Open Access Article
Journal of Biomolecular Structure & Dynamics, ISSN 0739-1102 Volume 28, Issue Number 2, (2010) ©Adenine Press (2010)
The authors, the publisher, and the right holders grant the right to use, reproduce, and disseminate the work in digital form to all users.
Insight Derived from Molecular Dynamics Simulation into Substrate-Induced Changes in Protein Motions of Proteinase K http://www.jbsdonline.com
Yan Tao1,2 Zi-He Rao3 Shu-Qun Liu1,3,* 1Laboratory
for Conservation and
Utilization of Bio-Resources, Yunnan Abstract Because of the significant industrial, agricultural and biotechnological importance of serine protease proteinase K, it has been extensively investigated using experimental approaches such as X-ray crystallography, site-directed mutagenesis and kinetic measurement. However, detailed aspects of enzymatic mechanism such as substrate binding, release and relevant regulation remain unstudied. Molecular dynamics (MD) simulations of the proteinase K alone and in complex with the peptide substrate AAPA were performed to investigate the effect of substrate binding on the dynamics/molecular motions of proteinase K. The results indicate that during simulations the substrate-complexed proteinase K adopt a more compact and stable conformation than the substrate-free form. Further essential dynamics (ED) analysis reveals that the major internal motions are confined within a subspace of very small dimension. Upon substrate binding, the overall flexibility of the protease is reduced; and the noticeable displacements are observed not only in substrate-binding regions but also in regions opposite the substrate-binding groove/pockets. The dynamic pockets caused by the large concerted motions are proposed to be linked to the substrate recognition, binding, orientation and product release; and the significant displacements in regions opposite the binding groove/pockets are considered to play a role in modulating the dynamics of enzymesubstrate interaction. Our simulation results complement the biochemical and structural studies, highlighting the dynamic mechanism of the functional properties of proteinase K.
University, Kunming 650091, Yunnan, P. R. China 2Library
of Yunnan University,
Kunming 650091, Yunnan, P. R. China 3National
Laboratory of Macromolecules,
Institute of Biophysics, Chinese Academy of Science, Beijing 100101, P. R. China
Key words: Molecular dynamics; Proteinase K; Essential dynamics; Dynamic pockets; Induced fit; Large concerted motion.
Introduction Proteases are enzymes that can hydrolyze peptide bonds in other proteins through catalytic reaction. Serine proteases (EC 3.4.21) are present in almost all organisms and possess diverse function (1). These enzymes exist as two families the trypsin-like (EC 3.4.21.4) and the subtilisin-like (EC 3.4.21.14) families, which have been independently evolved with a similar catalytic mechanism (1-4) and play important roles in protein digestion, blood coagulation, posttranslational processing of secreted proteins, neurotransmitters, and hormones (5). Proteinase K (EC 3.4.21.64) from the fungus Tritirachium album limber belongs to the subtilisin family (6), which has attracted considerable research interests from academic, industrial, and agricultural communities. These interests are inspired by the ready amenability of subtilases to structural and functional investigation (7), the application to biotechnology such as the removal of DNases and RNases when isolating DNA and RNA from tissues or cell lines (8, 9), and industrial and agricultural importance such as protein-degrading components in washing powders (10) and bio-control agents against parasites (11). As a result, such intense research activity
*Phone: +86 871 5035257 Fax: +86 871 5034838 E-mail:
[email protected] [email protected]
143
144 Tao et al.
has led to the subtililases becoming a model system for protein engineering studies thereby many of their properties such as structure, substrate specificity, catalysis, and stability to temperature and pH profile have been well probed (1, 2, 7, 11-16). A common catalytic mechanism, which involves an identical stereochemistry of the catalytic triad Asp-His-Ser and the oxyanion hole, has been established. In this mechanism, Ser functions as the primary nucleophile and His plays a dual role as the proton acceptor and donor at different stages in the reaction. The role of Asp is believed to bring His into the correct orientation to facilitate the nucleophilic attack by Ser. The role of the oxyanion hole is to stabilize the developing negative charge on the oxygen atom of the substrate upon the formation of the tetrahedral intermediate (17). As a typical representative of subtilases, structures of proteinase K both alone and in complex with peptide chloromethyl ketone inhibitors have been solved to 0.98-2.2 Å in a series of X-ray crystallographic studies (13, 15, 18, 19), with root mean square deviation (RMSD) values ranging from 0.2 Å to 0.4 Å. Typically, proteinase K presents a well-defined global fold which is composed of fifteen β strands, six α helices and one 3/10 helices (Figure 1A). The catalytic triad consists of Asp39, His69 and Ser224; the oxyanion hole is formed by Asn161 (N and Nδ) and Ser224 (N); and the substrate recognition site is primarily formed by two segments, Gly100-Tyr104 and Ser132-Gly136, where the P4-P1 segment of the substrate is accommodated as a central strand via intermolecular hydrogen bonding to form a three-stranded antiparallel β-pleated sheet with residues at the S4-S1 sites of the enzyme (15) (Figure 1B). Although the wealth of the static structural information provides invaluable insight into the function of proteinase K, the detailed aspects of the dynamic processes, i.e., substrate binding and release and how these processes are regulated, remain unstudied. It has been shown that dynamic processes are crucial steps in the function of enzymes (20-22). Even for classes of protein for which the mechanical action is not directly involved in function, molecular motions are still important (23). Therefore, the detailed information about the molecular motions/dynamics of the proteinase K, either bound to substrate or in its unliganded form, is necessary for a complete understanding of its function. Here molecular dynamics (MD) simulations can play very significant role for understanding and predicting motion and dynamics. Molecular dynamics is a powerful method to investigate structural and dynamical information of macromolecular structure in atomic details (24-36). In
Figure 1: Ribbon structures of (A) substrate-free proteinase K (PDB code: 1IC6) and (B) substrate-complexed proteinase K (PDB code: 3PRK). The structures of the enzyme are color-coded by the secondary structure elements with α helices in red, β sheets in yellow and loops in green. The residues of the catalytic triad (Asp39, His69 and Ser224), oxyanion hole (Asn161), and peptide substrate (Ala280-Ala281-Pro282-Ala283) are shown in stick models with carbon atoms in cyan, oxygen atoms in red and nitrogen atoms in blue.
