Response to chemical induced changes and their implication in yfdX proteins

Abstract

yfdX proteins occur in a large number of virulent bacteria. Recently we have shown that STY3178, a yfdX protein from Salmonella Typhi, exists in a trimeric state in solution which is capable of interacting with antibiotics, stable at elevated temperatures and undergoes reversible thermal unfolding. In this present study, we report the chemical response of STY3178. We monitor the stability of the protein in presence of chaotropes. It can regain the native-like structure from the chaotrope induced unfolded states. The structural stability of this protein is further studied in a wide pH range which reveals that the STY3178 trimer is stable in both acidic as well as basic media. We further show that the protein interacts with oxalate in vitro. Finally, we perform computational studies viz. modeling and molecular dynamics simulation to understand the stability of trimeric STY3178 over its monomer conformation. The conformational thermodynamic changes indicate that oligomerization induces stability via salt bridge interactions, present at the monomer interface.

---

Introduction

Salmonella Typhi (S. Typhi) is a pathogenic gram negative bacterium which uses the human body as a host and causes typhoid fever.^1–4 There are several strains of this bacterium with various degree of virulence, for instance, many which are resistant to antibiotics.^5–7 These strains of S. Typhi contain many uncharacterized proteins. One such protein is yfdX, first reported^8,9 in E. coli where the multidrug and acid response regulator EvgA regulates its expression. There are four other genes yfdU, yfdW, yfdE and yfdV in E. coli in the operon, which get activated by EvgA and are related to survival of cells in acidic environments.^10,11 yfdU is reported¹¹ as an oxalyl-CoA decarboxylase. Its deletion makes the cells vulnerable in presence of oxalate and at low pH, affecting the cell survival.¹¹ yfdW is a formyl-coenzyme A (CoA) transeferase and its deletion also reduces cell survival.¹¹ yfdE is an acetyl CoA : oxalate CoA-trasnferase.¹² All these proteins are involved in oxalate-induced acid tolerance response (ATR).^11,12 yfdV is also suggested to be involved in oxalate-induced ATR in E. coli.^11,12 However, no detailed report on yfdX protein in this direction is available. In our earlier studies,¹³ we have biophysically characterized the yfdX protein, STY3178 from S. Typhi.^6,7 This protein is in a trimeric state of oligomerization in solution.¹³ We have identified its antibiotic binding capability.¹³ We have further shown that STY3178 is quite stable at elevated temperature and undergoes reversible thermal unfolding.¹⁴ The stability of this protein in various chemical environments has not been probed.

With this backdrop we study the chemical response of STY3178. In particular, we check the chaotrope induced unfolding–refolding of the protein. Another important aspect of this protein is determining the stability in acidic or basic pH media, since neighboring genes of yfdX are all involved in cell survival in acidic environment. We show STY3178 gets stabilized initially in presence of urea compared to the native state. At higher concentration of chaotrope, it unfolds completely which upon reducing urea results in refolding of protein. In presence of guanidine hydrochloride (GndHCl), unfolding occurs gradually without any such initial stabilization of the protein. We find that STY3178 is stable in a wide range of pH starting from moderately acidic to strong alkaline medium. We further show that this protein is capable of interacting with oxalate. To understand the stability of STY3178 in solution, we propose a homology model for the trimer using our earlier reported monomer model.¹⁴ We estimate the relative conformational stability of STY3178 trimer with respect to its monomer by computing conformational thermodynamic changes using the molecular dynamics simulation trajectories. Conformational thermodynamics changes indicate oligomerization induces gain in stability and ordering for each monomer unit, where the interfacial residues contribute predominantly. These are acid–base residue pairs forming salt bridges between two monomer subunits. Our proposed trimer model could explain the experimental observations.

Methods

STY3178 protein has been overexpressed in E. coli and purified as reported in our earlier study.¹³

Sample preparation

Urea and guanidine hydrochloride (GndHCl) experiments. Unfolding experiments are carried out using 5 μM protein. The buffer condition used is 50 mM potassium phosphate (pH 7), 250 mM sodium chloride (NaCl) and 1 mM phenylmethanesulfonyl fluoride (PMSF) having urea or GndHCl. Different concentrations of urea or GndHCl starting from 0.2 M to 6 M are used for unfolding experiments. Protein in buffer is mixed with chaotropes and incubated over night at room temperature for equilibration.

For refolding experiments, 200 μM protein is first unfolded in 8 M of either urea or GndHCl and kept overnight for equilibration at room temperature. This unfolded protein is then diluted in different buffers to obtain the desired extent of dilution for urea or GndHCl, keeping protein concentration fixed at 5 μM. These protein samples with lower chaotrope concentrations are equilibrated overnight.

For instantaneous dilution experiments, the denatured protein (200 μM STY3178 in 8 M urea or GndHCl solution) is suitably diluted with a buffer containing no chaotrope resulting in 40, 30, 20 and 10 folds lower concentration of urea or GndHCl.

pH dependent measurements. We study the effect of pH ranging from 2.5 to 10 on STY3178 structure. We use the following buffers with 50 mM concentration for achieving the desired pH values: glycine–HCl (pH 2.5 and 3), sodium acetate–acetic acid (pH 4, 4.5 and 5.2), potassium phosphate (pH 6, 6.5 and 7), Tris–HCl (pH 7.5, 8 and 8.5) and glycine–NaOH (pH 9 and 10). STY3178 with 10 μM concentration is equilibrated over night in each of the aforementioned buffers prior to measurement. Samples for background correction are prepared similarly without adding protein.

