The project SENSOR groups 3 teams with expertise in chemistry (Team 3), molecular and structural biology (Team 2 and 1 respectively) in an integrative structure-function study of 4 key-sensors which belong to the large family of periplasmic binding proteins (PBP) and are involved in the perception of plant-derived compounds such as GABA and opines (resulting from condensation of amino acids, organic acids or sugars). These 4 PBPs and their ligands are in turn involved, or potentially involved, in quorum sensing and virulence of the plant pathogen Agrobacterium tumefaciens strains C58 and B6. A. tumefaciens B6 and C58 are model strains of which the complete genome sequence is available. The bacterial pathogen A. tumefaciens genetically engineers the plant host by transferring a piece of DNA (the T-DNA) from its tumor inducing (Ti) plasmid to the nuclear genome of plants. Proliferation of the transformed plant cells results in formation of a tumor (crown gall disease) in numerous plants of agronomical (poplar, tomato plant) and horticultural (rose) interest. In plant tumors, expression of the T-DNA redirects the metabolism of plant cells towards the production of opines, which are used by the pathogen as nutrients (C and N sources) and signals to control the expression of some virulence functions. In bacterial cells, opines stimulate the synthesis of a quorum-sensing (QS) signal which increases aggressiveness of Agrobacterium and activates the dissemination of the virulence Ti plasmid by horizontal transfer (by conjugation). In 2009, Team 1 and Team 2 joined their expertises (co-direction of a PhD, Sara Planamente 2009-2011) to understand how the A. tumefaciens C58 solves the paradox between opine-induced biosynthesis and GABA-induced biodegradation of quorum-sensing signals by studying a PBP (Atu2422) which binds GABA but also some other amino-acids. They demonstrated that free proline (Pro), which accumulates in the plant tumour, binds Atu2422, antagonizes import of GABA and consequently prevents the GABA-induced degradation of the bacterial quorum-sensing signal. In A. tumefaciens, other PBPs are involved in the perception and importation via ABC transporters of plant-derived signals including opines and GABA. Surprisingly, none PBP structure has been solved in complex with opines so far and a unique PBP structure in complex with GABA (but this PBP Atu2422 is not specific to GABA) was recently published by the teams implicated in the proposed SENSOR project. The three PBPs involved in opine binding are NocT and AccA in A. tumefaciens C58 and OccJ in A. tumefaciens B6. That involves in GABA binding only is Atu4243 from A. tumefaciens C58. Moreover, bacterial genome databases reveal that putative orthologs of these 4 PBPs are present in several animal- and plant-interacting bacteria including Pseudomonas, Burkholderia and Rhizobium. SENSOR project will (1) solve and compare the molecular structure and affinity of these 4 PBPs-sensors involved in the binding of opines and GABA, (2) evaluate their roles in quorum-sensing and virulence of Agrobacterium using appropriate KO-mutants and transcriptomics, and (3) will use them as references for defining structural-functional clusters among PBPs that are present in databases. This project will improve our basic knowledge on the structure-function relationship and virulence regulation of one the most abundant families of ligand-binding protein (PBP) in pathogenic and symbiotic bacteria. In addition, several actions for dissemination of these data to scientific and non-scientific communities are proposed.

