Biophysical Interaction Landscape of Mycobacterial Mycolic Acids and Phenolic Glycolipids with Host Macrophage Membranes

Abstract

Lipidic adjuvant formulations consisting of immunomodulatory mycobacterial cell wall lipids interact with host cells following administration. The impact of this cross-talk on the host membrane’s structure and function is rarely given enough consideration but is imperative to rule out nonspecific perturbation underlying the adjuvant. In this work, we investigated changes in the plasma membranes of live mammalian cells after exposure to mycobacterial mycolic acid (MA) and phenolic glycolipids, two strong candidates for lipidic adjuvant therapy. We found that phenolic glycolipid 1 softened the plasma membrane, lowering membrane tension and stiffness, but MA did not significantly change the membrane characteristics. Further, phenolic glycolipid 1 had a fluidizing impact on the host plasma membrane, increasing the fluidity and the abundance of fluid-ordered−disordered coexisting lipid domains. Notably, lipid diffusion was not impacted. Overall, MA and, to a lesser extent, phenolic glycolipid 1, due to minor disruption of host cell membranes, may serve as appropriate lipids in adjuvant formulations.

Introduction

Mycobacterium tuberculosis (Mtb), the causative pathogen for pulmonary tuberculosis (TB), has evolved for millennia within its human host. A major outcome of this coevolution has been a synchronized response of innate and adaptive host immune cells to mycobacterial cell wall lipids.^1,2 This in turn has triggered the therapeutic use of the immunostimulatory properties of the mycobacterial cell wall lipids; for instance, either as vaccine adjuvants or in drug delivery.^3–5 In this direction, one of the most explored Mtb lipid is Mycolic acids (MAs), which are long chain β-hydroxy fatty acids (FAs) bound to the cell wall via an ester bond.⁶ They also constitute the acyl chains for the noncovalently bound glycolipids present in the mycobacterial outer membrane. Three major classes of MA are alpha (α-) MA, keto-MA, and methoxy (MeO-) MA (Figure 1A)^7,8 and play an important role both as a physical barrier and in infection biology.⁹ Another recent addition is the phenolic glycolipid (PGL), shown to boost acquired immune responses as an adjuvant in vivo.¹⁰ PGL is a surface lipid and glycosylated version of phthiocerol dimycocerosate (PDIM) − the most extensively studied mycobacterial virulent lipid.^11,12 Its structure is composed of saccharide units (1−3 units) bound to a phenyl group, a mycocerosic acid and a phthiocerol region, Figure 1B.

Figure 1. Chemical structures of (A) alpha, keto, and methoxy MAs and (B) phenolic glycolipid, PGL1.

Red color in PGL1 structure depicts the saccharide units, and blue is the phenyl motif; absence of these generates PDIM. (C) Quantitative mass analysis of alpha, keto, and methoxy MAs (in mol %) in the extracted and purified fraction of Mycobacterium bovis. Lower panels depict the mol % abundance of constituting acyl chains in increasing order of carbon chain lengths within each class of MA.

In recent years, the need for efficient adjuvant formulations eliciting wanted cellular immune responses have triggered the generation of novel adjuvant combinations.¹³ Archaeosomes from the polar lipid portion of archaea triggering both a humoral and a cell-mediated immune response and cationic liposomal formulation of immunostimulatory total lipid extract of Mycobacterium bovis Bacillus Calmette-Guérin (BCG) and dimethyl dioctadecyl ammonium bromide (DDA), myco-somes, are few examples illustrating this progression.¹⁴ Cellular administration of such lipidic adjuvant formulations includes interaction of the same with host cells. Still, one aspect not given sufficient attention is the effect of these components on the host cell membranes, the first point of contact preceding the initiation of any cellular immune response. This is mainly driven by the lipidic nature of the cell wall components, which are likely to insert within the host cell plasma membrane (PM) and modify its properties, thereby modulating membrane-associated processes. These can have an influence on the desired immune responses expected from mycobacterial cell wall lipid-based adjuvants and merit proper investigation.

