Impact of PrsA on membrane lipid composition during daptomycin-resistance-mediated β-lactam sensitization in clinical MRSA strains

Abstract

Background

The cyclic anionic lipopeptide daptomycin is used in the treatment of severe infections caused by Gram-positive pathogens, including MRSA. Daptomycin resistance, although rare, often results in treatment failure. Paradoxically, in MRSA, daptomycin resistance is usually accompanied by a concomitant decrease in β-lactam resistance in what is known as the ‘see-saw effect’. This resensitization is extensively used for the treatment of MRSA infections, by combining daptomycin and a β-lactam antibiotic, such as oxacillin.

Objectives

We aimed: (i) to investigate the combined effects of daptomycin and oxacillin on the lipid composition of the cellular membrane of both daptomycin-resistant and -susceptible MRSA strains; and (ii) to assess the involvement of the post-translocational protein PrsA, which plays an important role in oxacillin resistance in MRSA, in membrane lipid composition and remodelling during daptomycin resistance/β-lactam sensitization.

Results

The combination of microbiological and biochemical studies, with fluorescence microscopy using lipid probes, showed that the lipid composition and surface charge of the daptomycin-resistant cells exposed to daptomycin/oxacillin were dependent on antibiotic concentration and directly associated with PrsA, which influenced cardiolipin remodelling/relocation.

Conclusions

Our findings show that PrsA, in addition to its post-transcriptional role in the maturation of PBP 2a, is a key mediator of cell membrane remodelling connected to the see-saw effect and may have a key role in the resensitization of daptomycin-resistant strains to β-lactams, such as oxacillin.

Introduction

Daptomycin, a cyclic anionic lipopeptide antibiotic, is one of the last resorts in the treatment of infections caused by Gram-positive pathogens, such as MRSA.^1,2 Daptomycin has a decapeptide lactone core and an N-terminal decanoyl fatty acid (FA) chain.³ For antimicrobial activity, Ca²⁺ ions are absolutely necessary to confer an overall amphiphilic character to the anionic nature of daptomycin, allowing the molecule to bind to the cell wall and cell membrane.^4–6 The detailed mechanism of action of daptomycin is still debatable. The latest studies have shown that a tripartite complex is formed between the Ca²⁺–daptomycin, the anionic phospholipid phosphatidylglycerol (PG) and undecaprenyl-coupled cell envelope precursors,⁷ but its ability to depolarize the membrane and to increase its permeability is still observed in the presence of palmitoyl lipids and cardiolipins.⁸

Daptomycin resistance is rare (0.1%–0.3% according to a systematic review and meta-analysis),⁹ but treatment failure may occur in a significant percentage of those cases.^10,11 In Staphylococcus aureus, daptomycin resistance is a multifactorial stepwise process, which includes membrane and cell wall adjustments.^12,13

We previously reported that the most prevalent mechanism of daptomycin resistance in a set of clinical MRSA strains derived from patients was through mutations in mprF.¹⁴ MprF is a new class of lipid-biosynthetic enzymes that synthesize lysyl-PG (Lys-PG) and facilitate its translocation.¹⁵ The cytosolic C-terminal domain of MprF uses lysyl-tRNA as a substrate to add lysine residues to negatively charged PG and the positively charged nature of the resulting Lys-PG could be a major contributor to the changes in cell surface charge necessary to repel daptomycin molecules.^13,16 On the other hand, the N-terminal membrane-embedded domain of MprF is necessary for the translocation of Lys-PG from the inner to the outer leaflet of the cytoplasmic membrane.¹⁵ The clinically relevant mutations that we found in daptomycin-resistant strains are those associated with mutations in mprF.^14,17

An interesting feature associated with daptomycin resistance in MRSA is that it is usually accompanied by a concomitant decrease in β-lactam resistance, in a process known as the ‘see-saw effect’.^18,19 This is used in clinical settings for the treatment of MRSA infections, by associating daptomycin and a β-lactam, particularly those targeting PBP 1 and PBP 2, such as oxacillin.^2,18 In this context, we showed that daptomycin induces cell wall changes primarily associated with delocalization of PBP 1 and PBP 2. Concomitantly, we discovered that PrsA, a lipoprotein acting as a post-translocational chaperone, involved in β-lactam resistance was affecting the amounts of PBP 2a in the cell membrane. Furthermore, we found that mutations in mprF impair PrsA chaperone functions, needed for the post-transcriptional maturation of PBP 2a, thus leading to a resensitization to β-lactams.^14,19

The impact of PrsA on membrane lipid composition and remodelling during daptomycin and β-lactam sensitization of daptomycin-resistant clinical MRSA has not, however, been explored. Therefore, the primary purpose of this study was to investigate the combined effects of daptomycin and oxacillin on the lipid composition of the cellular membrane of daptomycin-resistant strains, by comparing them with their corresponding daptomycin-susceptible isogenic clinical pair strains, and to assess how modifications triggered by PrsA could affect daptomycin-mediated β-lactam resensitization in S. aureus.

