Carnosine and Homocarnosine Degradation Mechanisms by the

Apr 22, 2016 - This belongs to the metallopeptidase M20 family, where a cocatalytic active site is formed by two Zn2+ ions, bridged by a hydroxide ani...
0 downloads 10 Views 2MB Size
Article

Subscriber access provided by SUNY DOWNSTATE

Carnosine and Homocarnosine Degration Mechanisms by the Human Carnosinase Enzyme CN1: Insights from Multiscale Simulations Matic Pavlin, Giulia Rossetti, Marco De Vivo, and Paolo Carloni Biochemistry, Just Accepted Manuscript • DOI: 10.1021/acs.biochem.5b01263 • Publication Date (Web): 22 Apr 2016 Downloaded from http://pubs.acs.org on April 24, 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 59

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

CARNOSINE AND HOMOCARNOSINE DEGRATION MECHANISMS BY THE HUMAN CARNOSINASE ENZYME CN1: INSIGHTS FROM MULTISCALE SIMULATIONS Matic Pavlin1,2, Giulia Rossetti2,3,4*, Marco De Vivo2,5, Paolo Carloni1,2

1Laboratory

for Computational Biophysics, German Research School for Simulation Sciences (GRS), Forschungszentrum Jülich – RWTH Aachen, 52425 Jülich, Germany

2Computational

Biomedicine section (INM-9), Institute for Neuroscience and Medicine (INM), and

(IAS-5), Institute of Advanced Simulation (IAS), Forschungszentrum Jülich, 52425 Jülich, Germany

3Jülich

Supercomputing Center (JSC), Forschungszentrum Jülich, 52425 Jülich, Germany

4Department

of Oncology, Hematology and Stem Cell Transplantation, University Hospital Aachen, RWTH Aachen University, Pauwelsstraße 30, 52074 Aachen, Germany

5Laboratory

of Molecular Modeling and Drug Discovery, Istituto Italiano di Tecnologia, Via Morego 30, 16163 Genoa, Italy

1 ACS Paragon Plus Environment

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

Abbreviations. Lcar, L-carnosine; Hcar, homocarnosine; hCN1, human serum carnosinase CN1; hCN2, human tissue carnosinase; MD, molecular dynamics; QM/MM, quantum mechanics / molecular mechanics; SAS, solvent accessible surface Footnotes. aIn human body homocarnosine is found in the central nervous system.14, 18 bThe study of the other two substrates of the enzyme, namely of anserine and ophidine, provides less information than that of Hcar: their affinity for hCN1 is similar to the one of Lcar, while their enzymatic hydrolysis is 3-4 fold slower17 (kcat values were not reported, see Tab. S1). Hence, these compounds are not further discussed in this work. cThe biggest difference between experimental and calculated 13C chemical shifts was observed at Cγ, where experimental values also differ from each other for 16.9 ppm. dAb initio model calculations show that these two tautomers are far more stable than any other tautomer in the binding site (see SI Section 3 and Tab. S16 and S17). eIn addition, the distances between both zinc ions and oxygen of bridging moiety are slightly shorter in both Michaelis complexes (see Tab. S19, S20). The changes of angle values of residues coordinating both zinc ions (reported in Tab. S21, S22) are consequences of above described changes in distances. fThe structure of LcarhCN1 complex at 303 K was very similar to the one obtained at 298 K (see Tab. S20 and S22).

2 ACS Paragon Plus Environment

Page 2 of 59

Page 3 of 59

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

Abstract. The endogenous dipeptide L-carnosine, and its derivative homocarnosine, prevent and reduce several pathologies like ALS, Alzheimer’s and Parkinson’s disease. Their beneficial action is severely hampered due to the hydrolysis by carnosinase enzymes, in particular the human carnosinase, hCN1. This belongs to the metallopeptidases M20 family, where a co-catalytic active site is formed by two Zn2+ ions, bridged by a hydroxide anion. The protein may exist as a monomer and as a dimer in vivo. Here we used hybrid quantum mechanics/molecular mechanics simulations based on the dimeric apo enzyme’s structural information to predict the Michaelis complexes with L-carnosine and its derivative homocarnosine. Based on our calculations we suggest that (i) Lcarnosine degradation occurs through a nucleophilic attack of a Zn2+-coordinated bridging moiety for both monomer and dimer. This mechanistic hypothesis for hCN1 catalysis differs from previous proposals, while it is in agreement with available experimental data. (ii) The experimentally measured larger affinity of homocarnosine for the enzyme relative to L-carnosine might be explained, at least in part, by more extensive interactions inside the monomeric and dimeric hCN1’s active site. (iii) Hydrogen bonds at the binding site, present in the dimer but absent in the monomer, might play a role for the experimentally observed higher activity of the dimeric form. Investigations of the enzymatic reaction are required to establish or disprove this hypothesis. Our results may serve as a basis for the design of potent hCN1 inhibitors.

3 ACS Paragon Plus Environment

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

The dipeptide L-carnosine (β-alanyl-L-histidine, Lcar hereafter) plays an important role for human biology. It is mainly present in skeletal muscles and in the central nervous system1, 2, where it acts as physiological buffer, wound-healing promoter, copper(II) and zinc(II)-chelating agent, antioxidant and free-radical scavenger.3-5 Lcar shows a variety of beneficial activities, including a protective ability against diabetes, osteoporosis, loss of vision and hearing, as well as, it shows immune function ability.6 It is an anti-inflammatory agent and it may reduce or prevent ALS, Alzheimer’s and Parkinson’s diseases.7-10 Its complexes with Zn2+ ions are effective for the repair of ulcers and other gut lesions.11 Unfortunately, therapeutic applications are drastically limited due to Lcar’s hydrolysis by the serum12-14 and tissue14-16 carnosinase enzymes, hCN1 and hCN2, respectively. hCN1, unlike hCN2, hydrolyses selectively Lcar and very few other dipeptides, including the close homologous homocarnosinea (γ-amino-butyryl-L-histidine, Hcar hereafter, Chart 1).12, 17 Hcar acts as antioxidant, free-radical scavenger and metal-chelating agent.18 Interestingly, hCN1 is the only known human enzyme that cleaves both Lcar and Hcar.2 Hcar binds stronger to hCN1 in vitro than Lcar (Km = 0.2 mM and 1.27 mM for Hcar and Lcar respectively).b12 The enzyme has been reported to be a homodimer in aqueous solution.12 It has the same multimeric state also in the X-ray structure of the apo-enzyme (PDB ID 3DLJ; article not published 4 ACS Paragon Plus Environment

Page 4 of 59

Page 5 of 59

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

yet, Fig. 1a). Each monomer unit (I and II hereafter) consists of a catalytic and a dimerization domain.12 The catalytic domain features a di-nuclear Zn2+ active site (Zn1 and Zn2 in Fig. 1b). Zn1 binds to one of the two carboxylate oxygens of Glu174 and the imidazole nitrogen of His452.14 Zn2 binds to the imidazole nitrogen of His106 and to one of the two carboxylate oxygens of Asp202. Each carboxylate oxygen of Asp139 coordinates one metal ion to form a bridge-like structure between the two metal ions.14 An unknown chemical group bridges the Zn2+ ions. Comparison across the available structures of members of the metallopeptidase M20 family (aminopeptidase from Aeromonas proteolytica19 (PDB ID 1AMP), allantoate amidohydrolase from E. Coli20 (PDB ID 1Z2L) and yeast β-alanine synthase in the free form and in complex with its substrate NCβA21 (PDB ID 2V8D and 2V8H)), leads us to suggest that this group is either a water molecule or a hydroxide anion, depending whether enzyme is in an apo or in a holo state (OW in Fig. 1b, se also Fig. S1 and S2). Although no structural information of the enzyme in complex with inhibitors is available, it is plausible to suppose that the hydrolysis reaction occurs as in other peptidases with identical active sites, namely via OW’s nucleophilic attack on substrates’ carbonyl carbon (Scheme 1).22, 23 Although hCN1 is a dimer in vitro12, 13, 24, the enzyme exists both as a monomer and as a dimer in vivo.2 The content in humans of the latter seems to increase with age, becoming predominant in

5 ACS Paragon Plus Environment

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

adults.2 The activity of the monomer is lower than that of the dimer, suggesting that possible rearrangement of the structure upon dimerization may affect the enzymatic function.25 In this paper, we used molecular simulations in an effort aimed at (i) providing a molecular basis for the stronger binding of Hcar relative to Lcar to hCN1 and (ii) elucidating the differences in structural determinants between monomer and dimer forms of hCN1, which contribute to explaining experimentally enzymatic activity in vivo.2 We performed classical molecular dynamics (MD) and hybrid Car-Parrinello QM/MM calculations of the adduct between Lcar and Hcar with either monomeric or dimeric hCN1, based on the available structural information. Anticipating our results, we find that Hcar forms a larger number of hydrogen bonds with the enzyme than Lcar. This might contribute to the Hcar’s stronger binding to hCN1 relative to Lcar. In addition, both substrates form two additional hydrogen bonds only when bound to the dimer. Notably, one of those two hydrogen bonds involves the carbonyl oxygen of the substrate. This hydrogen bond polarizes the electron density of the C=O, likely favoring substrate hydrolysis. Methods Lcar and Hcar in water. The four tautomers (A, B, C and D in Chart 1) were built using Molden.26 They first underwent geometry optimization using B3LYP functional27 and 6-31G(d,p) basis set in implicit solvent.28 IEF-PCM type was used for implicit solvent with water as solvent (ε = 78.3553). 6 ACS Paragon Plus Environment

