The worrisome increase in bacterial resistance to current generation of most efficacious antibiotics is a rapidly expanding global health problem. Currently, at least 700,000 people die each year due to antimicrobial drug-resistant pathogens, which is expected to rise exponentially to 10 million by 2050, which could displace more than 20 million people below the poverty line by 2030
[1][2][3]. The present United Nation and World Health Organization reports also cite the dire need for innovative and sustainable methods for the development of novel drugs to overcome antimicrobial resistance diseases
[4][5]. Antimicrobials either target the cell envelope or cytosolic components of bacteria to inhibit essential cellular events. While lipophilic antibiotics (e.g., macrolides) can diffuse directly through the bacterial plasma membrane (PM), low molecular weight hydrophilic antibiotics (e.g., β-lactams) having size exclusion limit < 600 Da use porins to reach their intracellular targets
[6]. The cell envelope in Gram-negative bacteria is made of an outer membrane (OM), peptidoglycan cell wall (CW), and inner membrane (IM) (
Figure 1). The OM is heterogeneous and asymmetric with an outer leaflet consisting of lipopolysaccharide (LPS) (or lipooligosaccharide) (LOS) and an inner leaflet consisting of phospholipids (PLs). However, the IM is a symmetric phospholipid (PL) bilayer. The cell envelope of Gram-positive bacteria such as
S.
aureus has a single plasma membrane made of PLs followed by a thick cell wall with no periplasmic space. Efficient permeation of antimicrobials through cell envelope is rate limiting in their activity
[7]. As the cell wall is porous and permeable to most antibiotics, permeability through the OM and IM are rate limiting factors in determining their cellular bioavailability in Gram-negative bacteria. However, in Gram-positive bacteria such as
S.
aureus, permeability through plasma membrane is crucial for antimicrobial efficacy
[8].
2. Role of OM in Gram-Negative MDR Bacteria
Four of the “ESKAPE” group of drug-resistant human pathogens, e.g., Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter sp. are Gram-negative. The OM of these bacteria is a sophisticated macromolecular assembly of LPSs, PLs, porins, and other OM-associated proteins. LPS, a peculiar glycolipid unique to Gram-negative bacteria is the hallmark of OM that consists of a glucosamine-based lipid (lipid A) attached to a core oligosaccharide followed by a distal polysaccharide chain (Figure 2). The OM is asymmetric with LPS residing exclusively to the exoplasmic (outer) leaflet and PLs residing to the cytosolic (inner) leaflet. The strain-specific immune-reactivity and virulence of most Gram-negative bacteria is because of the immensely variable composition and organization of the sugar moieties and fatty acyl tails of LPS molecules. This incredible variability in LPS that is attained through unique organization and adaptability to multiple secondary covalent modifications plays a key role in the multi-drug-resistant (MDR) of Gram-negative bacteria.
Figure 2. Lipid A modification in Gram-negative bacteria leading to altered membrane properties. Gram-negative bacteria modify phosphate groups with (a) phosphoethanolamine, (b) glucosamine, and (c) aminoarabinose. The fatty acyl tails add (d) positively charged glycine, (e) diglycine moieties, or (f) palmitate at the 3′-position of the glucosamine disaccharide. Fatty acyl chains such as laureate (g) and hydroxy–myristate (h) are added at positions 2 and 2′, respectively. Arrows indicate the position of the groups modified.
2.1. Asymmetric Organization of LPS in Gram-Negative OM LEADS to CAMP Resistance
The asymmetric distribution of LPS in the outer leaflet of Gram-negative OM provides an intrinsic resistance against most bulky hydrophobic and cationic antimicrobials. LPS has a nearly conical geometry with its membrane-embedded hexa-acylated lipid A and a highly variable bulky polysaccharide chain that extends ~30 A° beyond the plane of the OM (
Figure 2). The LPS layer is a robust and continually evolving antibiotic barrier leading to bacterial MDR and virulence
[25]. An in vitro reconstituted model of the Gram-negative OM suggests that the trans-bilayer asymmetry is crucial for stable interaction and insertion of cationic peptides
[26].
