.1007/BF00197809. [DOI] [PubMed] [Google Scholar]
  • 28.Johnson B.A., Blevins R.A. NMR View: a computer program for the visualization and analysis of NMR data. J. Biomol. NMR. 1994;4:603–614. doi: 10.1007/BF00404272. [DOI] [PubMed] [Google Scholar]
  • 29.Wu D.H., Chen A.D., Johnson C.S. An improved diffusion-ordered spectroscopy experiment incorporating bipolar-gradient pulses. J. Magn. Reson. A. 1995;115:260–264. [Google Scholar]
  • 30.Uedaira H., Uedaira H. Transitional frictional coefficients of molecules in aqueous solution. J. Phys. Chem. 1970;74:2211–2214. [Google Scholar]
  • 31.Mohan P.M., Chakraborty S., Hosur R.V. Residue-wise conformational stability of DLC8 dimer from native-state hydrogen exchange. Proteins. 2009;75:40–52. doi: 10.1002/prot.22219. [DOI] [PubMed] [Google Scholar]
  • 32.Santoro M.M., Bolen D.W. Unfolding free energy changes determined by the linear extrapolation method. 1. Unfolding of phenylmethanesulfonyl alpha-chymotrypsin using different denaturants. Biochemistry. 1988;27:8063–8068. doi: 10.1021/bi00421a014. [DOI] [PubMed] [Google Scholar]
  • 33.Mohan P.M., Chakraborty S., Hosur R.V. Hierarchy of local structural and dynamics perturbations due to subdenaturing urea in the native state ensemble of DLC8 dimer. Biophys. Chem. 2010;153:17–26. doi: 10.1016/j.bpc.2010.09.010. [DOI] [PubMed] [Google Scholar]
  • 34.DeLano W.L. DeLano Scientific LLC; Palo Alto, CA: 2002. The Pymol Users Manual. [Google Scholar]
  • 35.Case D.A., Darden T.A., Kollman P.A. University of California; San Francisco: 2012. AMBER 12. [Google Scholar]
  • 36.Lee M.C., Deng J., Duan Y. Large-scale conformational dynamics of the HIV-1 integrase core domain and its catalytic loop mutants. Biophys. J. 2005;88:3133–3146. doi: 10.1529/biophysj.104.058446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37.Humphrey W., Dalke A., Schulten K. VMD: visual molecular dynamics. J. Mol. Graph. 1996;14:33–38. doi: 10.1016/0263-7855(96)00018-5. 27–38. [DOI] [PubMed] [Google Scholar]
  • 38.Nasser M.W., Marjoram R.J., Richardson R.M. Cross-desensitization among CXCR1, CXCR2, and CCR5: role of protein kinase C-epsilon. J. Immunol. 2005;174:6927–6933. doi: 10.4049/jimmunol.174.11.6927. [DOI] [PubMed] [Google Scholar]
  • 39.Gangavarapu P., Rajagopalan L., Rajarathnam K. The monomer-dimer equilibrium and glycosaminoglycan interactions of chemokine CXCL8 regulate tissue-specific neutrophil recruitment. J. Leukoc. Biol. 2012;91:259–265. doi: 10.1189/jlb.0511239. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Mikhailov D., Young H.C., Mayo K.H. Heparin dodecasaccharide binding to platelet factor-4 and growth-related protein-alpha. Induction of a partially folded state and implications for heparin-induced thrombocytopenia. J. Biol. Chem. 1999;274:25317–25329. doi: 10.1074/jbc.274.36.25317. [DOI] [PubMed] [Google Scholar]
  • 41.Wishart D.S., Sykes B.D. Chemical shifts as a tool for structure determination. Methods Enzymol. 1994;239:363–392. doi: 10.1016/s0076-6879(94)39014-2. [DOI] [PubMed] [Google Scholar]
  • 42.Shen Y., Delaglio F., Bax A. TALOS+: a hybrid method for predicting protein backbone torsion angles from NMR chemical shifts. J. Biomol. NMR. 2009;44:213–223. doi: 10.1007/s10858-009-9333-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43.Rajagopalan L., Chin C.C., Rajarathnam K. Role of intramolecular disulfides in stability and structure of a noncovalent homodimer. Biophys. J. 2007;93:2129–2134. doi: 10.1529/biophysj.107.108761. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Chu R., Pei W., Bai Y. Relationship between the native-state hydrogen exchange and folding pathways of a four-helix bundle protein. Biochemistry. 2002;41:7998–8003. doi: 10.1021/bi025872n. [DOI] [PubMed] [Google Scholar]
