Chemotaxis: Navigating by Multiple Signaling Pathways

See allHide authors and affiliations

Science's STKE  24 Jul 2007:
Vol. 2007, Issue 396, pp. pe40
DOI: 10.1126/stke.3962007pe40


During chemotaxis, phosphatidylinositol 3,4,5-trisphosphate (PIP3) accumulates at the leading edge of a eukaryotic cell, where it induces the formation of pseudopodia. PIP3 has been suggested to be the compass of cells navigating in gradients of signaling molecules. Recent observations suggest that chemotaxis is more complex than previously anticipated. Complete inhibition of all PIP3 signaling has little effect, and alternative pathways have been identified. In addition, selective pseudopod growth and retraction are more important in directing cell movement than is the place where new pseudopodia are formed.

Chemotaxis is the ability of cells to move in the direction of an external gradient of signaling molecules. This process coordinates attractive and repulsive movement during development, guides the migration of immune cells toward sites of infection, induces the spread of metastatic cancer cells toward growth factors, or provides a mechanism for single-celled organisms to find niches with ample food supplies (13). Research on chemotaxis by eukaryotic cells has progressed substantially, mainly through the study of neutrophils or the amoeba Dictyostelium discoideum as model systems (4, 5). Studies in these models have led to the concept of a compass: The lipid second messenger phosphatidylinositol 3,4,5-trisphosphate (PIP3) accumulates at the front of chemotaxing cells, where it induces cytoskeletal rearrangements leading to protrusions in the direction of the chemoattractant gradient (58). A series of recent papers on Dictyostelium chemotaxis challenges this model of a PIP3-mediated compass.

The concept of a compass is overwhelmingly obvious when observing mammalian or Dictyostelium cells expressing a green fluorescent protein (GFP)–tagged pleckstrin homology (PH) domain that binds PIP3 moving toward a pipette releasing chemoattractant (911). In these steep gradients, cells move in nearly straight lines toward the pipette and exhibit a strong patch of PIP3 at their leading edge (Fig. 1A). Experiments in shallow gradients provide a different view of gradient sensing (12, 13). In shallow gradients, receptor occupancy is "noisy" and cells occasionally experience a "wrong" gradient of occupied receptors and move away from the source of chemoattractant. Under these conditions, it is more likely that the direction of cell movement is determined by a chemoattractant-induced bias of random pseudopod extension than by a compass that tells the cell where to go. Pseudopodia are self-organizing structures. The determination of the place and time where pseudopodia are extended is a stochastic process regulated by activators and inhibitors in the cytoplasm. In a shallow gradient, receptor activation could locally increase the concentration of one or multiple activators or inhibitors, thereby inducing a bias in the direction of locomotion. In this view, chemotaxis is a pure stochastic process with many "mistakes" that becomes increasingly accurate in steeper gradients.

Fig. 1.

(A) PIP3-directed protrusion in a gradient of chemoattractant (left), in uniform chemoattractant (middle), and in a gradient of chemorepellent (right). The asterisks indicate the source of attractant or repellent. (B) Signaling pathways leading to extension (blue) and retraction (red) of protrusions mediating chemotaxis in Dictyostelium. (C) The behaviors of the protrusions are random initiation (top left), selective retraction (top middle), selective extension (top right), suppression of pseudopodia in the rear (bottom left), and pseudopod splitting (bottom right).

The concept that PIP3 directs the extension of new pseudopodia is very instructive and is based on convincing experiments in Dictyostelium and neutrophils: Areas with local formation of actin filaments (f-actin) and actin-based projections occur in areas with elevated levels of PIP3 (Fig. 1). These protrusions can be associated with PIP3 at the front of the cell in a gradient of chemoattractant (9), with random PIP3 patches in cells stimulated with uniform chemoattractant (14) or even with PIP3 at the side of the cell pointing away from a chemical that induces repulsion of the cells (15). PIP3 is produced by phosphorylation of phosphatidylinositol 4,5-bisphosphate [PtdIns(4,5)P2] by a family of enzymes known as phosphatidylinositol 3-kinases (PI3Ks) and is degraded by a family of 5-phosphatases, such as SHIP, to PtdIns(3,4)P2, or by the 3-phosphatase PTEN to PtdIns(4,5)P2. Two positive feedback loops of PI3K binding to f-actin and PTEN binding to PtdIns(4,5)P2, respectively, lead to symmetry breaking and rapid amplification of the shallow chemoattractant signal to a steeper gradient of PIP3 in the cell (1618). The observation in pten-null cells of extended areas of PIP3 that are associated with large and multiple protrusions also strongly supports the idea that PIP3 may function as a compass (19).

The first signs that PIP3 may not be the compass come from experiments with Dictyostelium deficient in PI3K activity. Dictyostelium cells have five genes that encode PI3Ks. When PI3K1 and PI3K2 are knocked out, PIP3 is undetectable, but a small segment of PI3K2 expressed in these cells continues to be localized at the leading edge, suggesting that the compass is upstream of PI3K signaling (20). Furthermore, inhibition of PI3K by LY294002 has little effect on directional sensing, except in very shallow gradients (21). Hoeller and Kay (22) have inactivated all five genes encoding type I PI3Ks and observed that cells that cannot produce any PIP3 exhibit nearly normal chemotaxis. Although PIP3 is amplified in the front of chemotaxing cells where it induces pseudopod formation, PIP3 signaling is not essential and other parallel pathways mediating gradient sensing must be present.

