Molecular dynamics simulations and rigid body (TLS) analysis of aspartate carbamoyltransferase: evidence for an uncoupled R state

Identifying and Visualizing Macromolecular Flexibility in Structural Biology

Structural biology comprises a variety of tools to obtain atomic resolution data for the investigation of macromolecules. Conventional structural methodologies including crystallography, NMR and electron microscopy often do not provide sufficient details concerning flexibility and dynamics, even though these aspects are critical for the physiological functions of the systems under investigation. However, the increasing complexity of the molecules studied by structural biology (including large macromolecular assemblies, integral membrane proteins, intrinsically disordered systems, and folding intermediates) continuously demands in-depth analyses of the roles of flexibility and conformational specificity involved in interactions with ligands and inhibitors. The intrinsic difficulties in capturing often subtle but critical molecular motions in biological systems have restrained the investigation of flexible molecules into a small niche of structural biology. Introduction of massive technological developments over the recent years, which include time-resolved studies, solution X-ray scattering, and new detectors for cryo-electron microscopy, have pushed the limits of structural investigation of flexible systems far beyond traditional approaches of NMR analysis. By integrating these modern methods with powerful biophysical and computational approaches such as generation of ensembles of molecular models and selective particle picking in electron microscopy, more feasible investigations of dynamic systems are now possible. Using some prominent examples from recent literature, we review how current structural biology methods can contribute useful data to accurately visualize flexibility in macromolecular structures and understand its important roles in regulation of biological processes.

B-factor Analysis and Conformational Rearrangement of Aldose Reductase

The NADPH-dependent reduction of glucose reaction that is catalyzed by Aldose Reductase (AR) follows a sequential ordered kinetic mechanism in which the co-factor NADPH binds to the enzyme prior to the aldehyde substrate. The kinetic/structural experiments have found a conformational change involving a hinge-like movement of a surface loop (residues 213-224) which is anticipated to take place upon the binding of the diphosphate moiety of NADPH. The reorientation of this loop, expected to permit the release of NADP⁺, represents the rate-limiting step of the catalytic mechanism. This study reveals: 1) The Translation/Libration/Screw (TLS) analysis of absolute B-factors of apo AR crystal structures indicates that the 212-224 loop might move as a rigid group. 2) Residues that make the flexible loop slide in the AR binary and ternary complexes. 3) The normalized B-factors separate this segment into three different clusters with fewer residues.

240s Loop Interactions Stabilize the T State of Escherichia coli Aspartate Transcarbamoylase

Here the functional and structural importance of interactions involving the 240s loop of the catalytic chain for the stabilization of the T state of aspartate transcarbamoylase were tested by replacement of Lys-244 with Asn and Ala. For the K244A and K244N mutant enzymes, the aspartate concentration required to achieve half-maximal specific activity was reduced to 8.4 and 4.0 mm, respectively, as compared with 12.4 mM for the wild-type enzyme. Both mutant enzymes exhibited dramatic reductions in homotropic cooperativity and the ability of the heterotropic effectors to modulate activity. Small angle x-ray scattering studies showed that the unligated structure of the mutant enzymes, and the structure of the mutant enzymes ligated with N-phosphonacetyl-L-aspartate, were similar to that observed for the unligated and N-phosphonacetyl-L-aspartateligated wild-type enzyme. A saturating concentration of carbamoyl phosphate alone has little influence on the small angle x-ray scattering of the wild-type enzyme. However, carbamoyl phosphate was able to shift the structure of the two mutant enzymes dramatically toward R, establishing that the mutations had destabilized the T state of the enzyme. The x-ray crystal structure of K244N enzyme showed that numerous local T state stabilizing interactions involving 240s loop residues were lost. Furthermore, the structure established that the mutation induced additional alterations at the subunit interfaces, the active site, the relative position of the domains of the catalytic chains, and the allosteric domain of the regulatory chains. Most of these changes reflect motions toward the R state structure. However, the K244N mutation alone only changes local conformations of the enzyme to an R-like structure, without triggering the quaternary structural transition. These results suggest that loss of cooperativity and reduction in heterotropic effects is due to the dramatic destabilization of the T state of the enzyme by this mutation in the 240s loop of the catalytic chain.