this regard, we have performed MD simulations of proteinase K with and without the peptide substrate to investigate changes in molecular motions induced upon substrate binding. Conventional structural properties of these two forms proteinase K during simulations were checked and compared. Dynamic variations of the substrate-binding pockets/subsites were examined and their relevant functional implications were discussed. In addition, the mechanism underlying the conformational changes of substrate-binding regions was also discussed. Materials and Methods Structural Model Preparations Two models, the substrate-free proteinase K and the proteinase K in complex with a peptide substrate Ala-Ala-Pro-Ala are used for MD simulations. The initial coordinates were taken from high-resolution crystal structures, with PDB codes 1IC6 (at 0.98 Å resolution) for the free proteinase K (13) and 3PRK (at 2.2 Å resolution) for the inhibitor methoxysuccinyl-Ala-Ala-Pro-Ala-chloromethyl ketone-complexed proteinase K (15). For the free proteinase K 1IC6, all the hetero atoms such as NO3 and crystal waters were removed and only the protein atoms were retained. For the complexed structure 3PRK, all the hetero atoms including the crystal waters, the N-terminal methoxysuccinyl group and C-terminal chloromethyl group of the inhibitor substrate were removed, and only the protein atoms and peptide substrate AAPA were retained. Subsequently, a negatively charged C-terminus (COO−) was added to the C-terminal Ala of the peptide substrate. The final models of the free and complexed proteinases K are shown in Figure 1. Residues within proteinase K were orderly numbered 1-279 from the N- to C-termini; and for clarity, residues within the peptide substrate were numbered 280-283 with Ala280, Ala281, Pro282 and Ala283 designated respectively as the P4, P3, P2 and P1 substrate residues according to Schechter and Berger nomenclature (37). Molecular Dynamics Setup All MD simulations were performed with the GROMACS software package (38, 39) using the GROMOS96 43a1 force field. The two structural models were individually solvated with the single point charge (SPC) water model (40) in rectangular periodic boxes with a 1.4 nm solute-wall minimum distance. After a first steepest descent energy minimization with positional restraints on the solute, 1 and 2 chloride ions were added respectively in the free and complexed proteinases K systems to neutralize the overall system charge, leading to a total of 52364 and 58185 atoms for these two systems. The systems were subjected to the second energy minimization until no significant energy change could be detected, followed by 200 ps position-restrained simulations to remove steric clashes and to better “soak” the water molecules into the macromolecules. The production MD simulations were run on linux cluster with 16 CPUs used, achieving a simulation rate for a protein in a 8.5 nm solvated cubic box of 4.0 ns per day. The following protocols were used: a time step of 2 fs was used; center of mass motion was removed for every time step; non-bonded pair was updated for every 10 time steps; electrostatic interactions were treated with Partial Mesh Ewald (PME) (41) with interpolation order of 6 and fourierspacing of 0.15 nm and the coulomb radius was set to 1 nm; van der Waals (VDW) interaction treatment was “cut-off” with radius 1 nm; protein and non-protein (solvent and counterions) were independently coupled to a 300 K bath and pressure coupling was at 1 atmosphere (42). Velocities were generated randomly at startup according to a Maxwell distribution corresponding to a temperature of 300 K. LINCS algorithm (43) with order 4 was used to constrain the bond lengths to their equilibrium positions. The production simulation was performed for 40 ns, and coordinates were saved every 6 ps.
145 Changes in Protein Motions of Proteinase K
146
Analysis Techniques
Tao et al.
The geometrical properties of protein during simulations, including the number of hydrogen bonds (NHB), number of native contacts (NNC), number of residues in the secondary structure elements (SSE), radius of gyration (Rg), solvent accessible surface area (SASA) and RMSD, were calculated using the programs g_hbond, g_mindist, do_dssp (44), g_gyrate, g_sas and g_rms within the GROMACS package, respectively. The ED technique (45, 46) was used to study the large concerted motions in protein. This method is based on the diagonalization of the covariance matrix built from atomic fluctuations in a trajectory from which the overall translation and rotation have been removed, yielding a set of eigenvectors and corresponding eigenvalues. The eigenvectors represent directions in a high-dimensional conformational space and describe concerted fluctuations of atoms. The eigenvalues are a measure of the mean square fluctuation of the system along the corresponding eigenvectors. Projections of the trajectory onto individual eigenvectors provide information about the time dependence of protein motion along these eigenvectors. In this study, the required covariance matrix, eigenvectors and eigenvector projections were obtained by applying the programs g_covar and g_anaeig within the GROMACS package. Only the Cα atoms were included in the analyses, as it has been shown that this subset of atoms captures most of the conformational changes in proteins (21, 45). A method called combined ED analysis is a useful technique for comparing the ED properties of two simulations on similar systems (47). In this method, ED analysis can be performed on a combined trajectory constructed through concatenating individual trajectories of these systems. Analysis and comparison of the behavior of the different parts of the projection along the combined eigenvector provide a powerful tool for evaluating similarities and differences in essential motions between these different trajectories. There are two main effects to be investigated: i) the differences in the average values of the projections, which indicate that simulations have different average displacement along eigenvectors, i.e., a static difference in the equilibrium structure in that direction; ii) the differences in the mean square displacement (MSD) of the projections, which can be used to study difference in dynamics along this direction. Here in this study, the two 7-34 ns MD trajectories of the free and complexed proteinases K were concatenated and the covariance matrix was constructed and diagonalized. The merged trajectory was then projected onto the first 30 combined eigenvectors; and the average value and the MSD of the projection along a combined eigenvector were calculated separately for the two halves of the projection, which correspond to the two parts of the free and complexed proteinases K, respectively. Results Structural Stability Check During Simulations The stability and equilibration of MD simulations of the substrate-free and complexed proteinases K were examined by monitoring Cα RMSD as a function of time. As shown in Figure 2A, The Cα RMSDs of both forms of proteinase K show a steady increase during the first 7 ns. Afterward, these two curves exhibit relatively stable fluctuations, with larger RMSD observed in the free form than in the complexed form. The Cα RMSD curve of the bound substrate AAPA displays a rapid increase during the first several hundred picoseconds and then decreases gradually until a plateau has been reached since ~ 5 ns. Such a stable fluctuation lasts to ~ 34 ns, after which it begins to increase and fluctuates with larger amplitude. Further examination of the distances between the centers of mass of individual substrate residues and those of their corresponding binding sites (Figure 2B) reveals that after ~34 ns the P4 residue (Ala280) of AAPA begins to diffuse away from the S4 site and this is followed by diffusion of other substrate residues P3-P1 at ~ 37 ns,
147
leading finally to the disassociation of the entire peptide substrate from proteinase K. Considering the stability of both forms of proteinase K as well as the stable association between enzyme and substrate during the simulations, we finally selected 7-34 ns parts of trajectories of these two simulation systems for further structural property and ED analyses.
