Unit: Bacterial Membranes
Director: WANDERSMAN Cécile
Our work concerns membrane transports in Gram negative bacteria with four major research themes: heme acquisition systems requiring an extracellular hemophore, protein secretion by ABC exporters across the two membranes which delimit the Gram negative cells, a two component system which is involved in iron acquisition and the characterization of ABC proteins of unknown functions.
1) Hemophore dependent heme acquisition systems (Sylvie Létoffé, Laurent Debarbieux, Philippe Delepelaire, Francis Biville, Fréderic Huché, Hélène Cwerman and Najla Benevides Matos)
Heme which is a major iron source is uptaken in Gram negative bacteria by two principal pathways. One involves the direct contact between heme or heme-containing proteins and specific bacterial cell surface receptors. The second requires the secretion of hemophores, a family of proteins discovered in our laboratory in 1994. These proteins, capture free heme or extract heme from heme carrier proteins, owing to their higher affinity for heme and return it to hemophore-specific outer membrane receptors. The Serratia marcescens hemophore dependent heme acquisition system (Fig. 1) consists of the iron regulated has operon encoding HasR, the hemophore-specific outer membrane receptor, HasA, the hemophore, HasD and HasE, the specific inner membrane hemophore secretion proteins. The last gene of the has operon, hasB encodes a TonB homolog specific for heme. The activity of HasR is dependent on a protein complex comprising the inner membrane proteins ExbB, ExbD, and TonB or HasB. This is a property shared with several other outer membrane iron receptors whose 3D structures have been elucidated showing an N-terminal domain closing the receptor pore and exposed to the periplasm, where it can make contact with TonB.
Structural analysis performed in collaboration with the NMR laboratory of Institut Pasteur and genetical studies performed on the S. marcescens hemophore have revealed the 3D structure of holo and apo-hemophore and allowed to identify the heme iron ligands (Fig. 2). Both heme-free and heme-loaded HasA bind to HasR to the same or overlapping site with the same apparent Kd (5nM). The binding of HasA to HasR involves two beta sheets located on the same side of HasA. The synthetic peptide corresponding to one beta strand competes with the corresponding hemophore region for binding to the receptor, suggesting that the two binding regions are independent binding sites. We propose that this double binding distorts the protein allowing heme transfer to the receptor.
HasR also allows free heme uptake owing to two histidines conserved on other heme receptors. It binds heme albeit with a lower affinity, in the micromolar range. It is not known how heme is transferred from HasA a protein with a very high affinity (Kd : 10-11M) to HasR a protein with a lower affinity (Kd around 10-6M). Active wild type and histidine mutant HasR proteins were purified in detergent. Spectrophotometric analysis of purified HasA, HasR and HasA-HasR complexes using wild type and mutant HasA and HasR proteins showed that heme transfer from the hemophore to the receptor, occurs "in vitro" and is thus energy independent. We are presently trying to cristalize the heme-loaded HasA-HasR complex in collaboration with W. Wellte in Constance, Germany.
Whereas heme is uptaken as a whole through heme receptors, hemophores are not transported and have to be stripped off at the cell surface: only the heme moiety being uptaken, the apo-hemophore remains outside. Since apo and holo-hemophores have the same affinity for the receptor, it is not understood how exchange between the apo- and holo-forms occurs.
We showed that empty-hemophore release from the receptor is energy driven and concomittent with heme uptake.
In Gram negative bacteria, many outer membrane receptors involved in iron uptake are positively regulated by the availability of a specific iron source via sigma factors and membrane bound anti-sigma factors. An extra cytoplasmic signal releases the sigma from the anti-sigma inhibition and thereby leads to the transcription of the target operons.
The S. marcescens has system is subject to this type of regulation. The has operon is induced only by holo-HasA. The signaling process is dependent both on HasR and on its two upstream genes hasI encoding the sigma factor and hasS encoding the anti-sigma factor.
Using a collection of HasA and HasR mutants we showed that heme transfer from HasA to HasR is required for induction and that one of the two HasA binding sites (the 51-60 domain) is also required for induction. Using HasA mutants unable to bind heme and also mutated in either one or the other binding regions (51-60 or 93-106), we showed that the inducing stimulus can be brought by two molecules: heme and the 51-60 domain. We are presently testing induction by the synthetic 51-60 peptide in the presence of heme.
Like the other iron starvation sigmas, hasI is repressed by iron-loaded Fur protein, but is not auto regulated. Unlike known iron starvation ECF anti-sigma factors, the transcription of hasS is directly regulated by HasI and consequently also subject to the holo-hemophore signaling cascade. When there is enough heme, there is a storage of inactive HasS molecules which become active when HasR is not occupied by holo-hemophore ligand molecules: as soon as there is a heme shortage transcription is turned off.
We have undertaken the molecular dissection of HasR using a computer modeled 3D structure (Fig. 3). The most striking result is that HasR beta barrel alone forms a specific heme channel requiring the conserved histidine residue to be functional.
