Residual Membrane Fluidity in Mycobacterial Cell Envelope Layers under Extreme Conditions Underlines Membrane-Centric Adaptation

Abstract

One of the routes for adaptation to extreme environments is via remodeling of cell membrane structure, composition, and biophysical properties rendering a functional membrane. Collective studies suggest some form of membrane feedback in mycobacterial species that harbor complex lipids within the outer and inner cell wall layers. Here, we study the homeostatic membrane landscape of mycobacteria in response to high hydrostatic pressure and temperature triggers using high pressure fluorescence, mass and infrared spectroscopies, NMR, SAXS, and molecular dynamics simulations. Our findings reveal that mycobacterial membrane possesses unique and lipid-specific pressure-induced signatures that attenuate progression to highly ordered phases. Both inner and outer membrane layers exhibit phase coexistence of nearly identical lipid phases keeping residual fluidity over a wide range of temperature and pressure, but with different sensitivities. Lipidomic analysis of bacteria grown under pressure revealed lipidome remodeling in terms of chain length, unsaturation, and specific long-chained characteristic mycobacterial lipids, rendering a fluid bacterial membrane. These findings could help understand how bacteria may adapt to a broad spectrum of harsh environments by modulating their lipidome to select lipids that enable the maintenance of a fluid functional cell envelope.

Introduction

Membranes maintain cell shape, compartmentalize lipids and proteins, thus regulating numerous processes both in eukaryotes and prokaryotes.¹⁻³ By leveraging biophysical properties such as fluidity, packing, stiffness, curvature, and lateral organization,³⁻⁵ membranes alter diffusion, localization, and interactions of lipids and proteins, and subsequently their activity.^6,7 The structural function of lipid membranes is also underpinned by their ability to form various lipid phases.^8,9 Liquid crystalline—the most common and physiological lipid phase forms the cell membrane matrix—and the inverted hexagonal (HI), bicontinuous cubic, and ordered micellar phases are transient and involved in processes such as fusion, fission, etc.¹⁰⁻¹³ Interconversion of lipid phases depends on the preferred curvature of lipids, and the interplay between membrane curvature, elastic stress, headgroup charges, and chain-packing frustration.¹⁴ Careful maintenance of membranes in a functional state is thus essential because subtle changes in membrane structure or biophysical characteristics impact a wide range of cellular functions. Toward this view, diverse adaptive strategies are known to preserve optimal membrane functionality in harsh environments,^15,16 e.g., in extremophiles, on early Earth,¹⁷ deep sea,¹⁸ and within the host’s microenvironment.

Harsh environments to study membrane behavior can also be recapitulated using physicochemical perturbations such as temperature (T), pH, pressure (p), etc. Specifically, temperature and hydrostatic pressure have been extensively employed to probe the structure and dynamics of various membrane phases;^18,19 interestingly, membranes are one of the most pressure sensitive cellular components. Hydrostatic pressure drives structural changes in membrane, while offering significant advantages,^17,20,21 over perturbation by temperature. High pressure does not tend to disrupt intramolecular bonding <2 GPa; is reversible, can be applied and released extremely rapidly; and, due to its rapid propagation, equilibrates quickly. Pressure-induced homeostatic membrane response to maintain cellular functioning is a subject of great interest.

Homeostatic membrane response in Mycobacterial tuberculosis (Mtb), the causative agent of Tuberculosis, is relatively unknown, despite modifications in its cell wall architecture and composition under various stress conditions.^22,23 Mtb also restructures its membrane in response to high osmotic pressure via osmosensory signaling pathways within the host.^24,25 Further, some mycobacterial species (M. pinnipedii) infect fishes or sea lions living in high pressure regions within deep sea.²⁶ Notably, exposure to high hydrostatic pressure during medical sterilization does not affect viability of M. abscesses.²⁷ These all point on the ability of mycobacteria to adapt to extreme conditions. Mycobacterial lipidome is highly complex encompassing long-chained, branched, and bulky sugar-decorated lipids.²⁸⁻³⁰ This structural complexity could underline potential lipidic nodes balancing the demands of membrane physiology under extreme conditions. In fact, a recent study reported mechanical adaptation of mycobacterial cell surface softness when exposed to host microenvironment.³¹ Alternatively, investigating the mycobacterial membrane response to extreme conditions could furnish insights for developing membrane destroying strategies to target mycobacteria. Together, these factors prompted us to determine how mycobacteria may mount a membrane-centric response to extreme physiochemical conditions underlying adaptation for survival and pathogenesis.

In this work, we used experimental and computational methods to investigate protein-free model membranes composed of inner and outer cell wall lipids from Mycobacterium smegmatis, Msm (a lab model for Mtb³²). We discovered that these compositionally distinct membranes exhibited lipid phase coexistence regions having residual fluidity even under medium to high pressure (∼2 kbar) and low temperature (∼10 °C). Despite having structurally distinct lipids, the outer and inner model membrane layers exhibited similar p,T stabilities and remained in a fluid-like, liquid-crystalline, and functional state. However, the outer layer was slightly more ordered and the fluidity settled in at approximately 10 °C higher compared to the inner layer. This suggests lipid-mediated mechanisms of maintaining membrane phase homeostasis in mycobacteria and possibly other organisms with complex cell wall lipids. This work is the first account of mapping the p−T phase diagram of natural lipid membrane extracts from mycobacteria.

Material and Methods

Bacterial Growth

Mycobacterium smegmatis (Msm) mc²155 was grown in Middlebrook 7H9 supplemented with 10% albumin-dextrose-catalase (ADC) and 0.2% glycerol under shaking conditions at 37 °C. The cells were harvested and centrifuged at 5000 rpm for 15 min once the bacterial OD₆₀₀ reached 3.0 and washed with distilled water to remove traces of media and other growth supplements. To look at the lipid profile of bacteria under pressure, bacteria were grown to OD₆₀₀ 0.8 and then placed in a high-pressure cell at 37 °C at 1030 bar for 12 h. After 12 h, the bacteria were centrifuged, and total lipids were extracted for LC/MS.

Lipid Extraction and Identification

After being harvested, outer membrane and inner membrane lipids from bacteria were extracted using a modified version of a reported method.^28,32 The bacterial cell pellet for total lipids was extracted overnight using a solution of chloroform−methanol and water (CMW 2:1:0.1). For every 10 mg of dry cell mass (lyophilized), 3 mL of the CMW mixture was used. After the first extraction, the supernatant was saved, and three more re-extractions of 15 min each were performed. All extractions were carried out in monophasic solutions.

