Bacterial lipid biophysics and membrane organization

Abstract

The formation of lateral microdomains is emerging as a central organizing principle in bacterial membranes. These microdomains are targets of antibiotic development and have the potential to enhance natural product synthesis, but the rules governing their assembly are unclear. Previous studies have suggested that microdomain formation is promoted by lipid phase separation, particularly by cardiolipin and isoprenoid lipids, and there is strong evidence that cardiolipin biosynthesis is required for recruitment of membrane proteins to cell poles and division sites. New work demonstrates that additional bacterial lipids may mediate membrane protein localization and function, opening the field for mechanistic evaluation of lipid-driven membrane organization in vivo.

Introduction

The spatial organization of molecules is the essence of life. Lipids are central to this process, forming the membranes that define the cell. Further compartmentalization of cell membranes is the basis of cell polarity, and in bacterial models, involves the formation of so-called lateral membrane “microdomains”. These membrane microdomains are nanometer-scale patches of the bilayer with distinct molecular compositions^1–3, and the paradigms that govern their in vivo formation have yet to be uncovered. In this review, we take a lipid-centric approach to understanding membrane microdomain biogenesis in bacteria. We briefly outline the diversity of bacterial lipids, highlighting lipids shown to undergo phase separation in vitro, and summarize the evidence for in vivo bacterial lipid compartmentalization. We then critically examine recent studies on the effects of lipids on the function and localization of bacterial membrane proteins, as well as presenting key challenges for advancing the bacterial lipid field.

Bacterial lipid diversity

Bacterial lipids are remarkably diverse compared to those of eukaryotes. In part, this diversity is a consequence of the stark nutrient limitations that bacteria encounter in nature. Caulobacter, Sinorhizobum and Pelagibacter spp. can survive phosphate starvation by shutting down phospholipids synthesis^4–6, and Staphylococcus spp. similarly can synthesize nitrogen-free lipidome under nitrogen depletion⁷. Lipid transfer via vesicles or direct membrane exchange⁸ also is common in bacteria, suggesting that bacterial lipid diversity may also have evolved to facilitate inter-species signaling. This has been shown for the pro-inflammatory lipids of human pathogens^9–13, though little work has been done in other community contexts.

Most knowledge of bacterial lipidomes comes from growth of single strains under nutrient-replete conditions. In these settings, bacterial membranes are predominantly based on phospholipids, the most common being phosphatidylethanolamine (PE) and phosphatidylglycerol (PG)^14,15. These two phospholipids usually account for 50–90% of total phospholipid content^5,16–20 along with their lyso-, acyl-, and amino acid-conjugated derivatives. The major phospholipids in eukaryotes, phosphatidylcholine (PC) and phosphatidylserine (PS), are not abundant in bacteria²¹ except for a few species such as Sinorhizobium meliloti⁵, Rhodopseudomonas palustris²², and Methylobacterium extorquens¹⁹. Most bacteria also produce cardiolipin (CL), which typically comprises 5–10%^5,16–20,22 of total bacterial lipid extract.

Beyond these core phospholipids, only the lipid A domain of Gram negative’s lipopolysaccharide and the polyprenol lipid carriers of bacterial polysaccharides are widespread. Though sphingolipids were once believed to be rare in bacteria, new evidence suggests they also are present in a broad swath phylogenetically diverse organisms²³. Most other “secondary” lipids are restricted by phyla. Phosphorus-free glycerolipids based on a diacylglycerol backbone are perhaps the largest category, but they are uncommon outside of the alpha-proteobacteria and photosynthetic clades¹⁵. Ornithine lipids are enriched in proteo- and actino-bacteria¹⁵ but have a scattered distribution in other taxa. “Minor” secondary lipids are synthesized in only a few genera or are present only in certain environments. The genus-specific minor lipids, such as the mycolic acids of mycobacteria, have been best studied in pathogenic contexts where they are used as in situ biomarkers for infection^24–26. Environment-specific lipids include isoprenoids such as sterols, hopanoids, carotenoids, and sporulenes. Though sterol biosynthesis in bacteria is rare²⁷, the more common hopanoids and carotenoids are enriched under plant-associated nitrogen fixation^28,29 and phototrophic growth³⁰, respectively.

