SEPT2 (green) turns polyglutamated microtubules (red, top) into high speed tracks. Epithelial cells (middle) no longer polarize when SEPT2 is missing (bottom).
Septins were first linked to polarity in budding yeast, where they form diffusion barriers and scaffolds for congregating polarity proteins at the yeast cortex. But, based on the new findings, mammalian septins may work differently by altering microtubule properties.
In epithelial cells, septins bound along microtubules near the trans-Golgi network, at exit sites for vesicles carrying polarity proteins to the apical or basolateral plasma membrane. Disruption of one septin, SEPT2, caused vesicles to accumulate in the cytoplasm and thereby prevented cell polarization.
SEPT2 seemed to help polarity proteins get to the surface by removing impediments on the microtubule tracks. These cytoskeletal speed bumps are created by microtubule-associated proteins such as MAP4. But MAP4 was less likely to hop onto SEPT2-bound tubulin, the group found. The two proteins probably compete for similar binding sites on microtubules.
Without MAP4, the SEPT2-decorated tracks are presumably freed for high speed transport. Such speed is probably needed for the intense, dynamic membrane expansion that drives polarization.
Only some microtubule tracks harbored SEPT2, which preferred filaments containing polyglutamated tubulin. In turn, SEPT2 was needed to maintain polyglutamation. Increases in this posttranslational modification might boost trafficking in developing epithelia and in other cell types that perform heavy amounts of regulated vesicle secretion, such as neurons and immune cells.
In addition to the microtubule pool, some SEPT2 was also seen on vesicles. As septins form oligomers, the two pools might velcro together vesicles and microtubules. The authors are also considering the possibility that different mammalian septins, like the various Rab GTPases, may define distinct post-Golgi trafficking pathways—to the apical versus basolateral membranes, for example.
Reference: