Gram-negative bacteria are responsible for a large proportion of antimicrobial-resistant infections in humans and animals. Among this class of bacteria are also some of the most successful environmental organisms. Part of this success is their adaptability to a variety of different niches, their intrinsic resistance to antimicrobial drugs and their ability to rapidly acquire resistance mechanisms. These mechanisms of resistance are not exclusive and the interplay of several mechanisms causes high levels of resistance. In this review, we explore the molecular mechanisms underlying resistance in Gram-negative organisms and how these different mechanisms enable them to survive many different stress conditions.
Infections caused by antimicrobial-resistant bacteria are difficult or even impossible to treat and they are becoming increasingly common and are causing a global health crisis. Recently, a few new antibiotics have been approved for use against Gram-positive organisms such as Staphylococcus aureus . However, infections caused by Gram-negative pathogens prove much harder to treat due to their very high intrinsic drug resistance. Hence, the only new class of antibiotic discovered in the last 30 years, teixobactin, does not act against Gram-negative bacteria . Yet, Gram-negative bacteria pose an imminent threat to our health. According to the Medicines Agency (EMEA), approximately two-thirds of deaths caused by antibiotic-resistant bacteria in Europe are due to infections by Gram-negatives . Gram-negative bacteria are also responsible for 45–70% of ventilator-associated pneumonia (VAP) cases, 20–30% of catheter-related bloodstream infections and commonly cause other infections associated with intensive care units such as surgical site or urinary tract infections (UTIs) .
Nosocomial (hospital-acquired) Gram-negative pathogens such as Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, Escherichia coli and the emerging pathogen Burkholderia cepacia display high levels of antibiotic resistance [5,6]. Treatment of these life-threatening infections is becoming increasingly difficult, in some cases resulting in a complete lack of viable treatment options. K. pneumoniae, an important member of the family Enterobacteriaceae, is one of the major causes of nosocomial infections such as pneumonia, bloodstream infections, neonatal infections and infections in intensive care unit patients . P. aeruginosa, whose natural niche is soil, can also easily infect humans and cause life-threatening opportunistic infections, which primarily target immunocompromised individuals such as HIV patients, burn victims and cancer patients, with a fatality rate of 50% for the latter two cases . P. aeruginosa is also a predominant cause of infections in indwelling devices  and chronic infection by P. aeruginosa is the main cause of mortality in patients suffering from cystic fibrosis. Similarly, B. cepacia an organism that is used as a biocontrol agent in the environment, causes life-threatening infections in immunocompromised patients and in cystic fibrosis sufferers .
Drug-resistant Gram-negative organisms also cause community-acquired infections in otherwise healthy people. P. aeruginosa causes serious eye infections that could lead to blindness and an ear infection commonly known as ‘swimmer’s ear’ and B. cepacia is associated with ‘swamp foot’, an infection prevalent among military personnel . Another highly drug-resistant Gram-negative organism, Neisseria gonorrhoeae that causes the sexually transmitted disease gonorrhoea, is the second most prevalent community-acquired infection . Highly drug-resistant, food-borne, Gram-negative pathogens are also a major concern as was shown with the recent outbreak of multidrug-resistant Salmonella heidelberg in the United States .
These hospital- and community-acquired organisms, while able to cause a diverse range of infections, all have a common attribute; they exhibit extremely high levels of antimicrobial resistance (AMR). In the case of K. pneumoniae, acquisition of multiple β-lactamase enzymes has allowed resistance to develop against penicillins, cephalosporins and carbapenems. Carbapenems are a last resort treatment option for K. pneumoniae as well as other multiple-drug resistant Gram-negative pathogens. However, resistance to carbapenems has been reported throughout the world, with some countries reporting >50% of K. pneumoniae infections as impervious to all antibiotic treatment options . Similarly, due to high levels of AMR only one class of antibiotics, the cephalosporins, are still active against N. gonorrhoeae . Conversely, P. aeruginosa and B. cepacia are innately resistant to many antibiotics due to the selective permeability of the outer membrane (OM) and a large repertoire of drug efflux proteins . Additionally, P. aeruginosa has a large genome and has the ability to easily acquire foreign genes (including resistance genes) to allow survival in almost any environment . This adaptability also means that P. aeruginosa could be a potential reservoir of resistance, connecting the environment with resistance in the clinic.
