Biochimica et Biophysica Acta 1858 (2016) 980–987

Contents lists available at ScienceDirect

Biochimica et Biophysica Acta

journal homepage: www.elsevier.com/locate/bbamem

Molecular mechanisms of membrane targeting antibiotics☆

Richard M. Epand ⁎, Chelsea Walker, Raquel F. Epand, Nathan A. Magarvey
Department of Biochemistry and Biomedical Sciences, McMaster University, 1280 Main Street West, Hamilton, Ontario L8S 4K1, Canada

a r t i c l e i n f o a b s t r a c t

Article history: The bacterial membrane provides a target for antimicrobial peptides. There are two groups of bacteria that have
Received 21 August 2015 characteristically different surface membranes. One is the Gram-negative bacteria that have an outer membrane
Received in revised form 7 October 2015 rich in lipopolysaccharide. Several antimicrobials have been found to inhibit the synthesis of this lipid, and it is
Accepted 23 October 2015
expected that more will be developed. In addition, antimicrobial peptides can bind to the outer membrane of
Available online 26 October 2015
Gram-negative bacteria and block passage of solutes between the periplasm and the cell exterior, resulting in
Keywords:
bacterial toxicity.
Lipid A In Gram-positive bacteria, the major bacterial lipid component, phosphatidylglycerol can be chemically modified
Lipopolysaccharide by bacterial enzymes to convert the lipid from anionic to cationic or zwitterionic form. This process leads to in-
Cardiolipin creased levels of resistance of the bacteria against polycationic antimicrobial agents. Inhibitors of this enzyme
Phosphatidylethanolamine would provide protection against the development of bacterial resistance.
There are antimicrobial agents that directly target a component of bacterial cytoplasmic membranes that can act
on both Gram-negative as well as Gram-positive bacteria. Many of these are cyclic peptides with a rigid binding
site capable of binding a lipid component. This binding targets antimicrobial agents to bacteria, rather than being
toxic to host cells. This article is part of a Special Issue entitled: Antimicrobial peptides edited by Karl Lohner and
Kai Hilpert.
© 2015 Elsevier B.V. All rights reserved.

1. Introduction The molecular components and morphology of membranes from

Gram-positive bacteria are fundamentally different from those of
There is an urgent need for new drugs against microbial infections Gram-negative bacteria (Fig. 1). Gram-negative bacteria are surrounded
because of the increased incidence of microbial resistance to traditional by two membranes, the cytoplasmic cell membrane and the outer
antibiotics [1]. One of the potential targets of novel antibiotics that could membrane. The outer monolayer of the membrane contains lipopoly-
be effective against resistant bacteria is the bacterial cell membrane. A saccharide (LPS) as the major lipid component, a lipid species unique
recent review has emphasized the characteristics of antimicrobial pep- to Gram-negative bacteria [6]. Although Gram-positive bacteria do not
tides that interact with bacterial membranes [2]. There already are have the additional outer membrane layer which Gram-negative bacte-
many antibiotics that target the bacterial membrane, but new, more po- ria do, they share the commonality of having a cell wall surrounding the
tent ones are needed. One of the reasons why the bacterial membrane is cytoplasmic membrane of the cell. The cell wall is comprised of peptido-
an attractive target for new antibiotics is that the bacterial membrane is glycan. The thickness of the cell wall is much greater for Gram-positive
arranged differently from a mammalian membrane. An important dif- than for Gram-negative bacteria [7]. In addition, the cell wall of Gram
ference is that anionic lipids are exposed on the surface of bacterial positive bacteria contains teichoic acids. The major lipid components
membranes, while in eukaryotic membranes, anionic lipids are seques- of the inner monolayer of the outer membrane of Gram-negative bacte-
tered to the monolayer facing the interior of the cell or organelle. For ria, as well as both monolayers of the cell membrane of both types of
this reason many of the antimicrobial agents are cationic so that they bacteria include phosphatidylglycerol (PG), phosphatidylethanolamine
have greater selectivity for bacterial membranes [3–5]. (PE) and cardiolipin (CL) [8]. Different species of bacteria have very dif-
ferent proportions of these three lipids. Some bacterial species also have
significant amounts of glycosyl diglycerides in addition to a number of
other minor lipid components [9].
Abbreviations: CL, cardiolipin; LPS, lipopolysaccharide; PE, phosphatidylethanolamine; There are many modes of action of antimicrobial agents including
PG, phosphatidylglycerol; PS, phosphatidylserine. targeting of specific enzymes or of DNA, or indirectly by stimulating
☆ This article is part of a Special Issue entitled: Antimicrobial peptides edited by Karl
the immune system. However, this review will be limited to agents
Lohner and Kai Hilpert.
⁎ Corresponding author. that directly target the bacterial membrane or that inhibit the synthesis
E-mail address: epand@mcmaster.ca (R.M. Epand). of components of bacterial membranes.

http://dx.doi.org/10.1016/j.bbamem.2015.10.018
0005-2736/© 2015 Elsevier B.V. All rights reserved.
 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987 981

Fig. 1. Schematic diagram comparing Gram-positive (left) and Gram-negative (right) bacterial cell membranes.