Bacterial lipid biophysics and organization in model membranes

Bacteria actively maintain the biophysical properties of their membranes through a process known as homeoviscous adaptation. This process is commonly studied under cellular stress: responses to elevated temperatures³¹, osmotic shock^19,32, and pore-forming antibiotics^33,34 all include significant lipidome remodeling to fortify the bilayer. Many membrane-fortifying lipids exhibit spatial heterogeneity¹, suggesting that they may promote organization of membrane processes. In pure bilayers, lipid spatial organization is governed by the free energies of lipid-lipid association, functions of a lipid’s charge, hydrophobicity, and surface area. Compressive forces or changes in bilayer geometry alter which lipid-lipid associations are favorable, and lipid phase separation can occur depending upon the diversity and ratios of lipids present.

While the phenomenon of phase separation only recently became popular in studies of proteins and nucleic acids, it was established in the lipid field over half a century ago^35,36. Most work on lipid phase separation concerns the predominantly eukaryotic lipids cholesterol and sphingomyelin. These lipids can segregate phospholipid bilayers into spatially distinct liquid-disordered (L_d) and liquid-ordered (L_o) phases in vitro; lipid gel phases (L_β) can also be observed, albeit outside of physiologically-relevant conditions. The L_o phase is characterized by its altered biophysical properties, including increased resistance to external forces (elasticity^37,38), tighter lipid packing (condensation³⁹), and lower lateral diffusion (fluidity⁴⁰). Methods used to detect cholesterol- and sphingomyelin-induced L_o phases include direct physical measurements by atomic force microscopy⁴¹ (AFM) or pressure-area isotherms, and fluorescence-based studies using lipophilic reporters (summarized in Table 1).

Table 1.

Commonly used dyes for reporting on membrane biophysics. Note that excitation and emission maxima can be strongly affected by the surrounding solvent.

Reporter	λ ex, max	λ em, max	Mechanism	Ref.
C-laurdan / Laurdan	405 / 350	440 (Ld) 480 (Lo)	Increased lipid packing shifts laurdan fluorescence emission toward longer wavelengths. This is due to sensitivity of the fluorophore to water; water is more likely to be excluded from regions with tight lipid packing.	155–157
DiIC12 / Fast DiO	549 / 484	565 / 501	Short chain DiI and DiO probes are found only in high fluidity membrane regions. The short acyl chains cannot extend into the thick hydrophobic core of Lo domains and are thermodynamically excluded from these regions.	158
DPH	355	430	Increased lipid packing reduces the rotational mobility of DPH in bilayers. DPH mobility can be quantified by fluorescence anisotropy.	159–161
di-4-ANEPPDHQ	465	570 (Lo) 630 (Ld)	Increased lipid packing shifts di-4-ANEPPDHQ fluorescence emission toward shorter wavelengths and increases fluorescence lifetime. This dye is also affected by changes in membrane potential.	151,157, 162–164
PDA	350	395 (Ld) 470 (Lo)	PDA forms “excimers,” molecular complexes that only exist after fluorophore excitation. PDA excimers emit at a longer wavelength than PDA monomers and are more likely to form in low-fluidity membranes.	165,166

C-laurdan = N-methyl-N-[6-(1-oxododecyl)-2-naphthalenyl]glycine; Laurdan = 2-dimethylamino-6-lauroylnaphtalene; DiIC12 = 1,1-Didodecyl-3,3,3,3-tetramethylindocarbocyanine; DiO = 3,3′-Dioctadecyloxacarbocyanine; DPH = trimethylammonium-diphenylhexatriene; di-4-ANEPPDHQ = [3-[4-[(E)-2-[6-(dibutylamino)naphthalen-2-yl]ethenyl]pyridin-1-ium-1-yl]-2-hydroxypropyl]-(2-hydroxyethyl)-dimethylazanium;dibromide; PDA = pyrenedecanoic acid.

Much less is known about the biophysical properties of bacterial lipids. Most studies focus on lipids also found in eukaryotes, such as cardiolipin and carotenoids. In phosphatidylcholine (PC) membranes, cardiolipin increases membrane fluidity and generates micron-scale lipid domains⁴² that are enriched at areas of high membrane curvature⁴³. Studies in phosphatidylglycerol (PG) and phosphatidylethanolamine (PE) bilayers yield similar results, demonstrating that cardiolipin raises membrane fluidity, thickness, and lipid packing^44–47 in bacterial membrane-mimicking phospholipid compositions. These seemingly contradictory biophysical effects – L_d-like high fluidity alongside L_o-like tight lipid packing – likely relate to cardiolipin’s preference for curved membranes, as bilayer bending alters lipid headgroup proximity. The magnitude of these effects is much greater with PE than in PC alone⁴⁷, demonstrating that lipids’ in vivo biophysical effects depend on the organism’s phospholipid repertoire.

