Untargeted lipidomic differences between clinical strains of methicillin-sensitive and methicillin-resistant Staphylococcus aureus

Abstract

Background

Staphylococcus aureus is a common cause of infectious diseases in humans. It has become resistant to many antibacterial agents making management of infections difficult. A better understanding of differences among S. aureus strains that are sensitive and resistant to antibiotics may offer insights into the resistant phenotype and identify new antimicrobial targets. This study aimed at comparing general differences in lipid profiles among clinical strains of S. aureus sensitive and resistant to antibiotics. The cell wall thickness and cell surface charge were also compared.

Methods

Five methicillin sensitive (MSSA) and five methicillin resistant (MRSA) S. aureus strains were compared both individually and as MSSA and MRSA groups in the absence of antibiotics. Lipids were compared by ultra-performance liquid chromatography-mass spectrometry, cell wall thickness was compared by scanning transmission electron microscopy and whole-cell surface charge was compared using a cytochrome c binding assay.

Results

Twenty-two lipid species were identified in all ten strains of S. aureus. The abundance of three lipid species (two lysyl-phosphatidylglycerol and one diglycosyldiacylglycerol) were found to be different between MSSA and MRSA. Differences in cell wall thickness were identified between strains but not between MSSA and MRSA. No difference in whole-cell surface charge was observed between MSSA and MRSA.

Conclusion

This study shows differences in membrane lipids between antibiotic sensitive and antibiotic resistant clinical strains of S aureus that may affect resistance mechanisms related to cell membrane structure and fluidity. Further research on these differences may identify new drug targets against resistant strains.

Keywords:

• Staphylococcus aureus
• infectious disease
• antibiotic resistance
• MRSA
• lipidomics
• cell wall
• cell membrane

Introduction

Staphylococcus aureus is a Gram-positive cocci bacterium that can be found as a commensal on human skin and nares in approximately 30% of the population. However, it can also cause a variety of infections including impetigo, sepsis and pneumonia [1]. These infections can be more dangerous when caused by antibiotic resistant strains such as methicillin resistant S. aureus (MRSA) as they are resistant to many common antibiotic therapies thereby complicating treatment. It has been reported that as many as 60% of clinically isolated strains of S. aureus are resistant to methicillin [2]. Resistance to other antibiotics including penicillin and last resort antibiotics such as vancomycin and daptomycin have also been reported. Increasing resistance to various antibiotics in MRSA is limiting the therapeutic options and necessitating researchers to look for new targets for effective antimicrobial treatments.

Antibiotic resistance in S aureus has been attributed to many factors. The main difference between antibiotic resistant and sensitive strains is the presence of mecA gene in the resistant strains which provides resistance to methicillin. Other differentiating factors between antibiotic resistant and sensitive strains are related to proteins that are responsible for virulence and pathogenicity [3–6]. While the role of genes and proteins in antibiotic resistance of S. aureus are well characterised, differences in other cellular components such as cell membrane and cell wall between antibiotic resistant and sensitive strains are less well understood.

The bacterial cell membrane is responsible for regulation of passage of substances into the cell, generation of ATP, and cell-to-cell communications [7]. The cell membrane of S. aureus is made up of a variety of phospholipids and glycolipids including phosphatidylglycerols (PG), lysyl-phosphatidylglycerols (Lys-PG) monoglycosyldiacylglycerols (MGDG), diglycosyldiacylglycerols (DGDG), and cardiolipins (CL). The makeup of the membrane can change in response to external stimuli such as antibiotics, temperature, and pH [8].

