Institut Pasteur Rapport d'activité de l'unité Génétique moléculaire pour l'année 1999

CNRS URA 1773


Responsable : Anthony P. PUGSLEY (max@pasteur.fr)

Résumé du rapport

Les processus complexes d'activation de transcription et de sécrétion de protéines impliquent tous les deux un nombre impressionnant d'événements qui peuvent être analysés par des approches biochimiques et génétiques. Dans l'unité de Génétique Moléculaire, nous avons contribué de façon significative à la compréhension moléculaire de ces deux processus par nos études, chez la bactérie Escherichia coli, de l'activation de la transcription par la protéine MalT et de la sécrétion d'une amylase (pullulanase) dont la production est contrôlée par la protéine MalT. Dans les deux systèmes, nous mettons en route des analyses fonctionnelles, structurales, génétiques et biochimiques afin d'obtenir encore plus d'informations sur ces deux processus.

Abstract

The complex processes of transcription activation and protein secretion involve a large number of events that can be dissected and analysed by genetic and biochemical methods. The Molecular Genetics Unit has made a major contribution to the molecular understanding of these processes through its studies of two model systems: transcription activation by the MalT protein and secretion of a MalT-regulated amylase-like enzyme by the bacterium Escherichia coli. In both systems, the analysis has progressed to a stage at which structural, biochemical, biochemical and functional studies combine to give exciting new insights into the underlying mechanisms.

Texte du rapport

Pullulanase secretion (Anthony P. Pugsley)

Pullulanase is an amylase-like enzyme that is secreted by Klebsiella oxytoca, a close cousin of E. coli. The expression of the pullulanase structural gene, pulA, and of a minimum of 12 genetically-linked genes permits the production and secretion of the amylase in recombinant E. coli. Secretion occurs in two totally independent steps, the first involving the Sec machinery located in the cytoplasmic membrane and the signal peptide at the N-terminus of the pullulanase precursor protein. Once this first step in secretion has been accomplished, the enzyme folds into its final, active configuration in the periplasm. Ultimately, the enzyme is transported from the periplasm across the outer membrane in a process that involves the products of the 12 genes that were cloned from K. oxytoca.

Similar secretion pathways have now been identified in a large variety of Gram-negative bacteria. In many cases, the secreted proteins are toxins or enzymes involved in animal or plant pathogenesis. Thus, besides its considerable fundamental importance (this is one of very few translocation machineries that translocate fully-folded and even multimeric proteins through biological membranes), this so-called secreton or type II secretion pathway is of considerable medical importance.

Some of the major recent advances and current goals are listed below.

Secretion and type IV pili (Anthony P. Pugsley)

A set of proteins closely related to those involved in type II secretion is involved in the assembly of type IV pili in a variety of pathogenic bacteria. Among the genes that are common to the two systems are pilin or pilin-like (pseudopilin) proteins, a pilus assembly ATPase and a pilin-specific signal peptidase (prepilin peptidase). Recent studies show that under certain circumstances, the secreton can assemble extremely long pili composed of pseudopilin or pilin on the cell surface (in collaboration with Pierre Gounon, Station de Microscopie électronique). The physiological role and structure of these pili are currently under investigation. X-ray crystallography of purified pilins and pseudopilins will be combined with lower resolution analysis of purified pili. Among the possible physiological roles that will be investigated are anchoring of bacteria to cell surfaces or enzyme substrates or channel formation to permit the transport of proteins through and beyond the outer membrane and surface layers around the cell.

Secretin and its specific pilot protein (Ingrid Guilvout, Nicolas Bayan)

Two of the proteins found in both type II secretion and type IV piliation systems are secretin and its specific pilot / chaperone protein. In the pullulanase secretion system, 12 copies of each of these two proteins (PulD and PulS, respectively) form a barrel-like complex with radial spokes that has been visualised by electron microscopy using negatively-stained and unstained samples. Since these two proteins are the only components of the secreton that are firmly associated with the outer membrane, it seems reasonable to assume that they form the channel by which pullulanase crosses this membrane. Indeed, the channel within the PulD-PulS barrel is sufficiently wide to accommodate a globular protein the size of pullulanase. However, in vivo and in vitro experiments indicate that the channel is tightly closed, possibly because the large N-terminal domain of PulD blocks the channel opening on the periplasmic side of the membrane or even fills the barrel cavity. Current studies are devoted to identifying specific domains and subunits within the PulD-PulS complex (in collaboration with Eric Larquet, Station de Microscopie électronique) and to studying the channel gating process.