The in vitro biophysics of carotenoids have been examined only in PC. Carotenoids condense and rigidify PC bilayers, even at low concentrations^48–53, but whether they phase separate into L_o microdomains in vitro is un known. Hopanoids lower fluidity and increase packing in PC-based membranes, forming micron-scale L_o microdomains at physiological conditions^54–57. Minor structural modifications attenuate these biophysical effects^58,59, and similar results occur in reconstituted bacterial lipid extracts^59–61. Shockingly few other bacterial lipids have been examined biophysically, likely because most are not commercially available. The single study of ornithine lipid biophysics found that they form more ordered bilayers compared to PC and PG, but this required in-house ornithine lipid synthesis⁶². Progress in total organic synthesis of bacterial lipids and expanded commercial availability would enhance this area rapidly.

In vivo biophysics and organization of bacterial lipids

Whether lipid phase separation occurs in vivo remains a perennial subject of debate. Many groups have tried to address this topic by increasing the complexity of in vitro and in silico membrane reconstitution systems to capture additional physiological factors. From these studies, factors that affect L_o microdomain formation include linkages between lipids and the cytoskeleton, cell wall, or extracellular matrix^63–65; bilayer leaflet asymmetry⁶⁶; heterogeneous membrane topology^67–70; and lipid acyl chain saturation⁷¹. How these factors interact to inhibit or promote phase separation is difficult to predict. Regardless, there is a consensus that animal membranes do contain discrete membrane regions with distinct lipid compositions, typically on the order of tens of nanometers rather than the micron-sized L_o domains observed in vitro⁷².

It is similarly debated whether spatial heterogeneity in bacterial membranes is driven by lipid phase separation. However, bacterial lipids that induce L_o phase separation in vitro do affect whole cell membrane biophysics in a manner consistent with L_o phase formation. Loss of carotenoids in Staphylococcus, Synechocystis, and Pantoea spp. increases whole cell membrane fluidity and cell elasticity^73–77, and hopanoid biosynthesis mutants of Bradyrhizobium, Rhodopseudomonas, and Methylobacterium spp. exhibit reduced in vivo lipid packing^59,78,79. C arotenoid localization in the cell has not been examined beyond in silico studies⁸⁰, but the localization of exogenously applied hopanoids has been evaluated by nanoSIMs (Nanoscale Secondary Ion Mass Spectrometry). In this technique, isotope-labelled lipids are fed to bacterial cultures, which are then fixed and eroded with a high energy ion beam. The erosion of the cells releases “secondary” ions that are collected for isotopic composition analysis (i.e. ²H/¹H ratios) by mass spectrometry. NanoSIMS analyses of Nostoc punctiforme treated with isotope-labeled hopanoids suggest that hopanoids are enriched in microdomains at the junctions between filamentous vegetative cells⁸¹. Finally, an exciting recent study in E. coli identified membrane microdomains of 25–250 nm diameter via high resolution AFM that are thicker than the surrounding membrane and diffuse and coalesce similarly to liquid phase separation⁸². These regions exclude dominant outer membrane proteins and are regulated by LPS levels, raising the possibility that they constitute lipid A-rich, L_o-like microdomains.

There is strong evidence that cardiolipin occupies spatially distinct membrane microdomains in vivo. Cardiolipin can be detected by the lipophilic dye 10-N-nonyl acridine orange (NAO)⁸³, which binds several anionic phospholipids^84,85. Staining cells with NAO suggests that anionic phospholipids are present at cell poles and septa of E. coli^86,87, Streptococcus pneumoniae⁸⁸, Mycobacterium tuberculosis⁸⁹, and Chlamydia trachomatis⁹⁰ and at the septum and forespore membrane of Bacillus subtilis⁹¹. The poles and septa of E. coli and B. subtilis are areas of relatively high membrane fluidity^88,92, and deletion of cardiolipin synthase decreases membrane fluidity in Pseudomonas putida⁹³. There are opposing views on whether targeting of cardiolipin to specific membrane regions results from phase separation. In the one model, cardiolipin is synthesized uniformly across the membrane but phase separates at poles and septa, where higher membrane curvature favors the conical cardiolipin structure^43,86,94,95. Another model suggests a prior localization of cardiolipin biosynthesis enzymes that constrains cardiolipin insertion into the bilayer^96–98. Others propose that both mechanisms operate in vivo⁹⁹, with varying degrees of coordination, an interpretation we also favor.

