All in the Family: Kin Contact Leads to Outer Membrane Exchange

Abstract

The ability to recognize related cells in a population can confer evolutionary benefits. For example, some bacteria use contact-dependent inhibition proteins to distinguish kin from nonkin. Kinship recognition is taken to a new level in Myxococcus, which uses the dual-purpose TraA protein for kin recognition and outer membrane and lipoprotein exchange. In this issue of the Journal of Bacteriology, Wei et al. (X. Wei, C. N. Vassallo, D. T. Pathak, D. Wall, J. Bacteriol. 196:1807–1814, 2014) show that Tra-dependent exchange can be uncoupled from outer membrane vesicle/tube formation, reported elsewhere to mediate outer membrane exchange.

COMMENTARY

A broad range of microbes, including Gram-positive and Gram-negative bacteria, archaea, fungi, and protozoa (1, 2), produce bioactive membrane vesicles. Membrane vesicles from pathogenic and nonpathogenic Gram-negative bacteria, the focus of this commentary, have been known for decades. Typically, they are 30 to 100 nm in diameter, roughly the same size as exosomes from fungi and protozoa (3). The best-characterized vesicle-like structure is the outer membrane vesicle (OMV), which is secreted from the cell. OMVs are small, closed spheroids that form by a budding process at division septa or along the cell body (4, 5). Budding may be induced by exogenous compounds such as membrane-active antibiotics or signaling molecules such as the Pseudomonas quinolone signal (PQS) (6). The OMV can release its contents after lysis or upon fusion with another cell (4).

Although detailed information is just beginning to emerge about the regulation and genetic mechanisms that govern vesicle formation, it is clear that this is not a passive process. The marine photoautotroph Prochlorococcus can continuously release vesicles that are postulated to play roles in carbon cycling (7). Secreted OMVs have been reported to transport DNA, RNA, proteins, and small molecules that can serve as signals to nearby cells. Vesicles can be enriched for particular types of cargo, including insoluble materials, virulence factors, and antimicrobials, such as lytic agents and metabolites. Because virulence factors and antigens can be packaged in OMVs, interest in these vesicles for vaccine development has become an area of intense study. The ability of OMVs to promote cell adhesion, transport toxic compounds, and nucleate biofilm formation illustrates the wide-ranging impact of these structures (1, 8).

While many reports in the literature document OMVs that are simply released into the environment, it is clear that some microbes have devised novel ways to modify or enclose their OMVs. The soil bacterium Delftia sp. Cs1-4, unable to rely on aqueous environments for transfer of its OMVs, appears to transport them in structures that have been called nanopods (9). The nanopods, which resemble peashooters protruding from the cell surface, can exceed 6 μm in length. Electron cryotomography revealed that the nanopods contained structures that resemble OMVs. Another variation of the OMV structure is the nanotube identified recently in Bacillus (10). Like OMVs, nanotubes are composed of membranous material, have a diameter of roughly 30 to 100 nm, and appear to serve as conduits for both nucleic acid and protein.

Myxococcus xanthus, a Gram-negative soil bacterium, glides over surfaces preying on other organisms and, when starved for nutrients, forms a multicellular fruiting body containing heat-resistant spores. For decades, anecdotal reports of extracellular, vesicle-like structures in M. xanthus have persisted, along with speculation that these structures might be involved in predation, biofilm formation, sporulation, or cell signaling. Early evidence for potential vesicle cargo came from experiments documenting extracellular complementation of motility defects. Nudleman et al. (11) found that M. xanthus cells were able to transiently exchange CglB and Tgl, two outer membrane proteins needed for gliding motility, in a process that became known as outer membrane exchange (OME). Although OME required direct cell-cell contact, trace levels of protein exchange involving OMVs could not be ruled out. Hence, the quest to resolve the OME and OMV processes began.