It has been proposed that type IV pili are anchored in the outer membrane by the secretin complex. In the secreton, pilus formation requires the PulD-PulS secreton complex and one would therefore assume that the pili are located within the lumen of the PulD-PulS barrel. Indeed, the diameter of an individual pilus filament (6-8 nm) is similar to that of the secretin channel. However, occupation of the lumen of the PulD-PulS channel by pili would have two important consequences for secreton organisation and function. First, pilus filaments are bundled into complex fibres, implying that the secretin channels must be clustered on the cell surface. Second, the pilus filament would block the PulD channel completely, thereby preventing movement of pullulanase through the same channel. At present, however, we lack direct evidence that either the pilus or pullulanase crosses the outer membrane by the secretin channel. Therefore, part of our current efforts are devoted to examining the distribution of the secretin channel on the bacterial cell surface and to studying its interactions with pullulanase and the secreton pilus.

Substrate recognition and energetics of secretion (O. Mary-Possot)

Bacteria secrete only a limited number of proteins by the secreton, implying the existence of specific recognition between these proteins and the secreton. In the case of the system studied in our laboratory, only one protein, pullulanase, is secreted. Furthermore, secretons from one bacterium cannot secrete proteins normally secreted by secretons in unrelated bacteria. However, most of the components from the Out secreton of Erwinia can substitute for their homologues in the Pul secreton, implying that they are not involved in secretion specificity. Only two proteins, secretin PulD (see above) and PulC cannot be replaced by their homologous Out counterparts.

PulC is of particular interest because we have found that it interacts with PulD in the outer membrane while remaining anchored in the cytoplasmic membrane. We are currently purifying this protein to study its structure. We are particularly interested by its apparent trimeric organisation and the role of a putative protein-protein interaction (PDZ) motif located close to its C-terminal end. In addition, mutagenesis and gene fusion techniques will be used to try to identify regions of the protein involved in pullulanase binding.

We have shown that pullulanase secretion requires the proton motive force, implying the existence of a coupling mechanism that transduces a signal from the energised cytoplasmic membrane to the outer membrane. By virtue of its proposed contact with both the cytoplasmic and outer membranes, and particularly with PulD, PulC is one of the components of the secreton that could play a role in energy transduction. In addition, one of the secreton components, PulE, shares sequence characteristics with ATPases. The role of PulE in secretion has not been defined. One possibility is that ATP hydrolysis by PulE drives secretion. However, it seems more likely that its role is in the export or assembly of the secreton pseudopilins. Therefore, we will investigate both the energetics of pseudopilin export and assembly and examine the biochemical properties of PulE itself.

Protein-protein interactions in the secreton (Frank Ebel)

The majority of the secreton components are located in the cytoplasmic membrane. Various approaches used to probe protein-protein interactions have suggested that many of them form a complex that anchors PulE to the cytoplasmic face of this membrane. To probe these interactions in greater detail and in order to identify minor secreton components to which we have so-far devoted little attention, we are generating banks of monoclonal antibodies. In addition, epitope-tagged derivatives of secreton components are being used to extract protein complexes from detergent-solubilized bacteria in an attempt to purify them for structural and functional analyses.

One attractive idea that we wish to pursue is that the complex that includes PulE is involved in pseudopilus export and assembly. In vivo and in vitro studies are being used to determine whether the export of pseudopilins across the cytoplasmic membrane occurs by the Sec pathway or requires specific secreton components. In addition, we will study the coupling between export and assembly of pseudopilins and the possibility that secreton components involved in their assembly are located in specific parts of the cell membrane. Green fluorescent protein hybrids are currently being constructed for this purpose.

The E. coli secreton (Olivera Francetic)

Genome sequencing revealed the existence of an apparently complete set of secreton genes (called gsp) in the E. coli K-12 genome. We have shown that these gsp genes are transcriptionally silent because the global repressor H-NS is bound to their promoter region. Derepression of gsp transcription occurs when the hns gene is deleted. Despite extensive searches, other genes regulating gsp have not been found, and the genes remained transcriptionally silent under all tested growth conditions. However, 7 out of 9 gsp genes tested were able to complement mutations in the corresponding genes of the pullulanase secreton, indicating that they remain functional and suggesting that they might be controlled by a specific but as yet unidentified extracellular inducer.