The asymmetry of Gram-negative OM is acutely regulated by a set of three different proteins that work in concert, namely, the OM phospholipase A
2 (PldA), the LPS, palmitoyltransferase (PagP), and the maintenance of outer membrane lipid asymmetry (Mla) system
[27]. Both PldA and PagP are involved in degradation or removal of mislocalized PLs in the outer leaflet of OM. Mla is a six-component lipid flippase complex conserved in all Gram-negative bacteria that facilitates the ATP-dependent retrograde transport of mislocalized PLs from the outer leaflet of OM to the IM (
Figure 1). Mla complex contributes to the the broad-spectrum antibiotic resistance in several pathogens, most prominently in
A.
baumannii [28]. MlaA, the subunit that is localized to the inner leaflet of OM, has the shape of a doughnut that only allows the removal of mislocalized lipids in the outer OM leaflet without affecting the inner leaflet lipids
[29]. Structural elucidation of the Mla complex reveals that PLs from MlaA are translocated to MlaC that shuttles through the peptidoglycan layer and periplasm to deliver the PLs to the IM-localized MlaBDEF complex. PLs from the binding site of MlaC are transferred to the bowl-shaped domain formed by the hexameric D subunits of the MlaBDEF complex on its periplasmic site. PLs are then translocated to the inner-IM leaflet through a gated transmembrane tunnel formed by the helices of D and E subunits. The hydrophobic side chains of Leu
153/Leu
154 directed towards the lumen of the tunnel plausibly facilitates the movement of the non-polar fatty acyl tails of PLs. However, movement of the polar head group is facilitated by the polar cationic side chains of R14, R47, and R234 located toward the cytosolic end of the tunnel (
Figure 1)
[30]. PLs from the bifurcated cytosolic outlet of MlaBDEF complex diffuse into the inner-IM leaflet. PLs localized to the inner-IM leaflet is then equilibrated with the outer leaflet by ATP-independent biogenic membrane flippases
[31]. However, it remains to be explained how the PLs is transferred between MlaA, MlaC and MlaD. Although the Mla complex was initially proposed to be involved in the retrograde transport of PLs from the outer OM leaflet to the inner IM leaflet, it is becoming increasingly clear that the same complex also facilitates the anterograde translocation of PLs from the IM to the OM in LPS-deficient strains of
A.
baumanii [32]. Mla in OM of Gram-negative bacteria is essential for the maintenance of the barrier function of OM to protect the bacterium againstthe noxious compounds. Hence, Mla inhibitors could be explored as a potent antibacterial that may act alone or in combination
[33][34].
2.2. Loss of OM Asymmetry in Gram-Negative Bacteria
The negatively charged LPS of Gram-negative OM is the substrate for attachment of CAMPs such as polymyxins to bacteria. The cationic antibiotic polymyxin E (colistin) inhibits divalent cation-mediated bridging between adjacent LPS molecules, leading to their loss in the outer-OM leaflet
[35]. In
A. baumanii, a mutant non-functional
lpxA is responsible for the loss of LPS which otherwise acts as the target for colistin. It also exhibits increased sensitivity towards other clinically relevant antibiotics such as Cefepime, Teichoplanin, and Azithromycin
[36]. However, repeated exposure of
A. baumanii to colistin stimulates loss of LPSs in outer-OM leaflet, leading to their replacement with PLs resulting in loss of OM asymmetry
[28]. In wild-type
A. baumannii, combined action of the periplasmic lipid-shuttle complex Mla and phospholipase A (PldA) works to remove mislocalized PLs in OM and maintains its asymmetry. However, colistin resistant variants of
A. baumannii are nonfunctional mutants of
Mla and
PldA that compensate for the loss of LPS in OM outer leaflet by stably replacing them with PLs
[25]. These LPS-deficient phenotypes of
A. baumannii demonstrate enhanced resistance to colistin.
However, interaction of polymyxins with PLs in the inner IM leaflet is not yet completely understood. Recent investigation reveals that MCR-1-mediated colistin resistance in
Escherichia coli is due to modified LPSs at the cytoplasmic membrane (CM) rather than OM
[37]. LPS that is synthesized in the inner IM leaflet is flipped to the outer leaflet from which it is transported to the OM outer leaflet by a LPS transport complex known as LptABCDEFG (Li et al., 2020). Hence, murepavadin, an inhibitor of LPS transport machinery that leads to accumulation of LPS in CM, kills
E. coli by lysing CM
[38].
2.3. Covalent Modification of LPS in MDR
Bulky CAMPs such as polymyxin B (MW = 1.2 kDa) with size > 600 Da target bacterial OM, leading to its disruption resulting in cell lysis
[39][40]. Some Gram-negative bacteria like
E. coli can acquire antibiotic resistance by modifying the sugar content in the OM
[41]. OM of many Gram-negative bacteria such as
E.
coli,
S.
enterica,
S.
typhimurium, and
K.
pneumoniae modify their LPS with additional functional groups such as fatty acyl chains (e.g., palmitate and S-2-hydroxymyristate), phosphoethanolamine (pEtN), 4-amino-4-deoxy-L-arabinose (L-Ara4N), and glycosyl groups to enhance their resistance against membrane-targeted CAMPs such as polymyxins
[42] (
Figure 2). Although LPS can have varying number of sugar units, studies indicate that the core oligosaccharide portion, especially the first heptose in the LPS, is critical for conferring antibiotic resistance in gram-negative bacteria
[43].