Circular dichroism (CD)

All CD measurements are performed in Jasco J-815 CD spectrometer with 1 mm path length cuvette for measurements in far UV (200–250 nm) region and 10 mm cell for near UV-CD region (250–330 nm). All spectra are collected at room temperature with a scan speed of 100 nm min⁻¹. Background corrections are done with respective urea or GndHCl buffers and pH buffers. Unfolding and refolding data are collected as an average of minimum two scans.

The ellipticity values at 222 nm (θ₂₂₂) in presence of different chaotropes are used to estimate the fraction of folded protein (α_i) at a certain chaotrope concentration according to the following relation, α_i = (θ_i − θ_U)/(θ_N − θ_U), where θ_i is the ellipticity value (θ₂₂₂) of protein at different chaotrope concentrations. θ_N and θ_U represent the ellipticity value (θ₂₂₂) for the native and unfolded protein, respectively. Folding constant (K_F) is calculated from α_i using the following method¹⁵ described by Greenfield-

K_F = P_tα_i/(nP_t(1 − α_i))ⁿ,

where P_t is the concentration of protein expressed in moles per liter, n is number of subunits of protein. The value of K_F is then used to estimate the free energy associated with protein folding using the relation, ΔG_F = −RT ln K_F, where R and T denote the universal gas constant (1.98 kcal mol⁻¹) and absolute temperature (Kelvin), respectively.

Assuming the unfolding to be a two-state process and linear dependence of free energy of folding (ΔG) over denaturant concentration (D), the free energy (ΔG₀) in absence of denaturant can be estimated from the fit of ΔG as a function of D according to the relation, ΔG = ΔG₀ + m[D], where m is the slope.¹⁵ The error in the intercept of the fitted curve is taken to be measure of the uncertainty of the estimated ΔG₀.

In another experiment, near UV-CD (250–350 nm) measurements of 30 μM STY3178 at pH 7 (50 mM potassium phosphate, 250 mM NaCl and 1 mM PMSF) in presence of 100 μM rifampin (Rfp) are performed with and without urea using a 10 mm cell. Urea concentrations used in this experiment are 0.6 M, 1 M and 2 M. The reported data is averaged over two scans.

Intrinsic fluorescence measurements

Steady state fluorescence experiments are recorded using Jobin Yvon Horiba Fluorolog with a protein concentration of 10 μM. For both unfolding and refolding, protein in a buffer, having various concentrations of urea or GndHCl is excited at 280 nm. The excitation and emission slit widths are kept at 3 nm for urea and GndHCl related measurements. Protein incubated in buffers of different pH is excited at 280 excitation wavelength using 10 mm path length quartz cell and 2 nm slit width. In a different set of measurement, the emission of protein in presence of Rfp (20 μM) is monitored for 280 nm excitation, keeping the slit width 2 nm where the protein is pre-equilibrated in following urea concentrations 0.6 M, 1 M and 2 M. Fluorescence emission in presence of potassium oxalate (10 μM, 50 μM, 100 μM, 200 μM, 300 μM and 400 μM) is measured for 280 nm excitation with 3 nm slit width. The interaction of STY3178 with potassium oxalate is quantified using modified Stern–Volmer equation as reported earlier.¹³ All samples are incubated overnight prior to the measurements. Buffers having respective pH and chaotrope concentration are used for background correction.

Nuclear magnetic resonance spectroscopy

The uniformly ¹⁵N-labeled protein is prepared using M9 minimal media containing ¹⁵NH₄Cl as the only source of nitrogen along with other supplements as mentioned earlier.¹³ The ¹H–¹⁵N heteronuclear single quantum coherence (HSQC) experiment is performed at 35 °C in 600 MHz Varian spectrometer equipped with room temperature probe. The concentration of ¹⁵N-labeled protein used is 400 μM in 30 mM phosphate buffer (pH 7), 150 mM NaCl and 10% D₂O. All HSQC spectra are acquired as 64 scans. HSQC without and with urea are recorded for 0 M, 0.5 M, 1 M, 2 M and 3 M urea keeping the protein concentration fixed. The 3 M urea containing denatured protein sample is refolded by buffer exchange using a spin concentrator (10KDa cut off, Amicon). Finally refolded protein is mixed with 10% D₂O and HSQC is performed. All data are processed in NMRPIPE¹⁶ and analyzed in NMRVIEW.¹⁷

Dynamic light scattering

Effect of urea, GndHCl and pH on the hydrodynamic size of protein is monitored using dynamic light scattering in a Nano-S Malvern instrument at 20 °C. The wavelength of laser and angle of scattering used for all measurements are 632.8 nm and 173°, respectively. Protein concentration of 10 μM is used. All samples are equilibrated over night and measurements are performed using a 10 mm path length cell. The measurement for each sample is collected as an average of five scans.

Computational modeling and simulation

A trimer arrangement for the protein is modeled using ZDOCK server.¹⁸ The template used for trimer assembly is 100 ns simulated monomer structure¹⁴ reported in our earlier work. We further optimize the first five best score assembly structures. The energetically most favorable trimer conformation is then simulated for 400 ns.