Briefly, for extracting all noncovalently bound OM lipids, for every 10 mg of dry cell mass, 1 mL of RMS (10 mM AOT in n-heptane) was used. The solution was left for overnight extraction followed by three re-extractions for 15 min each. Column chromatography was performed using an alumina gel column with a 100−200 mesh particle size to remove the AOT from the extracted OM lipids. A 1−7% methanol gradient in chloroform was used as the mobile phase. Lipids were visualized on thin layer chromatography (TLC) using a developing solution of 1% anthrone in conc. H₂SO₄ and methanol (1:1). After extracting OM lipids, the bacterial cells were washed multiple times to remove any traces of RMS and prepped to extract the IM lipids.

IM lipids were extracted overnight using a solution of chloroform, methanol, and water (CMW 2:1:0.1). For every 10 mg of dry cell mass, 3 mL of the CMW mixture was used. After the first extraction, the supernatant was saved, and three more re-extractions of 15 min each were performed. All extractions were carried out in monophasic solutions. Each of the extracted lipids were concentrated and dried using nitrogen gas and finally reconstituted in chloroform to obtain final lipid stocks.

An Agilent 1260 infinity II UHPLC system (Agilent Technologies, USA) with an XBridge C18 column (Waters Corp, USA) was used for LC with dimensions of 2.1 mm × 150 mm, 3.5 μm, heated to 45 °C and was used with a binary solvent system and a flow rate of 0.2 mL/min. The system was equilibrated with 100% solvent A [0.1% formic acid in methanol−water (99:1; v/v)], and an aliquot of the lipid extract (5 μL) was applied to the column. Solvent A was maintained at 100% for 1 min, followed by a 30.0 min linear gradient to 100% solvent B [0.1% formic acid in acetonitrile–water (99:1; v/v)], and held at 100% solvent B for 2 min, followed by a decrease in gradient to 5% solvent B for 1 min, and held at 5% solvent B for 9 min. All solvents and chemicals purchased were MS or HPLC grade.

An Agilent 6545 time-of-flight (TOF) mass spectrometer equipped with an Agilent ESI source was used for accurate mass analysis of the LC eluent. Positive (+) and negative (−) ion data were generated by operation of the mass spectrometer in a mixed ESI mode with a capillary voltage of 4000 V, nebulizer of 45 psi, drying gas of 8 L/min, gas temperature of 300 °C,, fragmentor of 175 V, charging voltage of 1000 V, skimmer of 65 V, and octopole radio frequency voltage of 750 V. Mass spectra were acquired in high-resolution mode at a rate of 1.00 spectra/s and data were collected as profiled spectra over a mass range of 250−3200 Da. Data were collected with the Agilent MassHunter WorkStation Data Acquisition software, version B.09.00.

LC/MS data files were processed with the MassHunter Workstation Qualitative Analysis Software, version 10.0 (Agilent Technologies, USA). The lipids were identified using Mycomass and LipidDB databases as previously made available online³³⁻³⁵ and using the Target Screening (find by formula) algorithm of the quantitative analysis software. The algorithm parameters, extraction algorithm, small molecule; peak filters, ≥500 counts; ion species, +H, +Na, +NH₄, −H, and +HCOO only; match tolerance of 10 ppm; isotope model, common organic molecules; charge state, 1−2; compound filters, none; mass filters, none; and mass defect, none the identified and matched lipids were further analyzed using custom Matlab code. The LC peak areas were normalized against that of the internal standard reserpine. The lipid species at the class, subclass, or molecular species levels were normalized to the total lipid concentration (i.e., mol % total lipid) or total lipid-class concentration (i.e., mol % total lipid class). Two independent replicates were performed for all measurements.

Large Unilamellar Vesicle (LUV) Preparation

Stock solutions of extracted outer membrane lipids and inner membrane lipids were made in chloroform. Based on the experiment performed, desired amount of the membrane lipids stocks and dyes were taken and chloroform was evaporated with a nitrogen stream. All remaining solvent was subsequently removed under vacuum overnight. All buffers were filtered through a 0.2 μm pore size syringe filter before use. Lipid dispersions of LUVs were prepared by hydrating the film in 20 mM Tris and 5 mM MgCl₂ buffer pH 7.4 followed by several freeze and thaw vortex cycles.

High Pressure Fluorescence Spectroscopy Measurements

All fluorescence spectroscopy measurements were performed on the Chronos multifrequency phase and modulation fluorometer coupled with a stainless-steel high-pressure vessel (ISS, Champaign, IL). Briefly, 1 mg/mL of LUVs labeled with 1 mol % of Laurdan were used. Laurdan emission spectra was collected for generalized polarization measurements The samples were excited using a laser diode (λ_ex = 370 nm) and emission was collected using two band-pass filters −436HC10-25 and 488HC10-25, to collect intensity data at 440 and 490 nm range.

$$\text{GP}=\frac{I_{\text{440nm}}-I_{\text{490nm}}}{I_{\text{440nm}}+I_{\text{490nm}}}$$ (1)

The above equation was used for GP calculation at various pressures.

Pressure was exerted using a manual pump from 1 to 2206.3 bar at specified pressures and ethanol was used as pressurizing fluid. The temperature was controlled by a circulating water bath directly connected to the sample holder.

Nuclear Magnetic Resonance Spectroscopy

Samples were prepared by dissolving in chloroform 1-palmitoyl-2-(²H₃₁) oleoyl-sn-glycero-3-phosphocholine (²H₃₁−POPC), bought from Avanti Polar Lipid, Alabaster, USA, with outer and inner membrane lipid extracts in a molar ratio of 10% ²H₃₁−POPC and 90% for the lipid extracts. The chloroform was evaporated under a stream of nitrogen. The residual lipid film was dispersed in 1 mL of Milli-Q filtered water and freeze-dried overnight to remove all traces of chloroform. The resulting powder (ca. 27 mg for the outer membrane sample and 40 mg for the inner membrane sample) was suspended into 100 μL of deuterium-depleted water to obtain a hydration of 71 to 80%. Samples were sonicated for 15 min at 55 °C. After shaking into a vortex mixer, samples were frozen in liquid nitrogen for 30 s, heated at 50 °C for 10 min in a water bath, and shaken again for better sample homogeneity; this freeze–thaw-shaking cycle was repeated 5 times and the resulting milky dispersion transferred into a 4 mm diameter Zirconium NMR rotor.