2. Antimicrobials targeting the bacterial outer membrane and cell super-family, which acts as a flippase. Previous in vitro analysis has sug-
wall gested the use of nisin as an antimicrobial. However, nisin also has a
high potential to cause antibiotic resistance in Staphylococcus aureus
Gram-negative bacteria tend to be more resistant to antimicrobial [17]. The cluster responsible for nisin production encodes three genes
agents than Gram-positive bacteria, because of the presence of the addi- with an assumed role in resistance: nisE, nisG, and nisI [18]. NisE and
tional protection afforded by the outer membrane. In addition to this NisG are thought to be ABC transporters, based on their homology to
outer membrane, both Gram-positive and Gram-negative bacteria other known ABC transporters responsible for drug efflux [18]. Similar-
have a cell wall that can protect the cytoplasmic membrane. Antimicro- ly, it has been well established that the nisI gene encodes a lipoprotein
bial agents targeting the cell wall have been known for many years. that functions in providing immunity to nisin [18].
Among them, the β-lactam antibiotics (like penicillin and cephalospo-
rin), inhibit cell wall synthesis. However, the effectiveness of these 2.2. Loss of outer membrane permeability
agents is limited as a result of increased expression of bacterial β-
lactamases that inactivate these antibiotics. Glycopeptides (like vanco- The outer membrane is more permeable than the cell membrane be-
mycin) also inhibit cell wall synthesis by forming a complex with D- cause of the presence of porin proteins that allow facile permeation of
Ala-D-Ala at the carboxyl terminus of a nascent peptidoglycan chain; small molecules having a mass of about 500 Da or less. However, dam-
these are also affected by increased resistance mechanisms. Phosphonic age to the outer membrane or loosening of its structure by removal of
acids (like fosfomycin) inhibit the cell wall enzyme MurA and are used Mg2 + results in more facile passage of larger molecules through the
in some Gram-negative bacterial infections. Though, like many of the cell membrane, resulting in bacterial toxicity [19].
other antibiotics that target the membrane, several different forms of It has also been suggested that bacterial toxicity can result from the
resistance to fosfomycin have developed, such as decreased drug up- opposite change in permeability of the outer membrane of Gram-
take, and antibiotic inactivation [10]. Another agent, trifolitoxin, is pro- negative bacteria. In general, all bacteria uphold a certain degree of per-
duced by Rhizobium leguminosarum and trifolii T24, as a ribosomal meability to allow for exchange of ions and small molecules between
synthesized natural product with extensive post-translational modifica- the extracellular media and the cell interior. There is evidence that cer-
tions. Trifolitoxin has the inherent ability to specifically target a small tain antimicrobial agents can cause bacterial toxicity by blocking the
subset of Gram-negative bacteria including alpha-proteobacteria such permeability of the outer membrane [20,21].
as Rhizobium, Sinorhizobium, and Mesorhizobium, as well as plant patho- Peptide antibiotics (such as Polymyxin B or E) bind to lipid A, the an-
gens such as Agrobacterium without causing toxicity to itself [11–13]. Al- chor for lipopolysaccharide in Gram-negative bacteria. Polymyxin B ini-
though the specific mode of action is not well understood, it is possible tially accumulates in the outer membrane and subsequently penetrates
that trifolitoxin is able to specifically target a component of the outer into the inner membrane and finally enters the cytoplasm. Resistance to
membrane of Gram-negative bacteria. Trifolitoxin, while important polymyxin can occur by modification of the phosphate groups of lipid A
within the agricultural industry, has yet to be recognized as a valuable and by mutations in the two component regulatory systems PmrAB and
pharmaceutical agent. PhoP/PhoQ. Although this resistance mechanism is rare, it is increasing
[22–25].
2.1. Outer membrane and cell wall synthesis
3. Antimicrobials targeting phospholipids in the cell membrane
Nisin is a lantibiotic that binds strongly to lipid II, an essential pre-
cursor of peptidoglycan synthesis. This binding results in nisin causing There are several mechanisms by which agents that target phospho-
potent permeabilization of membranes containing lipid II [14,15]. An- lipids can alter membrane properties. One of such ways is to alter the
other enzyme involved in the synthesis of LPS is PagP that transfers a bulk physical properties [26–32] of the membrane (Fig. 2). Agents
palmitoyl chain to lipid A. Inhibition of this process has been suggested which alter bulk properties of the membrane do not bind to a specific
to be a strategy for damaging the outer membrane [16]. The assembly of membrane component, but affect one or more of the bulk properties.
the peptidoglycan layer in bacteria requires lipid II, which is synthesized Examples of such alterations that can occur include changes in the spa-
at the inner surface of the cytoplasmic membrane, before being tial distribution of molecules within the cell membrane or alterations of
translocated across the bilayer. Transport of the cell wall precursor, a bulk physical property such as intrinsic curvature or fluidity. Examples
lipid II in Escherichia coli requires MurJ, a member of an exporter of these effects are illustrated in Fig. 2. Alternatively, a particular lipid
982 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987

Fig. 2. Graphical representation of the different outcomes peptide antimicrobials can have on properties of the bacterial cell membrane. Antimicrobial peptides can affect the physical prop-
erties of the cellular membrane such as (A) induction of membrane physical curvature or intrinsic curvature strain (see [36]), (B) lipid clustering, (C) prompting packing defects resulting
in complete or partial loss of the permeability barrier, and (D) by directly targeting components of a membrane such as lipids leading to a variety of consequences.

class can be targeted that can affect the bacteria by various mechanisms, diverse relationships that can occur as a result of changes in the physical
including that of altering the bulk biophysical properties (see below). properties of membrane lipids [42].
Targeting specific phospholipids has the potential of being effective
against both Gram-positive and Gram-negative bacteria. These phe-
3.2. Specific phospholipids
nomena are not necessarily independent of each other. For example,
membrane physical curvature can drive clustering and vice versa, in ad-
The lipid composition of bacterial membranes varies considerably
dition, clustering will lead to packing defects at the boundary between
among different bacterial species. The predominant phospholipid clas-
domains. Directly targeting lipids can lead to any of the other three phe-
ses are CL, PG and PE. The phospholipids of bacteria have the same
nomena illustrated.
headgroup structure as the corresponding lipids in eukaryotes, but in
bacteria the acyl chains tend to be shorter and more saturated [9].
3.1. Bulk membrane properties There is also the difference that anionic lipids and PE are exposed on
the external surface of bacterial membranes, while they are largely se-
3.1.1. Clustering questered to the cytoplasmic surface of eukaryotic membranes. These
Lipid components of biological membranes are not uniformly dis- features contribute to the possibility of designing antibacterial com-
tributed in the membrane, but rather are enriched in certain domains pounds that target specific lipids of bacteria (Fig. 3). Agents that target
[33]. This distribution can be altered, for example, by clustering anionic specific lipids may or may not act by altering the bulk membrane prop-
lipids with cationic peptides having multiple positive charges [8,33–35]. erties. There is also an indirect way to inhibit specific lipids and this is by
This can be toxic to the bacteria by preventing the interaction of lipids inhibiting enzymes that are required for their synthesis.
with other molecular components of the cell membrane or by
disrupting existing natural domains or by forming phase boundary de-
3.2.1. Cardiolipin (CL)
fects between the clustered lipids and the bulk of the membrane.
CL is an important lipid that segregates into a domain in the pres-
ence of certain cationic antimicrobial peptides. We have shown that
3.1.2. Curvature this clustering phenomenon can define the bacterial species specificity
Biological membranes can have physical curvature by bending from of certain antimicrobial agents [35].
a flat position. Since the membrane is composed of a bilayer, each In the case of the antibiotic daptomycin, there is evidence that a
monolayer of the bilayer would have to adopt opposite types of curva- component of its mechanism of action is the redistribution of cardiolipin
ture in order to bend. In addition to physical curvature, there is also on the membrane surface. The mode of action of daptomycin is associ-
the intrinsic curvature of each monolayer of the bilayer. This is the ated with the direct interaction of the molecule with the bacterial divi-
shape the monolayer would acquire if it did not have to be juxtaposed sion septum of the cell membrane in a calcium-dependent manner [43].
opposite the other monolayer. These concepts are described in more de- In the case of daptomycin-resistant Enterococcus faecalis, mutations
tail in a recent reference [36]. Membrane curvature has been suggested have been found in cardiolipin synthase [44]. Furthermore, the single
to be involved both in the action of antimicrobial peptides perturbing amino acid mutations have been localized to occur within the LiaF
membrane morphology or properties, as well as membrane intrinsic gene [43]. This particular gene has been found to encode a transmem-
curvature modulating the potency of these peptides [31,37–40]. It brane protein which is a part of a three-component regulatory system
should be emphasized that membrane curvature is one of many physi- involved in the regulation of the bacterial cell envelope in response to
cal properties that affect the function of antimicrobial agents. There are stress [45]. As mutations occur within this particular region, daptomy-
examples in which membrane curvature properties do not predict anti- cin no longer interacts with the septum, and subsequently gets diverted
microbial potency [41]. There is an excellent recent discussion of the and trapped in distinct membrane regions.
 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987 983