Materials and methods

Bacterial strains, chemicals and growth conditions

The MRSA clinical strains used in the present study are listed in Table 1. Cells were grown in Mueller–Hinton broth (Sigma–Aldrich, St Louis, MO, USA) supplemented with 50 mg/L calcium, in 15 mL Falcon tubes (Thermo Fisher Scientific, Waltham, MA, USA) containing 3 mL of medium, and in 8-well tissue culture chambers for high-resolution microscopy (Sarstedt AG & Co. KG, Nümbrecht, Germany). The antibiotics used were daptomycin and oxacillin (Sigma–Aldrich), at concentrations adjusted according to the MICs for the parental strains and genetic mutants, determined as indicated in the Supplementary Materials and methods (available as Supplementary data at JAC Online). The CB1634 ΔmprF mutant strain was grown in the presence of 10 mg/L chloramphenicol (Sigma–Aldrich), as selection antibiotic marker as we previously described.¹⁴ The prsA deletion mutant was generated using a pKOR1 strategy. Plasmid pKOR1, an Escherichia coli/S. aureus shuttle vector, permits rapid cloning via lambda recombination and ccdB selection as previously described.²⁰

Table 1.

List of strains used in this study and their MICs of oxacillin and daptomycin

Strain	Description	MIC (mg/L)		Reference
		oxacillin	daptomycin
CB1631	clinical daptomycin-susceptible parent strain	32	0.5	14
CB1631 ΔprsA	CB1631 isogenic mutant with prsA deleted	8	0.5	this study
CB1634	clinical daptomycin-resistant isogenic derivative	0.5	4.0	14
CB1634 ΔmprF	CB1634 isogenic mutant with mprF deleted	32	0.75	14
CB1634 ΔprsA	CB1634 isogenic mutant with prsA deleted	32	0.5	this study

Lipid analysis

Lipids were extracted from the cell pellet by a modified Bligh and Dyer²¹ method, according to Findlay et al.,²² and adjusted to bacterial cells as described previously^23,24 and in the Supplementary Materials and methods. The FA were methylated to FA methyl esters (FAMEs), using the Instant FAME™ procedure from MIDI, Inc. (Newark, DE, USA), and analysed on a gas chromatograph with flame ionization detector (FID) controlled by the Sherlock software package, v.6.2 (MIDI, Inc.), and by gas chromatograph-mass spectrometry, as described in the Supplementary Materials and methods. Results are the average of lipid extractions of at least three independent cultures grown under each tested condition.

Visualization of the cells by fluorescence microscopy

Cells were grown in 8-well tissue culture chambers in the presence of different antibiotics. The following fluorescent dyes and kits were used according to the supplier’s instructions (Invitrogen, Paisley, UK): LipidTOX™ Phospholipidosis and Steatosis Detection Kit; Laurdan; BODIPY^® 581/591 C11; Nonyl Acridine Orange (NAO); and Nile Red. Each fluorescent dye/kit was used for comparison of control, parent and mutant strains under the same conditions of growth and cells were visualized by fluorescence microscopy. Details are described in the Supplementary Materials and methods. Results are the average of the data obtained from at least nine images taken per sample collected from each of three independent cultures made for each tested condition.

Zeta potential and surface charge density (SCD)

After preparing the cells as described in the Supplementary Materials and methods, the zeta potential of the S. aureus cells was determined using the electrophoretic mobility measured at 25°C in a Doppler electrophoretic light scattering analyser (Zetasizer Nano ZS, Malvern Instruments, Malvern, UK) using Zetasizer software 7.10 (Malvern Instruments). The SCD of the cells was determined using the equation σ = ηu/d, where η is the viscosity of the solution, u is the electrophoretic mobility and d is the diffuse layer thickness, as previously described.²² Results are the average of measurements made for three biologically independent experiments.

Statistical analysis

Statistical analysis was performed using Microsoft^® Excel and MINITAB^® 14 from MINITAB Inc. Significant differences between strains under conditions tested were determined by one-way ANOVA or by one-way ANOVA with multiple comparisons using the Dunnett method as detailed in the figure legends. A P value <0.05 was deemed significant.

