Antibiotic-induced accumulation of lipid II synergizes with antimicrobial fatty acids to eradicate bacterial populations

  1. Ashelyn E Sidders
  2. Katarzyna M Kedziora
  3. Melina Arts
  4. Jan-Martin Daniel
  5. Stefania de Benedetti
  6. Jenna E Beam
  7. Duyen T Bui
  8. Joshua B Parsons
  9. Tanja Schneider
  10. Sarah E Rowe
  11. Brian P Conlon

  • Curated by eLife

    eLife logo

    Evaluation Summary:

    The authors here present how specific fatty acids modulate the bactericidal effect of the antibiotic vancomycin. The authors find that palmitoleic acid significantly increases the bactericidal activity of vancomycin and investigate the mechanism responsible. The key finding will be of interest to a broad audience of researchers focused on microbiology, host-pathogen interactions, and antimicrobial development, as well as to clinicians that treat antibiotic-recalcitrant infections.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Antibiotic tolerance and antibiotic resistance are the two major obstacles to the efficient and reliable treatment of bacterial infections. Identifying antibiotic adjuvants that sensitize resistant and tolerant bacteria to antibiotic killing may lead to the development of superior treatments with improved outcomes. Vancomycin, a lipid II inhibitor, is a frontline antibiotic for treating methicillin-resistant Staphylococcus aureus and other Gram-positive bacterial infections. However, vancomycin use has led to the increasing prevalence of bacterial strains with reduced susceptibility to vancomycin. Here, we show that unsaturated fatty acids act as potent vancomycin adjuvants to rapidly kill a range of Gram-positive bacteria, including vancomycin-tolerant and resistant populations. The synergistic bactericidal activity relies on the accumulation of membrane-bound cell wall intermediates that generate large fluid patches in the membrane leading to protein delocalization, aberrant septal formation, and loss of membrane integrity. Our findings provide a natural therapeutic option that enhances vancomycin activity against difficult-to-treat pathogens, and the underlying mechanism may be further exploited to develop antimicrobials that target recalcitrant infection.

Article activity feed

  1. Version published to 10.7554/elife.80246 on eLife
  2. eLife

    Evaluation Summary:

    The authors here present how specific fatty acids modulate the bactericidal effect of the antibiotic vancomycin. The authors find that palmitoleic acid significantly increases the bactericidal activity of vancomycin and investigate the mechanism responsible. The key finding will be of interest to a broad audience of researchers focused on microbiology, host-pathogen interactions, and antimicrobial development, as well as to clinicians that treat antibiotic-recalcitrant infections.

    (This preprint has been reviewed by eLife. We include the public reviews from the reviewers here; the authors also receive private feedback with suggested changes to the manuscript. Reviewer #1 agreed to share their name with the authors.)

  3. eLife

    Reviewer #1 (Public Review):