Fig. 3. Chemical structures of the eight antimicrobial peptides and their respective lipid targets within the bacterial cell membrane. The following antimicrobial peptides have been well
characterized in regard to their specific interaction with membrane lipids; telomycin targetscardiolipin (CL), cinnamycin and duramycin target phosphatidylethanolamine (PE), lysocin E
targets menaquinone, and daptomycin targets phosphatidylglycerol (PG).

Telomycin is a member of the cyclic depsipeptide family of antibi- 3.2.2. Phosphatidylethanolamine (PE)
otics isolated from the culture broth of Streptomyces canus [46,47]. Al- The presence of PE in the membrane is important for its high affinity
though the mode of action of telomycin remains to be determined, it binding to several antimicrobial peptides [51]. Many bacteria have a
appears to be especially potent toward Gram-positive bacteria organ- high concentration of PE on their surface. In mammalian cells, PE is
isms. The biosynthetic cluster responsible for telomycin has recently largely sequestered to the cytoplasmic leaflet of the plasma membrane,
been established, and several semi-synthetic derivatives have dem- although there is a low concentration, about 5% of the total lipid, of PE
onstrated rapid, and specific bacterial lysis of organisms including on the surface of mammalian cells. Any drugs targeting the PE in bacte-
those which are multi-drug resistant [48]. Furthermore, an analog ria would have to be selective against the low concentration of PE on the
of telomycin, LL-A-0341β, has an assumed mode of action specifical- surface of mammalian cells. The first peptides that were suggested to
ly targeting the cytoplasmic membrane of Gram-positive organisms use PE as a receptor were several members of the class I, type B
[49]. Due to the structural similarities of these depsipeptide antibi- lantibiotics such as cinnamycin and duramycin [52,53]. Support for
otics, it could be presumed that telomycin may also have a specific this suggestion comes from the observation that mutation of Bacillus
mode of action toward a lipid component of the cytoplasmic subtilis that results in resistance to duramycin is accompanied by a de-
membrane of Gram-positive organisms. Possible lipid components crease in the amount of membrane PE and CL [54]. It has also been
could include cardiolipin, representing a new molecular target for shown that membrane curvature plays a role in the interaction of
antibiotics. duramycin to membranes [55]. This study also showed that duramycin
Recently a new synthetic antibiotic, cyclo(RRRWFW), has been can promote flip-flop of PE, that could be a cause of mammalian cell tox-
shown to target cardiolipin using live cell imaging and isothermal ti- icity [55]. The specific mode of action of duramycin (i.e. targeting phos-
tration calorimetry [50]. This cyclic peptide has a unique mechanism pholipase A2) has been shown using circular dichroism [56]. A study by
of antibacterial action that does not involve either membrane per- [56], suggested that duramycin is able to promote its mode of action by
meation or interaction with an intracellular target, but rather by indirectly inhibiting phospholipase A2 through the sequestering of PE
causing redistribution of membrane lipids into different domains within the membrane. Cinnamycin has particularly strong binding affin-
[50]. Thus, this peptide is an example of an antibiotic with a specific ity, in the nanomolar range, to membranes containing PE [57,58]. The
lipid target, but one that acts by affecting the biophysical properties binding is not simply a consequence of electrostatic partitioning, but
of the membrane. rather indicates a specific binding site on the peptide. An NMR study
984 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987