Changes in Protein Motions of Proteinase K
Figure 2: Structural stability assessment during MD simulations. (A) Time evolutions of Cα RMSDs of the substrate-free (black line), substrate-complexed proteinase K (grey line) and bound peptide substrate (black dotted line). RMSDs were calculated every 6 ps after superimposition on the Cα atoms of their respective starting structures. (B) Distances between the centers of mass of substrate residues and those of their corresponding binding sites. The distances of the P1-S1 site, P2-S2 site, P3-S3 site and P4-S4 site are shown in black, grey, black dotted and grey dotted lines, respectively.
Comparison of Structural Properties In order to evaluate the changes in structural characteristics and stability of proteinase K upon substrate binding, geometrical properties such as NHB, SASA, NNC, Rg, RMSD, and SSE were calculated over MD trajectories. Table I shows a comparison of average values of these properties between the free and complexed proteinases K during 7-34 ns simulations. The minor differences in these average values indicate that there are no large conformational changes in proteinase K upon substrate binding. Nevertheless, the observed subtle changes in certain geometrical properties can still reflect the effect of substrate binding on protein dynamics. For instance, the substrate-complexed proteinase K shows slightly increased NHB, NNC and SSE when compared to the substrate-free form, reflecting Table I Average geometrical property (Standard deviations are shown in parentheses) statistics during 7-34 ns MD simulations. RMSDe (Å) Proteinase K free complexed aNumber
SSEi
NHBa
SASAb(Å2)
NNCc
Rgd (Å)
All bbf
SS bbg
Loop bbh
α helix
β sheet
Turn
212 (8) 217 (8)
15388.0 (344.4) 15258.6 (321.3)
134288 (1244) 135547 (930)
16.7 (0.74) 16.8 (0.45)
1.76 (0.10) 1.55 (0.08)
1.12 (0.09) 1.00 (0.07)
2.23 (0.13) 1.94 (0.11)
68 (3) 73 (2)
61 (4) 62 (3)
28 (5) 28 (5)
of hydrogen bonds. A hydrogen bond is considered to exist when the donor-hydrogen-acceptor angle is larger than 120° and the donor-acceptor distance is smaller than 3.5 Å. bTotal solvent accessible surface area. cNumber of native contacts. A native contact is considered to exist if the distance between two atoms is less than 6 Å. dRadius of gyration. eRMSD relative to respective starting structures. The RMSD values were calculated through superposition on the secondary structure element backbones as defined by DSSP [44] in the starting structure. fBackbone RMSD values of all residues. gBackbone RMSD values of secondary structure elements. hBackbone RMSDs of loop regions. iNumber of residues in corresponding secondary structure elements.
148 Tao et al.
the enhanced inter-atomic interactions in protein when peptide substrate is present. Although both forms display almost the same Rg, the overall molecular surface is shrunk by the bound substrate as shown with the reduced SASA in the complexed proteinase K. In addition, the complexed form has lower backbone RMSD values than the free form, suggesting that the stability of the enzyme structure is enhanced upon substrate binding. For both forms of proteinases K, interestingly, the “Loop bb” (backbone of the loop), “All bb” (backbone of all residues) and “SS bb” (backbone of the secondary structure elements) show the largest, moderate and lowest RMSD values, respectively, an indication that the conformational fluctuations originate mainly from the loops linking the secondary structure elements (SSE) (Table I). It is noted that in the presence of the peptide substrate, the RMSD value of the “Loop bb” is decreased by 0.29 Å, slightly higher than the reduced amplitude of the “SS bb”, 0.12 Å, suggesting that upon substrate binding, the “rigidifying” in the loop regions makes more contribution to the reduced RMSD of enzyme structure than that in the SSEs. Taken together, these comparative analyses, combined with relatively lower standard deviations (SD) for certain geometrical properties (such as SASA, NNC, Rg, RMSD and SSE) in the complexed proteinase K than in the native proteinase K, suggest that the enzyme structure is on average in a more compact and stable conformational state when the peptide substrate is present. Essential Dynamics ED technique was used to investigate large concerted motions in both forms of proteinase K as well as the effect of substrate binding on these motions. The 7-34 ns trajectories of the free and complexed proteinases K were used to construct the covariance matrices. Figure 3 shows the eigenvalues and the cumulative contribution of eigenvectors to the total mean square fluctuation (TMSF) as a function of eigenvector index. The eigenvalue curve of the free proteinase K is very steep, with the first 30 eigenvectors contributing 74% to the TMSF. For the complexed proteinase K, the eigenvalue curve is less steep, and therefore more eigenvectors (the first 45 eigenvectors) are required to achieve the same level of approximation of the TMSF. It seems likely that the interactions between the peptide substrate and proteinase K lead to the increased complexity of molecular motions in the complexed proteinase K. However, as observed for several other proteins (20, 21, 48, 49), the actual number of eigenvectors is much smaller than the original 3N-dimensional conformational space formed by the Cα coordinates (here N = 279), reflecting that most of the internal motions of proteinase K are confined within a subspace of very small dimensions. In order to evaluate the influence of the substrate binding on the flexibility of proteinase K, TMSF and the mean square fluctuations within a subspace spanned by the first 30 eigenvectors were calculated. These two values of fluctuations for the complexed proteinase K (2.07 and 1.41 nm2) decrease with respect to their starting values (for the free proteinase K, 2.44 and 1.80 nm2) by 15.1% and 21.7%, respectively. This is in agreement with the results of structural property comparison described above, both suggesting that the substrate binding reduces the overall flexibility of the protease. Changes in Molecular Motions Upon Substrate Binding
Figure 3: Eigenvalues derived from ED analyses of MD simulations. The main plot shows the eigenvalues of only the first 30 eigenvectors. The inset shows the cumulative contribution of all the 837 eigenvectors to the total mean square fluctuation. The eigenvalues of the free and complexed proteinases K are shown in black and dashed lines, respectively.
