WO2002048709A1 - Membrane signalling networks - Google Patents

Abstract

The invention relates to an integrated system for the identification of the components of membrane signalling networks in both normal and pathogen-infected cells. The invention utilises preparation of an organelle termed the phagosome as an integrated system to allow the identification and manipulation of protein and lipid components of signalling pathways. The invention also relates to methods for the treatment and prevention of pathogen-mediated disease by enhancing the efficency of cellular defences against pathogen infection.

Description

MEMBRANE SIGNALLING NETWORKS

The invention relates to an integrated system for the identification of the components of membrane signalling networks in both normal and pathogen -infected cells. The invention utilises preparations of an organelle termed the phagosome as an integrated system to allow the identification and manipulation of protein and lipid components of signalling pathways.

All publications, patents and patent applications cited herein are incorporated in full by reference.

Eukaryotic cells are thought to possess up to 100,000 gene products. About 10,000 of these proteins are thought to be expressed and operate simultaneously in any one living cell. Many of these proteins are present in membranes, including the plasma membrane, in addition to an estimated 500 different lipid types. The biological role of membranes is crucial in numerous cell processes and in signalling networks that can link the cell to other cells and to the physiological environment outside the cell membrane via transmembrane signalling to the cell interior. However, membranes, and especially their hydrophobic interiors and cytoplasmic domains, are highly inaccessible for many experimental approaches.

On the plasma membrane alone, a vast cascade of molecules and lipids are now known to be operating simultaneously and their activities can be regulated by a multitude of other proteins and lipids, as well as by physical changes that occur within the bilayer (for example, effected by ion channels or activation of stretch receptors). However, our current understanding of these networks is fragmentary. Moreover, little is known of signalling networks that must undoubtedly exist in complex intracellular membrane compartments such as the endoplasmic reticulum (ER), Golgi, endosomes, lysosomes and mitochondria present inside each cell. The importance of membrane signalling in cellular function is widely appreciated. These networks form an important component of the global cellular network and are crucial for a multitude of processes ranging from cell motility to metastasis. Importantly for the understanding of disease, all these processes may be perturbed when pathogens interact with cellular membranes. Furthermore, pathogens utilise varied mechanisms to manipulate these processes.

Our molecular understanding of processes such as membrane fusion, actin and microtubule based organelle motility is fairly advanced as a result of the use of many different in vitro assays by the cell biological community. However, the intracellular membrane signalling networks underlying these, and other processes are very poorly understood. Furthermore, it has proved extremely difficult to follow any particular branch of a signalling network in one cell in any detail, in a holistic, integrated fashion. In particular, this is a consequence of the fact that the vast majority of published data on membrane signalling networks are the result of whole cell analyses. Any results generated from experiments performed in such systems are thus an amalgam resulting from the blending of multiple signalling pathways from a number of different organelles. Furthermore, often these systems use cell cultures in which the composite cells are not even in the same physiological state. These problems provide an insurmountable barrier to a holistic understanding of how membrane signalling is orchestrated in living cells and how these networks are linked to key processes such as actin nucleation, microtubule binding and motility, and membrane fusion.

Accordingly, there is presently an unfulfilled need for a system that allows the analysis and manipulation of membrane signalling networks in isolated membrane systems.

Summary of the invention

According to a first aspect of the invention, there is provided a method for the analysis of a membrane signalling network, the method comprising assaying the activity of a component of the membrane signalling network in a phagosome composition under first reaction conditions, and comparing the activity of said component with its activity in a phagosome composition under second reaction conditions.

The invention provides a simple but defined eukaryotic membrane system (the phagosome) as an adaptable model system to analyse, and thus identify and integrate membrane signalling networks. The system allows the integrated analysis of complex intracellular functions such as membrane fusion, actin nucleation and binding, microtubule binding and motility, and the oxidative burst phenomenon (including production of reactive oxygen intermediates (ROI) and reactive nitrogen intermediates (RNI)). The activity of many of the lipids and proteins that are present in this system can be regulated by the manipulation of the levels of simple metabolic components. The broad scope of the in vitro assays described below also allows links to be made in the signalling pathways that govern the interactions of membranes with other membranes (fusion), with microtubules or with actin, or with pathogens. Furthermore, by combining the assays with cytoplasmic extracts it is also possible to make links between membrane signalling networks and those operating in the cytoplasm.

Phagosomes are membrane organelles enclosing particles ranging from inert material, such as asbestos, carbon particles from cigarette smoking and latex beads (latex bead phagosomes), to bacterial or eukaryotic pathogens (pathogen phagosomes). The use of lμm latex beads is now a well-established procedure for preparing relatively pure phagosome fractions, via gradient flotation. Such preparations of phagosomes thus represent a well- characterised model system. Latex bead phagosomes (LBPs) may be prepared by any suitable method, such as by that described by Desjardins et al. (1994, J. Biol. Chem. 124: 677-688) and Jahraus et al. (1998, J. Biol. Chem.; 273(46): 30379-90). Pathogen phagosomes may be prepared according to the method of Alvarez-Dominguez and Stahl, (1998, J. Biol. Chem.; 273(51): 33901-4); Sturgill-Koszycki et al. (Electrophoresis, 1997, 18, 2558-2565); Mills and Finlay (1998, Eur. J. Cell. Biol. 77, 35-47).

The invention allows the analysis of any membrane signalling network that may be linked to a particular biological function. By following the biological function under different conditions, an observable phenotype is provided that gives an indication of the activity of the membrane signalling network that is implicated in the particular biological function tested. By assaying for the presence of a component of the membrane signalling network, or by assaying for its activity, and correlating this with the biological function, an indication may be obtained as to the relevance of the tested component in the signalling network. For example, if the activity or level of the tested component correlates closely with variations in the biological function, then this component may be considered as a functional node in the membrane signalling network concerned. Alternatively, the level or activity of the tested component may be found not to be linked with the biological function in any way. In one example of the system of the invention, an effector molecule that is specific for a component of a signalling pathway may be added to a phagosome system and its effect on the biological function monitored. If the biological function changes, then the target of the effector may be considered a functional node in the membrane signalling network concerned. Any biological function that may be associated with the phagosome system, and that can be assayed may be selected for use as a marker for the activity of a membrane signalling network according to the invention. However, the biological function should advantageously be one that is susceptible of assaying in vitro. Examples of such biological functions include actin nucleation, and F actin binding; the identity of certain of the components that are active in these systems is well known in the literature. However, it should be noted that the invention is not in any way limited to assaying these particular biological functions. Any biological function that is linked to membrane signalling, and which may be assayed is suitable for use with the method of the invention. When using an in vitro assay, complementary in vivo methods can be applied where these are necessary. Moreover, in vivo methods can be combined with in vitro methods by adding compounds to cells and subsequently isolating the phagosomes of those cells and testing them in vitro.

There is already a wealth of information in the published literature that relates to the identity of components that are implicated in a particular biological function. Therefore, when analysing a membrane signalling network according to the method of the invention, it may initially be instructive to focus on protein or lipid components are that implicated at early (upstream) points in the signalling hierarchy, that are expected to bind directly, or via a single intermediate, to receptors in the phagosome membrane. Examples of such components will be clear to the skilled reader and include proteins such as tyrosine kinases, trimeric G proteins, Rho proteins, PI3-kinase, PI4-kinases and MAP-kinases. By selecting effector molecules that are known to regulate the levels or activities of these components in the selected assay and comparing the phagosome system in the presence and absence of these effector molecules, the degree to which the targeted component is implicated in the membrane signalling network may be elucidated. In this way, any perturbations in the biological function that are generated in the course of the analysis may be linked to changes in the concentration or activity of that particular component of the signalling network. For example, the effector may be scored as a stimulation (+), an inhibition (-) or as having no effect (O), relative to standard, or to control conditions. As soon as an effector molecule is found to stimulate or inhibit a candidate component of a membrane signalling pathway, this molecule is then considered as a functional node in the network. Components that are known to interact in some way with the node that is identified as functional may then be tested systematically and a functional map may gradually be built up in this manner. An example of such as system is shown in Figure 1 , in which the signalling component PI₄₅P₂ is shown included as a functional node and connected to a large network based on known biochemical pathways.

In parallel to the functional analysis summarised above, protein or lipid phosphorylation analyses (or total phospholipid analyses) may be undertaken to test the effect of the effectors of the identified functional nodes on the levels or activities of other proteins and/or lipids that are present in the phagosomal membrane. This may even be undertaken at the single phagosome level, using approaches such as fluorescence resonance energy transfer (FRET) (reviewed by Wouters et al., Trends Cell Biol., 2001, ϋ, 203-211) Furthermore, a bioinformatic approach such as that described by Schuster et al (2000 Nature Biotech. 18(3): 326-332) may be applied to the system of the invention, in order to describe mathematically a model of the flux of reactions through the network. Such an approach allows the analysis of the flow of information through complex multi-component networks of biochemical interactions. Using this approach, a model will initially be established that is based on known biochemical reactions which will lead to experimental predictions that may be tested in the phagosome system of the invention. Such experiments will then lead to a refinement of the proposed model, so leading to the design of new experiments. By this iterative process, the analytical tool that the present invention provides will ultimately provide a mathematical description of membrane signalling networks and their dynamics.

Such an approach will be greatly facilitated by ongoing proteomic (Garin et al, 2001, J. Cell. Biol. 152. 165-180) and lipid analysis by thin layer chromatography (TLC), high pressure liquid chromatography (HPLC) and mass spectrometry. Once a detailed map has been built up for a given control situation (for example, 2 hours old, phagosomes enclosing beads conjugated with an inert protein such as fish-skin gelatin, hereafter referred to as 'naked latex bead phagosomes') the process may be repeated for any effector to be tested.

For example, immunoglobulin G- (IgG) coated beads can be used to enrich for Fc receptors on latex bead phagosomes (LBPs). The same effectors against nodes that were found to regulate the phagosome functions in control phagosomes (naked LBPs) can be applied to the IgG-coated bead phagosomes and the two systems can be compared as described above, in the pattern (+; -; O). Even a single difference may be exploited by evaluating the difference in the context of the whole network (for example, naked LBPs might stimulate actin in the presence of a Protein Kinase C inhibitor, such as staurosporine, whereas IgG beads might be unaffected). The bioinformatic analysis will lead to predictions about how the network will behave for the effector-treated membrane relative to the control membrane. The method

According to the invention, phagosomes may be prepared from a variety of different cell lines. Preferably, mononuclear phagocyte cells such as macrophage or neutrophil cells are used. These cells are the most prolific in their ability to phagocytose foreign particulate and colloidal material.

It should be noted that there are two types of phagocytic cells. First are the 'professional' phagocytes, that predominantly include macrophages and neutrophils. Second, there are non-professional phagocytes that are usually selective for live pathogens. For example, many pathogenic bacteria can be actively phagocytosed by cell types that are normally not phagocytic. For example, Shigella can induce phagocytic uptake into epithelial cells while Afipiafelis (the agent of cat scratch disease) can induce its uptake into endothelial cells that are not normally phagocytic. Such cells, for example, would not take up naked latex beads, although they may enter those cells if the appropriate effectors could be bound to the beads.

Already, a large proportion of the proteins that are present in this organelle as prepared from J774 mouse macrophages have been identified by proteomics projects carried out by collaborators of the Applicant for the present invention. A complete documentation of the phagosome proteins is predicted to be available within the near future. In parallel, a detailed lipid analysis is in progress; already close to twenty different species have been identified.

The method of the invention can be used in conjunction with certain biochemical assays for a number of purposes. For example, the method may be used to describe and identify the components of holistic membrane signalling networks, to analyse and define disease- specific pathways, to link signalling events induced by a pathogen to genetic differences in the host cells and to identify novel drug molecules that are capable of altering signalling networks. In its broadest sense, the invention therefore provides a method for the analysis of a membrane signalling network that involves the comparison of one or more phagosome compositions under differing reaction conditions. The term "reaction conditions" is intended to be construed broadly, and embraces a number of different possibilities, as will be described in further detail below. Reaction conditions

In one sense, the term "reaction conditions" may refer to the actual concentration of components present in the phagosomes. These components may be components necessary for the metabolic function of the signalling system in question, such as lipids (for example, sphingomyelin, ceramide, sphingosine, sphingosine-1 -phosphate (SIP), diacylglycerol, arachidonic acid (AA), phosphatidic acid, phosphatidyl choline, phosphatidyl inositol-4- phosphate (PLP;) and phosphatidyl inositide 4,5 Bis-orthophosphate (PI₄₅P₂)); proteins and protein families (such as protein kinase C (PKC), GTPases, rho family proteins (e.g. racl), phospholipase D (PLD), ezrin, moesin, profilin, gelsolin; co-factors; inhibitors; simple chemical compounds (ATP, GTP, ADP, GDP, IP₃, calcium ions); buffer solution and so on. By comparing the activity or level of a particular signalling component or components under different reaction conditions, the function of the signalling component(s) may be elucidated and/or analysed with respect to a given phagosome function, such as actin nucleation.

In another meaning of the term, "reaction conditions" may refer to the molecular constituents of the phagosomes in the phagosome preparation. Accordingly, this aspect of the invention provides a method for analysing a membrane signalling network, comprising the steps of: a) preparing i) a first phagosome preparation in which substantially all of the constituent phagosomes contain one or more components of said membrane signalling network, or contain the activities of one or more components of said membrane signalling network; and ii) a second phagosome preparation in which the components or their activities in the constituent phagosomes differ from the components or the activities in the phagosomes in the first phagosome composition; and b) comparing the membrane signalling network in the phagosomes of said first phagosome preparation with the membrane signalling network in the phagosomes of said second phagosome preparation.

In one embodiment of this aspect of the invention, the phagosomes may contain particles within them, and may thus be used as a simplified cell system. For example, latex bead phagosomes (Jahraus et al., 1998) or other particle-containing phagosomes (such as erythrocytes, apoptotic cells, silica particles or yeast phagosomes) may be used. Through covalent coupling of purified ligands to the surface of the particles, the membrane signalling network can be compared between two phagosome preparations. By over- expressing proteins in phagocytic cells, defined receptors may be targeted in large numbers to the plasma membrane of any cultured cell. The use of the ligands of these receptors on beads provides a researcher with a defined receptor-ligand system whose signalling behaviour can be compared with the naked latex beads. LBPs are also uniquely suited for binding studies, since after incubation the bound component can be assayed by floating (i.e. repurifying) the phagosomes. Such experiments are extremely difficult with other semi- purified organelle fractions.

A number of different types of membrane signalling network components may be utilised according to the method of invention. For example, a first phagosome preparation may comprise phagosomes containing beads with a first ligand type coupled thereto, and a second phagosome preparation may comprise phagosomes containing beads to which a second ligand type is coupled. The membrane signalling networks in the two systems can then be compared to assess the relative effects of the two ligand types on the membrane signalling network. Of course, the phagosomes of either or both phagosome preparations may be compared to beads without any ligand type coupled to them ('naked beads').

It will be clear to the skilled reader that a number of different ligand types may be conjugated to beads according to the method described herein, and thus their effect on a membrane signalling system studied. The term "ligand" is meant to include any molecular entity that exhibits binding affinity for a particular binding partner, and thus includes any molecule that is able to bind selectively, preferably selectively and stoichiometrically, to one or more sites on another molecule. Preferably, the ligand exhibits binding affinity for a binding partner exposed on the surface of phagocytic cells.

Examples of suitable ligands that are useful for coupling to particles in phagosomes include small chemical compounds such as hyalaruonic acid (which is a surface component of some species of streptococcus), small drug molecules, receptors (such as the CD44 receptor that binds to hyalaruonic acid), the extracellular domains of such receptors, enzymes, antibodies and antibody fragments, structural proteins (such as members of the ezrin-radixin-moesin (ERM) protein family), growth factors, glycoproteins, bioactive peptides and oligopeptides, bacterial or viral virulence factors (such as the Shigella surface protein IpAB which also binds to CD44), bacterial or viral structural proteins, lipid molecules and signalling molecules. Ligands for macrophage or neutrophil receptors are particularly preferred. Further examples of suitable ligands will be clear to those of skill in the art. The composition of the phagosomes themselves may be manipulated to express a binding partner for a ligand, such as a receptor, or a lipid molecule, preferably by incorporating a lipid molecule into the membrane of cells or of isolated phagosomes, in order that the effect of altering membrane composition may be compared. This may be necessary to ensure that the biological effect of the ligand is manifested in the membrane signalling network in the phagosome. For example, the phagosomes of both first and second phagosome preparations may contain beads to which the same ligand type is coupled, but may contain a different membrane composition. One phagosome type might express in its membrane the wild type receptor selective for the ligand, and a second phagosome type may express mutant receptor or no receptor whatsoever. In this way, membrane signalling networks may be compared between phagosome preparations in which the constituent phagosomes contain different ligand types. Since the physiological state of the phagosomes containing latex beads and pathogens changes significantly as these organelles mature (age) in the cell, it may also to be of interest to test this parameter with respect to the signalling networks that regulate any defined phagosome function.

Another possibility is to compare phagosomes from untreated cells with phagosomes from cells that have been treated with an effector molecule such as, for example, gamma interferon or another cytokine that can 'activate' cells such as macrophages. Activation is currently a poorly understood phenomenon in which many phagosome functions, such as the production of RNI or ROI is enhanced. Any effector that potentially changes the physiological state of a cell and that might have a bearing on the physiological state of the phagosome may be used in such a system. This approach could be used to address how cytokines such as gamma interferon affect phagosomal membrane signalling networks.

In a related system, membrane signalling networks may be compared between phagosomes that possess a different genetic composition. For example, latex bead phagosomes, with purified ligands coupled to the beads, may be prepared from different cell types or even from different species types (for example, mouse and human). In this fashion, the biological effects of different membrane constituents may be compared.

Particles other than synthetic particles may be contained within the phagosomes used in the present invention. For example, so-called "pathogen phagosomes" may be used, in which a phagosome contains a pathogenic organism such as a bacterium or a virus. A variety of such pathogenic organisms invade mammalian cells by the mechanism of phagocytosis. These pathogens are responsible for diseases such as tuberculosis, salmonellosis and leishmaniosis. Cell invasion by these pathogens is made possible only by the orchestrated evasion by the pathogen of the degradative mechanisms of their host cells.

For example, protozoan parasites of the genus Leishmania infect millions of people worldwide and are responsible for a wide spectrum of diseases termed leishmaniases. E. donovani, the causative agent of visceral leishmaniasis (Kala-azar), disseminates and infects macrophages of the liver, the spleen, and the bone marrow. This infection is chronic and may be fatal in untreated cases. So far, efficient prophylactic measures, including safe vaccines, are not available.