Simulation details

The all-atom molecular dynamics (MD) simulations for the two systems, namely STY3178 monomer¹⁴ and STY3178 trimer are carried out using the NAMD¹⁹ program in presence of explicit TIP3P water molecules and counter ions keeping the total number of atoms (N = 61 860) same for each of the cases. Simulations are performed considering the CHARMM27 ²⁰ force field parameters at temperature of 310 K and 1 atm pressure in isothermal–isobaric (NPT) ensemble, following standard protocol^21,22 with 1 femto second (fs) time step. We consider the equilibrated trajectories of monomer and trimer between 100–400 ns for thermodynamics calculations.

We calculate the radius of gyration (R_g) of simulated STY3178 trimeric assembly as the square of distance of C_α-atoms from the centre of mass (R⃑_CM), . The respective mass and position vectors of the i^th C_α-atom are represented by m_i and r⃑_i, respectively. The expression²¹ used for R_g calculation is, . The R_g is calculated over the equilibrated trajectory. We then estimated the hydrodynamic radius (R_H) of the trimer model using the R_g value in equation²³, R_g/R_H = 0.77.

Conformational thermodynamics changes

The conformational thermodynamics changes in a given state with respect to another is extracted using the histogram based method (HBM), detailed description of which can be found in a recent study.²² These two different states represent the monomer and trimer conformations of STY3178. We compute all the dihedral angles of protein across the equilibrated trajectory and generate probability distributions of individual dihedral angles in a given state. The normalized probability distribution of dihedral angle θ is given by the histogram P^X_i(θ) where X_i represents a particular conformational state. The peaks of P^X_i(θ) represent the equilibrium values of θ in that conformational state. The equilibrium free energy change of a protein dihedral angle, θ associated with change from one state, X₁ to another, X₂ is defined as

ΔG^conf(θ) = −k_BT ln[P_max^X₂(θ)/P_max^X₁(θ)], (1)

where ‘max’ represents the peak value. The conformational entropy for a dihedral in a particular state X_i is given by the Gibbs entropy formula,where j indicates histogram bins. Therefore, the entropy change for the dihedral between two conformational states is given by the difference,

In the present study X₁ refers to monomer, while X₂ denotes the trimer. The conformational thermodynamics changes of dihedral angles are additive owing to their independent distributions.

Results

Effect of chaotropes

The ellipticity at 222 nm (θ₂₂₂) is measured using circular dichroism and plotted against increasing urea concentration (Fig. 1A). We find a decrease in ellipticity initially up to 0.6 M urea (Fig. 1A) reaching a minimum. Similar behaviour is reported earlier for proteins such as CII²⁴ from bacteriophage and bovine serum albumin.²⁵ Above 0.6 M urea, θ₂₂₂ increases and reaches a native-like ellipticity value at 1.6 M urea. Beyond 1.6 M urea concentration, the ellipticity increases implying unfolding of the protein, until a constant value is attained in the range of 2.5 M to 6 M urea. In near UV-CD of STY3178, the characteristic broad shoulder around 275 nm is retained up to 1 M urea (Fig. 1B). In presence of 2 M urea, this broad shoulder disappears, indicating rearrangement of aromatic residues. Steady state fluorescence emission spectra show that emission peak position (∼342 nm) starts shifting towards red above 1 M urea (Fig. 1C). Nearly 11 nm red shift of the emission peak is observed for urea concentrations 2 M and above, compared to the native protein (Fig. 1C). The dynamic light scattering measurements of protein show that hydrodynamic size (∼6.5 nm) remains unchanged up to 2 M urea. It increases to 7.5 nm as urea concentration increases to 4 M (Fig. 1D). Aggregation size increases further to about 12 nm upon increasing urea concentration above 4 M urea (Fig. 1D).

Fig. 1 Effect of urea on protein structure. (A) Shows the ellipticity change at 222 nm (θ₂₂₂) in presence of urea during unfolding (triangle) and refolding (circle). (B) Near UV-CD (250–350 nm) spectral changes of STY3178 in 0.4 M (medium dash), 0.6 M (dash–dot–dot), 1 M (small dash), 2 M (dotted) and 6 M (dash–dot) urea concentration. (C) Shows the plot of fluorescence emission peak position in presence of (0–6) M urea concentration upon unfolding (triangle) and refolding (circle) for excitation wavelength 280 nm. (D) Shows the hydrodynamic size of STY3178 with increasing urea concentration during unfolding (triangle) and refolding (circle). (E) Shows the CD spectra upon instantaneous urea dilution of unfolded protein by 40 (solid), 30 (long-dash), 20 (dotted) and 10 (short-dash) folds. (F) Represents the near UV-CD spectra of protein upon decreasing urea concentration starting from 8 M to 6 M (dash–dot), 2 M (dotted), 1 M (short dash), 0.6 M (dash–dot–dot) and 0.4 M (long dash) during refolding. The native spectrum is shown in solid for (B), (E) and (F).