In order to investigate the motions of proteinase K associated with the substrate binding, combined ED analysis was performed on the merged trajectories of the free and complexed proteinases K simulations. Figure 4 shows the averages and mean square displacements (MSD) of the first 30 eigenvector projections as a function of eigenvector index. Since the analysis was performed on the combined trajectory of the two simulations, the two curves of the average value are symmetric
around zero (Figure 4A). The most significant difference in average values of these projections is observed only between the first eigenvectors, which correspond to the well separated free and complexed states of proteinase K (Figure 5A). The second eigenvectors show very similar average values, suggesting similar equilibrium structures between these two states. The almost identical average values for the projected eigenvectors with index greater than 10 suggest the restrained harmonic motions common to these two states. It is important to note that the cosine contents with 0.5 periods for the projections of the eigenvector 1 are close to zero (~ 0.003 and ~ 0.007 for the free and complexed proteinases K, respectively), indicating that complete sampling/equilibrium has been achieved via 7-34 ns MD simulations (50). Therefore, we mainly focus on the static conformational differences described by the first eigenvectors of these two forms of proteinase K.
149 Changes in Protein Motions of Proteinase K
The Cα root mean square fluctuations (RMSF) and the two extreme structures calculated along the first eigenvector of the merged trajectory are shown in Figure 5A and B, respectively. It is important to keep in mind that the linear interpolations between the two extremes are not the conformational transition pathway between the free and complexed states but emphasize merely the structural differences between them. As shown in Figure 5B, several large conformational arrangements are observed in proteinase K upon substrate binding. The regions showing relatively large shifts, which are arbitrarily defined as those with Cα RMSF greater than 0.06 nm, comprise residues 2-7, 14-22, 43-45, 56-62, 80-82, 99-104, 118-124, 134-135, 162, 165-168, 185-188, 195-197, 214-216, 241-248, 258-260, 262-268. It is interesting to note that almost all of these residues are located either in surfaceexposed loops/links or at the N- and C-termini. Among these, the most significant conformational shifts (RMSF > 0.12 nm) occur in regions of the N- and C-termini and the well-exposed residues 121-122, 166-168, 186, 196, 241-248 and 265-268. This suggests that, for the rigid global protein proteinase K, substrate binding has relatively minor influence on dynamics of the internal rigid core but large effect on external loops, which are located either far away from or within/close to the substrate binding site. Close inspection of the motions along eigenvector 1 (Figure 5A and supplementary animation 1) reveals that upon substrate binding, the structural changes originate mainly from the regions of residues 100-104 and 165-168. Residues 100-104 participate in the formation of the substrate-binding subsites S2-S4; and residues 165-168 are part of the subsite S1. As shown in Figure 5A, the concerted shifts of these two segments towards the peptide substrate lead to the closing of these substrate-binding sites. Accompanying this closing motion, the well-exposed loop residues 195-197 and 265-268 (which are located
Figure 4: Properties of the projections of the merged trajectory onto the “combined” eigenvectors. (A) Average values of projections of the first 30 eigenvectors as a function of eigenvector index. (B) MSD of projections of the first 30 eigenvectors as a function of eigenvector index. The average value and MSD of the projection along a combined eigenvector were calculated separately for the two halves of the projection that represent the free proteinase K half and the complexed proteinase K half. The values for the free and complexed proteinases K are indicated by black circles and grey squares, respectively.
150 Tao et al.
spatially close to residues 165-168) and 56-62 (which lie spatially adjacent to residues 100-104), move concertedly in the direction of the substrate-binding groove, resulting in the contractions of these surface loops. The concerted contractions involving loop residues 214-216 and 241-248, which move concertedly in the direction of protein internal core, are also observed. These loop contractions may account for the decreased SASA and increased inter-atomic interaction observed
Figure 5: Proteinase K conformational changes induced by substrate binding. (A) Two extreme structures obtained through projection of the merged trajectory onto the first “combined” eigenvector. The linear interpolations between these two extremes are colored from blue (the free proteinase K) to red (the complexed proteinase K) to highlight the primary structural differences between these two states, but do not represent the transition pathway. (B) Cα RMSF of the first eigenvector of the merged trajectory as a function of residue number.