Examples of pathogens that are suitable for use in this aspect of the invention (i.e. those that survive and replicate within phagosomes), include Mycobacteria (for example, Mycobacterium tuberculosis), Salmonella (for example, Salmonella typhimurium or S. typhi), enterotoxic E. coli, Campylobacter, Brucella Coxiella, Rhodococcus, Legionella, Nocardia, Chlamydia and eukaryote parasites such as Leishmania (for example, Leishmania donovani), Giardia and Cryptosporidium, Toxoplasma and Histoplasma. The evasion mechanisms used by pathogenic bacteria such as Leishmania are presently understood poorly at the molecular level (see Garcia-del Portillo and Finlay, 1995). One of the strategies that is used by pathogens to invade their host cells successfully is thought to be by altering the fusion of phagosomes with endocytic organelles, a process resulting in the inhibition of phagolysosome biogenesis (Finlay and Falkow, 1997 Microbiol. Mol. Biol. Rev. 61:136-169). This inhibition may be caused by the modulation of phagosome protein composition (Small et al, 1994, Science 263:637-639). It has been shown recently that L. donovani promastigotes reside in phagosomes that display poor fusogenic activity with endocytic organelles compared with phagosomes containing latex beads, which fuse extensively with endosomes and lysosomes (Desjardins et al. 1994; Desjardins and Descoteaux, 1997, J. Exp. Med. 185: 2061-2068; Jahraus et al. 1998; Claus et al. 1998, J Biol Chem 273:9842-9851). However, the nature of the alterations occurring on bmαm -containing phagosomes at the molecular level are mostly unknown at present.

The presence of pathogenic mycobacteria in phagosomes is known to block normal maturation of the phagosomes. This is manifested in a block in phagosome actin nucleation (Bos et α/-to be submitted) and fusion with lysosomes (Russell, Nature Rev. Mol. Cell. Biol. 2001, 12, 569-577), a reduction in RNI and ROI production (Flynn and Chan, Ann. Rev. Immunol., 2001, 19, 93-129) and in acidification (lysosomes (Russell, Nature Rev. Mol. Cell. Biol. 2001, 12, 569-577). It seems highly likely that all these alterations favour the intracellular survival of the pathogen. It is thought that mycobacterial proteins, and known that some of its lipids (Fischer et ah, J. Immunol, 2001, 167. 2187-2192), become incoφorated into the phagosomal membrane. However, the precise mechanisms responsible for the block in maturation are still unclear. Of relevance to the present application is that any treatment that can overcome the pathogen-induced block in phagosome maturation can be expected to favour the killing of the pathogens.

In one embodiment of this aspect of the invention, a method is provided for analysing the mechanism of pathogen action in host cells, comprising assaying the activity of a component of a membrane signalling network under first reaction conditions in a phagosome preparation in which substantially all of the phagosomes contain pathogenic organisms, and comparing the activity of said component with that of the component in a phagosome preparation under second reaction conditions.

As will be clear to the skilled worker from reading the embodiments of the invention that are set out above, there are a number of possible variations involving this method. For example, the membrane signalling networks for a defined phagosome function operating in vitro may be compared between phagosomes containing a certain pathogen type at different concentrations of the chemical components of the signalling pathway. Alternatively, reagents affecting signalling may be compared in vivo followed by isolation of phagosomes and subsequent testing in vitro. Furthermore, phagosome compositions that contain different pathogens may be compared. This aspect of the invention provides a method for analysing the mechanism of pathogen action in host cells, wherein a component of a membrane signalling network is compared between a first phagosome preparation in which substantially all of the phagosomes contain a first pathogenic organism, and a second phagosome preparation in which substantially all of the phagosomes contain a second pathogenic organism. In related embodiments, the same pathogen type may be used in phagosomes of differing genetic make-up, or isolated from cells of different species, for example, human and mouse.

Preferably, this aspect of the invention provides a method for analysing the mechanism of pathogen action in host cells, wherein a component of a membrane signalling network is compared between a first phagosome preparation in which substantially all of the phagosomes contain a first pathogenic organism, and a second phagosome preparation in which substantially all of the phagosomes contain a second non-pathogenic organism, preferably a genetically-related non-pathogenic organism. In this manner, signalling events induced by a pathogen may be linked to genetic differences in host cells as well as to other factors that can influence the process of phagocyte activation, such as cytokines. This process is well known to activate signalling cascades that lead to increased pathogen killing. The recent isolation of LBP from Dictyostelium disccoidum offers another useful variation on this system since the powerful genetic approaches, and the many mutants that are available with this organism can be used to test the role of specific proteins (see Cornillon at al 2000, J. Biol. Chem 275, 34297- 34292).

As one example of an important pathogen that may be addressed using the system of the invention is Mycobacterium tuberculosis (MT). Humans are highly susceptible to infection by this pathogen, and indeed, current estimates by the National Institute of Allergy and Infectious Diseases (NIAID) is that about 25% of the entire human population of the globe is infected. The mouse has been widely used as a model system for this microbe. Significantly, some mice strains are highly susceptible to infection by this pathogen, whereas others are quite resistant to infection (Medina and North, 1999, Immunology; 96(1): 16-21). The signalling networks that lead to phagosome acidification, actin nucleation, microtubular function, oxidative burst (including production of reactive oxygen intermediates (ROI) and reactive nitrogen intermediates (RNI)) and/or membrane fusion may thus be compared between MT phagosomes from mice and MT phagosomes from human macrophages, or alternatively MT phagosomes isolated from susceptible and resistant strains of mice, including cells derived from such humans and animals. Such a screen allows the identification of important differences in the behaviour of the membranes of the two different hosts. This is likely to lead to the identification and evaluation of medically-relevant mechanisms, such as mechanisms for boosting innate resistance and/or the bacterial killing capabilities of macrophages and neutrophils.

Another embodiment of the invention may involve the comparison between phagosomes containing dead and live pathogens. Generally, in the case of pathogens that inhibit the fusion of phagosomes with lysosomes (such as, for example, mycobacteria and salmonella), the bacterium needs to be alive for this effect; heat- or formaldehyde-killed pathogens invariably follow the route taken by latex beads, i.e. they fuse fully with lysosomes. Additionally, the role of the immune system can be evaluated, for example, by comparing IgG or complement-opsonised beads, or microbes, with non-opsonised particles. In a further aspect of the invention, the protein composition of a phagosome may be compared between different pathogen phagosomes. Such a comparison may facilitate an understanding of the molecular mechanisms through which pathogens alter phagosome properties. This analysis is now facilitated because a large number of molecules present in the phagosome system are known to the inventors and their collaborators. Furthermore, the activity of many of the known lipids and proteins that are present in this system can be regulated by the manipulation of the levels of simple metabolic components.

In one embodiment, this aspect of the invention therefore provides a method for the identification of a factor that is implicated in the mechanism of virulence in a pathogen. By "implicated in the mechanism of virulence in a pathogen" is meant that the particular factor participates somehow in the mechanism by which a pathogen avoids the defence mechanisms of the host cell and therefore survives and multiplies.

In one embodiment, this aspect of the invention may thus provide a method for the identification a factor that is necessary for virulence to be effective in a host, said method comprising: a) comparing the lipid composition and/or protein composition of a phagosome containing a virulent micro-organism with the lipid composition and protein composition of a phagosome containing a non-virulent micro-organism; and b) selecting as the factor, a lipid or protein factor that is only present in the phagosome that contains the virulent micro-organism and is not present in the phagosome that contains the non-virulent micro-organism.

In an alternative embodiment of this aspect of the invention, there is provided a method for the identification of a factor that is responsible for virulence in a micro-organism, said method comprising: a) comparing the protein and/or lipid composition of a virulent micro-organism in a phagosome with the lipid and/or protein composition of a non-virulent micro-organism in a phagosome; and b) selecting as said factor, a lipid or protein factor that is only present in the virulent microorganism and is not present in the non-virulent micro-organism. The factor that is implicated in the mechanism of virulence in the pathogen may thus be a pathogen-derived factor, or may be host-derived (e.g. TACO, see Cell, 1999, 97(4), 435- 47). In the former case, the pathogen-derived factor will generally be a protein whose effects contribute to the evasion of the cellular host defences against infection. For example, one of the strategies used by pathogens to invade host cells successfully is to alter phagosome fusion with endocytic organelles, a process resulting in the inhibition of phagolysosome biogenesis (Finlay and Falkow, 1997).

In the latter case, the host-derived factor will generally be a protein that is present in the phagosome membrane with which the pathogen interacts.

In preferred aspects of these embodiments of the invention, the method does not necessarily focus on identifying the actual factors that are responsible for the mechanism of virulence, but instead analyses the effects of these factors on signalling. In this way, signalling effectors that regulate a defined phagosomal membrane function (such as actin nucleation, actin binding, microtubule binding, microtubule motility, membrane fusion, the generation of ROI and RNI, and the altering of pH) can be studied.

In this manner, different phagosome types or conditions can be compared. For example, signalling pathways in phagosomes containing virulent pathogens can be compared to those in phagosomes that contain non-virulent pathogens. A specific example of a signalling pathway is that which governs actin nucleation; for the two conditions to be compared the list of effectors shown in Figure 1 may be tested and scored as +,- or 0. Differences are noted and so-called "elementary modes" (as well as other algorithms, such as those described by Schuster et al 2000 loc cit) may be used to make predictions about the behaviour of the signalling networks.

Of course, the same process can be repeated for any function for which an in vitro functional assay is available. In this way, key nodes of a signalling network relating to virulence, for example, may be identified and targeted by distinct molecular entities, such as drugs.

In this way, the signalling networks related to a specific receptor, or to a specific pathogen or pathogens are made accessible in a manner that extends well beyond current capabilities. Presently, state of the art technology in molecular cell biology does not offer the possibility to visualise broad networking connections in a single cellular signalling system. Current techniques at the whole cell level are perfectly able to analyse single, dimeric or trimeric (or even larger) interactions (which can be independently confirmed using purified components), but the complexity of living cells precludes a holistic understanding of any signalling network in a cell, and especially those events that occur within, or on the surfaces of membranes. Accordingly, despite the vast network of known potential interactions, in no membrane system can a complete network currently be visualised.

Another example of great medical relevance, to which the methods of the present invention may be applied, is the oxidative burst phenomenon, including production of reactive oxygen intermediates (ROI) and reactive nitrogen intermediates (RNI), that is especially prominent in neutrophil phagosome function. A large number of molecules that operate in this system have now been identified. However, no holistic model is presently available. Using latex bead phagosomes from neutrophils (which are known to be especially active in oxidative burst in vivo) one can take advantage of existing assays for monitoring oxidative burst events such as oxygen uptake, or free radical release by phagosomes in order to set up in vitro assays to study signalling involved in the oxidation burst phenomenon, including production of reactive oxygen intermediates (ROI) and reactive nitrogen intermediates (RNI), once in vitro assays become available. Methods for the development of such assays, as well as those for acidification, will be clear to those skilled in the art. One reason for the suitability of the phagosome systems described above in the methods of present invention is that the state of activation of membrane signalling in the system is easily manipulable, being greatly influenced by the ATP levels in the system. While physiological levels of ATP are generally in the range of l-5mM, this level can drop significantly during ischemia which, for example, is a common problem during surgical operations. Under conditions of low ATP concentration (around 0.2mM), it has been found that a vast cascade of membrane proteins and lipids are activated / synthesised by phosphorylation events that are part of the molecular machinery of membrane signalling regulating actin nucleation. At this concentration of ATP, the system is constitutively "on", but can be inhibited by negative effectors such as sphingosine-1 -phosphate (SIP) (Figure 1), and drugs that affect enzymes such as DAG-kinase (data not shown). At high ATP (around 5mM, corresponding to cellular ATP levels) the system is switched "off, unless it is activated via signalling. Among the positive effectors at high ATP concentration is SIP, showing that the effects of this lipid can be differentially regulated in vitro by the concentration of ATP. In contrast, the upstream precursor of SIP, sphingosine, behaves in the opposite manner, stimulating at low ATP and inhibiting at high ATP. Preliminary lipid analyses suggested that at low ATP SIP stimulated phospholipase D (PLD) to make more phosphatidic acid (PA) (summarised in Figure 1). However, more rigorous analysis revealed that, in fact, sphingosine activates DAG kinase to produce more PA (see Figure 3G). The knowledge of the K_m of the various enzymes for ATP is now giving the first hints of how selected enzymes can behave differently at low versus high ATP. For example, DAG kinase and choline kinase have a K_m for ATP in the order of lmM and can be expected to be much less active at low ATP. In contrast, phosphatidyl inositol 4-phosphate kinases (PI4 kinases) have a Km that is about ten-fold lower and are expected to be much more active at low ATP.

Phosphorylation studies have shown that many phagosomal proteins and lipids are phosphorylated in vitro by ATP. These phosphorylated components are currently being identified and are thought to include ezrin and moesin, that are essential for the actin nucleation. A number of lipids are also synthesised, including PLP, PI₄₅P₂ and PA. The levels of these compounds can be modulated by drugs affecting signalling pathways. Using technology currently available (see Examples herein), it is possible in principle to identify all of the phosphorylated proteins and lipids in this system.

According to a further aspect of the invention, there is provided a method of screening for a candidate drug molecule capable of altering a membrane signalling network comprising contacting a phagosome composition according to any one of the embodiments of the invention described above with the candidate drug molecule, and analysing a membrane signalling network in the phagosome composition. Using such a screen, it will be possible to identify drug molecules or combinations of drug molecules that are effective in regulating core signalling mechanisms that are shared between different cell types, including both inhibition and up-regulation of these signalling mechanisms. Furthermore, drug molecules may be identified that are effective in the prevention of successful host cell infection by pathogenic organisms, for example by activating or inhibiting specific nodes of signalling networks in order to activate phagosome 'killing' functions (such as actin nucleation, phagosome-lysosome fusion, phagosome acidification, and the oxidative burst phenomenon) thus preventing successful host cell infection by pathogenic organisms.

All existing therapies against microbial pathogens target the pathogen itself, (one example is the widespread use of antibiotics). The approach described herein targets the phagosomal membrane, and the signalling cascades within. For example, M. tuberculosis (TB) blocks fusion of phagosome with lysosomes; if one understood the networks in the phagosome membrane that regulate this fusion event, one or more reagents could be added to infected animals or humans, that can switch on this fusion process i.e. the signalling networks would be activated by simultaneously switching on several key regulatory hubs or nodes. According to the invention, these would be identified by the procedure outlined above, i.e. by a combination of functional assays, protein and lipid analyses and bioinformatic approaches. It is known from published data from cultured macrophage experiments that if the TB phagosome fuses with a lysosome, the bacterium is usually effectively destroyed. Alternatively, one could add effectors that boost the anti-microbial oxidative burst phenomenon or use our knowledge of these networks to overcome the inhibitory effects of some bacterial agents, including toxins, proteins and lipids, that block this process.

The ability to use beads coated with bacterial virulence factors and other proteins that alter the membrane trafficking routes of the host cells is a useful adaptation of this system. In addition to screening for novel drugs, the methods and screens described herein may be used for the further analysis of drug molecules that have proven to have interesting effects on cultured cell systems, or in animal models. For example, an interesting inhibitor or effector molecule may be screened in one or more of the numerous assay types described above. Then if, for example, a drug was found to affect the process of actin nucleation, a more refined series of conditions known to effect the signalling pathways leading to actin nucleation could be used, such as by using drugs that stimulate or inhibit PKC or PLD, to analyse the signalling network in a holistic fashion.

In one embodiment, the method of this aspect of the invention may comprise the steps of: a) preparing at least one phagosome composition in which the phagosomes contain one or more components that are in a particular active state of a membrane signalling network; b) contacting said phagosome composition with a candidate drug molecule; c) testing a component, or an effect of a component, of said membrane signalling network in order to assess the effect of said candidate drug molecule on the network for any given function; and d) selecting as said candidate drug molecule, a drug molecule that has a desired effect on said signalling network.

Preferably, in step c), the effect of the candidate drug molecule is assessed with respect to the membrane signalling network regulating actin nucleation. Any one of the methods of the aspects of the invention that are described above may include a step of outputting as a result, the identity of a component of a membrane signalling network that is implicated by the method of the invention in a membrane signalling pathway.

According to a further aspect of the invention, there is provided the use of a phagosome preparation to analyse a membrane signalling network, such as a specific ligand-induced membrane signalling network, to identify a component of a membrane signalling network. The use of the latex bead phagosome is particularly preferred in this respect. For example, beads to which hyaluronic acid, or the Shigella protein IpAB is coupled offer a more defined system for analysing actin nucleation since the CD44, ezrin, moesin and PIP₂ moieties provide the core of a potentially large actin nucleation machine in/on the phagosome membrane. The cytoplasmic domain of CD44 binds to Ezrin, which itself binds to PIP₂ and actin. The same combination of compounds is also expected to operate in concert on the plasma membrane of many cells. This makes the latex bead/hyaluronic acid or IpAB phagosomes a particularly powerful tool in a general method for assaying actin nucleation on a membrane surface. Alternative examples of ligands immobilised on beads for phagocytosis studies include host cell proteins such as IgG or complement components, as well as pathogen ligands such as lipopolysacchariden or lipoarabinomannan, as well as other ligands that bind the many well-characterised surface receptors on cells such as macrophages, neutrophils, endothelial and epithelial cells.

The invention also provides for the use of a phagosome preparation in a screen for the identification of a candidate drug molecule capable of altering a membrane signalling network, such as that relating to actin nucleation.

All of the systems discussed herein are highly amenable to a refined bioinformatic analysis. Such approaches will be greatly facilitated by ongoing proteomic (Garin et al, 2001, J.Cell.Biol. 152. 165-180) and lipid analysis by thin layer chromatography (TLC), high pressure liquid chromatography (HPLC) and mass spectrometry to identify all the lipid species in phagosomes and to evaluate how they are modulated by different effectors. The former approach has recently identified close to 150 distinct polypeptides in latex bead phagosomes (2hr of age) from J774 macrophages. The latter set of methods have identified close to 20 different lipids in the latex bead phagosome membrane. As discussed above, the intention of collaborators of the inventors is to map the proteome of a number of different phagosome types, so leading to an exhaustive directory of the 400 to 500 constituent proteins that are present in these organelles. The generation of a parallel lipidomic map is currently in progress. By comparing the membrane signalling networks between different phagosome types, and assessing any differences in protein or lipid composition between the compared phagosomes, it is envisaged that justification may be found for any signalling differences recorded. The use of proteomics, lipid analysis and in vitro functional bioassays on phagosomes has now provided a large catalogue of molecules that are essential for membrane signalling function. Moreover, complex links are emerging between phagosome membrane signalling networks and to known signalling networks. The number of known components is already more than 30 proteins/lipids. As discussed above, it is thought that mapping these networks will lead to signalling network models that are predictive. These models may be validated and developed by experimental testing. An algorithm recently published in the literature (Schuster et al, (2000) Nature Biotech. 18(3): 326-332) allows the analysis of the flow of information through complex multi-component networks of biochemical interactions. Initially, a model will be established based on known biochemical reactions that will lead to experimental predictions. The latter will then lead to a refinement of the model, so leading to the design of new experiments. By this iterative process, this analytical tool will ultimately provide a mathematical description of membrane signalling networks and their dynamics.