In this direction, we and others have shown unprecedented effects of structurally diverse mycobacterial cell wall lipids on host cell membranes.^11,12,15,16 For instance, while trehalose dimycolate (TDM) and PDIM stiffen the plasma membrane and increase membrane order, sulfoglycolipids show opposite effects; all these lipids reside within the same cell wall envelope layer though.^11,15–18 These clearly suggest that the chemical and structural makeup of mycobacterial lipids play a crucial role in dictating the modulation of the host membrane’s physical attributes. However, despite this, no studies detailing the interaction of MA and PGL with host cell membrane exist but are critical to fully evaluate the adjuvant potential of these lipids underlined by no perturbation to the host cell membrane function.

Here, we monitored the effect of MA and PGL1 on the THP-1 macrophage cell membrane. MA used for the study consisted of all three species; α-, keto-, and MeO-MAs in the same relative abundance (Figure 1C) as found physiologically within the bacterial cell wall.^19,20 Monitoring membrane elasticity and lateral domain organization, we noted that while no major change in the plasma membrane organization and dynamics was seen with MA, PGL1 induced membrane softening and enhanced membrane fluidity. The latter could enhance receptor clustering or favorably impact their structural conformation, orientation, and distribution for efficient immune activation,²¹ thereby substantiating their use in adjuvant therapy or delivery platforms. Comparison with other mycobacterial lipids furnished a fuller understanding of the role of exogenous bacterial lipids in affecting the host membrane surface. In the future, investigations into species-specific lipids would render a robust view for their use in therapeutic avenues.

Experimental Section

Cell Lines, Mycobacterial Lipids, and Reagents

Fetal bovine serum (FBS), Roswell Park Memorial Institute medium (RPMI-1640), 1× Dulbecco’s phosphate buffer saline (D-PBS), sodium bicarbonate, 100× antibiotic−antimycotic solution containing amphotericin-B, streptomycin, penicillin, 0.25% trypsin−EDTA, and cell freezing solution were purchased from HiMedia. Laurdan[1-(6-(dimethylamino)naphthalen-2-yl)dodecan-1-one,6-dodecanoyl-2 di-methylaminonaphthalene] was obtained from Sigma-Aldrich (USA) and Phorbol-12-myristate-13-acetate (PMA) and paraformaldehyde from MP Biomedicals (USA). THP-1 cells were a kind gift from Prof. Sarika Mehra (Chemical Engineering, IIT Bombay). THP-1 cells were cultured as per previous reports.^5,15 Before experiments, the THP-1 cells were differentiated using 50 nM PMA for 48 h, after which they were incubated in complete RPMI media containing no PMA for 24 h. A pure homologous mixture of MA from M. tuberculosis (bovine strain) was obtained from Sigma-Aldrich (USA). PGL1 was obtained through BEI Resources, NIAID, NIH: Mycobacterium leprae phenolic glycolipid-I (PGL-I), NR-19342.

Preparation of Mycobacterial Lipid Suspension

Lipid suspensions for cell-based assays were prepared by gentle hydration and the freeze−thaw method reported previously.^5,15,22

Fluorescence Lifetime Imaging Microscopy

Differentiated THP-1 cells were treated with MA and PGL1 at a concentration of 10 μg/mL for 4 h. For staining, 5 μM Laurdan was added to cells and incubated for 10−15 min at room temperature, and imaging was performed at 25 °C in phenol-free culture media supplemented with 25 mM 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid) (HEPES).

Fluorescence lifetime imaging microscopy (FLIM) was performed on a MicroTime 200 (Picoquant, Gmbh, Germany) using time-tagged time-resolved (TTTR) methodology. This setup is attached to an inverted microscope (IX71) equipped with a water immersion objective (UPlan SApo NA 1.2, 60×, WD = 0.28 mm). Laurdan was excited at 405 nm using a pulsed diode laser. Fluorescence emission was collected using a 460/60 bandpass filter. The collected fluorescence signal was directed to a single-photon avalanche photodiode (SPAD) detector. The images were acquired with 512 × 512 pixel resolution. The fluorescence decays were analyzed using the built-in software of MicroTime 200 and were fitted using iterative reconvolution to a triexponential function.

$$I_{\lambda }(t)=I_0(\lambda )\mathbf{\sum }_{i-1}^n{a}_i{\text{e}}^{-t/{\tau }_i}$$ (1)

where I(t) is the intensity of the fluorescence at time t, a_i is the pre-exponential factor for the fraction of the fluorescence intensity, τ_i is the fluorescence lifetime of the emitting species, and n is the number of exponentials used. The average fluorescence lifetime was calculated using the following relation.

$${\tau }_{\mathrm{avg}}=\frac{{{\displaystyle \mathbf{\sum }}}_{\mathrm{i=1}}^3{a}_i{\tau }_i}{{\alpha }_1+{\alpha }_2+{\alpha }_3}$$ (2)

where α_i is the pre-exponential factor representing the fractional contribution of the decaying component with a lifetime τ_i. In order to segregate the contribution of internalized dye, we used a particular feature in the FLIM software SymPhoTime64, provided by Picoquant, which enables quantification of lifetime signals within the user-defined regions of interest (ROIs) (cell plasma membrane).

