Dictyostelium, a Tractable Model Host Organism for Legionella

Dictyostelium, a Tractable Model Host Organism for Legionella

Heike Bruhn and Michael Steinert

Abstract

The haploid amoeba Dictyostelium has proven to be a suitable model system for studying cellular aspects of Legionnaires' disease. In this review we describe both Legionella and Dictyostelium factors and processes that are relevant to infection. Moreover, we summarize the results of the Dictyostelium transcriptional host cell response to infection and discuss how genetic and molecular biology techniques available for Dictyostelium can further improve our knowledge of Legionella-host cross-talk. read more ...

Introduction

Infection models in Legionella research

Susceptibility to a certain infection is an innate property of a given host species and is governed by a number of different factors such as anatomy, physiology, variations in tissue or cell surface receptors, effector molecules of the innate and adaptive immune system and others. To gain insight into specific aspects of the host-pathogen interaction, the selection of an appropriate model system with an appropriate level of complexity appears to be essential. Frequently studied host models for Legionella research are guinea pigs, inbred mice, mammalian cell culture systems and various protozoa species. Each system provides specific advantages for the analysis of Legionella infection. Guinea pigs are highly susceptible to infection with Legionella and have been widely used since they develop pulmonary disease symptomatically and histologically similar to that observed in humans (Edelstein et al., 1999; Wagner et al., 2007). Moreover, they allow reliable predictions of drug efficacies. Mice are genetically well characterized; however, they are relatively resistant to Legionella infections. Only intratracheally inoculated A/J mice seem to be a suitable model of a replicative L. pneumophila lung infection (Rossier et al., 2004). Accordingly, only peritoneal and bone marrow-derived macrophages from A/J mice are permissive for Legionella growth, whereas macrophages from other inbred strains are not (Chapter 5).

In addition to animal models, primary cell culture systems and permanent cell lines have provided important information on virulence properties of Legionella and corresponding host targets. In this regard, human monocytes, human alveolar macrophages, and the macrophages of guinea pigs, hamsters and rats have successfully been studied (Steinert et al., 2002). At least equally important are infection studies performed with host protozoa. Ubiquitously present in aquatic habitats, Legionella naturally exploits numerous species of protozoa, which provide nutrients, protect the bacteria from adverse conditions and serve as a vehicle for the colonization of new habitats.

Each of the aforementioned infection models suffers from one of three drawbacks: The animal models are complex; therefore, their analysis and manipulation is time-consuming and difficult. Cell lines lack the cross-talk and other responses typical of multicellular organisms. The single celled protozoa are only poorly investigated and hence insufficiently understood on the molecular level. Moreover, they are not genetically tractable.

In the last years, model organisms like Caenorhabditis elegans, Drosophila melanogaster and Dictyostelium discoideum became increasingly attractive for pathogenicity studies because of their ease to cultivate, the short generation times and the availability of cell biological and genetic tools. Particularly, the haploid social amoeba D. discoideum appears to be an ideal model system for Legionella infection due to its life style: It is an obligate phagocyte that feeds on bacteria - just as the natural protozoan hosts of the pathogen. This phenotypic relation is corroborated by the narrow phylogenetic distance between Dictyostelium and Acanthamoeba, which predicts a similar molecular repertoire. read more ...

Model organism Dictyostelium discoideum

The intriguing biology of D. discoideum has fascinated cell biologists for decades, making it a popular model for investigating molecular cell biology. The haploid cells feed on soil bacteria and replicate by binary fission. Under conditions of starvation, approximately 10⁵ cells aggregate by cAMP-relayed chemotaxis and differentiate in a complex progression into a fruiting body, which is built by spore cells and dead stalk cells (Fig. 12.1). Many fundamental processes underlying cellular procedures like cytokinesis, motility, chemotaxis and signal transduction observed in metazoans appear to have emerged in primordial precursor cells early in evolution. Therefore, these complex processes can be studied in unicellular organisms like D. discoideum. The life cycle of D. discoideum also encourages the investigation of social behaviours like cell differentiation and altruistic cell death, processes that rely on cell-cell communication and chemotactic cell sorting. Hence, it is not surprising that infection biologists have also been attracted to this organism as an experimental model (Steinert and Heuner, 2005).