Accordingly, a further aspect of the invention provides a relational database system comprising information relating to proteins and lipids that are present in a phagosome membrane system, the database comprising a plurality of tables containing information relating to at least two of the following: a) the identity of proteins and lipids present in any phagosome system; b) the regulatory mechanisms of proteins and lipids present in the phagosome system; c) the putative or proven function of proteins and/or lipids present in the phagosome system; d) post-translational modifications of proteins, and modifications of lipids, present in the phagosome system; and e) alteration of the level of expression or activity of a protein molecule and/or a lipid molecule present in the phagosome system as a result of infection of a host by a micro-organism. This database system may additionally comprise information relating to the proteome of a phagosome or comprising information relating to lipids present in a phagosome system.

Recent analyses of different kinds of networks (from the world wide web to metabolic networks) show clearly that such 'scale-free' networks have a natural tendency to self- organise into similar overall patterns, a consequence of the dynamics of the system (Phys. Rev. E. Stat. Phys. Plasmas. Fluids Relat. Interdiscip. Topics, 2001, 64(2-2), 026704). A key element of such networks is that they organise in such a manner that most 'nodes' in the system are relatively sparsely interconnected to other 'nodes'. However, a few 'key nodes' are heavily interconnected with other nodes. A consequence of this network structure is that if one inactivates any node at random, the chances are high that a poorly connected node would be 'hit', and that the overall robustness of the network would not be affected. If, however, by design or by chance a 'key node' is hit, there is a much higher chance that the overall behaviour of the network would be significantly affected. Such a phenomenon has been clearly demonstrated for the world-wide web (Albert et al., Nature, 406. 378-382). In the case of signalling networks regulating phagosome functions where one may wish to activate or inhibit a process (such as actin nucleation), it seems logical to propose that, by analysing the overall behaviour of the network (by the approaches described herein) one can identify such 'key nodes' that are more heavily interconnected to other nodes.

Figure 12 shows the connectivity of metabolites in the phagosomal membrane that regulate action nucleation. Evaluation of this diagram, as well as the 'elementary modes' shows, for example, that DAG kinase and choline transferase are key enzymes that occur in many modes. These would logically appear to be key enzymes that one could target in order to switch actin nucleation on or off.

In another application, suppose that a drug blocks actin nucleation by inhibiting Phospholipase C. The approach of the invention could pinpoint the site of this inhibition by network analyses. For example, in the presence of the drug, the addition of DAG (or IP3) would be expected to allow the network to proceed normally. It is the combination of the use of in vitro (and sometimes also in vivo), functional assays and bioinformatic approaches such as 'elementary mode' analysis that allows our approach to be successful in analysing networks. Thus, the present invention allows the identification and evaluation of key 'nodes' within phagosomal signalling networks via the combination of bioinformatic and practical approaches.

It has been discovered, using the techniques that are outlined above, that it is possible to modulate the ability of phagocytic cells to destroy pathogens by altering the concentration of certain lipid species to which the phagocytic cells are exposed. This aspect of the invention thus provides a method for modulating the ability of a phagocyte to destroy a pathogen, the method comprising contacting the phagocyte with lipid selected from the group consisting of arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1- phosphate and phosphatidyl-inositol-bis-phosphate. This is the first demonstration of such an effect and paves the way for the development of specific techniques and agents that enhance the body's natural defences against pathogens, particularly pathogenic microorganisms such as mycobacteria.

It has also been found that it is possible to improve the efficacy of pathogen killing by phagocytic cells by contacting the phagocyte with a modulating agent that stimulates phagosome maturation. This aspect of the present invention thus provides a method for modulating the ability of a phagocyte to destroy a pathogen, comprising contacting the phagocyte with a modulating agent which stimulates phagosome maturation and pathogen killing. In particular, it has been demonstrated herein that pathogen killing mediated by phagocytes may be significantly enhanced by increasing the degree of actin nucleation occurring in the phagocyte cell. This is the first time that the killing functions of the phagosome have been linked with the process of actin nucleation. Preferably therefore, the mechanism by which the modulating agent stimulates phagosome maturation and pathogen killing by phagocytes is that of actin nucleation. The modulating agent may also stimulate phagosome-lysosome fusion in phagocytes, lead to a reduction in phagosomal pH, and/or stimulate the release of ROI and RNI. The modulation agent may lead to a combination of the above effects, thereby enhancing phagosome maturation and pathogen killing.

Using the methods of the invention, libraries of potential modulating agents may be screened in any of a variety of drug screening techniques. Candidate modulating agents may be isolated from, for example, cells, cell-free preparations, chemical libraries or natural product mixtures. These compounds may be natural or modified substrates, ligands, enzymes, receptors or structural or functional mimetics. For a suitable review of such screening techniques, see Coligan et al., Current Protocols in Immunology l(2):Chapter 5 (1991 and other relevant texts that will be known to those of skill in the art).

In a preferred method, the modulating agent of the above-described methods is a lipid species. Most preferably, the modulating agent is selected from the group consisting of arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1 -phosphate and phosphatidyl-inositol-bisphosphate.

The invention also provides for the use of such a modulator agent in the manufacture of a medicament for the treatment or prophylaxis of phagocytic pathogen infection. Preferably, this use leads to the treatment or prophylaxis of phagocytic pathogen infection through the stimulation of pathogen killing by phagocytes. This stimulation of pathogen killing may be due to increased actin nucleation, to increased levels of phagosome-lysosome fusion, a reduction in phagosomal pH and/or the increased release of ROI and RNI.

Preferably, the pathogen described in the above methods and uses of the present invention is a mycobacterium, particularly a mycobacterium selected from the group consisting of M. tuberculosis, M. paratuberculosis, M. leprae, M. avium and M. bovis. These pathogens all exploit the mechanism of phagocytosis to infect susceptible cells.

The present invention also relates to a method of treating or preventing infection of a host cell by a pathogen, comprising contacting infected phagocytes in a patient with a therapeutically-effective dose of a modulating agent that stimulates actin nucleation, particularly a lipid species.

A further aspect of the invention provides a method of treating a human or animal having a disease or condition associated with pathogen phagosomes, particularly mycobacterial pathogen phagosomes, comprising administering to said human or animal a therapeutically or prophylactically effective amount of a modulating agent that stimulates actin nucleation. The term "therapeutically effective amount" as used herein refers to an amount of a therapeutic agent needed to treat, ameliorate, or prevent a targeted disease or condition, or to exhibit a detectable therapeutic or preventative effect. For any compound, the therapeutically effective dose can be estimated initially either in cell culture assays, for example, of neoplastic cells, or in animal models, usually mice, rabbits, dogs, or pigs. The animal model may also be used to determine the appropriate concentration range and route of administration. Such information can then be used to determine useful doses and routes for administration in humans.

The precise effective amount for a human subject will depend upon the severity of the disease state, general health of the subject, age, weight, and gender of the subject, diet, time and frequency of administration, drug combination(s), reaction sensitivities, and tolerance/response to therapy. This amount can be determined by routine experimentation and is within the judgement of the clinician. Generally, an effective dose will be from 0.01 mg/kg to 50 mg/kg, preferably 0.05 mg/kg to 10 mg/kg. Compositions may be administered individually to a patient or may be administered in combination with other agents, drugs or hormones, including pharmaceutically-acceptable excipients.

In a further aspect of the invention, the modulating agent used for the treatment of diseases or conditions associated with pathogen phagosomes, particularly mycobacterial pathogen phagosomes, is a lipid, and most preferably is selected from the group comprising arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1 -phosphate and phosphatidyl -inositol-bisphosphate.

Various aspects and embodiments of the present invention will now be described in more detail by way of example. It will be appreciated that modification of detail may be made without departing from the scope of the invention.

Brief description of the Figures Figure 1 shows the signalling component PI₄₅P₂ included as a functional node and connected to a large network based on known biochemical pathways that regulate actin nucleation. The circles (with (+), (-) and (0) notation) show the effect of adding the compound on the LBP actin nucleation relative to control phagosomes under standard assay conditions. Figure 2 shows a summary of known biochemical pathways that can be linked to the actin nucleation machinery that operates in vitro on the surface of a LBP

Figure 3A shows the visualisation of rhodamine actin (see arrows) on LBP having green fluorescent beads in the actin nucleation assay. At higher magnification, Figure 3B shows an example of a loose bundle of actin attached to the LBP. The beads are lμm. Figure 3C shows the percentage of actin-labelled LBP following co-incubation with increasing ATP concentrations (given in mM). The inset shows the difference in nucleating activity at low ATP, following pre-incubation of LBP (followed by gradient purification) with 0.2 mM versus 5mM ATP. All of these data exemplify what is shown on Figure 4. Figure 3D shows the effects of co-incubating LBP with 100 mM of SIP, or N,N-dimethylsphingosine (an inhibitor of Sp kinase) at high and low ATP. Figure 3E shows the effects of lOOnM sphingosine at high and low ATP, while Figure 3F shows the corresponding experiment with ImM ceramide. Figure 3G shows a TLC autoradiogram of the in vitro phosphorylated lipids from untreated phagosomes (lane 1) or following treatment with SIP (lane 2) or Sp (lane 3). Both lipids lead to alterations in the lipid pattern relative to control, with Sp addition inducing a prominent increase in SIP and in PA. Additional, yet to be identified lipids are also changed (see unlabeled arrows). Figure 4 Summary of the effects of lipids on LBP actin nucleation at high and low ATP. "+" indicates a stimulation, "-" an inhibition, and "0" no effect relative to the standard condition at low ATP. The notion that a PKC can inhibit ezrin-facilitated actin nucleation is based on a survey of the literature. The asterisk next to Sphm denotes the fact that this was tested only after the elementary mode analysis predicted it to be a positive effector of actin nucleation. The abbreviations used in Figure 4 and elsewhere are as follows:

Metabolites:- arachidonic acid (AA), lysophosphatidylic acid (LPA), phosphatidylcholine (PC), phosphatidyl inositol (PI), phosphatidyl inositol 4-phosphate (PIP), phosphatidyl inositol 4,5 bis-phosphate (PIP2), diacylglycerol (DAG), phosphatidic acid (PA), sphingosine (sph), sphingosine 1 -phosphate (S-l-P), inositol 1 ,4,5-trisphosphate (inositolTP), monoacylglycerol (monoAG).

Enzymes:- Dihydrosphingosine kinase (SPK), phospholipase A (PLA2), phospholipase C (PLC), phospholipase D (PLD), phosphatidate phosphatase (PAP), CDP-diacylglycerol— inositol 3-phosphatidyltransferase (CDPinotra), sphingomyelin phosphodiesterase (Sphmydias), ceramidase, lipase, DAG kinase (DAGkin), phosphatidyl inositol 4-phosphate kinase (PI4kin), phosphatidyl inositol 4,5-phosphate kinase (PI4P5kin), phosphatidyl inositol 4,5-phosphate diesterase (PI45dias), CDP-DAG synthase (CDPsynth), DAG- cholinephosphotransferase (DAGcholT), choline kinase (Cholinkin), phosphatidyl inositol diesterase (PIdiase), formation of Ezrin, PIP2 and its receptor (ERMform), choline- phosphate cytidylyltransferase (CholinPtf), actin nucleation (Actinnucl), actin depolymerization (Actindepoly), mono-acylglycerol-cholinphosphate-transferase

(mono AGcholinPtf) .

Figure 5 Actin nucleation by mycobacterial phagosomes. Figure 5A shows an example of a GFP-expressing, M.smegmatis enclosing phagosome having newly assembled rhodamine actin (red). The arrows in Figure 5A and 5B illustrate the actin (red when viewed under the microscope) that has been nucleated by the phagosome. Figure 5B shows the equivalent experiment with heat-killed M.avium surface-labelled with FITC (green under the microscope); the actin labelling here is especially prominent at the phagosome poles. Figure 5C indicates the percentage of actin-labelled M.smegmatis phagosomes under different conditions, dead versus live, low (L) versus high (H) ATP, and with or without co- incubation with the different lipids. For each lipid a mock treatment was used in which the standard assay at low ATP was supplemented with the lipid solvent without the lipid. A summary of the results in the simple +, -, 0 notation is given below the figure in which each treatment is compared to the mock control at low ATP. Figure 5D shows the equivalent experiment with killed M. αv m-containing phagosomes.

Figure 6 Effects of lipids on actin nucleation by live M.a v/wm-containing phagosomes. The abbreviations and conditions are the same as for Fig 5. A summary is given below the figure in the +, -, 0 notation. Figure 7 Testing predictions of the elementary mode analysis. The effects of CDP choline and Sphm on LBP, killed (K) or live (L) M.avium phagosomes (MaP) with respect to actin nucleation is shown at both high (H) and low (L) ATP. The control shows the value for the standard condition at low ATP wheras for the treatments with CDP-choline and Sphm a mock control is shown in which identical buffer conditions were used without the effectors. Figure 8 Effects of lipids on actin assembly in infected macrophages. Figure 8Ai and ii show labelling for rhodamine phalloidin (red under the microscope) around FITC-labelled (green under the microscope) bacteria in phagosomes. In Figure 8Ai, the phagosome having live M. avium is devoid of actin labelling whereas in Figure 8Aii the phagosome containing the killed pathogen is strongly labelled for actin and highlighted with an arrow. In Figure 8B a quantitation of this labelling is shown for the different live bacteria, as well as killed M.avium phagososmes from untreated cells (con, denotes the control), or following treatment with 125μM AA for 6, or 24 hrs. The percentage of phalloidin-labelled phagosomes is shown. Figure 8C summarises the same parameters for .tb-infected cells incubated with the different lipids. The concentrations used were, AA 125μM, PIP₂50μM, Sphm (SM) lOOμM, S IP lOOnM, SP lOOnM, ceramide ImM, DAG 20μM.

Figure 9 Effects of lipids on phagosome fusion. Figure 9A and Figure 9B show a quantification of the percentage of phagosomes in macrophages that acquired lOnm gold- BSA particles that had been pre-internalised into cells for lh pulse followed by a 2h chase before internalising the bacteria for lh, rinsed and then chased in the absence of bacteria but in the presence of lipids for a further 5h. The cells were embedded in epoxy resin and the examination done on thin sections by EM. The concentrations of the lipids were identical those given for Figure 8.

Figure 10 Effects of lipids on phagosome pH and on fusion. Figure 10 Ai and ii show the labelling of cells infected with killed M.tb (i) and live Mtb (ii) for lh followed a rinse and chase of lhr, followed by an overnight chase in the presence of lyso tracker red. The co- localisation of this pH indicator with the bacterial phagosome here is a sensitive indication of both fusion and acidification. A quantification of the percentage of labelled phagosomes is given for killed and live Mtb Fig 10B and M.avium Fig 10C following treatment with different lipids (used at the same concentrations as described for Fig 8). Figure 10D shows a quantification of the pH of phagosomes enclosing live and killed M.avium using FITC covalently conjugated to the bacteria before internalisation for lh followed by a 5h chase. The pH is estimated from the ratio of fluorescence emission at 460 and 495nm. On the left of the figure the pH values are given for the corresponding experiment done in the presence of lOμM nigericin, which dissipates the proton gradient, and standard pH buffers of the indicated pH in order to calibrate the pH. From these data, for example, it can be seen that the live M.avium phagosomes have a pH in cells of about 6.3 while the killed bacterial phagosome acidifies down to about pH around 5 and that treatment with ceramide significantly lowers the phagosomal pH.

Figure 11 Effects of lipid treatment on the killing of M.avium (Figure 11 A) M.tb (Figure 1 IB) in J774 cells. The bacteria were allowed to infect for lhr and, after rinsing, the lipids were added to the culture medium at the concentrations given in Figure 8. After 1, and 3 days the cells were scraped, pelleted and homogenised in the presence of 1% NP40 (which has no effect on the bacteria). A dilution series of the bacteria were plated on agar and the number of colony-forming units was estimated after two-three weeks of culture. The data are presented as the percentage of bacterial survival relative to the untreated cells.

Figure 12 shows the metabolites and reactions involved in the polymerisation of actin. Elementary mode analysis of this scheme allows us to identify the key nodes in the signalling networks regulating actin nucleation (the reactions that are most interconnected to other reactions). Examples

1. Biochemical assays

There are a number of biochemical assays available that allow the analysis of the membrane signalling networks that are described above. Some of these assays have already been reported in the literature in a different context to the methods of the invention, and some have been developed specifically for this purpose. A brief summary of these assays is given below.

1.1 Membrane Fusion Assay involving phagosomes and endocytic organelles

Two different membrane fusion assays have been developed. The first is a biochemical, content-mixing approach that uses thin sections to evaluate content mixing (see Jahraus et al. 1998, under "Biochemical in vitro fusion of phagosomes with endocytic organelles").

The second comprises two different electron microscopy (EM) approaches, that use negative staining or cryo EM to evaluate content mixing. One assay is described under "EM in vitro Fusion of Phagosomes and Endocytic Organelles" in Jahraus et al, 1998; Jahraus et al., 2000. These assays can be adapted to investigate the fusion of any potentially fusing organelles. As examples of their utility, these assays have been demonstrated to allow the identification of an important role of rab5 and the actin cytoskeleton in membrane fusions (Jahraus et al., 1998; Jahraus et al, 2001, in press in Mol. Cell Biol. 2001-Jan). Effectors of known signalling components can be added in vitro, as well as in vivo, and tested for their effects on a defined function, such as actin nucleation, by scoring for stimulation (+), inhibition (-) or no effect (0), as explained above. Simpler and more sensitive assays using fluorescent lipids described by Peyron et al. (J. Biol. Chem., 2001, 276, 35512-35517) for neutrophil LBP fusion assays will be adapted for macrophage phagosomes.

1.2 In vitro fusion assay between early endosomes A complementary fusion assay has recently been developed to monitor the homotypic fusion between early endosomes of J774 cells or BHK cells. This fusion process shows many molecular similarities to the phagosome-endosome fusion process. It is easier and more sensitive than the above assay and requires only a few microlitres of membrane organelle material per reaction (cf. to 50ul or more in the biochemical fusion assay discussed in Section 1.1.) This assay is a modification of an earlier chemiluminescence assay previously developed at EMBL - (see Horiuchi et al. 1997, Cell, 90; 1149-1159). The procedure for homotypic fusion of early endosomes from BHK cells, either from crude endosome preparations or from highly purified endosome preparation, was developed in collaboration with colleagues in Norway (Tjelle et al. 1998 J. Cell. Sci. 109,2905-2914)

Briefly, early endosomes from two sets of cells are labelled in vivo for 5 mins with biotinylated transferrin and sheep α-human transferrin antibody. The early endosomes may be purified from J774 use macrophages by density gradient centrifugation (Tjelle et al, 1997). The basal fusion reaction is carried out by incubating the two populations of endosome population for 30 min at 37°C in the presence of 4mg/ml J774 cytosol, an ATP generating system (17 mM creatine phosphate, 87 μg/ml creatine kinase, and 2.2 mM ATP), and unlabeled transferrin. The fusion is then quantified by incubating the fusion mixture in a wash buffer (50 mM Tris (pH 7.4), 100 nM NaCl, 1 mg/ml BSA, and 2% Triton X-100) and streptavidin-coated paramagnetic Dynabeads for one hour. After washing two times using a magnet to retain the beads the samples are incubated with a rabbit α-sheep secondary antibody coupled to a ruthenium trisbipyridene chelate (IGEN) and the chemiluminescence signal, indicative of the mixing of transferrin from one set of endosomes with anti-transferrin from the other set, is estimated using an ORIGEN analyser.

