Thermodynamics of the Purine Nucleoside ... - ACS Publications

Subscriber access provided by GAZI UNIV

Article

Thermodynamics of the Purine Nucleoside Phosphorylase Reaction Revealed by Computer Simulations Geir Villy Isaksen, Johan Aqvist, and Bjorn-Olav Brandsdal Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.6b00967 • Publication Date (Web): 15 Dec 2016 Downloaded from http://pubs.acs.org on December 20, 2016

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Biochemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Thermodynamics of the Purine Nucleoside Phosphorylase Reaction Revealed by Computer Simulations Geir Villy Isaksen,† Johan Åqvist,‡ and Bjørn Olav Brandsdal*,†

†

The Centre for Theoretical and Computational Chemistry, Department of Chemistry, University of Tromsø, NO-9037 Tromsø, Norway

‡

Department of Cell and Molecular Biology, Biomedical Center, Uppsala University, SE-75124 Uppsala, Sweden

*

Corresponding author:

E-mail:

Bjørn Olav Brandsdal

[email protected]

Phone:

+47 77 64 40 57

Fax:

+47 77 64 47 65

ACS Paragon Plus Environment

1

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 26

ABSTRACT Enzymes are able to catalyze chemical reactions by reducing the activation free energy, yielding significant increase in the reaction rates. This can thermodynamically be accomplished by either reducing the activation enthalpy or by increasing the activation entropy. The effect of remote mutations on the thermodynamic activation parameters for human purine nucleoside phosphorylase is here examined using extensive molecular dynamics and free energy simulations. More than 2700 independent reaction free energy profiles for six different temperatures have been calculated to obtain high precision computational Arrhenius plots. Based on these, the activation enthalpies and entropies were computed from linear regression of the plots with ∆G‡ as a function of 1/T, and the obtained thermodynamic activation parameters are in very good agreement with experiments. The Arrhenius plots immediately show that the 6oxopurines (INO and GUO) have identical slopes, whereas the 6-aminopurine (ADO) has a significantly different slope, indicating that the substrate specificity is related to the difference in thermodynamic activation parameters. Furthermore, the calculations show that the human PNP specificity for 6-oxopurines over 6-aminopurines originates from significant differences in electrostatic preorganization. The effect of the remote double mutation K22E and H104R (E:R) has also been examined, as it alters human PNP toward the bovine PNP. These residues are situated on the protein surface, 28 – 35 Å from the active site, and alters the enthalpy – entropy balance with little effect to the catalytic rates. It is thus quite exceptional that the empirical valence bond method is able to reproduce the enthalpies and entropies induced by these long – ranged mutations.


2

Page 3 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Enzymes display remarkable abilities to significantly increase the rates at which chemical reactions occur compared to the same reactions in water. Thermodynamically, this is accomplished by a significant lowering of the free energy of activation

∆ ‡ = ∆ ‡ − ∆ ‡

(1)

where T is the temperature and ∆H‡ and ∆S‡ the enthalpy and entropy of activation, respectively. The relationship between the free energy of activation (∆G‡) and the reaction rates (krxn) is according to transition state theory

=

∆ ‡

exp − =

∆ ‡

∆ ‡

exp −

(2)

where, κ is the transmission coefficient (generally assumed to be ~1), and kB and h are Boltzmann’s and Planck’s constants, respectively. It is well established that enzymes are able to significantly lower the energy barriers required for converting substrates to products through improved electrostatic stabilization of the transition state.1-7 That is, part of the enzyme folding energy is used to align the active site dipoles so that the enzyme becomes preorganized for stabilizing charge redistribution along the reaction coordinate.8,9 From the viewpoint of thermodynamics (eqs. 1 and 2) it is only the enthalpy-entropy balance that determines the free energy of activation, and thus the reaction rates, at a fixed temperature. What is perhaps not so obvious is that identical reaction rates for two enzyme orthologs (or for two different substrates), may in fact hide a very different balance between the activation entropy and enthalpy. This type of enthalpy-entropy compensation has for example in recent works proven to be the key to