Page 6 of 59

Page 7 of 59

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

The atomic radii were scaled by factor 1.1 from those of UFF force field. Absence of calculated imaginary frequencies confirmed that the optimized structures correspond to a local minimum on the potential energy surface.29 The quantum chemical calculations were performed using software package Gaussian09.30 Next, we performed µs-long classical molecular dynamics (MD) in aqueous solution at physiological pH and room temperature of all four tautomer in order to investigate their predominant protonation states. Simulations were based on the AMBER99SB31 force field with ILDN modification.32 The partial charges of both tautomeric forms for Hcar and Lcar (Chart 1) were derived from the DFT calculations using the RESP algorithm.33 The charges were fitted on the electrostatic potential grid calculated by CHelpG scheme34 using Antechamber module.35 The parameters for β-alanine and γ-amino-butyryl were obtained with the Antechamber module35, using as input the bond lengths and bond angles data from the DFT calculations (see Tab. S2-S4). The TIP3P potential was used for water.36 The four tautomers were inserted in an octahedron shaped box for MD simulation. The solute was solvated with at least 12 Å of water molecules in each direction. Periodic boundary conditions were applied. All simulations were performed at 298 K and 1 bar by coupling to Nosé-Hoover thermostat37, 38 and Parrinello-Rahman barostat.39 The long range electrostatic interactions were taken into account using particle mesh Ewald method.40 Time step in simulations was 1 fs. Each system was heated up to the final temperature of 298 K by using 7 ACS Paragon Plus Environment

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

20 steps of simulated annealing. In the first step, the temperature was increased from 0 to 10 K in 50 ps. In the following 8 steps, the temperature was increased for 10 K in each step of 25 ps. In the last 11 steps, the system was heated 20 K in each step of 25 ps. The convergence of sampling in our simulations was checked using cosine content of principal components (see SI Section 1 and Tab. S5).41 Further independent simulations for both A and B were performed to test the robustness of our computational protocol. The trajectories were clustered based on their RMSD value using the gromos method42 implemented in the g_cluster analysis tool of the Gromacs software.43, 44 This clustered the MD trajectory based on the RMSD values of the ligands’ atoms. First, neighbors of each data point (i. e. snapshot) are defined based on a selected RMSD cutoff value (in our case 1.1 Å for Lcar and 1.3 Å for Hcar). Next, the data point with the largest number of neighbors is defined as the central structure of the first cluster, then this point and its neighbors defining this cluster are removed. The algorithm is finally iterated until all the data points have been assigned to a cluster. By definition, the central point of each cluster is the one with the most neighbors (within a cutoff distance). Therefore, this point is a representative of all its neighbors within the given cut-off distance. The following properties were calculated: (i) the square of the gyration radius defined as Rg2 = Σi((rirg)2/N), where ri is the position of the atom i, rg is position of the molecular center and N is number of atoms.45 (ii) The free energies of representative structures of four highly populated clusters from 8 ACS Paragon Plus Environment

Page 8 of 59

Page 9 of 59

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

A and B for Lcar (models 1-4A and 1-4B) and from C and D for Hcar (models 1-4C and 1-4D) were calculated as single point energies using B3LYP functional27 and 6-311+G(2d,p) basis set in implicit solvent28 with Gaussian09 software package.30 So obtained free energy values were used in the calculation of Maxwell-Boltzmann statistics using Eq. 1: / = e^(-EX/kBT) / e^(-E0/kBT) (Eq. 1) where X is naming of the model, 0 stands for the lowest energy conformation, i.e. model 1B for Lcar and 2D for Hcar, is the number of particles in the state X, EX is the energy of corresponding conformation, kB is Boltzmann constant and T is temperature. We used the temperature at which the NMR experiments were performed45 (T = 298 K). (iii) 1H and 13C chemical shifts of each representative structure of the four highly populated clusters from all four tautomers were calculated (protocol of chemical shifts calculations is described in SI Section 2). They were weighted based on our Maxwell-Boltzmann statistics analysis and compared with experimental data.45-49 Monomeric and dimeric hCN1. The X-ray crystal structure of the human apo enzyme (solved at 2.26 Å, PDB ID 3DLJ, paper not published yet) is a homodimer. Each monomer contains an unknown residue, which coordinates both Zn2+ ions. Based on the arguments discussed in the Introduction, we suggest that this unknown atom is an oxygen atom of a water molecule in the apo state and a hydroxide anion in the holo state (Fig. 1b). 9 ACS Paragon Plus Environment

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

Monomer I lacks residues 1-3, 77, 78, 208, 209, 437, 438 and 481. The other (II) lacks residues 17, 77-79, 208, 209 413 and 481 (numbering as in the X-ray structure). They were added using the Modeller9.950 code. The 26 residues (-25 to 0) belonging to the signal peptide12 were not considered in our model, since this part is absent in the crystal structure and is believed not to play an active role in the activity of the enzyme. The histidine protonation states were determined using version 3.1 of H++ webserver.51-53 Tab. S6 reports the assignment of histidine protonation states as well as other non-standard protonation states. The best initial models according to the Modeller's DOPE score54 was optimized using loop refinement procedure.55 All the models were ranked using DOPE score and evaluated using PROCHECK software.56 The best structure of dimeric hCN1 according to both scores was used. Our initial model of monomer hCN1 was obtained by simply removing monomer II from the dimer. The adducts of Lcar and Hcar with the monomer and the dimer were then constructed by molecular docking, followed by classical MD and QM/MM simulations. Lcar and HcarhCN1 complex. The structures of the Michaelis complex of the monomer with the most favorable tautomer of Lcar (tautomer B) and Hcar (tautomer D), according to our simulations (and consistent with experimental data47), were determined with a two-step molecular docking calculations.

10 ACS Paragon Plus Environment

Page 10 of 59

Page 11 of 59

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

(i) Step 1. We docked both ligands to the refined enzyme structure and with water molecule bridging the two Zn2+ ions. Since it is reasonable to assume that position of the bridging water molecule would change upon the ligand binding with respect to the apo state of the enzyme, we used a code (GOLD57) able to treat water molecules flexibly.58 The complexes obtained by docking were ranked using Goldscore59, 60 and then rescored using ChemPLP scoring function.61 Six complexes with the highest score were used for further analysis and from these structures we determined average position of bridging water molecule/hydroxide anion (Tab. S7). So-obtained structure of the apo enzyme was used in the second step of docking. (ii) Protonation states of hCN1 active site. In the dipeptide hydrolysis in metallopeptidases with a similar catalytic active site, the bridging water molecule during the course of the reaction gives one of its protons to a nearby residue. Then, the resulting hydroxide anion performs nucleophilic attack on the peptide bond of the dipeptide.22 Based on a structural similarity analysis of hCN1 across dizinc peptidases for which structural information is available (namely, the aminopeptidase from Streptomyces griseus (PDB ID 1CP7)62 and aminopeptidase from Aeromonas proteolytica (PDB ID 1AMP19) and hCN1 (PDB ID 3DLJ)), the residue accepting the proton is expected to be Glu173 (Fig. 1b). To verify our hypothesis, we determined which residue preferably picks up the proton from the water molecule. Therefore, we defined our model system as constituted by all the residues and the Zn2+ ions in the active site of the hCN1 in apo state and with the position of the bridging water 11 ACS Paragon Plus Environment

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

molecule as obtained in the first step of docking. Next, we built modifications of the model system by varying the position of the hydrogen atom, subtracted from the water molecule, among all the possible nearby proton acceptors (Glu173, Glu174 and Asp202). Finally, we performed geometry minimizations and calculated potential energies of the model system and its modifications in implicit solvent.28 We used IEF-PCM model with atomic radii from UFF force field scaled by 1.1. Different dielectric constant (ε) values were used as suggested in ref.63 In these calculations all atoms except hydrogen ones were frozen in order to maintain correct geometry of the active site. The calculations were carried out at the B3LYP27 level of theory, using the 6-31+G(d,p) basis set . Our calculations confirm that Glu173 picks the proton (named Glh173 hereafter), either at Oε2 or Oε1 oxygen atom of carboxylic group (resulting in tautomers hCN1α and hCN1β; Fig. 2). The differences in potential energy by modifying the position of proton from Glh173 to Glu174 and Asp202 turned out to be large (15.8 kcal/mol or higher). Therefore it is reasonable to hypothesize (although it is not proven) that the entropic contribution would not modify such a trend. (iii) Step 2. Based on the results from previous step, we performed new set of docking calculations with hydroxide anion and starting from two different models of the monomer, monomer hCN1α and monomer hCN1β. Here we used AutoDock version 4.2 docking software64 for which an improved force field for docking of small molecules to zinc metalloproteins has recently been developed.65 12 ACS Paragon Plus Environment