1.3 Microtubule - Binding Assay

A fluorescence microscopy assay has been developed to quantify the binding of phagosomes to rhodamine-labelled microtubules in the presence of a low concentrations (2mg/ml) J774 macrophage cytosolic extracts (see under "Binding assay", in Blocker et al., 1996). This assay identified a MAP faction and demonstrated that this factor preferentially binds to the "plus" ends of microtubules (Blocker et al, 1996). This assay, as well as the actin-binding assay (see below) is easily adaptable for use with 96-well plates combined with a robotic, high content screening system that can automatically focus on, and record in each chamber the number of bound fluorescent phagosomes per field. This would facilitate large-scale (high-throughput) screening of multiple effectors or drugs.

1.4 Microtubule - Motility Assay

This is a similar assay to the assay discussed above in section 1.3, except that a relatively high concentration of cytosol is required for the bi-directional motility of phagosomes along microtubules (MTs) (see under "motility assay", in Blocker et al, 1998). Further video microscopy is employed in this assay, which demands a high level of technical expertise. This assay led to the identification of kinesin, kinectin, dynactin and dynein in bidirectional phagosome motility (Blocker et al., 1997).

1.5 Actin Binding Assay

The group of Sergia Kusnetzov (University of Rostock) in collaboration with the Applicant, has established a light microscopy assay that reconstitutes the binding of phagosomes to labelled filamentous actin. This assay, which requires cytosol, has identified myosin V as one essential binding protein in this process (Al Haddal et al. 2000, in revision for J. Cell. Biol.). This assay may also be adapted to analyse phagosome-actin-myosin motility. The details of this assay are as follows. Actin filaments are polymerized from rabbit skeletal muscle G-actin, prepared by the method of Spudich and Watt (1971, J Biol Chem., 246(15):4866-71) by the addition of 2 mM MgCl₂. After overnight polymerization, actin filaments were stabilized and labelled with rhodamine-phalloidin (Sigma) using a 1: 1 molar ratio.

Microscope chambers are built from a glass microscope slide (Menzel Super Frost, Gerhard Menzel GmbH, Braunschweig, Germany) and an 11 -mm circular glass coverslip (Menzel) sealed onto two pieces of double-sided tape (3M Scotch), forming a 2-3 μl chamber. All incubations are carried out in a moist chamber at room temperature. Filamentous actin (0.5 μM), stabilized and labelled with rhodamine-phalloidin, is perfused into the chamber and incubated for 5 min. Non-specific binding is blocked by perfusion of the chamber with 3 mg/ml casein (Sigma) in homogenisation buffer (HB). Excess actin filaments and casein are washed away by perfusion with 3 chamber volumes of HBS, (HB containing 10% sucrose).

One volume of phagosome binding reaction mixture, containing phagosomes [working concentration 0.001% (wt/vol)], 0.3 mg/ml casein and binding factors to be tested are perfused into the chamber and incubated for 20 min. Unbound phagosomes are washed away by perfusion with 3 volumes of HBS. Binding was analyzed by fluorescence microscopy with a Nikon Diaphot 300 microscope with a 10 x eyepiece and Nikon lOOx

PlanApo oil objective (field surface area of 22,000 μm^). The bound phagosomes are counted by eye, and in each experiment values from at least 10 fields from two separate, but identical, reactions are routinely averaged and the standard deviations calculated. 1.6 Actin Nucleation Assay

Actin nucleation on membrane surfaces is one of the most elusive of cellular processes. However, this process is crucial for a large range of essential cell functions, including cell motility, microvilli, filipodia and lamellipodia assembly, phagocytosis, endocytosis, differentiation development as well as a crucial, though poorly defined role, in metastasis and cancer. Three independent and powerful assays have now been developed that provide complementary information about this membrane-dependent actin nucleation process. These assays can be complemented by use of standard biochemical procedures to study actin or membranes.

1.6.1 • Pure system

In this assay (without cytsosol), phagosomes are mixed with 2μm muscle G actin, 6μM thymosin 64 (Tβ4) and a buffer. This use of T64 as an actin buffer, is a novel and crucial aspect of this assay. Actin nucleation is monitored by fluorescence microscopy (Defaque et al., 2000a; see under "In vitro actin assembly assay: fluorescence microscopy") or flow cytometry (Defaque et al. 2000b-submitted). This assay had led to the identification of ezrin and moesin as being essential for actin nucleation. It has also identified a vast signalling network that is operating in vitro in this system (see below).

1.6.2 • Torsional Rheometry

Two complementary approaches have been used for analysing the more complex process of membrane-dependant actin nucleation on phagosomes that occurs in cytosolic extracts. The first uses a biophysical method, torsional rheometry to follow the viscoelastic properties of the cytoplasmic extracts / phagosome system. The use of this system has confirmed that phagosomes nucleate actin and has provided a physical description of the process. The effects of drugs affecting this system can be followed directly and kinetically (Jahraus et al, Mol. Biol. Cell., 2000, 12, 155-170). A microrheometry system using laser traps has recently been built at the EMBL by Ernst-Ludwig Florin that requires only 20-50 μl of material, instead of the 500ul that is needed for the macrorheology approach. The theoretical background and experimental details (from Jahraus et al, Mol. Biol. Cell., 2000, 12, 155-170) are provided below.

1.6.3 Rheological Measurements

Quantitative measurements of the macroscopic viscoelastic properties of the fusion assay were made by using a rotating disc rheometer, described in detail by Mϋller et al. (1991, Macromolecules, 24: 3111-3120). The actin solution is contained in a cylindrical cuvette and is covered by a lipid monolayer to prevent denaturing of actin due to exposure to air. A silanized disc is placed on top of the solution. Torsional oscillations of the disc are excited by an oscillatory magnetic field acting on a small magnet fixed on the centre of the top of disc. Time-dependent measurements were conducted at one fixed frequency (ω = 0.2 rad Is) at 37°C to monitor predominantly the polymerization of actin in cytosol and to study the effect of the presence of phagosomes, PNS membranes and ATP on the polymerization.

As a reference, 4 μM pure G-actin in F-buffer was also tested. In all cases, our conventional fusion assay was scaled up to a volume of 500 μl while keeping the relative concentrations of reagents constant, except that only 1/10 of the usual phagosomes concentration was used to a final OD₆₀₀ of 0.024. When we used phagosomes at the usual concentration that is used for the fusion assay, an aggregation and sedimentation was observed, making rheological measurements impossible.

After mixing, the fusion assay was kept on ice for 5 min, and then carefully pipetted into the pre-warmed measuring chamber. Measurements were taken every 3 minutes during the 80 minute incubation time.

Polymeric fluids (like F-actin solutions) are described by the frequency-dependent complex viscoelastic modulus, G * (ώ) , where G * (ω) = θ((ύ) +

which is composed of a 'real' part, G'(ω), and an 'imaginary' part, iG"(ω) (i denotes the imaginary unit, i = V-T , and ω is the angular frequency). The real part, the so-called storage modulus, G'(co), is a measure for the elastic component of the network. The imaginary part, the so called loss modulus G"(ω), is related to the viscosity of the network by G"(ω) = ωη(ω) . The complex quantity G * (ω) is characterized by its absolute value \G *\ = ΛIG'²+G"² and by the phase shift φ = arctan(G"/ G') . \G^*\ is determined by measuring the angular deflection of the disc (as described above) as a function of the shear force and is simply given by the ratio of the shear stress within the solution and the angular deflection angle, φ is the phase shift angle between the oscillatory force and the angular deflection of the disc. G ώ) and

G"(ω) are finally obtained from the two quantities \G \ and φ by using the relations

G' = *| COS φ and G" = \G *| sin φ , respectively.

The viscoelastic moduli G'(ω) and G"(ω) are complex functions of frequency which depend on the concentration of polymerized actin and on the degree of cross-linking between filaments in a complex manner. For a full characterization of the viscoelastic 1 G* I1 and φ or G'(ω) and G"(ω) over several frequency decades. However, numerous studies of purely entangled and cross-linked actin networks showed that the degree of actin polymerization can be well characterized measuring the above parameters at a single frequency of ω = 0.2rad/s (Sackmann, E. 1997. Viscoelasticity, rheology and molecular conformational dynamics of entangled and cross- linked actin networks. In Modern optics, electronics, and high precision techniques in cell biology. G. Isenberg, editor. Springer- Verlag, Heidelberg, pp213-259).

The phase shift, φ , depends on the fluidity of the sample and is φ = π/2 for a pure fluid (such as G-actin solution) and φ = 0 for a solid (which responds instantaneously to a step- wise force). Therefore the formation of an interconnected network of F-actin results in a decrease of the phase shift tgφ = θ'/θ and the build-up of a finite elastic modulus G'(ω). Based on previous studies, (Sackmann, 1997) measurements at a frequency of ω = 0.2rad/s are chosen for such measurements.

It should be noted that we also measured full frequency curves of \G | and φ in some cases to ensure that measurements at ω = 0.2rad/s are well suited to study the build up of viscoelastic networks in our cell extracts. However, since these measurements require 90 minutes for one lowest frequency applied they are not generally suited to study the time evolution of the generation of interconnected actin filaments.

1.6.4 • Confocal microscopy

In this second assay to analyse actin nucleation and aggregation (docking) of fluorescently- labelled phagosomes (and, if required, differentially-labelled endosomes) the organelles are mixed with macrophage cytosol containing rhodamine G actin and an ATP regenerating system. Confocal microscopy is used to analyse the nucleation of actin phagosomal membranes and the supra-organisation of the newly assembled actin. As a consequence of actin nucleation by one phagosome, other phagosomes or labelled endocytic organelles (or both) are able to bind to the F actin bundles and then move along them towards the nucleating phagosomes. These data, in conjunction with data from EM fusion assays have led to a model in which this aggregation of organelles along self-nucleated actin facilitates membrane fusion (Morten Egeberg, PhD Thesis, Department of Biology, University of Oslo, 2001).

Experimental details are as follows. For the light microscopy assay, HRP is labelled with Oregon Green. 20 mg HRP dissolved in 9.5 ml of 0.1 M NaHCO3/Na2CO3 (pH 9.0) is mixed with 0.14 mg Oregon Green 488- X, succinimidyl ester 6-isomer (Molecular Probes, Leiden, The Netherlands) in DMSO overnight at room temperature. This corresponds to a 5: 1 molar excess of Oregon Green over HRP. The unreacted active groups are quenched with 0.5 ml of 0.2M glycine (pH 8) and mixed for 30 min more. The mixture is dialyzed against several changes of PBS and then internalization medium.

The 2 hour phagosomes used for light microscopy are made by internalization of 1 μm blue fluorescent carboxylate-modified latex beads (Molecular Probes) covalently coupled with BSA according to the manufacturer's recommendations.

4 mg/ml J774 macrophage cytosol is preincubated for 20 min at 37°C with muscle G-actin labelled with 2 mM 5-carboxytetrametyl rhodamine (Molecular Probes), as described in Kellogg et al. (1988), Development, 103: 675-686). For this assay, phagosomes contained fluorescent beads whereas the organelles of the endocytic pathway contained fluorescent HRP (see above). The assay is otherwise identical to the biochemical in vitro fusion assay described in Jahraus et al. (1998), except from the omission of biotinylated insulin and the addition of 0.017% DABCO in water. Experiments may be performed either with phagosomes alone or with endocytic organelles alone. BSA-conjugated beads may be used as a negative control. The incubation may be performed in microscope chambers built from a glass microscope slide and two pieces of double-sided Scotch tape onto which a 15mm circular glass coverslip (Menzel, Braunschweig, Germany) is sealed, forming a 7μl chamber that is lOOμm deep. The microscope slides and coverslips were coated with 0.5% fish skin gelatin (Sigma) in H₂O and air-dried prior to use. The samples are perfused in before the chamber was sealed with a 1:1:1 mix of vaselin, lanolin and solid paraffin (VALAP) preheated to 70°C. The samples are then incubated up to 80 min at 37°C. Confocal images where subsequently acquired on a Zeiss LSM 510 confocal microscope with a 40x oil Plan- Neofluar lens (NA 1.30; Carl Zeiss. Inc).

For dual colour experiments (blue beads or Oregon Green-HRP and rhodamine-actin) excitation lines from two lasers may be used (364nm or 488nm respectively and 543nm) and emission monitored using the respective filters (385 - 470 or 505-530 respectively and 560 long-pass). For triple colour experiments (blue beads, Oregon Green-HRP and rhodamine-actin), excitation lines from three lasers are used (364nm, 488nm, 543nm) and emission was monitored using the respective filters (385 - 470, 505 - 530, 560 long-pass). Samples are line scanned (typically 8.96 μsec/pixel, ~ 125 seconds/frame) and power settings minimized to avoid photobleaching. All micrographs represent one optical section acquired 50μm above the microscope slide using a pinhole size of 78μm, giving an optical section of 1 μm. This is done to avoid artifacts from surface-induced actin polymerization.

In dual colour experiments with phagosomes alone, the blue beads were visualized as green to facilitate visualisation of the co-localization with the rhodamine-actin.

1.6.5 • Fret Assays

Fluorescence resonance energy transfer (FRET) is a powerful method for investigating interactions between two molecules that can be differentially labelled with two suitable fluorescence dyes. In principle, any two molecules which can be labelled with two such dyes will give a FRET signal when they physically interact (Clegg 1996 In: Fluorescence Imaging and Microscopy; Wang, and Herman, Eds ppl79-251; Wiley); Squire and Bastiaens, (1999, J.Microscopy 193: 36-49); Bastiaens and Squire, 1999, Trends In Cell. Biol. 9: 48-52). The phagosome system is potentially an ideal system for investigating interactions between two protein types, two lipid types or a protein and a lipid since the cytoplasmic aspect of isolated phagosomes is accessible to many labelling strategies. Suitable dye combinations for executing FRET are cy3 and cy5; cy 3 and Green Fluorescent Protein (GFP), oregon green and GFP, and different combinations of sulfonated rhodamines (for example, alexa dyes from Molecular Probes Inc.). Examples of accessible approaches with the phagosome system would be to use cy3 labelled PKC (commercially available) as a FRET donor, and a receptor (such as the Fc- or, complement-receptor which has GFP incorporated into the cytoplasmic domain), and over-express these components in macrophages. Provided the GFP receptors are incorporated into phagosomes any interactions with PKC could be detected at the single phagosome level by adding the labelled PKC and measuring the FRET signal. A related approach for visualising PKC activation in cells was described by Ng et al (1999 Science 283,2085-2089). The use of fluorescently-labelled lipids, such as phosphoinositides (for example, see Ozaki et al 2000; PNAS USA, 97: 11286-11291) or GFP-labelled PH domains that bind specifically to selected phosphoinositides (eg. Peyrollier et al 2000 Biochem J: 352, 617-622). This approach is potentially very powerful for directly analysing (and proving) interactions that are inferred from a functional map (such as those set out in Figures 1 and 2), since the method can be performed on the level of a single phagosome. This is especially important for microbial phagosomes, since it has been demonstrated many times that even in any one cell infected with a bacterium (for example, a pathogenic bacterium), the phagosomes always exist in at least two functional states - those that can fuse with endocytic organelles, and those that cannot. If such phagosomes are analysed biochemically, such functional differences would be averaged out.

Example 2: actin nucleation assay.

Phagosomes may be prepared from cells after any desired internalisation time into J774 cells. For the standard assay, 2hr phagosomes are used.

The phagosomes are then mixed with 2μm rhodamine-conjugated monomeric (g) actin and 6μm thymosin beta 4 (a 43 amino acid polypeptide, chemically synthesised) and an F-actin buffer.

An effector is then added at this time, as desired. The mixture is layered onto a glass slide chamber (between slide and cover-slip) and incubated for 15-20 minutes at room temperature. The percentage of phagosomes that have red actin dots, patches, or occasionally a halo, are then counted using immunofluorescence microscopy (total 100-150 phagosomes from two slides).

This procedure may repeated for any desired reaction conditions e.g. high versus low ATP or in the presence of a kinase inhibitor. For many experiments such as lipid additions (for example, sphingosine) the lipid is incubated with phagosomes for 15-60 mins at room temp. The phagosomes are then re-isolated by gradient flotation-the phagosomes themselves partition at the interface between 10 and 25% sucrose, while the free lipids float to the top of the gradient. The basic analysis is thus simple. In all the light microscopy assays described above, the percentage of phagosomes that nucleate actin, that bind to actin or that bind to microtubules is estimated to give an indication of the efficacy of the effects in the actin nucleation assay. For the actin nucleation, another parameter is sometimes useful, such as the amount of actin signal per phagosome. This can be done either by Adobe photoshop on fluorescence micrographs or by FACS analysis (Defacque et al. Cytometry, 2000,41,46-54).

The signalling scheme presented as Figure 2, was developed using data obtained from the LBP phagosome. Figure 2 shows a summary of known biochemical pathways that can be linked to the actin nucleation machinery that operates in vitro on the surface of a LBP (in the absence of added components, except actin, a buffer and ATP). Reagents affecting precise nodes of this network were tested at both the standard low ATP and, in most cases, with high ATP. In each case, the actin nucleation was compared with a control (standard) condition (untreated low ATP). Significant stimulation (+), or, inhibition (-), no effect (0) was scored and was mapped accordingly.

Not shown in this scheme is another useful experimental variation. A drug may be added to phagosomes in living macrophages and the phagosomes subsequently isolated and functionally tested. In the present case, this has been done using staurosporine, a PKC inhibitor. Whereas this drug was found to have no effect with standard LBP at low ATP, it was stimulating at high ATP. However, LBP isolated after staurosporine treatment of cells, were strongly inhibited.

In parallel with these functional assays, the effect of a reagent can be tested for its effects on either total lipids of the phagosomes, for example, using mass spectrometry, while the newly synthesised lipids can be detected by TLC, by HPLC or by mass spectrometry.