AMR can be intrinsic, adaptive or acquired. Intrinsic resistance refers to the inherent properties of a microorganism that limit the action of antimicrobials e.g. the permeability barrier of the OM and constitutively expressed drug efflux pumps. Adaptive resistance is the ability of an organism to adapt and survive many stress conditions by rapidly altering their transcriptomes in response to environmental cues. These adaptations could help the organisms to overcome nutrient limitation, survive antibiotic stress etc. Acquired resistance occurs when an antibiotic-sensitive organism becomes resistant through the acquisition of genes or as a result of mutations. An example of acquired resistance is the spread of plasmids encoding β-lactamase genes, which allow the organism to acquire resistance to β-lactam antibiotics [17–20]. The boundaries among these mechanisms are somewhat blurred as drug efflux pumps, for example contribute to both intrinsic and acquired resistance and porins can be involved in both adaptive and acquired resistance.
In this review, we explore the intrinsic, adaptive and acquired resistance mechanisms that afford Gram-negative organisms their incredible ability to survive a plethora of external stresses. P. aeruginosa is an excellent example of an organism with a high level of intrinsic and adaptive resistance and thus we will discuss these resistance mechanisms with reference to P. aeruginosa. The molecular mechanisms behind acquired resistance to our most used class of antibiotics, the β-lactams i.e. drug inactivation, drug efflux and reduced drug entry, are examined. Finally, other acquired mechanisms of resistance such as drug alteration and target modification will also be discussed.
All Gram-negative bacteria are intrinsically more resistant than Gram-positives due to the OM, which acts as a permeability barrier and prevents antibiotics from reaching their targets . The OM is an asymmetric bilayer composed of lipopolysaccharides (LPS, outer leaflet) and phospholipids (inner leaflet) . LPS typically consists of lipid A, a short-core oligosaccharide and an O-antigen that may be a long polysaccharide. Some Gram-negative pathogens, such as members of the genera Neisseria and Haemophilus as well as Campylobacter jejuni contain lipooligosaccharides (LOS) instead of LPS. LOS share similar lipid A structures with LPS, but lacks the O-antigen units; instead the oligosaccharide structures are limited to ten saccharide units . Passive diffusion of hydrophobic compounds is relatively slow, while small hydrophilic compounds gain access through porins that are scattered throughout the OM. Larger hydrophilic antibiotics are effectively excluded. For instance, the glycopeptide antibiotic vancomycin that is the treatment of choice against methicillin-resistant S. aureus (MRSA), is completely infective against Gram-negative bacteria . Vancomycin is a relatively large antibiotic and is therefore ineffective against Gram-negative bacteria due to the inability to breach the OM permeability barrier. In P. aeruginosa, the permeability barrier is further reinforced by the lack of non-specific porins in the OM through which antimicrobials can permeate. In addition, P. aeruginosa also expresses a range of drug efflux pumps that confer resistance against many different classes of antibiotics and additionally, against biocides used in disinfectants [25–29]. Consequently, pan-drug-resistant strains of P. aeruginosa have been observed that are able to grow in 1% chlorhexidine disinfectant solutions that are used in hospitals . A more detailed description of the role of porins and efflux pumps in antibiotic resistance is given later in this review under ‘acquired mechanisms of resistance’.
