Eukaryotic chemotaxis

eukacryotte
Figure 1: Snapshots of preliminary time-courses of PIP3 patches that we obtained for eukaryotic directional sensing. Images show two different situations: the left image displays a relatively large patch covering about 1/3 of the cell membrane while the right image feature a (shorter-living) small patch.
     The model organism for eukaryotic chemotaxis is the slime mold Dictyostelium discoideum (dicty) but the phenomenology seems to apply quite generally to eukaryotic chemotaxis (van Haastert and Devreotes, Nature Rev. Mol. Cell Biol., 5, 626-634, 2004). It is observed that dicty is highly sensitive in orienting and responding to tiny gradients (reportedly 1-2% across the cell and possibly even below). Furthermore, adaptation processes allow the cell to respond faithfully over several orders of magnitude of chemoattractant concentration. Tools of molecular biology and genetics have led to major progress in the identification of molecules involved in the chemotactic signaling pathway. Experiments on the response to cAMP indicate that shallow extracellular gradients of chemoattractant are translated into an equally shallow gradient of receptor activation and that the nonlinear process of amplification takes place during the recruitment of the cytosolic enzyme phosphatidylinositol 3-kinase (PI3K) to the plasma membrane, where it phosphorylates PIP2 into the D3-phosphoinositide product of PI3K, PIP3. This process leads to a sharp separation in PIP2- and PIP3-rich phases that are spatially organized (directional sensing) with respect to the direction of the cAMP gradient. Snapshots of the time-evolution of PIP2-3 patches are obtained by fluorescent tags (Postma et al., Journ. of Cell Science, 117, 2925-35, 2004). PIP3 acts then downstream by inducing cell polarization, regulating cytoskeletal dynamics (Cullen et al., Curr. Biol., 11, R882-93, 2001) and eventually cell motion (Ridley et al., Science 302, 1704-09, 2003).
     This sequence of events where the cell detects and amplifies the external spatial cue, filters out the noise and internally propagates the signal to effectors constitutes the phase of directional sensing. This phase, which precedes cell polarization (when the cell itself starts to deform and move), is a general feature for the chemotactic response of eukaryotic cells (Devreotes and Janetopoulos, J. Biol. Chem., 278, 20445-48, 2003; Mortimer et al., Trends in Neurosciences, 31, 90-98, 2008). In dicty, cell polarization can be decoupled from directional sensing by using latrunculin A, an inhibitor of actin polymerization, so that cells are immobilized but still form patches of PIP2-3 as in untreated cells (Janetopoulos et al., PNAS, 101, 8951-56, 2004). Current methods for chemical stimulations are based on the diffusion of solutes released by micropipettes and suffer from poor spatial and temporal resolutions. In particular, by alternating release of cAMP from several micropipettes, it is observed that dicty is capable of relatively rapid reorientations (Devreotes and Janetopoulos, J. Biol. Chem., 278, 20445-20448, 2003). However, the method is qualitative and quantitative description of the time-dependency of the chemotactic response is missing.
     As for modeling, mechanisms proposed in the literature are reviewed in (Devreotes and Janetopulos, J. Biol. Chem., 278, 20445-48, 2003). ``First hit”, “pilot pseudopodia”, “positive feedback” and “mechanical restriction” mechanisms cannot explain the behavior of cells treated with latrunculin. The ``local excitation, global inhibition’’ model fails to account for quantitative aspects of the amplification and the qualitative behavior when the external gradient is shifted. This last point reiterates the importance of assaying variable stimuli, which is one of our objectives. An alternative interesting model for directional sensing was proposed more recently by Gamba et al. (PNAS 102 16927-16932 2005) based on models of statistical physics for coarsening processes.

Our objectives and projects for eukaryotic chemotaxis

     The directional sensing of Dictyostelium discoideum (dicty) to spatial and/or time periodic stimuli of cAMP will be measured with the specific goal to assay its response in the temporal-frequency domain. Patterns of cAMP with spatial periodicity comparable (or even smaller) to the size of the cell (about 10 μm) will be generated in collaboration with M. Dahan (ENS, Paris) and V. Studer (ESPCI, Paris) using the technology of μsticks that they recently presented (see Bartolo, D., et al., Lab on a Chip 8, 274-279, 2008). Each cell is then submitted to a local gradient that will induce the formation of an asymmetric distribution of PIP2-3 patches. The phase of the pattern will then be modulated so that the orientation of the stimulus is inverted. By varying the frequency of the modulation (in a range up to 10Hz) and simultaneously monitoring the localization of the patches, we shall be able to fully characterize the temporal response during directional sensing. This new information will be complemented by more standard experiments on spontaneous polarization of PIP2-3 patches, as in Fig. 1. Indeed, even in the absence of stimuli, patches are still forming and disappearing and we shall monitor their evolution with much better statistics so as to reliably estimate their dynamic time scales.

     As for the modeling of eukaryotic chemotaxis, we shall mostly concentrate on directional sensing and approach towards the end of the project the issue of cell polarization and actual motion. The goal will be to give a quantitative, mathematical description of the space-time dynamics of PIP2-3 patches on the cell membrane in the absence and in response to static and variable stimuli of cAMP concentration. The ``local excitation, global inhibition’’ model will be further tested on quantitative aspects of the amplification process and response to variable and non-homogeneous concentrations. The issue was already broached in (Devreotes and Janetopulos, J. Biol. Chem. 278 20445-48, 2003) and tests seem to point at the necessity of improving the model. In parallel, we shall also consider two models more specifically concerned with the PIP2-3 patch dynamics. The first is the work (Gamba et al., PNAS 102 16927-32, 2005) and the second (that we developed in collaboration with A. Gamba and G. Serini, at IRCC, Turin) is yet unpublished. The major difference between the two is as follows. In the former, the dynamics is the result of the competition among patches of different sizes that acts on pre-formed seeds of non-homogeneity. In the model, this seeding heterogeneity is attributed to microscopic mechanisms of nucleation that are left unspecified. The typical time for the dynamics of the patches varies as the square of its size. Conversely, the second unpublished model is based on an instability mechanism and the typical time-scale of variation of the patches is independent of their size, i.e. this mechanism ensures a faster directional sensing. In the language of physics, the two models are akin to Wagner-Lifchitz-Slyozov theory of phase separation and spinodal decomposition, respectively. Discrimination between the models can be achieved via the response to non-homogeneous and variable stimuli and by quantitatively measuring the typical dynamical times of PIP2-3 patches, which is the information provided by the experiments previously described. The two models produce similar asymptotic responses to a static gradient but strongly differ in their kinetics of orientation so that the experiments we propose are ideally suited to bring decisive clues.
A longer-term goal to be pursued in collaboration with N. Guillen and E. Labruyère (Inst. Pasteur) is to exploit information acquired on dicty for the human parasite Entamoeba histolytica, which is known to strongly rely on chemotaxis for the invasion of tissues (Blazquez et al., Cellular Microbiology, 10, 1676-86, 2008). PIP patches have in particular not been tracked yet in Entamoeba and the response of the protozoan to variable stimuli has not been assayed. Variable conditions are likely to be encountered by the parasite during the invasive amoebiasis and it will be of interest to compare its response to that of dicty.