Results

Oxacillin triggers greater cell membrane lipid changes than daptomycin in daptomycin-resistant MRSA

The presence of chemical substances, such as antibiotics, and FA supplementation is known to trigger changes in the lipid composition of the cellular membrane phospholipids of S. aureus, which may hamper the bactericidal activity of daptomycin, depending on growth state.^25,26 The cellular membrane of S. aureus is mainly composed of branched saturated FA (BSFA), branched mostly at the anteiso position, which means the branch point is located on the ante-penultimate carbon atom.^25,27 Initially, we assessed the changes in lipid composition in CB1631, CB1634 and mutant derivatives CB1631 ΔprsA, CB1634 ΔprsA and CB1634 ΔmprF during growth in Mueller–Hinton broth supplemented with 50 mg/L calcium in the absence of antibiotics (Figure 1). We observed that the isogenic susceptible CB1631 strain and the daptomycin-resistant counterpart CB1634 strain, grown under identical conditions, displayed an increasing percentage of BSFA (anteiso and iso BSFA) over time, reaching 86.0% and 80.6%, respectively, after 5 h, corresponding to the late exponential phase of growth (Figure 1). In the absence of antibiotics, no significant differences were seen between the FA composition of the phospholipids of CB1631 and CB1631 ΔprsA, and between CB1634 and CB1634 ΔprsA (Figure 1b; P > 0.05 by one-way ANOVA). In contrast, statistically significant changes were seen between CB1634 and CB1634 ΔmprF (P < 0.05 by one-way ANOVA).

Figure 1.

Growth and FA composition of phospholipids of S. aureus cells. (a) Growth curves of daptomycin-susceptible CB1631 and daptomycin-resistant CB1634 cells untreated and treated with daptomycin, oxacillin or daptomycin+oxacillin and corresponding mutant strains, CB1631 ΔprsA, CB1634 ΔprsA and CB1634 ΔmprF. (b) FA composition of the phospholipids of the strains at a timepoint of 5 h of cell growth in Mueller–Hinton broth. CB1634 DAP 1, CB1634 cells grown in the presence of 1 mg/L daptomycin; CB1634 OXA 0.5, CB1634 cells grown in the presence of 0.5 mg/L oxacillin; CB1634 DAP 1 OXA 0.5, CB1634 cells grown in the presence of both 1 mg/L daptomycin and 0.5 mg/L oxacillin. Data presented are the average of the results from three independent cultures grown under each tested condition. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; only significant differences are indicated). This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

Next, to assess whether the FA composition of the cell membrane changed when the strains were exposed to daptomycin, oxacillin or to a combination of both agents, cells of CB1634 were grown in 1 mg/L daptomycin, 0.5 mg/L oxacillin or a combination of both. Significant differences were observed in lipid composition between CB1634 cells exposed to antibiotics and non-treated cells (P < 0.05 by one-way ANOVA). We observed that, in the presence of oxacillin, CB1634 cells responded by making changes in the FA content; notably the BSFA to saturated straight FA (SSFA) ratio was reduced 36.5%, when compared with that of untreated cells, resulting in reduced membrane fluidity (Figure 1b). In contrast, when CB1634 cells were grown in the presence of 1 mg/L daptomycin, a BSFA/SSFA ratio (4.8 ratio) similar to that for untreated cells (4.4 ratio) was observed.

Importantly, when CB1634 cells were exposed to both 1 mg/L daptomycin and 0.5 mg/L oxacillin, an intermediate BSFA/SSFA ratio value of 4.6 was observed. When cells were grown in the presence of both antibiotics, they displayed a change in the proportion of anteiso and iso BSFA, without significantly changing the total amount of BSFA (Figure 1b). CB1634 cells displayed an anteiso/iso ratio of 1.7, whilst cells exposed to both daptomycin and oxacillin increased this ratio to 2.0. We found that the main FA contributing to this increase were 15:0 anteiso and 17:0 anteiso (data not shown). Of note, anteiso FA present phase transition temperatures below their iso counterparts, leading to an increase in cellular membrane fluidity.^28,29 Thus, these results indicate that the exposure of CB1634 to both daptomycin and oxacillin antibiotics leads to an increase in membrane fluidity.

The changes in phospholipid composition observed in the presence of antibiotics may require the expenditure of energy, resulting in lower biomass production, when compared with CB1634 cells grown in the absence of antibiotics (Figure 1a).

Changes in lipid composition during daptomycin-resistance-mediated β-lactam oxacillin sensitization were assessed in CB1631, CB1634 and mutant derivatives CB1631 ΔprsA, CB1634 ΔprsA and CB1634 ΔmprF grown in the presence of daptomycin, oxacillin or both agents combined (Figure 2). Cells were harvested during the mid-exponential phase in all assays, mostly to prevent the effect of age on FA composition that may occur as the cultures go from the exponential to stationary growth phase.^29,30

Figure 2.