Page 12 of 59

Page 13 of 59

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

Following the approach for docking of small ligands to zinc metalloenzymes used by Park et al.66, we determined all atomic charges of enzyme and substrates. Specifically, for the atomic charges of the hCN1 active site we followed the same approach as it is used in Metal Center Parameter Builder (MCPB)67 of AmberTools, where model system of the active site is created and its charges are determined using RESP methodology33 on B3LYP/6-31G(d) level of theory. So calculated charges are then assigned to all atoms considered in the model, except for the backbone heavy atoms (i.e. atoms forming peptide bonds), which are restrained to the values from AMBER94 force field.68 Our model systems of the active site consisted of both Zn2+ ions, hydroxide ion, all residues coordinating Zn2+ and the additional residue Glh173. These residues were capped with acetyl and Nmethylamine residues. The scheme used adjusts the charge of capped residues to an integer value, thus allowing the formal charge of the cluster to disperse over the metal and the bound ligands.67 Single point calculations were performed to preserve geometry of the active site within a computationally inexpensive procedure. For the rest of the enzyme atomic charges were taken from the AMBER94 force field.68 Substrates’ atomic charges were calculated using RESP method33 on the structure of the substrates optimized with B3LYP functional27 and 6-31G(d) basis set. The docking calculations were performed using Lamarckian genetic algorithm69 with 100 docking runs with the initial population of 150 individuals for each of the two substrates. The maximum number of generations and energy evaluations were set to 27,000 and 1×107. The enzyme and both, 13 ACS Paragon Plus Environment

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

the Zn2+ ions and the hydroxide anion, were kept rigid during the calculations, while in the substrates all bonds were defined as rotatable except the peptide one. The grid box had dimension of 51×51×51 points with spacing of 0.375 Å between them. The center of mass of two Zn2+ ions was selected as the grid box center. We considered the representatives of the highest populated clusters, based on the RMSD value, of Lcar and Hcar Michaelis complexes with monomeric hCN1α and monomeric hCN1β from this second step of docking simulations for the following calculations. To determine whether monomeric hCN1α or monomeric hCN1β is more stable in the presence of substrate, we performed additional ab initio scanning with the same setup as described above. The difference was that here active site model system, besides all the residues as in previously performed scanning, contained also the ligand. It is possible that the protonation state of ligands’ imidazole ring would change upon binding in the active site. Therefore, additional docking calculations to monomeric hCN1α were performed with the other two protonation states of both ligands (where imidazole ring is protonated on Nδ and with positively charged imidazole ring). The complexes obtained with docking calculations were then subject to 40 ns long classical MD simulations. Each system containing monomeric hCN1 was solvated using ~42,000 water molecules, with 135 Na+ and 121 Cl- ions ensuring neutrality and physiological conditions. Active 14 ACS Paragon Plus Environment

Page 14 of 59

Page 15 of 59

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

site residues as well as ligand, Zn2+ and hydroxide ions were constrained in order to preserve geometries from docking. For the rest of the protein and ions we used AMBER99SB31 force field with ILDN modification.32 TIP3P potential was used for water. Simulations were performed with periodic boundary conditions at 298 K and 1 bar by coupling to Nosé-Hoover thermostat37, 38 and Parrinello-Rahman barostat.39 The temperature of 298 K was selected as the structural ensemble of the ligands was obtained by performing simulations at this temperature. Electrostatic interactions were taken into account using particle mesh Ewald method.40 We applied the LINCS algorithm70 to constrain all bond lengths in order to achieve time step of 2 fs. From the plot of RMSD value of the proteins we assumed that relaxation runs were long enough (see Fig. S3a). The last snapshots were then taken for further QM/MM simulations. Here, each system was partitioned into two parts: one treated at the quantum mechanical level (the QM part), and another part treated at the force field level (the MM part). The AMBER99SB31 force field was used for the MM part. The QM region was identified by the proposed hCN1 active site (i.e. two Zn2+ ions, residues involved in the binding of these two ions – His106, Asp139, Glu174, Asp202 and His452 and Glh173, which is believed to be involved in the dipeptide hydrolysis22), the substrate Lcar/Hcar, hydroxide anion which acts as nucleophile in the dipeptide hydrolysis, and residues which can possibly form hydrogen bonds with the substrate (i.e. Asn203, Arg350, Ser423, Thr424, Asp421 and Glu451). Open valencies were saturated by introducing a bond capping pseudopotential at the 15 ACS Paragon Plus Environment

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

position of the first MM atom at the cutting border.71 The time step for the dynamics was 0.12 fs, and the fictitious electronic mass was chosen to be 600 au. An energy cutoff of 80 Ry and the gradient corrected PBE exchange-correlation functional were used.72 The core-valence electron interactions for the C, N, O, Zn and H atoms were described using non-conserving Martins and Trouiller pseudopotentials, corrected for a better description of van der Waals interactions.73, 74 Integration of the nonlocal parts of the pseudopotential was obtained via the Kleinman-Bylander scheme75 for all atoms except for Zn2+, for which a Gauss-Hermite numerical integration scheme was used.76 Electrostatic interactions between periodic images of the QM part were decoupled with the scheme of ref.77, while those of the classical part with periodic boundary conditions were treated by the particle-mesh method. A Nosé-Hoover thermostat37, 38 was used to maintain a constant temperature of 298 K, as in ref.45 in order to be consistent with the temperature of the simulation in water. In total, 2 ps of QM/MM simulations were performed. Additionally, 1 ps QM/MM simulation of Lcarmonomeric hCN1α was also performed at 303 K. This is the temperature at which Km and kcat were measured in vitro.12 The CPMD3.15.1 and GROMOS96 codes were used for the QM/MM calculations.78 The procedure for the docking to the dimer, followed by MD simulations and 2 ps QM/MM simulations. The substrates were docked in the active site of monomer I. For the molecular simulatios both dimeric complexes were solvated using ~93,000 water molecules, with 295 Na+ 16 ACS Paragon Plus Environment

Page 16 of 59

Page 17 of 59

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 267 Cl- ions ensuring neutrality and physiological conditions. The QM region in this case was the same as in monomeric complexes with additional His235 side chain from monomer II, which can form hydrogen bond with the substrate, accordingly to our docking results. The following properties were calculated: (i) the solvent accessible surface (SAS) was calculated using g_sas analysis tool of the Gromacs software.43, 44 (ii) The hydrogen bonds were identified with the Hydrogen Bonds analysis tool of VMD visualization software, respectively. Results Lcar and Hcar in water. These are zwitterions79 present in two tautomeric forms, depending on whether the histidine ring is protonated at the Nδ (tautomers A, C in Chart 1, for Lcar and Hcar, respectively) or at the Nε (tautomers B, D in Chart 1, for Lcar and Hcar, respectively).49, 80 The four systems first underwent ab initio geometry optimization. DFT geometry optimizations and subsequent free energies calculations (from frequency calculations) of ligand in implicit solvent performed at B3LYP/6-31G(d,p) level of theory, show that A and C tautomers of Lcar and Hcar are more stable than B and D ones (difference in free energies is 7.5 kcal/mol for both ligands). Namely, in A and C tautomers, the proton is located on the Nδ, which enables formation of an intramolecular hydrogen bond with the carboxylic group. On the other hand, in tautomers B and D, the proton is bound to Nε and the formation of an intramolecular hydrogen bond with the carboxylic group is sterically not possible. 17 ACS Paragon Plus Environment

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

Next, these four tautomers next underwent µs-long MD simulations. They were equilibrated already in the first ns of dynamics. The convergence of sampling in our simulations was checked using the Hess analysis41 (see SI Section 1 and Tab. S5 for details). Lcar’s and Hcar’s most populated tautomers turned out to be B and D, respectively, according to the Maxwell-Boltzmann distribution analysis (see Methods and Tab. 1). The difference between results from classical MD and DFT calculations can be explained, at least in part, by the different treatment of solvent environment. In classical MD simulations, explicit water molecules form hydrogen bonds with the ligand (see also Tab. S8). On the other hand, in the DFT calculations, solvent is described implicitly and therefore explicit solvent-ligand hydrogen bonds are absent. B and D tautomers assume predominantly an 'extended' conformation in water, as reported in ref.45 (see Fig. S5). This is defined in terms of the squared gyration radius Rg2: the tautomer is extended if Rg2 > 13.0 Å2 for Lcar and Rg2 > 15.0 Å2 for Hcar or 'semi-folded’ if 13 Å2 < Rg2 < 10.5 Å2 for Lcar and 12 Å2 < Rg2 < 15 Å2 for Hcar (Fig. 3, S5 and SI Section 3). The RMSD, distance between terminal NH3+ and COO-, accessible surface area, RMSF and radial distribution functions of A, compare well with previous classical MD study of this tautomer (Fig. S4S7, Tab. S8 and SI Section 3).45 The predictive power of our simulations was established by a quantitative comparison with solution NMR data. Lcar’s and Hcar’s theoretical chemical shifts are in fair agreement with experimental values (see Tab. S9-S15).45-49 They differ from experiments by 18 ACS Paragon Plus Environment

Page 18 of 59

Page 19 of 59

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

27.0 ppm for 13Cc and 1.8 ppm or less for 1H for Lcar and for 1.8 ppm or less for 1H for Hcar (see also SI Section 2 and Tab. S9-S15). Tautomers B and D, which were found to be the most favorable ones in aqueous solution, were therefore used for the subsequent calculations of the ligands in complex with the enzyme. Lcar and HcarhCN1 complexes. The structural determinants of apo hCN1 binding sites are very similar to those of the other dizinc peptidases for which structural information is available. These are the aminopeptidase from Streptomyces griseus (PDB ID 1CP7)62 and aminopeptidase from Aeromonas proteolytica (PDB ID 1AMP19, Fig. 1b and S1). The metal ions bind to an acidic residue (Glu174 for “Zn1” and Asp202 for “Zn2”, Fig. 1b), along with a histidine (Fig. 1b). An aspartate and water molecule bridges the two metal ions in the apo state. In the presence of the substrate, the bridging group is believed to be a hydroxide anion and the Glu173 (which hydrogen bonds to the dizinc-bound hydroxide anion) to be protonated (if protonated named Glh173; see also Fig. 1b and S1).62, 81 Either of the two carboxylic oxygens can be protonated (Oε1 and Oε2 in Fig. 2, forming hCN1β and hCN1α, hereafter).d The following models of the monomer and the dimer were considered: Lcar and Hcar in their most favorable tautomeric forms (B and D) in complexes with monomeric hCNα, monomeric hCN1β and dimeric hCN1α, as well as complexes between other tautomeric forms of both ligands with monomeric hCN1α (for details see SI Section 4, Fig. S8 and Tab. S18). The initial models of the 19 ACS Paragon Plus Environment

