All these results were obtained thanks to the HPC facilities of and the following founders
I am are interested in the multi-scale modeling and molecular dynamics (MD) simulations of biophysical processes. For my research, I develop and use atomistic and coarse-graining MD for studying protein stability and function in different environments such as in phospholipid membranes or in direct or inverted micelles. I also use quantum mechanical (QM) approaches to develop force field parameters for biomolecules ( e.g. detergents) widely used in experiments in the field of membrane protein (MP). A brief description of my current research interests is given below and the papers published from these research topics are listed in the publication section of this website.
More info

Reverse micelle systems

Reverse micelles (RM) are stable, isotropic, water microemulsions suspended in organic solvents (such as alkanes, scCO2). They were originally employed since decades for instance, with terahertz or NMR, spectroscopy techniques or MD to examine the structure and dynamics of the water and biomolecules in a confined environment mimicking the cellular interior. RMs have a great interest since they are capable to solubilize various biomolecules (such as proteins, poymers, sugars, etc.) inside their water pools. However, the solubilization or not of biomolecules depends significantly on a multitude of parameters such as the chemical nature of the surfactant, the use of cosurfactants or the water content inside the RM water pool and remain poorly understood, so far. To examine this aspect in details, we use atomistic MD and experimental approaches (e.g. time-domain THz spectroscopy, SAXS) to examine different parameters of the RM and confined water oroperties that may influence their solubilization capabilities such as the static and dynamical characteristics of the confined water, their structural changes with confined peptides, globular proteins (cytochrome C) and more recently with membrane protein (gramicidin A channel). These researches were made, in part, in collaboration with Dr. D. Laage at ENS Paris and Dr. T. Elsaesser, at U. Berlin.

Surfactant systems

Detergents (or surfactants) play an essential role in the membrane protein (MP) studies; their amphiphilic nature allows them to interact with the hydrophobic region of MPs to keep them water-soluble outside of their native bilayer environment. Unfortunately, solubility does not always “translate” to native structure and stability; a detergent that is useful for extraction may not be compatible with purification and/or biochemical studies. Furthermore, a detergent working on one membrane protein may not be suitable for a different one. Indeed, detergents are often qualitatively described as ‘harsh“ (e.g., SDS, DPC) and “mild“ (e.g., DDM and OG) based upon their relative propensity to denature proteins by interacting with non-membranous regions of proteins. Properties of detergents that render them denaturing include headgroup charge (a charged head group is more denaturing than a zwitterionic one which, in turn, is more denaturing than a neutral one), as well as acyl tail length (a shorter chain is more denaturing than a longer one). Therefore, there is not a set of “golden rules” for the use of detergents for MP applications. Understanding the physical-chemical properties associated with different classes of detergents is thus crucial for choosing which detergent may work best for a particular application. In this research, we use large scale atomistic and coarse-grained MD simulations combined with biophysical experiments such as scattering experiments (e.g. SAXS, SANS) and fluorescence spectroscopy to study structural and dynamical properties of various types of micelles with e.g. bDDM, LDAO, DPC, SDS or TX100 with often newly developped parameters and peptides, membrane proteins as well as with organic pollutants such as PAHs in the context of surfactant ehanced remediation. These researches were made, in part, in collaboration with Dr. François Yves Dupradeau U. of Amiens, France, Dr. Françoise Bonneté, IBPC, Paris or the group of Pr. Zhi Dang at the South China University of Technology, Guanzhou, China

Membrane-protein systems

Membrane proteins (MPs) play major roles in living organisms and participate to exchanges and communications between cells and their immediate environment as receptors, transporters, ion channels, etc. MPs are also involved in a large number of pathologies ( e.g. influenza, HIV …) and genetic diseases ( e.g. cystic fibrosis) and theirefore their physiological and biomedical relevance makes them major targets for a large part of the pharmaceutical molecules in development. Despite their importance, the 3D structures at atomic level of MPs represent barely 2% of biological macromolecular structures in the Protein Data Bank. This deficit is due to difficulties in the preparation of suitable samples for structural studies by NMR, cryo-EM or crystallography. Indeed, these structural techniques require the use of membrane-mimetic environments, such as detergents, amphipathic polymers or nanodiscs, to extract proteins from their native lipidic environments and to keep them stable, functional and monodisperse. For instance, in the case of crystallography, growing well-ordered diffracting crystals directly from detergent solutions remains the most largely used method, because of its ease of implementation similar to soluble proteins (vapor diffusion, batch, dialysis). However, this crystallization method is very dependent on the nature of the detergent used and thus remain nowadays challenging in particular for choosing the good detergent. In this context, we use large scale MD simulations combined with experimental approaches (SAXS, SANS, SEC-MALLS) to gain insights into how the selected detergent buoy mimic the natural membrane and influence MP properties is essential. Recent results were obtained for large MPs such as the light harvesting complex type 2 from a purple bacteria Rps. acidophila or with the TonB MP protein (ShuA) from Shigella dysenteriae . These activities of research is done, in part, in collaboration with Dr. Françoise Bonneté, IBPC, Paris
Models of the MP ShuA inserted in DDM micelles with various numbers of surfactants or in a gram-negative bacteria outer membrane.

