Report

Magnetosomes Are Cell Membrane Invaginations Organized by the Actin-Like Protein MamK

See allHide authors and affiliations

Science  13 Jan 2006:
Vol. 311, Issue 5758, pp. 242-245
DOI: 10.1126/science.1123231

Abstract

Magnetosomes are membranous bacterial organelles sharing many features of eukaryotic organelles. Using electron cryotomography, we found that magnetosomes are invaginations of the cell membrane flanked by a network of cytoskeletal filaments. The filaments appeared to be composed of MamK, a homolog of the bacterial actin-like protein MreB, which formed filaments in vivo. In a mamK deletion strain, the magnetosome-associated cytoskeleton was absent and individual magnetosomes were no longer organized into chains. Thus, it seems that prokaryotes can use cytoskeletal filaments to position organelles within the cell.

Prokaryotes are highly organized cells with many ultrastructural similarities to eukaryotic cells (1). In addition to a dynamic cytoskeleton composed of homologs of actin, tubulin, and intermediate filaments (2), some prokaryotes also possess intracellular membranous organelles (36). The magnetosome chains of magnetotactic bacteria are among the best studied examples of membranous prokaryotic organelles. Magnetosome chains contain 15 to 20 ∼50-nm magnetite crystals that together act like the needle of a compass to orient magnetotactic bacteria in geomagnetic fields, thereby simplifying their search for their preferred microaerophilic environments (3). The distinct properties of magnetosomal magnetite crystals have drawn attention to their potential use in biotechnology, bioremediation, and geobiology (3). Furthermore, the cell biological characteristics of magnetosomes make them an ideal model system for the study of organelle biology in prokaryotes. Each magnetite crystal within a magnetosome is surrounded by a lipid bilayer, and specific soluble and transmembrane proteins are sorted to the magnetosome membrane (7, 8).

Electron cryotomography (ECT) is an emerging technology that can reveal the three-dimensional (3D) ultrastructure of small, intact prokaryotic cells in a near-native, “frozen-hydrated” state (9, 10). Here, we used ECT to define the precise physical characteristics of the magnetosome chain and to identify the molecular factors controlling its subcellular positioning and organization.

To image magnetotactic bacteria by ECT, exponential-phase cultures of Magnetospirillum magneticum sp. AMB-1 were plunge-frozen and imaged in a transmission electron cryomicroscope. Dual-axis image tilt series were recorded (11), from which 3D reconstructions of individual cells were calculated (Movie S1). A slice through the middle of one typical reconstruction (Fig. 1A and Movie S2) clearly showed the inner and outer membranes and the peptidoglycan layer. The cytoplasm was full of ribosomes, recognized as the numerous, dense, ∼25-nm particles. Large, approximately spherical poly-β-hydroxybutyrate storage granules, outer membrane blebs, and deep invaginations in the membranes were also apparent in all cells.

Fig. 1.

Electron cryotomography of Magnetospirillum magneticum sp. AMB-1 reveals that magnetosomes are invaginations of the inner membrane. (A) General features of AMB-1 cells highlighted in a 12-nm-thick section of an ECT reconstruction. Outer membrane, OM; inner membrane, IM; peptidoglycan layer, PG; ribosomes, R; outer membrane bleb, B; chemoreceptor bundle, CR; poly-β-hydroxybutyrate granule, PHB; gold fiduciary marker, G; magnetosome chain, MG. Scale bar, 500 nm. (B to E) Representative magnetosomes containing no magnetite (B), small (C), mediumsized (D), and fully-grown (E) crystals are invaginations of the inner membrane. Scale bar, 50 nm.

Of greatest interest, however, were the long chains of high-contrast magnetite crystals present in every cell and clearly surrounded by the magnetosome membrane. ECT images of cells grown in iron-poor conditions also provided a clear view of chains of empty magnetosomes (Fig. 2B and Movie S3) (12). Surprisingly, ∼34% (97 out of 283, in 15 different cells) of magnetosomes were clearly invaginations of the inner membrane rather than freestanding vesicles. These included magnetosomes in different locations within the chain containing magnetite crystals ranging in size from undetectable to full (Fig. 1, B to E). All of the remaining 66% of the magnetosomes were positioned close enough to the membrane to be invaginations, but most were located just above or below the inner membrane with respect to the electron beam, where the missing “pyramid” of data in reciprocal space blurred details of their connectivity. Thus, invagination from the cell membrane appears to be a permanent characteristic of the magnetosome and not just a step in the development of a cytoplasmic vesicle. This contrasts with the generally accepted view that magnetosomes are separated from the cell membrane (3). This idea stems from freeze-fracture electron microscopic studies in which no obvious connections to the cell membrane can be observed (7). The narrow membranous invaginations we observed here, however, would likely be missed by most fracture angles.

Fig. 2.

Magnetosome chains are flanked by long cytoskeletal filaments. (A) Larger view of the magnetosome chain in Fig. 1A. (B) Similar view of a magnetosome chain grown in the absence of iron, which prevents the formation of magnetite crystals. Arrows point to the long filaments. (C) Three-dimensional organization of magnetosomes (yellow) and their associated filaments (green) shown in (B) with respect to the whole cell (blue). Scale bars, 100 nm.

Although the observation that magnetosomes are invaginations ensures that they will always be close to the inner membrane, the observed linearity of magnetosome chains implies that some additional structural constraint has to be present. By inspecting the volumes surrounding each magnetosome chain, we discovered networks of long filaments 200 to 250 nm in length running parallel to four to five individual magnetosomes along the chain in all wild-type cells regardless of the growth condition (Fig. 2, A to C). At any one position within the chain, up to seven of these filaments flank the magnetosome with no obvious spatial pattern. Recent genetic, genomic, and proteomic analyses of magnetotactic bacteria have led to the identification of a large genomic island containing many of the genes known or suspected to be involved in magnetosome functioning (1214). In ECT reconstructions of a spontaneous nonmagnetic AMB-1 mutant missing this genomic island (15), neither magnetosome invaginations nor comparable filaments were detected, which suggests that the two structures were functionally as well as spatially related (16).