of the interaction of cinnamycin with lysophosphatidylethanolamine bacterial membranes by causing membrane perturbations as well as in-
suggests that the peptide folds into a tight and specific binding site for ducing lipid flip-flop, and membrane leakage [67]. Similar to many
this lipid [59]. As the tight binding interaction occurs, cinnamycin is other clinically relevant antibiotics, increasing levels of daptomycin re-
able to elicit its potent antimicrobial action resulting in the disruption sistant organisms soon followed its introduction and use as a pharma-
of the bacterial cell membrane. ceutical agent. Model organisms such as Bacillus subtilis 168 have been
A similar system involves the plant antimicrobial peptides belonging utilized to demonstrate the genetic, genomic, and phenotypic conse-
to the cyclotide family and including kalata B1, kalata B2 and quences of daptomycin resistant isolates, including a resistance mecha-
cycloviolacin O2. These peptides also have a rigid, cyclic structure and nism involving the reduction in PG membrane content [76].
their membrane lytic activity is enhanced by the presence of PE in the
outer leaflet of the membrane [51,60]. The binding of these peptides 3.2.5. Metabolic alteration of lipid headgroups leading to resistance
to PE-containing membranes has not yet been studied experimentally In addition to the examples given above of peptides binding to spe-
in detail. However, there have been molecular dynamic simulations cific lipid components of a membrane, a common feature of antibacteri-
that suggest a specific interaction of these cyclic peptides with PE [60– al peptides is that many of them are polycationic. As a consequence,
62], although some of the groups found experimentally to be required they will be preferentially sequestered to the surface of bacteria
for specific interaction with PE was not supported by the simulations where anionic lipids are exposed. To protect against these antimicrobial
[51]. agents, bacteria can reduce the negative charge on their membrane by
Cyclotides are a large family of host defense peptides in the plant modifying the anionic lipids PG and sometimes also CL, by covalent at-
kingdom. It has been suggested that many of these peptides, that all tachment of lysine or alanine residues to the headgroup [77,78]. This re-
form stable ring structures, may be a growing class of PE-binding anti- action is catalyzed by the integral membrane protein, MprF. This
microbial agents that can serve as a membrane lipid that is exposed in enzyme catalyzes the transfer of lysine or alanine from aminoacyl-
some bacteria, but to a much lower extent in normal eukaryotic cells, tRNA on the cytoplasmic side of the membrane. MprF also has another
affording a means for targeting these membrane lytic agents toward in- domain with a flippase activity that can transfer the modified PG from
fectious bacteria [62,63]. the cytoplasmic side of the membrane to the exterior where it can
repel cationic antimicrobial peptides in the environment [79]. Another
3.2.3. Menaquinone modification to reduce the surface negative charge of Gram-positive
Menaquinones and other isoprenoid quinones are components of bacteria is to modify LPS with aminoarabinose [80].
the cytoplasmic membrane of Gram-positive, Gram-negative,
Cyanobacteria, and Mycoplasma organisms [64]. Their role within the 4. Summary
cytoplasmic membrane includes acting as an obligatory component of
the electron transport chain, promoting endospore formation, and par- The bacterial membrane has an important role in bacterial cell sur-
ticipating in active transport [65–67]. Lysocin E is a natural product vival and on the potency of antimicrobial peptides to protect the host.
identified from the culture supernatant of Lysobacter sp. RH2180-5 Gram-negative bacteria have an additional layer of protection of the
was developed by utilizing an in vivo approach to identify therapeutical- outer membrane. The outer leaflet of the outer membrane of these bac-
ly effective antibiotics, with potentially unique targets [68]. Lysocin E teria is comprised largely of LPS. Several potent antimicrobial agents act
caused rapid bacterolysis in S. aureus, and exhibited a higher degree of by inhibiting the synthesis of mature LPS at various stages. This action
potency than other commonly utilized antibiotics such as vancomycin renders these bacteria more vulnerable to destruction because the
[68]. Through a series of mutational analysis, lysocin E was determined outer membrane becomes leakier, making this additional protective
to directly bind to menaquinone within the bacterial membrane [66]. layer ineffective. Curiously, there are also cases in which antimicrobial
The recent discovery of lysocin E suggests that there are still additional agents block the permeability of solutes across the outer membrane of
antimicrobials to be discovered with the ability to target unique compo- Gram-negative bacteria, resulting in their lethality. Thus targeting the
nents of bacterial membranes such as menaquinone. synthesis of the outer membrane of Gram-negative bacteria or altering
A second scenario involving menaquinone as a potential target for the properties of this membrane provides a strategy for designing
antimicrobials arises in the form of indirectly targeting menaquinone agents targeted to Gram-negative bacteria.
through its biosynthetic pathway. Ro 48–071, is a chemically synthe- Phospholipids can also be involved in the action of antimicrobial
sized product initial found as a lanosterol cyclase inhibitor to effectively agents. There are three main phospholipids in most Gram-negative
lower cholesterol levels in certain mammals [69]. This was later extend- and Gram-positive bacteria. These are PG, PE and CL. PG is generally
ed to demonstrate Ro 48–071 as being able to inhibit the enzyme menA found to be the most abundant of these three lipids. It is an anionic
in menaquinone biosynthesis in Mycobacteria [70]. This leads to a de- lipid and hence will attract cationic peptides. Bacteria can protect them-
crease in oxygen consumption and disruption of the electron transport selves from these peptides by modifying the PG headgroup with the at-
within the bacterial membrane [70]. Due to the exhibited potency and tachment of lysine or alanine. This will reduce the negative charge on
potential for development as a pharmaceutical agent, Ro 48–071 en- the membrane and thus make the bacteria more resistant to the action
tered into the discovery optimization stage for the treatment of Myco- of cationic antimicrobial peptides. A strategy to combat this resistance
bacterium tuberculosis in [71]. mechanism is to inhibit the enzyme that modifies the PG headgroup
as well as inhibiting its flippase activity that transfers the modified
3.2.4. Phosphatidylglycerol (PG) lipid to the exterior of the cell. PE is generally abundant in Gram-
PG is another of the essential components of bacterial cell mem- negative bacteria and in most species of Gram-positive Bacilli and
branes, and assumes a role as providing bacterial membrane stability Clostridium difficile. There are several cyclic peptides that bind specifical-
[72]. PGs constitute on average, approx. 25% of the total lipid composi- ly to PE and this property can be used to target these bacteria. In addi-
tion, but can vary depending on bacteria type. In S. aureus the PG con- tion, mixtures of PE and an anionic lipid can form segregated clusters
tent averages around 80%, in comparison to E. coli with an average PG if the anionic lipid is preferentially bound to an antimicrobial peptide.
lipid content of 20% [73]. Daptomycin is a calcium-dependent antibiotic The clustering mechanism can also explain the antimicrobial activity
produced by Streptomyces roseosporus which inhibits Gram-positive of agents that bind to cardiolipin [50]. There are peptides that bind spe-
bacteria, including those which are multi-drug resistant [74,75]. Circular cifically to either CL or PE, providing a mechanism for targeting antimi-
dichroism studies revealed daptomycin to undergo two structural tran- crobial agents to specific bacteria.
sitions; one occurring through interaction with Ca2+, and the second by Thus the membrane is a versatile target for antimicrobial agents that
direct interaction with PG [75]. Daptomycin exerts its mode of action on can be used as targets for these agents or repel them in resistance, as
 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987 985

