The pictures reproduced here aren't just pretty — they reveal something fundamental about how cells move towards the source of an attractant chemical. The left image shows the behaviour of normal cells of Dictyostelium discoideum, a type of slime mould. The cells on the right lack an enzyme called PTEN. The colour scale, from blue to red, indicates how each cell moves over time. The normal cells are clearly moving towards a chemical source (asterisk), but the mutant cells seem less sure where they're going.

The images are taken from a paper by Miho Iijima and Peter Devreotes (Cell 109, 599–610; 2002). Together with work by Satoru Funamoto et al. (Cell 109, 611–623; 2002), this study reveals the importance to motility of enzymes — including PTEN — that alter the lipid content of cell membranes.

Many cells need to move up a concentration gradient of a given chemical ('chemoattractant'); in humans, for instance, immune cells migrate towards signals emitted by invading microorganisms. Cells must sense the gradient and then extend protrusions ('pseudopodia') in the direction of movement. Filaments of the protein actin form the skeleton of pseudopodia.

Studies with lipid-binding proteins suggested that the lipid content of cell membranes forms a gradient during chemoattraction, with 3-phosphoinositides being concentrated at the front edge. These are thought to bind signalling proteins, leading to the changes needed for movement. But how does the lipid imbalance come about? This question is tackled in the new papers.

Funamoto et al. started by studying phosphatidylinositol-3-OH kinases (PI(3)Ks), enzymes that produce 3-phosphoinositides, in D. discoideum. When cells were placed in a chemoattractant gradient, two different PI(3)Ks moved to the membrane at the leading edge, with similar kinetics to 3-phosphoinositide-binding proteins. Membrane-bound PI(3)Ks seemed to trigger the formation of pseudopodia.

So the localization of PI(3)Ks could explain the generation of 3-phosphoinositides at the leading edge. But the concentration of these lipids tails off sharply towards the rear of a cell, hinting that an enzyme that degrades them might be at work there. Both groups investigated a possible candidate — PTEN. When Funamoto et al. overexpressed this enzyme in D. discoideum, the cells moved more slowly and became less polarized in response to a chemoattractant. Iijima and Devreotes engineered D. discoideum lacking PTEN, and again the cells moved abnormally, taking a circuitous route to the chemoattractant. The cells also produced pseudopodia more erratically, sometimes at the back, and had more actin filaments than normal. Finally, both groups found that PTEN is usually located at the rear of moving cells.

This suggests that PTEN and PI(3)Ks help to produce a gradient of 3-phosphoinositides, which is much steeper than the chemoattractant gradient. The lipid gradient translates into the correct localization of key signalling proteins and actin filaments, and so into directed movement. But how these enzymes are moved about the cell, and how they interact with other chemoattractant-sensing pathways, remains to be seen.