We study refolding capability of STY3178 by diluting 8 M urea unfolded protein and monitoring far UV-CD spectra. We achieve refolding of protein upon 40 (0.2 M), 30 (0.27 M), 20 (0.4 M) and 10 (0.8 M) folds instantaneous dilution of urea (Fig. 1E). The ellipticity values of the refolded proteins are slightly less compared to the native value which can be due to inappropriate equilibration upon instantaneous dilution (approximately 2 minutes lag between dilution and measurement). In another set of experiment, we equilibrate the unfolded protein in a refolding buffer and record its CD. Protein starts regaining the native-like conformation, as urea concentration in the solution reaches below 2.5 M (Fig. 1A). The unfolding and refolding show very similar ellipticity values at a given concentration of urea, except for the region between 0.2 and 1.2 M urea (Fig. 1A). The pronounced minimum in this region during unfolding is replaced by a shallow and broad minimum during refolding. The signature of refolding is also captured from near UV-CD spectra upon dilution of urea to 1 M concentration (Fig. 1F). Similarly, a blue shift in emission peak position is observed in fluorescence emission upon dilution of urea, compared to the unfolded protein (Fig. 1C). Below 2 M urea concentration, refolding starts as indicated from the blue shift. We observe hydrodynamic size of very high order (∼400 nm) upon reducing urea concentration to 4 M from 8 M concentration. Upon further dilution of urea to 3 M and 2 M concentration, nearly 8 nm size is observed (Fig. 1D). Further reducing urea concentration to 1 M and below, a hydrodynamic size similar to native protein is observed (Fig. 1D).

Native protein HSQC spectrum shows well dispersed peaks in the amide region. Along with these peaks, there are large number of peaks in the HSQC spectrum between 7.5 and 8.5 ppm (Fig. 2A) which indicate presence of helices in the folded protein. HSQC in presence of 0.5 M urea (Fig. 2B) is very similar to native protein spectrum with hardly any noticeable change. In 1 M urea, few of the peaks show change in chemical shift compared to native HSQC (marked by arrows in Fig. 2C). Around 25% of well dispersed peaks in HSQC show shift in presence of 2 M urea (Fig. 2D). Similar to CD and fluorescence results, HSQC in presence of 3 M urea results in a complete collapsed spectrum with no peak dispersion indicating protein unfolding (Fig. 2E). Overall we find that STY3178 is stable in 1 M and has partial stability in 2 M urea concentration beyond which it loses the structure. HSQC spectrum of protein upon lowering urea concentration by buffer exchange shows native-like dispersion of peaks (Fig. 2F). This again clearly demonstrates a complete refolding of STY3178 in absence of urea starting from unfolded protein at 3 M urea.

Fig. 2 ¹H–¹⁵N HSQC spectra of STY3178. (A) HSQC spectrum (6–10.5 ppm) of singly labeled native protein is shown in red. (B), (C), (D) and (E) represent the HSQC spectra of STY3178 in presence of 0.5 M urea (black), 1 M (green), 2 M (magenta) and 3 M (blue) urea, respectively. (F) Represents the HSQC of refolded protein upon reducing urea concentration from 3 M to 0.013 M and shown in dark green. Peaks showing slight shift in presence of urea are marked with black arrows in panels (C) and (D). All spectra are acquired at 35 °C with 64 scans.

The unfolding in presence of GndHCl shows a slightly different trend. GndHCl being a more potent chaotrope can affect the secondary structure of STY3178 at lower concentration of 0.6 M as observed in far UV-CD (Fig. 3A). In presence of 0.8 M, complete unfolding of the secondary structures of protein is observed. The unfolding data do not show initial stabilization unlike the urea case. The initial spectral change occur at 0.6 M GndHCl followed by complete collapse of tertiary structure at 0.8 M GndHCl and above, where we neither find native-like shoulder in near UV-CD (Fig. 3B) nor native-like fluorescence maxima (Fig. 3C). A shift of nearly 12 nm in emission peak position is observed above 0.8 M GndHCl. The hydrodynamic diameter with increasing GndHCl concentration is shown in Fig. 3D (triangle). Native hydrodynamic diameter of 6.5 nm is retained up to 1 M GndHCl. Protein hydrodynamic size increases by 1 nm in 1 M to 3 M GndHCl concentration range. Above 3 M GndHCl, aggregation size further increases to about 12 nm. The increase in hydrodynamic size of protein indicates formation of higher order aggregates upon unfolding similar to that observed for urea unfolding.

Fig. 3 Secondary and tertiary structure of STY3178 in presence of GndHCl. (A) Shows change in ellipticity at 222 nm (θ₂₂₂) as a function of GndHCl concentration (0–3 M) during unfolding (triangle) and refolding (circle) of protein. (B) Near UV-CD (250–350 nm) spectra of protein in 0.4 M (medium dash), 0.6 M (dash–dot–dot), 1 M (small dash), 2 M (dotted) and 6 M (dash–dot) concentration of GndHCl are shown. (C) Represents the plot of fluorescence emission peak position versus GndHCl concentration (0–6 M) during unfolding (triangle) and refolding (circle). (D) The hydrodynamic diameter of protein plotted as a function of GndHCl concentration (0–6 M) during unfolding (triangle) and refolding (circle). (E) Shows the CD spectra of protein upon instantaneous dilution of GndHCl by 40 (solid), 30 (long-dash), 20 (dotted) and 10 (short-dash) folds starting from 8 M concentration. (F) Represents the near UV-CD spectra of protein upon reducing GndHCl concentration from 8 M to 6 M (dash–dot), 2 M (dotted), 1 M (short dash), 0.6 M (dash–dot–dot) and 0.4 M (long dash) during refolding.