Some of the first lipid microdomains in bacteria were identified in mycobacteria, based on the spatial heterogeneity of fluorescent lipophilic dyes¹⁰⁰. In Mycobacterium smegmatis, a lipid microdomain at the elongating cell pole, known as the inner membrane domain (IMD), is enriched in phosphatidylinositol mannosides, phosphatidic acid, lipid II, and related lipid biosynthesis enzymes^101–103. Although the biophysical properties of the IMD have not been evaluated directly, new work has shown that its abundance is sensitive to treatment with chemicals that raise membrane fluidity¹⁰². IMD formation is dependent on peptidoglycan synthesis¹⁰⁴ and the polar growth and division regulator DivIVA¹⁰², suggesting that this domain is initiated by prior localization of polarized growth proteins rather than lipid phase separation.

Mycobacterium spp. also may have distinct microdomain compositions in situ, due to their ability to incorporate host cell cholesterol. Though primarily has examined from the perspective of bacterial nutrition¹⁰⁵, cholesterol uptake is likely to affect membrane biophysics and occurs in many human pathogens including Mycoplasma¹⁰⁶ and Helicobacter pylori¹⁰⁷. Immunogold labeling of cholesterol in Borrelia burgdorferi cells isolated from mice suggested that cholesterol is present in ~50 nm outer membrane domains¹⁰⁸ that are enriched in outer surface lipoproteins¹⁰⁹. Many obligate intracellular pathogens manipulate host cholesterol pathways¹¹⁰, but it is unknown whether this manipulation is directly related to lateral membrane organization.

Effects of bacterial lipids on protein behavior

A recent study of 200 fluorescent protein fusions in B. subtilis estimated that 85% of membrane proteins have spatially heterogeneous membrane localizations¹¹¹. How many of these proteins are targeted to lipid microdomains is impossible to estimate, as the relationship between membrane protein localization and lipid microdomain formation does not follow a single paradigm. In some cases, bacterial membrane protein localization depends on direct interactions with specific lipid species. For example, cardiolipin determines the localization and directly regulates the E. coli osmotic pressure-sensitive proline importer ProP¹¹² and cell envelope stress proteins DolP (YraP)¹¹³ and PspA¹¹⁴, as well as the Sec translocon ATPases of both E. coli¹¹⁵ and Streptococcus pneumoniae¹¹⁶. There also is broad agreement that lipids affect the localization of bacterial flotillins. These proteins are homologs of the eukaryotic flotillin/reggie family of proteins, which form oligomeric scaffolds and are enriched in cholesterol and sphingomyelin microdomains. In bacteria, flotillins localize to in vivo microdomains in 80–90 nm sized clusters¹¹⁷ where they recruit diverse membrane proteins, including regulators of protein secretion^118,119 and biofilm formation^120–122. Early evidence in B. subtilis suggested that bacterial flotillins co-localize with cardiolipin by NAO staining¹²³, but more recent work contradicts this finding¹¹⁷. Knockout of the two flotillin homologs of B. subtilis raises membrane fluidity and increases lipid packing (though the lipid packing sensor used in the study, C-laurdan, was misinterpreted as a fluidity reporter)¹²⁴; as these phenotypes are the opposite of those resulting from cardiolipin loss, suggesting that flotillin is unlikely to stabilize cardiolipin domains. Other studies suggest that flotillin localization is dependent on isoprenoid lipids^121,125 such as the carotenoid staphyloxanthin, which appears to directly bind the flotillin FloA of Staphylococcus aureus¹²⁶. The dependence of bacterial flotillins on isoprenoid lipids for localization is consistent with the dependence of eukaryotic flotillins on cholesterol, which is also in the isoprenoid class. The flotillin homolog of Helicobacter pylori appears to increase cholesterol uptake from host tissues during mammalian infection¹²⁷, though direct cholesterol binding has not been reported.

