Fluidity of the bacterial outer membrane is species specific

Summary

The outer membrane (OM) is an essential barrier that guards Gram-negative bacteria from diverse environmental insults. Besides functioning as a chemical gatekeeper, the OM also contributes towards the strength and stiffness of cells and allows them to sustain mechanical stress. Largely influenced by studies of Escherichia coli, the OM is viewed as a rigid barrier where OM proteins and lipopolysaccharides display restricted mobility. Here we extend the discussion to other bacterial species, with a focus on Myxococcus xanthus. In contrast to the rigid OM paradigm, myxobacteria possess a relatively fluid OM. We conclude that the fluidity of the OM varies across environmental species, which is likely linked to their evolution and adaptation to specific ecological niches. Importantly, a fluid OM can endow bacteria with distinct functions for cell-cell and cell-environment interactions.

Graphical Abstract

The outer membrane (OM) protects Gram-negative bacteria from various chemical and mechanical insults. The OM is generally viewed as a rigid barrier where the motion of its components are restricted. Here we suggest that some bacterial species harbor a relatively fluid OM, which influences their cellular functions and how they interact with the external world.

Introduction

Bacteria inhabit almost every ecological niche on Earth. Their survival and proliferation relies on their ability to cope with rapidly changing environments. To protect themselves from various chemical and physical insults, bacteria evolved a barrier, the cell envelope. In Gram-negative bacteria, the cell envelope contains essential components, including the outer membrane (OM), which encases the inner membrane (IM, aka cytoplasmic membrane) with a thin layer of peptidoglycan sandwiched in between. Besides acting as a gatekeeper to selectively transport nutrients and other molecules, the OM allows cells to resist mechanical stress and contributes to the stiffness and integrity of cells.^[1–3] In addition, the OM harbors proteins that enable critical functions in motility, virulence, and adhesion, among many others.^[4–6]

The composition and biogenesis of the OM has been extensively studied over the last half century.^[7] These works provided fundamental insights into the chemical, physical and structural properties of the OM along with how the components assemble. More recently researchers have begun to revisit the mobility dynamics of OM proteins and lipids.^[8–11] Here we extend the existing view of OM dynamics to different bacteria and suggest OM fluidity is species specific. In addition, the variation in OM dynamics reflects different strategies bacteria use to adapt to different environments and how they interact with other cells.

An asymmetric membrane and the Escherichia coli paradigm

A central feature of the OM is its unique asymmetry; the inner leaflet is composed of phospholipids and the outer leaflet is typically composed of glycolipids (Fig.1). This strict asymmetry is important for the barrier functions and rigidity of the OM.^{[2, 12]} Specifically, lipopolysaccharide (LPS) fortifies the OM, hence protecting the cell against entry of antibiotics, detergents and other toxic molecules. In addition to LPS, the OM harbors integral OM proteins (OMPs) and lipoproteins that maintain membrane integrity and fundamental cellular functions. OMPs typically consist of antiparallel β-strands that fold into barrel-like structures that can serve as nutrient transporters, whereas lipoproteins are typically peripherally anchored in the phospholipid bilayer where they reside in the periplasm and, occasionally are found extending extracellularly. Recent research has advanced our understanding of OMP and LPS biogenesis and their assembly in the model organism E. coli.^[13] Parallel work has advanced our understanding of how these OM components are spatially and temporally organized.^[14] A key feature of the current paradigm is that the outer leaflet harbors components with restricted lateral mobility.^{[8, 14–16]} That is, the mobility of OMP and LPS is confined to relatively small patches or islands in the OM. This paradigm sharply contrasts with the fluid mosaic model found in the cytoplasmic membrane. However, this view of the OM is primarily derived from studies in E. coli. Here we focus on the fluidity of the best-studied components, LPS and OMPs, since the dynamics of other components, such as OM lipoproteins, remain relatively underexplored.

Figure 1. Structure and molecular dynamics of the outer membrane.

(1) LPS containing hexa-acylated lipid A and long O-antigen side chain is less fluid. Non-covalent cross-bridging of neighboring LPS molecules by divalent cations (e.g. Ca²⁺) contributes to the OM rigidity. (2) LPS containing penta-acylated lipid A and short O-antigen side chain is more fluid. (3) OMP islands formed by promiscuous interactions between OMPs display restricted mobility.^[8] Old OMP islands (predominantly at cell poles) and new OMP islands (predominantly at mid-cells) are color-coded. (4) Top view of the interaction between trimeric porin (blue) and LPS (red).^[39] One interaction site is stabilized by Ca²⁺ (yellow). (5) Gliding motility machinery in myxobacteria. Proton motive force indicated with an arrow. Helical movement of motility motors are shown. (6) TraA, TraB, OM lipoproteins and LPS are fluid in myxobacteria.

