Crowder-Induced Conformational Ensemble Shift in Escherichia coli Prolyl-tRNA Synthetase

Abstract

The effect of molecular crowding on the structure and function of Escherichia coli prolyl-transfer RNA synthetase (Ec ProRS), a member of the aminoacyl-transfer RNA synthetase family, has been investigated using a combined experimental and theoretical method. Ec ProRS is a multidomain enzyme; coupled-domain dynamics are essential for efficient catalysis. To gain insight into the mechanistic detail of the crowding effect, kinetic studies were conducted with varying concentrations and sizes of crowders. In parallel, spectroscopic and quantum chemical studies were employed to probe the “soft interactions” between crowders and protein side chains. Finally, the dynamics of the dimeric protein was examined in the presence of crowders using a long-duration (70 ns) classical molecular dynamic simulations. The results of the simulations revealed a shift in the conformational ensemble, which is consistent with the preferential exclusion of cosolutes. The “soft interactions” model of the crowding effect also explained the alteration in kinetic parameters. In summary, the study found that the effects of molecular crowding on both conformational dynamics and catalytic function are correlated in the multidomain Ec ProRS, an enzyme that is central to protein synthesis in all living cells. This study affirmed that large and small cosolutes have considerable impacts on the structure, dynamics, and function of modular proteins and therefore must be considered for stabilizing protein-based pharmaceuticals and industrial enzymes.

Significance

The impact of the crowding on protein structure and function is an emerging topic of biophysical chemistry. This study used kinetics, spectroscopy, and molecular dynamics simulations to probe the crowding mechanism on an enzyme system, Escherichia coli prolyl-transfer RNA synthetase, in which coupled-domain dynamics are integral to its function. This study investigated the impact of size and concentration of crowders on the catalytic efficiency, probed “hard” (steric) and “soft” (chemical) interactions between crowders and protein side chains, and explored the changes in conformational ensembles. The study joins a growing number of work that demonstrates how important the effects of “soft interactions” are in altering conformational ensembles, thereby impacting the catalytic function of a multidomain protein.

Introduction

Aminoacyl-transfer RNA (tRNA) synthetase (AARSs) are multidomain enzymes that play a critical role in protein biosynthesis in all living organisms; they catalyze the covalent attachment of amino acids to their respective tRNAs (1, 2). The AARS prolyl-tRNA synthetase (ProRS) catalyzes the ligation of proline to tRNA^Pro in a two-step reaction. The first step is the formation of the enzyme-bound prolyl-adenylate complex (Pro-AMP) and pyrophosphate (PP_i) in the presence of ATP (Eq. 1). In the second step, the activated amino acid is transferred to the 3′-end of tRNA^Pro (Eq. 2), forming aminoacylated tRNA (Pro-tRNA^Pro) as follows:

$$\text{ProRS}+\text{Pro}+\text{ATP}\rightleftharpoons \text{ProRS}\cdot \left(\text{Pro}-\text{AMP}\right){+\text{PP}}_{\text{i}},$$ (1)

$$\text{ProRS}\cdot \left(\text{Pro}-\text{AMP}\right)+{\text{tRNA}}^{\text{Pro}}\to {\text{Pro}-\text{tRNA}}^{\text{Pro}}+\text{ProRS}+\text{AMP}\text{.}$$ (2)

Many AARSs, including ProRSs, undergo a large conformational change upon substrate binding (3, 4, 5). Our research group has demonstrated that the dynamic coupling between distant domains is crucial for effective catalysis (6, 7). However, these experimental and theoretical studies were carried out in aqueous environments. different from what is found in vivo. The interior of a cell is crowded with many biological macromolecules and metabolites (8). The cytoplasm of cells such as Escherichia coli have a macromolecule concentration of up to 300–400 g/L, and 20–30% of the intracellular volume is occupied by macromolecules (9, 10). Recent studies revealed that crowding affects protein stability (11, 12, 13, 14), folding (15, 16), binding affinity (17, 18), and catalysis (15, 19, 20, 21, 22, 23). Specifically, the size, structure, concentration, and chemical nature of crowding agents contribute to changes in stability, folding, conformational equilibrium, and catalysis when compared with dilute conditions (24, 25, 26, 27).

The crowding is inherently a dynamic phenomenon both in terms of the diffusional aspect of the crowders and the conformational dynamics of the enzyme (27). The preferential exclusion or accumulation of cosolutes (crowders) from/on the protein surface, proposed by Timasheff (28, 29) in the early 80s, is considered to be the core thermodynamic mechanism (30) behind the crowding effect. However, the underlying molecular mechanism of the crowding on protein structure and function is still under debate (31). The effects of “steric repulsions” (excluded volume effect) and the “soft interactions” (chemical interactions) have been probed in many laboratories (32, 33, 34, 35). The alteration of solvent dielectric properties by crowders, which in turn modulates “soft interactions” between biological macromolecules and crowding agents, has also been proposed (36). The interplay of enthalpic-entropic changes stemming out of solvent-cosolute and cosolute-macromolecule interactions has only been partially revealed (31).

The proposed thermodynamic mechanism poses new questions about the role of molecular crowding on the structure and function of E. coli ProRS (Ec ProRS). There are two main questions: is the coupled dynamics among various domains impacted by crowding, and if impacted, does it alter the protein’s function? To gain a molecular-level understanding of the effects of molecular crowding on Ec ProRS function, an investigatory approach with three major components, namely, enzyme kinetics, intrinsic tryptophan fluorescence spectroscopy, and molecular dynamic (MD) simulations, has been employed. Kinetic assays were performed to explore the impact of size and concentration of crowding agents on the catalytic efficiency of Ec ProRS. In parallel, interactions between protein side chains and crowding agents were studied through intrinsic tryptophan fluorescence measurements and validated through approximate quantum chemical studies. Finally, the preferential interactions between cosolutes and the protein and the subsequent conformational dynamics were studied by long-duration MD simulation. The ensembles of evolved conformational populations were examined, and their impact on enzyme functions was assessed. Our study provides a comprehensive assessment of the impact of crowding on the conformation, dynamics, and catalytic function of Ec ProRS.

Materials and Methods

All crowding agents were purchased from Sigma-Aldrich (St. Louis, MO), except for polyethylene glycol (PEG) 8000 (Thermo Fisher Scientific, Waltham, MA). Proline (≥99%) was also from Sigma-Aldrich. Both [γ- ³²P] ATP and [³²P] PP_i were purchased through PerkinElmer (Shelton, CT).

