Imagerie intégrative : de la molécule à l'organisme

General context

The main concern of our team is to study the structure-dynamics-function relationship of some proteins in order to understand their mechanism, and also to be able to inhibit their function when necessary. For this purpose we use biochemical and biophysical techniques, as protein engineering, Nuclear Magnetic Resonance (NMR), thermodynamics (ITC), Circular Dichroism (CD) and molecular simulations [Molecular Dynamics (MD), normal mode analysis (NMA) and docking]. In addition, we collaborate with external laboratories that use crystallography and Small Angle X-ray Scattering (SAXS) techniques.

Centrins and calcium-binding proteins

Fig. 1: Crystal structure of HsCen2 (gray) in complex with XPC (orange). The two Ca²⁺ ions are in cyan. The figures were made with VMD, and the cartoons were drawn using STRIDE.

Centrins. For the last few years we were more specifically interested in centrins, a family of calcium-binding proteins (CaBPs) from the EF-hand superfamily, which are well conserved in the eukaryotic realm. These small proteins (∼20 kDa) are composed of two relatively independent EF-hand domains. Each domain possesses two potential Ca²⁺-binding sites with varying affinities that are determined largely by the amino acid sequence of the binding loops.

Centrins are ubiquitous proteins, with a fraction concentrated in the microtubule organizer center (MTOC). MTOC is structurally and functionally similar to a basal body and it is known as spindle pole body in yeast and as centrosome in higher eukaryotes. However, the role of centrins in cell cycle regulation is not yet known. More recently, it was found that centrin plays also a role in diverse cellular processes as DNA repair, the control of the calcium channel in the ciliary membranes, the duplication of the Golgi apparatus and the export of mRNA and proteins from the nucleus.

In our group we have resolved by NMR the structure of the N-terminal domain of the human centrin (HsCen2; pdb code: 1ZMZ) and the C-terminal domain of the same protein, either free (code: 1M39) or in complex with two different peptides taken from centrin-target proteins: XPC (xeroderma pigmentosum complementation group C protein, a nuclear centrin target from the DNA repair, code: 2A4J) and Sfi1 (Suppressor of fermentation-induced loss of stress resistance protein 1, a large centrosomal target with several repeated sequence motifs that can bind centrin; code: 2K2I).  We have also contributed to the resolution of the crystal structure of the integral HsCen2 (2OBH).

In addition we have done several studies concerning the centrin stability (using mainly CD), its affinity for ligands (ITC) and its conformational changes (SAXS).
 

Prediction of the global form of centrins and other calcium-binding proteins (CaBPs).
Fig. 2: LAH for the investigated proteins. The horizontal line delimits between the predicted extended structures (LAH > 1.4) and the predicted non-extended ones (LAH ≤ 1.4). Vertical lines delimit between the known extended structures (filled circles), the known non-extended structures (open diamonds) and the unknown structures of centrins (filled triangles). For the unknown-structure centrins, we indicate the phylogenetic subfamilies. Generally, EF-hand CaBPs may be divided into two broad classes: those that bind calcium to regulate its concentration (calcium-buffering and calcium-transporting proteins) and those that bind calcium to decode its signal (calcium-sensor proteins). The two functional classes also have different structural features: calcium-buffering and calcium-transporting proteins, usually have a compact tertiary structure and are not conformationally sensitive to calcium-binding, whereas calcium sensor proteins, such as calmodulin and troponin C, have extended tertiary structures and show important conformational changes upon calcium-binding. In the extended form, the linker between the two domains may be structured in a straight helix, as for HsCen2, whereas, in the non-extended form, the linker is unstructured leading to either a floppy conformation, as for yeast centrin (Cdc31), or a very compact one, as for recoverin. It is important to understand the physical reasons for these differences, which would provide tools to predict the form of the CaBPs in general and of centrins, in particular, from their sequences, and therefore indicate to which class they belong.

In our team we have created a bioinformatics method, based on the inter-domain Linker Average Hydrophilicity (LAH) that allows to predict the form of CaBPs. This method represents a simple and powerful means to discriminate between extended and non-extended forms of calcium-binding proteins. Indeed, when tested on 17 known-structure CaBPs and applied to 59 unknown-structure centrins, it discriminated well between all the extended and non-extended forms of the known-structure CaBPs, and its predictions concerning centrins reflected well their phylogenetic classification (Fig. 2).

 The LAH algorithm can be used at: http://u759.curie.u-psud.fr/modelisation/LAH

Microtubule

Fig. 3: Molecular modeling of a 13-protofilament microtubule used to study the movements occurring during microtubule dynamics considering all the atoms. Each protofilament is composed of three dimers of α (red) and β (blue) tubulins.Microtubules (MTs) are spatially organized, extremely dynamic structures of the cytoskeleton. They play an essential role in the cellular division due to their dynamics and their interactions with other proteins. Most common MTs are composed of 13 protofilaments (pfs) of αβ-tubulin dimers (Fig. 3); they interact with the Microtubule associated proteins (MAP) and Plus-end tracking proteins (+TIPs ) which can stabilize or destabilize the structure. The dynamics of MT is also governed by the presence of GDP or GTP at the (+)-end β-tubulin, which will favor respectively its depolymerization or polymerization process. 

The smallest part of MT that makes sense contains 3 αβ-tubulin dimers in each protofilament, which makes a total of 315,432 atoms for the GDP-bound MT. What is remarkable is that the additional 65 atoms of the GTP-capped MT are sufficient to stabilize the entire structure. To understand this mechanism, we have used molecular simulations.

The most crucial motions of a system like MT are those of large amplitude, but they are difficult to apprehend at the atomic level by molecular dynamics. For this reason we have used the most appropriate method to study large-amplitude motions of big proteins or protein assemblies, namely, the normal mode analysis (NMA). In this method the motions are decomposed in a sum of vibrations, the lowest frequency vibrations of compact systems corresponding to the most collective motions.

The application of NMA to GDP- and GTP-bound MTs, in all-atom representation, allowed us to decipher the mechanism of stabilization of the microtubule by GTP (manuscript in preparation). We project now to understand the mechanism of some MAPs.

Drug-design

Protein kinases play important role in cell cycle regulation. In some cases it is necessary to inhibit the activity of some kinases, especially those overexpressed in cancer cell proliferation. We collaborate with the group of Chemistry (UMR176) in the Curie Institute, to find new families of small molecules that would be able to inhibit CK2 or GSK-3β, two kinases involved in cancer. For this purpose we use in our group, docking and virtual screening techniques.

Fig. 4: GSK-3β in complex with an active inhibitor (red), which was synthesized in UMR176 and docked in our group. The color code is as follows: the G-loop is in yellow, the two following strands (3 and 4) of sheet A are in orange, the negatively charged residues are in blue, the loop B-C (that relates sheets B and C) is in green.	Fig. 5: Poses of some active compounds. Docking of BSN (a, e), FSA (b, f), nitrile (c, g) and bromine (d, h) compounds in CK2α. In the upper panel, oxygen atoms are represented in red, nitrogen in blue, carbon in cyan, sulfur in yellow, bromine in green and hydrogen in white. In the upper panel, hydrogen bonds are represented as dotted lines. In the lower panel, the cavity is represented as a solid surface in which hydrophobic residues are figured in white, negatively charged in red, positively charged in blue and polar residues in green. Some key structural elements of the protein surrounding the active site are shown (in brackets when there is no interaction with the ligand).