Dictyostelium shares many molecular features with macrophages, the human host of Legionella. The cytoskeletal composition of D. discoideum is similar to that of mammalian cells as are the processes driven by these components, such as phagocytosis, membrane trafficking, endocytic transit and vesicle sorting. Like leukocytes, D. discoideum possess chemotactic capacity. Hence, D. discoideum represents a suitable model system to ascertain the influence of a variety of host cell factors during Legionella infections. Furthermore, the abundance of well-characterized tools to observe the cell biology of this protozoan provides the opportunity to analyse the pathogen's mechanisms of infection by investigating the fate of various Legionella mutants defective in putative pathogenicity factors. read more ...

Dictyostelium-Legionella interactions at the cellular level

Most importantly, Legionella is capable of infecting D. discoideum and using this host for reproduction (Hägele et al., 2000; Solomon et al., 2000). The uptake and replication of the bacteria in D. discoideum resemble that which is known from amoebae and macrophages (Steinert et al., 2007). The pathogen itself triggers the phagocytosis and replicates within membrane-bound vesicles that prevent fusion with lysosomes (Hägele et al., 2000; Solomon et al., 2000). The endoplasmic reticulum (ER) contributes to the Legionella-containing vacuoles (LCV) in the natural protozoan hosts A. castellanii and Hartmanella (Abu Kwaik, 1996; Bozue and Johnson, 1996) and in macrophages (Swanson and Isberg, 1995; Kagan and Roy, 2002; Robinson and Roy, 2006) as well as in D. discoideum (Fajardo et al., 2004; Lu and Clarke, 2005). Furthermore, the growth-rates of wild-type and various Legionella mutants are comparable in the different hosts, demonstrating essentially similar replication mechanisms (Heuner et al., 2002; Schreiner et al., 2002; Skriwan et al., 2002). Taken together, these prerequisites are essential and sufficient to justify the use of the non-natural host D. discoideum as an infection model.

Legionella factors and function

The bacterial factors provoking pathogenicity can mainly be grouped into two categories: Regulatory factors, which control the transcriptional pattern in response to the environment, and effector molecules, which mediate interactions with the host cell. The regulatory factors comprise the alternative sigma factors RpoS (Hales and Shuman, 1999) and FliA (Heuner et al., 2002) as well as the two-component response regulator LetA (GacA) (Hammer et al., 2002; Gal-Mor and Segal, 2003; Lynch et al., 2003) which regulate the expression of proteins characteristic for the transmissive state (Molofsky and Swanson, 2004). The most prominent effector molecules are the type 4 secretion system (T4SS) Dot/Icm (Segal et al., 1998; Vogel et al., 1998) and several substrates secreted by it. These transferred proteins are mostly still of unknown function, but concertedly, they prevent the fusion of the Legionella-containing vacuole with the endosomal compartment (Chapter 10). Two additional secretion systems contributing to pathogenicity are the T2SS Lsp (Hales and Shuman, 1999) and a twin-arginine translocation secretion system (De Buck et al., 2005; Chapter 8). A protein identified as macrophage infection potentiator (Mip) (Cianciotto et al., 1989) is also involved in the infection of protozoan hosts (Cianciotto and Fields, 1992). It belongs to the family of FK506-binding peptidyl-prolyl-cis/trans-isomerases (Fischer et al., 1992) and contributes to the dissemination of the bacteria in the tissue in concert with a yet unknown serine proteinase (Wagner et al., 2007). Finally, type IV pili promote the adherence of the bacteria to cells (Stone and Abu Kwaik, 1998). Since all these factors are described in comprehensive details in particular chapters of this book, we here will concentrate on their impact on specific interactions with D. discoideum.