One example of these approaches is that sphingosine 1 phosphate (SIP) was predicted (from the literature) to activate DAG kinase and possibly expected also to influence the levels of PLP and/or PI₄₅P₂. Using P³² labelled ATP and TLC analyses, it was found that that sphingosine treatment at low levels of ATP increased the levels of labelled PA (the product of DAG kinase) seven-fold without an increase also in the level of PLP or PI₄₅P₂ that was synthesised in vitro (Figure 3G). In conjunction with bioinformatic approaches this system should produce the first large-scale membrane network of signalling events that is directly amenable to a sophisticated analysis. Such a system would be an ideal one to make highly sophisticated dry-tests. Example 3: Preliminary Investigation of Pathogen Phagosomes

The use of different Mycobacterial species has been established in the phagosome systems of the present invention. Sucrose gradient produces have been developed for producing enriched fractions of phagosomes containing the following bacteria: non-pathogenic M.smegmatis and the pathogenic M.avium (an increasing problem for Aids patients). Green fluorescent protein (GFP) labelled bacteria have been used to facilitate handling and analysis.

Materials and methods used in the below examples are as now follows. Preparation of a suspension of Mycobacterium smegmatis

M. smegmatis (strain mc2 155, a highly efficiently transformable strain derived from the American Type Culture Collection 607; Snapper et al, 1990 Mol. Microbiol. 4: 1911-1919) was transformed with an integrative plasmid (from the bacteriophage Ms6 site-specific recombination system, Anes et al. 1998. Microbiology 144:3397-3406) expressing GFP driven from a strong Ms6 promoter.

Bacterial cultures are grown on Myco broth (4.7 g Middlebrook's 7H9 broth Medium (Difco), 5 g of Nutrient broth (Difco), suplemented with 0.5% glucose and 0.05% Tween 80 per litre), until an exponential grown phase (OD600= 0.2), corresponding to approximately 10⁸ cells/ml.

After an overnight culture, cells are pelleted and washed twice in PBS pH 7.4. The cells are ressuspended in PBS such that they are 10⁹ or 10¹⁰ cells/ml (10-100 x concentrated). The suspension is treated for 2 min in a water-bath sonicator (room temperature) using 4x 30 sec pulses to disperse clumps. Then cells are passed through an 22-gauge needle to disrupt the rest of bacterial clumps. Before infection, residual bacterial aggregates were removed by low speed centrifugation (800 rpm) and a single cell suspension was verified by microscopy.

Infection of the J774 mouse macrophage line

The J774 cell line is cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% of FBS, 200 mM L-glutamine, streptomycin and ampicillin. For infection, cells are used after 2 days culture (almost presenting confluency). Before infection, cells are washed in PBS and DMEM medium (antibiotic free) is added. We use approximately 10 petri dishes of 20 mm per point phagosome preparation.

The ratio used for infection is usually 10-100: 1 bacteria:macrophage, depending on whether the phagosomes used are 24h- 8hr phagosomes or 3h-12hr phagosomes, respectively. Two hours post-infection, cells are washed again with PBS to remove free bacteria and then cells cultured in DMEM with 16μg/ml gentamycin (which enters into the cell) to prevent any remaining extracellular bacteria multiplying.

Phagosome isolation At a precise time post-infection, macrophage cells are washed in PBS, scraped and pelleted. Then the whole cell preparation is re-suspended in 1.5 ml of HB buffer (0.25 M sucrose, 3 mM Imidazole, pH 7.4) containing appropriate amounts of protease inhibitors (leupeptin, aprotinin, pepstatin A, l-chloro-3-tosylamido-7-amino-2-heptanone and DTT). Macrophages are lysed by approximately 30 passages through a syringe apparatus fitted with 22-gauge needles. For post nuclear supernatant separation the macrophage lysate is subjected to velocity sedimentation (200xg, 30 min, 4°C) through 15% (w/w) sucrose overlaid on a 25% (w/w) sucrose cushion. A second step gradient allows the separation of phagosomes from other organelles. The sucrose gradient is performed in the presence of protease inhibitors with a 1ml 62%, a 2 ml 50%, a 6 ml 37% (w/w) step cushion, and at the top the sample obtained on the first gradient. Centrifugation is performed at lOO.OOOxg during lh at 4°C in a SW40 rotor. We are able to obtain three bands, one in the top of the 37% gradient with the majority of the organelles, another in the interface between 50 % and 37%, with the bacterial phagosomes, and the last at the 62%/50% interface with a few larger phagosomes and some nuclei.

The actin nucleation assay developed for latex bead phagosomes (LBP) was directly and easily applied, taking less than two weeks of man time. A fraction of phagosomes that was significantly higher (30-40%) than that routinely seen with LBP (10-20%) were found to nucleate actin and the amount of actin assembled per phagosome was higher than what was observed with LBP. Moreover, many effectors of actin nucleation in the LBP system could be tested with the M. megmatis phagosomes. The following conditions were found to be identical in their effects, described simply as "stimulation", (+), inhibition (-) or no effect (0), referring to the effect of an effector relative to the standard control preparation (naked latex bead phagosomes of 2 hours). In both systems, the actin nucleation activity behaved accordingly, as disclosed below:

1. 2hr phagosomes (high), 12hr phagosomes (low), 24-36hr phagosomes (high) 2. low(5-10μm) ATP- (high); high (lmM)ATP-(low)

3. at low ATP sphingosine IP inhibits, therefore scored as (-)

4. at high ATP sphingosine IP stimulates, therefore scored as (+)

5. at low ATP sphingosine stimulates, therefore scored as (+)

6. at high ATP sphingosine inhibits, therefore scored as (-) 7. at low ATP staurosporine (PKC inhibitor) has no effect (0) but stimulates at high

ATP (+). Since inhibition of PKC stimulates actin at high ATP it is assumed that PKC normally inhibits and is scored as (-). In contrast with these results with M. smegmatis, preliminary observations using M.avium phagosomes show the following behaviour:

When the bacteria are alive within the phagosomes, actin nucleation is inhibited. However, dead bacteria-containing phagosomes nucleate actin well. These data collectively make the case that the technology developed for the LBP system can be successfully and easily transferred for use with phagosomes enclosing potentially any microbial phagosome, provided that the particle can be isolated intact and identified as a discrete particle.

The results obtained from the work with M. smegmatis phagosomes may be summarised as follows.

Two investigators tested (in a blind fashion) specific effectors that have altered LBP signalling (see Figure 2). All of the nodes indicated in the figure behaved identically in terms of their response to what was observed with LBP. This demonstrates that the actin nucleation system developed for the LBP is easily adaptable for use with phagosomes containing a bacterium. Preliminary data also show that M. smegmatis phagosomes can bind to microtubules.

In the following Examples 4-9, the methods and techniques described below were used in conjunction with, and in addition to, those described above:

Bacterial and Phacocytic cell lines and growth conditions

The mouse macrophage cell line J774A.1 (Ralph P., Nature, 1975, 257(5525), 393-4) was maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), 1% glutamine, 100 units/ml penicillin, 100 μg/ml streptomycin (referred to as complete medium) at 37°C and 8% CO2. For infection experiments, the cells were maintained in antibiotic-free complete medium for 48 h and grown to confluency in petri dishes or seeded on glass coverslips and grown to semiconfluency.

The field isolate M. paratuberculosis 6783 used in these examples has been described (Jark et al, Net. Microbiol., 1997, 57(2-3), 189-98). M. smegmatis mc²155 (ATCC 19420) was kindly provided by G. Auling (Institut fuer Microbiologie, University of Hannover). M. gordonae (ATCC 14470) and M. avium (ATCC 25291) were obtained from the DSMZ (German Collection of Microorganisms and Cell Cultures, Braunschweig). All mycobacteria were grown in Middlebrock 7H9 medium supplemented with 10% OADC, 0.5% glycerol and 0.05% Tween 80 at 37°C. Homogenisation of the mycobacteria was done as follows. The suspensions were vortexed in the presence of glass beads (3 mm of diameter) for 5 min and incubated for another 2 min in a sonic waterbath (GFL) at high intensity, then centrifuged for 5 min at 50 x g. Microscopy confirmed that the resulting supernatant contained mainly single bacteria. Viability of bacteria was examined using the RαcLight system (Molecular Probes). Bacterial suspensions from M. ptb contained between 70% to 80% live bacteria, those from M. avium, M. gordonae, and M. smegmatis approximately 95%. Heat-inactivation of bacteria was achieved by incubating bacterial suspensions with an optical density at 660 nm (OD₆₆o) of 1.0 for 15 min at 85°C (Zurbrick and Czubrynski, 1987). Loss of viability was confirmed by serial platings and colony growth.

Fluorescence labelling of mycobacteria

Mycobacteria were harvested as described above and diluted with ice cold PBS to an OD₆₆₀ of 1.0. To 1 ml of the cell suspension, 100 μl sulpho-NHS-biotin (5 mg/ml DMSO) was added and incubated on a shaker for 20 min at 4°C. Then, the bacteria were washed three times with 1 ml of cold antibiotic-free complete medium and centrifuged for 5 min at 13,000 rpm using a microcentrifuge. To determine labelling efficiency, labelled and non- labelled bacteria were treated with 2% paraformaldehyde for 15 min at room temperature, washed three times with PBS, and incubated in PBS containing 5% FCS for 20 min to block non-specific binding sites. After three washing steps with PBS, bacteria were incubated with 100 μl of PBS containing 1 μg/ml TexasRed-streptavidin for 20 min. After additional three washes with PBS the bacterial pellet was resuspended in 1 ml PBS and 10 μl of the mycobacterial suspension was mounted on a glass slide by heat fixation. Samples were then examined microscopically using a Leica DML microscope with an N2.1 filtersystem (Leica, Bensheim, germany). Approximately 200 bacteria were counted. The labelling efficiency was between 90% and 95%.

Viability assessment of intracellular mycobacteria

To determine mycobacterial viability during infections, J774 macrophages were infected as described above. Monolayers were washed two times with PBS and scraped off the plates in 1 ml of 1% Non-idet P40 in PBS. J774 cells were disrupted by 10 passages through a 24 gauge needle. Tenfold serial dilutions of the homogenates were prepared with PBS and 100 μl each plated on Middlebrook 7H10 agar plates with 10% Middlebrook OADC enrichment (Difco) (for M. ptb the medium was supplemented with 2 mg/1 of mycobactine J). After incubation for 2-3 weeks (2-3 months for M. ptb) at 37°C, colony forming units (CFU) were counted. Confocal laser scanning microscopy

For confocal laser scanning microscopy J774 cells grown on glass coverslips were washed two times with PBS and then incubated with antibiotic-free complete medium containing 1 mg/ml FITC-dextran for 30 min at 37°C and 5% CO . The cells were washed three times with warm PBS and incubated with antibiotic-free complete medium containing sulpho- NHS-FITC labelled mycobacteria (M.avium, M. tuberculosis or M. smegmatis) of an OD₆₆₀ of 0.1 at 37°C and 8% CO2. After 24h cells were fixed with 3% paraformaldehyde for 15 min at room temperature (the coverslips were washed three times with ice cold PBS and before fixing). Following, cells were treated with 0.1% Triton X100 for 5 min. Cells were washed two times with DMEM and incubated in a blocking solution (containing PBS 2% FCS, 1% BSA and Rhodamine-phalloidin 0.1 mg/ml diluted 1:300) for 20 min at room temperature. Then, PBS was replaced by 1 μg/ml TexasRed-streptavidin in PBS. After an additional 30 min, coverslips were washed with PBS and then mounted upside down on slides using 10 μl of Mowiol^® containing 10 mg/ml freshly prepared DABCO^®. The slides were then sealed with nail polish and stored at 4°C until further examination. Fluorescence images were collected using a Leica DM-IRBE inverted confocal laser scanning microscope equipped with an immersion oil Plan-Apo lOOx/1.4 n.a. lens and an argon/krypton laser (Leica). Settings allowed simultaneous co-localisation of FLTC- dextran-labelled endosomes and TexasRed labelled mycobacteria. Computer acquisition and analysis of pictures was done using the Leica confocal laser scanning standard software and Adobe Photoshop (Adobe Systems Incorporated, Amsterdam, Holland). Twenty fields were examined per slide, and representative pictures were taken.

The TexasRed fluorescence of the mycobacteria was analysed for co-localisation with the FITC-dextran fluorescence by overlaying the green and red pictures resulting in a yellow signal. Red fluorescent signals and yellow fluorescent signals which could be identified as single bacteria in terms of morphology were counted to a total of 200 events. Analysis of phagosome-lysosome fusion

Analysis of phagosome-lysosome fusion was performed as described by Oh Y.K. and Straubinger R.M., Infect. Immun., 1996, 64(1), 319-25. Cells were washed two times with PBS and incubated in the presence of 200μM calcein in antibiotic-free complete medium for 24 hours. Following, cells were washed two times in PBS and incubated in antibiotic- free complete medium containing sulpho-NHS -labelled mycobacteria as described above. After 1 h and 5 h coverslips were treated and analysed as described above. For the assessment of Lampl and Lamp2 in mycobacteria containing phagosomes (MCP) J774 cells were infected as described above and analysed for co-localisation of mycobacteria phagosomes with late endosomal and lysosomal markers using monoclonal antibodies from rat against murine Lampl (clone 1D4B) and Lamp2 (clone ABL-93) at a dilution of 1:50 in PBS following fixation with 2% paraformaldehyde and permeabilisation with 0.1% Triton X100. Secondary FLTC-conjugated goat anti-rat IgG antibody was used at a dilution of 1:300 in PBS. Samples were analysed with a Leica confocal scanning microscope (model SP2) with a xlOO oil immersion objective and the appropriate filter set. Cells and phagosomes were clearly identifiable under DIC optics. Images were digitally acquired and processed using Adobe Photoshop™. Phagosomes from at least 20 cells per time point were counted.

Isolation of Latex Bead Phagosomes (LBP) Isolation of LBP was performed as described in (Defaque et al., 2000)

Isolation of Mycobacteria Containing Phagosomes (MCP)

For isolation of MCP, 20x20cm petridishes with a confluent layer of J774 cells were infected with 50 ml DMEM containing either M. avium, M. paratuberculosis, or M. smegmatis (live or heat killed) either non-labelled or labelled with NHS -fluorescein at an OD₆₆₀ of 0.1 for 1 hour at 5 % CO₂ at 37°C. Non-internalised bacteria were removed by intensive washings with PBS and further incubated for another 1 or 24 hours. Then the medium was removed and cells were washed again with PBS. Cells were harvested by scraping in 20 ml PBS and collected in 50 ml Falcon tubes. Cells were centrifuged at lOOOx g for 10 min, resuspended in 5 ml of HB-buffer and pelleted again. Pellets were resuspended in a final volume of 1.5 ml of HB buffer and lysed using a 1ml syringe fitted to a 22-gauge needle. Nuclei and intact cells were removed by gentle centrifugation at 800 x g at 4°C and supernatants were layered on top of a discontinuous sucrose gradient consisting of 2ml 50% sucrose, 4ml 37% sucrose and 4 ml 25 % sucrose and centrifuged at 24k rpm using a SW40 rotor for 60 min at 4°C. The band forming between the 50% and 37% sucrose contains the mycobacteria-containing phagosomes and can be collected in approximately 200μl with a syringe by penetrating the tube wall with a 22 gauge needle. The phagosome suspension is diluted with HB to a final volume of 2 ml and layered on top of 1 ml 50% sucrose and centrifuged again in a SW55 rotor at 36k rpm for 30 min at 4°C. Mycobacteria- containing phagosomes are seen as a band above the 50% sucrose and are collected as described above. Phagosomes are aliquoted in 20μl portions and stored at -20°C until further use. In order to check the proportion of intact phagosomes, phagosomes containing dead mycobacteria were incubated with rabbit anti-mycobacterial antiserum and detected with a goat anti rabbit antibody. The ratio between phagosomes stained (not intact) and non-labelled phagosomes (intact) was determined and only preparations with above 65% intact phagosomes were used for further experiments.

Actin nucleation assay by fluorescence microscopy This assay was described above and in detail by Defacque et al (Cytometry, 2000, 41(1), 46-54). The percentage of rhodamine actin-labelled phagosomes was determined using a Zeiss Axioscope microscope. An advantage of most of the described assays is that they utilise light microscopy, and it is thus not necessary to have pure fractions of the phagosomes since the organelles are directly sampled by the microscope. Phagosome pH monitoring

Lyso-tracker was performed as described by Via et al. (J. Cell. Sci., 1998, 1 L1 (Pt 7), 897- 905). The acidotrophic dye Lysotracker Red DND-99 (Molecular probes, Eugene, OR) was added to the cells (final concentration of 50nM) for 2h before infection. Infection was carried out as described above. After uningested mycobacteria removal, additional Lysotracker was added to each well. After 24 h post-infection macrophages were washed and mounted as described before. The fraction of green fluorescent phagosomes that colocalized with lysotracker was determined by analysing more than 100 phagosomes (combined from at last five random fields) from three separate monolayers. Measurement of phagosomal pH was done using FITC labelled mycobacteria to infect cells. After 24 h of infection cells were harvested and measured in quartz kuvettes fluorometrically and compared with a standard curve derived from infected cells in nigericin buffers ranging from pH5-7 as described in Sturgill-Koszycki et al. (Science, 1994, 263(5147), 678-81). Thin Layer Chromatography (TLC)

The lipids were separated by thin layer chromatography (TLC) on Silica Gel G60 plates (pretreated with 1% potassium oxalate/2 mM EDTA in methanol/water (1: 1, v/v)) using a solvent mixture of chloroform/acetone/methanol/glacial acetic acid/water (80:30:26:24: 14, v/v/v/v/v) (Norris and Majerus, 1994). Phospholipid standards (PI, PI(4)P, PI(4,5)P₂, PI(3,4,5)P₃ and PA) were stained with iodine. The quantitation of ³²P-labelled phospholipids separated by TLC was performed using a Fujifilm Imaging Plate and Fujifilm Fluorescent Image Analyzer FLA-2000 equipment.

In all experiments described herein, the errors reported are the standard deviations from at least three separate experiments. For lipid treatments, the stock lipids were dissolved as follows: Arachidonic acid (AA) and ceramide (Cer) in water (lmg/ml), sphingosine (Sph) in ethanol (5mM), spingosine-1 -phosphate (SIP) in warm methanol (5mM) and phosphatidyl-inositol-di-phosphate (Pi(₄₎₅₎P ) in ethanol (5 mM). For performing either the actin nucleation assay or survival experiments the phagosomes were mixed with the respective lipid at the following final concentrations: 250 μM AA (Sigma), 100 μM Sph and SIP (Calbiochem), ImM ceramide (Sigma) and 50-150 μM Pi(₄,₅₎P₂ (Calbiochem). To analyse the effect of a lipid, results were routinely compared to controls containing only the indicated solvent as a reference control.

Example 4: Pathogenic mycobacterial phagosomes do not nucleate actin. Further experiments have now been performed in support of the conclusion that a host- evasion mechanism of pathogenic mycobacteria may be the prevention of phagosome actin nucleation, apparent from Example 3 above.