Fig. 4. The lipid components of bacterial membranes can play a multitude of functional roles. Such roles include the (A) formation of micelles (or exosome vesicles), (B) functioning as
receptors, (C) acting as primary messengers, (D) forming lipid domains and participating in protein recruitment or dispersion, (E) maintaining proton gradients formed by H+-ATPases
and providing energy storage, (F) providing a physical barrier for the cell, and (G) acting as secondary messengers.

well as provide an important site for toxic mechanisms. Bacterial lipids References
as well as bacterial proteins that synthesize or process membrane com-
ponents are attractive targets for antimicrobial therapy. [1] S. Reardon, US vows to combat antibiotic resistance, Nature 513 (2014) 471.
[2] V. Teixeira, M.J. Feio, M. Bastos, Role of lipids in the interaction of antimicrobial pep-
The bacterial membrane can have many roles affecting the action of tides with membranes, Prog. Lipid Res. 51 (2012) 149–177.
antimicrobial agents (Fig. 4). The membrane presents itself as a barrier [3] D. Alves, P.M. Olivia, Mini-review: antimicrobial peptides and enzymes as prom-
that drugs must breach in order to access intramolecular targets. How- ising candidates to functionalize biomaterial surfaces, Biofouling 30 (2014)
483–499.
ever, this membrane barrier is also responsible for establishing concen- [4] C.D. Fjell, J.A. Hiss, R.E. Hancock, G. Schneider, Designing antimicrobial peptides:
tration and electrical gradients between the cell and its environment. form follows function, Nat. Rev. Drug Discov. 11 (2012) 37–51.
These gradients can be dissipated by damage to the cell membrane. [5] R.M. Epand, H.J. Vogel, Diversity of antimicrobial peptides and their mechanisms of
action, Biochim. Biophys. Acta 1462 (1999) 11–28.
One gradient that is of particular importance is the transmembrane pro- [6] B. Beutler, LPS in microbial pathogenesis: promise and fulfilment, J. Endotoxin Res. 8
ton gradient that is also required for certain drug efflux mechanisms as- (2002) 329–335.
sociated with bacterial resistance. Several drugs or drug combinations [7] A. Wada, M. Kono, S. Kawauchi, Y. Takagi, T. Morikawa, K. Funakoshi, Rapid discrim-
ination of Gram-positive and Gram-negative bacteria in liquid samples by using
have been shown to be potent antimicrobial agents as a consequence NaOH-sodium dodecyl sulfate solution and flow cytometry, PLoS One 7 (2012),
of inhibiting drug efflux as well as being effective against resistant bac- e47093.
terial strains [81–85]. The membrane also contains many molecules [8] R.M. Epand, R.F. Epand, Bacterial membrane lipids in the action of antimicrobial
agents, J. Pept. Sci. 17 (2011) 298–305.
with unique chemical structure and properties that can be used as
[9] C. Ratledge, S.G. Wilkinson, Microbial Lipids, Academic Press, San Diego, 1988.
receptors for antimicrobial agents. Enzymes responsible for the biosyn- [10] D.E. Karageorgopoulos, R. Wang, X.H. Yu, M.E. Falagas, Fosfomycin: evaluation of the
thesis of essential membrane components are also potential targets for published evidence on the emergence of antimicrobial resistance in Gram-negative
antimicrobial therapy. Additionally, the membrane is in a dynamic state pathogens, J. Antimicrob. Chemother. 67 (2012) 255–268.
[11] E.W. Triplett, Isolation of genes involved in nodulation competitiveness from Rhizo-
with components being chemically modified and also translocated bium leguminosarum bv. Trifolii T24, Proc. Natl. Acad. Sci. U. S. A. 85 (1988)
across the membrane or clustering to form domains. Inhibition of 3810–3814.
these processes offers yet other strategies to design antimicrobial com- [12] E.W. Triplett, T.M. Barta, Trifolitoxin production and nodulation are necessary for
the expression of superior nodulation competitiveness by Rhizobium leguminosarum
pounds. Clearly this is a rich and complex field that is just starting to be bv. Trifolii strain T24 on clover, Plant Physiol. 85 (1987) 335–342.
exploited. [13] E.W. Triplett, B.T. Breil, G.A. Splitter, Expression of tfx and sensitivity to the rhizobial
peptide antibiotic trifolitoxin in a taxonomically distinct group of alpha-
proteobacteria including the animal pathogen Brucella abortus, Appl. Environ.
Microbiol. 60 (1994) 4163–4166.
Conflict of interest [14] S.T. Hsu, E. Breukink, E. Tischenko, M.A. Lutters, K.B. de, R. Kaptein, A.M. Bonvin, N.A.
van Nuland, The nisin–lipid II complex reveals a pyrophosphate cage that provides a
blueprint for novel antibiotics, Nat. Struct. Mol. Biol. 11 (2004) 963–967.
The authors have no conflict of interest. [15] I. Wiedemann, E. Breukink, K.C. van, O.P. Kuipers, G. Bierbaum, K.B. de, H.G. Sahl, Spe-
cific binding of nisin to the peptidoglycan precursor lipid II combines pore formation
and inhibition of cell wall biosynthesis for potent antibiotic activity, J. Biol. Chem. 276
(2001) 1772–1779.
Acknowledgements
[16] R.E. Bishop, The lipid A palmitoyltransferase PagP: molecular mechanisms and role
in acterial pathogenesis, Mol. Microbiol. 57 (2005) 900–912.
This work was supported by a grant from the Natural Sciences and [17] K.L. Blake, C.P. Randall, A.J. O'Neill, In vitro studies indicate a high resistance poten-
Engineering Research Council of Canada, grant 9848 (RME) and CIHR- tial for the lantibiotic nisin in Staphylococcus aureus and define a genetic basis for
nisin resistance, Antimicrob. Agents Chemother. 55 (2011) 2362–2368.
JPIAMR grant # 337866 (NAM). NAM is a Canada Research Chair in Nat- [18] K. Siegers, K.D. Entian, Genes involved in immunity to the lantibiotic nisin produced
ural Products and Chemical Biology. by Lactococcus lactis 6F3, Appl. Environ. Microbiol. 61 (1995) 1082–1089.
986 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987