Legionella Dot/Icm mutants replicate less efficiently than wild-type bacteria in D. discoideum and permit Dictyostelium growth on bacterial lawns (Solomon et al., 2000), incontrovertible evidence that the pathogen requires the secretion system to reproduce. The effect of only few of the known substrates of Dot/Icm has been investigated in D. discoideum thus far. Of these, SidC is the only one which provokes a specific effect in the slime mould. SidC accumulates on the LCVs by binding to phosphatidylinositol(4) phosphate (PI4-P) in a highly specific manner (Weber et al., 2006). This metabolite of the phosphoinositol pathway is generated by the action of phosphatidylinositol(3) phosphate-kinase, an enzyme involved in several signal transduction pathways. Inhibitory studies revealed that the enzyme is also required for replication of Legionella in Dictyostelium (Weber et al., 2006). Its activity might enrich the LCV membrane in PI4-P, which then functions as an anchor for transmitted bacterial proteins like SidC.

The only other Legionella protein investigated that influences the intracellular growth in D. discoideum is the alternative sigma factor FliA, which controls the flagellar regulon. Deletion mutants are aflagellate and are incapable of replication in D. discoideum (Heuner et al., 2002). Although the analysis of Legionella mutants in D. discoideum is a straightforward and powerful approach, the effect of some discrete host factors is not ready at hand due to the lack of homologues in the slime mould. Prominent examples are NF-kappa B, an observation that indicates some metazoan pathways of cell responses do not exist in D. discoideum. A second example is caspases, enzymes that are involved in programmed cell death (Golstein et al., 2003). The distinct mechanisms that orchestrate cell death in macrophages and D. discoideum might account for the observed cell-type specificity of the Dot/Icm substrate SdhA in the two hosts. SdhA apparently prevents host cell death in macrophages, thereby ensuring Legionella replication (Laguna et al., 2006). The sdhA deletion mutants are severely impaired in macrophages but not in D. discoideum. For other pathogenicity factors, the effect of their deletion is more pronounced in one host than the other. For example, the mutant of the Dot/Icm substrate VipD, which possesses a phospholipase A domain and is suggested to assist in the establishment of the LCV, replicates normally in macrophages but exhibits an increased growth rate in D. discoideum (VanRheenen et al., 2006). These results support the idea that the biochemical activity of these proteins is identical in the different hosts, but they may have evolved to operate more efficiently in specific cell types.

Dictyostelium factors and functions

The internalization of Legionella is actin-mediated and involves several cytoskeleton-associated proteins like myosin I (Solomon et al., 2000), coronin (Solomon et al., 2000; Fajardo et al., 2004), comitin (Schreiner et al., 2002; Skriwan et al., 2002), the actin cross-linker alpha-actinin/ABP120, the F-actin fragmenting and capping protein DAip1, the cytoskeleton-membrane interface proteins LimA-D and the F-actin associated protein villidin (Fajardo et al., 2004). Knock-out mutants of all these proteins show reduced uptake and/or increased intracellular growth of the bacteria. Interestingly, the only actin-binding protein which reduces bacterial internalization is profilin, a protein that sequesters G-actin (Hägele et al., 2000). Knock-out mutants of profilin exhibit a generally higher rate of phagocytosis and thus a higher rate of bacterial uptake whereas phosphatidylinositol(3) kinases are apparently not required for phagocytosis of Legionella (Weber et al., 2006), inhibitory studies revealed that the phospholipase C (PLC) pathway does contribute (Fajardo et al., 2004). PLC itself is activated by heterotrimeric G-proteins, and knock-out mutants of the Gβ subunit are impaired in uptake (Solomon et al., 2000; Fajardo et al., 2004).

