ATP-Binding Site Lesions in FtsE Impair Cell Division

FtsE and FtsX of Escherichia coli constitute an apparent ABC transporter that localizes to the septal ring. In the absence of FtsEX, cells divide poorly and several membrane proteins essential for cell division are largely absent from the septal ring, including FtsK, FtsQ, FtsI, and FtsN. These observations, together with the fact that ftsE and ftsX are cotranscribed with ftsY, which helps to target some proteins for insertion into the cytoplasmic membrane, suggested that FtsEX might contribute to insertion of division proteins into the membrane. Here we show that this hypothesis is probably wrong, because cells depleted of FtsEX had normal amounts of FtsK, FtsQ, FtsI, and FtsN in the membrane fraction. We also show that FtsX localizes to septal rings in cells that lack FtsE, arguing that FtsX targets the FtsEX complex to the ring. Nevertheless, both proteins had to be present to recruit further Fts proteins to the ring. Mutant FtsE proteins with lesions in the ATP-binding site supported septal ring assembly (when produced together with FtsX), but these rings constricted poorly. This finding implies that FtsEX uses ATP to facilitate constriction rather than assembly of the septal ring. Finally, topology analysis revealed that FtsX has only four transmembrane segments, none of which contains a charged amino acid. This structure is not what one would expect of a substrate-specific transmembrane channel, leading us to suggest that FtsEX is not really a transporter even though it probably has to hydrolyze ATP to support cell division.Cell division in Escherichia coli is carried out by ∼20 proteins that localize to the midcell, where they form a structure called the septal ring (also called the divisome or septalsome) (4, 27, 55, 57). One component of the septal ring is an apparent ABC transporter composed of the integral membrane protein FtsX and its associated cytoplasmic ATPase, FtsE (15, 47). FtsE and FtsX are widely conserved among gram-negative and gram-positive bacteria. ftsE and/or ftsX mutants exhibit division defects in E. coli, Neisseria gonorrhoeae, Aeromonas hydrophila, and Flavobacterium johnsoniae, indicating that FtsEX function in cell division is conserved in these organisms (33, 40, 43, 45). In contrast, FtsEX of Bacillus subtilis has no obvious role in cell division but instead regulates entry into sporulation (24).One interesting property of E. coli ftsEX null mutants is that they can be rescued by a variety of osmotic protectants (44). For example, when grown in LB containing >0.5% NaCl, an E. coli ftsEX null mutant is viable and only mildly filamentous, but upon shift to LB lacking NaCl, the cells become filamentous and die (20, 47). A shift to low-osmolarity medium is also accompanied by a dramatic slowing of the overall rate of growth (mass increase) (47). We suspect that ftsEX contributes to both cell division and growth, but it has proven difficult to exclude the possibility that the growth defect is caused by attempts at cell division that go awry.FtsEX contributes to cytokinesis by improving the assembly and/or stability of the septal ring. Septal ring assembly in an ftsEX mutant is fairly normal in LB that contains 1% NaCl but defective in LB that lacks NaCl (hereinafter referred to as LB0N) (47). More precisely, in LB0N, septal ring assemblies contain the “early” division proteins FtsZ, FtsA, and ZipA but lack the “late” proteins FtsK, FtsQ, FtsL, FtsI, and FtsN. The mechanism by which FtsEX contributes to septal ring assembly is still under investigation, but it probably involves protein-protein interactions, because FtsX has been shown to interact with FtsA and FtsQ in a bacterial two-hybrid system (31), while FtsE has been shown to interact with FtsZ in a coprecipitation assay (15). The FtsE-FtsZ interaction could be important for improving constriction rather than, or in addition to, septal ring assembly.Remarkably, nothing is known about FtsEX''s most obvious potential function—transporting a substrate involved in septum assembly. We are aware of only two studies that attempted to address this issue. The first concluded that FtsEX is needed for insertion of potassium transporters in the cytoplasmic membrane (54). However, in our view the data were not compelling and the connection to cell division, if any, is not obvious. The other study noted that ftsE and ftsX are cotranscribed with ftsY, which is a component of the signal recognition particle pathway for insertion of many proteins into the cytoplasmic membrane (20). That study therefore tested an ftsEX null mutant for defects in export of β-lactamase to the periplasm or insertion of leader peptidase into the cytoplasmic membrane. No such defects were found, so the authors concluded that FtsEX function is probably unrelated to FtsY. Besides these studies, at least one review article suggested that FtsEX might insert division proteins into the cytoplasmic membrane (8). Obviously, the finding that localization of several membrane proteins to the septal ring shows a leaky but pronounced dependence on FtsEX could be explained if the “missing” proteins were not getting into the membrane efficiently. Despite these speculations about potential FtsEX substrates, it is important to note that some members of the ABC “transporter” family do not move anything across cell membranes (reviewed in reference 19), so it cannot be taken for granted that FtsEX is a transporter at all.