In 2009, Palsdottir et al. (12) described small (30- to 60-nm) OMVs in the extracellular matrix of M. xanthus biofilm material. Electron tomography revealed that OMVs were tethered to each other and to the outer membrane. The organelle-like structures were either crescent or concentrically spherical. Subsequent proteomic analysis identified some of the components in these OMVs harvested from cells exposed to different environmental conditions. For example, protein subunits of the branched-chain keto acid dehydrogenase (BCKAD) complex were abundant in OMVs from cells that had been starved for nutrients (13). This is significant because BCKAD proteins generate a molecule called Esg that is required for cell-cell signaling during development (14). Hence, the finding that OMVs contain BCKAD proteins hinted that production or dissemination of the signal could be linked with OMV release during the starvation-induced development cycle (15).

To understand how Myxococcus cells transfer proteins, the Wall lab constructed a series of molecular tools to track the fate of CglB and Tgl, the outer membrane motility proteins known to undergo exchange (11, 16). When exchange of an mCherry fusion protein carrying a known OM type II signal sequence (OMII_ss-mCherry) was compared with an mCherry fusion carrying a known inner membrane signal sequence (IM_ss-mCherry), only the OMII_ss-mCherry fusion protein was exchanged between cells (16). While the original work showed efficient transfer of protein between cells (11), this was the first visual demonstration of rapid transfer of protein and OM lipid between M. xanthus cells (17). A significant breakthrough came when the Wall lab identified genes that are essential for OME. They showed that OME in myxobacteria requires the TraAB proteins, which must be present in both the donor and recipient, and that exchange of materials is bidirectional (17, 18). Exchange was rapid and efficient on a solid surface (Fig. 1A) but did not occur in liquid medium (Fig. 1B). Bidirectionality is a hallmark of Tra-mediated OME that distinguishes it from other transport systems, such as conjugation. Using tra mutants, the authors were able to demonstrate that OME can regulate cellular behaviors, including motility and developmental sporulation.

FIG 1.

Outer membrane exchange requires TraA. When donor and recipient cells with related traA alleles are mixed on a solid surface, outer membrane and lipoprotein are exchanged between cells in a process that may resemble SNARE-mediated fusion in eukaryotes (A). Exchange does not occur in liquid medium, a condition that favors production of OMVs/OMTs (B). OMVs/OMTs form in a traA mutant, but exchange does not occur (C and D). Exchange occurs only between M. xanthus strains that carry similar alleles of traA. Cells with incompatible TraAs do not exchange lipoprotein or outer membrane material (E).

Insight into the mechanism of lipoprotein and outer membrane transfer came, in part, from analysis of Tra proteins and the phenotype of M. xanthus strains overexpressing TraAB. TraA contains a PA14-like domain that may serve as a glycan-binding site. This domain is also conserved in floculins, the cell surface adhesins involved in flocculation of yeast. Similarity to floculins, plus the finding that cells with increased levels of TraAB stimulated cell-cell adherence, led the authors to propose that interaction between TraA and its ligand (presumably a glycan) could stimulate the formation of a channel between two cells (17). However, further analysis provided support for a different model in which adherence is mediated by TraA-TraA homotypic interactions (19). There were two critical findings that supported this model. First, TraA was shown to localize to the cell surface (19). Second, TraA was found to be strain specific. OME did not occur between myxobacterial strains with incompatible versions of TraA (Fig. 1E), but the OME fusion barrier could be overcome by swapping traA alleles. In other words, TraA-bearing cells undergo OME only with cells that express a highly similar form of TraA. This shows that part of TraA's role is kin recognition. In this capacity, traA acts as a “greenbeard” gene because its product is a receptor that recognizes cells that express the product of a related allele. The resulting OM fusion confers fitness to the partner cell because it gains beneficial characteristics such as bacteriocin resistance (19, 20).