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

Michaelis complexes were obtained by molecular docking. Because of the presence of metal ions, we used DFT-based QM/MM simulations to describe substrate/enzyme interactions as in references.82, 83 Hence, the complexes underwent 2 ps long relaxation QM/MM simulations, preceded by 40 ns long classical MD-based equilibration (see Fig. S3). Let us discuss first the monomer adducts. Both, Lcarmonomeric hCN1α and Hcar monomeric hCN1α Michaelis complexes feature metal ions’ coordination polyhedra similar to those of the apo enzyme. The metal-ligand atom distances are similar except those between Zn1 and the bridging aspartate (Asp139, see Tab. S19, S20), which is in both Michaelis complex longer than in the X-ray structure. The presence of weaker Zn1-Asp139 interaction may be consistent with observations of aminopeptidase from Aeromonas proteolytica reported in ref.19, where it is stated that, during the catalysis, the bridging aspartate residue can dissociate from either or both Zn2+ ions.e,f The substrates’ histidine rings form a direct and a water-mediated hydrogen bond with Glu174 carbonyl group, for Lcar and Hcar, respectively (Fig. 4 and Tab. 2). The substrates’ amino terminal hydrogen forms water-mediated and direct hydrogen bond with Asp202 backbone unit, for Lcar and Hcar, respectively. That of Lcar forms also a salt bridge with the side chain of this residue. The substrates’ carboxylic group forms salt bridge with Arg350. Both substrates’ backbone amide groups form hydrogen bonds with Thr424 backbone unit. The latter interactions are more persistent in Hcar than Lcar (Tab. 2), though the calculated occurrences should be taken with care 20 ACS Paragon Plus Environment

Page 20 of 59

Page 21 of 59

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

since they might strongly depend on the time scale of the simulations. The substrates’ backbone carbonyl forms a water-mediated and direct hydrogen bond with Thr424 backbone, for Lcar and Hcar, respectively (Tab. 2). Hcar’s amino group interacts with Asp421 and Glu451 through watermediated salt bridges for most of the simulation time, while the same interaction is completely absent in the case of Lcar. Same hydrophobic interactions are present for both ligands, namely between (CH2)2 chain in Lcar and (CH2)3 in Hcar and the side chain of Ile425. Overall, Hcar appears to form more intermolecular interactions with the hCN1 than Lcar (Fig. 4 and Tab. 2). Therefore Hcar appears to be tightly bound inside the enzymatic active site, compared to Lcar. In addition, while both ligands feature similar flexibility inside the active site, Lcar’s peptide bond appears to be more stable (see Fig. S10a for RMSF of atoms of both ligands inside monomeric hCN1 active site), which could contribute to the Lcar’s higher propensity for hydrolysis. The orientation of both substrates in Lcarmonomeric hCN1β and Hcarmonomeric hCN1β Michaelis complexes is less favorable for hydrolysis to occur (see Fig. S11, Tab. S23 and SI Section 5 for a discussion). Hence, our QM/MM calculations give evidence that the Lcarmonomeric hCN1α and Hcarmonomeric hCN1α are the most probable tautomers. We further investigated this finding by ab initio calculations. In a model of the active site, we varied the position of the hydrogen atom among all the possible nearby proton acceptors (Glu173, Glu174 and Asp202). We calculated the relative stability of so-obtained tautomers at different 21 ACS Paragon Plus Environment

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

dielectric constant values (same approach of ref.63, see Methods for details). Consistently with the QM/MM calculations, the monomeric hCN1α tautomers turn out to be more stable than the hCN1β ones by as much as ~13 and ~9 kcal/mol in Lcarmonomeric hCN1 and Hcarmonomeric hCN1, respectively (see Tab. S24). The other tautomers are even less stable than hCN1β. We next turn our attention to the LcarhCN1 and HcarhCN1 complexes in their dimeric states. The MD and QM/MM structures, based on the most favorable docking binding pose (see SI Section 6) show that the ligands’ binding pose and most of the interactions described for the monomer between each ligand and hCN1 are indeed preserved (see Tab. 2 and Fig. 4). However, here Lcar’s and Hcar’s imidazole rings form, respectively, direct and water mediated hydrogen bond with the backbone carbonyl group of Glh173 (Fig. 4). In addition, the ligands’ carbonyl group forms hydrogen bond with the side chain of Thr337 from the second monomer (Fig. 4). Finally, and most importantly, the ligands’ carbonyl group forms one hydrogen bond with His235, also from the second monomer (Fig. 4 and Tab. 2). This H-bond might play a role in the enzymatic hydrolysis through both the polarization of the substrate’s C=O group and the conservation of a proper orientation for catalysis. The absence of this hydrogen bond in the monomeric complexes might indeed contribute to lower the enzymatic activity compared to the dimeric form. Investigations of the enzymatic reaction are required to further establish this aspect of hCN1 catalysis.

22 ACS Paragon Plus Environment

Page 22 of 59

Page 23 of 59

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

Overall, also in the dimeric hCN1 state, Hcar forms more extensive hydrogen bond network with respect to Lcar (see Tab. 2 and Fig. 4). This suggests that binding of Hcar to hCN1 is stronger than Lcar. This finding is consistent with the lower Km value of Hcar compared to the Lcar (Tab. S1).12 However, as already mentioned, the hydrogen bond occurrence, especially that of the watermediated hydrogen bonds, should be taken with care as they strongly depend on the time scale of the simulation. Interestingly, the absence of the above discussed hydrogen bond between the ligand’s carbonyl oxygen and the imidazole ring of His235 in the monomeric complexes might contribute to the experimentally observed lower enzymatic activity compared to the dimeric form.2, 13, 24 The overall fold of hCN1. The solvent accessible surface of the monomer, calculated from the classical MD, is not too dissimilar to that of the dimer (Tab. S25). However, the two structures differ in regions RI – RV (see Fig. 5a). In region RI (located between alpha helices 2 and 3), alpha helix 3 starts at residue 42 in the monomeric form and at residue 45 in the dimeric form and therefore also the whole loop region features different conformation (Fig. 5a). The turn in region RII (residues 77 – 79) is shifted towards the solvent of ~5 Å in the dimeric form compared to the monomeric one. The loop forming region RIII (residues 228 – 245), between beta strand 11 and alpha helix 8, is entangled in the active site of the cognate momomer (and vice-versa) in the dimeric form. In the monomeric form this loop is solvent-exposed. Region RIV (residues 256 and 309) displays a 23 ACS Paragon Plus Environment

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

difference in the position of the backbone atoms within the helical region. The most significant difference comes in the region RV (residues 328 – 340), which is entangled in the cognate monomeric unit in the case of dimeric hCN1. In the monomeric form, this region is stabilized by the salt bridge between Asp331 and Lys336 while this is sterically not possible in the case of the dimeric forms. The distribution of intramolecular hydrogen bonds is different on passing from the monomeric to the dimeric forms. The average number of intramolecular hydrogen bonds is higher in the monomer than in the dimer (Fig. 5b). Parts of the hydrogen bonds engaged in the formation of the dimer are indeed replaced by intramolecular hydrogen bond in the monomer. The number of intramolecular hCN1 hydrogen bonds is similar in the dimeric adducts of hCN1 with Lcar and Hcar. It is larger in the monomeric hCN1 adduct with Lcar relative to that of the Hcar adduct (see Fig. 5b). Discussion Despite the biological relevance of hCN1 inhibition for a variety of human diseases3-11, the current structural information for this enzyme is scarce. The enzyme exists as a monomer or a dimer in vivo2, yet only the structure of the dimer is available (PDB ID 3DLJ). Structural information on the Michaelis complexes with the two dipeptides, Lcar and Hcar, selectively cleaved by hCN1, has so far relied on the elegant study by Vistoli et al.84, which was however performed on the hCN1 monomer and before the X-ray structure of the enzyme had been reported. 24 ACS Paragon Plus Environment

Page 24 of 59

Page 25 of 59

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

Here we have predicted structures of Michaelis complexes formed between Lcar and Hcar with both monomeric and dimeric hCN1, by molecular simulations. First, we have identified the most predominant tautomers of the two peptides in aqueous solution at physiological pH by µs-long MD simulations (B and D in Chart 1). This result is consistent with available NMR data.45-49 Then, we have docked B and D on monomeric and dimeric hCN1 using flexible docking simulations followed by MD simulations and finally QM/MM calculations. Comparison between monomer and dimer. The pose of both ligands in the active site shares similarities between monomeric and dimeric hCN1. However, the ligands form two hydrogen bonds with the second monomer in the dimeric hCN1 (Fig. 4 and Tab. 2). Notably, these hydrogen bonds are absent in the monomeric complexes. As a result, while in the monomer, the backbone oxygen is orientated towards the Zn ion, while in the dimer it is oriented toward His235. Such feature might play a role for the experimentally observed higher activity of the dimer hCN1 with respect the monomer.13, 24, 25, 84 Indeed, one of those two additional hydrogen bonds might shift the electron density in the C=O bond away the oxygen, therefore making the carbonyl carbon more prone to the nucleophilic attack of the bridging hydroxide anion. This hypothesis might be consistent with mutagenesis experiments on the structurally similar Aminoacylase-1 homodimer enzyme:25 the point mutation of His206 (corresponding to the His235 in hCN1) to Asn significantly reduces the efficiency of the enzyme.25 25 ACS Paragon Plus Environment