Past studies have shown that antibiotic resistance in S aureus can be related to differences in lipid content. Different lipid species may play different roles in antibiotic resistance. For example, PG’s are negatively charged while Lys-PG’s are positively charged, and their altered abundances can make the cell membrane more positively or negatively charged affecting resistance (such as in bacteria resistant to cationic antimicrobial peptides (CAMPs)) [9]. MGDG’s are non-bilayer forming and DGDG’s are bilayer forming and therefore are responsible for the stability and fluidity of the membrane and may change the permeability of the cell to antibiotics [10]. CL’s have been shown to stabilise liposomes against osmotic stress and to contribute to resistance to daptomycin by preventing membrane permeabilization [8,11]. Lipidomic differences observed in a previous study between clinical isolates of methicillin sensitive S. aureus (MSSA) and MRSA indicate that antibiotic susceptibility may be associated with the lipid composition of the bacterial cell membrane [12]. There is still limited knowledge on the lipidomic differences between antibiotic sensitive and resistant strains of S. aureus.

Some studies show that antibiotic resistance in S aureus can also be associated with changes in the cell wall. The bacterial cell wall is responsible for maintaining cell shape and it consists of a thick layer of peptidoglycan with lipopolysaccharides and wall teichoic acids in Gram positive bacteria [2]. Thickening of the cell wall in S. aureus has been associated with reduced susceptibility to vancomycin and daptomycin [13,14]. It has also been associated with adaptive resistance to amikacin in MRSA [15]. Alterations to the cell wall can therefore be important in antibiotic resistant strains of S. aureus.

In order to understand MSSA and MRSA better, many past studies have shown general differences in virulence factors, protein profiles, biofilm formation and clinical outcomes between MRSA and MSSA strains [5,6,16,17], but there is paucity of data on differences in the lipid profiles of methicillin sensitive and resistant strains. Our study therefore characterised and compared lipid profiles of clinical strains of MSSA and MRSA to better understand general differences between them. Our study was guided by many other reports in literature comparing non-isogenic strains of MRSA and MSSA isolated from patients and their samples. The lipid profiles were taken in the absence of antibiotic stress so that baseline measurements could be taken. In addition to cell membrane lipids, our study also compared cell wall thickness and cell surface charge between methicillin sensitive and resistant strains of S aureus. Successful identification of such differences between MSSA and MRSA strains may help in exploring potential drug targets as well as the ability to better tailor antimicrobial therapy based on cellular characteristics.

Materials and methods

Bacterial isolates and growth conditions

Ten clinical strains of S aureus (five MSSA and five MRSA) (Table 1) were obtained from Dr Iain Gosbell (Liverpool Hospital, NSW, Australia) whose lab had characterised the strains including typing and antibiotic resistance. The strains are described in literature [18,19] and used in our previous study [20]. Each strain was grown overnight on lysogeny broth (LB) agar plates at 35 ± 2 °C before each experiment. 2-3 colonies were suspended into Mueller-Hinton Broth (MHB) to OD₅₉₅ 0.09 (≈1 × 10⁸ colony forming unit/mL (CFU/mL) matching a 0.5 McFarland turbidity standard). These cells were then diluted to 1 × 10⁷ CFU/mL and appropriate volumes were used to have 0.5 × 10⁵ CFU/mL in the experiments.

Table 1. Characteristics of methicillin-sensitive and methicillin-resistant strains of Staphylococcus aureus used in the experiments.

Strain	Name	Type (HA/CA)^a	Origin	Resistance pattern
MSSA 1	MSSA (1)	CA	Australia	Typically no additional resistances
MSSA 2	MSSA (2)	CA	Australia	Lincosamides, Fusidic acid
MSSA 4	MSSA (4)	CA	Germany	Typically no additional resistances
MSSA 6	MSSA (6)	CA	Australia	Typically no additional resistances
MSSA 8	MSSA (8)	CA	Vietnam	Typically no additional resistances
MRSA 1	QLD1	CA	Australia	Typically no additional resistances (susceptible to non-β-lactam classes of antimicrobial)
MRSA 2	Aus-2	HA	Australia	Macrolides/Lincosamides Tetracyclines, Trimethoprim-sulfamethoxazole, Gentamicin, Fluoroquinolones
MRSA 8	USA300	CA	USA	Macrolides/Lincosamides Fluoroquinolones (variable)
MRSA 9	SWP	CA	New Zealand	Typically no additional resistances (susceptible to non-β-lactam classes of antimicrobial)
MRSA 13	WA-1	CA	Australia	Typically no additional resistances Macrolides/lincosamides variable Fusidic acid variable

^aHospital Associated/Community Associated.