Overexpression and purification of Ec ProRS

Wild-type (WT) Ec ProRS was overexpressed in SG13009 (pREP4) competent cells using 0.1 mM isopropyl β-D-thiogalactoside for 4 h at 37°C. Histidine-tagged WT Ec ProRS was purified using Talon cobalt affinity resin column; 100 mM imidazole was used to elute the protein (37, 38). The Bio-Rad protein assay (Bio-Rad Laboratories, Hercules, CA) was used to determine the total concentration of protein. An active site titration was performed to determine the concentration of active protein (39).

Enzyme kinetics

Crowding agent concentration variation

The ATP-PP_i exchange assay was performed, following the protocol described elsewhere, to examine the effect of increasing concentrations of crowding agents on proline activation (Eq. 1) by Ec ProRS (40). In the ATP-PP_i exchange assay, radiolabeled PP_i (³²PP_i, ≈10 μCi/mL) and nonradiolabeled ATP (2 mM) were used, and the percentage of product (Pro-AMP) formation at 20 min postinitiation of the reaction was measured in the presence of crowding agents. The Pro-AMP formed was indirectly measured by monitoring the amount of ³²P-ATP formed via the reverse reaction of Eq. 1. The percentage of product formation was calculated from the ratio of the Pro-AMP formation in presence and absence of the crowding agent (i.e., the regular buffer solution). Reactions containing 0.75 mM proline, 10 nM WT Ec ProRS, and the crowding agent were incubated at 37°C for 20 min before being quenched with 0.4 M PPi, 15% HClO₄, and 3% activated charcoal. A range of 100–300 g/L dextrose and sucrose was used, while PEG 8000 and Ficoll 70 were varied from 50–200 g/L. The same concentrations of proline and enzyme were used in dilute solution as well. The reaction mixture containing a given crowding agent and substrates (proline and ATP) in the absence of enzyme was considered as a control. For all crowding agents, the stability of the enzyme was examined by incubating the enzyme with a given crowding agent at 37°C for 30 min and analyzing for any degradation using polyacrylamide gel electrophoresis (Fig. S1).

Proline concentration variation to determine the kinetic parameters for proline activation

ATP-PP_i exchange assays were performed following the method described previously (40). The impact of the crowding agent on proline activation (Eq. 1) efficiency was ascertained by comparing the K_M and V_max of WT Ec ProRS (in the presence of crowders) relative to that in the absence of any crowding agent (i.e., the dilute solution). These assays employed proline concentrations between 0.125 and 1.00 mM, 10 nM of WT Ec ProRS, and 200 g/L of crowding agents (dextrose, sucrose, and Ficoll 70). Concentrations higher than 50 g/L for PEG 8000 resulted in a dramatic reduction in catalytic activity, preventing the consistent and accurate determination of kinetic parameters. Therefore, the final concentration was maintained at 50 g/L for PEG 8000. Reaction mixtures were incubated at 37°C for 4, 8, 12, 16, and 20 min before being quenched with 0.4 M PP_i, 15% HClO₄, and 3% activated charcoal (40). Ec ProRS follows Michaelis-Menten kinetics. The Michaelis-Menten equation consists of two parameters, V_max (the maximal reaction rate) and K_M (the Michaelis constant). These two parameters were obtained from Lineweaver-Burk plots (41, 42). In this method, the reciprocal of the initial velocity is plotted against the reciprocal of the substrate concentration. A straight line is fitted to the points; V_max is calculated as the reciprocal of the intercept of the line on the l/V axis, and K_M is obtained by multiplying the slope of the line, K_M/V_max by V_max. All kinetic assays were performed either in duplicate or triplicate (when the difference between the two measured parameters was >10%). A Student’s t-test with two-tailed statistical analysis was performed on the data from both the proline variation study and the crowding agent concentration variation study.

PEG size variation

To investigate the impact of the size of crowding agent, without changing the chemical nature of crowding agents, a size variation study was conducted that utilized PEG 600, 1500, 4000, 6000, 8000, and 20,000. Again, the ATP-PP_i exchange assay was performed (40), in which constant concentrations of 0.75 mM proline, 10 nM WT Ec ProRS, and 100 g/L PEG were maintained. Reactions were incubated at 37°C for 20 min before being quenched with 0.4 M PP_i, 15% HClO₄, and 3% activated charcoal (40). The quantity of Pro-AMP (nmol) formed was indirectly measured by monitoring the amount of ³²P-ATP formed and compared with the R_h (Å) of each PEG (Table S1; (43)). Reactions with dilute conditions were performed as a reference.

Intrinsic tryptophan fluorescence spectroscopy

Intrinsic tryptophan fluorescence was measured using a PerkinElmer LS 55 Fluorimeter (PerkinElmer, Waltham, MA). Excitation wavelength was 280 nm, and the emission spectra were collected from 300 to 400 nm. Baseline emission spectra (controls) were recorded with solutions that contained all reagents but the WT Ec ProRS enzyme (i.e., crowders, buffer, and salt) to eliminate the effect of impurities. Fluorescence solutions contained crowding agents of concentrations ranging from 25 to 400 g/L, 2 μM WT Ec ProRS, 10 mM phosphate buffer (pH = 7.4), and 100 mM NaCl. Crowding agents from the smallest hydrodynamic radius to the largest hydrodynamic radius are as follows: dextrose, sucrose, PEG 8000, and Ficoll 70 (Table 1). Experiments were also performed that utilized a combination of two different crowding agents, namely sucrose with Ficoll 70, sucrose with dextrose, and Ficoll 70 with dextrose. Samples were incubated at room temperature for 10 min. All experiments were performed in triplicate. The barycentric mean fluorescence wavelength (λ_bcm) was determined using the following equation (Eq. 3):

$${\lambda }_{\text{bcm}}=\frac{\sum _{\lambda }\lambda \times I\left(\lambda \right)}{\sum _{\lambda }I\left(\lambda \right)},$$ (3)

where λ is the wavelength, and I(λ) is the emission intensity at that given wavelength. The change in λ_bcm [Δλ_bcm = λ_bcm (crowder) − λ_bcm (water)] because of the presence of crowders was examined to monitor changes in the tryptophan’s local environment.

Table 1.

Tryptophan Intrinsic Fluorescence in the Presence of Crowding Agents

^a,b,cHydrodynamic radii were taken from literature for dextrose (96), PEG 8000 (43), and Ficoll 70 (25).