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

In our structural prediciction such mutation might reduce polarizability of the C=O group, possibly affecting the enzymatic reactivity. Finally, we noticed that the number of intramolecular hydrogen bonds in the monomer is larger than that of same subunit in the dimer. Comparison between Lcar and Hcar binding to hCN1. Hcar differs from Lcar only by the length of the alkyl chain of the non-histidine residue: it features three CH2 groups instead of two. In both, monomer and dimer, Hcar forms a larger number of hydrogen bonds than Lcar. This might play a role for the larger affinity in vitro of Hcar relative to Lcar. The proposed structures of our Michaelis complexes are radically different from the one previously suggested for hCN184, prior the X-ray structure was deposited in the PDB. The latter model did not indeed include the bridging water molecule (either in its ionic or neutral state) and considered different residues to be involved in the coordination of the two Zn2+ ions and in the interaction with Lcar (Fig. S13). Implication for drug design. Understanding the molecular basis of the increasing affinity on passing from Lcar to Hcar may help design hCN1’s inhibitors retaining Lcar’s beneficial effects but with far higher affinity for the target. These could turn out to be efficient therapeutic agents. Hence, based on the predicted structural determinants, it is tempting to propose some general considerations and principles for the rational design of selective monomeric and dimeric hCN1 inhibitors. The addition of a methylene group in Hcar can generate additional hydrophobic 26 ACS Paragon Plus Environment

Page 26 of 59

Page 27 of 59

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

interactions in the Hcarmonomeric and dimeric hCN1 complex relative to those of Lcar and might therefore be an important structural reason that converts the preferred substrate Lcar into the worse substrate Hcar. This may provide a molecular basis for the smaller Km for Hcar compared to that of Lcar. These interactions are formed mostly between ligand’s alkyl chain and side chain of Ile425 (Fig. 4). Thus, the retention of a propyl (or longer) amine chain attached at the ligand carbonyl might help generating analogs of Lcar that fit well into hCN1, although without undergoing hydrolysis. These analogs might therefore bind and block the enzyme, inhibiting its hydrolytic function.85, 86 Replacing the scissile amide bond with an electrophilic carbonyl element has been used to generate a stable but unproductive tetrahedral intermediate during amide hydrolysis in other enzymes that cleave amide bonds.87 The geometry of the ligand’s pose is not expected to change drastically. We propose therefore that the substitution of the hydrolyzed amide function with bioisosteres of this group (including for example 5-membered heteroaromatic groups such as triazole or tretrazole rings or trifluoroethylamines87) or with a carbamate or urea group might be an alternative strategy to generate inhibitors. In conclusion, we have proposed here structures of Michaelis complexes formed between Lcar and Hcar substrates with hCN1. Our observances from QM/MM simulations are consistent with the experimental evidence that Hcar has a strong competitive inhibitory effect on Lcar turnover 27 ACS Paragon Plus Environment

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

mediated by hCN1.2 Namely, despite Hcar’s Michaelis complex is better stabilized with respect to Lcar by stronger hydrogen bond network, Lcar’s position inside active site is better orientated for hydrolysis reaction. These findings may provide the means for accurate modeling of nonhydrolysable Lcar derivatives with high affinity for their target.

Acknowledgment

The authors would like to thank Laboratory for Biocomputing and Bioinformatics from National Institute of Chemistry Slovenia for many useful discussions and help with using GOLD docking software. The authors gratefully acknowledge the computing time granted by the John von Neumann Institute for Computing (NIC) and provided in the supercomputer JUROPA at Jülich Supercomputer Center (JSC) (NIC project 8307).

Appendix

Supporting Information

Km and kcat values for L-carnosine and its derivatives hydrolysis by hCN1 (Tab. S1); Comparison of hCN1 active site across the M20 family of dizinc-containing metallopeptidases (Fig. S1 and S2);

28 ACS Paragon Plus Environment

Page 28 of 59

Page 29 of 59

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

partial atomic charges, bond length and angle parameters of Lcar used in classical MD (Tab. S2-S4); description and validation of MD simulations with Hess analysis (Section 1 and Tab. S5); description of procedure for chemical shifts calculations (Section 2); protonation states of histidine residues and other non-standard protonation states (Tab. S6); structural characterization of Lcar and Hcar in water from classical MD simulations (Section 3, Tab. S8 and Fig. S4-S7) and comparison of calculated chemical shifts with experimental NMR data (Tab. S9-S15); RMSD plots of classical MD simulations of ligandhCN1 Michaelis complexes (Fig. S3); relative free energies of model systems of apo hCN1 from ab initio calculations (Tab. S16 and S17); characterization of ligandhCN1 Michaelis complexes obtained with docking calculations (Section 4, Tab. S7, S18, Fig. S9 and S12); comparison of distances and angles between zinc ions and their coordinating residues from QM/MM simulation and crystallographic data (Tab. S18-S22); RMSF plots of ligands inside hCN1 active site during QM/MM simulations (Fig. S10); characterization of ligandhCN1β Michaelis complexes from QM/MM simulations (Section 5, Tab. S23 and Fig. S11); comparison of our model of LcarhCN1 with previously proposed one (Fig. S13); relative free energies of model systems of holo monomeric hCN1 from ab initio calculations (Tab. S24); solvent accessible surface of monomeric and dimeric hCN1 (Tab. S25); Docking of Lcar and Hcar to the hCN1 dimer (Section 6)

29 ACS Paragon Plus Environment

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

Author Information

Corresponding Author

*Tel. +49 2461 61 8933 E-mail. [email protected]

Notes

The authors declare no competing financial interest.

30 ACS Paragon Plus Environment

Page 30 of 59

Page 31 of 59

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

References

1.

Bellia, F., Vecchio, G., Cuzzocrea, S., Calabrese, V., and Rizzarelli, E. (2011) Neuroprotective

features of carnosine in oxidative driven diseases, Molecular Aspects of Medicine 32, 258-266.

2.

Peters, V., Kebbewar, M., Jansen, E., Jakobs, C., Riedl, E., Koeppel, H., Frey, D., Adelmann, K.,

Klingbeil, K., Mack, M., Hoffmann, G., Janssen, B., Zschocke, J., and Yard, B. (2010) Relevance of allosteric conformations and homocarnosine concentration on carnosinase activity, Amino Acids 38, 1607-1615.

3.

Quinn, P. J., Boldyrev, A. A., and Formazuyk, V. E. (1992) Carnosine: Its properties, functions

and potential therapeutic applications, Molecular Aspects of Medicine 13, 379-444.

4.

Decker, E. A., Livisay, S. A., and Zhou, S. (2000) A re-evaluation of the antioxidant activity of

purified carnosine, Biochemistry (Moscow) 65, 766-770.

5.

Kang, J. H., Kim, K. S., Choi, S. Y., Kwon, H. Y., Won, M. H., and Kang, T. C. (2002) Carnosine

and related dipeptides protect human ceruloplasmin against peroxyl radical-mediated modification, Molecules and Cells 13, 498-502.

6.

Hipkiss, A. R. (2009) Chapter 3 carnosine and its possible roles in nutrition and health, In

Advances in food and nutrition research (Steve, L. T., Ed.), pp 87-154, Academic Press.

31 ACS Paragon Plus Environment

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

7.

Lelie, H. L., Liba, A., Bourassa, M. W., Chattopadhyay, M., Chan, P. K., Gralla, E. B., Miller, L. M.,

Borchelt, D. R., Valentine, J. S., and Whitelegge, J. P. (2011) Copper and zinc metallation status of copper-zinc superoxide dismutase from amyotrophic lateral sclerosis transgenic mice, Journal of Biological Chemistry 286, 2795-2806.

8.

Basun, H., Forsell, L. G., Wetterberg, L., and Winblad, B. (1991) Metals and trace elements in

plasma and cerebrospinal fluid in normal aging and alzheimer's disease, Journal of Neural Transmission / Parkinson's Disease and Dementia Section 3, 231-258.

9.

Davies, K. M., Bohic, S., Carmona, A., Ortega, R., Cottam, V., Hare, D. J., Finberg, J. P. M., Reyes,

S., Halliday, G. M., Mercer, J. F. B., and Double, K. L. (2014) Copper pathology in vulnerable brain regions in parkinson's disease, Neurobiology of Aging 35, 858-866.

10.

Jellinger, K. A. (2013) Chapter one - the relevance of metals in the pathophysiology of

neurodegeneration, pathological considerations, In International review of neurobiology (Kailash, P. B., and Susanne, A. S., Eds.), pp 1-47, Academic Press.

11.

Sakae, K., Agata, T., Kamide, R., and Yanagisawa, H. (2013) Effects of l-carnosine and its zinc

complex (polaprezinc) on pressure ulcer healing, Nutrition in Clinical Practice 28, 609-616.

32 ACS Paragon Plus Environment

Page 32 of 59

Page 33 of 59

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

12.