S. aureus response to daptomycin and oxacillin at the phospholipid level. (a) FA composition of the phospholipids of daptomycin-susceptible strain CB1631 and of its prsA-deleted mutant in the presence of 0.25 mg/L daptomycin (0.25 DAP), 0.25 mg/L oxacillin (0.25 OXA) or both (0.25 DAP 0.25 OXA). (b) FA composition of the phospholipids of daptomycin-resistant strain CB1634 and of its prsA-deleted and mprF-deleted mutants in the presence of 0.25 mg/L daptomycin (0.25 DAP), 0.25 mg/L oxacillin (0.25 OXA) or both (0.25 DAP 0.25 OXA). The data presented are statistically significant as determined by one-way ANOVA, when comparing each cell type in the presence and absence of antibiotics (P < 0.001, except for CB1634, where P = 0.002). Results are the average of lipid extractions from cells of at least three independent cultures grown under each tested condition. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; only significant differences are indicated). This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

In the presence of 0.25 mg/L daptomycin, the cells of CB1631 ΔprsA decreased the percentage of BSFA 2.5%, whilst the cells of CB1631 increased their content 3.3%, when compared with untreated cells (Figure 2a). By comparing the daptomycin-resistant CB1634 with its mutants CB1634 ΔprsA and CB1634 ΔmprF, it was observed that the latter presented membranes with a lower content of SSFA, regardless of the presence or absence of antibiotics (Figure 2b). When strain CB1634 was grown in the presence of 0.25 mg/L oxacillin, a significant change was observed in the SSFA content, which decreased 7.4% compared with untreated cells (P < 0.05). When CB1634 cells were grown in the presence of 0.25 mg/L oxacillin and 0.25 mg/L daptomycin, these FA increased 14.5%. In contrast, when PrsA was deleted in CB1634, an opposite effect in CB1634 ΔprsA cells grown in the presence of both agents was observed, with a significant decrease of 14.7% in the SSFA content (P < 0.05). These results suggest that PrsA in CB1634 has an effect on the SSFA cell membrane content in the presence of daptomycin and oxacillin. In support of these observations, the comparison between CB1631 and the CB1631 ΔprsA mutant revealed that, in the absence of PrsA (e.g. in CB1631 ΔprsA), the phospholipids exhibited an 8% higher content of SSFA and a 9% lower content of BSFA than CB1631, and that, in the presence of both daptomycin and oxacillin, a decrease in the amount of SSFA from approx. 43% to 39% was observed in both strains. Moreover, in the presence of antibiotics, the differences between treated and non-treated cells were statistically significant (P < 0.05 by one-way ANOVA).

In the case of CB1634 ΔmprF, the cells increased 17.8% and 22.5% the anteiso/iso BSFA ratio in the presence of oxacillin and both oxacillin and daptomycin, respectively. Taken together, these results suggest that: (i) both prsA and mprF are associated with changes in the anteiso/iso conformation in the presence of both daptomycin and oxacillin, determining an increase in membrane fluidity as a response to antibiotic exposure; and (ii) PrsA determines an increase in SSFA content in CB1634 ΔprsA, whilst an opposite effect was seen in the CB1631 ΔprsA mutant.

Impact of PrsA on surface charge during daptomycin-resistance-mediated sensitization to β-lactams

Daptomycin is negatively charged, but its activity depends on the presence of positive calcium ions.³¹ Changes in the net surface charge of the cells may affect the affinity of the complex Ca²⁺–daptomycin for negatively charged phospholipids.^7,32 This was traditionally determined indirectly by binding of positively charged cytochrome c to cells.^17,33 However, the cell surface charge can be deduced directly from the electrokinetic or zeta potential, which can be calculated from the electrophoretic mobility of cells.^25,34

The presence of oxacillin, daptomycin and both caused an increase in the magnitude of the negative zeta potential of CB1631 cells up to 1.5-fold, while a 1.3-fold decrease was observed in CB1634, when compared with untreated cells (Figure 3). CB1631 ΔprsA cells presented zeta potential values more negative than CB1631 cells for all the tested conditions except in the presence of both daptomycin and oxacillin. However, the decrease in zeta potential observed between antibiotic-exposed cells and those without antibiotic for each strain was less pronounced in CB1631 ΔprsA cells than in the parent CB1631 cells. In fact, changes in the net surface charge of the CB1631 ΔprsA cells were only statistically significant (P < 0.001) in the presence of oxacillin and both antibiotics, when compared with cells in the drug-free medium.

Figure 3.