^dΔλbcm and the percentage of intensity reduction in the presence of 300 g/L sucrose were calculated from the equation of the linear regression of Fig. S5

Computational setup

All gas-phase potential energies were calculated using self-consistent density functional tight-binding theory with dispersion corrections (44, 45). Protein structure was obtained from the protein database (46). Homology models were constructed using the web-based utility SWISS-MODEL (47, 48, 49). All structural manipulations were carried out using Visual MD (VMD) (50). Packmol was used to distribute molecules of sucrose and dextrose randomly throughout the simulation volume (51). MD simulations were carried out using the NAMD package using CHARMM36 force field (52, 53, 54, 55, 56). Electrostatic energy calculations were carried out using particle-mesh Ewald method (57). All SASA calculations were performed with VMD’s dedicated algorithm using a probe radius of 1.4 Å. Radial distribution function (RDF) was calculated from the MD trajectory data using VMD. The collective motion was analyzed using PCA of the MD trajectory data using CARMA (58, 59). Backbone root mean-square deviation (RMSD) calculations were performed with CARMA using the first frame of the trajectory as the reference.

MD simulation

All simulations were performed on a homology model structure of Ec ProRS constructed using Ec ProRS amino acid sequence and Enterococcus faecalis ProRS crystal structure (Protein Data Bank [PDB]: 2J3L) as a template. These simulations involved the dimeric form of the Ec ProRS and were carried out in the presence of two crowding agents, namely dextrose and sucrose in addition to that in dilute condition. The setup of these systems in this simulation has been carried out following protocols used in our previous study (7). All structures were explicitly solvated (with TIP3P model) and ionized (with 42 sodium atoms) with VMD plugins, resulting in solvent boxes with dimensions 138 × 114 × 95 Å for dilute (or sucrose) and 145 × 115 × 101 Å for dextrose (50). The sucrose and dextrose boxes were built in the same manner to a final crowder concentration of 200 g/L. Placements having an atom less than 2.0 Å from any other atom were deleted. After completion of the setup, there were a total of 383 sucrose molecules and 777 dextrose molecules. The total number of water molecules in dilute, sucrose, and dextrose systems were 41,334, 31,100, and 29,205, respectively.

All systems underwent 40 ps of minimization. The 70 ns of MD simulations were run using the NAMD package with a time step of 2.0 fs. All simulations were run at a constant temperature of 298°K and pressure of 1 bar via NAMD implementation of Langevin dynamics (60) within periodic boundary conditions. The electrostatic calculations involved a grid size of 150 × 150 × 180 Å. Switching was turned on for electrostatic and van der Waals interactions with a “switchdist” of 10 Å, a cutoff of 14 Å, and a “pairlistdist” of 16 Å.

Preferential interaction coefficients

The preferential interaction coefficient between a certain cosolute and protein was calculated as described by Ma et al. (61). In this method, the three components, namely water, protein, and the cosolute, are labeled as 1, 2, and 3, respectively. The relative accumulation or exclusion of a cosolute on/from the protein surface (i.e., local) can be computed using a two-domain model as proposed by Record et al. (62, 63) using Eq. 4

$${\mathit{\Gamma }}_{2\text{,}3}=\langle n_3^{local}-n_1^{local}\times \frac{n_3^{bulk}}{n_1^{bulk}}\rangle ,$$ (4)

where $n_3^{local}$and $n_1^{local}$ are the number of cosolute and water particles in the local domain, respectively. Similarly, $n_3^{bulk}$and $n_1^{bulk}$ are the number of cosolute and water particles in the bulk domain. The angular bracket represents the ensemble average quantity. Γ_2,3 is related to the partial molar chemical potentials μ_2,3 and μ_3,3 (61) as follows:

$${\mathit{\Gamma }}_{2\text{,}3}=-\frac{{\mu }_{2\text{,}3}}{{\mu }_{3\text{,}3}},$$ (5)

where μ_2,3 and μ_3,3 are the partial derivatives of chemical potential of the protein and the cosolute, respectively, with respect to the change in the cosolute’s molality (62):

$${\mu }_{2\text{,}3}={\left(\frac{\partial {\mu }_2}{\partial m_3}\right)}_{p,T,m_2},$$ (6a)

$${\mu }_{3\text{,}3}={\left(\frac{\partial {\mu }_3}{\partial m_3}\right)}_{p,T{\text{,m}}_{2,\mathrm{...}}}.$$ (6b)

The μ_3,3 can be calculated under ideal state approximation and is equal to RT/m₃, where R is the ideal gas constant, T is the temperature, and m₃ is the molality of the cosolute.

Principal component analysis

Essential dynamics (64) of the protein was calculated by principal component analysis (PCA), following procedures discussed earlier (6). Briefly, a modified trajectory file was prepared from the 70-ns trajectory by removing the overall translational and rotational motions and retaining the information of only C_α atoms’ fluctuations. The principal components (or modes) of the motion were obtained by diagonalizing the covariance matrix computed for C_α atoms. A clustering of conformations was carried out based on the contribution of the first three principal components, and the backbone fluctuations were studied using CARMA (58).

Potential energy surface calculation

The calculations of potential energy of interactions between tryptophan residues with solvent molecules were carried out in the gas phase. The interaction potential energy was calculated between a single indole molecule and a water molecule or a crowding agent (dextrose and sucrose). The structural models of these three interacting systems were generated using CHARMM (65) package. Each system was developed by moving the solvent/small molecule from 1 to 12 Å through a direction perpendicular to the plane of the indole by 0.1 Å. A total of 100 single-point potential energy per system were calculated with self-consistent charge-density functional tight binding with dispersion (44, 45) correction.

Results and Discussion

As described earlier, ProRS catalyzes a two-step aminoacylation reaction. This study has focused on the first step, which involves the formation of Pro-AMP intermediate. This first step of the reaction (Eq. 1) is tRNA^Pro independent for Ec ProRS (40), and it is the rate-limiting step for class II AARSs like ProRS (66). Therefore, we determined the kinetic parameters for the first step of the aminoacylation reaction to explore the impact of crowders on the enzymatic function. This narrowed the number of complications considerably that may have arisen because of the presence of tRNA^Pro. With the kinetic assays, we have considered the impact of concentration, size, and the chemical nature of a crowding agent on the Pro-AMP formation (i.e., proline activation).

Decrease in product formation with an increasing concentration of crowding agents

For all crowders, the general trend was a decrease in the product (Pro-AMP) formation with an increasing concentration of the crowding agent (Fig. 1). However, there was an observable increase in the product formation for solutions containing 100 g/L sucrose and 50–200 g/L Ficoll 70 as compared with the dilute solution. In contrast, PEG 8000 caused a significant decrease in product formation (Fig. 1). Even at a lower concentration of 50 g/L, the percentage of product formation was less than that observed with 300 g/L dextrose and sucrose (Fig. 1).

Figure 1.

Percentage of product formation in the presence of crowding agents. Crowding agent concentration ranged from 100 to 300 g/L for dextrose and sucrose, while a 50-200 g/L range was used for PEG 8000 and Ficoll 70. A concentration of 10 nM WT Ec ProRS and 0.75 mM proline were used. Reactions were incubated at 37°C and quenched at 20 min post-initiation. The percentage of product formation was calculated by comparing the nmol of Pro-AMP formed after 20 min in the presence of crowding agents to that formed in dilute conditions. A two-tailed Student’s t-test was performed to assess the significance of the difference between the percentage of product formation in the presence of each individual crowding agent of a specific concentration and the dilute condition. The significance of the results of the t-test are represented as follows: ^∗∗p < 0.01, ^∗∗∗p < 0.05, and ^∗∗∗∗p < 0.08.

The above results revealed that for a given crowder, the product formation decreased with increasing concentration, irrespective of their chemical nature or size. One can rationalize this trend by considering the steric effect, which can limit the accessibility to the enzyme active site and thereby impact the product formation. Alternately, the crowding may also influence enzyme conformational equilibrium and internal dynamics. Such variation in conformational equilibrium due to crowding has been reported earlier (24) and could have impacted the product formation by the multidomain Ec ProRS, in which coupled-domain dynamics is known to be important for catalysis (6). To have better insight into the impact of crowding agents on Pro-AMP formation, the kinetic parameters were determined.

Impact of crowding on kinetic parameters

Ec ProRS was observed to follow Michaelis-Menten kinetics in the presence of crowding agents. The K_M and V_max for proline activation by Ec ProRS were obtained from Lineweaver-Burk plots (Fig. S2; Table S2). A tighter substrate binding is evident in the presence of crowders as the K_M was found to be significantly lower (Fig. 2 a). The most significant decrease in relative K_M was obtained with 200 g/L dextrose, resulting in a relative K_M less than 0.5 (Fig. 2 a). The lowering in K_M values indicates no substantial interactions between substrates and crowders as observed in the study of Aumiller et al. (67).

Figure 2.

Kinetic parameters for Pro-AMP formation in the presence of various crowding agents: (a) all crowding agents affect the K_M of Ec ProRS; (b) Ficoll 70 was the only crowding agent resulting in no significant change in relative V_max; and (c) sucrose enhanced relative k_cat/K_M, whereas PEG decreased relative k_cat/K_M. These studies were performed with 10 nM WT Ec ProRS and 0.125–1.00 mM proline. Reactions were incubated at 37°C. A concentration of 200 g/L for PEG 8000 resulted in too significant of a reduction in catalytic activity to obtain accurate and consistent kinetic parameters such as K_M and V_max. Error bars represent differences between two trials. Values for K_M and V_max were determined using a Lineweaver-Burk plot of a proline concentration variation assay (Fig. S2). Relative K_M and V_max displayed on the y axis were normalized to the K_M and V_max determined under dilute condition. A two-tailed Student’s t-test was performed to assess the significance of the difference between the relative kinetic parameters in the presence of crowding agents compared with dilute condition. The significance of the results of the t-test are represented as follows: ^∗p < 0.005, ^∗∗p < 0.01, and ^∗∗∗p < 0.05.

The reaction rate (V_max) of the formation of a prolyl-adenylate decreased for all crowders. For example, in the presence of 50 g/L PEG 8000, the relative V_max was found to be significantly reduced (>75%) compared with the dilute condition (Fig. 2 b). Also, compared with the dilute condition, the relative V_max in 200 g/L dextrose and 200 g/L sucrose decreased by ∼50 and 25%, respectively. However, the relative V_max was not significantly different in the presence of 200 g/L Ficoll 70 (Fig. 2 b). Compared with the dilute condition, a decrease in k_cat/K_M was observed in the presence of 50 g/L PEG 8000, whereas an overall increased catalytic efficiency was noted in the presence of dextrose, sucrose, and Ficoll 70 (Fig. 2 c). This significant decrease in the catalytic efficiency is consistent with the decreased product formation in the presence of PEG 8000 (Fig. 1).

The alteration in binding (K_M) and reaction rate (V_max) of the formation of Pro-AMP can be explained by the excluded volume effect, in which the overall accessible space is decreased in the presence of crowders, resulting in an increase in local concentration of substrate (68). Alternately, the crowders are also known to impact the conformational equilibrium favoring the substrate-bound conformation (68). In both cases, the stronger enzyme-substrate interactions would decrease K_M, consistent with what is observed. Furthermore, the stronger interaction would stabilize the enzyme-substrate complex, lowering its potential energy, which in turn would increase the activation barrier. The above scenario is consistent with the reduction in V_max. However, a significant impact on K_M but not on V_max in the presence of 200 g/L Ficoll 70 could be due to a subtle conformational change in the enzyme, which resulted in a slight decrease in K_M. A similar observation was made for Enterobactin-specific isochorismate synthase (22).

Impact of variable-sized PEG on product formation

Changes in Ec ProRS kinetic parameters (Fig. 2; Table S2) did not correlate with the R_h (Å) of the crowding agents (Table 1). This is expected as the concentration of the crowders used to mimic the cellular interior was greater than their overlap concentrations (69). An impact of crowder size on proline activation was noticed when endpoint kinetics with variable-sized PEG molecules was performed using a concentration below their overlap concentrations (Fig. 3; Table S1; (70)). The main reason for choosing PEG is its simple chemical structure and the availability of variable-sized PEG with hydrodynamic radius comparable to the other crowders. A nonlinear trend was observed between the Pro-AMP formation (nmol) and the hydrodynamic radius (R_h) (Å) of variable-sized PEG (Fig. 3). It is apparent that when considering size alone, smaller crowding agents have more impact on the catalytic function of Ec ProRS; the smaller the R_h of the PEG, the less Pro-AMP was formed.

Figure 3.

The impact of of variable-sized PEG (hydrodynamic radii, Table S1) on the formation of Pro-AMP (nmol). A concentration of 100 g/L PEG, 0.75 mM proline, and 10 nM WT Ec ProRS was used. Reactions were quenched at 20 min. PEG sizes from furthest left to right on graph are as follows: 600, 1500, 4000, 6000, 8000, and 20,000. Dilute condition had an average formation of 4.23 ± 0.04 nmol Pro-AMP.

One explanation for the trend is that it is due to the poor accessibility of the enzyme active site in the presence of smaller PEG molecules. However, all PEG molecules used in this study are larger compared with the active site of Ec ProRS, which has an estimated volume of ∼500 Å³ (Fig. S3). Therefore, the reduction in product formation may not be due to the blocking of the active site. Rather, these molecules may have “soft interactions” with the enzyme. This is more likely because PEG interacts with proteins and is capable of inducing conformational changes (71, 72), which in this scenario could impact the conformational equilibrium. The above finding is consistent with a recent report by Shkel et al., in which the authors observed a dominant role of the “soft interactions” for the small PEGs, whereas the larger PEGs affect the chemical potentials of proteins by a combination of excluded volume effect and reduced chemical interactions (73).

Crowder-induced changes in intrinsic tryptophan fluorescence

To examine if “soft interactions” are responsible for the shift in conformational equilibrium, we assessed intrinsic tryptophan fluorescence study in the presence of crowders. Intrinsic fluorescence of tryptophan is sensitive to the local environment, and both fluorescence intensity and fluorescence wavelength are influenced by the extent of the interactions of a tryptophan with its neighboring species (74). It is also known that if a tryptophan becomes more exposed to the surrounding hydrophilic solvent, there is a shift in the barycentric mean fluorescence wavelength (λ_bcm) toward a higher wavelength (red shift/Stokes shift) (74, 75). Conversely, if a tryptophan becomes less exposed, the λ_bcm undergoes a shift to lower wavelength (blue shift/anti-Stokes shift) (74). Hence, changes in the tryptophan fluorescence serve as an excellent tool to assess “soft interactions,” conformational changes, substrate binding, and denaturation (74). Ec ProRS is a dimer with five tryptophan residues per subunit (vide infra). Any effect of crowding on the intrinsic tryptophan fluorescence is the combined effect resulting from the varied degrees of interactions of the five tryptophan residues with solvent/crowders molecules.

Alteration in fluorescence emission intensity in presence of crowders

The intensity of fluorescence emission decreased as the concentration of crowders increased (Figs. S4–S7 a a; Table 1). Here, dextrose was the exception, in which 100 g/L had the greatest decrease and 300 g/L had the least effect (Fig. S4 a). The underlying mechanism of this effect is not well understood. As we are observing the net effect of 10 tryptophans present in the dimeric Ec ProRS, it could be possible that a higher concentration of dextrose is sequestering some of the tryptophans and protecting them from being quenched as it has been observed in the case of surfactants (76). The greatest decrease in fluorescence emission intensity was produced by Ficoll 70 (76.7%), followed by sucrose (51.9%) and then PEG 8000 (43.3%); dextrose had the least effect (6.52% reduction) (Table 1). Although no definitive trend was observed between the size of the crowding agent and the percentage of reduction in fluorescence emission intensity, the larger crowders have more impact.

The quantum yield of tryptophan fluorescence decreases in a hydrophilic environment (77). Fluorescence quenching could be either due to collisions (dynamic quenching) or due to “soft interactions” (static quenching) with the crowders and solvent molecules (77). In either case, the hydrophobic tryptophan fluorophore must have been exposed to the surface, indicating moderate to strong interactions between Ec ProRS and crowding agents. This is supported by the fact that a similar extent of reduction in fluorescence intensity was also observed for free tryptophan in the presence of crowders. For free tryptophan, 300 g/L sucrose and dextrose resulted in a 30 and 2.5% decrease in fluorescence intensity, respectively.

A mixture of crowding agents also had an impact on fluorescence emission intensity. A combination of 100 g/L Ficoll 70 and 200 g/L sucrose resulted in the greatest decrease in emission intensity (Figs. S8–S10 a a; Table S3). Solutions with a combination of sucrose and dextrose had a lower percentage of decrease in the emission intensity compared with other mixtures of two crowding agents (Table S3). There was also a correlation between the percentage of emission intensity reduction and the total concentration of crowding agents, with higher total concentrations of crowders resulting in a higher percentage reduction than lower concentrations (Table S3). A closer analysis revealed that the impact of the mixture of two crowding agents on the fluorescence intensity (Table 1; Table S3) is additive, suggesting that crowding agents are acting in a noncompetitive manner.

Shift in barycentric mean fluorescence wavelength in the presence of crowders

All crowding agents, except for dextrose, showed increasing λ_bcm (red shift) with increasing concentrations of crowding agents; Ficoll 70 resulted in the largest increase in λ_bcm (Figs. S4–S7 b b; Table 1). Dextrose, being the exception, resulted in a Δλ_bcm that was slightly negative compared with dilute conditions (Table 1). There was no distinct trend between the R_h of these four crowding agents and Δλ_bcm (Table 1). However, it took approximately twice the concentration of sucrose to achieve the same increase in λ_bcm as observed for Ficoll 70 (Table S3). An increase in λ_bcm was observed in the presence of a mixture of crowders; the combination of 100 g/L Ficoll 70 and 200 g/L sucrose has the greatest effect on λ_bcm (Figs. S8–S10 b b; Table S3). Because the resultant fluorescence emission spectra is the sum total of emissions from all of the 10 tryptophan residues in the dimeric Ec ProRS, we can only say that the crowding effect led to a net increase/decrease in tryptophan exposure to the solvent (74). The changes in λ_bcm indicate considerable interactions between tryptophan residues and surrounding solvent/cosolute molecules, capable of inducing a change in the conformation of Ec ProRS. For free tryptophan, Δλ_bcm was ≤1 nm in 300 g/L sucrose, and it remained practically unchanged in the presence of 300 g/L dextrose.

Taken together, the alteration in λ_bcm and the fluorescence intensity in the presence of crowding agents is indicative of a crowder-induced conformational change in Ec ProRS. A similar observation has been made for other protein systems, in which protein folding and conformational equilibrium were affected by molecular crowding (24, 25, 29, 78, 79). To obtain a molecular-level understanding of how a crowding agent alters an enzyme’s conformational equilibrium, we took a closer look at the impact of the two crowding agents, dextrose and sucrose, on the Ec ProRS conformation and dynamics using MD simulations.

MD simulations to probe the crowder-induced conformational dynamics

Simulations of the dimeric Ec ProRS were performed using the conditions of the fluorescence experiments. Enzyme systems were built using the following three conditions: 1) in water or dilute, 2) in the presence of 200 g/L dextrose, and 3) in the presence of 200 g/L sucrose. As a representative system, the sucrose-containing solvated system is shown in Fig. S11. Although PEG has a significant impact on product formation, modeling a protein system with PEG was not straightforward. Instead, only dextrose and sucrose were chosen because of the simplicity in building the solvent box with these two crowding agents. To ascertain changes in collective domain motions, the essential dynamics of the protein were studied by PCA. In addition, solvent accessible surface area (SASA) calculations were carried out to explore the changes in the conformation of the protein over the course of entire simulations.