To identify the molecular components of the magnetosome cytoskeleton, we focused on the mamAB gene cluster that lies within the genomic region lost in the spontaneous nonmagnetic mutant (14). The products of the mamAB cluster are essential for magnetite production and are known to localize to magnetosomes (8, 13). One gene within this cluster, mamK, is predicted to code for a homolog of the bacterial actin-like protein MreB, which forms filaments and has been implicated in cell shape determination, establishment of cell polarity, and chromosome segregation (1719). Another homolog of MreB, ParM, also forms filaments and is involved in the proper segregation of certain naturally occurring plasmids (20). Although the genome of AMB-1 has not been fully sequenced, the closely related species Magnetospirillum magnetotacticum MS-1 and Magnetococcus MC-1 both contain a copy of mamK in their mamAB operons as well as a predicted mreB gene in operons with homologs of other genes involved in cell shape determination (16). MamK proteins are more similar to each other (the MS-1 and AMB-1 MamK proteins are identical in sequence) than they are to the MreB homologs found within their respective genomes (Fig. 3A). MreB, ParM, and MamK are also predicted to form three phylogenetically and functionally distinct groups of prokaryotic actin-like proteins.

Fig. 3.

MamK, a homolog of the bacterial actin-like protein MreB, forms filaments in vivo. (A) Phylogenetic relationship between MamK and other bacterial actin-like proteins demonstrated by an unrooted tree. These proteins separate into three distinct groups: MamK (green), ParM/StbA (red) and MreB (blue). (B) MamK fused to GFP (green) appears to form filaments in vivo localized to the inner curvature of the cell (cell membrane stained red with FM4-64).

MreB, ParM, and MamK also display different localization patterns within the cell. In Escherichia coli, Bacillus subtilis, and Caulobacter crescentus, MreB appears to form helical filaments adjacent to the cell membrane. In E. coli, ParM appears as an approximately straight line with no consistent relation to the membrane (2, 21). Here, green fluorescent protein (GFP) fusions to the C terminus of MamK were used to explore its subcellular localization. MamK-GFP appeared in straight lines extending across most of the cell approximately along its inner curvature (Fig. 3B), consistent with the magnetosome-associated filaments in both localization and extent. Notably, the magnetosome-associated filaments observed in the ECT reconstructions had a thickness of roughly 6 nm (∼2.5 voxels), just as expected for an F-actin-, ParM-, or MreB-like filament (21, 22). Thus, MamK is the most likely candidate for the magnetosome-associated cytoskeleton.

To test this idea, we generated a nonpolar, in-frame deletion of mamK in AMB-1. This mutant forms magnetite, turns in magnetic fields, and has no cell shape or growth defects compared with wild type (16). Whereas in ECT reconstructions the individual magnetosomes were still seen to be invaginations of the cell membrane, the mutant lacked the long, highly organized chains seen in wild-type AMB-1 (Fig. 4 and Movie S4). Instead, small groups of a few (two to three) neighboring magnetosomes separated by large gaps appeared dispersed throughout the cell, and no filaments comparable to those found in the wild-type cells were associated with the magnetosomes. To ensure that these defects were due to the absence of mamK alone, we complemented the ΔmamK cells with mamK-GFP and observed both the restoration of long, ordered magnetosome chains and the reappearance of magnetosome-associated filaments in approximately 15% of the cells (Fig. 4C). Many of the cells did not exhibit complete complementation, probably because of the expression of MamK-GFP from a heterologous promoter on a low-copy plasmid. Although we cannot eliminate the possibility that MamK is required for the filaments to form without actually being part of their structure, the simplest interpretation of all these results is that MamK is the long filament seen that organizes magnetosome membrane invaginations into a chain roughly parallel to the long axis of the cell.

Fig. 4.

MamK is required for the proper organization of the magnetosome chain. (A) Three-dimensional reconstruction of a wild-type AMB-1 cell. The cell membrane (gray), magnetosome membrane (yellow), magnetite (orange), and magnetosome-associated filaments (green) are rendered. (B) ΔmamK mutant, where magnetosomes appear disordered and no filaments are found in their vicinity. (C) ΔmamK cell expressing mamK-GFP on a plasmid showing full reversal of the mutant phenotype.

These results have important implications for the mechanism of magnetite synthesis within magnetosomes. The precursor of magnetite in magnetotactic bacteria is a ferrihydrite mineral that forms in the periplasm (23). The opening between magnetosomes and the periplasm might thus enable the direct transport of ferrihydrite between these two compartments. These results also expand our view of the bacterial cytoskeleton, showing that just as in eukaryotes, actin-like proteins are used to position membranous organelles (24). MamK could act in establishing the specific localization of the chain through the proper subcellular targeting of magnetosome biogenesis factors, or it could be important in the maintenance of the chain after the production of magnetosomes. Finally, our findings also present a problem in evaluating the evolutionary relationship between magnetosomes and eukaryotic intracellular organelles. Because magnetosomes do not separate from the cell membrane, they might be analogous to photosynthetic membranes in bacteria. They may also represent a step in the development of mechanisms for membrane bud formation before the invention of membrane fission.

Supporting Online Material

www.sciencemag.org/cgi/content/full/1123231/DC1

Materials and Methods

References

Movies S1 to S4

References and Notes

View Abstract

Navigate This Article