Lipid analysis

Lipid extraction

Whole lipid extracts were obtained using a Bligh and Dyer method modified from Hewelt-Belka et al. [12]. In brief, the OD adjusted cells suspended in MHB were incubated on a shaker (20 h, 35 °C, 100 rpm), pelleted and suspended in the UPLC grade solvent mixture (chloroform-methanol 1:2 v/v). The suspension was incubated on a vortex mixer (30 min, room temperature, 1000 rpm) to disrupt cells. Chloroform and Milli-Q water (1:1.5 v/v) were added and the solution centrifuged (9,000 rcf, 10 min) to separate into layers. The lower organic phase containing the membrane lipids was transferred to a glass chromatography vial and stored at −20 °C prior to analysis by UPLC-MS.

Ultra-Performance liquid chromatography – mass spectrometry for lipidomics

The whole lipid extracts were analysed by UPLC-MS using a method modified from Damen et al. [21]. UPLC was conducted using a Waters Acquity UPLC System fitted with a binary solvent manager and mass spectrometric detection was conducted using a Waters Xevo QToF-MS quadrupole time of flight mass spectrometer, fitted with an electrospray source. Separation consisted of two mobile phases. Mobile Phase A (10 mM ammonium acetate with 0.1% formic acid in Milli-Q water-acetonitrile (ACN) (40:60, v/v)), and Mobile Phase B (10 mM ammonium acetate with 0.1% formic acid in ACN-isopropanol (IPA) (10:90, v/v)). The weak wash consisted of ACN-Milli-Q water-IPA (30:30:40, v/v/v) and the strong wash consisted of IPA and 0.1% formic acid. Elution was applied at a flow rate of 0.2 mL/min through an Acquity UPLC HSS C18 column (2.1 mm x 150 mm) thermostatted to 45 °C with each sample run taking 20 min. The gradient was 0 min 40% B; 2 min; 43% B; 2.1 min 50% B; 11 min 54% B; 11.1 min 70% B; 17 min 99% B; 17.1 min 40% B and 20 min 40% B. Samples were diluted two-fold in IPA-ACN-Milli-Q water (2:1:1) and 5 µL was injected for each run. Data was obtained in technical triplicate.

Mass spectrometry was conducted with a capillary voltage of ± 0.6 kV and a sampling cone voltage of ± 30 V for both positive and negative ion mode electrospray ionizations. A desolvation nitrogen gas flow of 700 l/h and a desolvation temperature of 450 °C were used. Data acquisition was conducted over the mass to charge range of 50–1,200. Mass accuracy was obtained using a lockspray solution consisting of 200 ng/mL leucine encephalin in 50% aqueous ACN with 0.1% formic acid. The solution was infused at 5 µL/min and the lockmass correction updated every 30 s.

Lipid identification and analysis

Lipid species and their relative abundances were identified by a targeted lipid search using ion adduct masses on the total ion chromatograms (TICs) generated by the above UPLC-MS. This was done in Waters MassLynx v4.1. Ion adduct masses were calculated from a list of neutral masses of membrane phospholipids and glycolipids from the classes diacylglycerol (DAG), DGDG, MGDG, PG, CL and Lys-PG acquired from Hewelt-Belka et al. [22]. Protonated, ammoniated, sodiated and potassiated ion adducts were calculated by adding the mass of a hydrogen ion, ammonium ion, sodium ion and potassium ion respectively. After putative identification of a lipid species, its relative abundance was calculated from the area under the curve of the extracted ion chromatographic peak. Comparisons were made between the MSSA group and MRSA group by a Welch two sample t-test in R at a significance of p ≤ .05.