The allosteric activator Mg-ATP modifies the quaternary structure of the R-state of Escherichia coli aspartate transcarbamylase without altering the T<-->R equilibrium

The allosteric enzyme aspartate transcarbamylase from Escherichia coli (ATCase) displays regulatory properties that involve various conformational changes, including a large quaternary structure rearrangement. This entails a major change in its solution X-ray scattering curve upon binding substrate analogues. We show here that, in the presence of the nucleotide effector ATP, known to stimulate the enzyme activity, the scattering profiles show a marked dependence on the metal bound to ATP. Whereas ATP has no major effect on the scattering pattern of ATCase, a saturating concentration of Mg-ATP notably modifies the scattering profile of the enzyme, either in the absence or in the presence of the bisubstrate analogue N-(phosphonacetyl)-l-aspartate (PALA). The transition with PALA in the presence of this metal-nucleotide complex remains concerted. Furthermore, Mg-ATP, as already observed with ATP, has no detectable direct effect on the T to R transition. The experimental scattering curves in the presence of Mg-ATP were fitted by a modeling approach using rigid body movements of the regulatory subunits and the catalytic trimers in the crystal structures. While the differences observed in the T-state in the presence of Mg-ATP are essentially attributed to the binding per se of the nucleotide, the solution structure of the R-state complexed to Mg-ATP is even more extended along the 3-fold axis than the previously described R solution structure, which is already more stretched out along the same axis than the crystal R structure. Based on the crystal structure of the enzyme in the R-state complexed with free ATP, a proposal is made to account for the effect of magnesium.

Allosteric signal transmission involves synergy between discrete structural units of the regulatory subunit of aspartate transcarbamoylase

Previous studies have shown that the S5' beta-strand (r93-r97) of the regulatory polypeptides of the aspartate transcarbamoylases (ATCases) from Serratia marcescens and Escherichia coli are responsible for their diverged allosteric regulatory patterns, including conversion of CTP from an inhibitor in E. coli to an activator in S. marcescens. Similarly, mutation of residues located in the interface between the allosteric and the zinc domains resulted in conversion of the ATP responses of the E. coli enzyme from activation to inhibition, suggesting that this interface not only mediates but also discriminates the allosteric responses of ATP and CTP. To further decipher the roles and the interrelationships of these regions in allosteric communication, allosteric-zinc interface mutations (Y77F and V106A) have been introduced into both the native and the S5' beta-strand chimeric backgrounds. While the significance of this interface in the allosteric regulation has been confirmed, there is no direct evidence supporting the presence of distinct pathways for the ATP and CTP signals through this interface. The analysis of the mutational effects reported here suggested that the S5' beta-strand transmits the allosteric signal by modulating the hydrophobic allosteric-zinc interface rather than disturbing the allosteric ligand binding. Intragenic suppression by substitutions in the hydrophobic interface between the allosteric and the zinc domains of the regulatory chains resulted in the partial recovery of allosteric responses in the EC:rS5'sm chimera and reduced the activation by ATP in the Sm:rS5'ec chimera. Thus, it seems that there is a synergy between these two structural units.