Refolding is also achieved from GndHCl induced unfolded state of the protein. A native-like CD spectrum is obtained upon instantaneous dilution of GndHCl for 40 to 20 folds of dilution (Fig. 3E). However, upon 10 folds dilution resulting in 0.8 M GndHCl, no refolding is achieved in contrast to urea. In equilibrium measurements, we observe protein starts refolding at 0.6 M GndHCl and completely refolds at 0.4 M (Fig. 3A). Near UV-CD signature upon refolding is observed at 0.4 M GndHCl (Fig. 3F). A native-like fluorescence emission peak is observed at 0.4 M GndHCl (Fig. 3C). We observe the hydrodynamic size of protein upon reducing GndHCl concentration and find lowering of size from 12 nm at 6 M to 10 nm at 5 M and 4 M GndHCl (Fig. 3D). The size decreases further below 4 M GndHCl and approaches native-like size (∼6.5 nm) at 1 M and below (Fig. 3D). Thus, unfolding of STY3178 in presence of GndHCl is also completely reversible.

Effect of pH

Secondary structure of STY3178 in the far UV-region (200–250 nm) is observed for different pH ranging from 2.5–10 using CD. We observe that native protein at pH 7 has two minima around 209 and 222 nm, indicating an α-helical nature, which is in agreement with our previous results.¹³ We show the ellipticity at 222 nm (θ₂₂₂) as a function of pH in Fig. 4A. The plot shows that ellipticity does not change much compared to the native protein at pH 7. This indicates that STY3178 has a stable α-helical structure in both acidic (pH = 2.5) and alkaline media (pH = 10). Near UV (250–350 nm) CD spectrum of STY3178 in native state has a broad shoulder in the range of 250–280 nm (Fig. 4B). This shoulder remains unaltered in the pH range 6 to 10 (Fig. 4B). There is a slight decrease in ellipticity at pH 4.5 and the broad structure is disrupted completely in strong acidic buffer at pH 2.5 (Fig. 4B). This indicates that the protein tertiary or quaternary structure is stable in the pH range 4.5–10.

Fig. 4 Effect of pH and oxalate on STY3178 structure. (A) Shows the plot of ellipticity (θ₂₂₂) at 222 nm wavelength versus pH in the range 2.5–10. (B) Represents the near UV-CD spectra of protein in buffers having pH values 2.5 (dotted), 4.5 (dash), 6 (grey dot), 7 (solid), 8 (grey dash) and 10 (dash–dot). (C) Shows the emission peak position for excitation wavelength 280 nm for different pH (2.5–10). (D) The hydrodynamic size of protein is shown as a function of different pH. (E) Shows the fluorescence emission spectra of 10 μM protein for 280 nm excitation in presence of potassium oxalate with concentrations 10 μM (red), 50 μM (green), 100 μM (grey), 200 μM (blue) and 300 μM (magenta). The native protein spectrum is shown in black. (F) Represents the modified Stern–Volmer plot of log[F₀ − F/F] versus log[Q] for interaction of oxalate with STY3178.

The effect of pH is also monitored using steady state fluorescence emission. Fig. 4C shows the variation in peak position for 280 nm excitation wavelength. Native protein has an emission maximum around 342 nm as reported earlier.¹³ The peak position remains almost similar to that of the native protein for pH 7.5–10. We find a slight blue shift of around 339 nm in moderate acidic buffers (pH = 4, 4.5 and 5.2) compared to that of native protein. The emission shows red shift with peak position around 347 nm in strong acidic pH 2.5 and 3. The intensity of emission decreases for samples in buffers with acidic pH (=2.5 to 6.5) and increases for that in alkaline pH (=7.5, 8, 8.5, 9 and 10) compared to native state. The hydrodynamic size of STY3178 is measured for the pH range 2.5 to 10. The native protein at pH 7 has a hydrodynamic diameter of about 6.5 nm as reported in our earlier study.¹³ We observe this hydrodynamic size remains unchanged over the entire pH range (Fig. 4D). Our experimental observations suggest that yfdX protein, STY3178 is quite stable in wide range of pH.

Interaction with oxalate

We probe any possible interaction of STY3178 with oxalate since four other proteins yfdU, yfdW, yfdE and yfdV belonging to the same operon of E. coli are involved in oxalate-induced ATR¹² response. We monitor the steady state fluorescence emission of STY3178 at 280 nm excitation with increasing concentration of oxalate. We observe quenching of fluorescence emission in presence of oxalate as shown in Fig. 4E with increase in concentration. The binding parameter is estimated from the modified Stern–Volmer equation which is used as per the earlier protocol.¹³ The binding constant value obtained is 59 × 10⁻⁴ μM⁻¹ which is extracted from the intercept of log[F₀ − F/F] versus log[Q] as shown in Fig. 4F. Dissociation constant of 169.5 μM is obtained from the reciprocal of binding constant which indicates substantial interaction of STY3178 with oxalate.