Fluorescence Recovery after Photobleaching

Differentiated THP-1 cells were treated with the indicated mycobacterial lipids at a concentration of 10 μg/mL for 4 h. Later, they were treated with 2 μg/mL of N-Rh DHPE, and fluorescence recovery after the photobleaching experiments were performed on N-Rh DHPE as per our previous reports.¹⁵ For a detailed description and data analysis employed, please refer to refs 15 and 23.

Atomic Force Microscopy

Atomic force microscopy (AFM) force spectroscopy was done in contact mode with an MFP-3D atomic force microscope (Asylum Research, Santa Barbara, CA, USA.) Silicon nitride cantilevers (Oxford Instruments) at a resonance frequency and a spring constant of 22 kHz and 0.16 N/m, respectively, were used. Cantilever calibration was done as per previous report.¹⁵ Differentiated THP-1 cells were incubated with mycobacterial lipids at a concentration of 10 μg/mL, respectively, for 4 h, and force curves were recorded in the cell body region between the periphery and the nucleus. At least 70 cells were used for force spectroscopy under each condition. Igor software (Asylum Research) furnished the elastic modulus by fitting the force curve using the Hertz model.¹⁵ For analysis of membrane tethers, please refer to refs 15 and 24.

Laurdan Two-Photon Generalized Polarization Imaging

Laurdan imaging experiments were performed and quantified as per our previously published reports.¹⁶ Briefly, THP-1 cells were treated with indicated lipid suspensions at a concentration of 10 μg/mL for 4 h, followed by washing with phosphate buffer saline (PBS). Afterward, Laurdan (5 μM) in serum-free media for 30 min was added, and spectral imaging was performed on a Zeiss LSM 780 confocal microscope (Carl Zeiss, Germany). Excitation was set at 780 nm using a multiphoton (titanium sapphire) laser (coherent radiation, CA), and images were reordered between 400 and 600 nm with 9 nm steps. The images were analyzed using a reported Spectral imaging toolbox compatible with Matlab, and pseudocolored generalized polarization (GP) maps were generated.

Calibration images were collected to calculate the G factor using 5 μM Laurdan solution in DMSO and eq 3

$$G=\frac{G{P}_{\text{ref}}+G{P}_{\text{ref}}G{P}_{\text{mes}}-G{P}_{\text{mes}}-1}{G{P}_{\text{mes}}+G{P}_{\text{ref}}G{P}_{\text{mes}}-G{P}_{\text{ref}}-1}$$ (3)

GP_ref is the reported reference GP value (0.207) of Laurdan, and GP_mes is the measured value of the dye in pure DMSO.

Two channel images were used to construct the pseudocolored images using eq 4

$$G=\frac{I_{400-460}-G{I}_{470-530}}{I_{400-460}+G{I}_{470-530}}$$ (4)

The GP values cover a range of −1.0 (high fluidity) − +1.0 (lowest fluidity, ordered phase) in a custom color palette. The stacks of confocal images were converted to binary images. All GP images were corrected using the G-factor, as mentioned above. Laurdan pixel counts were obtained from the Matlab-based macro, and GP distributions were obtained from the histogram of the GP images. Histograms were fitted to Gaussian functions using Origin (Origin Pro.9.1). The plasma membrane GP images and distributions were obtained after segmentation of membrane pixels using previous method.²⁵

Results and Discussion

Mycobacterial PGL1 Not MA Modulates Membrane Stiffness and Tethers of THP-1 Macrophages

Migration, cell cycle, differentiation, and other cellular activities are finely controlled by membrane stiffness.²⁶ Modulation of membrane stiffness alters membrane rigidity and impacts molecular interactions within the membranes, rewiring cellular immune signaling. AFM was used to examine these nanomechanical characteristics of host membranes upon exogenous exposure to physiologically appropriate quantities of mycobacterial lipids.

