Roles of two-component regulatory systems in antibiotic resistance

Abstract

Two-component regulatory systems (TCSs) are a major mechanism by which bacteria sense and respond to changes in their environment. TCSs typically consist of two proteins that bring about major regulation of the cell genome through coordinated action mediated by phosphorylation. Environmental conditions that activate TCSs are numerous and diverse and include exposure to antibiotics as well as conditions inside a host. The resulting regulatory action often involves activation of antibiotic defenses and changes to cell physiology that increase antibiotic resistance. Examples of resistance mechanisms enacted by TCSs contained in this review span those found in both Gram-negative and Gram-positive species and include cell surface modifications, changes in cell permeability, increased biofilm formation, and upregulation of antibiotic-degrading enzymes.

The escalating issue of antibiotic resistance

Since the introduction of penicillin in 1942, antibiotics have been nothing short of miracle drugs, prolonging life through treatment of previously lethal infections and permitting advances in healthcare through surgery. However, within 2 years of clinical implementation of penicillin, healthcare workers were already beginning to see a loss of drug efficacy due to bacterial resistance. Now, over 70 years later, a post-antibiotic era is either here or uncomfortably close. Measures must be taken to institute better stewardship of current drugs to dedicate funding to research of new drugs as well as to research the mechanisms of antibiotic resistance. This last point may seem comparatively trivial or too little, too late, but an understanding of how bacteria evade antibiotics is crucial to the success of both the former measures.

The intent of this review is to introduce the primary mechanisms by which bacteria activate antibiotic resistance via two-component regulatory systems (TCSs). TCSs are ubiquitous in bacteria and crucial to the maintenance of homeostasis, allowing bacteria to sense changes in their environment and respond accordingly. While the primary purpose of many TCSs is to initiate responses to simple criteria such as environmental changes in nutritional availability and osmolarity, others directly respond to the presence of antibiotics or conditions induced by them. In many cases the identity of the signal is not known, and even TCSs that do not directly sense antibiotics may respond to conditions that overlap with those found inside hosts or may indirectly lead to an increased antibiotic-resistant state. Consequently, TCSs are major players in the realm of infectious disease caused by pathogenic bacteria, and so much so that recent years have seen an increase in research of identifying drugs that target TCSs [1–21].

Fundamentals of bacterial two-component signal transduction

Prototypical bacterial TCSs consist of a pair of proteins: a sensor histidine kinase (HK) and a response regulator (RR) (Figure 1). The HK constitutes a homodimeric integral membrane protein with an N-terminal domain containing a receptor that faces the extracellular or periplasmic space and a C-terminal kinase domain located in the cytosol. These domains are connected by a series of transmembrane helices, the number of which vary from system to system. The variable HK N-terminal domain senses environmental changes via binding of an extracellular ligand or through other conformational changes, which triggers autophosphorylation of a conserved C-terminal histidine residue through adenosine triphosphate hydrolysis. The histidine-bound phosphate is transferred to an aspartate residue in the conserved N-terminal domain of the RR, a homodimeric protein located in the cytosol, which activates the RR’s variable C-terminal output domain, allowing it to regulate genomic expression via targeting of DNA. Through this flow of information through phosphorylation, bacteria are able to effectively sense changes in their surroundings (nutrition, pH, osmotic pressure, antibiotics, etc.) and orchestrate their genetic expression in a way that permits a rapid, cohesive reaction to a dynamic environment [22–26]. Furthermore, the variable natures of the input domain of the HK and the output domain of the RR can allow adaptation to recognize new signaling molecules and/or regulate different genetic foci.

Figure 1. The basic process of two-component signal transduction.