Another mechanism of intrinsic antibiotic resistance is the lack of the target for the antibiotic. An example is the lipopeptidolactone, daptomycin, which is active against MRSA, vancomycin-resistant S. aureus (VRSA) and vancomycin-resistant enterococci (VRE), yet has no activity against Gram-negative bacteria . Daptomycin targets the cytoplasmic membrane where it inserts using a Ca2+-mediated process and subsequently causes membrane depolarization. The cytoplasmic membrane of Gram-negative bacteria has a significantly lower proportion of anionic phospholipids compared with that of Gram-positive bacteria. This difference in membrane composition reduces the efficiency of the Ca2+-mediated insertion of daptomycin into the cytoplasmic membrane and hence lowers bactericidal activity of the drug .
P. aeruginosa has many characteristics that account for its enormous ecological success in different environments. It has a very large genome that gives it adaptability to many different situations. It has extremely low nutritional requirements and are notorious for its ability to grow in distilled water. For example, withholding iron is a common defence of host organisms against pathogens. In response, pathogens excrete high affinity iron chelating molecules (siderophores) that can ‘steal’ iron from the storage proteins of its host . Most pathogens synthesize one or more of these siderophores. P. aeruginosa produces two siderophores itself, but its genome contains the genes for 34 different siderophore-acquisition porins, allowing the acquisition of siderophores from many different organisms . In this manner, P. aeruginosa is able to acquire the iron necessary for cellular viability, but without the metabolic expense of synthesizing the siderophores in the first place . P. aeruginosa also has the cunning ability of being able to use parabens, the most commonly used preservatives in non-sterile pharmaceutical products as a source of nutrients. Hence, it is a common contaminant of pharmaceutical products . Additional mechanisms of adaptive resistance are facilitated by the ability of bacteria to change their behaviour in response to extracellular environmental conditions. This is seen predominantly in the development of persister cells and biofilms. Persister cells are a subpopulation of cells that enter a quiescent state and stop actively growing, hence reducing the ability of antibiotics to inhibit cellular proteins required for bacterial growth . Conversely, biofilms are entire communities of microorganisms, attached to a surface and encased by a polymer matrix, which can render the bacteria up to a 1000 times more resistant to antibiotics compared with the planktonic organisms [37,38]. Biofilms are also much more resistant to biocides, sheer forces and host immune defences , therefore allowing the organism to persist in extremely harsh environments. Biofilm formation is largely driven by the excretion of small signalling molecules that allow microorganisms to communicate. This process is called quorum sensing. If, for example the first signal could be roughly translated as: “Is there anybody out there?”, the subsequent detection of an appropriate cell density (a quorum) would trigger a change in the message to: “Let’s settle down and form a community”. At this point, gene expression is altered and the bacteria undergo a remarkable change from the free swimming, planktonic form typical of an acute infection, to the biofilm mode found in chronic and device-related infections [40,41]. Bacteria within a biofilm are much more resistant to antimicrobials compared with free swimming bacteria .
In addition to intrinsic resistance, Gram-negative pathogens also display specific acquired molecular mechanisms of resistance to antimicrobials [17,18,42]. These can be classified as (i) antibiotic modification/inactivation, (ii) antibiotic target alteration, (iii) increased antibiotic efflux and (iv) reduced antibiotic uptake. In most cases, several of these mechanisms combine to give a high level of resistance against a particular antibiotic. To illustrate this phenomenon, the resistance mechanisms utilized against β-lactam antibiotics (drug inactivation, increased efflux and reduced entry) will be described first.