    The overarching objective of the study reported in this manuscript was to identify cell-membrane active agents (CMAAs) that potentiate antibiotic killing of Gram-positive pathogens by vancomycin. This work builds upon the research group's previous findings that a class of CMAAs, rhamnolipids, produced by Pseudomonas aeruginosa synergize with aminoglycosides in the killing of Staphylococcus aureus. In the current study, the group investigated the capacity of 7 additional CMAAs to potentiate vancomycin killing of S. aureus. The focus on vancomycin is appropriate given that this is the most commonly prescribed antibiotic for serious Gram-positive infections, and yet this antibiotic is also associated with a high rate of treatment failure that cannot be ascribed to resistance alone. The discovery of compounds that potentiate vancomycin activity or re-sensitize tolerant or resistant bacteria to the antibiotic is therefore of high significance to the field of infectious diseases. The authors discover that two unsaturated fatty acids, palmitoleic acid (PA) and linoleic acid (LA), potently synergize with vancomycin to kill S. aureus. Although the antimicrobial activity of select UFAs toward bacterial pathogens has long been recognized, the mechanisms by which these compounds kill bacteria or potentiate antibiotic activity are less clear. The authors, therefore, conducted a series of experiments aimed at understanding the mechanistic basis for UFA potentiation of vancomycin activity. Using a panel of antimicrobial compounds that target different steps of cell wall biosynthesis, the authors suggest that the accumulation of lipid II is necessary for the PA potentiation of vancomycin. To understand how dual PA/VAN treated cells are being killed, assays of membrane depolarization and permeability are conducted and reveal that PA/VAN treatment may increase permeability. However, in this particular experiment, it is unclear if this is different from PA treatment alone. The authors then utilize high-resolution fluorescence microscopy and TEM to demonstrate that PA/VAN treated cells have alterations in the membrane fluidity, septal architecture, and localization of cell division and peptidoglycan synthesis machinery. These findings lead the team to conclude that dual PA/VAN treatment delocalizes divisome and peptidoglycan biosynthesis machinery through the accumulation of lipid II moieties and subsequent effects on membrane fluidity. In a final experiment, the authors show that, excitingly, PA can re-sensitize vancomycin-intermediate and vancomycin-resistant Gram-positive bacteria to vancomycin.

    Strengths of the manuscript:

    1. The discovery of natural compounds that re-sensitize antibiotic-tolerant or antibiotic-resistant bacterial pathogens to commonly used antibiotics is a critical unmet need in the field of infectious diseases
    2. This study will be of broad interest to researchers focused on microbial pathogenesis, drug discovery, and antimicrobial-resistant bacteria.
    3. Multiple lines of evidence convincingly demonstrate the membrane and septal perturbations induced by dual PA/VANC therapy, suggesting a mechanism of action.
    4. Drug doses are carefully considered based on pilot data.
    5. High bacterial inocula are tested, which increases the confidence that dual therapy is capable of killing at least some tolerant/persister cells.
    6. The manuscript is extremely well-written and carefully considers prior studies reporting on the antimicrobial activities of UFAs.

    Weaknesses of the manuscript:

    1. Although dual PA and vancomycin therapy shows clear efficacy against vancomycin sensitive and resistant isolates of S. aureus and the high inocula used are likely to contain some amount of stochastically-developed persister cells, it is unclear what bacterial growth phase was tested in the killing experiments. Given that some forms of antibiotic tolerance, such as "tolerance-by-lag," will not be appropriately modeled in exponential phase bacteria, it would be important to know if bacterial cultures approaching the stationary phase are also effectively killed by UFAs plus vancomycin. Although the manuscript reports a 1:1000 dilution, it is unclear how long the bacteria are grown prior to challenge with dual therapy, and therefore conclusions about tolerant cells may need to be tempered.
    2. Although a strength of the manuscript is that multiple approaches are used to query membrane and septal effects of dual UFA/Vanc therapy, these studies only indirectly support the proposed mechanism regarding the accumulation of lipid II and alterations of membrane fluidity as the underlying reason for antibiotic sensitization. Future studies will be required to rigorously confirm this proposed mechanism, although this is considered a minor weakness.

  4. eLife

    Reviewer #2 (Public Review):

    Vancomycin is commonly used to treat infections caused by methicillin-resistant Staphylococcus aureus (MRSA) but treatment failure is common and there is evidence of the increasing frequency of strains with reduced vancomycin susceptibility. Therefore, there is interest in devising approaches to enhance treatment efficacy. In this work, the authors show that vancomycin activity is enhanced by the presence of a specific fatty acid. This has the potential to enhance topical antibiotic preparations and may also shed light on antibiotic activity in the host environment.

    The authors then go on to attempt to decipher the underlying mechanism responsible and provide some evidence that the combination of antibiotics and fatty acid disrupts the machinery that synthesises the cell wall. However, it is not clear if such disruption is causative of the enhanced killing phenotype. Furthermore, it is not clear whether non-potentiating fatty acids also disrupt this machinery, which is a key limitation.