The SWISS-MODEL homology server generated two models (one per monomer unit), each of which had Global Model Quality Estimation scores of 0.78 and coverage of >95% (data not shown). The resulting dimeric structure was analyzed using Procheck (80, 81). A total of 87.0% of the residues were in the most favored regions of the Ramachandran plot, 11.6% in additional allowed regions, and only five of (0.5%) residues in disallowed regions (Fig. S12). These results ensured that the model structure of the dimeric Ec ProRS is of reliable quality and suitable for MD simulations.

Impact of crowders on protein conformational distribution

Each subunit of Ec ProRS has three distinct domains—the editing domain (also known as insertion or insertion [INS] domain), the central catalytic domain (CD), and the anticodon-binding domain (Fig. 4 a). The two cosolutes, dextrose and sucrose, have a significant impact on the motion of the INS domain. In the optimized structures, the protein segment with residues 313–322 of INS domain was located ∼6 Å away from the small helical segment (residues 84–93) of the central CD (Fig. 4 a). The proximity between the above two segments was observed for all three systems (water, dextrose, and sucrose), indicating that the “closed” conformation occurs at the start of each simulated system. During the 70-ns simulation, the INS domain was moved away from the CD domain, forming an “open” state in the absence of any crowding agent (Fig. 4 a). The extent of opening was measured from the displacement of the INS domain relative to the CD. The interdomain distance was defined by the C_α-C_α distance of residues Q88 of CD and P318 of INS. From the evolution of the structure during 70-ns MD simulations, it is evident that the two domains, INS and CD, moved apart by ∼25 Å in the absence of any crowding agents (Fig. 4 b). On the other hand, smaller domain displacement (<15 Å) was observed in the presence of crowding agents, indicating that crowders favored the “closed” conformation.

Figure 4.

(a) Structures of Ec ProRS in “closed” and “open” states. The editing domain residues (residues 313–322) and CD residues (residues 84–93) are shown in green color; (b) the evolution of the distance (in Å) between C_α atoms of Q88 and P318 over the simulation period of 70 ns in three systems: dilute condition, dextrose (200 g/L), and sucrose (200 g/L); (c) the backbone root mean-square deviation (RMSD) of each frame was calculated from their starting conformation over the simulation period of 70 ns for the three systems.

The ratio of “open” to “closed” conformations was measured from the MD trajectories (Table 2). The conformations were defined as “closed,” in which the interdomain distance (i.e., Q88(C_α)-P318(C_α)) was ≤15 Å. Analysis of the trajectories revealed that the “open” conformation is favored in the dilute condition, whereas the “closed” conformation was predominant in the presence of dextrose and sucrose (Table 2). The plot of Q88(C_α)-P318(C_α) versus simulation time clearly demonstrated that the compact conformation was preferred in sucrose and dextrose and the extended conformation in dilute condition (Fig. 4 b). As discussed in the subsequent section, similar observations were made, in which the addition of cosolutes like sucrose and dextrose has increased compactness and the stability of the protein structure (29, 82, 83).

Table 2.

Change in Conformational Ensemble Along the 70-ns Trajectory of MD Simulations

Interdomain Separation in Terms of Q88(Cα)···P318(Cα) Distance	Conformations	Dilute	Dextrose (200 g/L)	Sucrose (200 g/L)
<15 Å	Closed	24	93	94
>15 Å	Open	91	7	6

The same concentration of sucrose and dextrose was used in MD simulations and kinetic assays. Thus, the changes in the simulated dynamics of the substrate-free enzyme are expected to be correlated to the kinetic assays and hence could provide insights into the catalytic changes. The shift in the conformational equilibrium from “open” to “closed” state in the presence of dextrose and sucrose correlates well with the increased binding affinity for the substrate. As the compact structure is more prevalent, certain segments of the editing domain remained closer to the CD in the presence of both crowders (Fig. 4 a). Thus, the substrate is expected to be more tightly bound, which is consistent with the decreased K_M as observed experimentally from the kinetic assays.

The plot of backbone RMSDs versus simulation time is shown in Fig. 4 c. In each case, the RMSDs were computed from the optimized starting structure of the solvated dimeric enzyme. The computed RMSDs indicate a strong damping effect on backbone fluctuation due to both crowders. The evolution of RMSDs demonstrated that the backbone RMSD fluctuated by only 3–4 Å in presence of crowders as compared with the dilute solution in which a 7 Å fluctuation was noted. As the backbone fluctuation is predominantly due to the motion of INS domain, the damping in backbone fluctuation correlates well with the “open” to “closed” population shift in the presence of dextrose and sucrose.

Preferential exclusion of dextrose and sucrose

As the name suggests, the preferential interaction coefficient is a measure of the preference of cosolute for the protein surface (local) relative to the bulk (84). The computed preferential interaction coefficients, Γ_2,3, for 1.11 M (200 g/L) dextrose and 0.58 M (200 g/L) sucrose are −16.00 mol dextrose/mol Ec ProRS and −7.73 mol sucrose/mol Ec ProRS, respectively (Table 3). A literature search revealed that these values are quite similar to the experimentally measured values for several proteins in sucrose (29) and dextrose (83). The negative values of the coefficients indicate preferential exclusion of both cosolutes (sucrose and dextrose) from the protein surface, which is consistent with the thermodynamic mechanism postulated by Timasheff (28). The computed μ_2,3 values (Eqs. 5 and (6a), (6b)), which represent the molar free energy change of the protein per mole of the cosolute, are −6.8 and −6.3 kcal for dextrose and sucrose, respectively (Table 3). The observed similar values of μ_2,3 in dextrose and sucrose indicate that these cosolutes induce a similar stability in Ec ProRS.

Table 3.

Preferential Interaction Coefficient and the Partial Molar Chemical Potential of the Enzyme in the Presence of Sucrose and Dextrose

Cosolutes	rlocal (Å)	rbulk (Å)	Γ2,3 (mol cosolute/mol Ec ProRS)	μ2,3 (kcal/mol cosolute/mol Ec ProRS)
Dextrose	6.0	>20	−16.00 ± 0.03	−6.82 ± 0.01
Sucrose	6.0	>20	−7.73 ± 0.02	−6.27 ± 0.02

The preferential interaction coefficient Γ_2,3, was computed using Eq. 4 and the last 2 ns of the MD simulation data. The μ_2,3 was computed from Γ_2,3 using Eq. 5 as discussed in the Materials and Methods.