In order to establish the LBP actin assay for pathogenic and non-pathogenic mycobacteria, we first infected J774 cells with the non-pathogenic M. smegmatis, as described in Example 3. In these cells the majority of M. smegmatis are killed within 24h and all of them within 48h of infection (Kuehnel et al, Cell Microbiol., 2001, 3, 551-566). GFP-labelled M. smegmatis were fed to J774 cells and after 2, 12, 24 or 48 hours and the phagosomes were isolated by two sequential gradient steps. In parallel, phagosomes containing killed M.avium were also prepared. As shown in Fig 5A, the live M. smegmatis phagosomes nucleated actin in vitro in a manner very similar to that seen with LBP, although more actin was evident per phagosome, and a higher percentage of labelled phagosomes was seen than with LBP (Fig 5C). There was often a significant concentration of nucleation sites at those membrane parts adjacent to the bacterial poles (Fig 5A). As for the LBP, no evidence for actin "comet" or "rocket" formation was observed (in the absence of cytosol or GTP). The phagosomes of live and killed M. smegmatis behaved similarly at high and low ATP (Fig 5B). Remarkably, for the M. smegmatis phagosomes, a number of the effectors tested behaved identically to their effects on the LBP-actin system (Fig 5B), with the exception of Ceramide which reacted in the opposite fashion to that observed in the LBP.

With the phagosomes containing killed M.avium, the morphological appearance of the actin was similar and the response to the effectors tested was the same as that for M. smegmatis phagosomes and LBP (Figs 5B and 5C). We next tested how the system would behave when a live pathogenic bacterium was enclosed within the phagosomes. The pathogenic M. avium bacterium can survive and replicate within phagosomes and eventually kill their macrophage host, including J774 cells. This is consistent with our findings that, in contrast to the phagosomes containing non-pathogenic M. smegmatis, those with live M.avium failed to nucleate actin significantly at either low or high ATP (Fig 6).

Example 5: Certain lipids stimulate LBP and MCP actin nucleation in vitro

When these organelles were co-incubated with physiological levels of ATP (l-5mM) the nucleation reaction was progressively inhibited (Fig.3C). Also when the LBP were pre- treated with high ATP, re-purified and then tested, their ability to nucleate actin was similarly impaired (Fig 3C; inset), presumably due to phosphorylation events (Emans et al., FEBS Letters, 1996, 398, 37-42. In a recent publication (Jahraus et al, Mol. Cell Biol., 2001, L2, 155-70) we showed that J774 cytosolic extracts are also strongly inhibited in their ability to nucleate actin (a membrane-independent process) at ImM ATP. However, at low ATP (0.2mM), the inhibition could be overcome. Our working hypothesis is that at low ATP (ischaemic conditions), both membrane and cytoplasmic actin nucleation may be constitutively "on", while at normal ATP levels both systems are "off, unless actively turned on by a signalling cascade.

Evidence is now accumulating that the sphingomyelin- and cholesterol- enriched membrane "rafts" prepared by cold Triton-X-100 treatment of whole cells, can be enriched in

PI(4,5)P₂, ezrin, and CD44 as well as in signalling molecules (Baird and Holowka, Annu.

Rev. Biophys. Biomol. Struct., 1996, 25, 79-112; Toomre and Simons, Nat. Rev. Cell Biol., 2000, 1, 31-9; Caroni, EMBO J., 2001, 20(16), 4332-6). Moreover, what are proposed to be the equivalent, specialised membrane sub-domains in cells may be favoured regions for actin nucleation (Rozelle et al., J. Cell Biol., 2001, 154(5). 1007-17; Caroni, EMBO J., 2001, 20(16), 4332-6). Recent biochemical data argue that rafts are also present on LBP (Dermine et al, J. Biol. Chem., 2001, 276(21). 18507-12). Sphingomyelin can be broken down to ceramide, which can be further degraded to sphingosine (Sp). In the presence of ATP, sphingosine kinase can phosphorylate Sp to make sphingosine- 1 -phosphate (SIP) from Sp. All these molecules have potent and highly complex effects on whole cells, that includes dramatic effects (both inhibitory and stimulatory) on the actin cytoskeleton (Bornfeldt et al, J. Cell Biol., 1995, 130(1); Spiegel and Milstein, FEBS Lett., 2000, 476, 55-7; Hannun et al, Biochemistry, 2001, 40(16), 4893-903). We therefore tested these lipids in our actin nucleation assay, both at low (0.2mM) and high (5mM) ATP.

At low ATP, co-incubation (Fig 3D) or pre-incubation (data not shown) of LBP with lOOnM SIP inhibited actin nucleation on LBP. When Sp kinase was inhibited, the level of actin nucleation was routinely elevated, suggesting that under standard conditions, a low- level synthesis of SIP leads to a partial inhibition of the system (Fig 3D). In contrast, at high ATP, SIP stimulated actin nucleation, whereas the Sp kinase inhibitor had no effect relative to the control (Fig 3D). In many cell systems, SIP behaves in an opposite fashion to its precursor Sp (Bornfeldt et al, J. Cell Biol., 1995, 130(1); Hannun et al, Biochemistry, 2001, 40(16), 4893-903. Review). In excellent agreement with this notion, treatment of LBP with lOOnM Sp led to effects that were the mirror image of the SIP results. Here, Sp stimulated actin nucleation at low ATP, but inhibited at high ATP (Fig 3F). The precursor of Sp, ceramide (a mixture of brain ceramides) behaved like Sp, activating actin nucleation at low ATP but inhibiting it at high ATP (Fig 3F). Treatment of J774 cells, that had internalised beads with SIP or the Sp kinase inhibitor before isolating LBP also led to significant alterations in their ability to nucleate actin (data not shown).

In order to show that these lipids indeed affect lipid metabolism in the LBP membrane, we followed the in vitro incorporation of P -labelled ATP into phagosomal lipids by thin layer chromatography (TLC), using a system already established (Defacque et al 2001, submitted). At low ATP (20-200mM unlabeled ATP with 10 μCi of P³², γ-labelled ATP), and in the absence of effectors, we can routinely resolve five lipids that incorporate phosphate label (SIP, two variants of PI(4)P having different hydrocarbon moieties, PI(4,5)P₂, a low level of PI(3,4,5)P₃ under some conditions, and a low level of phosphatidic acid (PA); Defacque et al 2001, submitted). As shown in Fig 3F, the addition of lOOnM Sp to LBP led to a significant increase in the synthesis of SIP as well as in PA, without significantly affecting the synthesis of PIPs. In contrast, the addition of lOOnM SIP did not affect the levels of PA, although they did increase the levels of some PIPs (see Fig 3F, figure legend). These data suggest that Sp (which stimulates actin nucleation at low ATP) can stimulate both Sp kinase, which leads to synthesis of SIP, as well as DAG kinase, which synthesises PA from DAG. However, neither of these increases by themselves can be the direct cause of the increase in actin nucleation by Sp at low ATP, since SIP by itself inhibits the process under these conditions (Fig 3D), while PA has no effect. These data can only be interpreted in the context of complex membrane signalling networks. It should be noted that, although DAG kinase has a Km for ATP around ImM, it nevertheless seems to be active under low ATP conditions.

The above results with ceramide, Sp and SIP prompted us to test a wide spectrum of reagents that affect signalling proteins and lipids that are known to be metabolically linked to these lipids. A summary of the data is given in Fig 4. In conjunction with Example 3, these results argue strongly that a large network of membrane signalling components that regulate membrane-bound actin nucleation can operate in vitro on LBP, as long as ATP is present. Both PKC and PKA appear to regulate this process. As is evident in Fig 4, five different nodes in the system were switched in their behaviour at low (ischemic) levels of ATP versus high (physiological) levels of ATP.

However, at low ATP AA, ceramide, Sp and PI(4,5)P₂ could significantly increase the percentage of M.avium phagosomes that nucleated actin, while A A, ceramide and S IP stimulated the process at high ATP (Fig. 6). In preliminary experiments actin nucleation was also successfully reconstituted from phagosomes containing M.paratuberculosis, an important disease of ruminants. Also in this example, the phagosomes enclosing the killed pathogens nucleated actin well, but those enclosing live M.paratuberculosis hardly nucleated actin at all (data not shown). These data show that the incorporation of specific lipids into the phagosomal membrane can activate a process that had been actively supressed by the pathogen. Example 6: Certain lipids can switch on mycobacterial phagosome actin nucleation and fusion in infected cells.

In the LBP, as in all other membrane systems, the actin that is nucleated by the membrane assembles into filaments that grow outwards from the organelle by insertion of monomers at the fast-growing (barbed) ends of filaments that are adjacent to the membrane surface (Defacque et al, EMBO J., 2000, 19(2), 199-212; Tilney L.G., in International Cell Biology, Rockefeller University Press, 1976, p388-402). A consequence of this polarity, that is universally seen on membranes (Tilney L.G., in International Cell Biology, Rockefeller University Press, 1976, p388-402), is that actin filaments and bundles emanate away from the membrane and any potential fusion partner organelle that has an appropriately bound myosin can bind to, and then move vectorially along the actin towards the actin barbed ends, and thus towards the nucleating organelle (Egeberg et α/-in prep). All myosins, except class VI (Wells et al 1999; this myosin is also present on LBP; Unpublished data), invariably move in this direction along actin in the presence of ATP. This "actin track" model predicts that an increase in membrane-dependent actin nucleation can lead to an increase in fusion. A further prediction of relevance here is that any effector that could switch on actin nucleation by the phagosomes of pathogenic mycobacteria might lead to an increase in membrane fusion, and consequently, in the killing of these pathogens. We followed up on these predictions at the cellular level, where it was also possible to focus on cells infected with virulent and attenuated strains of the human pathogenic M.tb.

As shown above, the live pathogen phagosomes were very poor at nucleating actin in vitro, whereas those with the killed pathogens showed a strong signal. As shown in Fig 8Ai and ii, the same pattern was observed in vivo: within J774 cells, the vast majority of phagosomes enclosing killed M.avium or Mtb, or live M. megmatis were strongly labelled for actin (Fig 8Aii). In contrast, in the cells infected with the live pathogens, the phagosomes were essentially devoid of labelling for actin (Fig 8Ai).

At the next level of the analyses we asked whether any of the lipids that would activate actin nucleation on live M.avium phagosomes in vitro could also have the same effect in infected cells. According to the actin track model an enhancement of actin around the phagosomes of the live pathogens by the "positive effector" lipids would be expected to enhance the level of the fusion of these phagosomes with late endocytic organelles. As shown in Fig 8B and 8C, treatment of J774 cells infected with live (or killed) M.avium, M. smegmatis or live M.tb with AA, ceramide, Sp, SIP or Sphm led to a significant increase in the percentage of phagosomes that labelled with rhodamine-phalloidin. As shown in Fig 9A and B, by use of an EM colloidal gold content-mixing assay, the lipids were found also to significantly increase the extent of fusion of phagosomes containing pathogenic mycobacteria with late endocytic organelles. However, ceramide treatment had no effect on the live pathogen phagosomes, although it did significantly increase the fusion of phagosomes with the killed pathogens (Fig 9B). Treatment of the infected cells with the "negative effector" lipid DAG had no effect on this fusion process (Fig 9A,B). It should be noted that in cells, as in vitro, the effective concentration of these lipids varies enormously; whereas SIP and Sp are effective at lOOnM the other lipids require 50-150 μM to be effective (except ceramide, which requires ImM). The low effective concentrations of Sp and SIP make these lipids particularly attractive for therapeutic testing in animal infection models. At the concentrations used in our experiments these lipids had no adverse effects on macrophage growth or viability (Results not shown).

Example 7: Lipids activate lowering of phagosomal pH

Whereas the phagosomes containing non-pathogenic mycobacteria or killed pathogenic mycobacteria acidify in infected cells to values around pH5, those of pathogenic mycobacteria acidify only to pH ~ 6.3 (Clemens, D. L., Trends Microbiol., 1996, 4, 113- 118; Russell D.G., Nature Rev. Mol. Cell. Biol. 2001, 12, 569-577; Kuehnel et al, Cell Microbiol., 2001, 3, 551-566). We have recently shown that, in addition to the phagosomes, the pH of the whole endocytic pathway is raised in cells infected with pathogenic mycobacteria (Kuehnel et al., Cell Microbiol., 2001, 3, 551-566).

For estimating phagosome pH, we first used lyso-tracker red which accumulates in low pH organelles. As shown in Fig lOAi the phagosomes of killed M.tb accumulated significant amounts of the dye (seen as a matching of the red dye with the green bacteria), indicating a low pH, whereas those enclosing the live pathogen failed to label (Fig lOAii). A quantitation of the percentage of phagosomes that labelled with this dye is shown for M. tuberculosis (Fig 10B) and for M.avium (Fig 10C). As shown in figure 10B, in infected macrophages, AA, ceramide, Sp and Sphm significantly increased the fraction of phagosomes containing live M.tuberculosis that labelled with the lyso-tracker dye, indicating a general lowering of pH. Fig 10C shows that the same lipids, as well as PI(4,5)P2 had the same effect on the phagosomes containing live M.avium. In order to provide a quantitative estimate of the pH live M.avium were coupled to FITC before internalising the bacteria. Using this approach, the pH of the phagosomal lumen can be estimated by analysing the ratio of the fluorescence emission at two different wavelengths using the procedure described above. As shown in Figs 10B and C the phagosomes enclosing live M.avium or M. tuberculosis had an estimated pH of 6.3 in untreated cells, as expected. In contrast, in the equivalent phagosomes in the cells treated with AA or ceramide the pH was reduced to about 5, the value expected when the phagosomes have been allowed to mature fully (Clemens, D. L., Trends Microbiol., 1996, 4, 113-118; Kuehnel et al, Cell Microbiol., 2001, 3, 551-566).

Example 8: Certain lipids stimulate pathogen killing.

We have described above the treatment of infected macrophages with lipids that enhance actin nucleation, fusion of phagosomes with late endocytic organelles and phagosome acidification. The Applicant now importantly also shows that treatment of infected macrophages with these lipids leads to an increase in pathogen killing. The treatment of cells infected with live M.avium (Figure 11 A) or with an attenuated or virulent strain of M.tb (Figure 11B, for virulent strain, data not shown) with AA, ceramide, Sp, or sphm led to a significant decrease in the viability of the pathogens. This increased killing was not restricted to pathogens since the killing of the non-pathogenic M. smegmatis was also significantly enhanced by arachidonic acid and all the tested sphingolipids. Whereas M. megmatis was normally completely killed by cells by 48h after infection, treatment of cells with any of the lipids killed all the bacteria within 24h (data not shown). In contrast, treatment of cells infected with M.avium or M.tb with the negative actin effector, DAG had no effect on pathogen killing (Fig 11A and 11B). At the concentrations used in our experiments these lipids had no adverse effects on macrophage growth or viability (data not shown).

In combination with the above examples, these data show convincingly that the treatment of infected cells with simple lipids can lead to activation of many phagosome functions leading to a significant increase in pathogen mortality in infected macrophages. It should be noted that mycobacteria are protected by an unusually thick and impervious cell wall (Barry, 2001). It is thus predicted that many other pathogens living in phagosomes may be even more susceptible to an activation of the phagosomal killing process. The following example (Example 9) demonstrates the feasibility of an approach termed 'elementary mode analysis' for the analysis of complex signalling networks. This method describes a complete set of the distinct biochemical modes of a network that can potentially operate in a given system as linear combinations of its "elementary modes".

Any metabolic path which can exist in steady state that balances all internal metabolites and that cannot be further decomposed is called an "elementary mode". It provides a mathematical tool to define and comprehensibly describe all metabolic routes that are both stoichiometrically and thermodynamically feasible for a group of enzymes. It has already been shown that this approach is useful for the analysis of complex networks, for instance to analyse the interplay between glycolysis and the pentose phosphate pathway in a well defined and predictable manner, and in agreement with experimental observations, both in eukaryotic and prokaryotic systems (Dandekar et al., Biochemical Journal, 1999, 343, 115- 124; Schuster et al., Nature Biotechnology, 2000, 18, 326-332). The examples below represent the first use of this approach to study signalling networks operating within a biological membrane.

The elementary mode analysis of the reaction scheme shown in Fig 12 yields 128 elementary modes (abbreviations used are those listed in the description of the figures section above):

1 CDPinotra CDPsynth irreversible 2 Actindepoly irreversible 3 Lipase -1 DAGcholT reversible 4 Lipase -1 PAP reversible 5 -1 DAGcholT PAP reversible 6 - 1 Lipase PIdiase irreversible 7 -1 Lipase -1 ERMform -1 ERMfform PI45diase irreversible 8 -1 DAGcholT PIdiase irreversible

-1 ERMform -1 ERMfform -1 DAGcholT PI45diase irreversible

10 -1 PAP PIdiase irreversible 11 -1 ERMform -1 ERMfform -1 PAP PI45diase irreversible

12 ERMform ERMfform -2 SPK PI4kin PI4P5kin reversible 13 -1 Lipase -1 DAGkin SPK reversible -1 DAGkin SPK -1 DAGcholT reversible

-1 DAGkin SPK -1 PAP reversible

-2 Lipase ERMform ERMfform -2 DAGkin PI4kin PI4P5kin reversible

ERMform ERMfform -2 DAGkin PI4kin PI4P5kin -2 DAGcholT reversible

ERMform ERMfform -2 DAGkin PI4kin PI4P5kin -2 PAP reversible

-1 SPK Actinnucl irreversible

-1 Lipase -2 SPK PI4kin PI4P5kin PI45diase irreversible

DAGkin -1 SPK PIdiase irreversible : -1 ERMform -1 ERMfform DAGkin -1 SPK PI45diase irreversible

DAGkin -3 SPK PI4kin PI4P5kin PI45diase irreversible : -2 SPK PI4kin PI4P5kin -1 DAGcholT PI45diase irreversible

-2 SPK PI4kin PI4P5kin -1 PAP PI45diase irreversible

-1 ERMform -1 ERMfform -1 PI4kin -1 PI4P5kin 2 Actinnucl irreversible : -1 ERMform -1 ERMfform 2 DAGkin -1 PI4kin -1 PI4P5kin 2 PIdiase irreversible

-3 ERMform -3 ERMfform 2 DAGkin -1 PI4kin -1 PI4P5kin 2 PI45diase irreversible

-1 Lipase -1 DAGkin Actinnucl irreversible

-3 Lipase -2 DAGkin PI4kin PI4P5kin PI45diase irreversible

-1 DAGkin -1 DAGcholT Actinnucl irreversible : -2 DAGkin PI4kin PI4P5kin -3 DAGcholT PI45diase irreversible

-1 DAGkin -1 PAP Actinnucl irreversible : -2 DAGkin PI4kin PI4P5kin -3 PAP PI45diase irreversible : -1 ceramidse -1 Sphmydias -1 CholinPtf reversible

-1 Lipase PLC -1 ceramidse -1 Sphmydias reversible

PLC -1 ceramidse -1 Sphmydias -1 DAGcholT reversible

PLC -1 ceramidse -1 Sphmydias -1 PAP reversible monoAGcholinPCtf -1 ceramidse -1 Sphmydias Cholinkin PLA2 reversible