  5. eLife

    Reviewer #3 (Public Review):

    The authors observe that unsaturated fatty acids in combination with vancomycin have a pronounced bactericidal effect on S. aureus with a >4 orders of magnitude greater killing at 20-fold MIC compared to vancomycin alone, which is specific to unsaturated fatty acids among several membrane-active agents. They observe that this phenomenon is not general to all antibiotics that target peptidoglycan pathways, but is only observed with vancomycin and bacitracin among the antibiotics they tested. The authors also observe that the combined treatment does not depolarize the bacterial cell membrane, but causes ATP levels to rise in the growth media. Microscopically, the combination results in membrane abnormalities not seen with either compound alone, which include foci that stain intensely with a dye that preferentially labels disordered regions of lipid bilayers. Furthermore, distorted septa and membrane defects are visible by EM only in cells treated with the vancomycin-palmitoleic acid combination; similar defects are also observed by light microscopy in the pattern of divisome protein localization and peptidoglycan biosynthesis. Finally, the authors demonstrate that the combination is active against vancomycin-resistant forms of several Gram-positive pathogens.

    The paper reports at least three remarkable observations. However, they are not pursued to a definitive resolution, which makes it difficult to determine what to conclude from the reported findings.

    One of the interesting observations made by the authors is the bactericidal effect of the combination of vancomycin and palmitoleic acid against apparent persisters (Fig. 1), which is a topic of considerable interest to the field of antibacterial therapy. This prompts the question of what the basis of this anti-persister activity is. The combination treatment could theoretically induce persisters to exit the persister state and thereby render them susceptible to killing by antibacterial mechanisms effective against the non-persister state. Alternatively, the combination may kill persister bacteria while they are still in the persister state. The combination might kill the persister and non-persister bacteria by the same mechanism, or the combination might act by multiple distinct mechanisms simultaneously, one or more of which is active against the persister state. Without further investigation, it is unclear what implications the anti-persister activity of vancomycin in combination with palmitoleic acid has for the general study of the phenomenon of bacterial persistence outside the specific context in which it was observed in this work.

    Another interesting observation is that palmitoleic acid only potentiates vancomycin and bacitracin among the several antibiotics that target peptidoglycan pathways (Fig.2 b-f), which argues against the potentiation resulting from compounded simultaneous generalized damage to the peptidoglycan (by an antibiotic) and to the cell membrane (by palmitoleic acid). The specificity of the effect is intriguing; however, it is not pursued to its resolution elsewhere in the manuscript: for example, it's not shown what phenotypes the peptidoglycan-targeting antibiotics other than vancomycin have in combination with palmitoleic acid in the experiments of figures 4-6 and if the phenotypes shown there are as distinct as in Fig. 2.

    Related to the above observation, the authors also invoke accumulation of lipid II as the mechanistic basis for the potentiation that they observe. Although it is a sound hypothesis, it is surprising that the authors limit themselves to inferring lipid II accumulation from effects observed under similar treatment conditions by other groups, without apparently assessing it directly and independently in their own experimental system under treatment, even though lipid II accumulation appears in the title of the manuscript. While it is highly likely that accumulation of lipid II will be observed in these experiments, the more important question is its causality in the antibacterial effects of vancomycin combined with palmitoleic acid. The killing of bacteria with antibiotics will produce a cascade of propagating effects in a cell, including membrane disturbances of various extent and types, and only further investigation can settle whether lipid II buildup is sufficient or incidental to achieve the antibacterial effects reported here. Assessing lipid II buildup under treatment conditions employed and correlating the degree of lipid II buildup to the extent of killing could at least strengthen the case for its role in the phenomenon if the degree of lipid II accumulation predicts the strength of the effect observed.