Impact on SASA

The cumulative average of SASA versus simulation time indicates a decrease in SASA of Ec ProRS for both dextrose- and sucrose-containing systems, relative to the dilute solution (Fig. 5 a). SASA decreased by ∼2% in dextrose (484 Å²) and ∼4% in sucrose (1111 Å²) as compared with that in dilute condition. As discussed by Kendrick et al. (85), the loss of sugar molecules from the protein surface causes an increase in the chemical potential of the protein. The chemical potential of a protein is directly proportional to its SASA, and following the Le Chateliar’s principle, the system counters this change by decreasing its SASA, thus favoring the compact conformation of the protein. Therefore, the decrease in SASA observed in this study indicates a compact structure of the Ec ProRS due to cosolutes, which is consistent with the results from previous reports (86, 87). Taken together, the reduction of SASA, the negative preferential interaction coefficients, and the conformational redistribution (Table 2) are the direct thermodynamic consequences caused by the preferential exclusion of cosolutes from the protein surface (61, 63).

Figure 5.

The effect of crowding agents on “soft interactions”: (a) cumulative average of the overall SASA (Ų) of the monomeric Ec ProRS; and (b) location of tryptophan residues shown as cyan beads on WT Ec ProRS monomer. The monomer is shown in “cartoon,” and the three distinct domains of Ec ProRS are labeled as INS (editing domain), catalytic domain (CD), and anticodon-binding domain; (c) computed SASA for tryptophans in the dilute condition (blue), 200 g/L dextrose (green), and 200 g/L sucrose (orange) are shown as histograms. All SASA were tracked every 10 ns of the simulation.

SASA of tryptophan residues

These studies also noted some increase in SASA for a few tryptophan residues. As mentioned earlier, there are five tryptophan residues in each subunit of the dimeric Ec ProRS; the locations of these residues are displayed in Fig. 5 b. The SASA of each tryptophan residue was calculated for every 10 ns (of the total 70 ns MD simulation) in three different conditions - dilute, dextrose, and sucrose (Fig. 5 c). The SASA of W375 of INS domain exhibited a significant change; the SASA increased from ∼25 Å² in the dilute to ∼70 and 90 Å² in the dextrose and sucrose, respectively. In contrast, a decrease in SASA was observed for W86 in the presence of crowding agents (Fig. 5 c). The alteration in SASA values confirmed the crowder-induced local conformational change of these residues, which is indicative of “soft interaction” between the crowders and these tryptophan residues. Although the effect of sucrose and dextrose on W375 in terms of the exposure to the solvent is similar, the reduction in Trp fluorescence intensity is not. This is because of the difference in the intrinsic nature of fluorescence quenching ability of sucrose and dextrose (30 and 2.5% decrease in Trp fluorescence intensity in the presence of sucrose and dextrose, respectively), as discussed earlier.

To further examine the effect of “soft interactions” between the crowder molecules and protein side chains, RDF of water, dextrose, and sucrose with respect to the two tryptophan residues, namely W86 and W375, were computed. These residues have higher SASA values (Fig. 5 c). The RDF plots show the probability of finding water or crowder molecules within the spherical region surrounding the tryptophan residue (Fig. 6, a and b). The center of the sphere represents the geometric average of all atoms of the respective tryptophan residue. The RDF plots revealed a flat curve for water but distinct peaks for both dextrose and sucrose, which indicates a higher probability of finding crowder molecules within a certain distance of the tryptophan (W86 and W375), relative to water. It is to be noted that the water molecules in these simulations greatly outnumbered the crowder molecules—for each dextrose molecule, there were ∼38 water molecules, whereas sucrose-to-water molecule ratio was ∼1:81. This observation coupled with the increased exposure of the tryptophan to the solvent (i.e., SASA) confirmed the presence of “soft interactions” between tryptophan and crowder molecules (dextrose and sucrose).

Figure 6.

RDF of water, dextrose, and sucrose molecules with respect to (a) Trp86 and (b) Trp375 of Ec ProRS, computed using the last 60 ns MD simulation data. The x axis represents the distance of separation between the center of mass of the tryptophan residue and the center of mass of water, dextrose, and sucrose in dilute, 200 g/L dextrose, and 200 g/L sucrose, respectively.

Potential energy surface calculation to probe “soft interaction” between tryptophan and sugar molecules

The higher RDF and SASA values of W375 and W86 indicate a stronger interaction between the tryptophan and dextrose/sucrose than water. The dispersion interaction or CH/π interaction between carbohydrates with aromatic amino acid residues have been reported earlier (88, 89). To examine if dispersion interaction prevails, the potential energy surfaces (PES) of the three systems were computed using approximate quantum chemical formalism (SCC DFTB-D module of CHARMM program) (44, 45, 65). The PESs depict the strength of the interaction of the indole ring (of the tryptophan) with water, dextrose, and sucrose in terms of the depth of the potential energy well. The calculated potential energies are plotted against the distance between the centroid of the benzene ring of indole and the centroid of the pyranose (sugar) molecule. In the case of water, the oxygen atom is used in the place of pyranose. The comparison of potential energy wells of the three systems is shown in Fig. 7. The lowest point of the potential energy well was found to be only −0.2 kcal/mol occurring at 3.6 Å for water. In contrast, much stronger interactions were noted for both dextrose and sucrose systems (−6.2 kcal/mol and −6.9 kcal/mol occurring at 3.1 Å, respectively; Fig. 7). These results indicate ∼30−35 times stronger interactions between the sugar ring of dextrose/sucrose and the indole ring of the tryptophan compared to the interaction between water and tryptophan. Therefore, the PES calculations revealed a molecular basis of how “soft interactions” are playing a role in altering the conformational equilibrium of Ec ProRS.

Figure 7.

Gas-phase potential energy of interaction between the tryptophan with water, dextrose, and sucrose, computed using approximate quantum chemical formalism (SCCDFTB-D module of CHARMM program). The intermolecular distance is the distance between the centroid of the benzene ring of indole and the centroid of the pyranose (sugar) molecule. In the case of water, the oxygen atom is used in the place of pyranose.