Solid-state ²H NMR were performed on a Bruker NEO 800 MHz SB spectrometer (Bruker Biospin, France) equipped with a triple 1H/2H/31P 4 mm MAS probe. ²H NMR spectra were acquired at 122.84 MHz by means of a quadrupolar echo pulse sequence³⁶ with a spectral width of 500 kHz, a π/2 pulse width of 4.7 μs, a 40 μs interpulse delay, 2k scans, a time domain of 2K points, and a recycle delay of 2 s. A Lorentzian noise filtering of 300 Hz was applied prior Fourier transformation from the top of the echo signal. Samples were allowed to equilibrate at least 20 min at a given temperature before the NMR signal was acquired.

High Pressure FTIR Spectroscopy

Pressure-dependent FTIR spectra was recorded using a Nicolet 6700 (Thermo Fisher Scientific) equipped with a liquid-nitrogen cooled MCT-detector. The sample chamber was continuously purged with CO₂-free and dry air. All measured spectra were averaged over 128 scans in a row at a spectral resolution of 2 cm⁻¹ and were processed with Happ-Genzel apodization by using Omnic 7.2 spectral processing software. The equilibration time before each spectrum was recorded for the pressure dependent studies, 5 min. The temperature for the pressure ramps were equilibrated for 15−20 min the start of the experiment. The setup of the pressure system consists of a membrane-driven diamond anvil cell (VivoDac Diacell) with type IIa diamonds, which is connected to an automated pneumatic pressure control (Diacell iGM Controller, Almax easyLab). To measure the pressure inside the cell, by adding BaSO4, the pressure-sensitive stretching vibration of SO₄ ²⁻ (∼983.5 cm⁻¹ ≙ 1 bar, 25 °C) was used as an internal pressure calibrator.³⁷ LUV solution (10 wt %) was made in 20 mM Tris and 5 mM MgCl₂ D₂O buffer. Depending on the sample, spectra of the buffer systems were subtracted from the spectra recorded and smoothed afterward. Processing and analysis of the spectra was carried out with the GRAMS AI 8.0 software (Thermo Fisher Scientific). The raw curves were first processed in the GRAMS software and the maxima for each curve was used for plotting the final data.

Differential Scanning Calorimetry (DSC)

The samples (inner and outer membrane lipids) were suspended in 20 mM Tris-HCl (pH 7.4), 5 mM MgCl₂. The samples were used at a concentration of 10 mg/mL. The scanning rate was 1 K min⁻¹; the measurements were performed in the 278−340 K temperature range in 3 cycles of heating and cooling in triplicates. The raw data was analyzed using the pyDSC program.³⁸

Small Angle X-ray Scattering (SAXS)

The SAXS experiments were performed at the P12 high brilliance beamline at European Molecular Biology Laboratory (EMBL) in Hamburg, Germany.³⁹ The lipids were dissolved in 20 mM Tris and 5 mM MgCl₂ buffer, yielding a 10% (w/v) dispersion, which was homogenized by freeze−thaw cycling. A total of 30 μL of the sample was filled into the sample cell. The medium X-ray energy was 16.5 keV, corresponding to a wavelength, λ, of 0.751 Å and a typical photon flux of 10³ photons per second. The sample−detector distance was 1.56 m, and the sample exposure time was−depending on the phase state of the sample−between 0.05 and 1 s. The temperature-dependent measurements at ambient pressure were conducted between 10 and 50 °C.

In order to extract the peak profile from the SAXS curves, background was subtracted from them. The background [BG(Q)] was represented by an exponential function (2)

$$\text{BG}(Q)=A{e}^{-BQ}$$ (2)

where A and B are constant. The net peak profiles were fitted by a log−normal function to estimate the peak position and the width of the scattering peak.

$$I(Q)=C\text{exp}\left\{-{\left[\frac{\text{ln}\left(Q/Q_0\right)}{w}\right]}^2\right\}+D$$ (3)

C, D: constant, Q₀: peak position, w: width. The peak positions and the widths were plotted as a function of temperature. Peak positions are converted to the real space spacing by

$$d=\frac{2\pi }{Q_0}$$ (4)

Molecular Dynamics (MD) Simulations

The initial structures of lipid molecules were built and optimized in GaussView (for details see ref 27, main text). Briefly, lipid14 force fields were used for the lipids of DOPC and DPPC. For lipids not included in the Lipid14, we calculated the partial charge by fitting the electrostatic potentials using the restrained electrostatic potential (RESP) method in the antechamber module of Amber16. The Generalized Amber Force Field (GAFF) and Lipid14 force fields were used to define the atom types, bonded interaction parameters, and van der Waals interaction parameters of lipids. For each lipid type, single molecules were dissolved in water, a short MD simulation was performed to optimize their structure. Membranes composed of different lipids were constructed to mimic the inner and outer membranes of the mycobacterial cell. Further, the composition of the modeled membrane was inspired by previous reports^28,32,33 and hence should be cautiously considered as only a first approximation of the mycobacterium membranes. The ratio of lipid components in the inner membrane was taken as AC₂PIM₂/Cardiolipin/DG/PI/PG/ PE = 50%:10%:10%:10%:10%:10%. The ratio of lipid components in the outer membrane was SL-1/TDM/PDIM/ L A M / D O P C / D P P C / M A = 10%:10%:10%:10%:15%:15%:30%. There were 800 lipids in the inner membrane and outer membrane, respectively. The lipids were packed together to create a lipid bilayer (400 lipids in each layer) using PACKMOL. The numbers of TIP3P water molecules added to build the solvent box for the inner and outer membranes were 56,000 and 78,000, respectively. The box sizes for the inner and outer membranes were 157 × 150 × 122 ų and 166 × 160 × 155 ų, respectively. Sodium ions were added to neutralize the net charge of the system. The steepest descent method was used to minimize the system until the root-mean-square of energy gradient was less than 0.0001 kcal/mol Å or the maximum iteration steps reached 10,000. The system was heated to 300 K linearly in periods of 100 psec in the NVT ensemble with the weak harmonic potential (10 kcal/mol Å) on the heavy atoms. Subsequently, a 1 nsec unrestrained equilibration with Langevin thermostat in the NPT ensemble was performed. Hydrogen bonds were constrained using the SHAKE algorithm. The same simulation parameters of NPT equilibration were used in the 500 nsec production runs performed using CUDA-version Amber16. The last 100 nsec trajectory was used for the analysis. To calculate the conformation freedom of AC₂PIM₂, we used the heavy-atom root-mean-squared fluctuation (RMSF) of AC₂PIM₂ in the last 100 nsec trajectory to derive the average RMSF.

