A bacterial actin unites to divide bacterial cells

Abstract

FtsA is an actin homolog required for cytokinesis in many bacterial species and has several key roles in cell division, including helping to tether FtsZ to the cytoplasmic membrane via a membrane-targeting sequence (MTS), recruiting other essential proteins to the divisome, and perhaps promoting divisome constriction (de Boer, 2010). Szwedziak et al (2012) recapitulate the FtsZ-FtsA-membrane association in vitro using liposomes with FtsZ and FtsA proteins from Thermotoga maritima. To get a closer look at the FtsA-FtsZ interface, the authors co-crystallize FtsA with the carboxy-terminal tail of FtsZ, which is known to interact with FtsA. Intriguingly, the crystal reveals an FtsA homodimer. Contrary to the previous bioinformatics model of FtsA self-interaction that proposed a 180° rotation between the two subunits (Carettoni et al, 2003), the FtsA-FtsA interface in the crystal structure shows no rotation, similar to F-actin. Szwedziak et al (2012) also show that FtsA can form longer, actin-like polymers in the presence of non-hydrolysable ATP or on lipid monolayers. These results are surprising because FtsA has a divergent subdomain architecture compared to other actin-family proteins (van den Ent and Lowe). A critical question now is whether FtsA needs to form polymers in vivo to function properly. Purified Streptococcus pneumoniae FtsA assembles into large polymers that are not like F-actin, and it remains unclear if these structures are relevant in vivo (Krupka et al, 2012). Wild-type FtsA proteins do not form detectable filaments in cells, but C-terminal truncations of FtsA that remove the MTS form polymers quite readily in cells when overproduced, although they are not functional (Pichoff and Lutkenhaus, 2007). Even so, starting with an MTS truncation derivative of FtsA to visualize in vivo polymers, Szwedziak et al (2012) design site-directed mutants of Bacillus subtilis FtsA based on the FtsA-FtsA interface of their crystals; these fail to assemble into polymers in vivo. Using a similar MTS truncation derivative, Pichoff et al (2012) created random mutations in Escherichia coli FtsA, and found that those mapping to the same interface found by Szwedziak et al (2012) also disrupted polymer formation. Together, these data suggest that these residues are needed for FtsA self-interaction. Perplexingly, when these mutants were subsequently tested for functionality in the context of full-length FtsA, the results were mixed. Pichoff et al (2012) showed that FtsA mutants deficient for self-interaction in E. coli have a gain-of-function phenotype, whereas Szwedziak et al (2012) report that analogous mutants in B. subtilis FtsA suffer a loss of function. These results support the idea that FtsA self-association is related to its activity (Shiomi and Margolin, 2007), yet understanding how self-interaction regulates FtsA function clearly requires further study. The ability of eukaryotic cytoskeletal proteins to form long polymers is essential to their function, but the physiological relevance of long polymer formation by bacterial cytoskeletal proteins is now a topic of debate (Figure 1). For example, it has been hypothesized that FtsZ protofilaments wrap around the entire circumference of the cell to form the cytokinetic ring. However, recent studies using photoactivated localization microscopy (PALM) and electron cryotomography reveal a different model in which FtsZ forms a series of very short polymers that overlap to encompass the diameter of the cell (Li et al, 2007; Fu et al, 2010). MreB was also originally thought to form long-range helical polymers extending the length of the cell, but recent data obtained with more sophisticated microscopic techniques suggest that MreB is distributed in patches that move circumferentially and independently (White and Gober, 2012). It is not yet clear which of these models represents the true cellular architecture of MreB, although it is likely that some degree of MreB polymerization is still needed for function. It is notable that other bacterial homologs of actin and tubulin used for generating scaffolds or partitioning plasmid DNA, but not for essential cellular processes such as cell division and growth, tend to form long polymers that extend throughout the cell (Pogliano, 2008). The continued combined use of microscopic, biochemical, and genetic methods, as demonstrated by Szwedziak et al (2012) will enhance future understanding of ancestral tubulin and actin proteins in prokaryotes. Figure 1 Bacterial actin and tubulin filaments involved in cell growth and division. (A) MreB (purple) has long been thought of as a spiral filament twisting along the cell length to control cell shape. Likewise, FtsZ protofilaments (blue) were once thought to ...