The fluidity of LPS

A key feature of LPS is that it contributes to OM barrier function and is essential for viability in most Gram-negative bacteria. The dynamic properties of LPS are intrinsically defined by its chemical composition, which includes lipid A, core oligosaccharide, and O-antigen (Fig. 1). LPS is generally negatively charged and the repulsive forces between adjacent molecules are stabilized by divalent cations (Mg²⁺, Ca²⁺) that form non-covalent cross-bridges between the core oligosaccharide regions (Fig. 1).^[17] These cross-bridges contribute to the rigid properties of LPS where, for example, after LPS patches are added in vitro to phospholipid bilayers, the LPS domains persist for days before becoming mixed with phospholipids.^[2] The fatty-acyl chains of lipid A also affect LPS fluidity. For example, penta-acylated lipid A provides increased fluidity compared to hexa-acylated lipid A,^{[12, 18–20]} where interchain forces regulate the mobility of fatty-acyl chains and hence the LPS monolayer (Fig. 1). Additionally, long O-antigen side chains decrease OM fluidity compared to LPS molecules with short or no side chains (Fig. 1).^{[2, 21]} In support of this, the lateral diffusion of OM proteins is faster in bacteria with truncated O-side chains.^{[2, 22]}

OMP fluidity

Advanced microscopic approaches revealed in E. coli that although OMPs exhibit local diffusion, they cannot undergo long-range diffusion across the entire surface of the OM.^{[8, 14]} One explanation for this is that promiscuous interactions between OMPs, mediated by aromatic residues in their β-barrels, restrict their long-range lateral diffusions.^[8] That is, although both the IM and OM contain a high concentration of integral membrane proteins, the β-barrel proteins, which are exclusively found in the OM, tend to have more non-specific interactions that self-associate and form large proteinaceous islands that restrict their long-range mobility (Fig. 1). In support of this, when the β-barrel assembly machine (BAM), which governs the insertion and folding of OMP in the OM, is depleted, the concentration of OMPs decreases with a corresponding increase in membrane fluidity.^[23]

The spatiotemporal dynamics of OMP islands have been investigated using a variety of approaches.^{[8, 14, 15]} From these and other studies,^[24–26] it is generally accepted that OMP distribution across the OM is not homogeneous, which is partially explained by the time dependence of OMP biogenesis and the restricted mobility of large OMP islands. Here, as found in E. coli, new OMP biosynthesis occurs predominantly at mid-cell or stochastically at non-pole areas and, as a result, old OMP islands are pushed and accumulate towards the poles as cells grow and deposit new OMPs (Fig. 1). This results in spatial segregation of old and new OMP islands. Consequently, after two generations some daughter cells contain new OMPs at both poles.^[8] However, daughter cells that retain old and damaged OMPs likely undergo cell aging and senescence, since there is no known mechanism to remove them,^[27] other than by shedding OM vesicles.^{[28, 29]}

The constrained motion of OMPs suggests that the OM may physically contribute to the mechanics and strength of the cell. In fact, recent works demonstrated that the OM, in conjunction with the peptidoglycan layer, function as fundamental load-bearing structures.^{[1, 3, 30–32]} In E. coli the OM and cell wall are extensively bonded together by hundreds of thousands of proteins (primarily consisting of covalently-bonded Lpp lipoproteins, and OM-embedded proteins harboring OmpA domains that non-covalently bind the cell wall, e.g. Pal), where the OM plays an important role in sustaining integrity and stiffness of the cell. This point raises a new question: How do different OM components contribute to the mechanics of the cell? A recent study suggests that OMPs are the most rigid component of the cell envelope,^[33] highlighting their importance in providing a scaffold to withstand physical stress.

