The article by Khemici et al. in this issue of Molecular Microbiology (Khemici et al., 2008) reports the first direct demonstration that ribonuclease E (RNase E) is a membrane-binding protein, as shown by its ability to directly bind phospholipid bilayers in vitro. Binding occurred via an amphipathic α-helix that anchors the protein in the bilayer structure. This finding is of special significance because of the recent discovery that RNase E and the other protein components of the RNA degradosome in intact cells are organized as cytoskeletal-like elements that coil around the cell cylinder to form long-range ordered helical structures (Fig. 1) (Taghbalout and Rothfield, 2007). The coiled structures are thought to be associated with the cytoplasmic membrane based on the peripheral cellular localization of fluorescently labelled degradosome proteins (Taghbalout and Rothfield, 2007) and the results of cell fractionation and immunoelectronmicroscopic studies of RNase E (Miczak et al., 1991; Liou et al., 2001). The present work by Khemici et al. lends strong support to this assumption and implies that the degradosome cytoskeletal-like structure is attached to the membrane by the direct interaction of the RNase E amphipathic anchor with the membrane bilayer (Fig. 2A). This is consistent with the previous evidence that RNase E can form the membrane-associated helical structures in cells that lack the other degradosome proteins (Taghbalout and Rothfield, 2008).
Previous studies showed that RNase E contains a cytoskeletal localization domain [RNase E (418–602)] that is necessary and sufficient for formation of the helical cytoskeletal structures (Fig. 2B). Significantly, the present work by Khemici et al. showed that the RNase E (565–585) domain (Segment A), which contains the amphipathic membrane anchor, lies within the RNase E (418–602) cytoskeletal localization domain (Fig. 2B). Deletion of Segment A or replacements of hydrophobic residues of the amphipathic helix caused a change in the localization pattern of RNase E, with either loss of peripheral localization or the appearance of clumps of RNase E in peripheral regions of the cell. This suggests that membrane association of the degradosome cytoskeletal structure is mediated, at least in part, by the amphipathic α-helix of RNase E Segment A. The role of the α-helix itself has not been completely elucidated and it will be of interest in higher resolution studies of the RNase E helical structure to define further the effects of the mutations within the amphipathic helix.
The RNA helicase B (RhlB) component of the degradosome can also assemble into helical cytoskeletal structures in the absence of RNase E and the other degradosome proteins (Taghbalout and Rothfield, 2008). The RhlB helices were indistinguishable from the RhlB-RNase E structures that were colocalized in wild-type cells. These and other experiments have suggested that the filamentous helical structures of the RNA degradosome are composed of separate RNase E- and RhlB-coiled filaments to which the other proteins [polynucleotide phosphorylase (PNPase) and enolase] are secondarily attached (Fig. 2A). The fact that the RhlB-coiled structures can be assembled in the absence of RNase E makes it likely that the RhlB component of the complete structure also interacts with the membrane. This might occur via direct interaction of RhlB with the membrane phospholipid bilayer, as is the case for RNase E, or by interaction with another as yet unidentified membrane component (Fig. 2A).
The use of a covalently attached amphipathic domain to attach a cytoskeletal-like structure to the cytoplasmic membrane appears to be a recurrent theme. The membrane association of the MinD helical cytoskeletal structure that is part of the division site localization system also occurs via a short amphipathic α-helical membrane targeting sequence (MTS) (Hu and Lutkenhaus, 2003; Szeto et al., 2003; Taghbalout et al., 2006). Similarly, the FtsZ cytokinetic ring is anchored to the membrane by an amphipathic segment within the FtsZ-associated FtsA protein that is closely related to the MinD membrane anchor (Pichoff and Lutkenhaus, 2005).
What is the functional significance of the membrane-anchored cytoskeletal organization of RNase E and the other components of the RNA degradosome? It has previously been shown that loss of the RNase E cytoskeletal localization domain is associated with significant defects in growth rate, cell division, chromosome segregation, cell morphology and autoregulation of rne expression (Taghbalout and Rothfield, 2007). The present work of Khemici et al. showed that deletion of Segment A alone or substitutions of key residues of the amphipathic domain also affected cell growth and abolished the autoregulation of rne expression. All of these observations imply that the membrane association and/or cytoskeletal-like organization of RNase E plays an important role in the life of the cell, presumably by affecting cellular RNA processing and degradation. Consistent with this idea, previous studies have shown an increased half-life of several mRNAs and a defect in the processing of rRNA and tRNA in cells that express only the RNase E (1–417) fragment, which lacks the cytoskeletal localization domain (Fig. 2B) (Ow and Kushner, 2002).
It is attractive to imagine that the defects associated with loss of the RNase E–membrane association reflect loss of the normal sequestration of the degradosome components within the cytoskeletal structures. However, there are several alternative explanations that could also explain these observations, and these are not mutually exclusive.
- (i) As suggested above, normal RNA decay and processing might require the formation of long-range ordered degradosomal cytoskeletal structures. These could act by compartmentalizing degradosomal enzymes within the cell, thereby facilitating the regulation of RNA processing and degradation by preventing free access of the enzymes to RNA substrates in the cytoplasmic compartment.
- (ii) Membrane association of RNase E might affect RNase E activity directly by inducing conformational changes that are required for normal RNase E function within the cell. The conformational changes could include or be in addition to the 2.5-fold increase in helical content that occurred when RNase E Segment A was mixed with phospholipids in vitro in the present studies of Khemici et al. However, this is made unlikely by their observation that the presence of phospholipids did not change the endoribonuclease activity of RNase E in vitro.
- (iii) The phenotypic defects might be due to a decreased intrinsic endoribonuclease activity of RNase E in which Segment A was altered by mutations or in which the cytoskeletal localization domain or Segment A was removed. These structural changes might alter the activity of RNase E, despite the fact that the minimal catalytic domain lies outside the deleted or altered regions (Caruthers et al., 2006).
The article by Khemici et al. marks a significant advance in understanding the organization of membrane-associated cytoskeletal structures within the bacterial cell. It is clear that additional work will be needed to characterize the functional role of these RNase E-associated structures in the life of the cell.