Teufel, M., Saudek, V., Ledig, J.-P., Bernhardt, A., Boularand, S., Carreau, A., Cairns, N. J.,

Carter, C., Cowley, D. J., Duverger, D., Ganzhorn, A. J., Guenet, C., Heintzelmann, B., Laucher, V., Sauvage, C., and Smirnova, T. (2003) Sequence identification and characterization of human carnosinase and a closely related non-specific dipeptidase, Journal of Biological Chemistry 278, 6521-6531.

13.

Lenney, J. F., George, R. P., Weiss, A. M., Kucera, C. M., Chan, P. W. H., and Rinzler, G. S. (1982)

Human serum carnosinase: Characterization, distinction from cellular carnosinase, and activation by cadmium, Clinica Chimica Acta 123, 221-231.

14.

Bellia, F., Vecchio, G., and Rizzarelli, E. (2014) Carnosinases, their substrates and diseases,

Molecules 19, 2299-2329.

15.

Lenney, J. F., Peppers, S. C., Kucera-Orallo, C. M., and George, R. P. (1985) Characterization of

human tissue carnosinase, Biochemical Journal 228, 653-660.

16.

Unno, H., Yamashita, T., Ujita, S., Okumura, N., Otani, H., Okumura, A., Nagai, K., and

Kusunoki, M. (2008) Structural basis for substrate recognition and hydrolysis by mouse carnosinase cn2, Journal of Biological Chemistry 283, 27289-27299.

33 ACS Paragon Plus Environment

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

17.

Pegova, A., Abe, H., and Boldyrev, A. (2000) Hydrolysis of carnosine and related compounds

by mammalian carnosinases, Comparative Biochemistry and Physiology B: Biochemistry and Molecular Biology 127, 443-446.

18.

Grasso, G. I., Arena, G., Bellia, F., Rizzarelli, E., and Vecchio, G. (2014) Copper(ii)-chelating

homocarnosine glycoconjugate as a new multifunctional compound, Journal of Inorganic Biochemistry 131, 56-63.

19.

Chevrier, B., Schalk, C., D'Orchymont, H., Rondeau, J.-M., Moras, D., and Tarnus, C. (1994)

Crystal structure of aeromonas proteolytica aminopeptidase: A prototypical member of the cocatalytic zinc enzyme family, Structure 2, 283-291.

20.

Agarwal, R., Burley, S. K., and Swaminathan, S. (2007) Structural analysis of a ternary

complex of allantoate amidohydrolase from escherichia coli reveals its mechanics, Journal of Molecular Biology 368, 450-463.

21.

Lundgren, S., Andersen, B., Piškur, J., and Dobritzsch, D. (2007) Crystal structures of yeast β-

alanine synthase complexes reveal the mode of substrate binding and large scale domain closure movements, Journal of Biological Chemistry 282, 36037-36047.

34 ACS Paragon Plus Environment

Page 34 of 59

Page 35 of 59

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

22.

Holz, R. C., Bzymek, K. P., and Swierczek, S. I. (2003) Co-catalytic metallopeptidases as

pharmaceutical targets, Current Opinion in Chemical Biology 7, 197-206.

23.

Navrátil, V., Klusák, V., and Rulíšek, L. (2013) Theoretical aspects of hydrolysis of peptide

bonds by zinc metalloenzymes, Chemistry- A European Journal 19, 16634-16645.

24.

Jackson, M. C., Kucera, C. M., and Lenney, J. F. (1991) Purification and properties of human

serum carnosinase, Clinica Chimica Acta 196, 193-205.

25.

Lindner, H. A., Lunin, V. V., Alary, A., Hecker, R., Cygler, M., and Ménard, R. (2003) Essential

roles of zinc ligation and enzyme dimerization for catalysis in the aminoacylase-1/m20 family, Journal of Biological Chemistry 278, 44496-44504.

26.

Schaftenaar, G., and Noordik, J. H. (2000) Molden: A pre- and post-processing program for

molecular and electronic structures, Journal of Computer-Aided Molecular Design 14, 123-134.

27.

Becke, A. D. (1993) A new mixing of hartree–fock and local density functional theories,

Journal of Chemical Physics 98, 1372-1377.

28.

Miertuš, S., Scrocco, E., and Tomasi, J. (1981) Electrostatic interaction of a solute with a

continuum. A direct utilizaion of ab initio molecular potentials for the prevision of solvent effects, Chemical Physics 55, 117-129.

35 ACS Paragon Plus Environment

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

29.

Thomas, S., Biswas, N., Malkar, V. V., Mukherjee, T., and Kapoor, S. (2010) Studies on

adsorption of carnosine on silver nanoparticles by sers, Chemical Physics Letters 491, 59-64.

30.

Frisch, M. J., Trucks, G. W., Schlegel, H. B., Scuseria, G. E., Robb, M. A., Cheeseman, J. R.,

Scalmani, G., Barone, V., Mennucci, B., Petersson, G. A., Nakatsuji, H., Caricato, M., Li, X., Hratchian, H. P., Izmaylov, A. F., Bloino, J., Zheng, G., Sonnenberg, J. L., Hada, M., Ehara, M., Toyota, K., Fukuda, R., Hasegawa, J., Ishida, M., Nakajima, T., Honda, Y., Kitao, O., Nakai, H., Vreven, T., Montgomery, J. A., Peralta, J. E., Ogliaro, F., Bearpark, M., Heyd, J. J., Brothers, E., Kudin, K. N., Staroverov, V. N., Kobayashi, R., Normand, J., Raghavachari, K., Rendell, A., Burant, J. C., Iyengar, S. S., Tomasi, J., Cossi, M., Rega, N., Millam, J. M., Klene, M., Knox, J. E., Cross, J. B., Bakken, V., Adamo, C., Jaramillo, J., Gomperts, R., Stratmann, R. E., Yazyev, O., Austin, A. J., Cammi, R., Pomelli, C., Ochterski, J. W., Martin, R. L., Morokuma, K., Zakrzewski, V. G., Voth, G. A., Salvador, P., Dannenberg, J. J., Dapprich, S., Daniels, A. D., Farkas, Foresman, J. B., Ortiz, J. V., Cioslowski, J., and Fox, D. J. (2009) Gaussian 09, revision a.02, Wallingford CT.

31.

Hornak, V., Abel, R., Okur, A., Strockbine, B., Roitberg, A., and Simmerling, C. (2006)

Comparison of multiple amber force fields and development of improved protein backbone parameters, Proteins: Structure, Function, and Bioinformatics 65, 712-725.

36 ACS Paragon Plus Environment

Page 36 of 59

Page 37 of 59

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

32.

Lindorff-Larsen, K., Piana, S., Palmo, K., Maragakis, P., Klepeis, J. L., Dror, R. O., and Shaw, D.

E. (2010) Improved side-chain torsion potentials for the amber ff99sb protein force field, Proteins: Structure, Function, and Bioinformatics 78, 1950-1958.

33.

Bayly, C. I., Cieplak, P., Cornell, W., and Kollman, P. A. (1993) A well-behaved electrostatic

potential based method using charge restraints for deriving atomic charges: The resp model, The Journal of Physical Chemistry 97, 10269-10280.

34.

Breneman, C. M., and Wiberg, K. B. (1990) Determining atom-centered monopoles from

molecular electrostatic potentials. The need for high sampling density in formamide conformational analysis, Journal of Computational Chemistry 11, 361-373.

35.

Wang, J., Wang, W., Kollman, P. A., and Case, D. A. (2006) Automatic atom type and bond

type perception in molecular mechanical calculations, Journal of Molecular Graphics and Modelling 25, 247-260.

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, Journal of Chemical Physics 79, 926-935.

37 ACS Paragon Plus Environment

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

37.

Nosé, S. (1984) A molecular dynamics method for simulations in the canonical ensemble,

Molecular Physics 52, 255-268.

38.

Hoover, W. G. (1985) Canonical dynamics: Equilibrium phase-space distributions, Physical

Review A 31, 1695-1697.

39.

Parrinello, M., and Rahman, A. (1981) Polymorphic transitions in single crystals: A new

molecular dynamics method, Journal of Applied Physics 52, 7182-7190.

40.

Darden, T., York, D., and Pedersen, L. (1993) Particle mesh ewald: An n⊕log(n) method for

ewald sums in large systems, Journal of Chemical Physics 98, 10089-10092.

41.

Hess, B. (2002) Convergence of sampling in protein simulations, Physical Review E 65,

031910.

42.

Daura, X., Gademann, K., Jaun, B., Seebach, D., van Gunsteren, W. F., and Mark, A. E. (1999)

Peptide folding: When simulation meets experiment, Angewandte Chemie International Edition 38, 236-240.

43.

Van Der Spoel, D., Lindahl, E., Hess, B., Groenhof, G., Mark, A. E., and Berendsen, H. J. C.

(2005) Gromacs: Fast, flexible, and free, Journal of Computational Chemistry 26, 1701-1718.

38 ACS Paragon Plus Environment

Page 38 of 59

Page 39 of 59

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

44.

Hess, B., Kutzner, C., van der Spoel, D., and Lindahl, E. (2008) Gromacs 4: Algorithms for

highly efficient, load-balanced, and scalable molecular simulation, Journal of Chemical Theory and Computation 4, 435-447.

45.

Zhang, R., Wang, D., and Wu, W. (2013) Conformations of carnosine in aqueous solutions by

all-atom molecular dynamics simulations and 2d-nosey spectrum, Chinese Journal of Chemical Physics 26, 67-72.

46.