Interface between LPS and OMPs in mediating OM fluidity

Cell membranes are crowded environments that harbor high concentrations of proteins.^{[34, 35]} However, there are gaps in our understanding of how lipid-protein interactions govern membrane organization.^[36] In the OM, LPS and phospholipids create a membrane layer where OMPs fold and are embedded. Interestingly, E. coli cells cannot simultaneously tolerate truncated LPS and defects in the OMP assembly machinery (BAM),^{[23, 37]} and synthetic antibiotics targeting both LPS and OMPs show increased potency,^[38] suggesting that LPS may synergize with OMPs to fortify the OM barrier. For instance, a study of trimeric porins in E. coli found they form high affinity complexes with LPS through specific binding sites that maintains OM barrier function (Fig. 1).^[39] Recently, a study proposed that the OM should be viewed as ‘a wall consisting of OMP bricks connected by lipid mortar’.^[33] Clearly, the interface between LPS and OMPs is important and how it mediates OM stability and dynamics has only recently begun to be unraveled.

Breaking the mold: A fluid OM

The modern view of OM fluidity is largely shaped by studies from E. coli. However, Gram-negative bacteria live in diverse environments and display diverse behaviors, each of which elicit specific selective pressures that impact the evolution of the composition and properties of the OM. Therefore, a broader understanding of the OM requires studies on divergent species. In this regard, we recently investigated another model Gram-negative bacterium, Myxococcus xanthus. M. xanthus is a soil-dwelling, predatory, and rod-shaped bacterium that displays complex social behaviors.^[40] Perhaps the best-known social trait is its ability to aggregate individual cells into multicellular fruiting bodies under starvation conditions. In contrast to the short rigid rods of E. coli, M. xanthus cells are half as wide and approximately 4× longer and exhibit the flexibility of wet noodles. This flexibility indicates M. xanthus harbors a less rigid cell envelope and that their OM display distinct properties. Other lifestyle differences further hint that M. xanthus OM are different than E. coli. For instance, gliding motility of myxobacteria showcases how harboring a less rigid OM facilitates their movements.^{[41, 42]} Moreover, gliding motility coincides with the finding that polystyrene beads bound to cell surface adhesins move rapidly around the cell.^[43] The gliding motors, which are multi-protein complexes spanning the IM and periplasm, are powered by the proton motive force^{[43, 44]} that is transformed into mechanical force on the cell surface. Although aspects of mechanotransduction of force are unknown, it appears that the OM needs to be flexible and fluid to support adhesin mobility and gliding motility. Similarly, Gram-negative members of the phylum Bacteroidetes, such as Flavobacterium johnsoniae,^[45] exhibit gliding motility that is again powered by rapidly motile cell surface proteins, hinting that their OMs are also relatively fluid.

Outer membrane exchange in myxobacteria

Our recent work directly observed that myxobacteria harbor a surprisingly fluid OM.^[41] Here, different OM proteins are found to be mobile, and unlike the mentioned gliding motility proteins, their motions are independent of the proton motive force. This novel discovery arose from studying a unique social behavior called outer membrane exchange (OME).^[46] That is, when individual cells make physical contact by gliding motility their OM components are subsequently exchanged. This behavior was implied over 40 years ago when different M. xanthus motility mutants were mixed together and, in some cases, their motility defects were transiently rescued by extracellular complementation,^[47] which later was demonstrated to be caused by cell-cell transfer of motility proteins.^{[48, 49]} Our recent studies showed that various OM components, including OM lipids, LPS, lipoproteins and OMPs, are exchanged efficiently between cells during OME.^{[46, 49–52]} These observations suggested that the OM in myxobacteria, unlike E. coli, must be fluid to allow OME by lateral diffusion through transient membrane fusion junctions between cells.

OME is governed by two OM proteins: TraA and TraB.^[46] TraA is a cell surface receptor, and it is the specificity determinant that directs recognition between sibling cells by homotypic binding.^{[53, 54]} TraA is prevalent across the order of Myxococcales, which belong to the Deltaproteobacteria, and it displays a high level of polymorphisms within its variable domain.^[55] These polymorphisms reflect relatedness among isolates to help ensure cargo is exchanged between close relatives. Exchanging OM components offers the benefit of promoting cooperation between kin by sharing cellular resources.^{[51, 56]} However, if exchange partners are not clonal, antagonism will likely ensue because polymorphic toxins are also transferred.^{[52, 57, 58]} TraB is not involved in specificity but is required to form a functional cell surface adhesin with TraA.^[54] TraB contains a predicted β-barrel domain that is embedded in the OM and a C-terminal OmpA domain predicted to bind non-covalently to the cell wall.^{[41, 54]}