Model of the trimeric assembly

We have reported earlier that STY3178 exists as a trimer in solution.¹³ Accordingly, we generate trimeric assemblies of protein in ZDOCK¹⁸ docking server using our simulated monomer model.¹⁴ We choose five best possible structures for the trimer based on the score and minimize them using NAMD¹⁹ with CHARMM27 ²⁰ force field and TIP3P²⁶ model of water to determine the energetically most favourable assembly. Most stable trimer model shows three monomeric chains symmetrically arranged in a disc-like conformation with a curvature having a concave and a convex surface (Fig. 5A). The concave surface has a central pore formed by the aromatic side chain of phenylalanine (F42) residues of three monomers. The electrostatic surface of trimer has both negative and positive charges exposed on the surfaces (figure not shown). The central pore on the concave surface is majorly positively charge. This assembly being symmetric, the interfacial residues are similar across the three monomeric subunits. The interface of any of the three subunits contains all residue types, acidic, basic, polar and hydrophobic (Table 1). Stabilization of the interface has major component from the three salt-bridges.

Fig. 5 Structure and conformational thermodynamics of STY3178 trimer. (A) Conformation of simulated model of STY3178 trimeric assembly showing a disc-like arrangement having a convex and a concave surface. Each monomeric subunits is coloured separately, A (green), B (magenta) and C (cyan). (B) and (C) show the conformational thermodynamics changes of STY3178 trimer with respect to monomers where changes in free energy, ΔG^conf_i for stabilized (green) and destabilized (red) residues are shown in (B) and residue wise changes in entropy, TΔS^conf_i for ordered (green) and disordered (red) residues are shown in (C). (D) and (E) quantify the contributions of interfacial charged, polar and hydrophobic residues where (D) shows ΔG^conf_interface and (E) TΔS^conf_interface of the trimeric assembly averaged over three chains.

Table 1 Interfacial residues of a monomeric subunit within the trimeric assembly

Residue type	Residue	Percentage
Acidic	D24, D85, E161, E187	9
Basic	R35, R41, K118, K122, R149, K150, K163	19
Polar	N46, N91, S92, S93, N114, Q146, Q147, T153, T154, Y164, Y165, Q166, Q175, S184, S188	47
Hydrophobic	W31, M39, F42, F45, I90, M117, G162, V185	25

---

We determine the radius of gyration (R_g) of trimeric assembly model from the equilibrated trajectory using the relation,²¹ , as described in method section. The calculated of trimeric model is 2.7 nm, the corresponding hydrodynamic radius (R_H) being 3.51 nm. The hydrodynamic diameter corresponding to the calculated R_H is 7.02 nm. This value is in good match with experimentally determined hydrodynamic diameter of about 6.5 nm, reported in our previous study.¹³ We also observe both intra-chain and inter-chain FRET pairs in the trimeric model. The respective pairs W70 and Y73, Y87; W71 and Y165, Y164 can explain the occurrence of FRET, observed¹³ in steady state fluorescence. Thus, the proposed trimer model can explain the experimental observations reasonably well.

Conformational thermodynamics

The relative stability of the trimer assembly with respect to its monomer is estimated using the thermodynamics of conformational changes.^{21,22,27–30} The conformational free energy and entropy changes are calculated from the equilibrium distributions of dihedral angles for each of the three sub-units (denoted by A, B and C) of trimer with respect to the monomer (Table 2). We find that the oligomerization of STY3178 into the trimeric assembly induces huge conformational stability and ordering with respect to the monomer (Fig. 5B and C). The three monomeric subunits within the assembly undergo similar gain in conformational stability and ordering. Owing to limited sampling we find small differences in free energy and entropy changes within the three monomeric subunits.

Table 2 Conformational free energy (ΔG^conf) and entropy (TΔS^conf) changes (in kJ mol⁻¹) of STY3178 trimer with respect to monomer

Chain id	ΔG^conf	TΔS^conf
A	−55.0	−345.4
B	−61.9	−308.6
C	−64.4	−276.6

---

The conformational thermodynamics changes of residues are calculated as an average over three subunits. The residues of STY3178 trimer that undergo a negative change in conformational free energy are stabilized, while those undergoing a positive change are destabilized. Similarly some residues show loss (ordered) or gain (disordered) in conformational entropy. Fig. 5B shows the distribution of conformational free energy changes having a predominance of stabilized residues (green) compared to few destabilized (red) residues. Similarly, the conformational entropy changes (Fig. 5C) indicates that most of the residues undergo conformational ordering (green) with sparsely distributed disordered residues (red). The interfaces are stabilized (Fig. 5D) as well as ordered (Fig. 5E). We find that the acidic and basic residues at the interface contribute largely to lower both conformational free energy and entropy followed by polar and hydrophobic residues. The interfacial thermodynamics changes indicate that inter-chain salt bridges between acidic and basic residues tend to provide stability to trimeric assembly (see Table 3). We find three such salt bridges at the interface: D24–K163, D85–K118 and E161–K122. These residues contributing to salt bridge formation show negative changes in both free energy and entropy. The large ordering of these acidic and basic residues indicates loss of flexibility due to salt bridge formation.

