Substrate Dynamics in Enzyme Action: Rotations of Monosaccharide Subunits in the Binding Groove are Essential for Pectin Methylesterase Processivity

Abstract

The dynamical behavior of biomacromolecules is a fundamental property regulating a large number of biological processes. Protein dynamics have been widely shown to play a role in enzyme catalysis; however, the interplay between substrate dynamics and enzymatic activity is less understood. We report insights into the role of dynamics of substrates in the enzymatic activity of PME from Erwinia chrysanthemi, a processive enzyme that catalyzes the hydrolysis of methylester groups from the galacturonic acid residues of homogalacturonan chains, the major component of pectin. Extensive molecular dynamics simulations of this PME in complex with decameric homogalacturonan chains possessing different degrees and patterns of methylesterification show how the carbohydrate substitution pattern governs the dynamics of the substrate in the enzyme’s binding cleft, such that substrate dynamics represent a key prerequisite for the PME biological activity. The analyses reveal that correlated rotations around glycosidic bonds of monosaccharide subunits at and immediately adjacent to the active site are a necessary step to ensure substrate processing. Moreover, only substrates with the optimal methylesterification pattern attain the correct dynamical behavior to facilitate processive catalysis. This investigation is one of the few reported examples of a process where the dynamics of a substrate are vitally important.

Introduction

MD is of crucial importance in the majority of biological processes. Advances in theoretical (1–7) and experimental (8–11) methods are constantly providing improved means to address on an atomic scale the MD of even larger molecules, on ever longer timescales. It is evident that conformational fluctuations influence key biological events, including protein-protein interactions and aggregation (12). Local and global protein dynamics have been implicated in recognition/binding and product release processes (13–18), including the case of ribonuclease A (19), and are therefore considered key determinants of enzymes catalytic activity. However, this model remains a matter of debate as a number of other studies draw opposite conclusions on the influence of collective protein dynamics (20–24).

In this study of the pectin methylesterase from the plant pathogen bacterium Erwinia chrysanthemi (Ec-PME), also known as Dickeya dadantii (25), we show how not only protein motions but also substrate dynamics can be crucial in optimizing complex enzymatic processes. PMEs are enzymes that catalyze the deesterification of O6 methylesterified α-D-galactopyranosyluronic acid residues (D-GalpAO6Me) that are part of the HG chains of pectic polysaccharides in the plant cell wall. Plant PMEs are crucial in many physiological processes occurring in plants including the stiffening and extension of the cell wall (26,27), cell division and seed germination (28,29), the determination of leaf polarity (30), and fruit ripening (31–33). Interestingly, PMEs are also vital participants in host-pathogen infections and are expressed by both bacteria and fungi that use them to breach the plant cell wall through an uncontrolled demethylesterification of the pectic polymers. We have used MD simulations to examine the effects of methylesterification patterns on substrate and protein dynamics for HG oligosaccharides in complex with Ec-PME.

Previous structural studies of PME proteins, including Ec-PME (34), tomato, carrot, and another bacterial PME (35–37) as well as bacterial and fungal pectate lyases (38–41), endogalacturonases (42–45), and rhamnogalacturonase (46) have revealed a triple β-helix fold featuring a functional cleft for the binding of HG chains. Additionally, x-ray structures of Ec-PME in complex with partially methylesterified galacturonides have provided clues to the mechanism by which bacterial PMEs process their substrates (47). To date, only the interactions with small HG chains (hexasaccharides) have been analyzed (Fig. S1 in the Supporting Material), due to both the difficulty in producing longer HG samples with controlled patterns of methylesterification and to the challenge of obtaining crystals of protein samples bound to polysaccharides. The structural studies with HG hexasaccharides revealed that the enzyme binding groove has 10 subsites, labeled −5 to +5, where subsites −1 to −5 are arranged in the direction of the nonreducing end of the HG chain and subsites from +1, the active site of methyl cleavage, to +5 are in the direction of the reducing end of the HG chain. Subsites −1 to −5 preferentially bind unsubstituted α-D-galacturonic acid residues of HG, whereas subsites +1 to +5 preferentially bind methylesterified residues. Interestingly, the monosaccharide subunits of the hexamers adopt an alternating (or mutually trans) orientation of the methylester or carboxylate groups, as a consequence of the irregular 2₁-helical conformation. In particular, for the central part of the binding grove (i.e., subsites −2 to +3), a complete helix turn is achieved every two residues, however, consecutive monosaccharides are not precisely oriented at 180° (Fig. S2).