In an attempt to identify proteins that might be secreted by the E. coli K-12 Gsp secreton, we screened the E. coli gene data base for genes coding for proteins that one would expect to be extracellular. One of the genes identified (chiA) was shown to code for a functional endochitinase. Since chitin is a large polymer, we reasoned that the chitinase should be secreted into the growth medium. Transcription studies demonstrated that the chiA gene encoding this chitinase is repressed by H-NS, as are the gsp genes (see above). However, chitinase remains largely periplasmic when chiA and the gsp genes are derepressed in an H-NS mutant. Searches for other potential substrates for the E. coli secreton are being pursued.

Pullulanase as a secretion locomotive (Olivera Francetic)

The secreton, and in particular pullulanase, have several features that appear to make them suitable for the secretion of hybrid proteins in E. coli. First, the secreton is able to secrete folded and even multimeric proteins. Thus, heterologous proteins that fold in the periplasm might be accommodated by this secretion machinery. Second, we have reported that two, relatively small regions of pullulanase (the proposed secretion signals) are necessary for pullulanase secretion and sufficient for the secretion of a heterologous protein, beta-lactamase, to which they are fused. Finally, pullulanase is a lipoprotein that remains anchored to the bacterial cell surface after it has crossed the outer membrane. Successful adaptation of this system would allow heterologous proteins to become surface-anchored, which might be advantageous for the construction of live vaccine strains or in the purification of recombinant proteins.

To test the potential applications of the pullulanase secretion system, we have constructed a series of protein hybrids containing different reporter proteins fused to complete pullulanase or the proposed secretion signals with or without the N-terminal fatty acid anchor sequence. The results of these studies show that only a very restricted range of heterologous reporter proteins can be secreted by this system. Furthermore, the fatty acylated N-terminus appears to be required for efficient secretion of reporter proteins fused to the secretion signals but is dispensable for secretion of full-length pullulanase.

Lipoprotein sorting in the cell envelope (Anthony Pugsley)

Studies of the lipoprotein pullulanase have recently led us into a new area of research on factors that determine how lipoproteins are targeted to their correct location in the cell envelope. In E. coli, all lipoproteins face the periplasm but they can be anchored to either the cytoplasmic - or outer membrane. We have shown that the essential sorting signal is the amino acid at position +2 of the signal peptidase processed polypeptide (position +1 is invariably a fatty acylated cysteine residue). If this amino acid is an aspartate, an aromatic amino acid or a proline, the lipoprotein will be located in the inner membrane. All other amino acids at this position except lysine or arginine (which interfere with export) direct lipoproteins to the outer membrane.

Current studies are aimed at determining whether the "+2 lipoprotein sorting rule" applies to other proteins and at designing genetic selection procedures to identify genes coding for protein components of the lipoprotein sorting machinery. In addition, we are exploring the use of this system as a target for bactericidal agents that could be used in the treatment of chronic infections caused by Gram-negative bacteria.

Transcription activation by MalT

MalT is the transcription activator of the E. coli maltose regulon, which is composed of 7 operons encoding proteins involved in maltose and maltodextrin transport and metabolism. In K. oxytoca, MalT also activates transcription of the pullulanase structural gene and of the divergently-transcribed operon of genes necessary for pullulanase secretion. MalT is the best-characterised representative of a new family of large (ca 100 kDa) transcription activators found in both Gram negative and Gram positive bacteria. One distinctive feature of MalT is that the protein itself is the target of multiple controls. Its activity is indeed modulated by at least two (and possibly three) negative effectors and by two positive effectors, ATP and maltotriose (the inducer of the maltose regulon), both of which are required for transcription activation. The aim is to understand how all of these regulatory signals are integrated at the level of the protein and control its activity.

Another characteristic feature of MalT is the structural diversity of the promoters controlled by the protein. Activation of transcription from the MalT-dependent promoters depends on MalT binding to an array of sites whose arrangement varies with respect to their number, their relative position and their orientation. Interestingly, MalT binds co-operatively all of the activating MalT sites present in each promoter whatever their configuration, which suggests that MalT binds promoter DNA as a multimer and that some special features allow the protein to recognise various arrays of MalT sites. All of the prokaryotic transcriptional regulators whose mode of action is well understood recognise regular arrays of their cognate sites. Therefore, we expect that the nucleoprotein complex that results from MalT binding to its promoter targets will have distinctive properties.