in the complexed proteinase K in comparison to the free form. The most prominent structural shift is observed in the loop residues 241-248, which is located opposite the substrate-binding pocket S1 and links helices α5 (whose N-terminus contains the catalytic residue Ser224) and α6 (whose C-terminus is linked to the surface-exposed loop residues 262-268 that show the second largest shift). The small fluctuations of these two helices indicate that the displacement of the segment 241-248 is capable of mediating structural change in the loop region 262268, which in turn can modulate structural change of the loop residues 165-168 via the segment 195-197 to accommodate the C-terminal portion of the incoming substrate. Additionally, the displacement of the residues 241-248 also has effect, to some extent, on fluctuation of the segment 221-223, which lies adjacent to the nucleophilic residue Ser224 and forms the “top” of the S1 pocket. Although relatively minor displacement is observed in the segment 221-223 upon substrate binding (Figure 5A and B), its flexibility may be important as the binding and orientation of the P1 substrate residue require conformational adjustment of this segment. Another large shift is observed in the loop region comprising residues 117-125, which resides between the α3 following segment 100-104 and the β6 preceding segment 132-136. Both the α helix and β strand show only small fluctuations, indicating that the displacements of the segments 100-104 and 132-136 are mediated by the structural changes in the loop region 117-125. Since the two segments 100-104 and 132-136 form the majority of the substrate-binding subsites S1-S4, their dynamic behaviors are of high importance in substrate recognition and binding and product release. Changes in Protein Dynamics/Flexibility Upon Substrate Binding For eigenvectors with index greater than 1, significant differences are observed in the MSD of the eigenvector projections (Figure 4B), which can be used to investigate the change in protein dynamics/flexibility between the free and complexed
151 Changes in Protein Motions of Proteinase K
Figure 6: Comparison of the Cα RMSFs of the selected eigenvectors. The black line shows RMSF of the eigenvector 2 obtained from ED analysis of the free proteinase K simulation. The grey line shows RMSF of the eigenvector 4 obtained from ED analysis of the complexed proteinase K simulation.
proteinases K. As shown in Figure 4B, the most significant conformational travels/ diffusions for the free proteinase K are observed in the subspace spanned by eigenvector 2, whereas those for the complexed proteinase K are found in the subspace spanned by eigenvector 4. Figure 6 shows the comparison of the RMSF values of these two eigenvectors. Substrate binding reduces the fluctuations of the N- and C-termini and most other regions such as residues 42-52, 65-88, 100-104, 132-136, 184-193, 200-231, and 243-256. Most of these regions are surface-exposed loops, whereas some of them are located in SSEs. For instance, residues 65-88 span the helix α2; residues 200-231 span the β11, β12, β13 and N-terminal portion of the α5; and residues 243-256 span the α6. However, most of these SSEs are located on the protein surface except for the α2 and α5, suggesting that the reduction in the overall protein flexibility upon substrate binding arises mainly from the reduced fluctuations of the surface-exposed regions, even including some surface-exposed SSEs. The observed reduction in fluctuations of the substrate-binding segments 100-104 and 132-136 is not surprising, as the interactions between the enzyme and substrate contribute to the stability of these two segments. The most significant increase in fluctuation upon substrate binding is observed in the loop residues 118124. This segment is located opposite the binding groove and therefore its increased flexibility is caused undoubtedly by the bound substrate due to the requirement to adapt to the interaction between substrate and protein segments 100-104 and 132136. The minor increase in fluctuation upon binding is observed in the segments 55-61, 139-143, 173-175, 193-198 and 234-241. It should be noted that the number of these segments is very limited and that the increased amplitude of fluctuation is small, suggesting that these regions have minor effect on the overall flexibility change of proteinase K upon substrate binding. It is interesting to note that all these segments are located either close to or opposite the substrate-binding subsites. For instance, the segment 55-61 lies spatially adjacent to the subsite S2 and S3; the segment 139-143 resides close to the bottom of the S4 pocket; the segments 173-175 and 193-198 are located close to the bottom and right-hand side of the S1 pocket, respectively; and the segment 234-241 is located opposite the S1 pocket. Overall, these observations suggest that the presence of the substrate not only reduces the conformational freedom of the segments that interact directly with the peptide substrate but also increases somewhat the fluctuations of structural regions that are located close to/opposite the substrate-binding sites. Dynamic Pockets Figure 7A shows that for the free proteinase K, the large fluctuations along the first few eigenvectors are mainly seen in the loops of residues 15-25, 40-50 and 185-200 and in the regions of N- and C-termini. It is the flexibility nature of these regions
152 Tao et al.
that give rise to these fluctuation. However, here we mainly focus on the relatively small fluctuations in segments 100-104 and 132-136 occurring along the first few eigenvectors (Figure 7A), because these collective fluctuations result in the opening and/or closing of the substrate-binding groove located between these two segments. For example for the motions described by eigenvector 1 and 4, the segments 100-104 and 132-136 move concertedly away from each other, resulting in the
Figure 7: Cα RMSFs of the first 4 eigenvectors obtained from ED analyses of (A) the substrate-free and (B) complexed proteinase K simulations. Individual plots are sequentially shifted by 0.2 nm.