Diffusion Analysis

The lateral diffusions of lipids in membrane planes were calculated using the mean square displacement (MSD)

$$\text{msd}(\tau )=\langle {[r(t+\tau )-r(t)]}^2\rangle$$ (5)

where r(t) is the position of the center of mass (COM) of lipids in the lateral direction at time t, and τ is the lag time to calculate the displacement of the position in the time step. The average displacements over time (t) and the number of particular lipids were calculated. The lateral diffusion constant (D) of lipids was calculated based on the trajectory 300−500 ns by fitting the MSD curve

$$D_{\tau }=\frac{\text{msd}(\tau )}{2d}$$ (6)

where d is the dimensionality (here, d = 2). The “stfcdiffusion” module in Amber16 was used to calculate the MSD curves and estimate the diffusion constants and their errors.

Cluster Size Analysis

To determine the cluster sizes of different lipids, a neighbor connectivity search analysis was done. The center of mass for any two similar lipids in neighboring grids having a distance <10.0 Å was considered to be in the same cluster. The DBSCAN (density-based spatial clustering of applications with noise) algorithm was used to identify the different clusters.

Order Parameter Calculation

Lipid tail-order parameters (S) were calculated for each lipid tail as follows

$$S=\frac{1}{2}\langle 3{\text{cos}}^2\theta -1\rangle$$ (7)

where θ is the angle between two vectors of the C−H bond of all carbons in a lipid tail. The average order parameter of a particular lipid tail was used to estimate the flexibility of the lipid. The “cpptraj” module in Amber16 was used to analyze the order parameter (SCD) of the SL-1 bilayer.

Lipid Shape Analysis

MD Analysis. The lipid shape was characterized by the density of heavy atoms of lipid head-groups and tails in the horizontal plane of the membrane. The positions of heavy atoms in the simulations are projected to the xy-plane, and the position distributions of the headgroup atoms and tail atoms give the shape outlines of the lipids. The area per lipid (APL) characterizes the average area of lipid in IM or OM, which is analyzed by the areapermol module of AmberTools.⁴⁰ The thickness measures the vertical distance (z-direction) between the headgroup atoms in the two layers of membranes, and the thickness plots were generated by APL@Voro.⁴¹ The heavy atom distribution analysis is implemented by in-house code, which account the number of heavy atoms of each lipid in the slices along the z-direction of the membrane. The two lipids with any heavy-atom distance less than 10 Å are considered to be in contact with each other, and the lipid neighbor probability is calculated by

$$p_{ij}\{\begin{array}{l}\frac{\langle N_{ij}\rangle }{n_i\times {{\displaystyle \text{Σ}}}_{k=1}^t\left(\langle N_{ik}\rangle /n_i\right)}\text{if}i\ne j\\ \frac{\langle N_{ii}^{\prime }\rangle }{n_i\times {{\displaystyle \text{Σ}}}_{k=1}^t\left(\langle N_{ij}^{\prime }\rangle /n_i\right)}\text{if}i=j\end{array}$$

in which ⟨N_ij⟩ is the average contact number between lipid i and j in the simulation, n_i is the number lipid i in the membrane, t is the type number of lipids. And

$$N_{ij}^{\prime }=\left(N_{ij}-n_i\right)/2$$

Results and Discussion

Msm Membranes Exhibit Lipid Phase Coexistence with Low Enthalpic Transitions

Mycobacteria harbors structurally unique lipids distinctly organized within the cell wall’s outer and inner membrane regions.^28,29 Some characteristic lipids in the outer membrane consist of free or covalently bound long-chain mycolic acids, phthiocerol dimycocerosate (PDIM), trehalose dimylcoate (TDM), sulfoglycolipids (SL), phosphatidylinositol mannosides (PIMs), lipoarabinomannan (man-LAM), phenolic glycolipid (PGL), and diacylglycerol (DG), Figure 1. The inner plasma membrane harbors diacylated phospho-myo-inositol dimannosides (AC₂PIM₂), phosphatidylinositol mannosides (PIM6), and other ACPIMs.

Figure 1.

(A) Quantitative distribution of various lipids in IM and OM fraction. (B) Representative structures from each lipid category. Lipid categories: FA—fatty acyls; GL—glycerolipids; GP—glycerophospholipids; PK—polyketides; PR—prenol lipids; SacL—saccharolipids. Lipid classes: DAT—diacyl trehalose; TAT—triacyl trehalose; SL—sulfolipids; GPL—glycopeptidolipids; PL—phospholipids; LPL—lysophospholipids; CL—cardiolipids; AcPIMs—acylated phosphatidylinositol mannosides; MA—mycolic acids; MAG—monoacyl glycerol; DAG—diacylglycerol; TAG—triacylglycerol.

First, protein-free model membranes composed of extracted Msm inner and outer membrane lipids were designed and lipid identities were quantified and characterized using mass spectroscopy for each fraction (Figure 1A).⁴² Major lipid categories identified in both membrane fractions were: glycerophospholipids (GPs; red), saccharolipids (SacLs; green), fatty acids (FAs; blue), glycerolipids (GLs; purple), polyketides (PKs, orange), prenol lipids (PR; gray), and others (black) with some lipid classes and species specific to each fraction. GPs consist of phospholipids (PLs), lysophospholipids (LPLs), cardiolipin (CL), and acylated phosphatidylinositol mannosides (AcPIMs) lipid classes, wherein AcPIMs are predominantly found in the inner membrane; GLs consist of triacylglycerol (TG), diacylglycerol (DG) and monoacylglycerol (MG) lipids. SacLs consist of diacyl trehalose (DAT), triacyl trehalose (TAT), sulfolipids (SLs), and glycopeptidolipids (GPLs) and are present in higher abundance in the outer membrane. FAs consist of mycolic acids (MA; alpha-MA, keto-MA, and methoxy-MA) and their conjugates and free fatty acids and their conjugates. Among the FAs, MA is specific to outer membranes only. Lipids constituting PK and PR categories are mycobactins and quinones, respectively. And last, some lipids were classified as others consisting of siderophores. The characteristic structures of essential lipids belonging to each lipid category for each fraction are shown in Figure 1B.