Table 3 Conformational free energy (ΔG^conf_i) and entropy (TΔS^conf_i) changes (in kJ mol⁻¹) of acidic and basic residues forming salt-bridge pairs at trimeric interface

Residue	ΔG^confi	TΔS^confi
D24	−1.2	−4.5
D85	−0.6	−5.8
E161	−0.5	−5.2
K118	−0.9	−7.4
K122	−1.25	−9.1
K163	−0.8	−8

---

Discussion

Our experiments suggest that STY3178 can refold back to its native-like state reversibly upon chemical denaturation. Hence, free energy change (ΔG) associated with its unfolding is estimated assuming a two-state model (native ⇆ denatured), as described in method section. We estimate the folding constant (K_F) from the fraction of folded protein (α_i) at a given chaotrope concentration from which ΔG is calculated (see Methods). ΔG shows a linear dependence over urea concentrations (Fig. 6A, triangle). The change in free energy at zero concentration of chaotrope (ΔG₀) is estimated to be around −23.4 (±0.9) kcal mol⁻¹. Similarly, we calculate the change in free energy (ΔG) as a function of different concentrations of GndHCl (Fig. 6A, circle). The value of ΔG₀ obtained upon extrapolating the plot to zero GndHCl is about −21 (±0.525) kcal mol⁻¹.

Fig. 6 Free energy of unfolding and interaction of rifampin in presence of urea. (A) The free energy change (ΔG) as a function of chaotrope concentration is shown in presence of urea (triangle) and GndHCl (circle). (B) and (C) Show the near UV-CD and fluorescence emission spectra of protein, respectively in presence of 0.6 M (green), 1 M (red) and 6 M (blue) urea. Spectra recorded after addition of 100 μM rifampin in (B) and 20 μM rifampin in (C) are shown using dotted lines in presence of both urea and rifampin. The native protein spectrum is shown in black for (B) and (C).

In an earlier report, we have shown that STY3178 interacts with antibiotics.¹³ We check whether STY3178 is able to interact with antibiotic in presence of urea. Near UV-CD spectrum of 30 μM protein contains a broad shoulder in the range 250–300 nm. We find decomposition of this shoulder into two peaks (∼260 nm and ∼280 nm) in presence of 100 μM Rfp similar to our earlier study.¹³ Similar decomposition of near UV-CD spectra of protein is observed in presence of 0.6 M and 1 M urea, after addition of 100 μM Rfp (Fig. 6B). This indicates that Rfp interacts with the protein in presence of urea as well. The shoulder completely diminishes in 2 M urea indicating change in protein structure. However, nature of near UV-CD spectrum of the protein in 2 M urea in the absence of Rfp differs from that in presence of 2 M urea and 100 μM Rfp. This suggests some interaction of protein with Rfp even in presence of 2 M urea.

Quenching of tryptophan emission in presence of Rfp upon 280 nm excitation in steady state fluorescence is also monitored. In presence of 0.6 M and 1 M urea, there is quenching of fluorescence emission upon addition of 20 μM Rfp as observed in case of 10 μM native protein (Fig. 6C). This suggests that protein interacts with Rfp in 0.6 M and 1 M urea. There is a red shift in emission peak position in 2 M urea due to exposure of tryptophan to more polar environment. We find quenching of tryptophan emission even in 2 M urea after adding Rfp. It indicates interaction of protein with the antibiotic, although tryptophan encounters more polar environment upon increasing urea concentration compared to the native state.

The surface exposed acidic and basic residues get affected through changes in their protonation states upon changes in pH. The trimeric assembly contains few such surface exposed acidic (D99, E138, D176 and D181) and basic (K25, K112 and K118) residues. The conformational thermodynamics of these residues shown in Table 4 indicate minor changes in free energy upon oligomerization. The major contribution to the stabilization of the trimeric state is due to salt bridges (Table 3). This may be the reason that the trimeric assembly remains stable across a wide range of pH starting from acidic to alkaline medium.

Table 4 Conformational free energy (ΔG^conf_i) and entropy (TΔS^conf_i) changes (in kJ mol⁻¹) of surface exposed acidic and basic residues of STY3178 trimer

Residue	ΔG^confi	TΔS^confi
D99	0.3	−0.3
E138	0.11	−2.4
D176	1.7	0.3
D181	1.2	0.6
K25	0.06	1.0
K112	0.04	0.6
K118	−0.9	−7.4

---

Conclusion

We monitor the structural changes of STY3178 in presence of urea and GndHCl. This protein can refold back to native-like structure from the chemically induced unfolded state. We show that STY3178 is stable over a wide range of pH. The α-helical structure is preserved in acidic and alkaline medium. This protein is also capable of interacting with oxalate. Our studies indicate that STY3178 might be involved in oxalate induced acid tolerance response of the bacteria. We perform MD simulation of its trimer and monomer models and estimate the conformational free energy and entropy changes upon trimerization. Our computational results indicate aggregation induced stability and ordering of STY3178 trimer through salt bridge interactions. Since STY3178 belongs to yfdX class of proteins, our work suggests that this protein may have functional roles in survival of bacteria in harsh chemical environment. It would be important to relate such functionalities to virulence of various bacterial strains.

Acknowledgements

M. G and P. S thank the high field NMR facility, TIFR, Mumbai. S. S thanks the UGC for funding. M. G and J. C thank the DST for funding.