opening of the groove; and for motion described by eigenvector 2, these two segments move concertedly towards each other, leading to the closing of the groove. Only in the case of eigenvector 2 can the segment 132-136 be found with larger fluctuation than the segment 100-104 (Figure 7A). This is not surprising, as the segment 100-104 is not engaged in secondary structure restriction and exposes well to solvent. In contrast, the N-terminus of the segment 132-136 is tightly anchored to the internal core of enzyme structure and its C-terminal portion forms a short antiparallel β-sheet with residues 168-170 (Figure 1), resulting in elevated rigidity of this segment. However, the flexibility of the segment 132-136 is still important as the accommodation of the large P1 and P4 residues requires conformational arrangement of this segment. Although the presence of the peptide substrate reduces the fluctuations of the segments 100-104 and 132-136 that make direct contact with the substrate, relatively large fluctuations of these two segments can still be found along the first few eigenvectors of the complexed proteinase K (Figure 7B). The collective fluctuations around the substrate-binding regions originate primarily from the fluctuations of the bound peptide substrate, resulting in the dynamic variation of the substrate-binding subsites/pockets as has been observed in the free proteinase K. Visualization of the motions along eigenvector 1 of the complexed proteinase K indicates that the most pronounced dynamic changes occur in the pockets S1 and S4 (Figure 8A and Supplementary animation 2). Both pockets are enlarged due to the concerted outward movements of the segments surrounding these two pockets, i.e., the segments 101-104 and 136-138 that form the two sides of the S4 pocket; and the segments 132-135 and 161-163 that form respectively the side and bottom of the S1 pocket (2, 11). In addition, the upper portion of the groove, where the subsites S2 and S3 reside, becomes narrowed due to the large inward displacement of the segment 132-134. For motions described by eigenvector 2 (Figure 8B and Supplementary animation 3), the largest downward displacement of the loop 117-123 causes the segments 99-102 and 132-136 to move away from each other resulting in the opening of the upper portion of the binding groove. By contrast, the lower portion of the binding groove (where the S4 pocket resides) becomes slightly narrowed due to the concerted inward movements of the segment
103-105 and 137-139. Simultaneously, the large leftward displacement of the segment 240-246 causes the segment 221-223 to move outwards, thus opening up the lid of the S1 pocket. Interestingly, the size of the bottom of the S1 pocket appears to be reduced by the inward movement of the segments 132-135 and 163-165. The large concerted motions within the subspace spanned by eigenvector 3 seem to lead to a twist of the binding groove and closure of the S1 pocket (Figure 8C and
153 Changes in Protein Motions of Proteinase K
Figure 8: Large concerted motions of the complexed proteinase K described by (A) eigenvector 1, (B) eigenvector 2, (C) eigenvector 3, and (D) eigenvector 4. The linear interpolations between the two extremes extracted from eigenvector projections are colored from blue to red to highlight the structural differences between these two extremes. The regions involving large concerted motions that cause dynamic variations of the substrate-binding pockets/subsites are labeled.
Supplementary animation 4). The obvious displacement observed in residues 100101 widens the S2 subsite and provides more space for the P2 residue, Pro282. In the subspace spanned by eigenvector 4 (Figure 8D and Supplementary animation 5), there is no apparent displacement in the segment 100-104 and 132-136. The largest displacement is observed in the segment 118-123, which mediates the segment 136-139 to move towards the helix α3, resulting in the reduced size of the bottom of the S4 pocket. These complicated dynamic variations of the substrate-binding subsites/pockets would most likely correlate with the complex catalytic processes of proteinase K and this will be discussed below. Discussion Although thoroughly studied in terms of structural, biochemical and biophysical properties, the detail information on the dynamics of proteinase K that is crucial for the understanding of its function remains unstudied. Here in this study, we performed two 40 ns MD simulations on the substrate-free and complexed
154 Tao et al.
proteinases K to investigate the influence of substrate binding on the dynamics/ molecular motions of this protease. Both forms of proteinase K displayed similar geometrical properties during MD simulations. However, the subtle geometrical differences can still reflect a more compact and stable conformational state of the enzyme structure when the substrate is present. This is further supported by comparison of mean square fluctuations within high-dimensional conformational space between these two forms of proteinase K, revealing that the complexed proteinase K shows the decreased overall structural flexibility. Such reduced conformational freedom upon substrate binding, which has been observed for many other enzymes (48, 49, 51-55), supports the proposed induced-fit mechanism of substrate binding (56). The attenuation of the loop motions, including not only the loops involved in direct contact with the peptide substrate but also those located far away from the substrate-binding sites, contribute to the reduced overall flexibility. It is interesting, however, to note that the substrate binding can enhance conformational flexibility for regions located relatively close to the substrate-binding sites, especially for loops located opposite the binding groove and S1 pocket. Such increased fluctuations are necessary, as the accommodation of the substrate into the binding groove and the subsequent catalytic steps require relatively large structural arrangements of the substratebinding regions, and these are regulated by dynamic behaviors of the regions located opposite the substrate-binding sites. Our combined ED analysis identified two such important loops (residues 117-125 and 241-248) in proteinase K showing large displacements upon substrate binding. The large structural displacements of the opposite