    Yet another set of curious observations are the microscopic abnormalities induced by the combination treatment that are not apparently caused by either of the agents alone (Figs 4-6). While these findings argue for a unique mechanism of bacterial killing by the vancomycin-palmitoleic acid combination, it is difficult to relate them to the other results reported in the manuscript and to the mechanistic hypothesis of lipid II buildup. It is unclear if the authors believe that DiI-C12 foci observed under vancomycin-palmitoleic acid treatment represent an enhanced extent of lipid II buildup under the combination treatment relative to vancomycin alone, or if the foci are a different phenomenon that is downstream from lipid II buildup, in other words, what the lipid composition of those membrane regions is exactly. It could be imagined that palmitoleic acid induces the confluence of accumulated lipid II into fewer regions of larger size or that it triggers a chain of events that result in a very different membrane composition. It is also unclear if any other antibiotics that target peptidoglycan produce similar foci in combination with palmitoleic acid, which would be important to know in order to assess their relevance to the unique effectiveness of the vancomycin-palmitoleic acid combination observed earlier. This consideration also applies to the EM images in Fig. 5 and the localization images in Fig. 6. In addition, it is unclear how observations in Fig. 6 relate to those in Fig. 4: aberrant PG incorporation could happen preferentially at the regions of increased fluidity that the authors observe, or peptidoglycan synthesis machinery could actively shun those regions. It seems important to connect those findings to each other in order to construct a mechanistic model. It is also unclear whether divisome proteins like EzrA gravitate towards domains of increased fluidity or if they are excluded from them. More generally, it remains unclear if there is a direct mechanistic connection between regions observed in Fig. 6 and those observed in Fig. 4, or if they are different consequences of the same upstream disturbance, and their locations are unrelated. For example, delocalizing the divisome by other means may or may not enhance vancomycin killing in the manner that the authors see, which would have implications for the mechanism of killing. Also, it is unclear if these disturbances appear in a defined time sequence or simultaneously or haphazardly, which can also shed light on their mechanistic relevance.

    Finally, the paper concludes by showing that this combination is effective in a range of Gram-positive species and is also active against vancomycin-resistant strains. This is potentially a very valuable finding because of the breadth of the spectrum and problems posed by antibiotic resistance; however how it relates to the preceding mechanistic hypotheses is left unexplored. As above, it would be useful to check that this new effect is still specific to palmitoleic acid and not observed with other membrane-active agents when combined with vancomycin. Importantly, vancomycin resistance seems to offer a unique opportunity to probe the mechanism of the bactericidal effect of the vancomycin-palmitoleic acid combination, both in terms of the dependence of the effect on the composition of the lipid II pool in the cell, how the effect changes under pretreatment with either agent alone, and the effect on phenotypes observed in Figs 4-6, but neither is pursued. It is not clear if the effectiveness of the combination depends on outpacing the modification of lipid II, or if it overcomes that mechanism of resistance, such as if palmitoleic acid somehow inhibits proteins that confer vancomycin resistance. Without knowing more about what underlies this activity against vancomycin-resistant bacteria and what its limitations are, it is difficult to assess its implications for the problem of vancomycin resistance in general.

    Methodological weaknesses:
    The implications of some of the reported results are difficult to interpret because of a lack of relevant comparisons. In Fig. 3, the performance of known peptidoglycan-targeting agents is not available as a benchmark to calibrate the performance of the vancomycin-palmitoleic acid treatment. The inclusion of gramicidin in Fig. 3a is helpful to indicate what signal a high degree of membrane depolarization would produce in this assay, and that does make the point that membrane depolarization is not the primary mechanism of killing by the combination treatment. However, it is unclear how the more relevant negative controls such as membrane-active agents of Fig. 1a or other peptidoglycan-targeting antibiotics perform in this assay, nor how degrees of depolarization (as assessed by this assay) correlate with bacterial viability. This would be needed in order to try to make inferences about its role in the effectiveness of the combination and how these data relate to the distinctions made elsewhere in the manuscript. Similar considerations apply to data in Fig. 3b. It indicates some amount of ATP leakage under combination treatment, but it is not clear how one can calibrate these values to the fitness of bacteria or how the relevant comparison treatments with established mechanisms of action against the cell envelope would perform in that assay. Literature reports of increased membrane permeability by other membrane-active agents are alluded to in the discussion, but it is unclear what signal they would produce in the specific assay the authors use to assess it in this manuscript. Similar points apply to Fig. 4 d-f.