Next, we investigated the thermal lipid phase transitions in these systems using differential scanning calorimetry (DSC). Outer membrane demonstrated two transitions at 1 bar centered at ∼21.7 and 30.8 °C, with ΔH of 0.073 J/g (Figure 2A). Conversely, the inner membrane showed transitions at relatively higher temperature (23.2 and 32.6 °C) with ΔH of 0.25 J/g.³⁸ This is attributed to highly long-chained and branched lipids in the outer membrane lipids accounting for the low membrane order and lower phase transitions.⁴² The inner membrane lipids, on the other hand, have more saturated lipid acyl chains as reported previously leading to a higher phase transition temperatures.⁴² Further, low ΔHs and broad peaks compared to standard phospholipids indicate inhomogeneous lipid mixing and possibly a “two-phase regime” of coexisting lipid phases that induce minor local rearrangements of lipids within the two nearly identical phases at ambient pressure. Multicomponent lipid mixtures as inner and outer Msm membranes studied here, generally show broad transitional lipid phase behavior.^31,42 Structural information on these systems was gained via small-angle X-ray scattering (SAXS) that revealed a diffuse scattering for both the inner and outer membranes and a bump in the range q ∼ 0.4−3 nm⁻¹ (Figure 2B). This could be due to either monolamellar vesicles or a nonregular alignment of adjacent membranes in the multilamellar vesicles, i.e., positionally uncorrelated bilayers.⁴³ For the outer membrane, the spacing derived from the SAXS data decreased with increasing temperature, while an increase in the iwidth above 30 °C suggests that a periodic order of the membrane structure decreases, i.e., it becomes more disordered than at lower temperature (Figure 2C). For the inner membrane at 10 °C, the sequence of two Bragg peaks with q_max (II order)/q_max (I order) = 2:1 arises from stacked bilayers from oligolamellar vesicles (Figure 2C). Furthermore, two Bragg peaks at 10 °C for the inner membrane lipids accounts for the two data points (for both the spacing and width of these two Bragg peaks) in Figure 2C. As the temperature rises above 20 °C, the width monotonically decreases, suggesting that the distribution of the spacings is widened in contrast to that observed for the outer membrane. However, similar to the outer membrane, the spacing increased with temperature but to a lesser extent suggesting a tighter regulation of the inner membrane structure.

Figure 2.

(A) DSC scans of outer (OM) and inner (IM) Msm membranes. (B) SAXS (small-angle X-ray scattering) peak profiles of IM (left) and OM (right). (C) The SAXS peak position and peak width profiles of IM and OM as a function of temperature. (D) 2H solid-state NMR spectra acquired at 122.84 MHz of pure ²H₃₁−POPC liposomes (blue) and ²H₃₁−POPC mixed with inner (IM, red) and outer (OM, black) at 25 °C and atmospheric pressure. (E) Laurdan GP profiles as a function of pressure at different temperatures. Data represented as mean and SD of 5 independent experiments. Significance was determined using one-way ANOVA with Tukey’s multiple comparison test and (F) representative pressure dependence of the symmetric CH₂−stretching mode wavenumber at different temperatures.

The lamellar nature of Msm membranes and lipid phase coexistence were further confirmed using NMR. Deuterium solid-state nuclear magnetic resonance (ssNMR) is a powerful, quantitative and noninvasive tool to study structure and dynamics of lipid assemblies.⁴⁴⁻⁴⁶ By using a small amount of a deuterium labeled lipid added to both inner and outer lipid extracts, the lipidic phases at 25 °C and atmospheric pressure were deciphered. Figure 2D displays ²H experimental spectra of 1-palmitoyl-(²H₃₁)-2-oleoyl-sn-glycero-3-phosphocholine (²H₃₁−POPC) in 3 samples, i.e., liposome of pure ²H₃₁–POPC (blue), and ²H₃₁−POPC mixed with inner (red) and outer (black) lipid extracts (see SI for sample preparation). ²H₃₁−POPC liposome (Figure 2D, blue) spectrum is typical of a lamellar phase with several quadrupolar splitting, the larger one corresponding to the CD₂ positions (carbons 2 to 5) close to the glycerol (25.7 kHz) and the smallest at the center, close to the chain methyl terminal. Intermediate splitting corresponds to the other CD₂ of the palmitoyl-(²H₃₁) chain. A small isotropic signal (traces of water, HOD) is also visible in the center of the spectrum and corresponds to less than 1% of the signal (estimated by spectral simulations, Figure S1). The lamellar phase was observed in both the inner and outer membranes, with a substantial proportion of isotropic peak; spectral simulations estimate it to be 20 ± 3% and 30 ± 3%, respectively. Although it is not characterized at the moment (isotropic or cubic lipid phase), the isotropic peak indicates the presence of another phase along with the lamellar one. This supports the thermotropic data indicating the presence of lipid phase coexistence regions in Msm model membranes.

Residual Fluidity in Msm Membrane Layers at Extreme Conditions

Hydrostatic pressure (HP) drives structural polymorphic changes in membranes,^10,21 accompanied by reduction in volume. HP minimizes the hydrocarbon acyl chain motions and concomitantly increases chain ordering.^47,48As the cross-sectional area of the lipid headgroup is less pressure-sensitive compared to the tails, pressure drives transitions to phases with spontaneous curvatures, and increases phase separation between coexisting phases contributing to lateral restructuring. Thus, pressure is a uniquely poised to investigate phase transitions between even subtly different lipid phases. We probed the variation of the membrane order and hydration in the model membranes with pressure and temperature using Laurdan, a well-known polarity-sensitive probe.^5,49Laurdan’s spectral shift, sensitive to membrane polarity, was quantified by ratio metric spectroscopy and generalized polarization (GP) that reports on the membrane packing/order and hydration. Higher positive GP values indicate a tightly packed ordered membrane and vice versa.

Slope changes in pressure-dependent Laurdan GP at various temperatures indicate lipid phase transitions impacting membrane hydration (Figure 2E). For the inner membrane, at ambient temperature (25 °C), the GP value increased steadily with pressure due to chain ordering up to ∼700 bar and plateaued with a negative slope at 0.5 (**p = 0.0024) indicating a coexistence of liquid-ordered phases, such as l_o1 +Al_o2. At 25 °C, the system at 1 bar already possesses a liquid-ordered (l_o)- liquid disordered (l_d) phase coexistence,³⁸ thus likely reflecting a transition from l_o + l_d to l_o1 + l_o2 with pressure. This is also supported by absence of sharp pressure-induced phase transitions (as seen for one component lipid Figure S2), thus implying transitioning to coexisting lipid phases with overlapping degree of membrane packing and hydration. At 10, 40, and 60 °C, slope changes were seen at 700 bar; 700 bar and 1.5 kbar; 1 kbar and 1.5 kbar, respectively. Slight downward slopes of GP curves with pressure suggest a hindrance to form an all-ordered lipid phase (e.g., a gel/solid-ordered s_o phase) in the inner membrane within 2 kbar and up to 60 °C. The downward GP slope occurred at even higher pressures with increasing temperatures indicating that a pressure-induced intermediate fluid phase is populated faster at lower temperatures. Thus, the inner membrane likely adapts to extreme conditions by regulating its optimal fluidity over a wide range of pressure and temperature. Similar analysis for outer membrane revealed relatively smooth pressure-induced transitions with slope changes at 500 bar and 1.3 kbar at 25 °C, 350 bar and 1.5 kbar at 40 °C, and ∼700 bar at 60 °C; at 10 °C, no significant changes in GP were seen. These indicate that phase transitions in outer membrane toward coexisting ordered lipid phase regimens are faster with pressure at higher temperatures. This was inferred from increasing (ΔGP/Δp; initial slopes) with increasing temperature.

