Unit: Bacterial Genome Plasticity

Director: MAZEL Didier

We study the mechanisms responsible of the bacterial genome variability, with a special interest for those involved in exogenous gene acquisition - the horizontal gene transfer. Our model system is the integron, a natural genetic engineering system involved in the development and dissemination of antibiotic resistance genes among Gram-negative species. We are also investigating other factors playing a role in the plasticity of the Vibrio species genomes, which are all constituted of two circular chromosomes that could have distinct dynamic characteristics.

A- the integrons.

Thirty years of molecular genetics, and moreover the "genomic era", have shown that the acquisition of exogenous genetic material, the "lateral gene transfers", plays a primary role in bacterial evolution.

Perhaps, the most striking example of the impact of gene exchange has been the development of multi-drug resistance over the 50 years. In this global phenomenon, the contribution of a particular class of mobile elements, the integrons, has been essential among Gram-negative bacteria. We are studying different aspects of this gene capture system: their distribution, their contribution to the adaptive capacity of their host and their recombination processes.

This natural genetic engineering system is composed of two basic elements: a gene coding an integrase of the site-specific tyrosine recombinase family and a primary recombination site, attI. The integrase activity allows the insertion of open reading frames, in the form of a circular cassette, at the recombination site. The tandem integration of cassettes leads to the assembly of multiple-resistance structures and integrons carrying up to 8 cassettes have been characterized. To date, more than 85 different resistance cassettes have been identified, allowing bacteria to evade most presently employed antibacterial agents. All these cassettes are composed of a single gene associated to a recombination site, the attC site (or 59-base element), indispensable for the integrase recognition and recombination with attI.

Five different "classes" of multi-resistance integrons have been described as of today, based on their integrase gene sequence, which are all carried on by mobile DNA elements (transposons and plasmids).

We have discovered an other type of integron, the super-integrons (SI), in the genome of a large number of γ-proteobacterial species. These integrons share several of the multi-resistance integron (MRI) characteristics, but they differ from them in 3 aspects: i) their size (several tens of cassettes), ii) the strong similarity observed among the attC sites of the cassettes belonging to a specific SI and iii) the fact that most SI cassettes, when known, are not linked to the antibiotic resistance..

The structural differences existing between the MRIs and the SIs led us to the proposal that MRIs evolved from SIs through the entrapment of SI integrase genes and their cognate attI sites by mobile DNA elements. The subsequent harvesting of cassettes from various SI sources then led to the establishment of contemporary MRIs, through the application of a constraint, the antibiotics used by man.

This year we focused our work on two objectives:

Establish the structural characteristics of the Vibrio Sis, especially the inventory of the poison-antidote cassettes.

We noticed the presence of multiple cassettes carrying a "post-segregational killing" (PSK or poison-antidote) system related to those found on plasmids or phage, in the different Sis characterized so far. For example, the V. cholerae N16961 SI contains 5 different families of PSK in 9 cassettes spread along the entire cassette array. The function of these cassettes in the SIs is still unclear. These cassettes may play an important role to stabilize these massive arrays of independently mobile genetic units by regulating the levels of random excision or deletion through homologous recombination between repeated cassettes, since a probability exists to lose or shut off expression of the PSK cassettes. Such an event would result in cell death.

We have already demonstrated the functionality of several of the PSK cassettes (the ccdAB cassette from the V. fischeri SI, the phd-doc from V. metschnikovii and V. cholerae…) in E. coli but also in their original strain. We also demonstrated that these cassettes carried a strong promoter allowing the expression (in collaboration with Laurence Van Melderen, ULB, Belgium).

Nature of the recombination partners and of the IntI recombination substrates.

We have undertaken the study of the real nature of the recombination partners in the integrons, and of their recombination mechanism. The IntI integrases mediate recombination between their specific attI site and a second type of recombination site, the attC site carried by a gene cassette. They can also catalyze recombination between two attC sites.

A unique trait of the integron recombination system resides in the structure of the recombination sites. Typical Y-recombinase core recombination sites consist of a pair of highly conserved 9- to 13-bp inverted binding sites separated by a 6- to 8-bp central region. The attI sites differ from this canonical organization, as one of the putative binding sites within the core site is always extremely degenerate and the central region differs greatly between the different attI sites. The structure of the attC site is more complex. It consists of two potential core sites, R''-L'' and L'-R' (respectively called 1L -2L and 2R-1R by Hall and coll.) separated by a central region. Only L'-R' is recombinogenic and if the central region can be highly variable in sequence and size, all structurally different attC sites can potentially form cruciform structures and are efficiently recombined by IntI1. In 1999, Francia and collaborators demonstrated that purified IntI1 bound specifically to the bottom strand (bs) of single-stranded attCaadA1 DNA but not to a double-stranded (ds) attCaadA1 site.