Enzymes that modify lipid A are regulated by two conserved component systems (TCS), PmrAB and PhoPQ, in response to specific environmental signals. PmrAB is activated in response to CAMPs, high Fe
3+ and acidic pH, whereas PhoPQ is activated in response to divalent cations (e.g., Ca
2+/Mg
2+) or CAMPs. PmrAB activates eptA (also known as pmrC) and arabinose (
arn) operon that encode pEtN and L-Ara4N transferases, respectively. PhoPQ phosphorylates the transcription factors resulting in transcription of PagL (only in
Salmonella) and PagP that, respectively, remove or add fatty acyl groups to lipid A. In addition, PhoPQ directly activates
arn expression in
Klebsiella and
Yersinia sp.
[44]. Both PhoPQ and PmrAB are interdependent and are connected through PmrD that binds to phospho-PmrA to prevent B-mediated dephosphorylation of transcription factors. Constitutive expression of PmrA dependent genes that increases modification of lipid A with pEtN and l-Ara4N, resulting in MDR of Gram-negative bacteria.
2.3.1. Fatty Acylation of Lipid A
Lipid A, the OM-embedded portion of LPS, also termed as endotoxin, is responsible for most of the pathophysiological effects associated with Gram-negative sepsis. Lipid A potently activates the host innate immune system leading to secretion of CAMPs, cytokines, clotting factors, and immunostimulatory molecules. It is a ubiquitous component of Gram-negative OM, making its modification a conserved resistance mechanism against most membrane-targeted antimicrobials. The length, number, and distribution of acyl chains are the key factors contributing to the biological activity of lipid A
[45]. The hexa-acylated lipid A can be modified by addition of a palmitate in acyloxyacyl linkage at position 2 of hexosamine sugars to produce hepta-acylated lipid A (
Figure 2). The hepta-acylated lipid A enhances the resistance of
A.
baumannii,
E. coli and
Salmonella to the last resort of CAMPs (e.g., colistin) by reinforcing the cell surface LPS
[46]. Enhanced acylation of lipid A (PagP-dependent or independent) is a general process reported in
E. coli,
Y. enterocolitica,
B. pertussis, and
A. baumannii. PagP cleaves the mislocalized PL in OM, restoring the outer leaflet composition, thereby increasing the bacterial resistance to aminoglycosides by 4-fold
[47]. Secondary acylation of lipid A with laureate and myristate in
K. pneumoniae increases its resistance against polymixin B and colistin. It also increases the virulence limit of bacterial pathogens against host immune response and survival on the desiccated surfaces of medical appliances
[48].
Hepta-acylation of lipid A in
E. coli and
Salmonella is regulated by the gene PagP. However,
A.
baumannii uses the two-component system LpxL and LpxM for lipid A acylation. In
E. coli, PagP-dependent lipid A palmitoylation increases the hydrophobic van der Waals forces of the LPS layer that prevent CAMP insertion. Hepta-acylation of lipid A protects
Salmonella from vertebrate CAMPs such as C18G, an α-helical membranolytic peptide
[49]. In
E.
coli and
S.
enterica, PagP transfers a palmitate from phosphatidylethanolamine to the R-2-hydroxymyristate of lipid A, leading to generation of hepta-acylated lipid A. However, the position of the palmitoylation may vary in different bacteria. PagP palmitoylates the
N-linked
R-3-hydroxymyristate chain at position 2 of lipid A in Enterobacteria
[50]. In
Bordetella, PagP palmitoylates the O-linked chains at position C–3′, C–2 and C–3 of the hexosamine residue in lipid A, which increases membrane rigidity.
Pseudomonas aeruginosa palmitoylates lipid A at position 3′
[51][52]. A
pagP homolog from
Legionella pneumophila was identified as
rcp (resistance to cationic antimicrobial peptides)
[53]. Host immune suppression is a key virulence strategy used by various drug-resistant bacteria. In
Salmonella enterica, lipid A palmitoylation impairs the production of CAMPs through the suppression of TLR4 pathway
[54]. In
Salmonella, myristoylation of lipid A increases its resistance against polymyxin. This secondary acylation of lipid A is important for the addition of L-Ara4N to its phosphate groups, leading to a twofold resistance against polymyxins
[55]. Asymmetric Langmuir model bilayers of outer leaflets and inner leaflets, consisting of LPSs and PLs, respectively, demonstrate that packing of the hydrophobic fatty acyl tails in LPS is the primary determinant of polymyxin B-induced OM disruption
[56]. The OM is made rigid by divalent cation interactions, which raises the transition temperature (Tm) of the membrane
[56].