(1) An extracellular ligand (orange) binds to the N-terminal receptor of the sensor histidine kinase (green), embedded in the cell membrane. Binding of the ligand causes (2) the C-terminal kinase domain to hydrolyze adenosine triphosphate and phosphorylate a histidine residue. (3) The phosphate (yellow) is then transferred to an aspartate residue on the N-terminal domain of the cytosolic response regulator (blue). (4) This phosphorylation event activates the response regulator’s C-terminal output domain, which leads to global transcriptional changes. (5) In some cases, the sensor histidine kinase also functions as a phosphatase, and terminates the response regulator’s activation by removal of the phosphate. Inset to right: intramembrane-sensing histidine kinases lack an extracytoplasmic sensory domain and have recently been shown to recruit other sensory proteins in order to function.

The number and organization of proteins comprising a TCS can vary from the archetypal arrangement described above. For example, certain systems utilize an intramembrane-sensing HK (IM-HK), which is significantly smaller than an average HK and does not extend into the extracellular or periplasmic space (Figure 1). The IM-HKs were originally predicted to detect stimuli within or at the surface of the membrane [27,28], but more recent studies indicated that these minimalistic HKs do not carry out sensory functions independently and instead recruit additional proteins that detect membrane stimuli [29–31]. Additionally, despite the name of ‘two-component system’, the domains that compose the HK and RR may exist on multiple separate proteins, forming a phosphorelay. In particular, it is not uncommon for an independent protein to provide the histidine-containing phosphotransfer (HPt) domain that transfers a phosphate to the RR aspartate [32,33]. Finally, an additional gene product may be required for proper functioning of the HK [34] or act as a second RR [35], forming a ‘three-component system’.

A subset of sensor kinases autophosphorylate on a serine or threonine residue, which is similar to signal transduction in eukaryotic cells and has resulted in their being termed eukaryotic-like Ser/Thr kinases (eSTKs). The eSTKs bring about cellular changes by different mechanisms than TCSs, and therefore are not considered to be substitutes of HKs in TCSs. However, interactions in signaling between eSTKs and TCSs can occur [36]. As the purpose of this review is to consider antibiotic resistance controlled specifically by TCSs, eSTKs will not be discussed further.

Mechanisms of TCS-induced antibiotic resistance

In many cases, TCSs respond directly to the presence of antibiotics to bring about a resistant phenotype. In other cases, TCS-enacted global regulation in response to general environmental stressors or changes results in physiological changes that indirectly render the bacteria more resistant to antibiotics. While the types of environmental signals are highly variable, the resulting mechanisms of antibiotic resistance brought about by TCSs will be divided into four categories for the purposes of this review: modification of the cell surface, decreased drug influx or increased drug efflux, upregulation of antibiotic-degrading enzymes and alternative forms of antibiotic resistance, including biofilm production and stress response-induced antibiotic resistance. The following sections will highlight the best characterized TCSs per category of resistance mechanism. Additional TCSs and those that affect antibiotic resistance through an undetermined mechanism can be found in Table 1.

Table 1. Summary of two-component regulatory systems that increase antibiotic resistance.

