| dc.description.abstract | The min system comprised of MinC, MinD, and MinE in Escherichia coli ensures that cell division occurs at the midcell position by preventing the assembly of FtsZ into a Z-ring at the poles. MinD is a member of the deviant walker A motif family that dimerizes on the membrane in an ATP-dependent fashion. MinC forms an inhibitory complex with MinD on the membrane to antagonize Z-ring formation. MinE functions as a spatial regulator that displaces MinC from MinD and activates MinD ATPase. The dynamic interplay of the Min proteins culminates in a pole to pole oscillation by which a time-averaged MinCD concentration is the lowest at mid cell, thus allowing Z ring assembly there. MinD is at the heart of the Min system since MinD-ATP on the membrane recruits both MinC and MinE. In this study, MinD-D40A d10, an ATP hydrolysis-deficient MinD truncated for the C-terminal amphipathic helix involved in membrane binding, was crystallized in the presence of ATP. The structure resolved at 2.4Å resolution showed that MinD-ATP is a dimer. Furthermore, our mutagenesis studies demonstrate that the MinC and MinE binding sites form upon MinD dimerization and that MinE has a higher affinity for MinD than MinC. Prior to this study, E.coli MinE was thought to consist of two functionally autonomous domains. The N-terminal domain called anti-MinCD (MinE-CD) that suppresses MinCD activity is a nascent alpha helix. The C-terminal domain, known as the topological specificity domain (MinE-TSD) required for cell division at midcell, exists as a 4 beta stranded structure. However, recently determined structures of H. pylori and Neisseria gonorrhoeae MinE revealed that MinE exists as a 6 beta stranded form and part of MinE-CD is sequestered at the dimeric interface as a beta strand, thus raising question on how MinE interacts with MinD. We isolated MinE suppressor mutants that overcome some MinD mutants. These MinE mutants have substitutions at I24 position. Through a series of genetic and biochemical approaches we demonstrated that these substitutions for I24 release the sequestered part of MinE-CD, thereby converting the 6 beta to a 4 beta structure. The structures of MinD-D40A d10-MinE I24N and MinD-D40A d10-MinE peptide12-31, resolved at 4.2Å and 2.6Å resolution respectively, verified that MinE releases the beta strand upon MinD binding which is stabilized as an alpha helix at the MinD dimeric interface. In addition, we show that the N-terminal region of MinE-CD is a membrane targeting sequence (MTS) that is released during MinD-induced conformational alteration of MinE. Finally, we propose the Tarzan of the jungle model to explain how MinE can sequentially interact with multiple MinDs. MinE binding to MinD-ATP on the membrane triggers MinD ATPase activation, however, the mechanistic basis of the activation is still elusive. To get a sense of how MinE induces ATP hydrolysis in MinD, we compared the structure of MinD-D40A d10 with MinD-D40A d10 complexed with MinE12-31 peptide. Our analysis shows that MinE binding to MinD causes alterations in switch regions and conformational changes in some residues constituting MinDE binding interface. The MinD ATPase activation by MinE requires the binding of MinE-CD to the dimeric interface of a MinD dimer. Nonetheless, it was unknown whether MinE-CD binding to one side of the two dimeric interfaces is sufficient to stimulate MinD ATPase. To test this possibility, we created a MinD heterodimer composed of wild type MinD and a mutant form of MinD deficient in MinE binding. Our results show that both ATP molecules bound to a MinD heterodimer are hydrolyzed, suggesting that MinE-CD binding to one side of a MinD dimer induces ATP hydrolysis in both MinD subunits. Moreover, ATP hydrolysis was also observed in a heterodimer of the hydrolytic-deficient MinD-D40A mutant and the MinD mutant deficient in MinE binding. Taken together, we propose an asymmetric activation model where MinD hydrolyzes ATP upon MinE-CD binding to one side of the MinD dimer. | |