3

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 26

understand enzyme temperature adaptations in terms of actual structure-activity relationships.10,11 The fingerprint of enzymes with respect to thermodynamics is therefore not just the free energy barrier, but also the enthalpic and entropic components. The importance of enthalpy-entropy compensation is also easily overlooked since mutations far away from the active site generally do not significantly affect the catalytic rates.12 Recent work suggests that it is the protein-water interface that regulates the thermodynamic activation parameters involved in temperature adaptation of trypsin.10 Mutations of key residues remote to the active site were found to soften the protein surface in mesophilic bovine trypsin rendering the enthalpy-entropy balance remarkably similar to that of the cold-adapted salmon trypsin. The opposite effect was observed when rigidifying the enzyme surface in cold-adapted trypsin making the thermodynamics similar to that of the warm-adapted enzyme.13 What is rather remarkable is that the calculations predict significant changes in ∆H‡ and ∆S‡, but also that these are nearly perfect compensating. Interestingly, a recent work by Ghanem et al. showed experimentally that the enthalpy-entropy balance can change upon distant surface mutations for human purine nucleoside phosphorylase (HsPNP).14 Purine nucleoside phosphorylase (PNP) catalyzes the reversible cleavage of the glycosidic bond of ribo- and 2’-deoxyribonucleosides yielding the corresponding purine base and (2’deoxy)ribose-1-phosphate as products.15 HsPNP belongs to the low molecular mass (low-mm) PNPs that generally are specific towards 6-oxopurines, e.g. inosine (INO) and guanosine (GUO), and thus frequently referred to as “Ino-Guo phosphorylases”.16-22 In available crystal structures with transition state inhibitors, Y88, F200, E201, N243 and H257 interact with the nucleoside, while S33, H64, R84, A116, S220 and two conserved waters interacting with the phosphate group are indicated as key catalytic site residues.17,23-25 In addition, F159* from one adjacent


4

Page 5 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

subunit is oriented in the active site so that it makes contact with the nucleoside and shields it from the solvent. However, despite being studied for decades, the catalytic mechanism for low-mm PNPs has remained controversial and somewhat confusing.20,22,25-27 In a recent work, density functional theory (DFT) calculations and extensive empirical valence bond (EVB) free energy simulations were used to investigate the glycosidic bond cleavage step of HsPNP.28 In the reported reaction mechanism a singly protonated phosphate was predicted for the nucleophilic attack on the neutral ribonucleoside in the chemical step. The corresponding purine base is thus rendered anionic towards the product state. This is consistent with experiments showing that the purine base binds to PNP in the anionic form in the reverse synthesis direction.29 The EVB model successfully reproduced the experimental rates with GUO, INO and ADO (Figure 1) as substrates and, most importantly, the fact that HsPNP is highly specific for 6-oxopurines and shows only negligible affinity for 6-aminopurines. The validity of the predicted reaction mechanism was further strengthened by reproducing the reversal of the 6-oxo vs. 6-aminopurine specificity through the N243D mutation with the EVB method. The rate-limiting step of some PNPs is furthermore known to be the release of purine,14,30,31 but prior to this, the anionic base must be protonated. In our previous work, we computed a possible proton shuttle mechanism from ribose-1-phosphate to the base, which clearly demonstrated that it would occur at a significantly faster rate compared to the glycosidic bond cleavage step. Thus, with respect to any potential protonation steps in phosphorolysis, the bond cleavage step is rate limiting. In this work, we apply extensive EVB simulations to investigate the thermodynamic activation parameters for the glycosidic bond cleave step in native HsPNP using our previously reported reaction mechanism. It is noteworthy here that the experimental activation entropies are positive


5

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 26

and that these are nicely reproduced by the calculations. We have also examined the double surface mutation (K22E and H104R), altering HsPNP toward bovine PNP (BtPNP), which was reported by Ghanem et al. to modify the enthalpy-entropy balance.14 It is rather exceptional that the calculated thermodynamic activation parameters reproduce the experimental trends for these distant mutations. Our simulations thus provide valuable insight to how protein surface residues can govern the enthalpy-entropy balances without significantly affecting the catalytic rates.

METHODS Model systems of human purine nucleoside Phosphorylase (HsPNP) in complex with inosine (INO), guanosine (GUO) and adenosine (ADO) were created by modifying the transition state inhibitor Immucillin (ImmH) obtained from the crystallographic structure with PDB entry code 1RR6.32 Coordinates for the phosphate loop (residues 57 – 65) oriented inwards to the active site were obtained from PDB entry code 4EAR.33 The glycosidic bond cleavage step was modeled with the EVB method2,8 utilizing the reaction mechanism recently described by Isaksen et al.28 Reaction free energy profiles were calculated using the free energy perturbation (FEP) umbrella sampling approach as described elsewhere.2,8 Each enzyme and water reaction free energy profile involved 500 ps simulation and compromised 51 discrete FEP steps. Activation enthalpies and entropies were obtained from Arrhenius plots where the activation free energy is plotted as function of 1/T so that ∆ ‡

= ∆ ‡ − ∆ ‡

(3)

Qgui34 was used to efficiently prepare input files and analyze the simulations carried out with the molecular dynamics package Q.35 Both reference and enzyme reactions with ADO, GUO and INO as substrates were immersed into a spherical droplet of water molecules centered at the


6

Page 7 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

phosphate anion with a 25 Å radius. Simulations with GUO as substrate were in addition performed with the entire homotrimer soaked in a spherical droplet of water with a 53 Å radius for native HsPNP and the K22E/H104R (E:R) double mutant. Water molecules were described using the TIP3P potential.36 A direct cutoff of 10 Å was used for non-bonded interactions, with long-range electrostatics treated using a multipole expansion method.37 No cutoffs were, however, applied to interactions involving the reacting fragments.

The two valence bond states in Figure 1B represent the interconversion of the resonance structures ( → ) describing the reactant and product states. A specific potential energy function,

!,

describing the charge distribution and bonding arrangement for each structure, was

used to calculate the energetics associated with the given resonance form. The energies of the different resonance structures are obtained from the diagonal terms of the EVB Hamiltonian, while the ground-state energy is obtained by adiabatic mixing of these structures via the offdiagonal matrix elements.2,8 Each diagonal energy function is given by an analytical force field of the form

! ! ! ! ! !! = "! = #$ % + #' ( + #)*

+ # $,

+ # $, , + #,, + - !

(4)

where the subscripts bnd, ang, tor and nb denote bond, angle, torsion and non-bonded terms, whereas r and s denote the reacting fragments and their surroundings, respectively. The bond ! potential energies, #$ % , for the reacting atoms are described by Morse potentials. Charges for

the EVB region were assigned in accordance with atomic electrostatic surface potential (ESP) charges obtained from single-point B3LYP/6-31G**+ calculations using Jaguar38 as previously


7

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 26

described by Isaksen et al.28 Remaining atom parameters and charges for the potential energy function,

!,

were assigned in accordance with the OPLS-AA force field.39,40 The last term of the

Hamiltonian, - ! , represents the intrinsic gas-phase energy of the given resonance structure with all fragments at infinite separation.2,8 A constant (Aij) was used for describing the off-diagonal matrix elements, !. , representing the adiabatic mixing of the VB states. The EVB potential energy surfaces for the reference reactions were fitted to the DFT activation and reaction energies reported by Isaksen et al.28 by adjusting Aij and - ! .

Prior to the production phase the systems were heated from 1 K to the target temperature (280 - 310 K) during 31 ps, using a stepwise scheme, followed by an equilibration period of 100 ps. Bonds and angles of solvent molecules were constrained using SHAKE.41 For the production phase a time step of 1 fs was used, and the temperature was maintained at the target temperature using a weak coupling to an external bath. All reference reactions were averaged over 10 independent simulations (total of 5 ns), whereas the enzyme simulations where averaged over 50-100 independent simulations (total of 25 – 50 ns) at six temperatures in the range 280 – 310 K (total of 150 – 300 ns). Altogether, the presented work utilized 2730 independent simulations for INO, GUO and ADO, resulting in a total simulation time of 1.4 µs.


8

Page 9 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

RESULTS AND DISCUSSION The main goal of the present work is to investigate the enthalpy-entropy balance for long-range mutations in HsPNP using the reaction mechanism recently predicted by Isaksen et al.28 The thermodynamic activation parameters provide valuable insight not only into the mechanisms responsible for the observed enthalpy-entropy compensation, but also into the energetics responsible for the differing HsPNP activity with GUO, INO and ADO (Figure 1A) as substrates.

Glycosidic bond cleavage step in HsPNP The glycosidic bond cleavage step in HsPNP was examined as previously described28 with GUO, INO and ADO (Figure 1A) using the EVB resonance structures illustrated in Figure 1B. To attain sufficiently high precision the enzyme calculations were averaged over 100 independent runs at each temperature (see below). The average EVB activation free energy at 298 K of 12.8 ± 0.1 kcal/mol with GUO as substrate is in good agreement with the experimentally obtained ∆G‡ of ~14 kcal/mol for the chemical step.14 The obtained activation free energies of 13.0 ± 0.1 kcal/mol and 20.2 ± 0.2 kcal/mol for INO and ADO, respectively, are furthermore fully compatible with the kcat values reported by Steockler et al.42 where substrate release is considered rate-limiting. As is evident from the reaction free energy profiles plotted in Figure 2A, HsPNP lowers the activation free energy for all three substrates with respect to the reference reactions in water. The lowering is, however, significantly smaller for ADO, which reflects the fact that HsPNP catalyzes phosphorolysis with ADO as substrate at much lower rate compared to GUO and INO.42