Determination of cell wall thickness

Cell wall thickness was measured using a method modified from Kawai et al. [23]. The cells were washed three times with 0.1 M Na-phosphate buffer (pH 7.4) and incubated (overnight, 4 °C) in the primary fixative (2.5% glutaraldehyde in 0.1 M Na-phosphate buffer). Cells were washed three times with Na-phosphate buffer and incubated (room temperature, 2 h) in the secondary fixative (1% osmium tetroxide). The cells were washed three times with Milli-Q water and dehydrated in an ethanol series (30%, 50%, 70%, 95%, and 100%) with the 100% ethanol step repeated three times. Cells were then dehydrated again with 100% acetone twice, pelleted and a loopful streaked directly into resin (10 g ERL 4221, 6 g Der 732, 26 g nonenyl succinic anhydride, 0.4 g dimethylaminoethanol) in a vertical plastic micromold. The resin was placed in an oven to polymerise and harden (overnight, 60 °C).

The resin blocks were sectioned (Leica EM UC7 ultramicrotome with a glass blade) into ultrathin slices (90 nm) and mounted on a 200-mesh copper grid.

Cells were examined by STEM (Zeiss Merlin FE-SEM) at 30 kV with an aperture size of 20 µm and at a working distance of 2.7 mm. The cell wall of 20 cells were imaged per strain and the cell wall thickness was measured at four points using ImageJ. A one-way ANOVA was conducted in R at a significance of p ≤ .05 to compare the average cell wall thickness of the MSSA group and MRSA group and a pairwise comparison of means was conducted in R at a significance of p ≤ .05 to compare the cell wall thickness of each strain.

Analysis of Whole-Cell surface charge by cytochrome C assay

Whole cell surface charge was determined using a cytochrome c biding assay based on Matsuo et al. [24] and Meehl et al. [25]. In brief, cells were incubated overnight in MHB on a shaking incubator (100 rpm, 35 °C) to a final concentration of ≈10⁹ CFU/mL. The cells were pelleted and washed twice in 20 mM morpholinepropanesulfonic acid (MOPS), pH 7, resuspended in 0.25 mg/mL cytochrome c, in MOPS and incubated in the dark at room temperature for 10 min. The cells were centrifuged and the amount of cytochrome c in the supernatant determined at OD₅₃₀ using a standard curve. Three independent experiments were run for all ten strains. The percent bound cytochrome c was compared between the MSSA group and MRSA group using a Welch Two Sample t-test conducted in R at a significance of p ≤ .05 and between each strain using a one-way ANOVA at a significance of p ≤ .05.

Results

Lipidomics

In this study, 22 lipid species were putatively identified in all ten strains of S. aureus and included members of the classes DAG, DGDG, PG, and Lys-PG. The relative abundance of each lipid was calculated by the area under the curve of the extracted ion chromatographic peak and from this the average abundance in the MSSA group and the MRSA group determined (Table 2).

Table 2. Lipid species identified in all ten S. aureus strains.