[19] N.H. Lam, Z. Ma, B.Y. Ha, Electrostatic modification of the lipopolysaccharide layer: Studies of Telomycin Reveal New Lipopeptides with Enhanced Activity, J. Am. Chem.
competing effects of divalent cations and polycationic or polyanionic molecules, Soc. 137 (2015) 7692–7705.
Soft Matter 10 (2014) 7528–7544. [49] B. Oliva, W.M. Maiese, M. Greenstein, D.B. Borders, I. Chopra, Mode of action of the
[20] R.F. Epand, B.P. Mowery, S.E. Lee, S.S. Stahl, R.I. Lehrer, S.H. Gellman, R.M. Epand, cyclic depsipeptide antibiotic LL-AO341 beta 1 and partial characterization of a
Dual mechanism of bacterial lethality for a cationic sequence-random copolymer Staphylococcus aureus mutant resistant to the antibiotic, J. Antimicrob. Chemother.
that mimics host-defense antimicrobial peptides, J. Mol. Biol. 379 (2008) 38–50. 32 (1993) 817–830.
[21] R.F. Epand, H. Sarig, A. Mor, R.M. Epand, Cell-wall interactions and the selective bac- [50] K. Scheinpflug, O. Krylova, H. Nikolenko, C. Thurm, M. Dathe, Evidence for a novel
teriostatic activity of a miniature oligo-acyl-lysyl, Biophys. J. 97 (2009) 2250–2257. mechanism of antimicrobial action of a cyclic R-,W-rich hexapeptide, PLoS ONE 10
[22] J.V. Farizano, M.L. Pescaretti, F.E. Lopez, F.F. Hsu, M.A. Delgado, The PmrAB system- (2015), e0125056.
inducing conditions control both lipid A remodeling and O-antigen length distribu- [51] D.A. Phoenix, F. Harris, M. Mura, S.R. Dennison, The increasing role of phosphatidyl-
tion, influencing the Salmonella typhimurium–host interactions, J. Biol. Chem. 287 ethanolamine as a lipid receptor in the action of host defence peptides, Prog. Lipid
(2012) 38778–38789. Res. 59 (2015) 26–37.
[23] A.D. Gutu, N. Sgambati, P. Strasbourger, M.K. Brannon, M.A. Jacobs, E. Haugen, R.K. [52] A. Fredenhagen, F. Marki, G. Fendrich, W. Marki, j. Gruner, J. van Oostrum, F.
Kaul, H.K. Johansen, N. Hoiby, S.M. Moskowitz, Polymyxin resistance of Pseudomo- Raschdorf, and H.H. Peter, Duramycin B and C, two new lanthionine-containing an-
nas aeruginosa phoQ mutants is dependent on additional two-component regulato- tibiotics as inhibitors of phospholipase A2, and structural revision of duramycin and
ry systems, Antimicrob. Agents Chemother. 57 (2013) 2204–2215. cinnamycin, in: G. Jung, H.G. Sahl (Eds.), Nisin and Lantibiotics, ESCOM, Leiden
[24] T.L. Harris, R.J. Worthington, L.E. Hittle, D.V. Zurawski, R.K. Ernst, C. Melander, Small 1991, pp. 131–140.
molecule downregulation of PmrAB reverses lipid A modification and breaks colistin [53] H.G. Sahl, R.W. Jack, G. Bierbaum, Biosynthesis and biological activities of lantibiotics
resistance, ACS Chem. Biol. 9 (2014) 122–127. with unique post-translational modifications, Eur. J. Biochem. 230 (1995) 827–853.
[25] A.K. Miller, M.K. Brannon, L. Stevens, H.K. Johansen, S.E. Selgrade, S.I. Miller, N. [54] E.A. Dunkley Jr., S. Clejan, A.A. Guffanti, T.A. Krulwich, Large decreases in membrane
Hoiby, S.M. Moskowitz, PhoQ mutations promote lipid A modification and poly- phosphatidylethanolamine and diphosphatidylglycerol upon mutation to
myxin resistance of Pseudomonas aeruginosa found in colistin-treated cystic fibrosis duramycin resistance do not change the protonophore resistance of Bacillus subtilis,
patients, Antimicrob. Agents Chemother. 55 (2011) 5761–5769. Biochim. Biophys. Acta 943 (1988) 13–18.
[26] D. Marquardt, N. Kucerka, J. Katsaras, T.A. Harroun, α-Tocopherol's location in mem- [55] K. Iwamoto, T. Hayakawa, M. Murate, A. Makino, K. Ito, T. Fujisawa, T. Kobayashi,
branes is not affected by their composition, Langmuir 31 (2015) 4464–4472. Curvature-dependent recognition of ethanolamine phospholipids by duramycin
[27] M. Herrmann, E. Schneck, T. Gutsmann, K. Brandenburg, M. Tanaka, Bacterial lipo- and cinnamycin, Biophys. J. 93 (2007) 1608–1619.
polysaccharides form physically cross-linked, two-dimensional gels in the presence [56] F. Marki, E. Hanni, A. Fredenhagen, J. van Oostrum, Mode of action of the
of divalent cations, Soft Matter 11 (2015) 6037–6044. lanthionine-containing peptide antibiotics duramycin, duramycin B and C, and
[28] N. Voievoda, T. Schulthess, B. Bechinger, J. Seelig, Thermodynamic and biophysical cinnamycin as indirect inhibitors of phospholipase A2, Biochem. Pharmacol. 42
analysis of the membrane-association of a histidine-rich peptide with efficient anti- (1991) 2027–2035.
microbial and transfection activities, J. Phys. Chem. B 119 (2015) 9678–9687. [57] G. Machaidze, A. Ziegler, J. Seelig, Specific binding of Ro 09-0198 (cinnamycin) to
[29] K. Hu, Y. Jiang, Y. Xie, H. Liu, R. Liu, Z. Zhao, R. Lai, L. Yang, Small-anion selective phosphatidylethanolamine: a thermodynamic analysis, Biochemistry 41 (2002)
transmembrane “holes” induced by an antimicrobial peptide too short to span 1965–1971.
membranes, J. Phys. Chem. B 119 (2015) 8553–8560. [58] G. Machaidze, J. Seelig, Specific binding of cinnamycin (Ro 09-0198) to phosphati-
[30] A. Mularski, J.J. Wilksch, H. Wang, M.A. Hossain, J.D. Wade, F. Separovic, R.A. dylethanolamine. Comparison between micellar and membrane environments, Bio-
Strugnell, M.L. Gee, Atomic force microscopy reveals the mechanobiology of lytic chemistry 42 (2003) 12570–12576.
peptide action on bacteria, Langmuir 31 (2015) 6164–6171. [59] K. Hosoda, M. Ohya, T. Kohno, T. Maeda, S. Endo, K. Wakamatsu, Structure determi-
[31] K.M. Scherer, J.H. Spille, H.G. Sahl, F. Grein, U. Kubitscheck, The lantibiotic nisin in- nation of an immunopotentiator peptide, cinnamycin, complexed with
duces lipid II aggregation, causing membrane instability and vesicle budding, lysophosphatidylethanolamine by 1H-NMR1, J. Biochem. 119 (1996) 226–230.
Biophys. J. 108 (2015) 1114–1124. [60] C.K. Wang, H.P. Wacklin, D.J. Craik, Cyclotides insert into lipid bilayers to form mem-
[32] E. Strandberg, A.S. Ulrich, AMPs and OMPs: is the folding and bilayer insertion of brane pores and destabilize the membrane through hydrophobic and
beta-stranded outer membrane proteins governed by the same biophysical princi- phosphoethanolamine-specific interactions, J. Biol Chem. 287 (2012) 43884–43898.
ples as for alpha-helical antimicrobial peptides? Biochim. Biophys. Acta 1848 [61] R. Burman, A.A. Stromstedt, M. Malmsten, U. Goransson, Cyclotide-membrane inter-
