The primary concern of our research group is to understand the dynamic interplay of protein interactions in multicomponent ABC transporters, taking the maltose transporter in E. coli as a paradigm.. We also make phylogenetic studies on the superfamily of ABC systems.

The ATP-Binding Cassette (ABC) superfamily is composed of systems that are widespread in all living organisms. These systems form the largest family of paralogues ever found. They are involved in a large number of living processes including primarily but not only transport. In humans, 15 severe genetic diseases including cystic fibrosis are caused by the dysfunction of ABC transporters (Mourez et al., 2000). The transporters are made of four structural domains: two very hydrophobic membrane spanning or transmembrane domains (TMD) and two hydrophilic cytoplasmic nucleotide-binding domains (NBD) peripherally associated to the cytoplasmic membrane (Higgins, 1992). The primary sequence of the hydrophilic cytoplasmic domains is highly conserved, displaying conserved Walker motifs A and B common to ATPases and another motif characteristic of ABC transporters, the LSGGQ or signature motif. These domains were found to bind and to hydrolyze ATP, thereby coupling transport to ATP hydrolysis (Schneider and Hunke, 1998). ABC transporters are involved in the export or in the import of a wide variety of substrates ranging from small ions to macromolecules. Import ABC systems are found only in prokaryotes and their four constitutive domains are carried in general by independent polypeptides. By contrast, export systems are found in all living organisms and have in general the TMD's fused to the NBD's.
ABC import systems are also called binding protein-dependent (BPD) transport systems(Boos and Lucht, 1996) (Saurin and Dassa, 1994).  In addition to the basal core structure of ABC transporters, they require for a proper function an extracytoplasmic substrate-binding protein, located in the periplasmic space of Gram-negative bacteria. BPD transporters are scavenging systems, able to extract trace elements from the environment.
The ABC maltose import system of Escherichia coli is a model of ABC transporters. Its functional characteristic are investigated in our group since the early 70's (Hofnung, 1974). We contributed significantly to the physiological, genetic and more recently to the molecular characterization of this system. A working model consistent with these studies is currently tested in our laboratory. According to this model, maltose and maltodextrins enter the periplasm by facilitated diffusion through a specific outer membrane porin coded for by the lamB gene, which is specifically required for maltose transport at sub-micromolar concentrations and for maltodextrins at all concentrations (Szmelcman and Hofnung, 1975). In the periplasm, maltose-binding protein (MBP) binds substrates at high affinity (KD = 1 ???. Upon binding of substrates, MBP undergoes a conformational change (Szmelcman et al., 1976) and interacts with a cytoplasmic membrane complex made of MalF, MalG (Dassa, 1993; Dassa and Hofnung, 1985; Dassa and Muir, 1993) and two subunits of MalK (MalFGK2), very likely with hydrophobic membrane proteins MalF and MalG (Hor and Shuman, 1993) that carry also a substrate binding site(s) (Treptow and Shuman, 1985). Upon substrate binding, MBP transmits through MalF and MalG a signal to MalK allowing it to hydrolyze ATP (Davidson et al., 1992). The questions which are addressed in our laboratory are the following: i) what are the sites on membrane proteins for interacting with each others; ii) how does ATP binding and hydrolysis result in conformational changes that mediate substrate translocation across the membrane and iii) how do ABC systems evolved and what are the rules that govern their functional diversity.

1. Interactions between cytoplasmic membrane partners
Integral membrane proteins of BPD transporters display a short hydrophilic conserved motif, the EAA motif located at about 100 residues from the C-terminus (Dassa and Hofnung, 1985). The motif is hydrophilic and was found to reside in a cytoplasmic loop (Saurin et al., 1994) so that it could constitute an interaction site with the conserved cytoplasmic ABC subunits (Dassa and Muir, 1993). Mutations in conserved residues lead in general to a transport negative phenotype (Dassa, 1990; Köster and Böhm, 1992). The different phenotypes of mutations suggested that EAA regions were involved in two kind of interactions, binding of NBD's to the membrane and transmission of functional signals to NBD's. Suppressor mutations restoring transport were found in the malK gene. They were mapped mainly into the so-called helical domain. The conclusion of these studies are in favor of the idea that EAA regions of TMD's are in close contact with the helical domains of NBD's (Mourez et al., 1997). Hence, we propose that the EAA region-helical domain interface would couple the energy of ATP hydrolysis to transport.