β-lactams are the most frequently prescribed antibiotics. In Australia, β-lactam antibiotics accounted for 68% and 82% of the ten most commonly supplied antibiotics in the community and hospitals respectively (Figure 1). These antibiotics inhibit the transpeptidase enzymes (penicillin binding proteins; PBPs) involved in cross-linking of the peptidoglycan strands, leading to inhibition of cell wall synthesis. Under these conditions, endogenous autolytic enzymes are activated through a two-component system, VncR/S, which destabilizes the cell wall and predisposes the bacterial cell to osmotic rupture . Penicillins, carbapenems, monobactams, cephalosporins and cephamycins are the major categories of β-lactam antibiotics . Unfortunately, resistance to β-lactam antibiotics is widespread and increasing rapidly. Gram-negative organisms can be resistant to β-lactam antibiotics through three different mechanisms (Figure 2): (i) the production of β-lactamase enzymes that hydrolyse the β-lactam ring thereby inactivating the antibiotic (an example of antibiotic inactivation), (ii) the production of antibiotic efflux pumps that remove the antibiotics from the periplasm, hence reducing their concentration to subtoxic levels (drug efflux) and (iii) alteration of the pores in the OM of the bacterium to prevent the entry of the antibiotic (reduced entry). Often, all three mechanisms will be involved in high-level resistance to β-lactam antibiotics (Figure 2).
β-lactams are the most frequently prescribed antibiotics
Acquired mechanisms of resistance to β-lactam antibiotics
Antibiotic inactivation: β-lactamase enzymes catalyse hydrolysis of β-lactam antibiotics
An integral part of β-lactam antibiotics is the β-lactam ring (Figure 2A). This four-membered ring is prone to hydrolysis (and subsequent deactivation) by β-lactamase enzymes. These enzymes are highly diversified with different spectrums of activity. As we expanded the arsenal of β-lactam antibiotics, up to the fourth generation of cephalosporins and carbapenems, bacteria have adapted a variety of β-lactamases to inactivate and overcome these antibiotics. The latest β-lactamase to be discovered, NDM-1, is able to inactivate our last line of carbapenem antibiotics and is resistant to almost all β-lactams. This enzyme is thought to have originated in New Delhi and its rapid worldwide spread was precipitated by medical tourism [48,49]. In order to combat the ever-increasing issue of bacterial β-lactamases, β-lactamase inhibitors (clavulanate, sulbactam and tazobactam) are now often used in combination with β-lactam antibiotics to protect them from hydrolysis (Figure 1). Although lacking significant intrinsic antibacterial activity, these molecules are able to inhibit a number of plasmid-mediated β-lactamases. Figure 3 provides a very brief summary of some of the most problematic β-lactamases.
A selection of β-lactamase enzymes produced by Gram-negative organisms
Antibiotic efflux: drug efflux pumps lower the level of antibiotics inside the cell
Drug efflux pumps are proteins that extrude antibiotics from the bacterial cell, thereby lowering their concentration to sublethal levels (Figure 2B). An interesting feature of these efflux pumps is their ability to extrude a wide variety of structurally different compounds [27,50–53]. The clinical consequence of this substrate promiscuity is the subsequent development of multidrug resistance [27,28,54,55]. Drug efflux pumps are the first line of defence of the cell. When the organism is challenged with an antibiotic, there is a transient up-regulation of efflux pump expression, which results in a subtoxic concentration of antibiotic in the cell. This allows the cell to survive until a specific resistance mechanism is acquired. Therefore, an active efflux pump is both necessary and sufficient for the selection of novel drug-resistant mutations [56–59]. In Gram-negative bacteria, clinically relevant levels of AMR are conferred by efflux pumps of the resistance-nodulation division (RND) family [27,60,61]. These complexes are tripartite protein assemblies that span the IM, the OM and the periplasm to extrude antibiotics over the double membrane of Gram-negative organisms (Figure 2B) . The best-studied tripartite drug efflux complexes are the AcrA–AcrB–TolC and MexA–MexB–OprM transporters from E. coli and P. aeruginosa respectively. In these complexes, AcrB/MexB are the IM proteins that utilize the proton motive force to expel drugs from the cytoplasm or the periplasm . TolC/OprM are the OM proteins that allow the drug to pass to the exterior of the cell, and AcrA/MexA are the membrane fusion proteins (also known as periplasmic adaptor proteins) that connect the inner and outer membrane proteins [62,64,65].