In this investigation, we employed MD simulations to probe in silico the interactions between Ec-PME and HG decasaccharides, initial models that are formed by merging the coordinates of two enzyme-bound HG hexamers (PDB codes 2nsp and 2nt9, Fig. 1 a) (47). Our study provides evidence that i), the substrate dynamics in the binding groove are strongly influenced by the pattern of methylesterification of the HG chain, ii), the ability of the enzyme to demethylesterify HG chains processively is tuned toward specific methylesterification patterns, and iii), after demethylesterification, rotations of HG monosaccharide subunits about glycosidic bonds orient the next methylester group (most importantly that at site +2) for the subsequent cycle. The results suggest that these rotations are due to the repulsion between the nascent HG carboxylate group and the carboxylate of Asp-199 in the active site, although, a specific pattern of methylesterification is required to achieve the optimal substrate dynamics for relatively facile rotations. Concerted rotations of this kind, which preorients the next methylester group for demethylesterification, solves, in part, the topological problems associated with achieving consecutive contiguous catalytic events along the polymer when the orientation of the carboxylate or methylester groups, as noted previously, has been observed to alternate (47) (Fig. S1 and Fig. S2).

Figure 1.

Docking of HG oligomers with Ec-PME. (a) Modeling the decameric HG chimera from x-ray structures of HG hexamers (PDB codes: 2nsp and 2nt9). The RMSD of C_α atoms of the protein in the two structures is 0.13 Å. A decameric HG chimera was obtained by merging the coordinates of the two hexameric chains that overlap in the subsites −1 and +1. The plots at the foot of the molecular diagrams present schematically the occupation of subsites observed in the x-ray structures and in the chimera model. (b) Representation of the methylesterification states investigated in the present MD simulations. A representative fully methylesterified chain (HM) is drawn. Abbreviations are: M, methylesterified subunits; U, unmethylesterified subunits; FM, fully methylesterified HG chain; HM, half methylesterified HG chain composed of unmethylesterified monosaccharides at binding subsites −5 to −1 and methylesterified monosaccharides at binding subsites +1 to +5; FU, fully unmethylesterified HG chain; FXM, fully methylesterified HG chain with a demethylesterified HG subunit binding at the subsite +1. HXM, half methylesterified HG chain (HM) with an additional demethylesterified HG subunit binding at subsite +1.

The results underline the importance that the conformational dynamics of the substrate can have in determining the functional interactions in enzyme-substrate complexes (48–52), both in general, and in particular in the enzymatic modifications of pectic polysaccharides by PME.

Methods

MD Setup

MD simulations of the Ec-PME both in the free state and in complex with homogalacturonan decasaccharides were performed using the GROMACS package (53) by explicitly simulating all protein and carbohydrate atoms and explicit TIP4Pew (54) water molecules. Initial structures of the HG decasaccharides were obtained by combining the coordinates of two x-ray crystal structures of HG hexasaccharides bound to Ec-PME in which the subsites −5 to +1 and −1 to +5 are occupied (PDB codes: 2nsp and 2nt9, respectively, Fig. 1 a). In these structures the two hexamers show overlapping monosaccharides at subsites −1 and +1, which closely superimpose (Fig. S1 and Fig. 1 a), and thereby allow for an accurate reconstruction of the coordinates of an HG decamer docked in the Ec-PME binding groove. For the overlapping −1 and +1 monomers, the coordinates from the 2nsp structure were used. The various patterns of HG methylesterification analyzed were obtained by adding or removing methyl groups from the HG decamer. Within the catalytic residues, Asp-178 was protonated as suggested by Fries et al. (47). The systems were accommodated in cubic boxes of 6.28 × 6.28 × 6.28 nm³. Oligosaccharide chains were treated with the GLYCAM06 force field (55,56) and protein atoms with the AMBER03 force field (57,58). This combination of force fields is highly effective for studying protein-carbohydrate interactions as shown by Kasson and Pande (59). When necessary, Cl⁻ or Na⁺ ions were added to neutralize the system (Table S1). The system was energy minimized and subsequently the solvent was equilibrated for 100 ps during which the protein and decasaccharide atoms were restrained to the energy-minimized Cartesian coordinates by a 1000 kJ·mol⁻¹·nm⁻² spring constant. Subsequently, unrestrained MD simulations were carried out for 50 ns using periodic boundary conditions and adopting LINCS as a constraint algorithm (60). Electrostatic interactions were treated by using the particle mesh Ewald method (54). Simulations were performed in the NPT ensemble by coupling the system to weak external pressure and temperature baths (1 atm and 300 K), with coupling constants of 0.1 and 1.0 ps, respectively. The simulations were analyzed using in-house programs and analysis tools from the GROMACS package (53). To achieve a statistically relevant sampling of the conformational dynamics of the various protein-decasaccharide complexes, each system was simulated six times for 50 ns (Table S1) using different random seeds to generate initial velocities. Consistency between the six independent simulations was verified for all the reported analyses.