(2015) 1944–1954. actions: defining factors of membrane binding, depletion and disruption, Biochim.
[33] R.M. Epand, R.F. Epand, Domains in bacterial membranes and the action of antimi- Biophys. Acta 1808 (2011) 2665–2673.
crobial agents, Mol. BioSyst. 5 (2009) 580–587. [62] S.T. Henriques, Y.H. Huang, M.A. Castanho, L.A. Bagatolli, S. Sonza, G. Tachedjian, N.L.
[34] R.F. Epand, G. Wang, B. Berno, R.M. Epand, Lipid segregation explains selective tox- Daly, D.J. Craik, Phosphatidylethanolamine binding is a conserved feature of
icity of a series of fragments derived from the human cathelicidin LL-37, Antimicrob. cyclotide–membrane interactions, J. Biol Chem. 287 (2012) 33629–33643.
Agents Chemother. 53 (2009) 3705–3714. [63] S.T. Henriques, D.J. Craik, Importance of the cell membrane on the mechanism of ac-
[35] R.M. Epand, S. Rotem, A. Mor, B. Berno, R.F. Epand, Bacterial membranes as predic- tion of cyclotides, ACS Chem. Biol. 7 (2012) 626–636.
tors of antimicrobial potency, J. Am. Chem. Soc. 130 (2008) 14346–14352. [64] M.D. Collins, D. Jones, Distribution of isoprenoid quinone structural types in bacteria
[36] R.M. Epand, K. D'Souza, B. Berno, M. Schlame, Membrane curvature modulation of and their taxonomic implication, Microbiol. Rev. 45 (1981) 316–354.
protein activity determined by NMR, Biochim. Biophys. Acta 1848 (2015) 220–228. [65] E. Lemma, G. Unden, A. Kroger, Menaquinone is an obligatory component of the
[37] J.C. Bozelli Jr., E.T. Sasahara, M.R. Pinto, C.R. Nakaie, S. Schreier, Effect of head group chain catalyzing succinate respiration in Bacillus subtilis, Arch. Microbiol. 155
and curvature on binding of the antimicrobial peptide tritrpticin to lipid mem- (1990) 62–67.
branes, Chem. Phys. Lipids 165 (2012) 365–373. [66] S.K. Farrand, H.W. Taber, Changes in menaquinone concentration during growth
[38] D. Koller, K. Lohner, The role of spontaneous lipid curvature in the interaction of and early sporulation in Bacillus subtilis, J. Bacteriol. 117 (1974) 324–326.
interfacially active peptides with membranes, Biochim. Biophys. Acta 1838 (2014) [67] A. Das, J. Hugenholtz, H. Van Halbeek, L.G. Ljungdahl, Structure and function of a
2250–2259. menaquinone involved in electron transport in membranes of Clostridium
[39] B.S. Perrin Jr., A.J. Sodt, M.L. Cotten, R.W. Pastor, The curvature induction of surface- thermoautotrophicum and Clostridium thermoaceticum, J. Bacteriol. 171 (1989)
bound antimicrobial peptides piscidin 1 and piscidin 3 varies with lipid chain 5823–5829.
length, J. Membr. Biol. 248 (2015) 455–467. [68] H. Hamamoto, M. Urai, K. Ishii, J. Yasukawa, A. Paudel, M. Murai, T. Kaji, T. Kuranaga,
[40] K. Matsuzaki, K. Sugishita, N. Ishibe, M. Ueha, S. Nakata, K. Miyajima, R.M. Epand, Re- K. Hamase, T. Katsu, J. Su, T. Adachi, R. Uchida, H. Tomoda, M. Yamada, M. Souma, H.
lationship of membrane curvature to the formation of pores by magainin 2, Bio- Kurihara, M. Inoue, K. Sekimizu, Lysocin E is a new antibiotic that targets
chemistry 37 (1998) 11856–11863. menaquinone in the bacterial membrane, Nat. Chem. Biol. 11 (2015) 127–133.
[41] S. Bobone, D. Roversi, L. Giordano, Z.M. De, F. Formaggio, C. Toniolo, Y. Park, L. Stella, [69] O.H. Morand, J.D. Aebi, H. Dehmlow, Y.H. Ji, N. Gains, H. Lengsfeld, J. Himber, Ro 48-
The lipid dependence of antimicrobial peptide activity is an unreliable experimental 8.071, a new 2,3-oxidosqualene:lanosterol cyclase inhibitor lowering plasma cho-
test for different pore models, Biochemistry 51 (2012) 10124–10126. lesterol in hamsters, squirrel monkeys, and minipigs: comparison to simvastatin, J.
[42] B. Bechinger, The SMART model: soft membranes adapt and respond, also transient- Lipid Res. 38 (1997) 373–390.
ly, in the presence of antimicrobial peptides, J. Pept. Sci. 21 (2015) 346–355. [70] R.K. Dhiman, S. Mahapatra, R.A. Slayden, M.E. Boyne, A. Lenaerts, J.C. Hinshaw, S.K.
[43] J. Pogliano, N. Pogliano, J.A. Silverman, Daptomycin-mediated reorganization of Angala, D. Chatterjee, K. Biswas, P. Narayanasamy, M. Kurosu, D.C. Crick,
membrane architecture causes mislocalization of essential cell division proteins, J. Menaquinone synthesis is critical for maintaining mycobacterial viability during ex-
Bacteriol. 194 (2012) 4494–4504. ponential growth and recovery from non-replicating persistence, Mol. Microbiol. 72
[44] T.T. Tran, D. Panesso, N.N. Mishra, E. Mileykovskaya, Z. Guan, J.M. Munita, J. Reyes, L. (2009) 85–97.
Diaz, G.M. Weinstock, B.E. Murray, Y. Shamoo, W. Dowhan, A.S. Bayer, C.A. Arias, [71] J.G. Hurdle, A.J. O'Neill, I. Chopra, R.E. Lee, Targeting bacterial membrane function: an
Daptomycin-resistant Enterococcus faecalis diverts the antibiotic molecule from the underexploited mechanism for treating persistent infections, Nat. Rev. Microbiol. 9
division septum and remodels cell membrane phospholipids, MBio. 4 (2013) (2011) 62–75.
e00281-e00213. [72] W. Zhao, T. Rog, a.A. Gurtovenko, I. Vattulainen, M. Karttunen, Role of
[45] S. Jordan, M.I. Hutchings, T. Mascher, Cell envelope stress response in Gram-positive phosphatidylglycerols in the stability of bacterial membranes, Biochimie 90
bacteria, FEMS Microbiol. Rev. 32 (2008) 107–146. (2008) 930–938.
[46] M. MISIEK, O.B. FARDIG, A. GOUREVITCH, D.L. JOHNSON, I.R. HOOPER, J. LEIN, [73] W. Dowhan, Molecular basis for membrane phospholipid diversity: why are there
Telomycin, a new antibiotic, Antibiot. Annu. 5 (2001) 852–855. so many lipids? Annu. Rev. Biochem. 66 (1997) 199–232.
[47] J.C. Sheehan, D. Mania, S. Nakamura, J.A. Stock, K. Maeda, The structure of telomycin, [74] M. Debono, M. Barnhart, C.B. Carrell, J.A. Hoffmann, J.L. Occolowitz, B.J. Abbott, D.S.
J. Am. Chem. Soc. 90 (1968) 462–470. Fukuda, R.L. Hamill, K. Biemann, W.C. Herlihy, A21978C, a complex of new acidic
[48] C. Fu, L. Keller, A. Bauer, M. Brönstrup, A. Froidbise, P. Hammann, J. Herrmann, G. peptide antibiotics: isolation, chemistry, and mass spectral structure elucidation, J.
Mondesert, M. Kurz, M. Schiell, D. Schummer, L. Toti, J. Wink, R. Muller, Biosynthetic Antibiot. 40 (1987) 761–777.
 R.M. Epand et al. / Biochimica et Biophysica Acta 1858 (2016) 980–987 987