Orthogonal studies using pressure dependence of lipid vibrational IR bands, particularly the symmetric CH₂ vibration (νCH₂,_sym) were undertaken. νCH₂,_sym exhibited a linear dependence on pressure and deviation from this behavior manifested as slope changes or discontinuities indicating phase transitions (Figure 2F).²¹ For inner and outer membranes, an initial decrease of νCH₂,_sym (slope change) with pressure at all temperatures suggested a transition to lipid phase with lower abundance of gauche conformations. This transition required higher pressure at higher temperature due to the membranes being in fluid liquid crystalline phase at temperature >30 °C. At 60 °C, the fluid liquid crystalline phase remained stable until almost 2 kbar. Beyond 3 kbar, the νCH₂,_sym increased with minor slope changes at 4.4 kbar (10 °C); 4.0 kbar (40 °C) and 4.2 kbar (60 °C) for the inner membrane, and at 4.3 kbar (40 °C), and 5.7 kbar (60 °C) for the outer membrane underlying gradual increase in gauche conformers. It indicates rearrangement of lipids within two-phase regions toward a relatively rigid phase under pressure; l_o1 + l_o2 to l_o + s_o or l_o + s_o to s_o. The all ordered s_o phase accumulates only under every high pressures. This series of pressure-induced structural phase transitions involves modifications in the acyl chain packing and to some extent in the headgroup region, driven by changes in the mutual accommodation between the area of the headgroup and the acyl chains due to pressure-induced fluctuations. The discontinuities in the GP and νCH₂,_sym were used for constructing the tentative p−T phase diagrams for Msm membrane models.

High Pressure-Induced Lipidome Remodelling and Cell Viability in Intact Bacteria

Next, we mapped the observed pressure-induced lipid phase changes in Msm model membranes to pressure-induced lipidome alterations in intact bacteria. We subjected live bacteria to high hydrostatic pressure of 1030 bar for 12 h at 37 °C. The bacteria grown for the same duration at 1 bar at 37 °C was used as a control. From these, total lipids were extracted, and lipidomic analysistiwas performed. Upon high pressure treatment, quantitative analysis revealed that FA and GP lipids increased. In contrast, PK and SacL lipids decreased in abundance (Figure 3A). The significant contributors for the increased FAs were mycosanoic acids and for increased GPs were lyso-PG and PE. Likewise, major contributing lipids in PK and SacL were dideoxy mycobactins (DDeMB), phosphomyciketides (PMk), and glycopeptidolipids V/VI (GPL V/VI), respectively (Figure 3B). The decrease of MA (Figure 3C) accounts for reduced ordering or higher fluidity at higher pressures due to the high ordering in MA chain conformations.^50,51 This could be the main factor for residual fluidity in OM model membrane upon pressurization. Next, we classified the relative lipidome changes in terms of degree of unsaturation (DOU) and acyl chain lengths. We found a significant increase in the mol % abundance of lipid species having 1 double bond coming from Lyso-PG, PE, and nonribosomal PKs like mycobactins. While the mol % abundance of saturated lipids also increased (due to free FAs, MG, Lyso-PI, PI, AcSGLs, and other lipids like siderophores) (Figure 3E), the fold change was higher for monounsaturated lipids thus governing fluid membrane at high pressures (Figure 3D). No changes were found for lipids having >3 double bonds. Another factor regulating membrane fluidity is lipid acyl chain length, wherein very long chains display higher hydrophobic and van der Waals interactions, thus ordering the membrane bilayer via tighter packing.^52,53 Reduction of lipids with chain lengths longer >50 with concomitant increase of species having short chain lengths (<40) thus underlines residual fluidity under pressure in mycobacterial membranes (Figure 3F). Finally, we accessed the bacterial viability under pressure and found no major effect on Msm viability reinforcing a presence of functional membrane (Figure 3G). The same is supported by residual fluidity in the in vitro Msm membrane models under these conditions.

Figure 3.

(A) Mol % total lipid abundance at the lipid class level shows differential distributions as a function of pressure. [ANOVA with Fisher’s LSD test, *P < 0.05, **P < 0.005, ****P < 0.0005, ****P < 0.0001]. (B) The mol % of lipids within each subclass contributing to pressure-induced lipidome changes. (C) The change in the mol % of mycolic acids due to pressure perturbation (D) the mol % total lipid abundance distribution as per the Degree of Unsaturation (DOU varying from 0, 1, 2, >5 double bonds). (E) The mol % of lipids contributing to pressure-induced unsaturation changes. (F) The mol % of lipids with varying lengths in their acyl chain components. (G) The effect of pressure on bacterial cell viability as a function of time. Data represented is mean and SEM of 3 independent experiments. Statistical analysis was determined using Unpaired t-test [*P < 0.05, **P < 0.005, ****P < 0.0005, ****P < 0.0001]. All data shown are Mean ± SD unless specified. Abbreviations MA&C—mycolic acids and conjugates, CL—cardiolipin, LPE—lyso phosphatidylethanolamine, PE—phosphatidylethanolamine, LPG—lyso phosphatidylglycerol, DDeMB—dideoxy mycobactins, PMk—phosphomyciketides, GPL V/VI—glycopeptidolipids V/VI, FAC—free fatty acid and conjugates, MG—monoacylglycerol, LPI—lysophosphoinositols, PI phosphoinositols, AcSGLs—acylated sulfoglycolipids, PG Phosphatidylglycerol, NKPK—nonribosomal polyketide (usually mycobactins).