The cells treated with PGL1 exhibited decreased PM cortical stiffness (Median Elastic modulus; MEM: 1.24 kPa in control cells; MEM: 0.78 kPa PGL1), while MA induced no significant change (Figure 2A,B). PGL1 possesses quadruple CH₃ branches, which could reduce host membrane lipid packing due to steric hindrances and hence contribute to softer membranes, as seen. This has indeed been previously seen with Mtb sulfolipid-1^15,27 harboring extensive CH₃ branches. Interestingly, despite being structurally related to PDIM, PGL1-induced host membrane softening implies opposite behavior to that of virulent PDIM¹⁵ and reveals an important role of the phenolic glycosylation in impacting the host membrane structure. The terminal trisaccharide of PGL1 drives its binding to complement receptor 3 (C3), invoking an immune response. This binding is likely to restrict PLG-1 lateral localization on the host cell membranes to within distinct lipid domains and induce local disordering of these regions,²⁸ contributing to enhanced softness.

Figure 2. Effect of mycobacterial lipids on the THP-1 membrane properties. Representative force−indentation curves of control and mycobacterial lipid-treated cells.

First, 0.5 μm of the force curves were fitted with Hertz equation to obtain estimates of cortical stiffness. (B) Elastic modulus distribution of THP-1 cells with and without mycobacterial lipid treatment, MA and PGL1 (10 μg/mL) for 4 h (n = 70 cells per condition) fitted with a single log−normal distribution. Median elastic modulus is shown in the insets. (C) Relative frequency of the membrane tether forces in cells with (n = 855 elements) and without (n = 633 elements) treatment with mycobacterial lipids fitted with a log−normal distribution. The median values are provided in the inset. For data in B,C, P < 0.05, Mann−Whitney test was used. (D) Changes in the tether number with the median values in the inset and (E) changes in tether length of cells in the presence and absence of mycobacterial lipids (10 μg/mL, 4 h). (***P < 0.005 unpaired student t-test.)

On the other hand, no effect seen with MAs, despite having long alkyl chains, suggests the involvement of the high conformational flexibility of MA chains within each subtype (α, keto, and methoxy) in mitigating rigidification or softness of the host membranes. It has been reported that each MA possesses distinct physicochemical attributes leading to diverse lipid phases spanned by these species. MA chains adopt various conformations;^29,30 the four-chain W-shaped parallel chain configuration (highest packing efficiency), folded-extended conformation, and fully extended states (least packing efficiency). While both alpha and MeO-MA (accounting for >50% abundance) possess low packing efficiency and form liquid condensed lipid phases, keto-MA exhibits highest rigidity forming solid condensed lipid phases.³¹ The CO group of keto-MA remains hydrated at the water surface. This facilitates keto-MA to adopt a relatively constrained W-conformation. Since the methoxy group of MeO-MA has some water affinity and cyclopropanes in α-MA have no hydrophilicity, these MA subtypes adopt more flexible conformations. Overall, it is very likely that different mycolates adopt different conformation in the host cell membrane, canceling their effect or leading to only minor changes in the membrane stiffness, which remained undetected. It will be tempting to investigate whether the various conformationally constrained MAs partition into specific host lipid domains, thereby locally changing the membrane stiffness. This could be rendered by monitoring the colocalization of labeled MA (of specific chemical type) within phase-segregated model lipid membranes harboring ordered and disordered regions visualized separately using specific lipid probes with selective affinity to such domains.

As the elasticity of membranes is also governed by the underlying actin cytoskeleton, we next investigated actin-free membrane tethers. Tethers are thin membrane nanotubes that aid in cellular adhesion and communication. While, tether force is associated with membrane stiffness, tether length and number indicate membrane resistance to bending.^32,33 MA increased the median tether force (59 pN) compared to the control (55 pN), Figure 2C. This is in contrast to other mycobacterial lipids, specifically those associated with virulence, such as TDM, PDIM and lipoarabinomannn (LAM-v).¹⁵ PGL1 on the other hand decreased the tether force (51 pN, Figure 2C) and again implicated the role of terminal trisaccharide in PGL1 in dictating an opposite behavior to PDIM.¹⁵ PGL1-induced reduced tether force suggests attenuated force requirement for tether extension, reduced surface tension, and bending rigidity. MA did not affect the tether number per cell or tether length, but PGL-1 decreased both.