loops have been observed in other enzymes (49, 51, 52), and are believed to serve to modulate dynamic behaviors of the substrate-binding regions according to the requirement of the substrate binding, orientation and product release. Here the rigid SSEs linking the substrate-binding regions and the opposite loops may serve as hinge segments, allowing the transmission and communication of the conformational changes between these segments. Interestingly, despite the large concerted shifts observed in many surface-exposed loops upon substrate binding, there is no significant structural shift in the catalytic triad residues, revealing that the architecture of the triad is well maintained whether the peptide substrate is present or not. The architecture maintenance of the catalytically important residues during MD simulation has been observed for other enzymes, i.e., the proteintyrosine phosphatase (48, 52). We consider that the conserved interactions such as hydrogen bonding and electrostatic interactions between these conserved residues, together with their positional invariability in the enzyme structure, contribute to the stability of the catalytic triad. However, the side chain flexibility and/or thermal motions of the catalytic residues, especially the flexibility of Ser224 hydroxyl group, is crucial for the catalytic reaction as it needs to assume different orientations appropriate for either proton transfer or nucleophilic attack and, subsequently, for release of the cleaved peptide product (57). Like the results of ED analyses on the other proteins (20, 21, 48, 49, 58), the large concerted motions of both forms of proteinase K are also captured by the first few eigenvectors. The structural arrangements surrounding the ligand-binding site can cause the dynamic variations of the binding pockets and therefore are often proposed to be related to the functional properties of the proteins (21, 22, 51-55, 57-59). As described above, the ED analysis yields individual large concerted motions. However, all the motional modes interplay with each other in a complex manner, and therefore not necessary all individual modes will correspond to a specific functional task. For purpose of clarity, the individual modes described by the first 4 eigenvectors will be discussed to link to the putative biological functions of proteinase K. In the case of the substrate-free proteinase K, the oscillatory nature of the segment 100-104 and to a less extent, that of the segment 132-136, causes random opening/ closing of the binding groove. For serine proteases, the breathing motions of their
substrate-binding grooves have been reported in (16, 57, 58) and are likely to be related to substrate recognition and binding. In the case of the substrate-complexed proteinase K, the collective motions described by the first eigenvector enlarge the pockets S1 and S4. We consider that the enlargements of these two important pockets provide more space for the P1 and P4 substrate residues to suit the needs for accommodation and subsequent orientation of these two residues. The simultaneous closing of the subsites S2 and S3 could grip the corresponding substrate residues P2 and P3 to prevent release of the peptide substrate before catalysis. The second ranked motions described by the eigenvector 2 not only narrow the bottoms but also open the lids of both S1 and S4 pockets. In conjunction with the opening of the upper portion (where the subsites S2 and S3 reside) of the binding groove, these motions lead to loose contacts with the substrate, thus facilitating the substrate release. The twist motion of the binding groove described by eigenvector 3 may facilitate position adjustments of the P2-P4 substrate residues in their respective binding subsites/pockets. On the other hand, the closing of the S1 pocket is likely to be responsible for the precise positioning of the P1 substrate residue, which is necessary for the nucleophilic attack to take place. Interestingly, we also observe a shift of the S1 pocket lid (segment 221-223) which leads to further approach of Ser224 to the P1 substrate residue, implying that this motional mode may be in the stage ready for the nucleophilic attack. The large concerted motions described by eigenvector 4 reduce the bottom size of S4 pocket and as thus can be linked to the release of the S4 substrate residue. Among the first four predominant motional modes, two modes, i.e., described by eigenvectors 2 and 4, cause reduced size of the bottom of the S4 pocket. This may explain the first dissociation of P4 substrate residue from its binding pocket observed in our MD simulation (Figure 2B). Accordingly, we speculate that the release of the N-terminal peptide product is possibly initiated by the dissociation of the P4 from the enzyme. Conclusions We have investigated the effect of substrate binding on the dynamics/molecular motions of proteinase K using MD simulation method. The binding of the substrate makes the enzyme structure adopt a more compact conformation, which can be attributed to the increased inter-atomic interactions and decreased loop flexibility. Combined ED analysis reveals that upon substrate binding, the concerted motions occur not only in the substrate binding regions leading to the closing of the binding groove, but also in surface-exposed loops located relatively far away from substrate-binding sites, leading to the contractions of these loops. The noticeable concerted motions upon binding are observed in the two loop segments 117-125 and 241-248 that are located opposite the substrate binding sites, which play a role in modulating dynamics of the substrate-binding regions to suit the needs of accommodating the bound substrate. Furthermore, the individual ED analyses of the complexed proteinase K simulation also suggest that these two segments play a role in mediating dynamic variations of the substrate-binding pockets/subsites according to the requirement of substrate orientation and release. The dynamic pockets within the first few essential subspaces are characterized and proposed to be linked to the functional properties of proteinase K. The results in this work provide the detailed information about the dynamic aspects of enzymatic mechanism of proteinase K. Supplementary Material Supplementary computer animations dealing with substrate binding-induced conformational changes of proteinase K and large concerted motions of substrate-complexed proteinase K are posted at this article’s web site at www. jbsdonline.com.