Time autocorrelation function of the dihedral angles

The dihedral configuration of the six consecutive dihedral angles defined by the glycosidic bonds linking HG residues docked in the enzyme subsites +1 to +3 has been examined and the time autocorrelation function, P(τ), adopted as a representation of the dynamical behavior of HG chains, has been calculated. The expression is a function of the characteristic time τ and is given by

$$P\left(\tau \right)=\sum _t\delta \left(\Xi \left(t\right),\Xi \left(t+\tau \right)\right),$$ (1)

where the function delta, δ(Ξ(t), Ξ(t + τ)), assumes values of 0 or 1 depending on whether or not a set of six dihedral angles differs or not at the time t and t + τ. A tolerance of 30° has been used in the comparison of angles.

$$P_{fit}\left(\tau \right)=A_1{e}^{-\frac{\tau }{{\tau }_1}}+A_2{e}^{-\frac{\tau }{{\tau }_2}}.$$ (2)

The autocorrelation curves were fitted with a double exponential decay that accounts for a fast (τ₁) and slow (τ₂) relaxation (Eq. 2).

Results

Modeling of HG decamers in the Ec-PME binding groove

In this work, to our knowledge, the dynamics of HG decasaccharides bound to the Ec-PME were studied for the first time by combining the coordinates of two x-ray crystal structures of HG hexasaccharides bound to Ec-PME (PDB codes: 2nsp and 2nt9, respectively, Fig. 1 a). First, the dynamics of three different HG chains, FM, FU, and HM, were investigated (Fig. 1 b). The latter is composed of unmethylesterified monosaccharides at subsites −1 to −5 and methylesterifed monomers from subsites +1 to +5. HM has been proposed to be the optimal substrate for Ec-PME (47). Second, to investigate the ability of the enzyme to perform processive catalysis along HG chains, we simulated the systems FXM and HXM, which were respectively composed of the FM and the HM decasaccharides, but demethylesterified at subsite +1, to mimic decasaccharides freshly demethylesterified at the active site (Fig. 1 b).

The dynamics of HG oligomers as a function of their degree of methylation

The analysis of the simulations indicates that the pattern of methylesterification strongly affects the fluctuations of HG chains within the binding cleft of Ec-PME. In particular, the RMSF profiles, calculated for the carboxyl moieties of HG subunits (Fig. 2 a), suggest that the dynamics of the HG chains is inversely related to their degree of methylesterification. Indeed, the FM shows in general low RMSF values, except at subsites ±4 and ±5, in which a reduced number of interactions are shown (Fig. S3), as consistently observed in the x-ray crystal structures of PME in complex to HG hexasaccharides (47). By reducing the level of methylesterification (i.e., from FM to HM to FU), however, the overall RMSF values increase, with the FU having RMSF values in the central regions comparable to those at the chain extremities. Moreover, FU simulations sampled dissociation events from the enzyme cleft (Fig. S4), which is in line with this oligosaccharide not being a substrate for Ec-PME. It is noteworthy that on comparing the dynamics of FM and HM chains, different values of the RMSF are found in subsites containing the same local methylesterification pattern (+1 to +5), suggesting that local dynamics are transmitted along the HG chains with residues at subsites −5 to −1 influencing the dynamical behavior of monosaccharides docked at subsites at +1 to +5. The conformational properties of the HG chains were also investigated by calculating the autocorrelation functions (Eq. 1) of the six dihedral angles for the glycosidic bonds that link the HG subunits docked in the subsites +1 to +3 (Fig. 2 b). These were fitted with double exponential decays (Eq. 2, Fig. 2 b) that, in agreement with the RMSF profiles, show the fastest relaxation for FU (τ₁ = 0.025 ± 0.01 ns, τ₂ = 0.84 ± 0.03 ns), intermediate relaxation for HM (τ₁ = 0.038 ± 0.002 ns, τ₂ = 1.87 ± 0.03 ns), and longer relaxation for FM (τ₁ = 0.087 ± 0.005 ns, τ₂ = 2.12 ±0.02 ns). As seen with the RMSF the dynamics of the monosaccharide subunits at the reducing end of the HG decasaccharide are affected by the methylesterification state at the nonreducing end of the chain. Rotations around the glycosidic bonds did not perturb the ring conformations in our simulations, with all monosaccharides remaining in ⁴C₁ chairs throughout all the simulations (Fig. S5).