Phagocytosis and replication of the bacteria are also modulated by cytosolic calcium. The influence of two ER-localized calcium sequestering proteins has been examined, namely the luminal calreticulin and the transmembrane calnexin (Fajardo et al., 2004; Lu and Clarke, 2005). According to Fajardo et al, both proteins extend into the phagocytic cup and accumulate permanently on the LCV, consistent with a contribution of these proteins to uptake and vacuole formation (Fajardo et al., 2004). By contrast, Lu and Clarke described the LCV captured by the ER with a concurrent accumulation of calnexin about one hour into infection whereas the LCV maintains its distinct identity (Lu and Clarke, 2005). The bacterial vacuoles are transported in the cell along microtubules (Lu and Clarke, 2005). Vesicle trafficking is a pivotal process for Legionella reproduction. Mutants in RtoA, a protein mediating vesicle trafficking, showed severely depressed intracellular growth of Legionella (Li et al., 2005). The Legionella vacuoles avoid lysosomal fusion, as demonstrated by the exclusion of lysosomal marker proteins such as DdLIMP (Hägele et al., 2000). In spite of the morphological similarity of the vacuoles to autophagosomes, macroautophagy is not required for bacterial replication in Dictyostelium (Otto et al., 2004). In mouse macrophages, a metal ion transporter protein called NRAMP1 (natural resistance associated membrane protein) decreases the susceptibility to intracellular infections. Likewise, Dictyostelium mutants that lack the homologous protein exhibit enhanced replication of Legionella; the phenotype can be rescued by NRAMP1 over-expression. The protein is located on macropinosomes, phagosomes and exocytic vesicles. In the presence of ATP, the protein acts as a proton/iron ion symporter thereby depleting the vacuole of iron, which is pivotal for intracellular pathogens. Since the protein itself does not bind ATP, the activity suggests a concerted reaction of the vesicular H⁺-ATPase and NRAMP1 (Lu and Clarke, 2005; Peracino et al., 2006).

The genome sequencing of Dictyostelium enabled the application of comprehensive strategies. The transcriptional response of the amoeba upon infection with wild-type or avirulent, DotA-deficient L. pneumophila and the less virulent species L. hackeliae was analysed using microarrays comprising approximately half of the Dictyostelium genes (Farbrother et al., 2006). Most of the transcriptional changes occur 24 h post infection when 131 genes are differentially regulated in response to all three Legionella species. Genes encoding proteins that promote degradation like hydrolases, lysozymes, lipid modifying enzymes and several calcium binding proteins were down-regulated, as were genes important for the biosynthesis of proteins, whereas genes that contribute to tRNA metabolism, nucleotide biosynthesis and the cytoskeleton are apparently up-regulated, presumably to ensure the pathogen's supply of nutrients. Consistent with conclusions from mutational studies, expression of a number of amoebal genes encoding proteins primarily involved in vesicle-trafficking processes like RtoA, ARF1 and β'COP were also induced after infection. read more ...

The Dictyostelium tool box opens new perspectives

Due to the interest of cell biologists in particular, multifarious tools have been developed for genetic manipulation and generation of mutants in D. discoideum (Table 12.1). Random mutagenesis by insertion of extrachromosomal vectors into the genome can be used to create large libraries of mutant strains by applying an elegant method called restriction enzyme mediated integration (REMI) (Kuspa and Loomis, 1992; Kuspa, 2006). The cells are electroporated with a linearized plasmid in combination with a frequently cutting restriction enzyme that produces the same overhanging ends that the plasmid offers. The enzyme mediates the insertion of the plasmid into the genome at the corresponding restriction sites. Southern Blot analysis performed with a second enzyme that does not cut inside the plasmid reveals the genomic region of integration. The fragment is isolated, recircularized and amplified in E. coli for sequencing. Phenotypic analysis of such libraries of mutants identified a large number of genes involved in cytokinesis, motility, aggregation and later stages of development. Moreover, directed mutagenesis of non-essential genes can be easily achieved by homologues recombination, which is also used to verify the phenotype of random mutants obtained by REMI. The constructions of double and triple knock-out mutants is also facilitated by parasexual genetics, which refers to the formation of diploid cells by fusion of pairs of haploid cells and the subsequent spontaneous haploidization by random loss of chromosomes, (King and Insall, 2003; King and Insall, 2006). All these mutants are listed in the 'Dicty Stock Center', the central repository for D. discoideum strains and plasmid. Currently, there are 1020 different strains, including natural isolates, random mutants, null mutants and GFP-labelled strains, and 290 plasmids are also recorded. Waiting to be analysed for their impact in the cell biology of the amoeba is the whole catalogue of mutant D. discoideum strains defective in cytoskeleton proteins, vesicle sorting, or signal transduction - exactly those processes that are of great importance to uptake and replication of Legionella.