In contrast to E. coli,^[14] we found that both TraA and TraB display long-range diffusion across the cell surface.^[41] This was shown by fluorescently tagging TraA and TraB and monitoring their movements. In fluorescence recovery after photobleaching (FRAP) experiments, we found marked recovery of fluorescence of these proteins within a minute.^[41] Interestingly, the OM dynamics of TraA/B changes when individual cells transition into multicellular groups.^[41] That is, when two kin make contact, their TraA/B receptors coalesce into clusters at cell-cell contacts, directed by homotypic interactions between TraAs from neighboring cells. This homotypic binding then restricts the mobility of TraA/B within clusters. However, when gliding cells brush past one another, the clusters move along the cell junctions and dissipate when the cells disengage contact. OM mobility was also demonstrated by the addition of polyvalent antibodies that triggered TraA proteins to coalesce into aggregates in real time.^[41] In addition, fluorescent OM lipoprotein cargo is even more fluid than TraA/B and is rapidly transferred between cells, after the formation of TraA/B intercellular junctions.^[41] Similarly, fluorescent OM lipids are also rapidly transferred between cells.^{[46, 59]} Together, we showed that various OM components are fluid to different degrees. Our observations argue that the fluidity of myxobacteria OMPs and LPS differs markedly from E. coli. Finally, it should be noted that the fluidity of the E. coli inner leaflet phospholipids and lipoproteins is not well understood, but their fluidity is likely different than components in the outer leaflet.

The physiochemical makeup underlying OM fluidity in myxobacteria is less well understood. However, we suggest that differences in the chemical composition of the OM between myxobacteria and other organisms provide clues. For example, some myxobacteria lack LPS and contain alternative lipids in their OM, which are also present in M. xanthus, that are less bulky and more fluid.^{[60, 61]} This point is consistent with the fact that differences in the O-side chain and fatty-acyl chains change the fluidity of LPS.^[2] In addition, M. xanthus lacks the known quality control systems that removes aberrant phospholipids accumulated in the outer leaflet that presumably compromises OM rigidity.^{[62, 63]} In turn, we expect that the accumulation of phospholipids in the outer leaflet would also make the OM more fluid.

The occurrence of a relatively fluid OM is beginning to be uncovered in other Gram-negative bacteria, which we hypothesize is a common trait that depends on a species ecological niche and lifestyle. As mentioned, other gliding bacteria appear to possess a fluid OM. In another recent example from Caulobacter crescentus, a protein called OmpA2, which contains β-barrel and cell wall binding domains, was shown to be mobile in the OM by FRAP after its cell wall binding domain was removed.^[64] In contrast, the full-length OmpA2 exhibited restricted mobility. The OmpA2 protein has an analogous domain architecture as TraB, but in the case of TraB, we found by FRAP analysis that the full-length protein was mobile.^[41] To explain this mobility, we speculate that the anchoring of TraB to the cell wall is regulated through its interactions with TraA.

Does OM fluidity regulate social behaviors and multicellularity?

Changes and variations in OM fluidity are functionally meaningful to bacteria. For example, Gram-negative bacteria remodel their OM, and hence alter membrane fluidity, in responses to changing environments. Studies have shown that OM proteomes and LPS chemical composition are modulated in response to antimicrobials,^[65] temperature changes^[66] or transitions in lifestyle (e.g. planktonic to biofilm).^[67] Such remodeling alters OM fluidity and presumably facilitates bacteria to adapt to environments, however, the underlying mechanisms are often not well understood. As OM fluidity also varies across species, we suggest that bacteria with a fluid OM display distinct functions that are relevant to their ecological needs. Below we highlight how a fluid OM plays a role in shaping bacterial social behaviors.