[75] D. Jung, A. Rozek, M. Okon, R.E.W. Hancock, Structural transitions as determinants of [80] J.S. Gunn, Bacterial modification of LPS and resistance to antimicrobial peptides, J.
the action of the calcium-dependent antibiotic daptomycin, Chem. Biol. 11 (2004) Endotoxin Res. 7 (2001) 57–62.
949–957. [81] R.F. Epand, J.E. Pollard, J.O. Wright, P.B. Savage, R.M. Epand, Depolarization, bacterial
[76] A.B. Hachmann, E. Sevim, A. Gaballa, D.L. Popham, H. Antelmann, J.D. Helmann, Re- membrane composition, and the antimicrobial action of ceragenins, Antimicrob.
duction in membrane phosphatidylglycerol content leads to daptomycin resistance Agents Chemother. 54 (2010) 3708–3713.
in Bacillus subtilis, Antimicrob. Agents Chemother. 55 (2011) 4326–4337. [82] K. Goldberg, H. Sarig, F. Zaknoon, R.F. Epand, R.M. Epand, A. Mor, Sensitization of
[77] C.M. Ernst, A. Peschel, Broad-spectrum antimicrobial peptide resistance by MprF- Gram-negative bacteria by targeting the membrane potential, FASEB J. 27 (2013)
mediated aminoacylation and flipping of phospholipids, Mol. Microbiol. 80 (2011) 3818–3826.
290–299. [83] F. Zaknoon, K. Goldberg, H. Sarig, R.F. Epand, R.M. Epand, A. Mor, Antibacterial prop-
[78] A. Peschel, R.W. Jack, M. Otto, L.V. Collins, P. Staubitz, G. Nicholson, H. Kalbacher, erties of an oligo-acyl-lysyl hexamer targeting Gram-negative species, Antimicrob.
W.F. Nieuwenhuizen, G. Jung, A. Tarkowski, K.P. van Kessel, J.A. van Strijp, Staphylo- Agents Chemother. 56 (2012) 4827–4832.
coccus aureus resistance to human defensins and evasion of neutrophil killing via [84] L. Ejim, M.A. Farha, S.B. Falconer, J. Wildenhain, B.K. Coombes, M. Tyers, E.D. Brown,
the novel virulence factor MprF is based on modification of membrane lipids with G.D. Wright, Combinations of antibiotics and nonantibiotic drugs enhance antimi-
l-lysine, J. Exp. Med. 193 (2001) 1067–1076. crobial efficacy, Nat. Chem. Biol. 7 (2011) 348–350.
[79] C.M. Ernst, S. Kuhn, C.J. Slavetinsky, B. Krismer, S. Heilbronner, C. Gekeler, D. Kraus, S. [85] G. Kaneti, H. Sarig, I. Marjieh, Z. Fadia, A. Mor, Simultaneous breakdown of multiple
Wagner, A. Peschel, The Lipid-modifying Multiple Peptide Resistance Factor Is an antibiotic resistance mechanisms in S. aureus, FASEB J. 27 (2013) 4834–4843.
Oligomer Consisting of Distinct Interacting Synthase and Flippase Subunits, MBio
6, 2015.