155 Changes in Protein Motions of Proteinase K
156
Acknowledgements
Tao et al.
The authors thank High Performance Computer Center of Yunnan University for computational support. This work was funded by the National Natural Science Foundation of China (approved number 30860011) and partially supported by grants from Yunnan Province (2006C008M, 2007C163M, 07Z10756 and 2007PY-22) and Project 211 and Innovation Group Project from Yunnan University (KL070002). References 1. R. J. Siezen, W. M. de Vos, J. A. Leunissen, and B. W. Dijkstra. Protein Eng 4, 719-737 (1991). 2. R. J. Siezen and J. A. M. Leunissen. Protein Sci 6, 501-523 (1997). 3. D. M. Blow. Acc Chem Res 9, 145-152 (1976). 4. H. Tsukuda and D. Blow. J Mol Biol 184, 703-711 (1985). 5. H. Neurath. J Cell Biochem 32, 35-49 (1986). 6. W. Ebeling, N. Hennrich, M. Klockow, H. Metz, H. D. Orth, and H. Lang. Eur J Biochem 47, 91-97 (1974). 7. J. A. Wells and D. A. Estell. Trends Biochem Sci 13, 291-297 (1988). 8. U. Wiegers and H. Hilz. Biochem Biophys Res Commun 44, 513-519 (1971). 9. H. Hilz, U. Wiegers, and P. Adamietz. Eur J Biochem 56, 103-108 (1975). 10. W. V. Shaw. Biochem J 246, 1-17 (1987). 11. S. Q. Liu, Z. H. Meng, J. K. Yang, Y. X. Fu, and K. Q. Zhang. BMC Struct Biol 7, 33 (2007). 12. M. W. Pantoliano, R. C. Ladner, P. N. Bryan, M. L. Rollence, J. F. Wood, and T. L. Poulos. Biochemistry 26, 2077-2083 (1987). 13. C. Betzel, S. Gourinath, P. Kumar, P. Kaur, M. Perbandt, S. Eschenburg, and T. P. Singh. Biochemistry 40, 3080-3088 (2001). 14. A. Müller, W. Hinrichs, W. M. Wolf, and W. Saenger. J Biol Chem 269, 23108-23111 (1994). 15. W. M. Wolf, J. Bajorath, A. Müller, S. Raghunathan, T. P. Singh, W. Hinrichs, and W. Saenger. J Biol Chem 1991, 17695-17699 (266). 16. J. R. Martin, F. A. A. Mulder, Y. Karimi-Nejad, J. van der Zwan, M. Mariani, D. Schipper, and R. Boelens. Structure 5, 521-532 (1997). 17. G. Dodson and A. Wlodawer. Trends Biochem Sci 23, 347-352 (1998). 18. C. Betzel, G. P. Pal, M. Struck, K. D. Jany, and W. Saenger. FEBS Lett 197, 105-110 (1986). 19. C. Betzel, G. P. Pal, and W. Saenger. Eur J Biochem 178, 155-171 (1988). 20. B. L. de Groot, S. Hayward, D. M. F. van Aalten, A. Amadei, and H. J. C. Derendsen. Proteins 31, 116-127 (1998). 21. D. M. F. van Aalten, D. A. Conn, B. L. de Groot, J. B. C. Findlay, H. J. C. Berendsen, and A. Amadei. Biophys J 73, 2891-2896 (1997). 22. L. V. Mello, B. L. de Groot, S. Li, and M. J. Jedrzejas. J Biol Chem 277, 36678-36688 (2002). 23. C. P. Barrett, B. A. Hall, and M. E. M. Noble. Acta Cryst D 60, 2280-2287 (2004). 24. G. H. Braun, D. M. M. Jorge, H. P. Ramos, R. M. Alves, V. B. Da Silva, S. Giuliatti, S. V. Sampaio, C. A. Taft, and Carlos H. T. P. Silva. J Biomol Struct Dyn 25, 347-356 (2008). 25. J. Yoon, J. Park, S. Jang, K. Lee, and S. Shin. J Biomol Struct Dyn 25, 505-516 (2008). 26. A. Cordomí, E. Ramon, P. Garriga, and J. J. Perez. J Biomol Struct Dyn 25, 573-588 (2008). 27. C. Y.-C. Chen, Y.-F. Chen, C.-H. Wu, and H.-Y. Tsai. J Biomol Struct Dyn 26, 57-64 (2008). 28. J.-H. Zhao, C.-T. Yang, J. W. Wu, W.-B. Tsai, H.-Y. Lin, H.-W. Fang, Y. Ho, and H.-L. Liu. J Biomol Struct Dyn 26, 65-74 (2008). 29. B. N. Macchion, R. Strömberg, and L. Nilsson. J Biomol Struct Dyn 26, 163-174 (2008). 30. U. B. Sonavane, S. K. Ramadugu, and R. R. Joshi. J Biomol Struct Dyn 26, 203-214 (2008). 31. F. Mehrnejad and N. Chaparzadeh. J Biomol Struct Dyn 26, 255-262 (2008). 32. J. Yoon, S. Jang, K. Lee, and S. Shin. J Biomol Struct Dyn 27, 259-270 (2009). 33. D. Ilda, C. Giovanni, and D. Alessandro. J Biomol Struct Dyn 27, 307-318 (2009). 34. A. Cordomí and J. J. Perez. J Biomol Struct Dyn 27, 127-148 (2009). 35. H. R. Bairagya, B. P. Mukhopadhyay, and K. Sekar. J Biomol Struct Dyn 27, 149-158 (2009). 36. J. Zhang. J Biomol Struct Dyn 27, 159-162 (2009). 37. I. Schechter and A. Berger. Biochem Biophys Res Commun 27, 157-162 (1967). 38. C. Kutzner, D. van der Spoel, M. Fechner, E. Lindahl, U. W. Schmitt, B. L. de Groot, and H. Grubmuller. J Comp Chem 28, 2075-2084 (2007). 39. E. Lindahl, B. Hess, and D. van der Spoel. J Mol Model 7, 306-317 (2001).
40. H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, and J. Hermans. Interaction models for water in relation to protein hydration. In: Intermolecular forces, B. Pullman, Eds., D. Reidel Publishing Company, Dordrecht, the Netherlands, pp. 331-342 (1981). 41. T. A. Darden, D. M. York, and L. G. Pedersen. J Chem Phys 98, 10089-10092 (1993). 42. H. J. C. Berendsen, J. P. M. Postma, W. F. van Gunsteren, A. DiNola, and J. R. Haak. J Chem Phys 81, 3684-3690 (1984). 43. B. Hess, H. Bekker, H. J. C. Berendsen, and J. Fraaije. J Comp Chem 18, 1463-1472 (1997). 44. W. Kabsch and C. Sander. Biopolymers 22, 2577-2637 (1983). 45. A. Amadei, A. B. M. Linssen, and H. J. C .Berendsen. Proteins Struct Funct Genet 17, 412-425 (1993). 46. M. A. Balsera, W. Wriggers, Y. Oono, and K. Schulten. J Phys Chem 100, 2567-2572 (1996). 47. D. M. F. van Aalten, A. Amadei, A. B. M. Linssen, V. G. H. Eijsink, and G. Vriend. Proteins Struct Funct Genet 22, 45-54 (1995). 48. G. H. Peters, T. M. Frimurer, J. N. Andersen, and O. H. Olsen. Biophys J 77, 505-515 (1999). 49. C. P. Barrett and M. E. M. Noble. J Biol Chem 280, 13993-14005 (2005). 50. B. Hess. Phys Rev E 62, 8438-8448 (2000). 51. G. H. Peters and R. R. Bywater. Protein Eng 12, 747-754 (1999). 52. G. H. Peters, T. M. Frimurer, J. N. Andersen, and O. H. Olsen. Biophys J 78, 2191-2200 (2000). 53. G. H. Peters, D. M. F. van Aalten, A. Svendsen, and R. Bywater. Protein Eng 10, 149-158 (1997). 54. O. P. Kuipers, M. Vincent, J. C. Brochon, H. M. Verheij, G. H. de Haas, and J. Gallay. Biochemistry 30, 8771-8785 (1991). 55. J. R. Collins, S. K. Burt, and J. W. Erickson. Nat Struct Biol 2, 334-338 (1995). 56. D. E. Jr. Koshland. Natl Acad Sci USA 44, 98-104 (1958). 57. S. Q. Liu, Z. H. Meng, Y. X. Fu, and K. Q. Zhang. J Mol Model 16, 17-28 (2010). 58. D. W. Miller and D. A. Agard. J Mol Biol 286, 267–278 (1999). 59. S. Q. Liu, C. Q. Liu, and Y. X. Fu. J Mol Graphics Modell 26, 306-318 (2007).
Date Received: March 6, 2010
Communicated by the Editor Ramaswamy H. Sarma
157 Changes in Protein Motions of Proteinase K