Tether formation involves induction of curvature in membranes and hence signals a plausible role of the exogenous lipids in its formation. Tether formation involves generation of positively and negatively curved membranes^34,35 at the tether circumference and base, respectively. The data thus indicate that PGL1 insertion into host cell membranes may hinder the formation of regions with high positive curvature, i.e., at the tether circumference after insertion and local enrichment at this site. This is attributed to the plausible conical shape of PGL1, as already shown for the parent PDIM molecule.⁶ This is expected to decrease the frequency of membrane nanotube generation along with their mean length, despite softening of the membrane, thus underlying the experimental observations (Figure 2B−E). Taken together, the insertion of exogenous MA into the host membrane did not change host membrane nanomechanics, while PGL-1 insertion slightly enhanced the softness of the membrane. This enhancement in membrane softness is contrary to that observed with other virulent mycobacterial lipids and could elicit an efficient immune response. This could be achieved via facile immune cell adhesion, migration, and activation by optimal tuning of plasma membrane mechanical properties.²¹ A softer membrane platform could also favorably impact receptor clustering, diffusion, and orientation to evoke an efficient immune response; however, it merits further investigations.

PGL1 Increases the Plasma Membrane Fluidity, Attenuates Ordering, and Alters the Lateral Organization

To gain more insights into the host membrane modulation by MA and PGL1, we evaluated the membrane heterogeneity in terms of bilayer hydration/order using Laurdan, a fluorescent solvochromic probe. Laurdan FLIM rendered visualization of lifetime-segregated membrane lateral heterogeneity/domains (Figures 3 and S1) on live THP-1 membrane surface. The low-lifetime regions likely represent the disordered domains, and the high-lifetime regions represent ordered domains³⁶ on the THP-1 membranes. Upon the exogenous addition of MA and PGL1, the lifetime distribution was modified. For quantitative comparison, fluorescence lifetime distributions (Figure 3, right) were deconvoluted using log−normal fitting, furnishing the lifetimes of the two identified populations; 2.2 ± 0.02 ns (low lifetime) and 5.5 ± 0.01 ns (high lifetime) in control cells.

Figure 3.

Fluorescence lifetime images of the THP-1 cell plasma membrane in the absence and presence of indicated mycobacterial lipids along with associated histograms (no of cells = 42(control), 35 (MA) and 32 (PGL-1), N = 2.) Scale bar 10 μm.

A decrease in the lifetime of the high-lifetime population with PGL1 (5.1 ± 0.009 ns) but not MA (5.4 ± 0.02 ns) indicates increased fluidity of the THP-1 membranes aligning with the elasticity data. The same was also evident from a significant decrease in the average fluorescence lifetime of Laurdan in PGL1-treated cells obtained by fitting the average lifetime decay (Table 1). Interestingly, PGL1 decreased the order and fastened the relaxation of Laurdan in the ordered regions by a larger magnitude (Table 1) compared with that by the low-lifetime component (reflective of disordered regions).

Table 1. Fluorescence Lifetimes of Laurdan in THP-1 Cells with and without Exogenous Lipid Treatment at Room Temperature Calculated from FLIM; n = 42 (Control), 35 (MA), and 32 (PGL-1).

	τ1 (ns)	α 1	τ2 (ns)	α 2	⟨τ⟩ (ns)
control	6.0 ± 0.14	0.72	2.2 ± 0.15	0.28	5.52
MA	6.3 ± 0.16	0.61	2.8 ± 0.10	0.39	5.53
PGL1	5.8 ± 0.07	0.68	2.1 ± 0.06	0.32	5.26

This could be attributed to the plausible interaction site of PGL1 on the host PM, i.e., ordered raft-like regions housing its cognate receptor. Furthermore, an increased PM fluidity could enhance receptor diffusion and clustering as well as modify the receptor’s conformations essential for efficient receptor−ligand binding and downstream signaling transduction amounting to a stronger immune activation. These aspects need further verification but could serve as an important framework for correlating the membrane effect of exogenous bacterial immunomodulatory lipids and the immune response.