Figure 2.

Conformational dynamics of different methylesterified homogalacturonan decamers in complex with Ec-PME. (a) RMSF of the carboxylate groups of FM (blue), HM (red), and FU (green) decamers in complex with Ec-PME. The RMSF values have been averaged over the six simulations performed on each complex and the standard deviations are reported by the bar graphs. On top of the graph a representative FU chain is shown. (b) Autocorrelation function calculated for FM (blue), HM (red), and FU (green) decamers. The function has been fitted (black lines) using a double exponential equation (Eq. 2). Fitting parameters are FM, τ₁ = 0.087 ± 0.005 ns, τ₂ = 2.115 ± 0.02 ns; HM, τ₁ = 0.038 ± 0.002 ns, τ₂ = 1.869 ± 0.03 ns; FU, τ₁ = 0.025 ± 0.01 ns, τ₂ = 0.842 ± 0.03 ns.

When analyzing the conformations of the protein, we found a significant reduction in RMSF values calculated on the Cα atoms of unbound and bound states. However, within the bound states, different methylation patterns of HG chains did not affect the enzyme fluctuations (Fig. S6). The unperturbed enzyme dynamics in FM, HM, and FU suggest that the observed HG fluctuations are determined by the interactions between the carbohydrate and the enzyme, and that there is weak coupling of motions between HG chains and the enzyme backbone.

Three key groups of interactions govern the structural dynamics of HG chains in the Ec-PME binding groove

The observed differences in the dynamical behavior of FU, HM, and FM chains suggest that the degree of methylesterification has a key role in stabilizing the PME-substrate complex, most likely mediated by hydrophobic interactions with the protein. Among the intermolecular interactions of methylesterified HG residues at the reducing end (i.e., at subsites +1 to +5), two hydrophobic pockets on the protein surface, at subsites +1 and +3, appear to be key for anchoring the HG chain to the enzyme. It is likely that the highly dynamical behavior of the FU chain primarily arises from the lack of methylesterified monosaccharides docked in the +1 and +3 subsites. For subsite +1, which features the catalytic residues Gln-177, Asp-178, and Asp-199, the simulations showed tight hydrophobic interactions between the side chains of Phe-202, Trp-269, and Ala-233 and the methylester group of the HG monosaccharide subunit (Fig. 3 a), as also reflected by the trend of the glycosidic angles formed between the residues +1 and +2 (Fig. S7). In addition to stabilizing the complex through hydrophobic interactions, the methylester group at the subsite +1 also mitigates the electrostatic repulsion between the carboxyl moiety of the carbohydrate and the side chain of Asp-199. For FM and HM, which dock a methylesterified monosaccharide at position +1, a narrow distribution of the distances is observed because of this interaction, whereas for FU a wide distribution of distances is observed (Fig. 3 a). Thus, although the methylesterification of the monosaccharide subunit occupying subsite +1 is fundamental in maintaining the right orientation for catalysis, repulsive charge effects and the lack of hydrophobic interactions upon demethylesterification are expected to destabilize the docking of the product HG residue in the +1 subsite. In fact, this might be seen as a prerequisite for substrate movement and subsequent catalysis on the consecutive monosaccharide.

Figure 3.