2. Conformational changes linked to ATP hydrolysis
The interactions of NBD's with the TMD's were analyzed in vitro. MalK was shown to bind to everted membrane vesicles made from a strain lacking this protein and to reconstitute a functional system for maltose transport and ATP hydrolysis (Mourez et al., 1998). MalK binding to vesicles was dependent on the presence of MalF and MalG. ATP binding and hydrolysis were shown to increase the binding of MalK, suggesting that its interaction with MalF and MalG was tighter in the presence of nucleotides. When incubated with nucleotides, MalK undergoes a large conformational change, detected by changes in intrinsic fluorescence and in the proteolytic pattern, that renders the helical domain accessible to proteases (Schneider et al., 1995). We have shown that a single proteolytic site, unveiled by ATP binding in purified MalK was not accessible when MalK is reconstituted with MalF and MalG. This suggests that the MalK helical domain would insert into the MalFG complex in the presence of ATP, a notion which is consistent with the higher efficiency of binding and with the observation that HisP and MalK were accessible to proteases from the exterior  of spheroplasts when the cognate TMD's were present (Baichwal et al., 1993; Schneider et al., 1995). These studies strengthen the idea that the helical domain of NBD's is crucial for interactions with TMD's (Liu and Ames, 1998).
We have investigated the proximity of residues in the conserved EAA sequence of MalF and MalG to residues in the helical segment of the MalK subunits by means of site-directed chemical cross-linking (Hunke et al., 2000). The results suggest that residues A85, K106, V114 and V117 in the helical segment of MalK, to different extent, participate in constitution of asymmetric interaction sites with the EAA loops of MalF and MalG. Furthermore, both MalK monomers in the complex are in close contact to each other through A85 and K106. These interactions are strongly modulated by MgATP, indicating a structural rearrangement of the subunits during the transport cycle. The number and the quality of contacts between residues of the MalF-MalG EAA motifs and the MalK helical domain was increased in the presence of ATP. Thus, it is tempting to speculate that the observed changes in subunit-subunit interactions may reflect structural rearrangements that are required to initiate ATP hydrolysis and transport.

3. Evolution of ABC systems in living organisms
We were among the first groups to analyze the sequence relationships between constituents of ABC systems. The sequences of BP's display very little overall sequence similarity in contrast with their closely related tertiary structures. A careful sequence analysis of 52 BP's revealed that they can be grouped in eight families of more strongly related proteins (Tam and Saier, 1993). Remarkably, clustering of sequences was in agreement with the chemical nature of the substrate. The degree of sequence divergence was too high to establish rigorously that all binding proteins are homologous, i.e. that they are of common origin, but this idea is supported by their common structural organization.
TMD's are in general more conserved than BP's but it is nevertheless difficult to demonstrate that they are homologous. It is possible to group the proteins in clusters of strongly related sequences and here also clustering reflects the substrate specificity of the system (Saurin and Dassa, 1994). The clusters defined in such analyses are strongly similar to those defined for substrate binding proteins. All integral membrane proteins of BPD transporters display the EAA conserved motif (Saurin et al., 1994).
NBD's display a region of homology that extend over 200 residues. A phylogenetic analysis of the sequence of BPD transporters NBD's revealed that these proteins descend from an ancestor protein (Kuan et al., 1995). Moreover the proteins segregate into clusters strongly related to those found for TMD's and for the BP's. This finding support the view that operons encoding partners of BPD transporters evolved as a whole unit by duplication.. Since similar BPD transporters are found in archaea and eubacteria, the putative common ancestor of such systems appeared probably before the separation of the two phyla, about 4 billions years ago. Recently, by analysing about 200 NBD's from eukaryotes and prokaryotes, we showed that ABC importers and exporters segregated in two different groups of sequences, independently of their origin. This remarkable disposition strongly suggests that the divergence between these two functionally different types of ABC systems occurred once in the history of these systems and probably before the differentiation of prokaryotes and eukaryotes (Saurin et al, 1999).