In myxobacteria, a fluid and less rigid OM contributes toward gliding motility, which in turn lays the foundation for social behaviors that include fruiting body development and predation (Fig. 2A). We hypothesize that a rigid OM may hinder the gliding apparatus from moving properly. A rigid OM may inhibit the ability of cells to contort into different multicellular arrangements and their ability to produce OM-enclosed tubes.^{[41, 68]} Finally, the social behavior of OME requires a fluid OM so the TraA/B adhesins can coalesce upon cell contact and for cell cargo exchange (Fig. 2A). This sharing of cellular components is thought to facilitate the emergence of homogeneous populations and to maintain homeostasis as a cooperative multicellular ‘tissue.’^{[41, 56]} Heterogeneity in the OM can be caused by cellular aging (e.g. OMP oxidation/unfolding) and adaptions to different microenvironments. In response to these physiological incongruences, OME can dissipate cell to cell variations and enabling neighboring cells to better communicate and cooperate.^[41] The fluidity of the OM also allows myxobacteria cells to ‘repair’ envelopes that contain damaged LPS or proteins by exchange with sibling cells.^{[46, 51, 56]} That is, the damaged materials is diluted throughout the populations, which decreases acute toxicity, while healthy siblings can also replenish missing components to damaged cells. When fluid TraA/B adhesins coalesce into clusters upon cell-cell contacts, they resemble gap junctions in eukaryotic tissues (Fig. 2B), that similarly facilitate component exchange and communication between cells.^[69] Another example of gap junction-like structures in bacteria comes from filamentous cyanobacteria,^[70] where these junctions mediate the exchange of cytoplasmic molecules between adjacent cells, while OME junctions involves cell envelop content exchange (Fig. 2). Finally, we showed that cellular cargo delivered by OME can be serially transferred across a population,^[58] suggesting siblings assemble into a cohesive community where their OM is fluid and shared dynamically. For these reasons, a fluid OM plays important roles in mediating social behaviors and multicellularity in myxobacteria.

Figure 2. A fluid OM facilitates multicellular behaviors.

a) Myxobacteria glide and interact with neighboring cells. The even distribution of TraA/B adhesins (black dots) in individual cells transition into clusters upon cell-cell contact. Within the tissue-like communities, myxobacteria exchange OM contents bidirectionally through cell-cell junctions (right box). b) Representative gap junctions found in muscle tissue. c) The septal junction (analogous to ‘gap junction’) found in multicellular cyanobacteria.^[70]

The risks and benefits of a fluid OM

Variations in OM fluidity across different bacterial species likely reflect how they adapt to specific ecological niches. That is, physiochemical properties of the OM that are advantageous under one condition may be deleterious under different conditions.^[12] For instance, a fluid OM can render cells susceptible to external insults such as detergents or antibiotics. E. coli and other enterics residing in the mammalian intestinal tract, use their rigid and well-structured LPS to form a formidable barrier that excludes toxic factors, such as high concentration of bile detergents, from entering the cells by slowing their passive diffusion.^[2] Their robust OM permeability barrier reflects how E. coli and other enterics evolved to live in a common habitat. Similarly, mycobacteria harbor a unique OM (mycomembrane) that serves as an impermeable barrier against many antibiotics, and correspondingly their glycolipids lack fluidity.^[71] Interestingly, other bacteria with mycomembranes are instead fluid, again highlighting that membrane fluidity is species specific.^[71]

In the case of myxobacteria, they inhabit the soil and have a porous OM as they are exquisitely sensitive to various detergents.^[72] On the other hand, harboring a fluid OM offers them benefits. We suggest a fluid OM provides fitness advantages for their multicellular lifestyle. Although the fluidity of OM components in myxobacteria are noticeably higher than E. coli, we note they are nevertheless not as fluid as the inner membrane,^{[73, 74]} suggesting that our observations are broadly in line with the findings on that OM contributes to the structural integrity of cells and is needed to maintain a minimal level of rigidity.

Conclusions and outlook

Historically the OM was viewed as a defensive barrier against the cytotoxic effects of external chemicals. Over the past decade, studies on OM dynamics have revealed that the OM also functions as a rigid mechanical barrier. Built on the prevailing view from E. coli that the OM is rigid and harbors components with restricted mobility, we extended the scope to other species, particularly myxobacteria, whose OM is relatively fluid. Our recent work^[41] also provides insights into how a fluid OM facilitates the emergence of complex multicellular traits. We put forth the idea that other Gram-negative species also possess a relatively fluid OM, which is distinct from E. coli and its relatives. We suggest that variations of OM fluidity in different Gram-negative bacteria are closely associated with their adaptations to particular ecological niches and lifestyles, which remains a largely underexplored and exciting area of future research. By understanding OM fluidity, among other properties, we will gain valuable insights into OM functions and how different bacteria interact with their external world.

Abbreviations

OM

outer membrane

IM

inner membrane

LPS

lipopolysaccharide

OMP

outer membrane protein

BAM

β-barrel assembly machinery

OME

outer membrane exchange

FRAP

fluorescence recovery after photobleaching