Zeta potential of daptomycin-susceptible (CB1631) and daptomycin-resistant (CB1634) S. aureus cells when grown in the absence and in the presence of 0.25 mg/L daptomycin (0.25 DAP), 0.25 mg/L oxacillin (0.25 OXA) or both antibiotics (0.25 DAP 0.25 OXA). Data from three independent cultures are presented as mean±SD. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ***P < 0.001; ns, no significant difference). This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

CB1634 ΔmprF cells presented values 5% to 20% more negative, whilst CB1634 ΔprsA cells presented zeta potential values 10% to 26% less negative, than CB1634 cells in the presence of antibiotics. CB1634 ΔmprF cells are unable to produce MprF and therefore could not decrease the negative character of the phospholipids in the presence of the antibiotics, when compared with CB1634 (Figure 3). In contrast, CB1634 ΔprsA cells were still able to decrease their negative charge, since MprF was still present in this mutant.

Interestingly, when CB1634 ΔprsA was exposed to both daptomycin and oxacillin, a decrease in negative charge was observed, when compared with CB1634 cells. All together, these results indicate that PrsA contributes to the cellular membrane surface charge during daptomycin/oxacillin exposure and that MprF and PrsA may have coordinate effects on the net surface charge.

Modifications in cell membrane surface charges and FA content of daptomycin-resistant MRSA strains are dependent on antibiotic concentration

To further elucidate the effect of daptomycin and oxacillin on S. aureus strains, the SCD was measured during their growth in the presence of increasing initial antibiotic concentration. In general, daptomycin and oxacillin caused a dose-dependent increase in the negative SCD of the tested strains (Figure 4). However, when CB1631 and CB1634 ΔprsA cells were exposed to daptomycin, they displayed less variation (slopes of −0.003 and −0.050 Cmg⁻¹, respectively) than their counterparts CB1631 ΔprsA and CB1634 (−0.060 and −0.111 Cmg⁻¹, respectively). CB1631 ΔprsA presented a slope 20 times larger than that shown by CB1631 and 1.8-fold larger than its daptomycin-resistant counterpart CB1634. The opposite effect was observed with increasing oxacillin concentrations: oxacillin-susceptible CB1634 presented a slope, respectively, 38- to 1.8-fold larger than oxacillin-resistant CB1631 and CB1631 ΔprsA. Additionally, we found that less negative SCD values were mainly observed in cells exposed to the antibiotic to which they are susceptible to, i.e. CB1631 and CB1631 ΔprsA presented lower values when exposed to daptomycin, whilst CB1634 cells were less negative when exposed to oxacillin (Figure 4a and b).

Figure 4.

SCD of S. aureus cells exposed to antibiotics. (a) SCD of daptomycin-resistant CB1634 and daptomycin-susceptible CB1631 strains of S. aureus and their respective prsA-deleted mutant strains when exposed to increasing concentrations of daptomycin. (b) SCD of daptomycin-resistant CB1634 and daptomycin-susceptible CB1631 strains of S. aureus and their respective prsA-deleted mutant strains when exposed to increasing concentrations of oxacillin. (c) SCD of daptomycin-resistant CB1634 and daptomycin-susceptible CB1631 strains of S. aureus and their respective prsA-deleted mutant strains when exposed to increasing concentrations of oxacillin and 0.25 mg/L daptomycin. Data from three independent cultures are presented as mean±SD. Statistical significance was determined by one-way ANOVA and Dunnett’s multiple comparison tests (*P < 0.05; **P < 0.01; ***P < 0.001). DAP, daptomycin; OXA, oxacillin. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

When both 0.25 mg/L daptomycin and increasing oxacillin concentrations were added to the cultures, the SCD values of daptomycin-susceptible strains followed a pattern mostly influenced by oxacillin exposure (1% to 6% difference; Figure 4c). However, CB1634, which displays a ‘see-saw’ phenotype (resistant to daptomycin and susceptible to oxacillin), presented SCD values much closer to those observed in the presence of only daptomycin (1% to 9.5% difference; Figure 4).

Regarding lipids, the amount of anteiso and iso BSFA presented the largest changes in response to increasing antibiotic concentrations. A dose-dependent increase in the anteiso/iso BSFA ratio was observed for all strains with daptomycin, except for CB1631 (Figure 5a; the difference between CB1634 and CB1631 response to daptomycin concentration was statistically significant with P < 0.05, but not between these strains and their respective prsA-deleted mutants). The largest changes were observed with the CB1634 strain, with a 23% increase between the cells exposed to the highest daptomycin concentration and unchallenged cells, while its corresponding CB1634 ΔprsA mutant presented only an ∼9% difference between those cells.

Figure 5.