Cytoskeletal motors are integral to all forms of eukaryotic life, with roles ranging from intracellular transport to cell division. They use the energy released upon ATP hydrolysis to produce work and most often move either along actin filaments, in the case of myosins, or along microtubules for dyneins and kinesins. There is a tight coupling between the nucleotide cycle (binding, hydrolysis and release) and the structural changes at the basis of the motor mechanism. Atomic structures of functional motor domains of all classes of cytoskeletal motors (kinesin, dynein and myosin) detached from their respective filaments are known. Besides, electron microscopy has provided low resolution models of decorated filaments. So far, structural methods have not yielded atomic-resolution descriptions of the motor-filament complexes. The biochemical properties of cytoskeletal motors change considerably upon interaction with their cognate filament. Consequently, available structural models do not account for many essential properties of the cytoskeletal motors. Our main goal is to establish the first complete structural cycle for a cytoskeletal motor by determining by X-ray crystallography the atomic structure of motile kinesin functional motor domains in complex with tubulin, the microtubule building block, along with the kinesin nucleotide cycle. We will also study microtubule depolymerising kinesins. To reach this objective, this proposal relies on the development of tubulin sequestering tools, which are in fact ‘chaperones’ for crystallization. We have selected anti-tubulin DARPins in the course of the ‘mec-tub’ ANR project. Whereas these DARPins have already proven useful and are being used as such, their affinities for tubulin (Kd in the 0.1 µM range) are limiting in some cases. Our first step serves as a basis for the whole project. We will screen available DARPins to identify those that do not interfere with kinesins for tubulin binding and evolve the selected ones towards higher affinities. The resulting DARPins will be submitted to crystallization experiments as ternary complexes with tubulin and functional motor domains of several motile kinesins, in their three main nucleotide states (ATP, ADP and nucleotide free) as well as bound to a transition-state analogue of nucleotide hydrolysis (ADP-AlFx). Our preliminary results clearly demonstrate this is possible. This will fulfil our main goal. A second objective is the determination of the structural mechanism of a microtubule-depolymerising kinesin complex. These kinesins, in the kinesin-13 family, target microtubule ends and induce disassembly instead of walking along microtubules. Their nucleotide cycle, adapted to their function, differs from that of motile kinesins. In particular, they interact with tubulin only when they are ATP-bound whereas motile kinesins bind to tubulin in the ATP- and nucleotide-free states. In this case, the main target is the structure of a kinesin-13(ATP state)-tubulin complex. We have expressed functional domains of the three human class 13 kinesins but, as they all have similar physico-chemical properties, we also plan to study more divergent ones, e.g. from diatoms. These kinesins will be characterized biochemically to ascertain their depolymerase activity and, in case this is indeed verified, structurally.

Little is known today about the structure of full-length prions. We recently established, using solid-state NMR studies of the Ure2p and HET-s proteins, that prions show different architectures, and that the building principles between them can differ considerably. We here aim at the study of the Sup35 prion, which poses a formidable challenge considering its large size. Having demonstrated the important role of the globular domain, we here aim to study assembly in the context of the entire proteins, as opposed to fragments, and investigate on a molecular level the role of the functional C-terminal domain of the protein in fibril assembly. We aim thus to compare the structures of fibrillar Sup35NM and full-length Sup35p by solid state NMR. The fragment Sup35NM is often used as a model for the full-length prion. We will compare the structures of fibrillar Sup35NM and full-length Sup35p by solid state NMR and determine whether the structure of this fragment is indeed preserved in the context of the full-length fibril, or if it is different. This question is particularly relevant in the light of our previous results on the full-length HET-s and Ure2p prions, which show that the structure of the isolated prion and globular domains are not conserved in the context of the full-length prions. We will address this question by solid-state NMR, which is an emerging technique for the structural study of insoluble proteins. Important developments over the last decade have pushed this fast evolving technique to become a serious partner in structural biology. If it has been shown in proof-of-principle experiments that amyloid and prion structures can be determined by solid-state NMR methods, it is still difficult to obtain structures of full-length prions. The main difficulties are caused by the large size of these proteins, a property they also share with other insoluble proteins. However, many elements of the technology needed to tackle these problems are now available in the applicants’ laboratories and we propose to fully develop the NMR methodology and, in parallel, apply it to the yeast prion Sup35p that is made of 685 amino-acid residues. The key NMR methodology developed in the context of this proposal will allow us to establish structural models which will be confronted with the large body of biophysical and biochemical and functional knowledge to work out the relationship between structural and biological features of Sup35p. Besides, the protocols and techniques developed will also be applicable to other large proteins and should allow to open structural studies on other insoluble proteins.