-1 PLC ceramidse Sphmydias -1 DAGkin SPK reversible

-2 PLC ERMform ERMfform 2 ceramidse 2 Sphmydias -2 DAGkin PI4kin PI4P5kin reversible

-1 Lipase PLC CholinPtf reversible

PLC CholinPtf -1 DAGcholT reversible

PLC CholinPtf -1 PAP reversible monoAGcholinPCtf CholinPtf Cholinkin PLA2 reversible

-1 PLC -1 CholinPtf -1 DAGkin SPK reversible

-2 PLC ERMform ERMfform -2 CholinPtf -2 DAGkin PI4kin PI4P5kin reversible monoAGcholinPCtf Lipase -1 PLC Cholinkin PLA2 reversible monoAGcholinPCtf -1 PLC Cholinkin DAGcholT PLA2 reversible monoAGcholinPCtf -1 PLC Cholinkin PLA2 PAP reversible

-1 PLC ceramidse Sphmydias PIdiase irreversible

-1 PLC -1 ERMform -1 ERMfform ceramidse Sphmydias PI45diase irreversible

-1 PLC ceramidse Sphmydias -2 SPK PI4kin PI4P5kin PI45diase irreversible

-1 PLC ceramidse Sphmydias -1 DAGkin Actinnucl irreversible

-3 PLC 3 ceramidse 3 Sphmydias -2 DAGkin PI4kin PI4P5kin PI45diase irreversible

-1 PLC -1 CholinPtf PIdiase irreversible

-1 PLC -1 ERMform -1 ERMfform -1 CholinPtf PI45diase irreversible

-1 PLC -1 CholinPtf -2 SPK PI4kin PI4P5kin PI45diase irreversible

-1 PLC -1 CholinPtf -1 DAGkin Actinnucl irreversible

-3 PLC -3 CholinPtf -2 DAGkin PI4kin PI4P5kin PI45diase irreversible monoAGcholinPCtf -1 PLC Cholinkin PLA2 PIdiase irreversible monoAGcholinPCtf -1 PLC -1 ERMform -1 ERMfform Cholinkin PLA2 PI45diase irreversible

-1 PLD monoAGcholinPCtf SPK PLA2 reversible

-2 PLD 2 monoAGcholinPCtf ERMform ERMfform PI4kin PI4P5kin 2 PLA2 reversible

PLD -1 monoAGcholinPCtf -1 Lipase -1 DAGkin -1 PLA2 reversible

PLD -1 monoAGcholinPCtf -1 DAGkin -1 DAGcholT -1 PLA2 reversible

PLD -1 monoAGcholinPCtf -1 DAGkin -1 PLA2 -1 PAP reversible

PLD -1 PLC Cholinkin -1 DAGkin reversible

PLD -1 monoAGcholinPCtf -1 PLC ceramidse Sphmydias -1 DAGkin -1 PLA2 reversible

-1 PLD ceramidse Sphmydias -1 Cholinkin SPK reversible

-2 PLD ERMform ERMfform 2 ceramidse 2 Sphmydias -2 Cholinkin PI4kin PI4P5kin reversible

PLD -1 Lipase -1 ceramidse -1 Sphmydias Cholinkin -1 DAGkin reversible

PLD -1 ceramidse -1 Sphmydias Cholinkin -1 DAGkin -1 DAGcholT reversible

PLD -1 ceramidse -1 Sphmydias Cholinkin -1 DAGkin -1 PAP reversible

PLD -1 monoAGcholinPCtf -1 PLC -1 CholinPtf -1 DAGkin -1 PLA2 reversible

-1 PLD -1 CholinPtf -1 Cholinkin SPK reversible

-2 PLD ERMform ERMfform -2 CholinPtf -2 Cholinkin PI4kin PI4P5kin reversible

PLD -1 Lipase CholinPtf Cholinkin -1 DAGkin reversible

PLD CholinPtf Cholinkin -1 DAGkin -1 DAGcholT reversible

PLD CholinPtf Cholinkin -1 DAGkin -1 PAP reversible

-1 PLD -1 Lipase PLC -1 Cholinkin SPK reversible

-2 PLD -2 Lipase 2 PLC ERMform ERMfform -2 Cholinkin PI4kin PI4P5kin reversible 83: -1 PLD PLC -1 Cholinkin SPK -1 DAGcholT reversible

84: -2 PLD 2 PLC ERMform ERMfform -2 Cholinkin PI4kin PI4P5kin -2 DAGcholT reversible 85: -1 PLD PLC -1 Cholinkin SPK -1 PAP reversible

86: -2 PLD 2 PLC ERMform ERMfform -2 Cholinkin PI4kin PI4P5kin -2 PAP reversible 87: monoAGcholinPCtf -1 PLC Cholinkin -1 DAGkin SPK PLA2 reversible 88: 2 monoAGcholinPCtf -2 PLC ERMform ERMfform 2 Cholinkin -2 DAGkin PI4kin

PI4P5kin 2 PLA2 reversible

89: -1 PLD monoAGcholinPCtf PLA2 Actinnucl irreversible

90: -2 PLD 2 monoAGcholinPCtf -1 Lipase PI4kin PI4P5kin 2 PLA2 PI45diase irreversible

91 : -1 PLD monoAGcholinPCtf DAGkin PLA2 PIdiase irreversible

92: -1 PLD monoAGcholinPCtf -1 ERMform -1 ERMfform DAGkin PLA2 PI45diase irreversible

93: -3 PLD 3 monoAGcholinPCtf DAGkin PI4kin PI4P5kin 3 PLA2 PI45diase irreversible

94: -2 PLD 2 monoAGcholinPCtf PI4kin PI4P5kin -1 DAGcholT 2 PLA2 PI45diase irreversible

95: -2 PLD 2 monoAGcholinPCtf PI4kin PI4P5kin 2 PLA2 -1 PAP PI45diase irreversible

96: -2 PLD 2 monoAGcholinPCtf -1 PLC ceramidse Sphmydias PI4kin PI4P5kin 2 PLA2

PI45diase irreversible

97: -1 PLD ceramidse Sphmydias -1 Cholinkin Actinnucl irreversible

98: -2 PLD -1 Lipase 2 ceramidse 2 Sphmydias -2 Cholinkin PI4kin PI4P5kin PI45diase irreversible

99: -2 PLD -1 PLC 3 ceramidse 3 Sphmydias -2 Cholinkin PI4kin PI4P5kin PI45diase irreversible 100: -1 PLD ceramidse Sphmydias -1 Cholinkin DAGkin PIdiase irreversible 101: -1 PLD -1 ERMform -1 ERMfform ceramidse Sphmydias -1 Cholinkin DAGkin PI45diase irreversible

102: -3 PLD 3 ceramidse 3 Sphmydias -3 Cholinkin DAGkin PI4kin PI4P5kin PI45diase irreversible 103: -2 PLD 2 ceramidse 2 Sphmydias -2 Cholinkin PI4kin PI4P5kin -1 DAGcholT PI45diase irreversible 104: -2 PLD 2 ceramidse 2 Sphmydias -2 Cholinkin PI4kin PI4P5kin -1 PAP PI45diase irreversible 105: -2 PLD 2 monoAGcholinPCtf -1 PLC -1 CholinPtf PI4kin PI4P5kin 2 PLA2 PI45diase irreversible 106: -1 PLD -1 CholinPtf -1 Cholinkin Actinnucl irreversible 107 -2 PLD -1 Lipase -2 CholinPtf -2 Cholinkin PMkin PI4P5kin PI45diase irreversible

108 -2 PLD -1 PLC -3 CholinPtf -2 Cholinkin PMkin PI4P5kin PI45diase irreversible

109 -1 PLD -1 CholinPtf -1 Cholinkin DAGkin PIdiase irreversible

110 -1 PLD -1 ERMform -1 ERMfform -1 CholinPtf -1 Cholinkin DAGkin PI45diase irreversible

111 -3 PLD -3 CholinPtf -3 Cholinkin DAGkin PMkin PI4P5kin PI45diase irreversible

112 -2 PLD -2 CholinPtf -2 Cholinkin PMkin PI4P5kin -1 DAGcholT PI45diase irreversible

113 -2 PLD -2 CholinPtf -2 Cholinkin PMkin PMP5kin -1 PAP PI45diase irreversible

114 -1 PLD -1 Lipase PLC -1 Cholinkin Actinnucl irreversible

115 -2 PLD -3 Lipase 2 PLC -2 Cholinkin PMkin PMP5kin PI45diase irreversible

116 -1 PLD PLC -1 Cholinkin -1 DAGcholT Actinnucl irreversible

117 -2 PLD 2 PLC -2 Cholinkin PMkin PI4P5kin -3 DAGcholT PI45diase irreversible

118 -1 PLD PLC -1 Cholinkin -1 PAP Actinnucl irreversible

119 -2 PLD 2 PLC -2 Cholinkin PMkin PI4P5kin -3 PAP PM5diase irreversible

120 -2 PLD 3 monoAGcholinPCtf -1 PLC Cholinkin PMkin PMP5kin 3 PLA2 PI45diase irreversible

121 PLD -1 PLC Cholinkin -1 SPK PIdiase irreversible

122 PLD -1 PLC -1 ERMform -1 ERMfform Cholinkin -1 SPK PI45diase irreversible

123 PLD -1 PLC Cholinkin -3 SPK PMkin PMP5kin PI45diase irreversible

124 2 PLD -2 PLC -1 ERMform -1 ERMfform 2 Cholinkin -1 PMkin -1 PMP5kin 2 PIdiase irreversible

125: 2 PLD -2 PLC -3 ERMform -3 ERMfform 2 Cholinkin -1 PMkin -1 PMP5kin 2 PI45diase irreversible

126: monoAGcholinPCtf -1 PLC Cholinkin -2 SPK PMkin PMP5kin PLA2 PI45diase irreversible

127: monoAGcholinPCtf -1 PLC Cholinkin -1 DAGkin PLA2 Actinnucl irreversible 128: 3 monoAGcholinPCtf -3 PLC 3 Cholinkin -2 DAGkin PMkin PMP5kin 3 PLA2 PI45diase irreversible

The overall reactions of these elementary modes are:

1 inositol + PA + CTP = CMP + PPi + PI 2 Actinpoly = 2 Actin 3 CDP-cholin + TAG = CMP + fatac + PC 4 P + TAG = fatac + PA 5 CDP-cholin + PA = P + CMP + PC fatac + PI = IcP + TAG

ERM-PIP2 + fatac = inositolTP + Ezrin + TAG + TM

CDP-cholin + PI = CMP + IcP + PC

CDP-cholin + ERM-PIP2 = inositolTP + Ezrin + CMP + PC + TM

P + PI = IcP + PA

P + ERM-PIP2 = inositolTP + Ezrin + PA + TM

Ezrin + 2 S- 1 -P + TM + PI = ERM-PIP2 + 2 sph fatac + DAG-3P + sph = TAG + S-l-P

CDP-cholin + DAG-3P + sph = CMP + S-l-P + PC

P + DAG-3P + sph = S-l-P + PA

Ezrin + 2 fatac + 2 DAG-3P + TM + PI = ERM-PIP2 + 2 TAG

2 CDP-cholin + Ezrin + 2 DAG-3P + TM + PI = ERM-PIP2 + 2 CMP + 2 PC

2 P + Ezrin + 2 DAG-3P + TM + PI = ERM-PIP2 + 2 PA

2 Actin + S-l-P = P + Actinpoly + sph fatac + 2 S-l-P + PI = inositolTP + TAG + 2 sph

S-l-P + PI = IcP + DAG-3P + sph

ERM-PIP2 + S-l-P = inositolTP + Ezrin + DAG-3P + TM + sph

3 S-l-P + PI = inositolTP + DAG-3P + 3 sph CDP-cholin + 2 S-l-P + PI = inositolTP + CMP + PC + 2 sph P + 2 S-l-P + PI = inositolTP + PA + 2 sph

4 Actin + ERM-PIP2 = 2 P + 2 Actinpoly + Ezrin + TM + PI ERM-PIP2 + PI = Ezrin + 2 IcP + 2 DAG-3P + TM

3 ERM-PIP2 = 2 inositolTP + 3 Ezrin + 2 DAG-3P + 3 TM + PI

2 Actin + fatac + DAG-3P = P + Actinpoly + TAG

3 fatac + 2 DAG-3P + PI = inositolTP + 3 TAG CDP-cholin + 2 Actin + DAG-3P = P + Actinpoly + CMP + PC 3 CDP-cholin + 2 DAG-3P + PI = inositolTP + 3 CMP + 3 PC

2 Actin + DAG-3P = Actinpoly + PA

3 P + 2 DAG-3P + PI = inositolTP + 3 PA CDP-cholin + fatac + PPi + sph = Sphmy + CTP 2 fatac + PC + sph = Sphmy + TAG CDP-cholin + fatac + sph = Sphmy + CMP

P + fatac + PC + sph = Sphmy + PA

PC + sph = Sphmy + AA

Sphmy + DAG-3P = fatac + S-l-P + PC

2 Sphmy + Ezrin + 2 DAG-3P + TM + PI = ERM-PIP2 + 2 fatac + 2 PC + 2 sph fatac + PC + CTP = CDP-cholin + TAG + PPi

CTP = CMP + PPi

P + PC + CTP = CDP-cholin + PA + PPi

PC + CTP = CDP-cholin + fatac + PPi + AA

CDP-cholin + DAG-3P + PPi + sph = S-l-P + PC + CTP

2 CDP-cholin + Ezrin + 2 DAG-3P + TM + 2 PPi + PI = ERM-PIP2 + 2 PC +2CTP TAG = 2 fatac + AA

CMP + PC = CDP-cholin + fatac + AA

PA = P + fatac + AA

Sphmy + PI = fatac + IcP + PC + sph

Sphmy + ERM-PIP2 = inositolTP + Ezrin + fatac + PC + TM + sph

Sphmy + 2 S-l-P + PI = inositolTP + fatac + PC + 3 sph

Sphmy + 2 Actin + DAG-3P = P + Actinpoly + fatac + PC + sph

3 Sphmy + 2 DAG-3P + PI = inositolTP + 3 fatac + 3 PC + 3 sph CDP-cholin + PPi + PI = IcP + PC + CTP

CDP-cholin + ERM-PIP2 + PPi = inositolTP + Ezrin + PC + TM + CTP

CDP-cholin + 2 S-l-P + PPi + PI = inositolTP + PC + CTP + 2 sph

CDP-cholin + 2 Actin + DAG-3P + PPi = P + Actinpoly + PC + CTP

3 CDP-cholin + 2 DAG-3P + 3 PPi + PI = inositolTP + 3 PC + 3 CTP

PI = fatac + IcP + AA

ERM-PIP2 = inositolTP + Ezrin + fatac + TM + AA

PA + sph = fatac + S-l-P + AA

Ezrin + 2 PA + TM + PI = ERM-PIP2 + 2 fatac + 2 AA

2 fatac + DAG-3P + AA = TAG + PA

CDP-cholin + fatac + DAG-3P + AA = CMP + PC + PA

P + fatac + DAG-3P + AA = 2 PA

DAG-3P = PA

Sphmy + DAG-3P + AA = PC + PA + sph

Sphmy + PA = fatac + S-l-P + PC

2 Sphmy + Ezrin + 2 PA + TM + PI = ERM-PIP2 + 2 fatac + 2 PC + 2 sph

2 fatac + PC + DAG-3P + sph = Sphmy + TAG + PA

CDP-cholin + fatac + DAG-3P + sph = Sphmy + CMP + PA

P + fatac + PC + DAG-3P + sph = Sphmy + 2 PA

CDP-cholin + fatac + DAG-3P + PPi + AA = PC + PA + CTP

CDP-cholin + PA + PPi + sph = S-l-P + PC + CTP

2 CDP-cholin + Ezrin + 2 PA + TM + 2 PPi + PI = ERM-PIP2 + 2 PC + 2 CTP 78 fatac + PC + DAG-3P + CTP = CDP-cholin + TAG + PA + PPi 79 DAG-3P + CTP = CMP + PA + PPi 80 P + PC + DAG-3P + CTP = CDP-cholin + 2 PA + PPi 81 fatac + PA + sph = TAG + S- 1 -P 82 Ezrin + 2 fatac + 2 PA + TM + PI = ERM-PIP2 + 2 TAG 83 CDP-cholin + PA + sph = CMP + S-l-P + PC 84 2 CDP-cholin + Ezrin + 2 PA + TM + PI = ERM-PIP2 + 2 CMP + 2 PC 85 P + sph = S-l-P 86 2 P + Ezrin + TM + PI = ERM-PIP2 87 DAG-3P + sph = fatac + S-l-P + AA

Ezrin + 2 DAG-3P + TM + PI = ERM-PIP2 + 2 fatac + 2 AA

89 2 Actin + PA = P + Actinpoly + fatac + AA 90 2 PA + PI = inositolTP + fatac + TAG + 2 AA 91 PA + PI = fatac + IcP + DAG-3P + AA 92 ERM-PIP2 + PA = inositolTP + Ezrin + fatac + DAG-3P + TM + AA 93 3 PA + PI = inositolTP + 3 fatac + DAG-3P + 3 AA 94 CDP-cholin + 2 PA + PI = inositolTP + CMP + 2 fatac + PC + 2 AA 95 P + PA + PI = inositolTP + 2 fatac + 2 AA 96 Sphmy + 2 PA + PI = inositolTP + 3 fatac + PC + 2 AA + sph 97 Sphmy + 2 Actin + PA = P + Actinpoly + fatac + PC + sph 98 2 Sphmy + 2 PA + PI = inositolTP + fatac + TAG + 2 PC + 2 sph 99 3 Sphmy + 2 PA + PI = inositolTP + 3 fatac + 3 PC + 3 sph 100 Sphmy + PA + PI = fatac + IcP + PC + DAG-3P + sph 101 Sphmy + ERM-PIP2 + PA = inositolTP + Ezrin + fatac + PC + DAG-3P +TM + sph 102 3 Sphmy + 3 PA + PI = inositolTP + 3 fatac + 3 PC + DAG-3P + 3 sph 103 2 Sphmy + CDP-cholin + 2 PA + PI = inositolTP + CMP + 2 fatac + 3 PC + 2 sph 104 P + 2 Sphmy + PA + PI = inositolTP + 2 fatac + 2 PC + 2 sph 105 CDP-cholin + 2 PA + PPi + PI = inositolTP + 2 fatac + PC + CTP + 2 AA 106 CDP-cholin + 2 Actin + PA + PPi = P + Actinpoly + PC + CTP 107 2 CDP-cholin + fatac + 2 PA + 2 PPi + PI = inositolTP + TAG + 2 PC + 2 CTP 108 3 CDP-cholin + 2 PA + 3 PPi + PI = inositolTP + 3 PC + 3 CTP 109 CDP-cholin + PA + PPi + PI = IcP + PC + DAG-3P + CTP 110 CDP-cholin + ERM-PIP2 +PA +PPi = inositolTP +Ezrin+PC+DAG-3P+TM+CTP 111 3 CDP-cholin + 3 PA + 3 PPi + PI = inositolTP + 3 PC + DAG-3P + 3 CTP 112 3 CDP-cholin + 2 PA + 2 PPi + PI = inositolTP + CMP + 3 PC + 2 CTP 113 P + 2 CDP-cholin + PA + 2 PPi + PI = inositolTP + 2 PC + 2 CTP 114 2 Actin + fatac + PA = P + Actinpoly + TAG 115 3 fatac + 2 PA + PI = inositolTP + 3 TAG 116 CDP-cholin + 2 Actin + PA = P + Actinpoly + CMP + PC 117 3 CDP-cholin + 2 PA + PI = inositolTP + 3 CMP + 3 PC 118 2 Actin = Actinpoly 119 3 P + PI = inositolTP + PA 120 2 PA + PI = inositolTP + 3 fatac + 3 AA 121 S-l-P + PI = IcP + PA + sph 122 ERM-PIP2 + S-l-P = inositolTP + Ezrin + PA + TM + sph 123 3 S-l-P + PI = inositolTP + PA + 3 sph 124 ERM-PIP2 + PI = Ezrin + 2 IcP + 2 PA + TM 125 3 ERM-PIP2 = 2 inositolTP + 3 Ezrin + 2 PA + 3 TM + PI 126 2 S-l-P + PI = inositolTP + fatac + AA + 2 sph 127 2 Actin + DAG-3P = P + Actinpoly + fatac + AA 128 2 DAG-3P + PI = inositolTP + 3 fatac + 3 AA

These represent a complete set of all metabolic modes that can operate in this limited system. Any observed in vivo state is a linear combination of these elementary modes. No additional information (other than just standard textbook information on the substrates and products of the involved enzymes) was used to predict the resulting elementary modes.