We are presently looking whether such specific heme porins might correspond to naturally occurring systems to assimilate heme for bacterial species growing in heme rich environments.
Heme uptake in S. marcescens requires either TonB or HasB, a TonB homolog specific for heme uptake that cannot complement the other TonB-dependent functions. S. marcescens has at least two heme acquisition systems, the Has system working at very low heme concentration (25nM) and a second one which requires higher heme concentrations. Sequence analyses, using the Db11 sequence database under construction at the Sanger Institute (Cambridge, UK) shows the presence of a HemR homolog (a receptor present in several Gram negative bacteria and allowing hemophore-independent heme acquisition).
The S. marcescens hemR homolog has been cloned together with the adjacent genes hemP and hemS in an E. coli hemA mutant and the recombinant strain is able to acquire exogenous heme.
We are presently testing heme acquisition in tonB mutants and in tonB mutants expressing HasB6, the HasB mutant functional in E. col, to determine whether HasB is functional with a heme receptor other than HasR. This will help to understand the molecular basis of HasB specificity .
2) Protein secretion by Gram negative ABC transporters (Sandra Cescau, Laurent Debarbieux, Annick Paquelin and Philippe Delepelaire)
Four secretion pathways have been actually described in Gram-negative bacteria. Type II pathway requires the universal N-terminal signal peptide and the general export pathway to cross the inner membrane. Type I and III pathways are independent of the signal peptide and sec system.
The Type I transporter is characterized by the presence of the ABC protein belonging to the ATP Binding Cassette class of proteins which are implicated in the vectorial movement of solutes across biological membranes and several play essential functions. The transporter comprises two other membrane proteins, one in the inner membrane and one in the outer membrane (Fig.1).
Our work has largely contributed to show that bacterial ABC protein exporters are widespread in Gram-negative organisms. We have shown that in several cases (such as the E. coli hemolysin transporter) the outer membrane component belongs to the TolC family and is a multifunctional protein involved in drug efflux and colicin import. All the proteins following this pathway, including the hemophore, have a C-terminal secretion signal which remains accessible owing to cytoplasmic chaperones which are required for secretion. The signal interacts with the ABC protein, modulates its ATPase activity and induces the formation of a secretion multiprotein complex.
We have shown that a hemophore mutant deleted for its C-terminal secretion signal still interacts with the ABC protein. We are presently trying to localize this second interaction site on HasA and testing whether this second interaction site drives the multiprotein complex formation. We are presently purifying the substrate bound ABC transporter multiprotein complex to study it by cryomicroscopy. This is done in collaboration with the laboratory of Electronic Microscopy of the Institut Pasteur and abroad with Glaeser, Berkeley (USA).
Our future researches are directed towards the study of the mechanisms of chaperonesubstrate interactions and the reconstitution of a complete secretion system " in vitro ".
Pentapeptide scanning mutagenesis of the ABC protein is undertaken to define its domains of interaction with the substrate and the other transporter proteins.
3) The two component system YgiX/YgiY (Francis Biville)
This system is present in many Gram negative bacteria. YgiY shares homologies with two component sensor kinases and YgiX with two components transcriptional activators. It appears to be involved in regulation of several E. coli functions such as motility, biofilm formation and iron acquisition. We have shown that inactivation of YgiY allows iron bound to pyrophosphate utilization even in the absence of enterochelin suggesting that YgiY/ YgiX regulates a yet not identified iron transport system. YgiY has a EXXE motif present on several iron binding proteins. We are presently searching whether YgiY binds radioactive iron and trying to chemically characterize the potential iron chelator produced by YgiY mutant.
4) Functional and phylogenetic analysis of ABC systems (Dorothée Murat, Laurent Goncalves et Elie Dassa)
ABC proteins lacking transmembrane domains are involved in gene expression regulation in eukaryotes. The function of homologous ORFs in prokaryotes is unknown. We inactivated two such ORFs, uup and yheS in E. coli.
Although mutants do not display obvious phenotypic changes in standard laboratory conditions, they are strongly disfavored in stationary phase when grown in mixed cultures (ratio 1 : 1) with the parental strain. They are not detectable after 4-5 days. The effects of mutations are additive since the double mutant is undetectable after 2 days. The genes would be involved in the stationary phase fitness of bacteria. This decrease in fitness is not observed in mixed cultures when a filtration membrane (0.2 μm) separates bacteria. This result suggests that a cell contact is needed to obtain mutant killing. We hypothesize that a parent strain cell surface component induces the killing of mutant cell.
We improve our phylogenetic analysis of ABC systems in living organisms and we update the ABCISSE database ( http://www.pasteur.fr/recherche/unites/pmtg/abc/database.iphtml). We recently observed that plants share with bacteria a few ABC systems that are not detected in animal genomes. These systems might originate from the bacterial ancestors of chloroplasts.
Figure 1 : Has systems
Figure 2 : 3D structure of Apo and Holo HasA
Figure 3 : Computer modeled Has R structure
Keywords: Membrane transport, iron acquisition, hemophore, ABC protein