Key interactions influencing the dynamics of HG decamers bound to Ec-PME. The histograms show the distance distributions between the carboxylate/methylester groups of HG monomers and specific amino-acid carbons in the subsites +1 (a), +3 (b), and −2 (c) for FM (blue bars), HM (red bars), and FU (green bars). In the top panels, representative structures of the oligosaccharide monomers and the hydrophobic pockets at the three different subsites are shown. The distances calculated are between the O6B carboxylic oxygen of each HG subunit and W269:CZ (subsite +1), V198:CG (subsite +2), and F202:CZ (subsite −2). The reported frequencies are the cumulative results from six independent simulations performed on each system.

Along the binding groove, a second tight interaction occurs between Val-198, Val-227, and Tyr-230 and a methylesterified HG monomer docked at the subsite +3 (Fig. 3 b). Different from subsites +1 and +3, which show direct anchoring intermolecular interactions, the inherent dynamics of the monosaccharide subunit at subsite +2 are not affected by the methylesterification state and mainly reflect the dynamics of the flanking HG residues.

In addition to the direct interactions at +1 and +3, the simulations revealed tight hydrophobic contacts at subsite −2. These interactions, which are consistently observed in all replicate simulations, are key to the stabilization of docking of the FM chain at the nonreducing end (Fig. 3 c) and occur between a methylester group of the monosaccharide residue and the side chains of residues Thr-109, Ala-110, Phe-202, Tyr-158, and Tyr-181. The HM chain, which lacks methylesterification at subsite −2, is therefore not able to establish these tight interactions, resulting in the increased dynamics observed when compared with the FM chain. The crystallographic studies of HG hexasaccharides in complex with Ec-PME did not include any oligosaccharide bearing a methylesterified residue in position −2 (47). Consequently, only electrostatic interactions between the unmethylesterified HG monosaccharide at subsite −2 and the main-chain amide of Ala-110 or with Thr-109 were described.

Overall, our analysis suggests that highly methylated HG chains adopt reduced dynamics, in part because of the more restricted space within the Ec-PME binding groove arising from the methyl substituents and in part because of several specific HG-protein interactions, in particular the hydrophobic interaction at the subsite −2.

Structural dynamics of HG post demethylesterification are of key importance for the processive action of the enzyme

Although the catalytic mechanism exploited by Ec-PME to achieve demethylesterification is well understood (47), the processivity exhibited by Ec-PME warrants further attention. Such processivity implies that, after the demethylesterification of a monosaccharide docked at subsite +1, the enzyme-substrate complex rearranges without dissociation of the HG chain, and the monosaccharide that had resided in subsite +2 moves into the active site (+1 subsite) to allow for a new cycle of catalysis to take place. Analysis of the hexasaccharide conformations collected by the x-ray crystallographic study (47) suggests the occurrence of a topological problem when a simple linear processive mechanism is considered for successive methylester groups. As the carboxylate or methyl galacturonate groups adopt an alternating (or mutually trans) orientation with respect to the active site (Fig. 1 and Fig. S1), a simple linear sliding motion (or a straightforward biased random walk) of the HG chain along the binding groove (or of the protein along the polysaccharide) would not allow for the correct positioning of the methylester that previously resided at subsite +2 in the appropriate conformation at the +1 subsite for a subsequent catalytic event. Moreover, as HG chains adopt an irregular 2₁-helix conformation in the central region of the enzyme’s binding groove (61–63), a rigid rototranslation would not be sufficient to position the polysaccharide in the right orientation for the next catalytic cycle (Fig. S2). As a result, more complex conformational rearrangements of the HG chain are required from one catalytic cycle to the following.