Crowder-induced alterations in coupled-domain dynamics

To explore the impact of crowding agents on the collective movements of the protein backbone, the essential dynamics of Ec ProRS in dilute and crowded conditions were examined by computing their principal components of motions. In particular, the dynamical changes in the first three principal components were examined for each system. The PCA revealed an alteration in the collective domain dynamics in the presence of dextrose and sucrose (Fig. 8 a). The editing domain and CD move in an anticorrelated manner. However, their motions were restricted in dextrose and sucrose (Fig. 8 a, left to right). In particular, the collective dynamics of the editing domain and the key secondary elements at the interface of the editing domain and the CD, which includes proline-binding loop (residues 195–210), the helix-loop β (residues 305–322) and the helix (residues 84–93) were altered in the crowded environment (Fig. 8 b). An increased compactness surrounding these structural elements, which comprise the active site pocket of Ec ProRS, was observed in the presence of dextrose and sucrose. Earlier, we reported that the coupled dynamics between distant structural elements are important for efficient catalysis by Ec ProRS (6, 7). In this study, a change in the dynamics of proline-binding loop was noted. As observed in the dilute condition, the movement of proline-binding loop (red to blue ribbon in Fig. 8, a and b) was more pronounced compared with that in the presence of dextrose and sucrose. Other structural elements also moved to a lesser extent because of crowding, creating a shift in the conformational equilibrium and thus impacting the enzyme function. The decrease in V_max can also be rationalized as the limited opening of the active site pocket leading to a reduced frequency of effective collisions with substrates as well as decreased product release. Moreover, when considering the dimeric Ec ProRS, the separation between INS domains of the two subunits were also altered in the presence of crowders (Fig. 9). The distance between the two INS domains decreased considerably in the presence of dextrose and sucrose, which could also alter the substrate binding in the presence of crowders. Therefore, results of MD simulations support the changes in kinetic parameters in the presence of crowders.

Figure 8.

(a) PCA of the dilute, dextrose (200 g/L), and sucrose (200 g/L) systems using the 70 ns simulation data. The red-colored ribbons depict the position of the backbone C_α atoms of Ec ProRS at the beginning, and the blue-colored ones represent the same at the end of the 70-ns simulation. (b) Conformational change was observed during 70-ns MD simulations of Ec ProRS in dilute, 200 g/L dextrose, and 200 g/L sucrose. The monomeric structure with the secondary structural elements surrounding the catalytic pocket, highlighted in green, is shown on the left. The displacements of these secondary elements during 70-ns simulations are shown on the right, where red and blue colors represent the starting and the ending conformations, respectively.

Figure 9.

Varying degrees of compactness in the dimeric structure of Ec ProRS in crowded environments: (a) dilute; (b) 200 g/L dextrose; and (c) 200 g/L sucrose.

Conclusions

This study probed the effect of crowding on the function of Ec ProRS using experimental and computational methods. Results indicate that crowding impacts the conformational equilibrium and dynamics of the enzyme, which leads to an alteration in its catalytic function. The concentration, size, and the chemical nature of crowders have a noticeable impact on Ec ProRS function. Analysis of these variables suggests that crowders influence enzyme function through a combination of “hard interactions” (steric hindrance) and “soft interactions” (chemical) with the protein. Simulated systems with sucrose and dextrose further demonstrated that the molecular mechanism of crowding involves primarily preferential interactions between cosolutes and the protein that is also supported by the results of kinetic and fluorescent studies.

A molecular-level picture of the key crowding mechanism emerges from this study. Kinetic studies using variable-sized PEG molecules demonstrated that smaller PEG molecules had a stronger impact on the product formation, indicating a strong interplay between “soft interactions” and kinetics (Fig. 3). MD simulations and subsequent PCA confirmed that the addition of small-sized crowders like sucrose and glucose resulted in an alteration of dynamics and a conformational shift in favor of the compact or “closed” state (Fig. 4, a and b; Table 2). The damped backbone fluctuation (Fig. 4 c) and reduction in protein’s overall SASA (Fig. 5 a) unambiguously established the favorability of the “closed” conformation in the presence of dextrose and sucrose, which is also consistent with the tighter proline binding and reduced V_max (Fig. 2, a and b). Scrutiny of protein-solvent and protein-crowder interactions further revealed that crowders are preferentially excluded from the protein surface, causing the conformational ensemble to alter. This is evident from the negative values of the computed preferential interaction coefficients for sucrose and dextrose (Table 3), which are similar to those observed experimentally for these crowders (29, 83). The exclusion of sucrose and glucose resulted in ∼ 6–7 kcal/mol decrease of the protein’s chemical potential (Table 3), providing clear thermodynamic evidence of the stability of the compact structure.

This study, however, found a perceptible reduction in fluorescence intensity and a shift in the λ_bcm in the presence of crowders (Table 1), which are comparable to those of a free tryptophan in the presence of these crowders. This indicates interactions between crowders and tryptophans on the protein surface. This is also supported by the simulated protein systems in the presence of dextrose and sucrose, which exhibited an increase in SASA for two tryptophans, W86 and W375, located close to the surface (Fig. 5, b and c). Furthermore, RDFs of these two tryptophans (Fig. 6, a and b) indicated a higher affinity for each colsolute when compared with water. Finally, the approximate quantum chemical computations demonstrate that the interaction between the indole ring of the tryptophan and dextrose/sucrose is more attractive compared to the interaction with water (Fig. 7).

The “in-cell” biophysical research has grown steadily in recent years (8, 27, 90, 91). Although information regarding the impact of molecular crowding is still emerging, there remains many unanswered questions about its core molecular mechanism of action on the structure and function of modular proteins. This study has added additional insights into the mechanism of crowding on the function of modular enzymes like Ec ProRS. It has demonstrated that crowding agents like dextrose and sucrose alter the conformational and dynamical properties of multidomain Ec ProRS through “soft interactions.” Taken together, this study is expected to serve as a valuable guide for predicting crowding-induced changes in the conformation and dynamics of modular enzymes in the cellular environment. Furthermore, the study highlights the significance of considering the molecular crowding effect, when probing the catalytic mechanism of modular enzymes.

To our knowledge, this is the first report of a crowder-induced shift in conformational ensemble in modular ProRS, an important member of the AARS family. AARSs have emerged as important targets for anti-infective drug development because of their essential role in protein biosynthesis (92, 93, 94, 95). The documented effects of molecular crowding on the conformation and function may have future implications in drug design. This study suggested that the screening of potent drug molecules for pathogenic enzymes requires a thorough investigation of the impact of molecular crowding on the conformational ensemble and catalytic function of these enzymes. Finally, this study has exemplified the need for a multifaceted investigatory approach involving kinetic, spectroscopic, and computational techniques to elucidate the overall effect of molecular crowding on enzyme structure and function. Further studies with crowding agents mimicking the intracellular biomolecules and metabolites should be carried out by employing both experimental and computational approaches as it provides a greater insight into the structure-dynamic-function relationship of multidomain enzymes in the intracellular environment.

Author Contributions

L.M.A., H.L.S., and S.H. designed and performed the wet-lab experiments. L.M.A. and R.J.A. performed fluorescence studies. R.J.A., Q.H.H., and S.B. designed and performed the computational experiments. L.M.A., R.J.A., S.H., and S.B. analyzed the data and wrote the manuscript.

Editor: Margaret Cheung.