Simulations of the T <--> R conformational transition in aspartate transcarbamylase

Aspartate transcarbamylase (ATCase) from Escherichia coli is one of the best known allosteric enzymes. In spite of numerous experiments performed by biochemists, no consensus model for the cooperative transition between the tensed (T) and the relaxed (R) forms exists. It is hypothesized, however, that changes in the quaternary structure play a key role in the allosteric properties of oligomeric proteins such as ATCase. Previous normal mode calculations of the two states of ATCase illustrated the type of motions that could be important in initiating the transition. In this work four pathways for the transition were calculated using the targeted molecular dynamics (TMD) method without constraint on the symmetry of the system. The most important quaternary structure changes are the relative rotation and translation of the catalytic trimers and the rotations of the regulatory dimers. The simulations show that these quaternary changes start immediately and finish when about 70% of the transition is completed whereas there are tertiary changes throughout the transition. In agreement with the work of Lipscomb et al., it was found that the relative translation between the catalytic trimers appears to play a central role in allowing the transition to occur. In all the simulations differences are observed in the opening and closing behaviours of the domains in the catalytic and regulatory chains that could provide a structural interpretation for the results of certain site-directed mutagenesis experiments. Overall the motions of the subunits are concerted even though the constraint imposed on the TMD method does not explicitly require that this be so.

Tertiary and quaternary conformational changes in aspartate transcarbamylase: a normal mode study

Aspartate transcarbamylase (ATCase) initiates the pyrimidine biosynthetic pathway in Escherichia coli. Binding of aspartate to this allosteric enzyme induces a cooperative transition between the tensed (T) and relaxed (R) states of the enzyme which involves large quaternary and tertiary rearrangements. The mechanisms of the transmission of the regulatory signal to the active site (60 A away) and that of the cooperative transition are not known in detail, although a large number of single, double, and triple site-specific mutants and chimeric forms of ATCase have been obtained and kinetically characterized. A previous analysis of the very low-frequency normal modes of both the T and R state structures of ATCase identified some of the large-amplitude motions mediating the intertrimer elongation and rotation that occur during the cooperative transition (Thomas et al., J. Mol. Biol. 257:1070-1087, 1996; Thomas et al., J. Mol. Biol. 261:490-506, 1996). As a complement to that study, the deformation of the quaternary and tertiary structure of ATCase by normal modes below 5 cm(-1) is investigated in this article. The ability of the modes to reproduce the domain motions occurring during the transition is analyzed, with special attention to the interdomain closure in the catalytic chain, which has been shown to be critical for homotropic cooperativity. The calculations show a coupling between the quaternary motions and more localized motions involving specific residues. The particular dynamic behavior of these residues is examined in the light of biochemical results to obtain insights into their role in the transmission of the allosteric signal.

Role of allosteric: zinc interdomain region of the regulatory subunit in the allosteric regulation of aspartate transcarbamoylase from Escherichia coli

The hydrophobic interface between the allosteric and the zinc domains of the regulatory subunit of aspartate transcarbamoylase has previously been implicated in the heterotropic ATP activation of the enzyme. The present work shows that this interface also affects CTP and CTP-UTP inhibition and proposes a structural explanation for the effects. Mutant enzymes derived from nonselective mutagenesis of residues r101-r106 (residues that contribute part of the interface) displayed a variety of homotropic and heterotropic effects. The cooperative behavior of the enzymes was affected, as indicated by reduced aspartate S0.5 values and apparent Hill coefficient values for V106L, V106L/N105S, and I103F/R102C. In addition, both ATP activation and CTP inhibition were significantly reduced and CTP+UTP synergistic inhibition was decreased in these mutants. The D104G mutant enzyme was subject to inhibition by CTP andCTP+UTP, but was not activated by ATP. Finally, the I103T mutant enzyme had an increased S0.5 value of 11.5 mM and displayed altered effector responses: ATP acted as an inhibitor, and the CTP+UTP synergistic inhibition was reduced. Most of these allosteric variations can be explained in terms of perturbations to the "tongue and groove" hydrophobic interface between the allosteric and the zinc domains and a consequent impact on a second interface ("reg1:cat4") between regulatory and catalytic subunits.