Branham, M. L., Singh, P., Bisetty, K., Sabela, M., and Govender, T. (2011) Preparation,

spectrochemical, and computational analysis of l-carnosine (2-[(3-aminopropanoyl)amino]-3-(1himidazol-5-yl)propanoic acid) and its ruthenium (ii) coordination complexes in aqueous solution, Molecules 16, 10269-10291.

47.

Friedrich, J. O., and Wasylishen, R. E. (1986) A 1h and 13c nuclear magnetic resonance study

of carnosine, Canadian Journal of Chemistry 64, 2132-2138.

48.

Gaggelli, E., and Valensin, G. (1990) 1h and 13c nmr relaxation investigation of the calcium

complex of β-alanyl-l-histidine (carnosine) in aqueous solution, Journal of the Chemical Society, Perkin Transaction 2, 401-406.

39 ACS Paragon Plus Environment

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

49.

Govindaraju, V., Young, K., and Maudsley, A. A. (2000) Proton nmr chemical shifts and

coupling constants for brain metabolites, NMR in Biomedicine 13, 129-153.

50.

Šali, A., and Blundell, T. L. (1993) Comparative protein modelling by satisfaction of spatial

restraints, Journal of Molecular Biology 234, 779-815.

51.

Anandakrishnan, R., Aguilar, B., and Onufriev, A. V. (2012) H++ 3.0: Automating pk

prediction and the preparation of biomolecular structures for atomistic molecular modeling and simulations, Nucleic Acids Research 40, W537-W541.

52.

Gordon, J. C., Myers, J. B., Folta, T., Shoja, V., Heath, L. S., and Onufriev, A. (2005) H++: A

server for estimating pkas and adding missing hydrogens to macromolecules, Nucleic Acids Research 33, W368-W371.

53.

Myers, J., Grothaus, G., Narayanan, S., and Onufriev, A. (2006) A simple clustering algorithm

can be accurate enough for use in calculations of pks in macromolecules, Proteins: Structure, Function, and Bioinformatics 63, 928-938.

54.

Shen, M.-y., and Sali, A. (2006) Statistical potential for assessment and prediction of protein

structures, Protein Science 15, 2507-2524.

40 ACS Paragon Plus Environment

Page 40 of 59

Page 41 of 59

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

55.

Fiser, A., and Šali, A. (2003) Modeller: Generation and refinement of homology-based

protein structure models, In Methods in enzymology (Charles W. Carter, Jr., and Robert, M. S., Eds.), pp 461-491, Academic Press.

56.

Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) Procheck: A

program to check the stereochemical quality of protein structures, Journal of Applied Crystallography 26, 283-291.

57.

Verdonk, M. L., Cole, J. C., Hartshorn, M. J., Murray, C. W., and Taylor, R. D. (2003) Improved

protein–ligand docking using gold, Proteins: Structure, Function, and Bioinformatics 52, 609-623.

58.

Verdonk, M. L., Chessari, G., Cole, J. C., Hartshorn, M. J., Murray, C. W., Nissink, J. W. M.,

Taylor, R. D., and Taylor, R. (2005) Modeling water molecules in protein−ligand docking using gold, Journal of Medicinal Chemistry 48, 6504-6515.

59.

Jones, G., Willett, P., and Glen, R. C. (1995) Molecular recognition of receptor sites using a

genetic algorithm with a description of desolvation, Journal of Molecular Biology 245, 43-53.

60.

Jones, G., Willett, P., Glen, R. C., Leach, A. R., and Taylor, R. (1997) Development and

validation of a genetic algorithm for flexible docking, Journal of Molecular Biology 267, 727-748.

41 ACS Paragon Plus Environment

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

61.

Page 42 of 59

Korb, O., Stützle, T., and Exner, T. E. (2009) Empirical scoring functions for advanced

protein−ligand docking with plants, Journal of Chemical Information and Modeling 49, 84-96.

62.

Gilboa, R., Greenblatt, H. M., Perach, M., Spungin-Bialik, A., Lessel, U., Wohlfahrt, G.,

Schomburg, D., Blumberg, S., and Shoham, G. (2000) Interactions of streptomyces griseus aminopeptidase with a methionine product analogue: A structural study at 1.53 å resolution, Acta Crystallographica Section D: Biological Crystallography 56, 551-558.

63.

Liao, R.-Z., Yu, J.-G., and Himo, F. (2011) Quantum chemical modeling of enzymatic reactions:

The case of decarboxylation, Journal of Chemical Theory and Computation 7, 1494-1501.

64.

Morris, G. M., Huey, R., Lindstrom, W., Sanner, M. F., Belew, R. K., Goodsell, D. S., and Olson, A.

J. (2009) Autodock4 and autodocktools4: Automated docking with selective receptor flexibility, Journal of Computational Chemistry 30, 2785-2791.

65.

Santos-Martins, D., Forli, S., Ramos, M. J., and Olson, A. J. (2014) Autodock4zn: An improved

autodock force field for small-molecule docking to zinc metalloproteins, Journal of Chemical Information and Modeling 54, 2371-2379.

42 ACS Paragon Plus Environment

Page 43 of 59

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

66.

Park, H., Kim, S., Kim, Y. E., and Lim, S.-J. (2010) A structure-based virtual screening

approach toward the discovery of histone deacetylase inhibitors: Identification of promising zincchelating groups, ChemMedChem 5, 591-597.

67.

Peters, M. B., Yang, Y., Wang, B., Füsti-Molnár, L., Weaver, M. N., and Merz, K. M. (2010)

Structural survey of zinc-containing proteins and development of the zinc amber force field (zaff), Journal of Chemical Theory and Computation 6, 2935-2947.

68.

Cornell, W. D., Cieplak, P., Bayly, C. I., Gould, I. R., Merz, K. M., Ferguson, D. M., Spellmeyer, D.

C., Fox, T., Caldwell, J. W., and Kollman, P. A. (1995) A second generation force field for the simulation of proteins, nucleic acids, and organic molecules, Journal of the American Chemical Society 117, 5179-5197.

69.

Morris, G. M., Goodsell, D. S., Halliday, R. S., Huey, R., Hart, W. E., Belew, R. K., and Olson, A. J.

(1998) Automated docking using a lamarckian genetic algorithm and an empirical binding free energy function, Journal of Computational Chemistry 19, 1639-1662.

70.

Hess, B., Bekker, H., Berendsen, H. J. C., and Fraaije, J. G. E. M. (1997) Lincs: A linear

constraint solver for molecular simulations, Journal of Computational Chemistry 18, 1463-1472.

43 ACS Paragon Plus Environment

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

71.

von Lilienfeld, O. A., Tavernelli, I., Rothlisberger, U., and Sebastiani, D. (2005) Variational

optimization of effective atom centered potentials for molecular properties, Journal of Chemical Physics 122, 14113.

72.

Perdew, J., Burke, K., and Ernzerhof, M. (1996) Generalized gradient approximation made

simple, Physical Review Letters 77, 3865-3868.

73.

Troullier, N., and Martins, J. (1991) Efficient pseudopotentials for plane-wave calculations,

Physical Review B 43, 1993-2006.

74.

von Lilienfeld, O., Tavernelli, I., Rothlisberger, U., and Sebastiani, D. (2004) Optimization of

effective atom centered potentials for london dispersion forces in density functional theory, Physical Review Letters 93, 153004.

75.

Kleinman, L., and Bylander, D. (1982) Efficacious form for model pseudopotentials, Physical

Review Letters 48, 1425-1428.

76.

Davis, P. J., and Rabinowitz, P. (1983) Methods of numerical integration, 2nd ed., Academic

Press Inc., London.

44 ACS Paragon Plus Environment

Page 44 of 59

Page 45 of 59

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

77.

Martyna, G., and Tuckerman, M. (1999) A reciprocal space based method for treating long

range interactions in ab initio and force-field-based calculations in clusters, Journal of Chemical Physics 110, 2810-2821.

78.

Shytle, R. D., Ehrhart, J., Tan, J., Vila, J., Cole, M., Sanberg, C. D., Sanberg, P. R., and Bickford, P.

C. (2007) Oxidative stress of neural, hematopoietic, and stem cells: Protection by natural compounds, Rejuvenation Research 10, 173-178.

79.

Tanokura, M., Tasumi, M., and Miyazawa, T. (1976) 1h nuclear magnetic resonance studies

of histidine-containing di- and tripeptides. Estimation of the effects of charged groups on the pka value of the imidazole ring, Biopolymers 15, 393-401.

80.

Matsukura, T., Takahashi, T., Nishimura, Y., Ohtani, T., Sawada, M., and Shibata, K. (1990)

Characterization of crystalline l-carnosine zn(ii) complex (z-103), a novel anti-gastric ulcer agent: Tautomeric change of imidazole moiety upon complexation, Chemical and Pharmaceutical Bulletin 38, 3140-3146.

81.

Cerdà-Costa, N., and Xavier Gomis-Rüth, F. (2014) Architecture and function of

metallopeptidase catalytic domains, Protein Science 23, 123-144.

45 ACS Paragon Plus Environment

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

82.

Dal Peraro, M., Ruggerone, P., Raugei, S., Gervasio, F. L., and Carloni, P. (2007) Investigating

biological systems using first principles car–parrinello molecular dynamics simulations, Current Opinion in Structural Biology 17, 149-156.

83.

Laio, A., VandeVondele, J., and Rothlisberger, U. (2002) A hamiltonian electrostatic coupling

scheme for hybrid car–parrinello molecular dynamics simulations, Journal of Chemical Physics 116, 6941-6947.

84.