Reduction in antibiotic uptake through altered OM permeability
As mentioned earlier, the OM of Gram-negative organisms acts as a permeability barrier and first line of defence. The diffusion of small hydrophilic antibiotics, such as β-lactams, over the Gram-negative OM is facilitated by porin proteins . Porins are characterized by a β-barrel structural motif that form a pore with a central hydrophilic region (Figure 2C). Bacterial porins can either be non-specific (diffusion porins) or exert substrate specificity. For example, the iron acquisition porin, FepA, has an additional ‘plug’ domain, which independently enhances recruitment of the specific cargo . Conversely, diffusion porins are only able to restrict cargo on the basis of size or interaction of the compound with an inwardly folded loop (loop 3) containing charged residues [22,68,69]. The properties of constitutively expressed porins are hugely significant for the intrinsic level of antibiotic resistance in Gram-negative bacteria. For example, P. aeruginosa has a much higher intrinsic resistance to a range of distinct antimicrobials than the Enterobacteriaceae. This is partly due to the fact that P. aeruginosa does not produce high permeability classical porins, but instead, expresses ‘slow’ porins with reduced diffusion rates . The extremely large genome also allows P. aeruginosa to express several specific porins that allow the diffusion of specific, small nutrients, but not the bulkier antibiotics, such as cephalosporins, which were developed to be insensitive to deactivation by β-lactamase hydrolysis .
Acquired antibiotic resistance through porins can be developed through (i) mutations that down-regulate the expression of porins (e.g. the loss of OmpF confers β-lactam resistance to E. coli), (ii) replacement of a large diameter porin with a porin of a smaller channel size (e.g. in β-lactam resistant K. pneumoniae isolates, OmpK35 is replaced with OmpK36 of much smaller channel size) and (iii) mutations that cause a modification that impairs porin function (e.g. the addition of two negatively charged amino acids in the channel-constricting loop 3 of the PenB porin of N. gonorrhoeae results in significantly reduced permeation of penicillin) (Figure 2C) [22,69,70].
There is a significant interplay between reduced OM permeability and efflux in AMR. Together, these two mechanisms give rise to resistance against many other classes of antimicrobials such as fluoroquinolones, aminoglycosides, erythromycins, choloramphenicol, tetracyclines etc. However, resistance to antibiotics is often multifactorial, relying on multiple molecular mechanisms operating simultaneously. Resistance to the aforementioned antimicrobials is conferred by reduced permeability and increased efflux combined with other mechanisms such as drug modification and target alteration.
An important example of antibiotic alteration is the modification of the aminoglycoside antibiotics. Currently, the most common clinical application of aminoglycosides is in the treatment of infections caused by aerobic Gram-negative bacteria (Enterobacteriaceae ssp, A. baumannii and P. aeruginosa) such as the treatment of uncomplicated UTIs caused by carbapenem-resistant Enterobacteriaceae ssp . Aminoglycosides exert their antibacterial activity by interfering with protein synthesis at the initiation complex level. They act by binding to the aminoacyl site of 16S rRNA within the 30S ribosomal subunit, leading to misreading of the genetic code and inhibition of translation [72,73]. However, the structure of the aminoglycosides renders them susceptible to modification by enzymes such as aminoglycoside N-acetyltransferases (AACs), aminoglycoside O-nucleotidyltransferases (ANTs) and aminoglycoside O-phosphotransferases (APHs), which can modify the antibiotic and therefore render it inactive .