Lipid species	Ion adduct	Theoretical mass	Observed mass	Average relative abundance in MSSA Group	Average relative abundance in MRSA Group	p Value
DAG (30:0)	[M + K]+	579.439	579.453	4235	4056	.7633
DAG (32:0)	[M + K]+	607.470	607.482	2801	2745	.8721
DAG (34:0)	[M + NH4]+	614.572	614.574	1870	2073	.613
DAG (36:0)	[M + NH4]+	642.604	642.59	1836	1670	.6055
DGDG (30:0)	[M + NH4]+	882.615	882.603	1073	862	.0997
DGDG (31:0)	[M + Na]+	901.586	901.57	1155	1395	.3094
DGDG (32:0)	[M + NH4]+	910.647	910.609	3219	2856	.3868
DGDG (33:0)	[M + NH4]+	924.662	924.632	1280	1076	.2757
DGDG (34:0)	[M + NH4]+	938.678	938.627	2220	1838	.003428
DGDG (35:0)	[M + NH4]+	952.694	952.648	1340	970	.1043
Lys-PG (18:0)	[M + Na]+	677.375	677.359	870	879	.9863
Lys-PG (25:0)	[M + H]+	753.503	753.505	743	741	.9939
Lys-PG (27:0)	[M + H]+	781.534	781.540	1243	1264	.8847
Lys-PG (29:0)	[M + H]+	809.566	809.56	160	178	.5489
Lys-PG (32:0)	[M + H]+	851.613	851.611	1914	1122	.000003282
Lys-PG (33:0)	[M + NH4]+	882.655	882.603	1073	1176	.5925
Lys-PG (35:0)	[M + NH4]+	910.686	910.609	3520	3745	.3636
Lys-PG (36:0)	[M + Na]+	929.657	929.627	740	608	.0912
Lys-PG (37:0)	[M + Na]+	943.673	943.620	1160	983	.1118
Lys-PG (38:0)	[M + Na]+	957.688	957.658	810	634	.0041
PG (32:0)	[M + NH4]+	740.544	740.524	2278	2149	.5967
PG (35:0)	[M + NH4]+	782.591	782.567	367	370	.9398

I dentified ion adduct, theoretical mass, observed mass, average abundance in MSSA group, average abundance in MRSA group and calculated p-value from a t-test looking for significant difference between the MSSA group and MRSA group.

A Welch Two Sample t-test in R at a significance of p ≤ .05 looking for a statistically significant difference between the average abundances of each lipid species in the MSSA group and MRSA group identified three lipid species present at significantly different abundances between the MSSA group and MRSA group. The three lipid species were DGDG (34:0) (p = .003428), Lys-PG (32:0) (p = .000003282) and Lys-PG (38:0) (p = .0041). All three lipid species were present at higher abundances in the MSSA group (Table 2).

Cell wall thickness

The sectioned cells of each strain appeared healthy in the electron micrographs (Figure 1). The average cell wall thickness was determined from 20 individual cells measured at four points for each strain and ranged from 20.185 nm to 31.546 nm. (Figure 2(A)). The cell walls in order from thinnest to thickest were MSSA 2, MRSA 8, MSSA 4, MRSA 1, MSSA 1, MRSA 13, MSSA 8, MRSA 2, MRSA 9 and MSSA 6. No clear pattern in cell wall thickness and antibiotic susceptibility is evident from this order (Figure 2(A)).

Figure 1. Scanning transmission electron micrographs of S. aureus cells embedded in resin and sectioned on an ultramicrotome. Cells imaged using a Zeiss Merlin FE-SEM at 30 kV and a working distance of 2.7 mm. (A) MSSA 1. (B) MSSA 2. (C) MSSA 4. (D) MSSA 6. (E) MSSA 8. (F) MRSA 1. (G) MRSA 2. (H) MRSA 8. (I) MRSA 9. (J) MRSA 13.

Figure 2. Average cell wall thickness S. aureus cells with standard deviation of the mean error bars. (A) Cell wall thickness of all ten strains. (B) Cell wall thickness of the MSSA group and MRSA group.

A one-way ANOVA conducted in R at a significance of p ≤ .05 and the calculated p-value of 2.24 × 10⁻¹⁰ indicated there was a significant difference between strains. A pairwise comparison of means conducted in R determined how strains differed from each other. MSSA 2 was thinner than MRSA 2, MRSA 9, MSSA 6 and MSSA 8. MSSA 6 was thicker than MSSA 2, MSSA 4, MRSA 1, MRSA 8 and MRSA 13. MRSA 9 was thicker than MRSA 8. All significant strain pairs had a p-value < .01 except for MRSA 9/MRSA 8 which had a p-value of 0.0480 (Table 3).

