Cryo-Electron Tomography Reveals the Cytoskeletal Structure of Spiroplasma melliferum

See allHide authors and affiliations

Science  21 Jan 2005:
Vol. 307, Issue 5708, pp. 436-438
DOI: 10.1126/science.1104031


Evidence has accumulated recently that not only eukaryotes but also bacteria can have a cytoskeleton. We used cryo–electron tomography to study the three-dimensional structure of Spiroplasma melliferum cells in a close-to-native state at ∼4-nanometer resolution. We showed that these cells possess two types of filaments arranged in three parallel ribbons underneath the cell membrane. These two filamentous structures are built of the fibril protein and possibly the actin-like protein MreB. On the basis of our structural data, we could model the motility modes of these cells and explain how helical Mollicutes can propel themselves by means of coordinated length changes of their cytoskeletal ribbons.

One of the key functions of a cytoskeleton is to determine and maintain the shape of cells. In bacteria, the cell wall is generally considered to be the primary determinant of cell shape. However, this has been called into question, most recently by the discovery of the Mollicutes (Mycoplasma, Acholeplasma, and Spiroplasma) (1). Mollicutes are the smallest and simplest free-living and self-replicating cells within prokaryotes, and they are enveloped only by a cholesterol-containing cell membrane. Despite the lack of a cell wall, these cells have distinct morphologies, and their peculiar mode of movement that occurs in the absence of appendages normally implicated in motility (e.g., flagella or secretion organelles) makes the existence of an internal cytoskeleton likely.

Williamson (2) first reported the isolation of a cytoskeleton upon cell lysis and sodium deoxycholate extraction of Spiroplasma isolated from Drosophila. Cytoskeletal elements have also been released from Spiroplasma cells by repeated freezing and thawing (3). Townsend et al. (3) also developed methods for purification of the membrane-associated fibrils from S. citri and suggested that they may be involved in motility. Electron microscopy and SDS gel electrophoresis have shown that S. melliferum possesses a cytoskeletal ribbon comprosed of ∼4- to 5-nm-wide fibrils with an axial repeat of ∼9 nm; these fibrils form pairs within the ribbon and are composed of the 55-kD fibril protein (3, 4). Townsend and Plaskitt (5) further used immunogold staining of thin sections with an antibody to p55 to localize the ribbon built of the fibril protein within S. melliferum cells. More recently, the structure and localization of this cytoskeleton was described as a flat, monolayered, membrane-bound ribbon composed of six or seven pairs of fibrils following the inner, shortest helical path of the cell (6, 7). The fibril protein, which is completely unrelated to any other eukaryotic or prokaryotic protein and has been found exclusively in the genome of Spiroplasma (8), was identified as the only component of this ribbon (7). Because this protein would have a diameter of about 5 nm, assuming a spherical shape, it was proposed that the subunit of the fibrils is a tetramer and that the filaments form pairs to give the 10-nm axial and lateral spacings in the cytoskeletal ribbon (9).

Our tomographic studies (10, 11) of intact vitrified S. melliferum cells, however, reveal a more complex pattern of filamentous structures just underneath the cell membrane (Fig. 1). Two types of filaments (thin and thick) are arranged in parallel, anchored to each other and to the cell membrane, and jointly form three ribbons that span the entire cell from the blunt to the tapered end. By calculating the Hough transformation (12) of the filament regions in several tomograms and finding the position of the highest contrast, we determined the number and spacings of the individual filaments (Fig. 2). The two outer ribbons consist of five thick filaments with a spacing of ∼11 nm. These two ribbons are joined together by nine thinner filaments with a spacing of ∼4 nm. Depending on the extent of twisting of the cell and the position of the filaments with regard to the direction of the electron beam during data acquisition, five or locally fewer than five thick filaments are visible in the tomograms. Using three-dimensional (3D) visualization (11), we were able to illustrate the arrangement and path of the two outer ribbons underneath the cell membrane through the entire cell body (Fig. 3 and movie S1). Although the ribbons cannot be visualized as continuous bands along the cell membrane of the whole cell because of the “missing wedge” problem (13), it becomes clear that the ribbons follow, in parallel, a helical path from one end of the cell to the other. By calculating the geodetic line (14) (i.e., the fastest connection between two arbitrary points on the 3D cell membrane), we could also show that one of the two outer ribbons is locally shorter; indicating differential length changes of the two ribbons (Fig. 3 and movie S1).

Fig. 1.

Superimposed slices from different z heights of a tomogram (A) (z heights indicated by bounding box) and two corresponding 3D visualizations (B and C) of part of a S. melliferum cell showing the arrangement and course of the cytoskeleton. (A) The cytoskeleton is composed of two outer ribbons of thick filaments and a ribbon of thin filaments sandwiched in between. (B) Simplified 3D representation of the filament ribbons (green, purple, and red) that wind in parallel helically around the cell just underneath the cell membrane (blue), showing, in this case, the left-handed course of the cytoskeleton. (C) Idealized visualization of the filaments with smooth transition to the original data (yellow).

Fig. 2.