Ligand-induced self-association of MalT and promoter binding (Evelyne Richet)

To understand how MalT binds promoter DNA and to delineate the role of ATP and maltotriose as positive effectors of MalT, we have recently analysed the quaternary structure of the liganded and unliganded forms of MalT in solution through the use of several biophysical and biochemical approaches. This study showed that the protein is monomeric in the absence of ligands and that ATP and maltotriose induce MalT multimerization when present together. The size of the multimers formed increases with increasing protein concentration. Cryo-electron-microscopy analysis revealed distinct, highly-structured particles associated in long chains of various length. In collaboration with Eric Larquet (Station Centrale de Microscopie électronique), we are presently analysing the structure of these particles by image processing and we are examining ways to determine their size and to localise domains of MalT (see below). We are also using cryo-EM to analyse the nucleoprotein complexes resulting from MalT binding to promoter DNA to determine the oligomeric structure of MalT that is the basis of these complexes.

Repressors of MalT activity (Evelyne Richet)

Three proteins, MalY, MalK and Aes, negatively control MalT activity in vivo, bringing an additional level of complexity to this already intricate regulatory system. One of our goals is to determine how each of these proteins affects MalT activity. We have begun by analysing repression by MalY, which we have found to inhibit the binding of maltotriose to MalT by interacting directly with MalT. We are now analysing the mechanism by which MalK down-regulates MalT.

Structure-function analysis of MalT (Olivier Danot)

Early attempts to obtain crystals of MalT for high-resolution X-ray analysis were unsuccessful. Therefore, we resorted to the analysis of MalT domains obtained, initially, by limited proteolysis of the protein. Different proteolytic fragments were observed when proteolysis was performed in the presence or absence of maltotriose, confirming the conformational changes that lead to MalT oligomerisation (see above). Genetically-engineered fragments of MalT corresponding to the domains identified by limited proteolysis are currently being analysed to determine their oligomeric state in solution, their ability to interact with each other, to locate the binding sites of the positive or negative effectors of MalT, and to determine their high resolution structure.

Transcription activation at the MalE promoter (E. Richet)

The maltose regulon promoters whose activation depends on the concerted action of MalT and CRP, proved to be useful models to study the mechanisms underlying synergistic activation in prokaryotes. One of these MalT- and CRP-dependent promoters is malEp whose activation requires formation of a higher-order structure wherein MalT binds co-operatively to two sets of MalT and which is stabilised by CRP bound to three sites located in between. CRP facilitates co-operative binding of MalT to the two sets of MalT sites simply by bending the DNA when bound to its three cognate sites. The function of this higher-order structure is to ensure high occupancy of the proximal MalT binding sites from which MalT activates initiation of transcription, most likely through a direct contact with RNA polymerase. We have recently shown that, besides its architectural role in the stabilisation of the activation complex, CRP participates directly in the activation process by recruiting RNA polymerase via a direct contact with the C-terminal domain of one of RNA polymerase a subunits. This function is carried out by the most distally-bound molecule of CRP. Hence, the malEp promoter provides the example of a novel and complex mechanism for transcriptional synergy in prokaryotes whereby one activator both helps the primary activator to form a productive complex and participates directly in RNA polymerase recruitment.

Personnel de l'unité

Secrétariat de l'unité

Lavenir Armelle

Chercheurs de l'unité

Pugsley Anthony
Richet Evelyne
Possot Odile
Danot Olivier

Stagiaires de l'unité

Bayan Nicolas
Ebel Frank
Francetic Olivera
Robichon Carine
Vignon Guillaume

Autre personnel de l'unité

Vidal-Ingigliardi Dominique
Guilvout Ingrid
Nadeau Nathalie
Reyngoud Maria

Publications de l'unité

En cas de problèmes ou de remarques concernant ce serveur Web, écrire à rescom@pasteur.fr. Cette page a été modifiée pour la dernière fois le Lundi 17 Juillet 2000 à 10 h 3 (Temps Universel) .