To study the ability of the enzyme to catalyze consecutive reactions along the polymer without dissociating from the chain, we sampled the dynamics of the FM and HM decamers carrying demethylesterified HG subunits at the catalytic subsite +1, respectively designated as FXM and HXM (Fig. 1 b). The conformations explored for FXM and HXM have been analyzed by monitoring the distribution of dihedral angles around the glycosidic bonds linking monosaccharide subunits docked at subsites +1 and +2 (Fig. 4), denoted here as ϕ_+1/+2 and ψ_+1/+2 (Fig. S8) (61). These angles have initial values of +147.4° and −90.8°, respectively. Although FM and HM maintain their initial conformations throughout the simulations, HXM and FXM show transitions leading to different ϕ_+1/+2 and ψ_+1/+2 values. Transitions within the ϕ/ψ space are very informative of the conformational properties of polysaccharides (64). In HXM and FXM, relevant transitions are promoted by the electrostatic repulsion between the carboxylate group in position +1 and the side chain of Asp-199, however, whereas both HXM and FXM are affected by this repulsion, they show different conformational behaviors. In particular, FXM adopts equilibrium values of +174.6° and −140.8° for ϕ_+1/+2 and ψ_+1/+2, respectively (Fig. 4). Conversely, HXM converts from anti-ϕ to syn-ϕ and from syn-ψ to anti-ψ (equilibrium values of +74.9° and −163.4° for ϕ_+1/+2 and ψ_+1/+2, respectively). Interestingly, in HXM ϕ_+1/+2 and ψ_+1/+2 are highly correlated, although showing no correlation in all other constructs. HXM also shows unique conformational behaviors around the glycosidic bonds connecting monosccharide subunits that bind at subsites +2 and +3 (Fig. 5), with ϕ_+2/+3 changing from a starting value of 70.0° to an equilibrium value of 124.1° and ψ_+2/+3 from −123.3° to −93.3°. For these angles, HM, FM, and FXM maintain the initial conformation. Interestingly, the equilibrium values of ϕ_+2/+3 and ψ_+2/+3 in HXM are close to the initial values of ϕ_+1/+2 and ψ_+1/+2, suggesting that residue at subsite +2 tends to adopt the initial conformation of residue +1.

Figure 4.

Distributions along ϕ_+1/+2 and ψ_+1/+2 glycosidic dihedral angles. These are calculated along the bond vectors linking the HG subunits at the subsites +1 and +2 (Fig. S8). The distributions are extracted from individual simulations of FM (a), HM (b), FXM (c), and HXM (d) bound to Ec-PME. Consistent results are found in all the other control simulations. FM and HM conformations remain close to the initial values of ϕ and ψ angles.

Figure 5.

Distributions along ϕ_+2/+3 and ψ_+2/+3 glycosidic dihedral angles. These are calculated along the bond vectors linking the HG monosaccharide subunits at the subsites +2 and +3 (Fig. S8). The distributions are extracted from individual simulations of FM (a), HM (b), FXM (c), and HXM (d) bound to Ec-PME. Consistent results are found in all the other control simulations. FM, FXM, and HM conformations remain close to the initial values of ϕ and ψ angles.

This conformational transition provides a key molecular step for the functional redocking mechanism necessary for a processive catalysis (Fig. S9). Indeed, the correlated rotations of the HG residues at subsites +1 and +2 suggest a pathway to rearrange appropriately the carbohydrate chain within the binding cleft for the subsequent catalytic cycle (Fig. S10). These correlated rotations are only observed in HXM, which shows the optimal conformational dynamics. On the contrary, FXM appears too rigid to allow these concerted motions to occur. Although the subsequent sliding motion of the HG chain, which would place the subunit at +2 into the active site at +1, is unlikely to be sampled in the timescales accessible to full atomic MD simulations, the rotations of the monosaccharides around the glycosidic bonds, which is crucial to reproduce the steric conditions necessary for a successive catalysis, was elucidated with an appropriate statistical significance.

Discussion

In addition to being of vital biological (65) and biotechnological (66) importance, pectin methylesterase enzymes represent an intriguing system on which to investigate the fundamental issue of the role of structural dynamics in complex biomacromolecular processes, in particular where the substrate is itself also a polymer. Previous crystallographic investigations of the Ec-PME in complex with hexameric oligosaccharides with different patterns of methylesterification have provided strong evidence for a preferential asymmetric specificity in the binding groove, where at one end of the enzyme cleft there is a preference for methylesterified residues and at the other unmethylesterified residues are preferred (47). Although strongly suggestive that HG chains are processed in a specific direction, a mechanistic understanding of how the relative motions of the binding partners facilitate the repositioning of the HG substrate in the active site has been notably lacking. Oligosaccharides might conceivably diffuse out from the binding groove of PME between each demethylesterification event (resulting in a so-called multiple chain mechanism and a random pattern of demethylesterification), or be processed via a mechanism of sliding within the binding cleft of the enzyme generating a limited number of contiguous demethylated monosaccharide subunits (multiple attack mechanism), or long demethylated portions along the chain (single-chain mechanism (SCM)). Experimental evidence from the study of PMEs from mung bean hypocotyl cell walls, suggests that a key factor in discriminating between a multiple attack mechanism or SCM is the pH of the solution (67), which is consistent with the importance of electrostatic interactions between the protein and the oligosaccharide demonstrated here.