The description of these parallel pathways comes from two studies, using very different methodologies, that identified phospholipase A2 (PLA2) as a chemotactic pathway parallel to the PI3K pathway. Chen et al. (23) took a genetic approach based on the assumption that chemotaxis of Dictyostelium cells toward adenosine 3′,5′-monophosphate (cAMP) would still occur when the PI3K and hypothetical parallel pathway were deleted separately, whereas chemotaxis should be severely inhibited when both pathways were disrupted. They screened some 7000 random integration mutants searching for cells that showed normal chemotaxis in buffer but failed to show chemotaxis when PI3K was inhibited by mild concentrations of LY294002. In this way, they isolated a gene that encodes a patatin-type PLA2. In an alternative approach (24), biophysical analysis indicates that the optimal second messenger to transduce a chemotactic signal has a lifetime of 2 to 8 s and a diffusion rate constant of ~1 um2/s (12). Although PIP3 fits this profile, other lipid products of phospholipase C (PLC), PLA2, or phospholipase D (PLD) are also potential candidates. In ~500 chemotaxis assays with Dictyostelium, with all possible combinations of available genetic disruption and pharmacological inhibitors of these second-messenger pathways, PI3K and PLA2 were revealed as the two key pathways. Inhibition of one pathway has significant effects only for shallow gradients, whereas the inactivation of both pathways blocked chemotaxis toward steep gradients of cAMP. The effects of inhibition of PI3K and PLA2 on chemotaxis were much smaller at later stages of Dictyostelium development, when cells become intrinsically polarized in the absence of chemoattractant (23, 24). This could imply a third signaling pathway that becomes operational in polarized cells. This third pathway may involve soluble guanylyl cyclase (sGC), because the sGC protein (sGCp; Fig. 1B) localizes to the leading edge where it supports f-actin–mediated protrusions in the direction of the gradient (25), and guanosine 3′,5′-monophosphate (cGMP) synthesis stimulates the formation of myosin filaments in the rear of the cell, which inhibits pseudopod extensions (26).

Finally, what is the cellular response leading to movement in the direction of the gradient? The general view of chemotaxis is that pseudopod extension occurs at the leading edge, and suppression of pseudopodia occurs at the side and in the rear of the cell. However, pseudopod extension may be a more random process than previously anticipated. Andrew and Insall (27) show that new pseudopodia are predominantly generated by splitting of an existing pseudopod, followed by survival of one pseudopod and retraction of the other pseudopod. The direction of the split is not influenced by the cAMP gradient, but the pseudopod that is furthest from the cAMP gradient is retracted. The informed-choice model of pseudopod retraction proposed by Andrew and Insall provides a simple model of chemotaxis for cells with an intrinsic polarity axis. Such cells strongly suppress pseudopodia in the posterior part of the cell and tend to make pseudopodia at the front. Stimulated by Andrew and Insall, we have observed mutants that do not suppress pseudopodia in the rear of the cell. As in mammalian cells, suppression of pseudopodia in the rear is caused by myosin filaments, which is mediated by Rho kinases in mammalian cells (28) and cGMP in Dictyostelium (29). A Dictyostelium mutant defective in cGMP signaling extends protrusions all over the cell body (29). Motivated by Andrew and Insall, we reanalyzed movies of guanylyl cyclase–null cells (25) showing that many protrusions do not arise by pseudopod splitting (30). In a cAMP gradient, mutant cells make nearly as many pseudopodia in the front as in the rear of the cell. Nevertheless, those cells exhibit good chemotaxis in natural cAMP gradients (25), suggesting that the pseudopodia formed in the direction of the cAMP gradient are extended over a longer distance or persist longer, leading to the efficient translocation of the cell body in the direction of the cAMP gradient. The collective data suggest that during chemotaxis, pseudopod initiation is more or less in random directions and that cells move in the direction of the gradient because of a mixture of four cellular events (Fig 1C): (i) Cell polarity suppresses the formation of new pseudopodia in the rear, thereby restricting the random pseudopod initiation to the front of the cell. (ii) Pseudopodia have a tendency to split instead of being produced by de novo pseudopod formation. Because existing pseudopodia are preferentially localized at the front, new pseudopodia that arise from splitting therefore are also preferentially formed at the front of the cell. (ii) Pseudopodia pointed in the direction of the chemoattractant gradient extend longer processes. (iv) Pseudopodia pointing away from the gradient are preferentially retracted.

The different signaling pathways may act in parallel to modulate all of these aspects of pseudopod behavior. Alternatively, each pathway may regulate one aspect of pseudopod behavior more strongly than other aspects. It is clear that cGMP mediates suppression of pseudopodia in the rear (29) and that cGMP alone in the absence of front signaling cannot induce chemotaxis (that is, in the absence of PI3K activity, PLA2 activity, and sGC localization). Pseudopod extension may be regulated by PIP3, because pseudopodia that originate from areas of the cell enriched in PIP3 are extended about twice as far from the cell body as other pseudopodia (31). The function of each pathway in chemotaxis may be revealed by studying cell motility in shallow and steep gradients with mutants in which all except one of the pathways are disrupted. Chemotaxis by Dictyostelium involves a surprisingly variety of signaling molecules and cellular responses. Comparison of chemotaxis in bacteria and eukaryotes suggests that during evolution, chemotaxis in bacteria has been improved by fine-tuning of the relatively simple signaling cascade for temporal signaling (32, 33), whereas in eukaryotes additional signaling systems have been implemented to gradually improve chemotaxis to the limit of what is physically possible in spatial sensing.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
View Abstract

Navigate This Article