Domain motions in dihydrofolate reductase: a molecular dynamics study

Molecular dynamics simulations have been carried out on the enzyme dihydrofolate reductase from Lactobacillus casei complexed with methotrexate, NADPH and 264 crystallographic water molecules. Analysis of correlations in atomic fluctuations reveal the presence of highly correlated motion (correlation coefficient > 0.6) in the region between residues 30 to 35 and 85 to 90 leading to the identification of two domains, an "adenosine-binding domain" and a "large domain", which rotate by 3 to 4 degrees with respect to each other. The strongest correlation (> 0.6) within the large domain involves a coupling between the motions of the "teen-loop", and the spatially contiguous loops linking beta 6-beta 7 and beta 7-beta 8. Moreover, there is a significant correlation (approximately 0.5) between the adenosine fragment of NADPH and the pteridine and p-aminobenzoyl fragments of methotrexate, which are separated by approximately 17 A, and is lost on removal of "rigid-body" motion from the original trajectory. This provides support for the idea that the relative motion of the two domains is a means by which the occupation of the binding site for the adenosine end of the coenzyme can affect methotrexate binding and vice versa. Quasiharmonic vibrational analysis of the trajectory reveals that the overall dynamics of the system are governed by domain motions whose contributions are dominant at low frequencies. In addition, different low-frequency modes are responsible for separately coupling the adenosine-binding site and parts of methotrexate.

Demonstration that 1-trans-epoxysuccinyl-L-leucylamido-(4-guanidino) butane (E-64) is one of the most effective low Mr inhibitors of trypsin-catalysed hydrolysis. Characterization by kinetic analysis and by energy minimization and molecular dynamics simulation of the E-64-beta-trypsin complex

1-trans-Epoxysuccinyl-L-leucylamido(4-guanidino)butane (E-64) was shown to inhibit beta-trypsin by a reversible competitive mechanism; this contrasts with the widely held view that E-64 is a class-specific inhibitor of the cysteine proteinases and reports in the literature that it does not inhibit a number of other enzymes including, notably, trypsin. The K1, value (3 x 10(-5) M) determined by kinetic analysis of the hydrolysis of N alpha-benzoyl-L-arginine 4-nitroanilide in Tris/HCl buffer, pH 7.4, at 25 degrees C, I = 0.1, catalysed by beta-trypsin is comparable with those for the inhibition of trypsin by benzamidine and 4-aminobenzamidine, which are widely regarded as the most effective low Mr inhibitors of this enzyme. Computer modelling of the beta-trypsin-E64 adsorptive complex, by energy minimization, molecular dynamics simulation and Poisson-Boltzmann electrostatic-potential calculations, was used to define the probable binding mode of E-64; the ligand lies parallel to the active-centre cleft, anchored principally by the dominant electrostatic interaction of the guanidinium cation at one end of the E-64 molecule with the carboxylate anion of Asp-171 (beta-trypsin numbering from Ile-1) in the S1-subsite, and by the interaction of the carboxylate substituent on C-2 of the epoxide ring at the other end of the molecule with Lys-43; the epoxide ring of E-64 is remote from the catalytic site serine hydroxy group. The possibility that E-64 might bind to the cysteine proteinases clostripain (from Clostridium histolyticum) and alpha-gingivain (one of the extracellular enzymes from phyromonas gingivalis) in a manner analogous to that deduced for the beta-trypsin-E-64 complex is discussed.

Diffuse scattering in protein crystallography

The understanding of flexibility and deformability in proteins is one of the current major challenges of structural molecular biology. The knowledge of the average atomic positions of three-dimensional folding of proteins, which is obtained either by X-ray diffraction or n.m.r. spectroscopy, is generally not sufficient to explain their functional mechanisms. Very often it is necessary to consider the existence of other concerted atomic motions as, for example, in the well-known case of the CO molecule fixation at the active site of myoglobin which requires the concerted displacement of a large number of atoms in order to open a channel down to this site. This opening, which depends on the physico-chemical conditions, plays the role of a regulator in the biochemical reactions (Janin & Wodak, 1983; Tainer et al. 1984; Westhof et al. 1984; Ormos et al. 1988).