The enzyme network describes several different conversions: Actin depolymerization (mode 2) numerous, partly overlapping, routes for fatty acid metabolism and intercon versions (113 modes) and 14 modes supplying in some way energy to the actin nucleation system. Of special relevance to the present study are the 14 modes that predicted a conversion of monomeric to filamentous actin (see below). For the analysis of the phagosome membrane system, this approach has the advantage that it handles the data in a simple binary fashion. Lists of potential substrates (taken as external metabolites) can be provided that can support actin nucleation (+) or not (-). The notation 0 (no effect) provides additional information about the testing of some of these predictions by actin nucleation assays and other experiments, as described below.

Example 9: Elementary modes predicted to be involved in actin nucleation

In each example given below, the first line lists the enzymes involved. Abbreviations used are those listed in the description of the figures section above. An enzyme is given a minus sign if it operates in the opposite direction under this mode (e.g. if a kinase releases a phosphate in this mode it is given a minus sign). The second line gives the overall transformation achieved by the mode. Every reaction within a mode is stochiometrically balanced with all other reactions and the numbers denote the number of moles of the metabolite that needs to be converted in order to maintain equilibrium. The second number [in brackets] denotes numbering in the context of the total 128 elementary modes generated by elementary mode analysis of the reaction scheme shown in Fig 12, as listed above:

Table 1. The 14 predicted actin nucleation modes.

1 [19]: -1 SPK Actinnucl

1 [19]: 2 Actin + S-l-P = P + Actinpoly + sph

2 [26] : - 1 Ezrin Rec form - 1 PMkin - 1 PI4P5kin 2 Actinnucl

-1 Ezrin Rec form Ezrin-Rec-PIP2 = PIP2 + Ezrin + Rec

- 1 PMkin PIP2 + ADP+P = PIP + ATP

- 1 PI4P5kin PIP + ADP+P = PI + ATP 2 Actinnucl 4 Actin + 2ATP = 2Actinpoly + 2ADP + 2P

Ezrin-Rec-PIP2 + (PIP2 + ADP+P) + (PIP + ADP+P) + (4 Actin + 2ATP) = (PrP2+Ezrin+Rec)+(PIP +ATP)+(PI+ATP)+(2Actinpoly+2ADP+ 2P) simplifies to: 2 [26] : 4 Actin + Ezrin-Rec-PIP2 = 2 P + 2 Actinpoly + Ezrin + Rec + PI

3 [29]: -1 Lipase -1 DAGkin Actinnucl

3 [29]: 2 Actin + fatac + DAG-3P = P + Actinpoly + TAG

4 [31]: -1 DAGkin -1 DAGcholT Actinnucl

4 [31]: CDP-cholin + 2 Actin + DAG-3P = P + Actinpoly + CMP + PC

5 [33]: -1 DAGkin -1 PAP Actinnucl

5 [33]: 2 Actin + DAG-3P = Actinpoly + PA [54]: -1 PLC ceramidase Sphmydias -1 DAGkin Actinnucl [54]: Sphmy + 2 Actin + DAG-3P = P + Actinpoly + fatac + PC + sph

[59] : - 1 PLC - 1 CholinPtf - 1 DAGkin Actinnucl [59]: CDP-cholin + 2 Actin + DAG-3P + PPi = P + Actinpoly + PC + CTP

[89] : - 1 PLD monoAGcholinPCtf PLA2 Actinnucl [89] : 2 Actin + PA = P + Actinpoly + fatac + AA

[97]: -1 PLD ceramidse Sphmydias -1 Cholinkin Actinnucl [97]: Sphmy + 2 Actin + PA = P + Actinpoly + fatac + PC + sph

[ 106] : - 1 PLD - 1 CholinPtf - 1 Cholinkin Actinnucl [ 106] : CDP-cholin + 2 Actin + P A + PPi = P + Actinpoly + PC + CTP

[ 114] : - 1 PLD - 1 Lipase PLC - 1 Cholinkin Actinnucl [ 114] : 2 Actin + fatac + PA = P + Actinpoly + TAG

[ 116] : - 1 PLD PLC - 1 Cholinkin - 1 DAGcholT Actinnucl [ 116] : CDP-cholin + 2 Actin + PA = P + Actinpoly + CMP + PC

[ 118] : - 1 PLD PLC - 1 Cholinkin - 1 PAP Actinnucl

-1 PLD PA + cholin = PC

PLC PC = DAG + cholinP

-1 Cholinkin cholinP + ADP = cholin + ATP

-1 PAP DAG + P = PA

Actinnucl 2 Actin + ATP = Actinpoly + ADP + P

(PA + cholin) + PC + (cholinP + ADP) + (DAG + P) + (2 Actin + ATP) = PC + (DAG + cholinP) + (cholin + ATP) + PA + (Actinpoly + ADP +P); simplifies to: 13 [ 118] : 2 Actin = Actinpoly

14 [ 127] : monoAGcholinPCtf - 1 PLC Cholinkin - 1 DAGkin PLA2 Actinnucl 14 [127]: 2 Actin + DAG-3P = P + Actinpoly + fatac + AA

Since four of the fourteen elementary modes predicted a role for CDP-choline in the assembly of actin, we tested the effects of this precursor of phosphatidyl choline in the actin nucleation assay, using both LBP and the phagosomes of M. smegmatis, and of live and killed M.avium. As shown in Fig 7, the addition of CDP-choline led to a significant stimulation of the nucleation of actin by LBP, as well as both the phagosomes enclosing killed and live M.avium. We also tested the effects of a second predicted effector, sphingomyelin (Sphm). This lipid was also able to stimulate actin nucleation by all the phagosome types tested, in agreement with the predictions (Fig 7). Whereas the LBP were stimulated at both high and low ATP with this lipid, the killed and live M. avium were stimulated only at low ATP (Fig 7).

We have shown that PI(4)P and P(4,5)P₂ greatly facilitate LBP actin nucleation and that an interaction between ezrin/moesin and PI4,5P₂ on the LBP membrane facilitated the assembly of actin (Defacque et α/,submitted). These proteins are known to bind to both membrane receptors (such as CD 44; although this is present on LBP in low amounts it remains to be determined whether it serves as a receptor for ezrin or moesin in this system), and to PI(4,5)P₂, via distinct regions of their N-terminal regions. In that study, we provided evidence that both pre-existing and newly sythesised PIPs are required for actin assembly and suggested that PIP turnover might be important for the ezrin/moesin-mediated insertion of actin on the membrane surface. Elementary mode 2 (above) is in excellent agreement with this notion; it predicts a pathway in which four moles of G-actin are inserted into filaments by the ezrin-receptor-PI(4,5)P₂ complex in conjunction with the breakdown of one molecule of PI(4,5)P₂ into PI. This suggests that actin assembly could be driven by ezrin/moesin molecules that are stably bound to membrane receptors alternatively binding to, and dissociating from PI(4,5)P₂ and that synthesis and breakdown of this lipid are coordinated with these binding and dissociation processes. It also implies that PI4 and 5- kinases and phosphatases must be integral components of the membrane-bound actin nucleation machinery. An unexpected result is that, even with only a limited sub-set of molecules that interact, in or on the membrane there appear to be many different pathways for providing energy for actin nucleation. Whether there is only one, or more distinct mechanisms, besides that involving ezrin/moesin, for the nucleation/insertion of monomers into filaments on the membrane surface remains to be determined. It should be noted that one elementary mode (mode 14 above) predicts a kinase-independent pathway for actin assembly involving phospholipase A2.

The above examples therefore clearly demonstrate the applicability of elemental mode analysis to the described phagosome assay systems. The elementary mode analysis gives an indication of how the different lipid pathways may activate actin nucleation. Although regulatory interactions and kinetic details of the enzymes are not taken into account, the full range of metabolic paths operative in the network is concisely described by the elementary modes. Both choline kinase and DAG kinase have K_ms for ATP in the mM range, and would thus be expected to be significantly less active under low ATP conditions. In the case of DAG this enzyme shows activity under low ATP conditions (see above). Presumably, the conversion of limited quantities of the metabolites is sufficient to stimulate the end reaction, actin nucleation.

Claims

1. A method for the analysis of a membrane signalling network, said method comprising assaying the activity of a component of said membrane signalling network in a phagosome composition under first reaction conditions, and comparing the activity of said component with its activity in a phagosome composition under second reaction conditions.

2. A method according to claim 1, wherein the concentration of at least one component of the membrane signalling network differs between said first and second reaction conditions.

3. A method according to claim 1 or claim 2, comprising the steps of: a) preparing i) a first phagosome composition in which substantially all of the phagosomes contain one or more components of said membrane signalling network; and ii) a second phagosome composition in which the components of the constituent phagosomes differ from the components of the phagosomes in first phagosome composition; and b) comparing the membrane signalling network in the phagosomes of said first phagosome composition with the membrane signalling network in the phagosomes of said second phagosome composition.

4. A method according to any one of the preceding claims, wherein the phagosomes in said first and second phagosome compositions contain latex beads.

5. A method according to claim 4, wherein in at least one of said phagosome compositions, a ligand is coupled to each of said latex beads.

6. A method according to claim 5, wherein a different ligand type is coupled to the latex beads present in said first and second phagosome compositions.

7. A method according to either of claims 5 or 6, wherein said ligand is a bacterial virulence factor.

8. A method according to any one of claims 5-7, wherein phagosomes in at least one of said phagosome compositions express a receptor for said ligand.

9. A method according to claim 8, wherein a different receptor type is expressed in the phagosomes in each of said first and second phagosome compositions.

10. A method according to claim 9, wherein a wild type receptor protein is expressed in the phagosomes of said first phagosome composition and a mutant receptor protein is expressed in the phagosomes of said second phagosome composition.

11. A method according to any one of claims 2-10, wherein the genetic composition of the phagosomes in said first phagosome composition differs from that of the phagosomes in the second phagosome composition.

12. A method according to any one of claims 2-10, wherein substantially all of the phagosomes in at least one of said phagosome compositions contain a pathogenic organism, preferably, a pathogenic micro-organism.

13. A method for analysing the mechanism of pathogen action in host cells, comprising assaying the activity of a component of a membrane signalling network under first reaction conditions in a phagosome preparation in which substantially all of the phagosomes contain pathogenic organisms, and comparing the activity of said component with that of the component in a phagosome preparation under second reaction conditions.

14. A method according to claim 13, wherein the component of the membrane signalling network is compared between a first phagosome preparation in which substantially all of the phagosomes contain a first pathogenic organism, and a second phagosome preparation in which substantially all of the phagosomes contain a second pathogenic organism.

15. A method according to claim 13, wherein the component of the membrane signalling network is compared between a first phagosome preparation in which substantially all of the phagosomes contain a first pathogenic organism, and a second phagosome preparation in which substantially all of the phagosomes contain a second, genetically- related, non-pathogenic organism.

16. A method for the identification of a factor that is necessary for virulence to be effective in a host, said method comprising:

a) comparing the lipid composition and/or protein composition of a phagosome containing a virulent micro-organism with the lipid composition and protein composition of a phagosome containing a non-virulent micro-organism; and

b) selecting as said factor, a lipid or protein factor that is only present in the phagosome that contains the virulent micro-organism and is not present in the phagosome that contains the non-virulent micro-organism.

17. A method for the identification of a factor that is responsible for virulence in a micro- organism, said method comprising:

a) comparing the protein and/or lipid composition of a virulent micro-organism in a phagosome with the lipid and/or protein composition of a non-virulent micro-organism in a phagosome; and

b) selecting as said factor, a lipid or protein factor that is only present in the virulent micro-organism and is not present in the non-virulent micro-organism.

18. A method of screening for a candidate drug molecule capable of altering a component of a membrane signalling network comprising contacting a phagosome composition with said candidate drug molecule, and analysing said signalling network using a method according to any one of claims 1-16.

19. A method of screening for a candidate drug molecule capable of altering a membrane signalling network, said method comprising the steps of: a) preparing at least one phagosome composition in which substantially all of the phagosomes contain one or more components that are in a particular active state of a membrane signalling network; b) contacting said phagosome composition with said candidate drug molecule; c) testing a component, or an effect of a component, of said membrane signalling network in order to assess the effect of said candidate drug molecule on the membrane signalling network for any given function; and d) selecting as said candidate drug molecule, a drug molecule that has a desired effect on said membrane signalling network.

20. A drug molecule identified by a method according to claim 18 or claim 19.

21. A system for analysing a membrane signalling network using a method according to any one of claims 1-17, comprising i) a first phagosome in which a ligand molecule is coupled to a moiety enclosed within the phagosome, wherein the phagosome expresses a receptor for said ligand molecule; and ii) a second phagosome in which the reaction conditions differ from those of the first phagosome.

22. Use of a phagosome composition to identify a component of a membrane signalling network.

23. Use of a phagosome composition in a screen for the identification of a candidate drug molecule capable of altering a membrane signalling network.

24. A relational database system comprising information relating to proteins present in a phagosome system, said relational database comprising a plurality of tables containing information relating to at least two of the following: a) the identity of proteins and/or lipids present in the phagosome system; b) the regulatory mechanisms of proteins and/or lipids present in the phagosome system; c) the putative or proven function of proteins and/or lipids present in the phagosome system; d) post-translational modifications of proteins and/or modifications of lipids present in the phagosome system; and e) alteration of the level of expression or activity of a protein or a lipid molecule present in the phagosome system as a result of infection of a host by a micro-organism.

25. A relational database system according to claim 24, additionally comprising information relating to the lipids present in the phagosome system.

26. A method for modulating the ability of a phagocyte to destroy a pathogen, said method comprising contacting the phagocyte with a lipid selected from the group consisting of arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1 -phosphate and phosphatidyl-inositol-di-phosphate.

27. A method according to claim 26, wherein said modulating agent stimulates phagosome- lysosome fusion in phagocyte cells.

28. A method according to claim 26 or claim 27, wherein said modulating agent causes a reduction in phagosomal pH.

29. A method according to any one of claims 26 to 28, wherein said modulating agent stimulates the release of reactive nitrogen intermediates (RNI).

30. A method according to any one of claims 26 to 29, wherein contacting the phagocyte with said lipid results in an alteration in the proportion of lipids in the phagosome membrane bilayer.

31. A method for modulating the ability of a phagocyte to destroy a pathogen, said method comprising contacting the phagocyte with a modulating agent that stimulates phagosome maturation and pathogen killing.

32. A method according to claim 31, wherein the mechanism that stimulates phagosome maturation and pathogen killing is that of actin nucleation.

33. A method according to claim 31 or claim 32, wherein said modulating agent stimulates phagosome-lysosome fusion in phagocyte cells.

34. A method according to any one of claims 31 to 33, wherein said modulating agent causes a reduction in phagosomal pH.

35. A method according to any one of claims 31 to 34, wherein said modulating agent stimulates the release of reactive nitrogen intermediates (RNI).

36. A method according to any one of claims 31 to 35, wherein said modulating agent is a lipid species.

37. A method according to claim 36, wherein said lipid is chosen from the group comprising arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1- phosphate and phosphatidyl-inositol-di-phosphate.

38. A method according to any one of claims 31 to 37, wherein contacting the phagocyte with said modulating agent results in an alteration in the proportion of lipids in the phagosome membrane bilayer.

39. Use of a lipid selected from the group consisting of arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1 -phosphate and phosphatidyl-inositol-di- phosphate in the manufacture of a medicament for the treatment or prophylaxis of phagocytic pathogen infection.

40. Use as recited in claim 39, wherein said treatment or prophylaxis of phagocytic pathogen infection is mediated through the stimulation of pathogen killing by phagocyte cells.

41. Use according to claim 40, wherein said stimulation of pathogen killing is mediated through increased actin nucleation.

42. Use according to claim 40 or claim 41, wherein said stimulation of pathogen killing by phagocytes is due to increased levels of phagosome-lysosome fusion.

43. Use according to any one of claims 40-42, wherein said stimulation of pathogen killing by phagocytes is due to a reduction in phagosomal pH.

44. Use according to any one of claims 40-43, wherein said stimulation of pathogen killing by phagocytes is due to increased release of reactive nitrogen intermediates (RNI).

45. Use as recited in any one of claims 39-45, wherein said pathogen is a mycobacterium.

46. Use according to claim 45, wherein said mycobacterium is selected from the group consisting of M. tuberculosis, M. paratuberculosis, M. leprae, M. avium and M. bovis.

47. A method of treating or preventing infection by a pathogen, said method comprising contacting infected phagocytes in a patient with an effective dose of a modulating agent, wherein said modulating agent stimulates actin nucleation.

48. A method of treating a human or animal having a disease or condition associated with pathogen phagosomes, comprising administering to said human or animal a therapeutically or prophylactically effective amount of a modulating agent, wherein said modulating agent stimulates actin nucleation.

49. A method of claims 47 or 48, wherein said modulating agent is a lipid species.

50. The method of claim 49, wherein said lipid is chosen from the group consisting of arachidonic acid, sphingosine, ceramide, sphingomyelin, sphingosine- 1 -phosphate and phosphatidyl-inositol-di-phosphate.