9

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 26

Thermodynamics behind the 6-oxopurine specificity Since the catalytic rates at room temperature are well reproduced for the three substrates, we can now start to examine the thermodynamic activation parameters for the glycosidic bond cleavage step. Six different temperatures were chosen in the range 280 to 310 K and 100 independent reaction free energy profile calculations were carried out at each temperature to obtain high precision Arrhenius plots. Activation enthalpies and entropies were computed from linear regression of the plots with ∆G‡/T as a function of 1/T. From the resulting Arrhenius plots (Figure 2B) it can immediately be seen that the 6-oxopurines (INO and GUO) have identical slopes, whereas the 6-aminopurine (ADO) has a significantly different slope. The calculated activation parameters at 298 K for GUO are ∆H‡ = 15.0 kcal/mol and T∆S‡ = 2.3 kcal/mol (Table 1). These values are similar to the experimentally obtained ∆H‡ of 18.6 ± 0.7 kcal/mol and T∆S‡ of 4.1 ± 0.2 kcal/mol for the glycosidic bond cleavage step with GUO as substrate,14 and thus almost quantitatively reproduce the enthalpy-entropy balance. The deviation in the computed enthalpy and entropy is here reflected by the lower ∆G‡ of 12.8 kcal/mol obtained from the EVB simulations, compared to the experimentally observed value of 14.5 ± 1 kcal/mol. If the EVB free energy surface for the enzyme reaction is fitted directly to the experimental free energy barrier, instead of to the reference solution reaction, the calculated ∆H‡ and T∆S‡ become 16.9 ± 0.4 kcal/mol and 2.3 kcal/mol at 298 K, respectively. It should also be noted that the experimental ∆H‡ and T∆S‡ are derived using an Arrhenius plot with only two points, which makes it difficult assess the error inherent in the experiments. The calculated values of ∆H‡ and T∆S‡ at 298 K with INO as substrate are 15.0 kcal/mol and 2.3 kcal/mol, respectively, which is essentially equivalent to the results obtained for GUO (Table 1). The fact that the predicted enthalpy-entropy balance is identical for the 6-oxopurines (GUO


10

Page 11 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

and INO) indeed seems reasonable, given that they are both natural substrates for HsPNP with similar reaction rates.42 HsPNP is furthermore able to catalyze the reaction with ADO, but at much lower rates compared to the 6-oxopurines.42 Interestingly, our results for the 6aminopurine (ADO) show an activation enthalpy which is more than twice that of GUO and INO (Table 1). The high value of ∆H‡ = 31.4 kcal/mol in this case is, however, largely counterbalanced by the positive activation entropy term of T∆S‡ = 11.0 kcal/mol at 298 K. The consistently positive activation entropies for all three substrates is an intrinsic feature of the dianionic phosphate attack, where the double negative charge on the reactant becomes much more delocalized in the transition state. This was further verified by computing the thermodynamic activation parameters also for the uncatalyzed reference reaction in solution (Table S1 and Figure S1). As can be seen from Table S1, the glycosidic bond cleavage by inorganic phosphate dianion for all three substrates (INO, GUO and ADO) is predicted to have a large positive activation entropy, which is largely associated with a reduction of solvent polarization in the transition state. In the case of ADO it can be noted that HsPNP is still able to reduce the enthalpy of activation by 17.6 kcal/mol compared to the calculated value in water. The higher activation enthalpy for ADO compared to INO and GUO further predicts that the reaction rate with ADO is significantly more temperature sensitive compared to the 6-oxopurines. This is, however, not surprising given the low activity towards 6-aminopurines at room temperature. Elevated temperatures act beneficially on the reaction rates by reducing the unfavorable contribution from the enthalpy of activation. In our previous work we found that the transition state with ADO is less stabilized compared to INO and GUO.28 This largely originates from unfavorable interactions between the base and E201, which is repelled away from the original N1 position observed for the 6-


11

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 26

oxopurines (Figure 3). We have also previously shown that HsPNP is significantly more electrostatically preorganized for catalysis of the 6-oxopurines compared to the 6-aminopurine.28 Thus, the higher enthalpy of activation with adenosine originates from less enthalpic stabilization of the transition state.

Remote mutations affect the activation enthalpy-entropy balance The fact that the EVB simulations are able to reproduce the enthalpy-entropy balance for the glycosidic bond cleavage step with GUO is very encouraging and allows us to move on to investigate the enthalpy-entropy compensation induced by remote mutations. Ghanem et al. recently reported that the double mutation K22E and H104R (E:R) significantly changes the thermodynamics of HsPNP with GUO as substrate.14 As illustrated in Figure 4A, K22 and H104 are far away from the subunit’s active site with distances of around 28 and 35 Å from the purine base of the nucleoside in the active site. Capturing the effect of these mutations therefore made it necessary to expand our original simulation sphere with a radius of 25 Å (see above) to cover the entire homotrimer. This was achieved by soaking the enzyme in a spherical droplet of water molecules with a radius of 53 Å, which significantly increased the computational costs of the EVB simulations. Nevertheless, in order to obtain high precision Arrhenius plots, a total of 100 and 50 individual reaction free energy profiles were calculated at six temperatures in the range 283 to 308 K for HsPNP and HsPNP-E:R, respectively. The resulting Arrhenius plots are shown in Figure 4B. It is noteworthy that increasing the size of the simulation sphere has very little effect on the calculated thermodynamic activation parameters for the native enzyme with GUO as substrate (Table 2). In a recent study we demonstrated that the free energies of activation are literally


