Structure and Dynamics of an Archeal Monoglyceride Lipase from Palaeococcus ferrophilus as Revealed by Crystallography and In Silico Analysis

The crystallographic analysis of a lipase from Palaeococcus ferrophilus (PFL) previously annotated as a lysophospholipase revealed high structural conservation with other monoglyceride lipases, in particular in the lid domain and substrate binding pockets. In agreement with this observation, PFL was shown to be active on various monoacylglycerols. Molecular Dynamics (MD) studies performed in the absence and in the presence of ligands further allowed characterization of the dynamics of this system and led to a systematic closure of the lid compared to the crystal structure. However, the presence of ligands in the acyl-binding pocket stabilizes intermediate conformations compared to the crystal and totally closed structures. Several lid-stabilizing or closure elements were highlighted, i.e., hydrogen bonds between Ser117 and Ile204 or Asn142 and its facing amino acid lid residues, as well as Phe123. Thus, based on this complementary crystallographic and MD approach, we suggest that the crystal structure reported herein represents an open conformation, at least partially, of the PFL, which is likely stabilized by the ligand, and it brings to light several key structural features prone to participate in the closure of the lid.


SM1. Technical details regarding the ligand building
To generate the Gromos54a7 FF parameters of the LDAO ligand, the ligand structure was submitted to the ATB server [6]. Topology files describing the ligand and the glycerol molecules at the united-atom level were directly retrieved from the ATB server, with the MolID 13206 and 30887, respectively.
To generate the enzyme-MAFP covalently bound complex, the PDB structure 1mt5 was aligned with chain A of PFL using Pymol (Pymol, version 1.8) [7]. It allowed to retrieve an orientation of the methyl arachidonyl fluorophosphonate ligand in adequation with the pocket of the chain A crystal structure. From that configuration, the residue Ser87 was linked to the ligand to generate the moiety reported in Figure SM1_1. The Serine N and C terminal groups were capped with electrically neutral CH3(C=O)-(ACE) and -N(CH3)2 (NME) groups. Figure SM1_1. Serine-MAFP C28H57N2O5P conformation submitted to the ATB server [6]. H atoms are not shown for clarity.
The moiety topology was retrieved at the united-atom level (ATB MolID 368680) and added to (or replacing some of) the Ser87 topology generated by the Gromacs5.1.4 building tools, while preserving an electric charge of zero (Table SM1_1). An in-house program was written for the renumbering of the ligand atoms and the generation of the additional 1-4 interactions.

SM2. Description of the Molecular Dynamics (MD) calculations
MD trajectories of the solvated systems were run using the Gromacs5.1.4 program package [8] with the Gromos54a7 force field [9] under particle mesh Ewald periodic boundary conditions and a Coulomb cut-off distance of 1.2 nm. The Newton equations of motion were numerically integrated using a leap-frog integrator. The van der Waals cut-off distance was set equal to 1.2 nm. Long-range dispersion corrections to energy and pressure were applied. The systems were optimized using a steepest descent algorithm with an initial step size of 0.10 nm.
The whole systems were again optimized, using a steepest descent algorithm with an initial step size of 0.10 nm, to eliminate large forces and then heated to 50 K through a 10 ps canonical (NVT) MD, with a time step of 2 fs and LINCS constraints acting on bonds involving H atoms. The trajectory was followed by two successive 20 ps heating stages, at 150 K and at the final temperature of 300 K (or 343 K), under the same conditions. Next, each system was equilibrated during 50 ps in the NPT ensemble, at P = 1 bar, to relax the solvent molecules, and for a further 60 ns MD equilibration run. The 'V-Rescale' and 'Parrinello-Rahman' algorithms were selected to constrain T and P, respectively. A final production run of 300 ns (150 x 10 6 steps) was performed for the evaluation of the structural, energetics, and dynamical properties of each system. Trajectory data were saved every 20 ps. It was further decided to consider the last 200 ns of the simulations only to avoid the largest fluctuations of the systems. The various stages of the simulations are reported in Table SM2_1.