Table 3. S. aureus strain pairs with significant difference in cell wall thickness.

Strain Pair	Strain with Thicker Cell Wall	p Value
MSSA 2 - MRSA 2	MRSA 2	<.01
MSSA 2 - MRSA 9	MRSA 9	<.01
MSSA 6 - MSSA 1	MSSA 6	<.01
MSSA 6 - MSSA 2	MSSA 6	<.01
MSSA 6 - MSSA 4	MSSA 6	<.01
MSSA 6 - MRSA 1	MSSA 6	<.01
MSSA 6 - MRSA 8	MSSA 6	<.01
MSSA 6 - MRSA 13	MSSA 6	<.01
MSSA 8 - MSSA 2	MSSA 8	<.01
MRSA 9 - MRSA 8	MRSA 9	.0480

The strain possessing the thicker cell wall is indicated and respective p-values as identified from ANOVA test looking for significant difference between strains cell wall thickness are also presented.

In addition to comparing the average cell wall thickness of each strain individually, the average cell wall thickness of the MSSA group and MRSA group were also compared (Figure 2(B)). A Welch Two Sample t-test was conducted in R at a significance of p ≤ .05 looking for a statistically significant difference between the averages of the resistant strains and the sensitive strains. The calculated p-value for this test was 0.9502 and indicated no significant differences in cell wall thickness between the MSSA group and MRSA group.

Whole-Cell surface charge

The average percent bound cytochrome c from three independent cytochrome assay experimental runs were used to compare the whole-cell surface charge of each strain. The average percent bound cytochrome c ranged from 28.231% to 40.460%. The percent bound cytochrome c in each strain from least to most were MRSA 1, MSSA 4, MSSA 2, MSSA 8, MSSA 6, MRSA 2, MSSA 1, MRSA 13, MRSA 9 and MRSA 8. From this it seems that the MSSA strains were less negatively charged compared to the MRSA strains (Figure 3(A)). A one-way ANOVA conducted in R at a significance of p ≤ .05 and the determined p-value of 0.875 indicated there was no significant difference between the whole-cell surface charge of each strain and therefore there was no need to conduct a pairwise comparison of means.

Figure 3. Average percent binding of cytochrome c to S. aureus cells with standard deviation of the mean error bars. (A) Percent binding of cytochrome c to all ten strains. (B) Percent binding of cytochrome c to the MSSA group and MRSA group.

The average percent bound cytochrome c in the MSSA group and MRSA group was determined from the three independent cytochrome assay experimental runs (Figure 3(B)). A Welch Two Sample t-test conducted in R at a significance of p ≤ .05 looking for a statistically significant difference between the average percent bound cytochrome c in each group and the calculated p-value of 0.3437 indicated there was no significant difference in the whole-cell surface charge between the MSSA group and MRSA group.

Discussion

As a common cause of infectious disease in hospitals and the community, S. aureus and its resistance to antibiotics is of considerable concern. While the role of genes in methicillin resistance in MRSA is well described, the roles of other cell components are not as well characterised in this context. This study exploring the lipid profile, cell wall thickness and whole-cell charge of S aureus clinical strains, identified differences in two cell membrane lipids, Lys-PG and DGDG, between antibiotic sensitive and antibiotic resistant S aureus strains but did not observe differences in cell wall thickness and surface charge between the two groups.