Mainly three independent paths are inviting the interest of scientists keen to characterize the host side of the interaction with bacteria. The first one comprises the cellular responses to the invasion. In this respect, a great boon was the publication in 2005 of the genome sequence of D. discoideum (Eichinger et al., 2005). The six chromosomes of D. discoideum comprise 33.8 Mb DNA coding for approximately 13.500 genes. The genome has a higher content of simple sequence repeats and homopolymeric traits like polyglutamine in coding regions than that of any other known organism. A large number of the genes is likely to be common in metazoans and D. discoideum but to be absent in S. cerevisiae. The genome is well-annotated by now and has enabled the application of comprehensive strategies. For example, the proteome of subcellular structures and alterations thereof have become amenable, along with new approaches to analyse protein-protein interactions. Knowledge of the complete set of predicted genes also allows the use of microarrays to analyse transcriptional changes during infection. Additionally, on the basis of the sequences, the proteome of subcellular structures and alterations thereof become amenable and even points the way to protein-protein interactions.

The second major track of investigation is the visualization of the dynamic cell behaviour, organelles or even single host cell proteins in vivo by time-lapse imaging. Sophisticated approaches using 2D and 3D Dynamic Image Analysis Systems (Soll et al., 2000; Wessels et al., 2006) or quantitative fluorescence microscopy and fluorescence resonance energy transfer (Sekar and Periasamy, 2003; Xu et al., 2005) were designed to analyse the chemotactic response of the cells initiating differentiation. The application of the green fluorescent protein and variants thereof as non-invasive tags facilitates the tracking of host cell factors (Knetsch et al., 2002; Müller-Taubenberger, 2006). Null mutants rescued by GFP fusions ensure the functionality of the fusion. Finally, GFP fusions can be used to investigate the behaviour of cells under various conditions.

Lastly, exploiting the high frequency of homologous recombination, custom-tailored knock-out mutants of D. discoideum and the applicability of RNA interference (Martens et al., 2002) allow the characterization of the impact of single proteins in the replication pathway. read more ...

Concluding remarks

Owing to remarkable similarities in cell structure and behaviour to the natural hosts, D. discoideum has been proven a powerful model for Legionella infection. The analysis of bacterial virulence in this model reflects many mechanisms that are relevant for the infection of macrophages. Particularly the applicability of genetic manipulation of both the pathogen and the host enables the detailed characterization of their interactions by the combined analysis of attenuated bacteria mutants and the corresponding host mutants. Moreover, high-throughput approaches like functional and structural genomics and proteomics of pathogen and host will extend the knowledge of their interrelations. Additionally, in the search for yet unknown functions of putative pathogenicity factors, like the Dot/Icm substrates, the abundance of cell biological tools and the ease of handling make D. discoideum an optimal experimental system. Since the idea to exploit the social amoeba as an infection model and the appreciation of its versatility emerged only a few years ago, the number of studies exploiting this system will undoubtedly increase dramatically in the near future.

from Legionella: Molecular Microbiology

Biology Blogs

Conferences

Journals

Open Access Biology Home | PCR | Microbiology Books | Molecular Biology Books | Molecular Biology Gateway | PCR Protocols | Real-Time PCR |