Influence of antibiotics on the composition of branched FA. (a) Ratio of anteiso/iso branched FA in daptomycin-susceptible CB1631 and daptomycin-resistant CB1634 strains of S. aureus and their respective psrA mutant strains when exposed to increasing concentrations of daptomycin. (b) Ratio of anteiso/iso branched FA in daptomycin-susceptible CB1631 and daptomycin-resistant CB1634 strains of S. aureus and their respective psrA mutant strains when exposed to increasing concentrations of oxacillin. Data from three independent cultures are presented as mean±SD. Statistical significance was determined by one-way ANOVA and Dunnett’s multiple comparison tests (*P < 0.05; **P < 0.01; ***P < 0.001; only significant differences are indicated). DAP, daptomycin; OXA, oxacillin. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

When oxacillin was added, larger changes in CB1631 isolates were observed than in CB1634 isolates: the anteiso/iso ratio increased 29.9% and 68.9% in CB1631 and CB1631 ΔprsA cells, respectively, and 6.2% and 13.6% in, respectively, CB1634 and CB1634 ΔprsA cells (Figure 5b; significant differences between strains indicated in the figure). Additionally, in CB1631, the largest changes in the anteiso/iso ratio resulted mainly from a lower content of iso BSFA, whilst, in CB1634, an increase in anteiso BSFA was observed. Thus, we conclude that PrsA does not have a major impact on the BSFA composition of the cellular membrane phospholipids in CB1634 (Figure 6).

Figure 6.

Summary of the major changes observed in the composition of the FA of the cellular membrane in the different MRSA strains tested. Blue indicates an increase in membrane fluidity, while green indicates an increase in membrane rigidity; colour intensity is a relative indication of the amount of changes observed, with white representing no changes in BSFA or SSFA content and a dark colour indicating large changes. DAP, daptomycin; OXA, oxacillin. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

PrsA is directly associated with lipid domain redistribution

The function, localization and heterogeneity of bacterial membrane-embedded proteins, such as PrsA, PBP 1 and PBP 2, which play an important role in the see-saw effect between β-lactams and daptomycin,^19,35 may be strongly influenced by phospholipid microdomains and thus antibiotic resistance may also be affected.^36,37 To address these questions in the studied strains herein, NAO was used to show the location of the phospholipid cardiolipin. Cardiolipin is an anionic lipid and its headgroup is a major contributor to SCD and to the bilayer electrostatic profile in bacterial membranes, while having a major role in the binding and positioning of proteins in the membrane.³⁸

Both daptomycin-resistant CB1634 and daptomycin-susceptible CB1631 strains, and their corresponding PsrA mutant strains, presented a regular division septum during the mid-exponential phase without antibiotics in the growth medium (Figure 7a–d and a1–d1). However, in the presence of 0.5 mg/L daptomycin, the growing cells remodelled the membrane and lipid microdomains were visible at cell pole/division sites (Figure 7e–h and e1–h1).

Figure 7.

NAO-stained cells. Daptomycin-susceptible CB1631 (a and e) and daptomycin-resistant CB1634 (c and g) isolates and respective prsA-deleted mutants (b and f, d and h) stained with NAO. Close-ups of the cells show cells with a regular division septum (yellow arrows), lipid microdomains at cell poles (white arrows) and lipid microdomains along the membrane (red arrows). Scale bars = 10 μm (a–d, e–h) and 1 μm (a1–d1, e1–h1). (i) Percentage of cells with cardiolipin-enriched microdomains; data are presented as mean±SD. Statistical significance was determined by one-way ANOVA (*P < 0.05; **P < 0.01; ns, no significant difference). CL, cardiolipin. DAP, daptomycin; w/o, without. This figure appears in colour in the online version of JAC and in black and white in the print version of JAC.

In daptomycin-resistant CB1634, the presence of daptomycin caused a significant 9.5-fold increase in the number of mid-exponential growing cells with cardiolipin microdomains at the pole/division septa, while a 4.3-fold increase was observed in CB1634 ΔprsA, when compared with untreated cells (Figure 7i). In contrast, a 1.8-fold increase in the number of cells with cardiolipin microdomains was observed in daptomycin-susceptible CB1631 and a 3.1-fold increase was attained in CB1631 ΔprsA when the cells were exposed to daptomycin, in comparison with unchallenged cells (Figure 7i). These findings strongly demonstrate that PrsA is directly linked to lipid remodelling at the cellular membrane during daptomycin resistance.

PrsA impacts membrane fluidity in daptomycin-resistant strains during exposure to daptomycin/oxacillin

To assess the effects of daptomycin and oxacillin on membrane fluidity, the daptomycin-susceptible, daptomycin-resistant and ΔprsA mutant cells were stained with Laurdan and Nile Red. Laudan is a probe for studying membrane order, that suffers a shift in emission wavelength as a response to an increase in local membrane fluidity and these spots can be colocalized with Nile Red.³⁰ The combination of both dyes thus allows the visualization of highly fluid membrane regions.³⁹

Both CB1631 and CB1634, and their respective prsA-deleted mutants, presented a larger number of highly fluid spots in the presence of oxacillin, when compared with cells exposed to daptomycin (Figure 8). Furthermore, the number of spots increased with increasing concentrations of daptomycin and oxacillin. Additionally, CB1634 cells presented, in general, a higher number of highly fluid spots than CB1634 ΔprsA and CB1631 cells. Importantly, in CB1634 ΔprsA cells exposed to both daptomycin and oxacillin, spots marking highly fluid regions of the membrane were markedly different, when compared with CB1634 cells under similar growth conditions. Taken together, these results suggest that PrsA influences cellular membrane fluidity during daptomycin-resistance-mediated β-lactam sensitization.