In other cases, protein localization is proposed to depend on membrane properties that can be controlled by multiple lipid classes – or to be largely lipid-independent, as shown by a thorough recent analysis of the B. subtilis membrane fission protein FisB¹²⁸. Carotenoids and environmental fatty acids have been proposed to indirectly regulate the activity of the S. aureus type VII secretion system via their effects on membrane fluidity¹²⁹, whereas unsaturated fatty acids are proposed to regulate the E. coli F₁F₀ ATPase through a similar mechanism¹³⁰. In Chlamydia trachomatis, cardiolipin is hypothesized to stabilize high inner membrane curvature, which in turn stabilizes binding of the curvature-sensitive cytoskeletal protein MreB⁹⁰. These proposed mechanisms of lipid-mediated protein localization are difficult to evaluate, however; fatty acids in particular are central to bacterial metabolism, and membrane proteins themselves can modify membrane fluidity and curvature, independently of specific lipids¹³¹.

A property of L_o-like membranes that can be exploited to identify potential associated proteins is their resistance to detergent-based solubilization. Separation of animal cell membranes into fractions of detergent-sensitive membranes (DSM) and detergent-resistant membranes (DRM) led to the discovery that GPI-anchored proteins are targeted to cholesterol- and sphingomyelin-rich microdomains¹³², ultimately creating lipid raft theory. Many groups have now performed studies of DSM and DRM cmponents in model bacterial strains^{123,125–127,133–139}. We summarize the functional categories of proteins that are most consistently found in DRMs in Table 2. A caveat with this approach is that DRM fractions will include any detergent insoluble protein assemblies, regardless of associated with microdomainforming lipids; this may explain why membrane proteins that bind unfolded protein substrates, and are necessarily resistant to highly hydrophobic environments, are so commonly identified. Additional verification of microdomain association is necessary for any proteins identified by this method, as has now been done for bacterial flotillins and the Sec and F₁F₀ ATPases discussed earlier in this section.

Table 2.

Selected bacterial proteins commonly enriched in detergent-resistant membrane (DRM) fractions. Categories were identified by collating the DRM proteomes reported for Agrobacterium tumefaciens (Ref. 134), Bacillus anthracis (Ref. 125) Bacillus subtilis (Refs. 123,137–139), Borrelia burgdorferi (Ref. 136), Escherichia coli (Ref. 139), Helicobacter pylori (Ref. 127), and Staphylococcus aureus (Ref. 126). Example proteins represent category members that were found in DRMs of at least two organisms.

Protein function	Organisms
Amino acid transport (i.e. MetN, GlnQ, GltT/P, LysP, YdjN)	E. coli, S. aureus, B. subtilis, B. burgdorferi
ATP synthesis (i.e. AtpAB, AtpFG)	E. coli, S. aureus, B. subtilis, H. pylori, B. burgdorferi
Cardiolipin synthesis (i.e. ClsABC)	E. coli, S. aureus, B. subtilis
Electron transport (i.e. QoxAB, CyoAB, NuoL/NuoG, HydA, SdhAB)	E. coli, S. aureus, B. subtilis, A. tumefaciens, H. pylori
Flotillins (i.e. FloT, FloA, FlotP)	E. coli, S. aureus, B. subtilis, H. pylori, B. anthracis
Membrane quality control (i.e. DegSP, FtsH, QmcA, HflCK, PrsA, HtrA)	E. coli, S. aureus, B. subtilis, B. burgdorferi, A. tumefaciens
Multidrug efflux (i.e. MsbA)	E. coli, S. aureus, A. tumefaciens
Oligopeptide transport (i.e. OppABC, OppF)	E. coli, S. aureus, B. burgdorferi
Peptidoglycan synthesis & cell division (i.e. MurJ, PBP5/DacA, PBP1A/MrcA, PBP4, CidB, RodA, FtsW, FtsA, FtsX)	E. coli, S. aureus, B. subtilis
Protein secretion (i.e. LolDCE, TatABC, Sec/YajC/YidC, PrsA)	E. coli, S. aureus, B. subtilis, B. burgdorferi, A. tumefaciens
Two-component signaling (i.e. RseE, WalK, PhoR)	E. coli, S. aureus, B. subtilis
Chemotaxis and motility (i.e. MotAB, CheV, MCPs, flagellins)	E. coli, B. subtilis, A. tumefaciens, H. pylori, B. burgdorferi

Interestingly, recent studies from two groups have suggested a connection between bacterial lipids and the formation of biomolecular protein condensates in proximity to the membrane. Vesicles made from E. coli total lipid extract promote the in vitro self-segregation of FtsZ¹⁴⁰ and its regulator SlmA^140,141, and condensates of the nucleoid occlusion protein Noc is enhanced on PC:PG unilamellar vesicles and is positively correlated with anionic lipid abundance. Further, doping vesicles with cholesterol to induce phase separation shows that Noc condensates form exclusively on L_d microdomains, suggesting bidirectional regulation of protein phase separation and lipids at the membrane¹⁴². This work provides a bridge between the membraneless organelle and lipid phase separation fields that will be an exciting avenue for exploration.