Lipid Shape and Interdigitation Fluctuations under Pressure Reveal Specific Lipids in the Msm Layers as Key Membrane Fluidity Regulators

To delve deeper into molecular details, we performed all-atom molecular dynamics simulations and analyzed the head and tail shape changes of constituent lipids within the inner and outer membrane fractions with pressure using previously characterized Msm membrane models.³⁰ For DG within the inner membrane, pressure induced a qualitatively higher conical shape (Figure 4A). Quantitatively, the conical shape was analyzed by ratioing the areas of heads and tails to estimate the area of density maps of headgroup atoms (H) and tail atoms (T); H/T (Table 1).

Figure 4.

Pressure-dependent changes in the (A) shapes of representative lipids from the inner (IM) and outer membranes (OM). The 2D heavy atom density projections of the headgroup and tail atoms for indicated lipids in the x−y membrane plane is shown. (B) Changes in the area occupied per lipid in inner and outer membranes and (C) bilayer thickness.

Table 1. Density Map Area Ratio of Lipid Head Groups and Tails (Head/Tail: H/T) for Specific Lipids in Msm Inner and Outer Membranes.

pressure (bar)	MA	TDM	DG	PG
1	0.07	0.28	0.14	0.52
1654	0.09	0.53	0.11	0.50

Smaller values of H/T mean higher conical shape. Higher conical shape of DG could account for the relatively fluid lipid phase under pressure in the inner membrane (as seen in Figure 2E). For phosphatidylglycerol (PG), head compactness was observed reducing the H/T value and increasing the conical shape propensity, while the acyl tail conformations exhibited irregular shape (this irregularity was also observed for other inner membrane constituent lipids). In outer membrane, most dramatic change was observed with TDM wherein the conical shape at ambient pressure transformed to a cylindrical shape at high pressure (increased H/T value, Table 1), affording tighter packing (possible liquid ordered phases) at high pressure. For other lipids, the shape was marginally affected, but tail density became irregular with pressure; as seen for MA and LAM (Figures 4A and S3). The irregularity could arise from intrachain conformational and interchain dynamics at higher pressure, and thus underline the molecular changes in lipids within the membrane phases to retain residual fluidity under pressure-induced ordered phases.

Area per lipid slightly increased for outer membrane (by ∼1 Ų) followed by a marginal decrease. For inner membrane, a nonuniform trend indicative of transitioning to possible phases with noncanonical properties (Figure 4B) was observed. Specifically, the dip around 800 bar suggests that this pressure-populated state has a similar area/lipid as observed under ambient pressure (1 bar), supporting a fluid-like nature of lipid phase induced under intermediate pressures. This reduction at intermediate pressure could also be due to lipid acyl chain interdigitation. This finding also supports the downward slopes in GP with pressure for inner membrane (Figure 2E). Both inner and outer membrane showed a decreased bilayer thickness with pressure. This underlies a fluidic nature of intermediate pressure-induced phases as thinner membranes have higher fluidity.⁵³

Next, pressure sensitivities of lipids were evaluated at molecular levels (Figure 4C). Nonintuitively, at higher pressures, S_CD (lipid order parameter) decreased for AC₂PIM₂, cardiolipin (CL) and DG (Figure S4) within inner membrane; other phospholipid counterparts exhibited either no change or marginal increase, but only at very high pressures. These reflect reduced lipid order within selective inner membrane lipids and changes in orientation of their chains that tend to become less aligned to the direction perpendicular to the bilayer plane. AC₂PIM₂, CL, and DG thus likely contribute toward the fluid-like nature of inner membrane lipid phases under pressure. For outer membrane lipids, changes were seen with PDIM, SL-1, and MA wherein the S_CD decreased again implying fluid-like characteristics under pressure. Interestingly, lipid order was chain-length dependent having distinct pressure sensitivities. For example, S_CD substantially decreased with pressure between C20−C35 in MA and SL-1 lipids having long chains (Figure S5); in same lipids the shorter chains, if any, remained stable. For PDIM, the decrease was also seen in shorter chains (∼C20−C22). Thus, MA, SL-1 and PDIM probably contribute toward the fluid-like nature of the lipid phases in outer membrane under pressure; TDM acting otherwise. This was supported by the lower diffusion rates (D) of both membranes under pressure. However, DG within inner and SL-1, MA and PDIM within outer exhibited the most negligible reduction (Figure S6, Table S1). The modeled diffusion rates of these membrane layers have been experimentally verified before.³⁰ Phospholipids within Msm layers exhibited the most considerable reduction in D with pressure underlying a lipid-specific pressure-sensitivity and resistance mounted by noncanonical Msm lipids to adapt to typical pressure-induced ordered lipid phases/conformation. With temperature, DG and PDIM showed the maximum increase compared to other membrane lipids (Tables S2 and S3). These data show that under pressure, both membranes adapt to physicochemical stresses by modulating the geometry and conformation of specific lipids, eventually regulating membrane fluidity.

Next, heavy atom distributions along the membrane z-direction revealed that many lipids from one leaflet stretched into the other layer, such as CL, phosphatidylethanolamine (PE), PG, phosphatidylinositol (PI) of inner membrane, and MA, PDIM, and SL-1 of the outer membrane even at ambient pressure, with minor modulation under pressure (Figure 5). The intermonolayer lipid chain extension was higher in outer membrane attributed to the long-chained MA, PDIM and SL-1. Under pressure, a higher intermonolayer chain extension was observed for DG in the inner membrane. A more extended chain conformation in DG at higher pressure together with higher conical shape could signify pressure-induced interdigitation contributing to fluidity.

Figure 5.

Pressure-dependent changes in the lipid chain extension from the top and bottom leaflet within the inner (A) and outer (B) membrane bilayers. Lower panel shows lipid chain extension of specific lipids within each layer.

Lipid Sorting in Response to Pressure and Temperature Stress

Lipid sorting is critical for membrane homeostasis, as well as for orchestrating membrane-associated signaling and trafficking events.⁵⁴ Motivated by these, we modeled lipid clustering in Msm membrane layers under pressure and temperature perturbation (Figure 6). Under ambient conditions, large clusters were observed (in order) for MA > PDIM > LAM > TDM. At the same time, SL-1 mainly existed as small dimer-pentamer clusters in outer membrane.

Figure 6.

Average number of lipid clusters with different sizes in Msm inner and outer membranes and their variation with (A) pressure and (B) temperature. Values were averaged over the last 100 ns of the simulation.