12

Page 13 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

independent of the simulation sphere size as long as the active site mobility is not affected.13 This shows that it is possible to obtain good estimates of the activation free energy with reduced simulation systems. However, as demonstrated in our previous study on cold-adapted trypsin, the enthalpy – entropy balance is dependent on the mobility of the enzyme surface.13 Thus, calculating the effects of remote mutations on the enthalpy – entropy balance makes it necessary to include a sufficiently large part of the enzyme surface in the simulation system. The EVB calculations yield an activation free energy of 14.4 ± 1.2 kcal/mol for HsPNP-E:R at 298 K, which is similar to that of native HsPNP and also consistent with the experimental rates (Table 2).14 This again underscores the general notion that mutations far away from the active site do usually not significantly affect the catalytic rates.12 The corresponding ∆∆H‡ and T∆∆S‡ values of 3.6 kcal/mol and 2.2 kcal/mol, respectively, are furthermore in very good agreement with the experimentally obtained ∆∆H‡ = 1.9 kcal/mol and T∆∆S‡ = 2.1 kcal/mol at 298 K. Structurally there are no observed differences along the reaction coordinate in the active site region between the wild-type and mutated enzyme. However, the calculated residual root mean square fluctuations (RMSF) at the reactant state reveals surface regions that have become more rigid in the mutated enzyme (Figure S2). In average, the enzyme flexibility is reduced from 0.48 Å to 0.44 Å, which corresponds to ~9 %, upon going from HsPNP to HsPNP-E:R. Thus, the mutations render the protein surface less flexible along with an increase in the activation enthalpy and entropy. Remarkably, this is the exact same effect as previously found for different temperature adapted enzymes. 10,11,13

The activation parameters of HsPNP-E:R are not altered towards BtPNP


13

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 26

As also recognized by Ghanem et al., the K22E and H104R mutations do not render the enthalpy and entropy of HsPNP towards bovine PNP (BtPNP).14 The E:R mutation leads to a minor increase in the experimentally measured reaction rates for the glycosidic bond cleavage step at 298 K (∆∆G‡ ≈ -0.2 kcal/mol) which is in the right direction with respect to BtPNP. However, the fact that the ∆G‡ becomes more similar does not mean that its thermodynamic components are also altered in the expected direction. The ∆∆G‡ obtained from the Arrhenius plots using the experimentally measured reaction rates (Table 2) is ~0 kcal/mol, which just reflects the negligible impact on the barrier within the experimental error bars. Moreover, ∆∆H‡ and T∆∆S‡ in BtPNP are −2.5 kcal/mol and 2.0 kcal/mol, respectively, compared to HsPNP, which is the exact opposite to the enthalpy-entropy compensation obtained with HsPNP-E:R. Nevertheless, these long-ranged mutations demonstrate how the protein surface residues can be used to tune the enzyme enthalpy-entropy balance without altering the activation free energy.

CONCLUSIONS We have in the present study examined the thermodynamic activation parameters for the glycosidic bond cleavage step in the reaction catalyzed by human PNP with three different substrates using extensive computer simulations. Actual calculations of the activation enthalpy and entropy for enzymatic reactions are still rather scarce, which mainly pertains to the difficulties related to sufficient configurational sampling. That is, the large number of degrees of freedom involved in condensed-phase systems requires extensive configurational sampling to obtain sufficient accuracies, typically below 0.2 kcal/mol. The approach pursued here involves running extensive molecular dynamics simulations and calculations of up thousands of free energy profiles, and then to obtain activation enthalpies and entropies analogous to the way


14

Page 15 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

experimental Arrhenius plots are derived. This way of calculating thermodynamic activation parameters turns out to be very efficient and yields near-quantitative agreement with experiments. The extensive MD simulations and free energy calculations presented here show that the thermodynamic activation parameters for the native substrate (GUO) for HsPNP are well reproduced. Predictions with INO as a substrate, which is also a 6-oxopurine, yield as expected virtually identical thermodynamic activation parameters as observed for GUO. In contrast, the activation enthalpy is increased by ~16 kcal/mol for the glycosidic bond cleavage with ADO as substrate, partly counteracted by a more favorable T∆S term, when compared to GUO and INO. The increased specificity for 6-oxopurines compared to 6-aminopurines can thus be explained in terms of electrostatic preorganization of the active site, favoring the changes in the charge distribution along the reaction coordinate for the former substrates. Mutations distant to the active site typically have very small effect on the catalytic rate. The double mutant of HsPNP examined here catalyzes the glycosidic bond cleavage of GUO with virtually identical rate as the native enzyme, but with an enthalpy-entropy compensation of about ~2 kcal/mol. It is quite remarkable the simulations reproduce the changes between enthalpy and entropy that take place with mutations ~30 Å from the active site. Furthermore, the work presented here provides strong support for the recently proposed reaction mechanism for HsPNP, where a singly protonated phosphate acts as the nucleophile to attack the neutral ribonucleoside.28 Not only are the thermodynamic activation parameters well reproduced, but also the enthalpy-entropy compensations occurring upon long-range mutations that conserve the catalytic rate.