Target alteration – where the target of the antibiotic is changed so that the antibiotic no longer has any activity against it – is a common mechanism of resistance against several classes of antibiotics. We will look at modification of the cellular target of fluoroquinolones as an example. Fluoroquinolones had a very obscure start as an impurity isolated during the synthesis of the antimalarial drug chloroquine . Nowadays, up to fourth-generation fluoroquinolones are used to treat infections caused by both resistant Gram-positive and Gram-negative bacteria . Epidemiological evidence suggests a strong correlation between resistance to fluoroquinolones and other most challenging resistance phenotypes. For instance, it has been shown that higher portions of extended-spectrum β-lactamase (ESBL) producing K. pneumoniae are concomitantly resistant to fluoroquinolones . These antibiotics act by interfering with DNA replication by targeting essential bacterial enzymes, specifically type II topoisomerases (gyrase and topoisomerase IV). During replication, the unwinding of the dsDNA results in supercoiling in front of the replication fork. Type II topoisomerases cut the DNA that allows it to unwind prior to religation. Fluoroquinolones interact with the DNA–topoisomerase complex, preventing the progression of the replication fork and resulting in DNA fragmentation and ultimately cell death . Mutations in the isomerase gene (most notably gyrA) result in amino acid substitutions that alter the protein structure and subsequently, the fluoroquinolone-binding affinity of the enzyme, leading to drug resistance [78,79].
Gram-negative bacteria are some of earth’s most successful organisms and can survive in countless different environments. The remarkable survival techniques utilized by these organisms have allowed them to develop mechanisms to resist all of the different classes of antibiotics that we have developed to target them. Mechanisms of resistance to these antibiotics can be broadly classified as intrinsic, adaptive or acquired resistance. However, AMR is often multifactorial, employing several mechanisms simultaneously to circumvent the activity of particular antimicrobials. Currently, we have a good comprehension of the biochemistry that underpins AMR, and this knowledge will be crucial for the development of new antibiotics to target these resistance strategies . For example, understanding the mechanistic detail of how β-lactam antibiotics act and how microbes resist them allows the continuous development of new molecules that could circumvent β-lactamase-driven resistance. Resistance determinants can also be targeted directly. The OM permeability barrier could be circumvented by combining antibiotics with permeabilizers such as nanosilver or polymyxin B nonapeptide [81–84]. The development of drug-resistant biofilms could be avoided by targeting quorum sensing [40,85,86] or by preventing biofilm attachment to surfaces [87–89].
The biochemistry of persister cells has also been studied and targets against them are in development . As drug efflux pumps play such an integral role in both intrinsic and acquired resistance, as well as in resistance development, they are attractive targets for inhibition [28,91–94]. However, clinical application of the above therapies has not yet been accomplished. In addition to the significant lag time between discovery of an antimicrobial and development as a clinical drug , research into novel antimicrobials is generally poorly funded and not well supported by the pharmaceutical industry. The ultimate consequence of this, is further perturbation of the rate at which new antibiotics can be developed.
The impact of diseases caused by AMR Gram-negative bacteria should not be underestimated. Resistance should be continuously monitored and the biochemical insights gained should inform the development of novel therapies targeting AMR mechanisms as this is the only way we would be able to stem the tide of drug-resistant, untreatable infections.
Infections caused by antimicrobial-resistant, Gram-negative organisms pose a serious threat to our healthcare and veterinary industries.
AMR can be intrinsic, adaptive or acquired.
Intrinsic resistance is the innate ability of an organism to withstand antimicrobial treatments such as the permeability barrier of the additional OM in Gram-negatives or the increased expression of non-specific drug efflux pumps.
Adaptive resistance is driven by differential gene expression that allows the bacteria to respond to external stresses. This could result in lifestyle changes such as the formation of persister cells or biofilms.
Acquired resistance is characterized as the emergence of a previously antibiotic-sensitive organism, which has acquired the genes or mutations that render it resistant to a particular antibiotic. Examples are the acquisition of the genes coding for potent β-lactamase enzymes or mutations in bacterial target proteins that render them resistant against the antibiotic.
Knowledge of the molecular mechanisms of resistance is very important as this allows the development of new therapies to combat multidrug-resistant Gram-negative organisms.
We thank Dr Stephanie Begg for critical reading of the manuscript.
The authors declare that there are no competing interests associated with the manuscript.