  • 45.Maity H., Maity M., Englander S.W. Protein folding: the stepwise assembly of foldon units. Proc. Natl. Acad. Sci. USA. 2005;102:4741–4746. doi: 10.1073/pnas.0501043102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 46.Chamberlain A.K., Handel T.M., Marqusee S. Detection of rare partially folded molecules in equilibrium with the native conformation of RNaseH. Nat. Struct. Biol. 1996;3:782–787. doi: 10.1038/nsb0996-782. [DOI] [PubMed] [Google Scholar]
  • 47.Li R., Woodward C. The hydrogen exchange core and protein folding. Protein Sci. 1999;8:1571–1590. doi: 10.1110/ps.8.8.1571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Bajaj K., Madhusudhan M.S., Varadarajan R. Stereochemical criteria for prediction of the effects of proline mutations on protein stability. PLOS Comput. Biol. 2007;3:e241. doi: 10.1371/journal.pcbi.0030241. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Prajapati R.S., Das M., Varadarajan R. Thermodynamic effects of proline introduction on protein stability. Proteins. 2007;66:480–491. doi: 10.1002/prot.21215. [DOI] [PubMed] [Google Scholar]
  • 50.Herning T., Yutani K., Kikuchi M. Role of proline residues in human lysozyme stability: a scanning calorimetric study combined with X-ray structure analysis of proline mutants. Biochemistry. 1992;31:7077–7085. doi: 10.1021/bi00146a008. [DOI] [PubMed] [Google Scholar]
  • 51.Nakamura S., Tanaka T., Nakai S. Improving the thermostability of Bacillus stearothermophilus neutral protease by introducing proline into the active site helix. Protein Eng. 1997;10:1263–1269. doi: 10.1093/protein/10.11.1263. [DOI] [PubMed] [Google Scholar]
  • 52.Matthews B.W., Nicholson H., Becktel W.J. Enhanced protein thermostability from site-directed mutations that decrease the entropy of unfolding. Proc. Natl. Acad. Sci. USA. 1987;84:6663–6667. doi: 10.1073/pnas.84.19.6663. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Nicholson H., Tronrud D.E., Matthews B.W. Analysis of the effectiveness of proline substitutions and glycine replacements in increasing the stability of phage T4 lysozyme. Biopolymers. 1992;32:1431–1441. doi: 10.1002/bip.360321103. [DOI] [PubMed] [Google Scholar]
  • 54.Watanabe K., Masuda T., Suzuki Y. Multiple proline substitutions cumulatively thermostabilize Bacillus cereus ATCC7064 oligo-1,6-glucosidase. irrefragable proof supporting the proline rule. Eur. J. Biochem. 1994;226:277–283. doi: 10.1111/j.1432-1033.1994.tb20051.x. [DOI] [PubMed] [Google Scholar]
  • 55.Choi E.J., Mayo S.L. Generation and analysis of proline mutants in protein G. Protein Eng. Des. Sel. 2006;19:285–289. doi: 10.1093/protein/gzl007. [DOI] [PubMed] [Google Scholar]
  • 56.Deechongkit S., Nguyen H., Kelly J.W. Context-dependent contributions of backbone hydrogen bonding to beta-sheet folding energetics. Nature. 2004;430:101–105. doi: 10.1038/nature02611. [DOI] [PubMed] [Google Scholar]
  • 57.Gao J., Bosco D.A., Kelly J.W. Localized thermodynamic coupling between hydrogen bonding and microenvironment polarity substantially stabilizes proteins. Nat. Struct. Mol. Biol. 2009;16:684–690. doi: 10.1038/nsmb.1610. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Document S1. One table and nine figures
mmc1.pdf (2.4MB, pdf)
Document S2. Article plus Supporting Material
mmc2.pdf (4.2MB, pdf)

Articles from Biophysical Journal are provided here courtesy of The Biophysical Society

RESOURCES