Article

Similarly, large Ac₂PIM₂, PI, PG, and DG clusters (11−20) were found in the inner membrane, while other lipid clusters were limited to dimers and pentamers. For inner membrane, pressure increased the abundance of small-sized clusters of PG and DG and diminished larger clusters of Ac₂PIM₂ (Figure 6A), supporting least reduction in the diffusion of DG. These underline mitigation to highly ordered lipid phase state at intermediate pressures by favoring small sized lipid clusters. Temperature had a distinct effect, wherein intermediate sized cluster (6−10) of Ac₂PIM₂ appeared at higher temperature while the larger (>20) and small clusters remained intact (Figure 6B). This implies that at low temperatures, clustering pattern of Ac₂PIM₂ regulates membrane fluidity. Similar effect was seen with PI clusters, and for DG, larger clusters (>20) broke into clusters of 11−20, while dimers diminished, probably balancing the membrane fluidity/other properties at low temperatures.

In outer membrane, upon pressurization, large clusters for PDIM decreased (supporting lipid diffusion) but increased for TDM, intermediate (3−5) sized clusters for LAM increased, and population of small clusters of SL-1 increased; MA showed no change (Figure 6A). With temperature, while SL-1 clusters remain unchanged, bigger clusters of PDIM and TDM reduced, and intermediate-sized (6−10) clusters for LAM increased, but decreased for MA (Figure 6B). The decrease of MA clusters underlines tight regulation of lipid clustering modulating diffusion/fluidity under low temperatures and intermediate high pressure.

To further evaluate the relative enrichment or depletion of specific lipids within each lipid cluster under pressure or temperature, we calculated the average number of neighbors for different lipids (Figure 7). In the inner membranes, PG and DG were excluded from Ac₂PIM₂ clusters at 1 bar. At the same time, CL and PI were enriched in these clusters. Under pressure, the cluster organization remained stable with no significant loss of specific lipid−lipid interactions, and this probably contributes to mitigating pressure-induced ordering. On the other hand, with an increase in temperature (Figure S7), exclusion of CL around Ac₂PIM₂ was observed, while DG was selectively enriched around PG. For the outer membrane, 1,2-Dioleoyl-sn-glycero-3-phosphocholine (DOPC) and TDM enrichment around MA increased with pressure, and PDIM/LAM/SL-1: DOPC contacts decreased. Pressure-mediated conical to cylindrical shape transformation of TDM and MA could underline more permissive interactions between these lipids. With temperature, PC’s enrichment around MA and LAM’s around PDIM decreased (Figure S7). Thus, specific lipid−lipid interactions were rewired in outer membrane at both high pressure and temperature. These indicate that at higher pressures and temperatures, the Msm membranes mainly consist of small/intermediate-sized lipid clusters with stable lipid−lipid contacts that could enable it to maintain a fluid-like lipid phase.

Figure 7.

Pressure-dependent change in the contact neighbors among the inner (IM) and outer membrane (OM) lipids. Number of neighboring lipids (within 1.5 nm) in the Msm membrane models for the last 100 ns of the simulation is shown. Values were normalized to the weighted average number of neighbors for each type. Blue indicates lipid enrichment, and orange indicates lipid depletion.

Conclusions

Fourier transform infrared spectroscopy, NMR, DSC, Laurdan fluorescence spectroscopy, fluorescence microscopy, and atomic force microscopy were used to delineate the temperature- and pressure-dependent changes in the phase behavior of the described Msm inner and outer membranes to constitute their tentative p,T-phase diagram for the first time (Figure 8). Both membranes exhibited a series of pressure-induced transitions between 10 and 60 °C and up until 5−6 kbar. The fluid disordered → (l_d) → l_d + l_o, l_d + l_o → l_o1 + l_o2, l_o1 + l_o2 → l_o + s_o → s_o phase boundaries exhibit transition slopes of about ∼14−20 °C/kbar in both membranes. Of note, the designation of Msm lipid phases as l_d or l_o or s_o should be taken cautiously as these phases have been defined and characterized in standard phospholipids or their mixtures, including cholesterol, which are structurally distinct from mycobacterial lipids.

Figure 8.

Tentative p, T-phase diagram of the (A) inner and (B) outer mycobacterial membrane lipids as obtained from pressure-dependent FTIR spectroscopy (red □) and GP (O) at various temperatures and temperature-dependent FTIR (♦), GP (▾), and DSC (×) at 1 bar.

The findings collectively indicate that Msm inner and outer lipid membrane layers composed of diverse and structurally distinct lipids exhibit a two-phase regime of coexisting lipid phases over varied pressure and temperature, with shallow phase transitions to other two-phase coexistence regime with overlapping degrees of membrane properties. The mycobacterial cell often encounters diverse environmental conditions and stressors, including pH changes, osmotic pressure, and antibiotic exposure. The bacteria’s responses to these challenges are mediated by protein functions, which rely on the fluidity and order of the lipid membrane. Under pressure stress, specific Msm lipids within each layer undergo shape and conformational fluctuations to resist pressure-induced triggers to highly ordered lipid phases and maintain residual fluidity underlying nonattenuated bacterial viability. This is further enforced by remodelling of the bacterial lipidome under pressure by enriching lipids with higher unsaturation, shorter chain lengths and reduced abundance of long-chained mycolic acids. This enables maintenance of residual fluidity in otherwise order-inducing environment. The bacteria’s ability to regulate and maintain membrane fluidity is crucial for lipid mobilization and efficient downstream signaling, both of which are essential for its survival and infectivity. Thus, it can be suggested that mycobacteria adapt to extreme conditions including that within the host intracellular environment by tightly regulating its membrane fluidity and hence regulating subsequent membrane-associated cellular functions to foster survival.

Lamellar liquid lipid phases, considered the most biologically relevant in natural cell membranes, extend over a wide temperature−pressure range in Msm lipid membranes. Subtle lipid remodeling underlying modulation of membrane fluidity, conformational changes, lipid clustering, sorting, and lipid–lipid contacts under pressure and temperature implies that Msm lipid membranes span a spectrum of nearly identical lipid phases with similar fluidity, likely differing in the local composition and orientation of lipids within these phases. This enables mounting a robust response to varied extreme conditions to maintain functional membrane layers with optimal fluidity but with different sensitivities within the two layers. The exact functional relevance of this interbilayer membrane homeostatic behavior warrants further studies. Integrating our findings on the unusual fluid-like membrane characteristics in mycobacteria under extreme conditions has the potential to reveal new functional insights on their unique lipids within the context of host−pathogen and drug interactions as well as membrane responses during various stages of infection. This study is a proof-of-concept multimode analysis of biologically derived lipids and their phase properties under extreme conditions. Observations made in Msm would require verification in mycobacterial species infecting deep sea organisms for a comprehensive understanding of the membrane physiology under high pressure and low temperatures.