In the social amoeba Dictyostelium discoideum, cyclic AMP activates a mitogen-activated protein kinase (MAPK) cascade, which operates through MAPK and extracellular signal-regulated kinase (ERK) kinase — MEK — to induce chemotaxis and cell aggregation. Here, Firtel's group shows that MEK SUMOylation and ubiquitylation are necessary for regulating chemotaxis.

Using mek1-null cells, the authors showed that a tagged version of MEK1 complemented the null phenotype (defective chemotaxis and small-aggregate formation), but that cAMP also shifted the electrophoretic mobility of MEK1 — with the maximal change occurring 5–15 seconds after stimulation. This result, plus the presence of two putative SUMOylation sites, led the authors to ask if MEK1 was SUMOylated. Biochemical studies showed that it was, but only on one of the sites — lysine (K) 105. A K105R (arginine) mutation (MEK1K105R) couldn't complement the chemotaxis defect of mek1-null cells, which implied that SUMOylation was needed for MEK to function properly.

So what is SUMOylation doing? One function, it seems, is to translocate MEK to the cell cortex in response to cAMP. Normally, a proportion of MEK moves to the cell cortex — again, maximally after 5–15 seconds — concomitant with the ability to detect SUMO immunoreactivity here and with a decrease in the amount of MEK in the nucleus. But this didn't occur in MEK1K105R cells.

The authors then investigated whether there is a relationship between MEK activation and SUMOylation — that is, would SUMOylation translocate non-activatable MEK? The answer was no; non-activatable MEK stayed in the nucleus and was not SUMOylated. Conversely, constitutively activated MEK was found mainly in the cytosol and was constitutively SUMOylated.

In searching for MEK1-interacting proteins, Firtel's group identified MEK1-interacting protein (MIP1), which contains a RING finger that is common in SUMO ligases (enzymes that add SUMO to their target proteins). But whereas MIP1 did not mediate MEK SUMOylation (MEK1 was still strongly SUMOylated in mip1-null cells), MIP1 turned out to be an E3 ubiquitin ligase. Indeed, a higher molecular weight MEK 'smear' that was seen after prolonged cAMP stimulation (3–30 minutes) was ubiquitin immunoreactive.

Further investigation showed that SUMOylation is not required for MEK1 ubiquitylation and that, in contrast to SUMOylation, ubiquitylation does not require MEK1 activation. So the authors propose a situation in which, in response to cAMP, MEK1 is activated, SUMOylated and transported to the cell cortex. MEK1 inactivation — by an unknown process — precedes deSUMOylation and re-localization to the nucleus, where it is ubiquitylated, and thereby terminates the signal.