A multiple chain mechanism might seem more likely because HG carboxylate or methylgalacturonate groups in the binding groove are observed in the crystal structures to alternate in orientation, such that these groups bound in the +5, +3, +1, −2, −4 sites face into the binding groove, whereas those bound in the intervening subunits face out to the solvent. With such an arrangement, processive activity is accompanied by a topological problem: a simple linear translation of the chain along the binding groove would not reproduce the steric conditions for a second catalytic event. Nevertheless, it has been experimentally shown that bacterial and plant PMEs are able to act processively by deesterifying HG chains in blocks (67). Our simulations of Ec-PME bound to HXM decasaccharide show concerted rotations of the nascent demethylesterified subunit at +1 and the methylesterified subunit at +2 that places these subunits in the correct conformation to slide into the −1 and +1 binding sites. These motions are also transferred to the monosaccharide at the site +3 that in turn partially reorients in the opposite direction, as can be seen from the changes in ψ_+2/+3 (Fig. 5) and from the structures at equilibrium (Fig. S10). The complete conformational rearrangement required for a new catalytic cycle in the SCM model would also include an n-1 sliding of the HG chain (e.g., with residue +2 sliding into the +1 site, etc.). As noted earlier, this sliding is unlikely to occur on the timescale of these simulations. We do not observe with statistical significance any coupling of the rotation about the glycosidic bond for the pair of HG subunits at the +1/+2 sites with sliding events. Therefore, although the rotations around the glycosidic bonds solve the topological problem associated with placing the HG subunit at +2 into the correct orientation for processing at site +1, dynamics that favor sliding of subunits of the HG chain from sites n to n-1 must occur at longer timescales than those required for rotations. To sample sliding motions (and associated rotations) of other HG subunits to reestablish the mutually trans orientation of HG subunits and methyl ester group in the binding groove of the protein, biased dynamics such as steered MD or metadynamics (4) methods could be used.

Moreover, the present MD investigation, which allows assessing the binding modes of a decasaccharide HG chain bound in the enzymatic binding cleft, revealed that the distribution of methyl substituents on the polysaccharide chain has a significant influence on the dynamics of the substrate in the protein-carbohydrate complex (Fig. 3). In specific subsites (i.e., −2, +1, and +3), tight hydrophobic interactions are established between the binding partners if the monosaccharide is methylesterified. These interactions provide key anchoring points that strongly influence the dynamical behavior of the decasaccharide in the binding groove. Our results reinforce the crystallographic evidence that the preferred specificity along the binding groove involves demethylesterified HG subunits at the nonreducing end (−1 to −5) and methylesterified subunits at the reducing end (HM). Indeed, the simulations suggest that the lower dynamics of FM compared to HM chains (Fig. 2) result from the additional hydrophobic interactions at the subsite −2 and the protein. In turn the increased dynamics, due to the demethylation at the subsite −2 in HM, appear to be important for the processive activity of the enzyme (Fig. 3 c). In particular, the analysis of the initial reaction products, FXM and HXM, suggests that, because of the interactions at subsite −2, FXM has additional constraints that reduce the likelihood of rotations necessary for the establishment of the steric conditions essential for a subsequent consecutive demethylesterification step. Conversely, the HM possesses the correct dynamical behavior to favor a processive SCM mechanism after the first demethylesterification event on the chain. Overall, these findings are consistent with experimental evidence of slower enzymatic kinetics measured when PMEs are incubated with highly methylesterified pectins (67,68) and suggest that the processive enzymatic activity of Ec-PME is sensitive to pectic substrates with different degrees and patterns of methylesterification, as has previously been suggested (69).

Conclusions

To our knowledge, the MD study of Ec-PME-carbohydrate complexes has provided unexpected new insights into the active role of structural dynamics of macromolecules in biological processes. Our investigation is one of the few reported examples where the substrate dynamics are key in the action pattern of an enzyme. Because of the influence on the substrate dynamics, the methylesterification state of the HG chains is relevant for enzyme processivity and therefore the kinetics. These results also underline the importance of studying the functional conformational transitions that allow proteins and other macromolecules to perform their biological function. To this end, the combination of experiments and MD simulations (17,70) or advanced computations (18) provide a powerful means of advancing our understanding of such events.