Sections (2.7-nm thick) of parts of S. melliferum cells derived from the tomograms (left) and corresponding profiles of the filament grey values (right) showing the number and spacings of the individual filaments. (Top) The two outer ribbons are composed of five thick filaments with a spacing of about 11 nm (highlighted in green and red). Depending on the extent of twisting of the cell, five or locally fewer than five thick filaments are visible in the tomogram. (Bottom) The region in between is composed of nine thin filaments with a spacing of about 4 nm (highlighted in purple). Each cross in the graphs corresponds to one pixel (pixel size: top, 1.72 nm; bottom, 0.68 nm). (In the bottom graph, the values on each side of the inner ribbon were set to zero.) Scale bars, 100 nm.

Fig. 3.

Three-dimensional visualization of a S. melliferum cell. The localization and course of the two outer ribbons within the cell are illustrated in red and green. The position of the geodetic line is shown in yellow. The red ribbon is locally overlapping with the geodetic line, which indicates that, in this case, it is shorter than the green one. This is most probably accomplished by simultaneous conformational changes of adjacent subunits of the five filaments. The length differences explain the helicity of the cell, as well as the ability to alternate handedness between left- and right-handed (movie S1).

What is the protein composition of the ribbons? We have been able to show that isolated and purified filaments (11) composed of the fibril protein exist in pairs with a total thickness of about 10 nm; hence, we suppose that the two outer ribbons of the S. melliferum cytoskeleton are made of the fibril protein. Our structural data showing two different filaments in the cytoskeleton of S. melliferum raises the question of the identity of the second protein. Until recently, there has been an ongoing controversy about the existence of actin or actin-like proteins in prokaryotes, including Mollicutes [for a review, see (15)]. Bacterial homologs of actin were first identified when the cell-shape determinants MreB and Mbl (MreB-like) were shown to assemble into helical filamentous structures that run in spirals around the periphery of the cell under the cytoplasmic membrane in Bacillus subtilis (16). Depletion of MreB induced the formation of rounded inflated cells that was ultimately lethal. This phenomenon was also reported for Escherichia coli (17, 18). MreB has been found only in bacterial species with rod-shaped, filamentous, or helical cells (16). Consistent with this, the helical cells of S. citri have five MreB homologs (19). Using Western blot and sequence analysis (11), we showed that both the fibril protein and MreB exist in S. melliferum cells and both are homologous to the sequenced fibril and MreB proteins of S. citri (supporting online material text). Hence, we suggest that the inner ribbon of the S. melliferum cytoskeleton is composed of MreB. Because of the instability of MreB filaments, the isolation of intact triple ribbons proved difficult and therefore, direct and unambiguous proof for this assignment by immunolabeling was not possible. Nevertheless, there is evidence supporting our assumption about the protein composition of the cytoskeleton. Purified MreB from Thermotoga maritima has been shown to form ∼4-nm-wide protofilaments in vitro (20), and a linkage between an actin-like protein and the fibrils in Spiroplasma cells was already proposed by Williamson et al. (4) but has not yet been proven.

How does this cytoskeleton enable the movements of S. melliferum cells? Trachtenberg and Gilad (7) suggested that the filaments can change their length in a coordinated manner, driven by conformational changes of their tetrameric subunits from nearly circular to elliptical. In this regard, the assumptions for our computer simulations (11), which are based on light microscopy experiments (21) and the existence of three ribbons, are (i) the ribbons are connected to each other, (ii) the inner thin filament ribbon functions elastically during cell movement, and (iii) length changes of the five filaments of the outer two ribbons occur simultaneously by switching between two conformational states of the filament subunits. Hence, we can simulate on a molecular level how the different motility modes are generated. If length changes in the two outer ribbons are coordinated such that one ribbon becomes gradually tense by shortening while the second one concurrently relaxes, the cell would alternate handedness between left- and right-handed, with handedness switching at the position where the state of tension and relaxation of the ribbons reverses (Fig. 4 and movie S2). This point of transition between left- and right-handedness travels throughout the whole cell, leading to a motion that resembles that of a bacterial flagellar bundle or of a single eukaryotic flagellum or cilium. A similar motion can be generated if the distance between the two outer ribbons is changed locally, thereby creating a deformation or kink, which propagates throughout the cell body. Thus, the whole Spiroplasma cell functions as a dynamic helical propeller. Coupling of a biochemical cycle (e.g., adenosine 5′-triphosphate hydrolysis) to the dynamics of the filaments could enable these filaments to propagate deformations that generate propulsive forces which, in turn, can drive cell motion (22). We assume that MreB filaments give the cell a rodlike shape by forming the inner, elastic ribbon, whereas the outer filaments composed of the fibril protein enable the formation of a helix, as well as movement, by interaction with the MreB ribbon.

Fig. 4.

Computer simulation explaining the change of handedness from left-handed (right) to right-handed (left). If the cell has a certain handedness over its entire length, then one of the two outer ribbons (red and green) is tense, and the other one is relaxed. If the tense, short ribbon gradually lengthens while the other one concurrently shortens its length (inset), the handedness switches, and the point of transition between the two states travels through the whole cell, producing a thrust in the opposite direction. The inset shows the length differences between the two outer filament ribbons (red and green) on each side of the point of transition (cross-over point). The inner ribbon, which does not change its length, is shown in purple (movie S2).

In the future, it will be necessary to elucidate in more detail the nature of the interaction between the individual filaments, as well as between the cytoskeleton and the cell membrane. Finally, the fine structure of the filaments and the mechanisms underlying changes in filament length and distance, as well as their driving force, remain to be clarified.

Supporting Online Materials

Materials and Methods

SOM Text

Movies S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article