References

1. M. B. Sztein, Clin. Infect. Dis., 2007, 45, S15–S19.
2. S. Ghosh, K. Chakraborty, T. Nagaraja, S. Basak, H. Koley, S. Dutta, U. Mitra and S. Das, Proc. Natl. Acad. Sci. U. S. A., 2011, 108, 3348–3353.
3. P. Everest, J. Wain, M. Roberts, G. Rook and G. Dougan, Trends Microbiol., 2001, 9, 316–320.
4. J. E. Galan, Curr. Top. Microbiol. Immunol., 1996, 209, 43–60.
5. C. M. Parry, V. A. Ho, T. Phuong le, P. V. Bay, M. N. Lanh, T. Tung le, N. T. Tham, J. Wain, T. T. Hien and J. J. Farrar, Antimicrob. Agents Chemother., 2007, 51, 819–825.
6. J. Parkhill, G. Dougan, K. D. James, N. R. Thomson, D. Pickard, J. Wain, C. Churcher, K. L. Mungall, S. D. Bentley, M. T. Holden, M. Sebaihia, S. Baker, D. Basham, K. Brooks, T. Chillingworth, P. Connerton, A. Cronin, P. Davis, R. M. Davies, L. Dowd, N. White, J. Farrar, T. Feltwell, N. Hamlin, A. Haque, T. T. Hien, S. Holroyd, K. Jagels, A. Krogh, T. S. Larsen, S. Leather, S. Moule, P. O'Gaora, C. Parry, M. Quail, K. Rutherford, M. Simmonds, J. Skelton, K. Stevens, S. Whitehead and B. G. Barrell, Nature, 2001, 413, 848–852.
7. W. Deng, S. R. Liou, G. Plunkett 3rd, G. F. Mayhew, D. J. Rose, V. Burland, V. Kodoyianni, D. C. Schwartz and F. R. Blattner, J. Bacteriol., 2003, 185, 2330–2337.
8. K. Nishino and A. Yamaguchi, J. Bacteriol., 2001, 183, 1455–1458.
9. K. Nishino, Y. Inazumi and A. Yamaguchi, J. Bacteriol., 2003, 185, 2667–2672.
10. N. Masuda and G. M. Church, J. Bacteriol., 2002, 184, 6225–6234.
11. E. M. Fontenot, K. E. Ezelle, L. N. Gabreski, E. R. Giglio, J. M. McAfee, A. C. Mills, M. N. Qureshi, K. M. Salmon and C. G. Toyota, J. Bacteriol., 2013, 195, 1446–1455.
12. E. A. Mullins, K. L. Sullivan and T. J. Kappock, PLoS One, 2013, 8, e67901.
13. P. Saha, C. Manna, S. Das and M. Ghosh, Sci. Rep., 2016, 6, 21305.
14. P. Saha, C. Manna, J. Chakrabarti and M. Ghosh, Sci. Rep., 2016, 6, 29541.
15. N. J. Greenfield, Nat. Protoc., 2006, 1, 2733–2741.
16. F. Delaglio, S. Grzesiek, G. W. Vuister, G. Zhu, J. Pfeifer and A. Bax, J. Biomol. NMR, 1995, 6, 277–293.
17. B. A. Johnson and R. A. Blevins, J. Biomol. NMR, 1994, 4, 603–614.
18. B. G. Pierce, K. Wiehe, H. Hwang, B. H. Kim, T. Vreven and Z. Weng, Bioinformatics, 2014, 30(12), 1771–1773.
19. J. C. Phillips, R. Braun, W. Wang, J. Gumbart, E. Tajkhorshid, E. Villa, C. Chipot, R. D. Skeel, L. Kale and K. Schulten, J. Comput. Chem., 2005, 26, 1781–1802.
20. B. R. Brooks, C. L. Brooks 3rd, A. D. Mackerell Jr, L. Nilsson, R. J. Petrella, B. Roux, Y. Won, G. Archontis, C. Bartels, S. Boresch, A. Caflisch, L. Caves, Q. Cui, A. R. Dinner, M. Feig, S. Fischer, J. Gao, M. Hodoscek, W. Im, K. Kuczera, T. Lazaridis, J. Ma, V. Ovchinnikov, E. Paci, R. W. Pastor, C. B. Post, J. Z. Pu, M. Schaefer, B. Tidor, R. M. Venable, H. L. Woodcock, X. Wu, W. Yang, D. M. York and M. Karplus, J. Comput. Chem., 2009, 30, 1545–1614.
21. S. Sikdar, J. Chakrabarti and M. Ghosh, Mol. BioSyst., 2014, 10, 3280–3289.
22. A. Das, J. Chakrabarti and M. Ghosh, Biophys. J., 2013, 104, 1274–1284.
23. B. M. Tande, N. J. Wagner, M. E. Mackay, C. J. Hawker and M. Jeong, Macromolecules, 2001, 34, 8580–8585.
24. A. B. Datta, S. Roy and P. Parrack, Eur. J. Biochem., 2003, 270, 4439–4446.
25. U. Anand, C. Jash and S. Mukherjee, Phys. Chem. Chem. Phys., 2011, 13, 20418–20426.
26. M. W. Mahoney and W. L. Jorgensen, J. Chem. Phys., 2000, 112, 8910–8922.
27. A. Das, S. Sikdar, M. Ghosh and J. Chakrabarti, J. Phys.: Conf. Ser., 2015, 638, 012013.
28. A. Das, J. Chakrabarti and M. Ghosh, Chem. Phys. Lett., 2013, 581, 91–95.
29. A. Das, J. Chakrabarti and M. Ghosh, Mol. BioSyst., 2014, 10, 437–445.
30. S. Sikdar, J. Chakrabarti and M. Ghosh, Mol. BioSyst., 2016, 12, 444–453.

---