Vistoli, G., Pedretti, A., Cattaneo, M., Aldini, G., and Testa, B. (2006) Homology modeling of

human serum carnosinase, a potential dedicinal target, and md simulations of its allosteric activation by citrate, Journal of Medicinal Chemistry 49, 3269-3277.

85.

Lodola, A., and De Vivo, M. (2012) Chapter 11 - the increasing role of qm/mm in drug

discovery, In Advances in protein chemistry and structural biology (Christo, C., and Tatyana, K.-C., Eds.), pp 337-362, Academic Press.

86.

De Vivo, M. (2011) Bridging quantum mechanics and structure-based design, Frontiers in

Bioscience (Landmark Edition) 16, 1619-1633.

87.

Meanwell, N. A. (2011) Synopsis of some recent tactical application of bioisosteres in drug

design, Journal of Medicinal Chemistry 54, 2529-2591.

46 ACS Paragon Plus Environment

Page 46 of 59

Page 47 of 59

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

Table 1. Maxwell-Boltzmann statistics and relative free energies (in kcal/mol) for representative structures of highly populated clusters from MD simulations of Lcar and Hcar tautomers.a

/b

Relative free energy with respect to 1B in kcal/mol

Hcar representative structure

/c

Relative free energy with respect to 2D in kcal/mol

1B

1

0

2D

1

0

3B

0.045

1.8

3C

0.001

4.3

3A

0.014

2.5

3D

1.393*10-6

8.0

1A

0.012

2.6

1D

3.895*10-8

10.1

4B

6.007*10-6

7.1

1C

8.080*10-11

13.8

2A

3.716*10-7

8.8

2C

3.499*10-11

14.3

2B

3.203*10-8

10.2

4D

1.417*10-13

17.5

4A

1.060*10-8

10.9

4C

9.503*10-16

20.5

Lcar representative structure

a Please

bAll

note that the clusters 1-4 are not the same for two ligands and their protonation states.

the structures are weighted against the structure with the lowest free energy – 1B. Index X

stands for the name of the model (1-4A and 1-4B). cAll

the structures are weighted against the structure with the lowest free energy – 2D. Index X

stands for the name of the model (1-4C and 1-4D).

47 ACS Paragon Plus Environment

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 48 of 59

Table 2. Probability of hydrogen bonds between ligands and the hCN1α α active site from our 2 ps long QM/MM simulations. WM indicates water mediated hydrogen bond. Hydrogen bond donor

Hydrogen bond acceptor

Ligand imidazole ring

Glh173 backbone oxygen

Ligand imidazole ring

Glh173 carboxyl group

Ligand imidazole ring

Glu174 backbone oxygen

Ligand imidazole ring

Ser177 hydroxyl group

Ligand amino group

Asp202 backbone oxygen

Ligand amino group

Asp202 carboxyl group

100%

/

/

/

Ligand carboxyl group

Asn220 amino group

/

/

79%WM

/

a) /

a) /

a) 85%

a) 100%

b) 49%

b) 88%

b) /

b) /

c) 4%

c) 97%

c) 86%

c) /

d) 40%

d) /

d) /

d) 87%

Arg350 side chaina

Ligands carboxyl group

LcarhCN1α HcarhCN1α LcarhCN1α HcarhCN1α

69% /

/

71%WM 4%WM 1%

/

/

12% 4%WM

56%

63%

18%WM

31%WM

/

/

4%

/

/

/

30%WM

26% 100%

96%

89%WM

51%WM

48 ACS Paragon Plus Environment

Page 49 of 59

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

Ligand amino group

Asp421 carboxyl group

/

83%WM

63%WM

76%WM

Ligand amino group

Ser423 hydroxyl group

/

/

/

5%

Ligand peptide NH group

Thr424 backbone oxygen

70%

100%

95%

Thr424 peptide NH group

Ligand carboxyl group

86%WM

16%

7%

58%

Ligand amino group

Glu451 carboxyl group

/

69%WM

87%WM

8%WM

Ligand backbone oxygen

His235 imidazole ringb

/

/

96%

75%

Ligand carboxylic group

Thr337 hydroxyl groupb

/

/

96%

90%

aBetween

3% 16%WM

Arg350 and substrate’s carboxylic group are 4 possible hydrogen bonds (marked with,

a), b,) c) and d), see also Fig. 4); also salt bridge is formed here. bResidues

His235 and Thr337 are from monomer II.

49 ACS Paragon Plus Environment

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

Chart 1. Lcar tautomers A and B and Hcar tautomers C and D.

50 ACS Paragon Plus Environment

Page 50 of 59

Page 51 of 59

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

Scheme 1. Proposed general mechanism for the hydrolysis of a peptide catalyzed by a metallopeptidase with di-zinc co-catalytic active site as suggested by Holz et al.22

51 ACS Paragon Plus Environment

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

Figure 1. (a) homodimer of hCN1 shown in ribbon diagram, with monomers colored in pink (monmomer I) and turquoise (monomer II). Zn2+ ions in both active sites are shown as dark purple spheres and residues coordinating Zn2+ ions are colored in yellow. (b) Superimposition of the apo hCN1 active site residues (residues colored in yellow, Zn2+ ions in gold and bridging moiety (OW) in orange with the ones of aminopeptidase from Streptomyces griseus62 (residues colored in light blue, Zn2+ ions in dark blue and bridging moiety in turquoise) and aminopeptidase from Aeromonas proteolytica19 (residues colored in purple, Zn2+ ions in dark purple and bridging moiety in magenta). Oxygen and nitrogen atoms are colored in red and blue, respectively. Residues are numbered according to the hCN1. Zn2+ ion coordinated by Asp139, Glu174, His452 and water molecule is named Zn1, while the one coordinated by His106, Asp139, Asp202 and water molecule is named Zn2.

Figure 2. Tautomers α and β of hCN1 active site. Proton that changes position is encircled in orange for tautomer α (protonated on Oε2) and in pink for β (protonated on Oε1). Coordination bonds between Zn2+ ions and hCN1 active site residues are shown as dashed green lines.

Figure 3. Representative structures of the largest populated clusters of Lcar tautomers A and B (see Chart 1). The ‘extended’ and ‘semi-folded’ representative conformations differ in the value of χ1 52 ACS Paragon Plus Environment

Page 52 of 59

Page 53 of 59

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

dihedral angle of histidine side chain (defined as dihedral angle between N-Cα-Cβ-Cγ). Namely, in all ‘extended’ conformation, the histidine side chain is in trans conformation, while in ‘semi-folded’ representatives it can be either in gauche+ (3B) or in gauche- (2A, 4A and 4B). Each ‘extended’ and ‘semi-folded’ conformations can be further distinguished by χ2 dihedral angle (dihedral angle between Cα-Cβ-Cγ-Nδ), which defines the orientation of the imidazole ring of His in sense of its rotation in the axis of Cβ-Cγ bond. Each conformation is colored in distinct color and it is shown from to different angles, as side view and as front view, with marked χ1 and χ2 dihedral angles.

Figure 4. Structures of Lcarmonomeric hCN1α (a), Hcarmonomeric hCN1α (b), Lcardimeric hCN1α (c) and Hcardimeric hCN1α (d) as obtained by our QM/MM simulations. On the left side of each panel is schematic 2D representation of the Michaelis complex (bulk water molecules are not shown due to clarity reasons), while on the right side there is matching 3D figure (non-polar hydrogen atoms and bulk water molecules are not shown due to clarity). Lcar is shown in green and Hcar in orange. Hydrogen bonds between ligand and environment are shown as dashed turquoise lines in 2D schemes and as full turquoise lines in the 3D representations (for occupancy of hydrogen bonds between ligands and hCN1 see Tab. 2; hydrogen bond naming (a, b, c and d) between ligand’s carboxylic group and side chain of Arg350 correspond to the one in Tab. 2).,Water-mediated hydrogen bonds are shown as dashed purple lines in 2D schemes (due to the 53 ACS Paragon Plus Environment

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

clarity, water molecules are omitted in these schemes) and as full purple lines in the 3D figures. Coordination bonds between Zn2+ ions and hCN1 active site residues are shown as dashed green lines on all figures. Zn2+ ions are shown as violet spheres in the 3D representation.

Figure 5. (a) Superimposition of monomeric hCN1 (light blue) with monomer I of dimeric hCN1 (light pink) shown in chain trace representation. Monomer II is shown in light grey in ribbons representation and the two Zn(II) ions (Zn1 and Zn2) are shown as purple spheres. Regions RI – RV are shown in blue for monomeric hCN1 and in pink for momoner I. (b) Distribution of intramolecular hydrogen bonds in the monomeric and momoner I of dimeric complexes with ligands during 40 ns classical MD.

54 ACS Paragon Plus Environment

Page 54 of 59

Page 55 of 59

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 1.

Figure 2.

55 ACS Paragon Plus Environment

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

Figure 3.

56 ACS Paragon Plus Environment

Page 56 of 59

Page 57 of 59

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

57 ACS Paragon Plus Environment

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

Figure 4.

Figure 5.

58 ACS Paragon Plus Environment

Page 58 of 59

Page 59 of 59

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

For Table of Contents Use Only

CARNOSINE AND HOMOCARNOSINE DEGRATION MECHANISMS BY THE HUMAN CARNOSINASE ENZYME CN1: INSIGHTS FROM MULTISCALE SIMULATIONS

Matic Pavlin, Giulia Rossetti, Marco De Vivo, Paolo Carloni

59 ACS Paragon Plus Environment