Made with Magix

© Tous droits réservés 2021

Developments of force fields for biomolecules Most of the empirical force fields used in classical MD simulations use the fixed-charge model when computing electrostatic forces and energies. In this approach, a single value for the atomic charge is assigned to each atom independently of the electrostatic environment. Atomic partial charges are derived and validated by using both experimental data and QM calculations and e ach   force   field   provides   its   own   set   of   atomic charges   for   standard   residues   or   other   biolecules   (e.g.   nucleaic   acids,   surfactant,   etc.)   in   its specific    format.    The “problem” of the atomic partial charges is that they are not observables   as   is   an   electron   density   and   that   there   are   no   standard   criteria   to   evaluate the   quality   of   them.   It   has   been   required   that   the   charges   should   be   independent   of   the computational   methods,   QM   basis-sets,   conformations   of   the   molecule   and   should   be transferable    and    conform    to    the    atom    electronegativity.    There    exist    several    different methods   for   deriving   partial   charges   such   as   Mulliken,   Löwdin   population   analysis   or empirical   approaches   to   reproduce   crystallographic or   liquid   data   (e.g.,   liquid   data,   or   crystal properties ), the electrostatic potential (ESP) derived charges using semi-empirical or ab- initio methods and the AM1-BCC approach. Unfortunately, none of the charge models has proved to be the best in all respects. For instance in case of the ESP charges are not easily transferable between common groups of homologous molecules and depends significantly on molecular orientation and conformation. In collaboration with Dr François Dupradeau (Unversité of Amiens, France), we used the “building blocks” approach implemented in the REDserver ( Development/ ) where a complex molecule (i.e. with different chemical groups) is split into different elementary chemical fragments with well-defined and controlled conformations. This approach has many advantages over the whole molecule approach to derive RESP charges: (i) the computational time required for geometry optimization and molecular electrostatic potential (MEP) computation is drastically reduced, (ii) the optimized geometry of the conformation(s) of each molecular fragment is fully defined and controlled, (iii) optimized conformations presenting intramolecular hydrogen bonds are discarded from charge derivation to avoid over-polarization effects, and finally (iv) a large set of analog molecules can be simultaneously involved in charge derivation leading to a homogeneous Force Field Topology DataBase ( FFTopDB ). This approach was successfully used to obtain RESP atomic partial charges for various biomolecules used in my MD simulation works including glucose polymers, photosynthetic pigments, modified amino- acids, and various types of surfactant fully compatible with the Amber force fields for proteins and sugars.
More info
Models of a pure LDAO micelles and a fragment of the lipid kinase PIK4A (DI)
More info
Models of aggregation process of two AOT RMs in isooctane with W 0 =5

Glycolipid based membrane systems

Bio-glycolipids prepared from the modified strain S. bombicola ΔugtB1 commonly called glucolipids (GL) present several interesting properties that can be useful for the development of several types of materials such as: hydrogels, vesicles, ultra-resistant foams, development of biocidal surfaces, stabilizers of nanoparticles for applications in the biomedical field, etc... Acidic glucolipid, characterized by a COOH group and a β- glucose unit, opposite each other show astonishing properties: the spontaneous formation of interdigitated lipid membranes whose charge is variable according to the pH. It has also been shown that the nature of the COOH group plays a very important role, in particular for the contribution of a negative charge which has the role of softening the elastic properties of the membrane. The effects of the pH and ions on these characteristics are not well understand, so far. To gain insights into these aspects and in particular their mechanical properties, MD simulations are used with various neutron scatterings approaches. T hese activities of research is done in collaboration with Dr. Niki Baccile, Laboratoire de Chimie de la Matière Condensée de Paris, Sorbonne Université, CNRS et Collège de France
More info
RESP charge derivation for the α- and β-anomers of glycolipids using the bulding block approach and applying INTRA- and INTER-MCC constraints during the fitting step are represented using fine dashed and dashed lines, respectively. This allows defining the molecular fragments required for the construction of the α- and β-anomers of DDM -surfactants.
Coarse grained model of a interdigitated membrane.with 100% of protonaned glycolipids The glucose and COOH heads are in yellow and red, respectively whereas the alkyl chain is in grey color.