The bacterial cell membrane acts as a selectively permeable barrier, maintaining appropriate concentrations of nutrients and waste inside the cell as well as protecting the cell from the entry of harmful substances such as antibiotics. The cell membrane of S. aureus is complex and contains a variety of straight chain and branched chain fatty acids. These fatty acids can be odd or even numbered and the location of the branch varies. This branching diversity and the existence of odd and even numbered fatty acids can lead to isobaric overlap between fatty acids. This overlap can make it difficult to identify individual lipid species during mass spectrometry experiments [26]. Additional complexity in the S. aureus cell membrane arises through the fatty acids combining into the lipid species PG, Lys-PG, MGDG, DGDG and CL. The abundance of these lipids in the membrane has been shown to vary in antibiotic strains [27]. This study, looking at differences in the lipid species in S. aureus strains using chromatographic peaks and parent ion mass is in support of the hypothesis that there are lipids that are different in abundance between antibiotic sensitive and resistant bacteria. Some previous studies have also shown differences in lipid content between antibiotic sensitive and antibiotic resistant strains of S aureus. For example, a higher level of PG has been reported for methicillin-susceptible strains in comparison with methicillin resistant strains of S. aureus [28]. Ratio of PG to Lys-PG has been associated with resistance to antibiotic fusidic acid in S. aureus [29]. Differences in PG and Lys-PG contents have also been reported in S. aureus strains susceptible and resistant to daptomycin, gentamycin and vancomycin [30,31].

This study identified three lipid species at significantly different abundances between antibiotic sensitive and antibiotic resistant strains of S aureus. These species were a DGDG (DGDG (34:0)) and two Lys-PG’s (Lys-PG (32:0) and Lys-PG (38:0)), all three present at higher abundance in MSSA compared to MRSA. Lys-PG and DGDG are known to be constituents of the cell membrane. The observed differences in lipid contents of Lys-PG and DGDG between antibiotic sensitive and antibiotic resistant S. aureus strains may offer insights into the resistance mechanisms that are related to the cell membrane structure and its fluidity. Lys-PG is an important cell membrane lipid in S aureus and changes in its amount are correlated with change in membrane fluidity [32]. It is also notable that lipids in the class Lys-PG contribute a positive charge to the membrane to balance the negative charge contributed by the cell wall and negatively charged membrane lipid PG [9]. A higher abundance of Lys-PG with no apparent change in the negatively charged lipids (such as PG) may indicate that the membrane of the sensitive strains may be positively charged with partial neutralisation of the cell surface than the resistant strains. An increase in Lys-PG is also associated with resistance to positively charged antibacterial compounds such as CAMPs as well as the antibiotic daptomycin [9]. Therefore, the tested MRSA strains may be more susceptible to CAMPs and daptomycin than the MSSA strains. Further testing on this would help in determining the effectiveness of available antibiotics against these strains. Lipids in the class DGDG, along with MGDG lipids play a role in maintaining the fluidity and stability of the membrane [10]. In our study, a higher abundance of one DGDG with no change in MGDG shows that the fluidity of the sensitive strains may be different from the resistant strains. This may play a role in affecting the entry of antibiotics into the cell [12]. Further studies expanding on the lipid identification described here and on the fluidity of the membrane will help to confirm this result and determine its effect on the entry of antibiotics.

In addition, the ratio of straight chain and branched chain fatty acids in the S. aureus membrane affects membrane fluidity. Branched chain fatty acids increase membrane fluidity while straight chain fatty acids reduce fluidity [33]. Therefore, characterising the ratio of the straight chain and branched chain fatty acids in S. aureus and how these may differ in antibiotic resistance bacteria can allow for insights into cell membrane properties and may assist in identifying effective antibacterial therapy. This was however beyond the scope of this current project.