These characteristics led us to propose a new model for the recombination in integrons, which only involved the attC bottom strand folded in a stem-and loop based on its symmetrical structure, and a canonical double-strand (ds) attI site. Recognition and recombination by the IntI integrase of such a structure with a canonical ds-attI site would lead to a Holliday junction (HJ) intermediate which may be resolved by a replication step. We developed an in vivo recombination assay based on a conjugation system, as a mechanism to independently deliver either the top strand (ts) or bs of the different integron recombination sites in single-stranded form into a recipient strain. Regardless of the conjugation system used, we found that recombination rates of the different substrates varied over a wide range. We determined that the rate of recombination following transfer of the attC-bs with a ds-attI site carried on a plasmid in the recipient was 1000 fold higher than the rate after conjugation of the attC-ts. Furthermore, the recombination rate measured after delivery of the attC-bs was found to be identical to that obtained in a classical assay using ds-attC and ds-attI sites carried on plasmids and co-maintained in bacterial cells expressing IntI1. To explain their unique properties, we propose a recombination model for the insertion of integron cassettes at the attI site that involves only the attC-bs strand folded in a stem-and-loop structure and a resolution of the HJ intermediate generated through replication (figure). In collaboration with F.-X. Barre (CNRS, Gif/Yvette), we have been able to show that a similar single-strand recombination mechanism was in charge of the CTX phage genome (which encodes the cholera toxin) integration at the chromosome 1 dif site of V. cholerae. The crystal structure of an integron integrase tetramer bound to an attC site, which has been determined in collaboration with D. Gopaul (Inst. Pasteur), strengthen our model. Indeed it shows that only 2 of the 4 subunits have an active conformation, while the other 2 non-attacking subunits are involved in interactions with two attC extrahelical bases.

B-. Sequencing the V. splendidus MEL 32 genome.

The second project is to investigate other factors involved in genome plasticity of the complex genome of Vibrio species. The Vibrio group includes a large number of pathogenic species whose hosts range from human to aquatic animals. The few species so far characterized have been found to carry two circular chromosomes. Despite an apparent long evolutionary history of this partition in two molecules, the comparison of chromosome content in the related cholerae and parahaemolyticus species indicates a high variability. The selective advantage conferred by such an organization is unknown. This also raises a number of fundamental questions, in particular for chromosome segregation, and offers a unique opportunity to experimentally manipulate chromosomes.

To increase our knowledge, we initiated in collaboration with C. Bouchier (sequencing platform 1, Institut Pasteur) the genome sequencing of a Vibrio species, which is only remotely related to the Vibrio species sequenced so far. Our strain, V. splendidus MEL 32 is an oyster pathogen, which has been responsible for high mortality rates in oyster beds in France since 1991.

V. splendidus is only remotely related to V. cholerae and V. vulnificus, and belongs to the other major branch in the Vibrio radiation, which also contains V. fischeri and V. salmonicida (a fish pathogen) It is then very likely that the data collected from this genome project will allow to better understand the evolutionary constraints exerted on the Vibrio genomes as a whole and will give new clues to understand the two chromosomes partition in this bacterial group, as well as understanding the specific pathogenic properties of this oyster pathogen.

We have now achieved the complete genome sequence, which is composed of two chromosomes of size 1676 and 3299 kb, respectively.

The comparative analysis with the other sequenced Vibrio genome is now in process, we expect better understanding of the rules governing the overall organization and the gene partition between the two chromosomes in Vibrio. We have also undertaken the identification by several approaches (in silico, substractive hybridization,…) of the genes which could play a role in the oyster disease development.


Recombination model for the attC bottom strand X attI double strand (from Bouvier et al 2005. EMBO Journal 24:4356-67).


Keywords: integron, site-specific recombination, horizontal gene transfer, evolution, Vibrio, cholerae

Activity Reports 2005 - Institut Pasteur

Page Top research Institut Pasteur homepage

If you have problems with this Web page, please write to rescom@pasteur.fr