    Some of the points made in the discussion section would require additional work to substantiate. The very specific statement that "the insertion of palmitoleic acid into the microenvironment around lipid II increases the disordered phospholipid environment surrounding the peptidoglycan monomer, subsequently creating large RIFs", would require something like a demonstration in an in vitro bilayer reconstituted from purified components (such as in ref 76, "Ca(2+)-Daptomycin targets cell wall biosynthesis by forming a tripartite complex with undecaprenyl-coupled intermediates and membrane lipids.") because there are far too many components in a cell to rule out all of them as contributors. The authors speculate that the vancomycin-palmitoleic acid combination acts too rapidly to allow enough time for the induction of vancomycin resistance in vancomycin-resistant strains. This could be tested by using strains with constitutive vancomycin resistance or by pre-exposure to vancomycin. The authors conclude by speculating about the possibility of using the vancomycin-palmitoleic acid combination in antibacterial therapy. However, several difficulties involved in doing so would be worth addressing. Getting two compounds with such different physical properties to distribute similarly in order for both to reach efficacious concentrations at infection sites is likely to be challenging. Also, how to deploy it in the types of infections where Gram-positive organisms are most dangerous: bacteremia, endocarditis, abdominal infections, and UTIs is far from straightforward. Thus, it would appear that it is the understanding of the mechanism that would be the finding of the greatest value as the guide to the clinical development of a pharmacologically tractable treatment. It seems that the authors believe that the combination acts through a mechanism that is unique to the combination, rather than through an enhanced version of the standalone mechanism of one of the agents with the other agent merely facilitating that mechanism, but a clearer analysis of that point would make the narrative easier to follow. It is unclear whether adding one of the agents allows the other to have the same effect at lower doses that would only be achievable at its higher doses if used alone.

    All in all, the reported results are potentially very valuable, and the amount of work done is substantial. However, it is hard to determine what the main conclusion of the manuscript is. If, on the one hand, it were the efficacy of the vancomycin-unsaturated fatty acid combination against clinically relevant Gram-positive pathogens, then the evidence of Figs 2-6 could mean that a variety of antibacterial effects result from the combination treatment that does not result from either agent alone. This could explain its effectiveness, but that would then require demonstrating experimental evidence for its translational potential using clinically relevant infection model(s). If, on the other hand, the main conclusion is the mechanism of how the combination works, which also appears to be the authors' intent from the title and abstract, then a clear mechanistic model has to be proposed, as well as all the likely alternative models, and both the narrative and the figures should be structured to buttress the proposed model for the mechanism of the UFA-vancomycin combination and to rule out possible alternative mechanisms. However, from the perspective of a model/mechanism-centric narrative, it is hard to see why the figures after Fig. 1 appear in the order in which they do as opposed to appearing in a different order and how a given figure builds on the preceding figures and leads into the next one. More generally, the experiments do not attempt to modulate the strength of the antibacterial effect of the combined treatment on the basis of what the authors believe its mechanism to be, and if a mechanistic model for an effect does not lead to predictions of what one can do in order to alter that effect, which can be then experimentally tested, it is difficult to build on its basis, such as for example to try to develop this into a clinical treatment.

    Thus while reporting promising observations, the paper does not achieve ultimate resolution of the questions raised by these observations, and difficulty in connecting the results of various reported experiments to each other makes it hard to evaluate the authors' claims about the mechanistic basis of their observations.

  6. Version published to 10.1101/2022.05.03.490474v1 on bioRxiv