15

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 26

AUTHOR INFORMATION Corresponding Author *[email protected] Author Contributions The manuscript was written through contributions of all authors. Funding Sources This work has been supported by the Research Council of Norway through a Centre of Excellence Grant (Grant No. 179568/V30). Notes The authors declare no competing financial interests. Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org. Glycosidic bond cleavage in water; average residual fluctuations (PDF)

REFERENCES (1) Fuxreiter, M., and Warshel, A. (1998) Origin of the catalytic power of acetylcholinesterase: Computer simulation studies. J. Am .Chem. Soc. 120, 183-194. (2) Åqvist, J., and Warshel, A. (1993) Simulation of enzyme-reactions using valence-bond forcefields and other hybrid quantum-classical approaches. Chem. Rev. 93, 2523-2544. (3) Åqvist, J., and Fothergill, M. (1996) Computer simulation of the triosephosphate isomerase catalyzed reaction. J. Biol. Chem. 271, 10010-10016. (4) Fothergill, M., Goodman, M. F., Petruska, J., and Warshel, A. (1995) Structure-energy analysis of the role of metal-ions in phosphodiester bond hydrolysis by DNApolymerase-i. J. Am. Chem. Soc. 117, 11619-11627. (5) Hwang, J. K., and Warshel, A. (1996) How important are quantum mechanical nuclear motions in enzyme catalysis? J. Am. Chem. Soc. 118, 11745-11751. (6) Adamczyk, A. J., Cao, J., Kamerlin, S. C., and Warshel, A. (2011) Catalysis by dihydrofolate reductase and other enzymes arises from electrostatic preorganization, not conformational motions. Proc. Natl. Acad. Sci. U.S.A. 108, 14115-14120.


16

Page 17 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

(7) Pisliakov, A. V., Cao, J., Kamerlin, S. C. L., and Warshel, A. (2009) Enzyme millisecond conformational dynamics do not catalyze the chemical step. Proc. Natl. Acad. Sci. U.S.A. 106, 17359-17364. (8) Warshel, A. (1991) Computer modeling of chemical reactions in enzymes and solutions, Wiley, New York. (9) Warshel, A. (1978) Energetics of enzyme catalysis. Proc. Natl. Acad. Sci. U.S.A 75, 52505254. (10) Isaksen, G. V., Åqvist, J., and Brandsdal, B. O. (2014) Protein surface softness is the origin of enzyme cold-adaptation of trypsin. PLoS Comput. Biol. 10, e1003813. (11) Bjelic, S., Brandsdal, B. O., and Åqvist, J. (2008) Cold adaptation of enzyme reaction rates. Biochemistry 47, 10049-10057. (12) Morley, K. L., and Kazlauskas, R. J. (2005) Improving enzyme properties: When are closer mutations better? Trends. Biotechnol. 23, 231-237. (13) Isaksen, G. V., Åqvist, J., and Brandsdal, B. O. (2016) Enzyme surface rigidity tunes the temperature dependence of catalytic rates. Proc. Natl. Acad. Sci. U.S.A. 113, 7822-7827. (14) Ghanem, M., Li, L., Wing, C., and Schramm, V. L. (2008) Altered thermodynamics from remote mutations altering human toward bovine purine nucleoside phosphorylase. Biochemistry 47, 2559-2564. (15) Stoeckler, J. D., Cambor, C., and Parks, R. E., Jr. (1980) Human erythrocytic purine nucleoside phosphorylase: Reaction with sugar-modified nucleoside substrates. Biochemistry 19, 102-107. (16) Ealick, S. E., Rule, S. A., Carter, D. C., Greenhough, T. J., Babu, Y. S., Cook, W. J., Habash, J., Helliwell, J. R., Stoeckler, J. D., Parks, R. E., Jr., and et al. (1990) Threedimensional structure of human erythrocytic purine nucleoside phosphorylase at 3.2 a resolution. J. Biol. Chem. 265, 1812-1820. (17) Narayana, S. V., Bugg, C. E., and Ealick, S. E. (1997) Refined structure of purine nucleoside phosphorylase at 2.75 a resolution. Acta Crystallogr.. Sect. D: Biol. Crystallogr. 53, 131-142. (18) Bzowska, A., Luic, M., Schroder, W., Shugar, D., Saenger, W., and Koellner, G. (1995) Calf spleen purine nucleoside phosphorylase: Purification, sequence and crystal structure of its complex with an n(7)-acycloguanosine inhibitor. FEBS Lett. 367, 214-218. (19) Koellner, G., Luic, M., Shugar, D., Saenger, W., and Bzowska, A. (1997) Crystal structure of calf spleen purine nucleoside phosphorylase in a complex with hypoxanthine at 2.15 a resolution. J. Mol. Biol. 265, 202-216. (20) Mao, C., Cook, W. J., Zhou, M., Federov, A. A., Almo, S. C., and Ealick, S. E. (1998) Calf spleen purine nucleoside phosphorylase complexed with substrates and substrate analogues. Biochemistry 37, 7135-7146. (21) Wielgus-Kutrowska, B., Bzowska, A., Tebbe, J., Koellner, G., and Shugar, D. (2002) Purine nucleoside phosphorylase from cellulomonas sp.: Physicochemical properties and binding of substrates determined by ligand-dependent enhancement of enzyme intrinsic fluorescence, and by protective effects of ligands on thermal inactivation of the enzyme. Biochim. Biophys. Acta 1597, 320-334. (22) Tebbe, J., Bzowska, A., Wielgus-Kutrowska, B., Schroder, W., Kazimierczuk, Z., Shugar, D., Saenger, W., and Koellner, G. (1999) Crystal structure of the purine nucleoside phosphorylase (pnp) from cellulomonas sp. and its implication for the mechanism of trimeric pnps. J. Mol. Biol. 294, 1239-1255.