Lipid identification in this study was done manually and this may have limited the number of lipids identified making it difficult to conclusively state the effect of differences in lipid abundances on the methicillin resistance. However, this study does show that the lipid extraction process is effective and that differences do exist between the membrane lipids in S. aureus strains. The data can also serve as a baseline for future research using tandem mass spectrometry in a more targeted search of lipids with altered abundances in sensitive and resistant strains of S. aureus. In addition, recent work has identified a methodology to distinguish lipids containing one, two or no branched chain fatty acids [26]. Such a methodology would be useful in our future experiments for further categorising the differences in S. aureus membranes in the context of antibiotic resistance. Whether differences in lipid abundances are influenced by the presence of antibiotics will also need to be investigated in the context of identifying differences between antibiotic sensitive and resistant strains of S. aureus. Further research is also needed to determine if these differences impact the strains’ ability to survive or cause infection. This research, therefore, brings to attention the benefits of a deeper understanding of differences between antibiotic sensitive and resistant strains beyond the differences in genes. Furthermore, characterisation of lipids present and their abundance in clinically relevant strains would allow for more effective tailoring of antibacterial therapy. For example, if a particular strain has a lower amount of positively charged lipids (such as Lys-PG) it may be more susceptible to CAMPs and daptomycin and thus these may be a better treatment method. Past work has also identified that the growth medium used to culture bacteria has an effect on the ratio of branched chain and straight chain fatty acids. This may serve as another area to investigate in our future research [33].

The cell wall of staphylococci is reported to range from 20 nm to 40 nm in thickness [34]. The results presented here coincide with this with no cell wall below 20 nm in thickness and none above 32 nm. Acquisition of methicillin resistance allows S. aureus to continue cell wall synthesis using PBP2A [35]. This alteration in the cell wall synthesis machinery may impact the cell wall itself. Thickening of cell wall is indeed linked with antibiotic resistance in S. aureus [13,14]. In this study, the comparisons of the thickness of the cell walls of MSSA and MRSA strains indicated significant difference between the individual strains, but not between the MSSA group and the MRSA group. The lack of any significant difference may be because the cells were grown in the absence of methicillin resulting in PBP2A not being used, or there being no difference between the cell wall thickness of MSSA and MRSA as groups. Further studies testing the first hypothesis may help in confirming or rejecting the second. Future studies may also involve examining cell wall thickness as a function of resistance after exposing sensitive strains to sublethal concentration of antibiotics until resistance is developed [15].

Due to the presence of teichoic acids in the cell wall, the overall surface charge in Gram-positive bacteria is negative. Bacterial cell surface charge is known to affect susceptibility to positively charged antimicrobial agents [24]. The cytochrome c assay used in this study showed that the negative charge did not vary significantly between the ten strains. The difference in cell surface charge between the MSSA group and MRSA group was also not significant. The lack of a difference may have been due to the lack of antibiotic challenge or that alterations to cell surface charge may not assist in methicillin resistance and so was not developed in MRSA strains. This second hypothesis is supported by methicillin being a non-charged molecule. Future experiments involving a charged antimicrobial can identify if exposure to sublethal concentrations can induce changes in the whole-cell surface charge and affect resistance.

Conclusions

In conclusion, this study shows differences in cell membrane lipids between methicillin sensitive and methicillin resistant strains of S. aureus that may affect the resistance mechanism of bacteria by altering membrane structure and fluidity. These differences highlight the potential for research into characteristics of bacteria beyond known resistance mechanisms. Future work would benefit from characterising these changes in isogenic strains, in the presence of antibacterial compounds and by inducing resistance in sensitive strains and comparing them before and after development of resistance.

Author contributions

PN conducted the experiments, analysed the results and wrote the manuscript. PM conceptualised the project, designed methodology, supervised and administered the project, and helped in the writing and revising the manuscript. DH assisted in the ultra-performance liquid chromatography-mass spectrometry experimental design and execution, and helped revise the manuscript. JW helped in revising the manuscript. All authors read and approved the final manuscript.

Acknowledgements

The authors thank Dr Richard Wuhrer, Western Sydney University, Advanced Materials Characterisation Facility (AMCF), for training on scanning electron microscopy and the use of their facilities; Meena Mikhael, Western Sydney University, Mass Spectrometry Facility for assistance in ultra-performance liquid chromatography; and Tony Romeo, University of Wollongong, for training and use of their ultramicrotome.

Disclosure statement

Author report no conflict of interest.