After the initial discovery of TraA, potential channel-like structures were described by several labs within a short time. Remis et al. (21) described a network of OMVs and OM tubes (OMTs) that appeared to connect cells. Interestingly, proteomic analysis of samples enriched for OMVs yielded CglB and Tgl, the proteins previously shown to be transferred between cells via OME (11, 18). This extensive OMV/OMT network, shown in Fig. 2, was revealed by imaging techniques, including three-dimensional focused ion beam scanning electron microscopy. The network structures, which resemble a molecular freeway interchange, were highlighted in a 2013 New York Times science discovery video (http://www.nytimes.com/2013/09/24/science/scientific-discovery-frame-by-moving-frame.html?_r=0).

FIG 2.

Molecular tethers. A surface rendering of M. xanthus cells (yellow) in a biofilm reveals intercellular appendages (red) that arise from outer membrane vesicles (OMV) and outer membrane tubules (OMT). Appendages may enable cells to connect to one another in biofilms. These structures might facilitate recognition and allow cells to exchange molecules that are perceived as signals. (Reprinted from Environmental Microbiology [21] with permission of the publisher.)

Ducret et al. (22) also identified OMTs in Myxococcus while conducting experiments to examine the mechanism of OME. Following on the work of Wei et al. (16) and using the same molecular tools, they focused on videomicroscopy of individual cells to investigate OME. The authors found that OM proteins and lipids were transferred efficiently after transient cell contact, confirming the results from the Wall lab. Furthermore, they noted that tubular structures occasionally appeared when two cells that had undergone OME had moved apart from one another by gliding motility. The authors proposed that an outer membrane synapse (tube) was required for productive exchange of OM materials and argued that tubes form as a consequence of the tight membrane connection.

In their paper in this issue of the Journal of Bacteriology, the Wall lab weighs in on the subject of OMTs in Myxococcus (23), which they had observed during their OME investigations. Using a fluorescent lipoprotein reporter for the OM and lipid dyes, the authors observed ∼50-nm-wide tubes extending from the surface of M. xanthus cells. Their results confirm that M. xanthus cells form OMTs and further show that these structures sometimes appear as chains of OMVs. OMTs were also found in other Myxococcus isolates and species, which suggests that OMTs are likely a general property of myxobacteria. However, in contrast to previous reports, they provide genetic and environmental data that argue against a role for OMTs in OME. First, they show that environmental conditions that favor production of OMTs are incompatible with OME. OMTs were produced when cells were grown in liquid medium (without agitation), conditions that do not support OME. OMT levels were reduced on agar plates, whereas an agar surface is required for OME (Fig. 1A). Second, OME is abolished in tra mutants, but OMTs still form in abundance in liquid medium (Fig. 1D). Hence, the only known host proteins required for OME are not required for OMT production. These experiments show that OME can be uncoupled from OMT production, which argues strongly that OMTs are not required for OME. To date, screens for metabolic inhibitors that block OMT formation and OMT mutants have been unsuccessful. Based on the results presented by Wei et al. (23), it is reasonable to speculate that OMTs might be produced as a mechanism to relieve membrane stress, unlike the genetically based OME process.

The OME mystery is starting to unravel, and it is revealing unexpected results. Wei and coworkers have shown that TraA-dependent fusion events enable Myxococcus cells to share lipoproteins and outer membrane components needed for processes including motility, development, and adaptation to environmental stresses. Furthermore, they have obtained insights into a novel mechanism by which TraA enables a cell to distinguish between kin and nonkin. This TraA-dependent feature facilitates forging a coherent unit from a mixed population. In this regard, the myxobacteria exhibit a behavior typically thought to be reserved for development in eukaryotes. In addition, Wei and coworkers have noticed intriguing similarities between Tra-mediated OME and eukaryotic membrane fusion mediated by SNARE proteins that suggest that these processes may be analogous. SNARE proteins form large complexes called SNAREpins that can catalyze selective fusion events (24). Both SNAREs and TraAs recognize and interact with a specific target to promote membrane fusion, leading to transfer/exchange of materials. A deeper understanding of the Tra mechanism may reveal whether OME is functionally related and perhaps an evolutionary precursor to eukaryotic membrane fusion.