Figure 8.

Membrane fluidity analysis of S. aureus cells. Daptomycin-susceptible CB1631 and daptomycin-resistant CB1634, and respective prsA mutant strains, stained with Laurdan and Nile Red showing white spots (some indicated by white arrows) indicative of highly fluid loci on the membrane when exposed to daptomycin, oxacillin and both. Scale bar = 1 μm. DAP, daptomycin; OXA, oxacillin. Concentrations indicated (0.25–8) in mg/L.

Discussion

The mortality related to S. aureus bloodstream infections is approx. 17%.⁴⁰ The combination of daptomycin and a β-lactam antibiotic is considered the best option to manage challenging patients, notably those with recurrent MRSA bacteraemia.^41,42

The activity of daptomycin is dependent on the presence of Ca²⁺ and several modes of action have been proposed, involving: (i) inhibition of lipoteichoic acid biosynthesis;^6,43 and (ii) disruption of membrane potential.⁵ In the presence of calcium, the anionic daptomycin molecule acquires its active cationic peptide form and increases the exposed hydrophobic surface, facilitating insertion of daptomycin into the negatively charged cell membrane.⁴ However, several studies have shown that the in vitro bactericidal activity of daptomycin is independent of lipoteichoic acid biosynthesis,⁴⁴ membrane depolarization occurs after cell death, being a consequence and not a cause of daptomycin bactericidal activity,⁴ and membrane potential dissipation occurs without disrupting membrane integrity.^28,29 Although the precise mode of action is not fully understood, the interactions between daptomycin and Ca²⁺,³¹ and between daptomycin and the anionic phospholipid PG,⁴⁵ are required for a bactericidal effect.

Müller et al.³² showed that daptomycin binding in Bacillus subtilis causes a drastic rearrangement of fluid lipid domains, causing delocalization of the membrane-associated lipid II synthase MurG and phospholipid synthase PlsX and hydrophobic mismatches between fluid and rigid membrane areas. This results mainly from daptomycin binding and clustering to fluid lipids, such as short, branched and/or unsaturated FA. While delocalization of the proteins may elucidate why daptomycin can block cell wall synthesis, the latter changes may explain proton leakage and cell membrane depolarization. The conclusion that daptomycin may influence fluid lipid domains and impair multiple cellular processes through interference of the lipid organization of the cell membrane opened a new view on the daptomycin mode of action. Antimicrobial agents affecting the cell membrane, such as amphiphilic aminoglycosides and polycationic compounds, affect the formation of lipid microdomains formed by the anionic phospholipid cardiolipin at the cell poles and division plane.^46,47

Recently, Grein et al.⁷ showed that Ca²⁺–daptomycin interacts with undecaprenyl-coupled cell envelope precursors in the presence of PG, forming a tripartite complex. When different bactoprenyl-lipid precursors were used in combination with 0.1% PG, binding of daptomycin to the membrane increased substantially and daptomycin complexes with lipid II could only be observed in the presence of PG. In the model proposed, the tripartite complex blocks cell wall synthesis and favours the delocalization of the machinery necessary for cell wall biosynthesis.

Among the phenotypic adaptation mechanisms described to confer bacteria the ability to survive when exposed to antibiotics are modifications in the FA composition of the phospholipids of the cellular membrane.^12,25,48 In the present study, we observed that the daptomycin-resistant CB1634 maintained a BSFA/SSFA ratio similar to that of untreated cells, but changed the anteiso/iso ratio. However, when exposed to oxacillin, these cells reduced the BSFA/SSFA ratio, indicating reduced membrane fluidity. The β-lactam oxacillin thus induced larger modifications in the FA composition of the phospholipids of daptomycin-resistant MRSA than daptomycin.