Next, to orthogonally explore the lateral membrane organization of THP-1 membranes, we performed spectral imaging using Laurdan. The solvochromic nature of Laurdan allows for a red-shifted emission spectrum in loosely packed disordered membranes due to enhanced polarity near the probe. This spectral shift is quantified by the ratio metric parameter, GP, and the reports on the membrane packing or fluidity. Higher values indicate a more tightly packed membrane.³⁷ Pseudocolored GP images demonstrated a heterogeneous membrane lipid order distribution in control cells for both global, Figure 4A, and only plasma membrane regions, Figure 4B. A few irregularly distributed high-GP areas (colored red to orange) were also evident. Plasma membrane GP histograms (segmented using published reports)²⁵ showed a trimodal Gaussian distribution (for control cells) centered at distinct GP values (C1 = −0.09, C2 = +0.32, and C3 = 0.69, Table 2, Figures 4C), indicative of fluid, coexisting disordered and ordered and highly ordered membrane regions. Upon treatment, both the global (Figures 4A and S2) and plasma membrane regions (Figure 4B) demonstrated altered lateral organization with PGL1 but not with MA. The mean GP values for the distinct GP populations increased, though minor, for MA (C1 = 0.02; C2 = 0.40; C3 = 0.72) and decreased for PGL1 (C1 = −0.12; C2 = 0.33; C3 = 0.63). These indicate that PGL1 incorporation causes the disordered and fluid regions to become more fluid. The surface coverage of high-GP domains (highly ordered) remained intact but increased for the populations reflecting fluid and disordered−ordered phase coexistence.

Figure 4. Mycobacterial PGL1 regulates the host cell membrane order and hydration.

(A) Representative GP images of THP-1 cells in the absence and presence of the indicated lipids (4 h, 10 μg/mL) (B) Segmented images derived from (A) showing only the plasma membrane regions. (C) GP distribution associated with the stack of GP images (n = 45, N = 2) fitted to Gaussian function. It shows peak centroids for each identified population. The surface coverage of each of these populations (area under the curve) is indicated as shown. Scale bar: 10 μm, 60× oil objective.

Table 2. Data from GP Imaging Detailing Peak Centroids (C1, C2, and C3) Representing Three Different Domains and the Area Occupied by Each of These for Treated and Control Cells.

	C1	% C1	C2	% C2	C3	% C3
control	−0.09	15.0	0.32	48.8	0.69	36.2
MA	0.02	15.4	0.40	59.3	0.72	25.3
PGL1	−0.12	11.3	0.33	56.3	0.63	32.4

Next, we decided to monitor the impact of membrane remodeling on lipid diffusion but found no change in the lateral lipid diffusion (Figure S3), implying that either membrane reorganization does not impact in-plane lipid diffusion substantially or the effect is within the detection limit of the experimental design. It would be interesting to monitor the diffusion dynamics within specific domains, i.e., the three distinct GP populations, and correlate to other mycobacterial lipids explored so far.

Conclusions

In this work, we evaluated the membrane-centric host interaction profile of two mycobacterial lipids, MA and PGL1. Both these lipids, particularly the MAs, are strong candidates for combinatorial vaccine adjuvant formulations, and hence investigating their effect upon contact with the host cell membrane is of utmost relevance. This is mostly controlled by the fact that the plasma membrane uses its biophysical characteristics, such as fluidity, stiffness, packing, curvature, and lateral domain organization, to regulate a variety of functions. Lipid/protein transport, localization, lipid−protein interactions, and ultimately their activity are all affected by changes in these characteristics. As a result, it imposes exquisite control on regulating cellular processes and hence are critical factors to be considered while deciding the choice of lipid-based vaccine adjuvants. Using various biophysical techniques, we monitored the changes in the plasma membranes of live THP-1 cells upon exposure to mycobacterial MA and phenolic glycolipid. Particularly, we observed that while MA did not modulate the membrane properties substantially and phenolic glycolipid 1 softened the plasma membrane, reducing the membrane tension and rigidity. This was further supported by observing a fluidizing effect of PGL1 on the host plasma membrane, wherein the fluidity was enhanced and the abundance of fluid/ordered−disordered coexisting phase regions was increased. Notably, this was not correlated with lipid diffusion. Based on these findings and previous reports on virulent mycobacterial lipids, it can be stated that both MA and PGL1 may serve as rational choices for either vaccine adjuvant formulations to boost the desired immune response during treatment and/or as a part of delivery platforms consisting of immunomodulatory lipids due to their minimal perturbation to host cell membranes.