17

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 26

(23) Murkin, A. S., Birck, M. R., Rinaldo-Matthis, A., Shi, W., Taylor, E. A., Almo, S. C., and Schramm, V. L. (2007) Neighboring group participation in the transition state of human purine nucleoside phosphorylase. Biochemistry 46, 5038-5049. (24) Kicska, G. A., Tyler, P. C., Evans, G. B., Furneaux, R. H., Shi, W., Fedorov, A., Lewandowicz, A., Cahill, S. M., Almo, S. C., and Schramm, V. L. (2002) Atomic dissection of the hydrogen bond network for transition-state analogue binding to purine nucleoside phosphorylase. Biochemistry 41, 14489-14498. (25) Fedorov, A., Shi, W., Kicska, G., Fedorov, E., Tyler, P. C., Furneaux, R. H., Hanson, J. C., Gainsford, G. J., Larese, J. Z., Schramm, V. L., and Almo, S. C. (2001) Transition state structure of purine nucleoside phosphorylase and principles of atomic motion in enzymatic catalysis. Biochemistry 40, 853-860. (26) Canduri, F., Fadel, V., Basso, L. A., Palma, M. S., Santos, D. S., and de Azevedo, W. F., Jr. (2005) New catalytic mechanism for human purine nucleoside phosphorylase. Biochem. Biophys. Res. Commun. 327, 646-649. (27) Erion, M. D., Stoeckler, J. D., Guida, W. C., Walter, R. L., and Ealick, S. E. (1997) Purine nucleoside phosphorylase. 2. Catalytic mechanism. Biochemistry 36, 11735-11748. (28) Isaksen, G. V., Hopmann, K. H., Åqvist, J., and Brandsdal, B. O. (2016) Computer simulations reveal substrate specificity of glycosidic bond cleavage in native and mutant human purine nucleoside phosphorylase. Biochemistry 55, 2153-2162. (29) Porter, D. J. (1992) Purine nucleoside phosphorylase. Kinetic mechanism of the enzyme from calf spleen. J. Biol. Chem. 267, 7342-7351. (30) Bzowska, A. (2002) Calf spleen purine nucleoside phosphorylase: Complex kinetic mechanism, hydrolysis of 7-methylguanosine, and oligomeric state in solution. Biochim. Biophys. Acta 1596, 293-317. (31) Ghanem, M., Saen-oon, S., Zhadin, N., Wing, C., Cahill, S. M., Schwartz, S. D., Callender, R., and Schramm, V. L. (2008) Tryptophan-free human pnp reveals catalytic site interactions. Biochemistry 47, 3202-3215. (32) Shi, W., Ting, L. M., Kicska, G. A., Lewandowicz, A., Tyler, P. C., Evans, G. B., Furneaux, R. H., Kim, K., Almo, S. C., and Schramm, V. L. (2004) Plasmodium falciparum purine nucleoside phosphorylase: Crystal structures, immucillin inhibitors, and dual catalytic function. J. Biol. Chem. 279, 18103-18106. (33) Suarez, J., Haapalainen, A. M., Cahill, S. M., Ho, M. C., Yan, F., Almo, S. C., and Schramm, V. L. (2013) Catalytic site conformations in human pnp by 19f-nmr and crystallography. Chem. Biol. 20, 212-222. (34) Isaksen, G. V., Andberg, T. A., Åqvist, J., and Brandsdal, B. O. (2015) Qgui: A highthroughput interface for automated setup and analysis of free energy calculations and empirical valence bond simulations in biological systems. J. Mol. Graph. Model. 60, 1523. (35) Marelius, J., Kolmodin, K., Feierberg, I., and Åqvist, J. (1998) Q: A molecular dynamics program for free energy calculations and empirical valence bond simulations in biomolecular systems. J. Mol. Graph. Model. 16, 213-225, 261. (36) Jorgensen, W. L., Chandrasekhar, J., Madura, J. D., Impey, R. W., and Klein, M. L. (1983) Comparison of simple potential functions for simulating liquid water. J. Chem. Phys. 79, 926-935. (37) Lee, F. S., and Warshel, A. (1992) A local reaction field method for fast evaluation of longrange electrostatic interactions in molecular simulations. J. Chem. Phys. 97, 3100-3107.