Outlook and challenges

Bacterial membrane microdomains have received new attention in biomedicine based on studies of membrane-targeting antibiotics. Common amphiphilic antibiotics such as daptomycin have been suggested to function by targeting cardiolipin-rich^143–145 and high fluidity microdomains¹⁴⁶. Proof-of-concept studies also show that targeting metabolic pathway enzymes to microdomains can enhance natural product synthesis^147,148. Despite these promising applications, we are in an early stage of our understanding of how lipids mediate membrane organization in vivo. Though abundant evidence shows that disrupting biosynthesis of certain lipid classes can affect whole cell biophysics and the location and activity of some proteins, in most cases we do not know whether these are direct or indirect consequences, nor whether they reflect roles of lipids themselves or of their biosynthetic enzymes. Complicating these interpretations is evidence that disrupting the synthesis of a single lipid can drive significant lipidome remodeling²², altering levels of dozens of other lipid species that may contribute to cellular phenotypes. Though these concerns can be addressed by lipid complementation, these assays are technically challenging; exogenous lipids may not be commercially available or readily taken up by cells, and they can only be quantified with advanced chromatography instrumentation.

A second major knowledge gap is the extent to which bacterial lipids act via modified membrane biophysics versus specific binding interactions. There are remarkably few studies on the lipid preferences for bacterial membrane binding domains¹²⁸, outside of bacterial lipid transporters. This again can be attributed partly to the dearth of commercially available bacterial lipids, as well as the general difficulty of studying membrane proteins in vitro. Progress in understanding the lipid specificities of bacterial membrane binding domains, particularly in the area of protein structural biology, could enable disruption of specific protein-lipid interactions by protein mutagenesis. Such studies could help tease apart how bacteria balance the demand for specific protein-lipid interactions versus general biophysical homeostasis.

Finally, a unique problem in Gram- negatives is that we do not know whether whole cell biophysical measurements or fluorescence microscopy studies of lipid localization reflect the inner versus outer membrane. Only a handful of bacterial membrane dyes have predicted inner and outer membrane preferences (summarized in Table 3), based on their localization in plasmolyzed cells. Unfortunately, plasmolysis and spheroplast generation can perturb lipophilic dye localization; in B. subtilis, cell wall disruption re-localizes the lipophilic dye FM 4–64 within organism’s sole, cytoplasmic membrane¹⁴⁹. These concerns in Gram- negatives reflect broader issues with relying on lipophilic dyes and fluorescent lipid analogs for biophysical measurements. Coupling fluorophores to lipids dramatically affects their size and physicochemical properties, and these labelled lipids can behave differently than their label-free counterparts¹⁵⁰. Lipophilic reporter dyes can yield inconsistent measurements¹⁵¹, and it is difficult to determine how much the dyes themselves perturb the membrane.

Table 3.

Localization preference of common membrane dyes to the inner membrane (IM) or outer membrane (OM) of E. coli cells after plasmolysis.

Membrane dye	Localization	Ref.
FM 4–64	OM	167–170
FM 1–63	OM	169
Kdo-Alexa488	OM	171
BODIPY FL-C12 / C10	IM	172
DPH	IM	167,173
FITC-poly-L-lysine	IM	174,175
Mitotracker Green	IM	168
NAO	IM	176
Nile Red	IM	167
Pyrene	IM	167

Mass spectrometry-based imaging such as nanoSIMs, direct in vivo biophysical measurements via AFM, and cryo-electron microscopy are promising solutions to these issues^{81,82,152–154}. A chieving the spatial resolution necessary for microdomain analysis requires operating at the limits of these technologies, and advances in these areas have the potential to revolutionize bacterial lipid cell biology. With imaging becoming easier, labelling becoming less biologically and biophysically disruptive, and advancements in lipid and lipid binding partner purification, the bacterial lipid field is poised for an explosion of discovery.