Both CB1631 and CB1634 cells increased their zeta potential in the presence of oxacillin, daptomycin and both. The CB1634 ΔmprF and CB1634 ΔprsA mutant cells presented, respectively, 5%–20% more negative and 10%–26% less negative zeta potential values than CB1634 cells. The Ca²⁺–daptomycin complex has an increased affinity for negatively charged phospholipids. Gain of function in the lysyl-PG synthase MprF leads to a reduction in negatively charged PG, which is necessary for daptomycin–membrane interactions.^32,49 Previous studies showed that mprF mutations favour excess cellular membrane synthesis of lysyl-PG and positive surface charge.^15,28 In fact, the primary response of S. aureus cells to daptomycin is, apparently, the induction of a more positive overall cell surface charge to favour electrostatic repulsion and avoid insertion of the positively charged daptomycin–calcium complex,⁵⁰ by efficient translocation of lysyl-PG to the outer leaflet of the cytoplasmatic membrane.¹⁵ In the present study, CB1634 ΔmprF cells, unable to produce MprF, could not decrease their negative character in the presence of antibiotics, when compared with the parent strain (Figure 3). In contrast, CB1634 ΔprsA cells were still able to decrease the negative character of their surface, probably because MprF was still present as wild-type in the CB1634 ΔprsA mutant. However, the daptomycin-susceptible CB1631 strain and the CB1631 ΔprsA mutant presented more negative values when exposed to the tested antibiotics than untreated cells. The results suggest that cells produced larger changes in the cell surface charge when exposed to the antibiotic to which they are more susceptible to.

This could be further elucidated by placing the cells in different concentrations of daptomycin and oxacillin. In general, daptomycin and oxacillin induced a dose-dependent increase in the negative SCD of MRSA cells, but CB1631-type strains presented larger variations when exposed to daptomycin, while CB1634-type strains presented larger variations when exposed to oxacillin. Similar conclusions could be drawn when analysing cell response at the phospholipid composition level.

In this study, different proportions of cells with relocated cardiolipin microdomains were also observed in response to daptomycin in comparison with untreated cells. The anionic cardiolipin is a major contributor to SCD, has a major role in the binding and positioning of proteins in the membrane³⁸ and its content increases as a response to antibiotic compounds.⁵¹ In enterococci, cardiolipin domain redistribution has been considered a major mediator of daptomycin resistance.^32,39

Moreover, our results suggest that, in daptomycin-resistant strain cells, the cardiolipin remodelling/relocation is strongly affected by PrsA, since its deletion was associated with a concomitant decrease in cardiolipin microdomains. The results regarding cardiolipin relocation are in agreement with the zeta potential and SCD data; PrsA has a pivotal role in daptomycin-resistance-mediated β-lactam sensitization in daptomycin-resistant strains, such as CB1634. This notion is reinforced by the following observations:

(i) CB1634 ΔprsA was less negative and showed lower SCD values than the CB1634 isolate; in contrast, CB1631 ΔprsA presented more negative zeta potential values and higher negative SCD values than CB1631 (Figures 3 and 4a). Although cardiolipin is an anionic lipid, the phosphates of the headgroup can ionize independently as strong acids, with ionization being dependent on lipid environment and protein interactions.³⁸

(ii) The cells presented a larger number of highly fluid membrane spots when exposed to oxacillin than to daptomycin and at higher antibiotic concentrations (Figure 8). The number of these spots was higher in the parent 1634 cells than in 1634 ΔprsA cells. Similarly, LiaX, a protein homologous to PrsA in enterococci resistant to daptomycin, has been shown to promote cell membrane remodelling with anionic phospholipids diverted from the septum, resulting in daptomycin binding away from its septal target.⁴⁸

In summary, we have highlighted the major role of PrsA in orchestrating the cell membrane lipid adaptation associated with the see-saw effect, in addition to its function required for the post-transcriptional maturation of PBP 2a; these effects may account for the resensitization of daptomycin-resistant strains to cell wall-specific β-lactams. The combination of daptomycin and β-lactams has gained increased acceptance for the treatment of MRSA infections caused by daptomycin-resistant strains, resulting in clinical successes. This study contributes greatly to the understanding of the impact of daptomycin/oxacillin on the cell membrane/cell wall machinery and provides fundamental insights into MRSA biology that may potentially be translated into the discovery of new therapeutic targets.

Funding

This work was supported by Fundação para a Ciência e a Tecnologia, I.P. (FCT, Portugal) through programme ‘Investigador FCT 2013’ (IF/01203/ 2013/CP1163/CT0002, awarded to C. C. C. R. de Carvalho) and by national funds from FCT in the scope of the project UIDB/04565/2020 and UIDP/04565/2020 of the Research Unit iBB-Institute for Bioengineering and Biosciences. This study was partly supported by Merck (formerly Cubist Pharmaceuticals), Lexington, MA, USA, and by a National Institutes of Health grant (NIH-R56AI102503-01A1) awarded to A. E. Rosato.

Transparency declarations

None to declare.

Supplementary data

Supplementary Materials and methods are available as Supplementary data at JAC Online.