18

Page 19 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

(38) Bochevarov, A. D., Harder, E., Hughes, T. F., Greenwood, J. R., Braden, D. A., Philipp, D. M., Rinaldo, D., Halls, M. D., Zhang, J., and Friesner, R. A. (2013) Jaguar: A highperformance quantum chemistry software program with strengths in life and materials sciences. Int. J. Quantum. Chem. 113, 2110-2142. (39) Jorgensen, W. L., Maxwell, D. S., and TiradoRives, J. (1996) Development and testing of the opls all-atom force field on conformational energetics and properties of organic liquids. J. Am. Chem. Soc. 118, 11225-11236. (40) Kaminski, G. A., Friesner, R. A., Tirado-Rives, J., and Jorgensen, W. L. (2001) Evaluation and reparametrization of the opls-aa force field for proteins via comparison with accurate quantum chemical calculations on peptides. J. Phys. Chem. B 105, 6474-6487. (41) Ryckaert, J. P., Ciccotti, G., and Berendsen, H. J. C. (1977) Numerical-integration of cartesian equations of motion of a system with constraints - molecular-dynamics of nalkanes. J. Comput. Phys. 23, 327-341. (42) Stoeckler, J. D., Poirot, A. F., Smith, R. M., Parks, R. E., Jr., Ealick, S. E., Takabayashi, K., and Erion, M. D. (1997) Purine nucleoside phosphorylase. 3. Reversal of purine base specificity by site-directed mutagenesis. Biochemistry 36, 11749-11756.


19

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 20 of 26

Tables Table 1. Thermodynamic activation parameters (25 Å sphere) calculated at 298 K for the glycosidic bond cleavage step in native human purine nucleoside phosphorylase. ∆ ‡

∆ ‡

∆ ‡

/'

INO

15.2 ± 0.7

2.4 ± 0.7

12.8 ± 1.0

15.8 ± 0.7

GUO

15.0 ± 0.4

2.3 ± 0.4

12.7 ± 0.6

15.6 ± 0.4

ADO

31.4 ± 0.7

11.0 ± 0.7

20.4 ± 1.0

32.0 ± 0.7

Energies in kcal/mol.

Table 2. Calculated thermodynamic activation parameters (53 Å sphere) at 298 K for the glycosidic bond cleavage step with guanosine bound to HsPNP, E:R-PNP together with experimental data including BtPNP. HsPNP Parameter

E:R-PNP

BtPNP

EVB

expt*

EVB

expt*

EVB

expt*

∆ ‡

14.7 ± 0.8

18.6 ± 0.7

18.3 ± 1.0

20.5 ± 0.7

-

16.1 ± 0.7

∆ ‡

1.6 ± 0.8

4.1 ± 0.2

3.8 ± 1.0

6.2 ± 0.2

-

2.1 ± 0.1

/'

15.3 ± 0.8

19.1 ± 0.7

18.9 ± 1.0

20.8 ± 0.7

-

16.7 ± 0.7

∆ ‡

13.1 ± 1.2

14.5 ± 1.0

14.4 ± 1.2

14.3 ± 1.0

-

13.9 ± 0.7

Energies in kcal/mol *Experimental data adapted from Ghanem et al. 14.


20

Page 21 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figures legends

Figure 1. A Molecular structures of inosine (INO), guanosine (GUO) and adenosine (ADO) with the general atom-numbering scheme illustrated for INO. B Resonance structures (φ1 and φ2) utilized to model the glycosidic bond cleavage step in HsPNP. Figure 2. A Calculated reaction free energy profiles at 298 K for the glycosidic bond cleavage step in water (dashed lines) and in HsPNP (solid lines) with adenosine (red), guanosine (black) and inosine (blue) as substrates. B Computed Arrhenius plots for guanosine (black), inosine (blue) and adenosine (green) for the glycosidic bond cleavage step in HsPNP. Figure 3. Snapshot of the transition state in HsPNP with adenosine as substrate. The red dashed lines illustrate the reaction coordinate. Figure 4. A Illustration of the relative position of the distant residues K22 and H104 relative to the active site in HsPNP. B Calculated Arrhenius plots for the glycosidic bond cleavage step in native and mutated (E:R) HsPNP with guanosine as substrate.


21

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 22 of 26

Figure 1


22

Page 23 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 2


23

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 24 of 26

Figure 3


24

Page 25 of 26

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Biochemistry

Figure 4


25

Biochemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 26 of 26

TOC Graphics


26

Thermodynamics of the Purine Nucleoside ... - ACS Publications

Recommend Documents