As amide linkages are more rigid, planar, and stereochemically constrained, an increased number of amide-linked acyl chains in lipid A enhances membrane rigidity. For instance, the addition of three amide-linked acyl chains and only one ester-linked acyl chain instead of two ester- and two amide-linked acyl chains to lipid A in
C. jejuni increases membrane rigidity. Accordingly, additional amide linkage influences the biological activity of lipid A by altering TLR4 response and enhanced CAMP resistance
[57]. The palmitoylation of lipid A allows for increased hydrophobic interactions between neighboring LPS molecules, leading to increased resistance to membranolytic antimicrobials
[58].
2.3.2. Amino Glycosylation
Decoration of lipid A with cationic or zwitterionic groups such as L-4-aminoarabinose and phosphoethanolamine (PE) that shield the negative charge on phosphates leads to drug resistance in Gram-negative bacteria. First, it reduces binding of CAMPs to the outer leaflet of OM by reducing the anionic charge on the OM lipids. Second, it reduces the requirement of divalent cations (e.g., Ca
2+ and Mg
2+) that is essential for the intermolecular bridging between LPSs, leading to its stabilization
[59]. In members of enterobacteriaceae such as
E.
coli and
Salmonella, the PmrAB two-component system directly activates L-Ara4N biosynthesis, leading to addition of the cationic amine moiety to lipid A. However, some members of enterobacteriaceae family such as
Enterobacter cloacae use the PhoPQ two-component system that contributes to colistin heteroresistance. PhoPQ directly binds the ArnB promoter to activate the L-Ara4N biosynthesis, its covalent attachment to Lipid A in IM, and transport to the outer leaflet of OM
[44]. Although several factors, such as low Mg
2+ concentration, low pH, presence of antibiotics, and high temperature are proven to modify lipid A, de novo expression of colistin resistance conferring peptides (Dcr) in
E.
coli have been proven to activate the
PmrAB conferring colistin resistance
[60]. In
E.
coli and
K.
pneumoniae strains, the mobile colistin-resistant gene 1 (mcr1) regulates collistin resistance to overcome destabilization of the bacterial outer membrane and prevents cell lysis
[61]. However, absence of
mcr1 in colistin and carbapenem resistant strains of
K.
pneumonia with mutations in several
Arn genes (e.g.,
ArnA_DH/FT, UgdH, ArnC and
ArnT) reveals altered L-Ara4N biosynthetic pathways. The L-Ara4N precursor, undecaprenyl phosphate-α-L-Ara4-formyl-N is synthesized on the cytosolic side of IM from the undecaprenyl phosphate and UDP-L-Ara4-Formyl-N that is flipped to the periplasmic side. However, the flippase protein that facilitates this essential biosynthetic step remains elusive. Once on the periplasmic side of IM, the undecaprenyl phosphate-α-L-Ara4-formyl-N is deformylated and the L-Ara4-N moiety is transferred to lipid A of LPS
[62]. Transfer of L-Ara4N to lipid A to the inner IM leaflet is reversibly catalyzed by the IM-associated integral protein ArnT
[63]. ArnT exhibit a conserved topology in both
Burkholderia cenocepacia and
Salmonella enterica with its N-terminal domain forming 13 transmembrane helices and a globular periplasmic C-terminal domain
[64]. From the IM, L-Ara4N-LPS is vectorially translocated to the OM through the periplasm that is catalyzed by the ATP-dependent ABC-transporter. This translocation requires an inter-membrane protein bridge, which connects both IM and OM that facilitates the unidirectional transport of L-Ara4N-LPS to the OM
[65]. L-Ara4N is cationic and results in lowering of the negative charge on LPS leading to its reduced binding affinity to CAMPs. As the LPS-modification enhances pathogenicity and innate immunity evasion, ArnT that catalyzes transfer of L-Ara4N to lipid A can be a virulence factor which is an ideal target for development of membrane-targeted therapeutics. Overexpression of
pmrA (polymyxin resistance gene
A) that modifies lipid A with L-Ara4N results in 75% lowering of polymixin binding
[66]. Similarly, deacetylation, another key modification of lipid A, is facilitated by several membrane-associated enzymes such as PagL, LpxR, NaxD, and YdjC. The glycylation, another lipid A modification that adds glycine or diglycine residues is reported to enhance drug resistance in Gram-negative bacteria. Phosphorylation or phosphate modification leads to secondary alteration of the core oligosaccharide structure and acylation status of lipid A, resulting in bacterial resistance against CAMPs like Polymixin B and Colistin
[67].