Two-component system	Regulation increasing antibiotic resistance	Antibiotic resistance(s)	Example species	Ref.
AarG	Increased expression of an aminoglycoside acetyltransferase Global transcriptional changes leading to increased intrinsic resistances	Aminoglycosides Tetracycline Chloramphenicol Ciprofloxacin	Providencia stuartii	[37]
AdeRS	Upregulation of AdeAB(C) efflux pump	Intrinsic resistance	Acinetobacter baumannii	[38–44]
AmgRS	Upregulation of MDR efflux pumps Activation of stress response protects membrane Mutations can give rise to SCVs	Aminoglycosides Intrinsic resistance	Pseudomonas aeruginosa	[45–48]
BaeSR	Upregulation of MDR efflux pumps	Ceftriaxone Intrinsic resistance	A. baumannii Escherichia coli Salmonella enterica	[38–40,42,49–52]
BfmRS	Increases formation of biofilm	Chloramphenicol Intrinsic resistance	A. baumannii	[1,38,39,53–56]
BlrAB	Activation of three β-lactamase genes	β-lactams	Aeromonas species	[57–59]
BraRS (BceRS, NsaRS)	Upregulation of MDR efflux pumps	Nisin Bacitracin	Staphylococcus aureus Bacillus subtilis	[60–66]
CesRK	Upregulation of cell-envelope related genes	β-lactams	Listeria monocytogenes	[67–69]
CesSR	Upregulation of cell-envelope related genes	Bacitracin Bacteriocins	Lactococcus lactis	[70]
CiaRH	Upregulation of cell-envelope related genes	β-lactams Cefotaxime Polymyxin B	Streptococcus pneumoniae Group B Streptococcus	[71–78]
ComDE	Increases formation of biofilm	Intrinsic resistance	Streptococcus mutans	[79,80]
CopRS	Decreased porin expression	β-lactams Carbapenems	P. aeruginosa	[81,82]
CprRS	Modification of lipid A by 4-aminoarabinose (via arn operon)	Polymyxins Aminoglycosides	P. aeruginosa	[83–88]
CpxAR	Decreased porin expression Upregulation of MDR efflux pumps	Chloramphenicol Amikacin Nalidixic acid Tetracycline	Klebsiella pneumoniae E. coli	[8,89–95]
CreBC	Activation of β-lactamase gene Increases formation of biofilm	β-lactams Intrinsic resistance	P. aeruginosa	[96–99]
CroRS	Upregulation PBP5	β-lactams	Enterococcus species	[100–106]
CzcRS	Decreased porin expression	β-lactams Carbapenems	P. aeruginosa	[82,107]
EnvZ/OmpR	Decreased porin expression Upregulation of MDR efflux pumps	β-lactams Intrinsic resistance	E. coli A. baumannii Yersinia pseudotuberculosis S. enterica	[89,90,108–122]
EvgAS	Upregulation of MDR efflux pumps	Intrinsic resistance	E. coli Klebsiella pneumoniae Vibrio cholera P. aeruginosa	[95,123–126]
fsr	Increases formation of biofilm via production of gelatinase	Intrinsic resistance	Enterococcus faecalis	[127]
GacSA	Increases formation of biofilm	Intrinsic resistance	A. baumannii	[128]
GraRS (aps)	Modification of teichoic acids via D-alanylation which reverses bacterial surface charge Upregulation of VraFG ABC transporter	Daptomycin Vancomycin Cationic antibiotics	S. aureus	[61,129–131]
hk11/rr11	Increases formation of biofilm	Intrinsic resistance	S. mutans	[132]
LiaSR	Regulation of cell wall stress responses Increases formation of biofilm	Vancomycin Bacitracin β-lactams Polymyxin B Nisin Intrinsic resistance	B. subtilis Streptococcus species	[34,60,71,133–137]
LisRK	Mechanism unknown	Nisin Cephalosporins	L. monocytogenes	[60,138,139]
LsrRS	Activation of ABC transporter LctFEG	Nukacin ISK-1	S. mutans	[60,61]
MisSR	Mechanism unknown; hypothesized upregulation of cell-envelope related proteins to protect the cell	Polymyxins Aminoglycosides	Neisseria gonorrhoeae Neisseria meningitidis	[140–142]
MtrAB	Upregulation of efflux pumps Regulation of a genetic switch to an MDR colony morphotype	Isoniazid Rifampicin Ethambutol Vancomycin	Mycobacterium tuberculosis Mycobacterium smegmatis Mycobacterium avium	[143–146]
NsrRS	Activation of NsrX, a gene hypothesized to interfere with nisin binding to lipid II	Nisin	S. mutans	[60,61]
ParRS	Modification of lipid A by 4-aminoarabinose (via arn operon)	Polymyxins Aminoglycosides	P. aeruginosa	[86–88,147]
PdtaRS	Mechanism unknown; hypothesized modification of 30S subunit of ribosome	Tetracycline Penimepicycline Blastocidin S	M. smegmatis	[148,149]
PhoBR	Decreased porin expression	Chloramphenicol Amikacin Nalidixic acid Tetracycline	K. pneumoniae	[150]
PhoPQ	Modification of lipid A by 4-aminoarabinose, phosphoethanolamine (via PmrAB), or palmitate (via PagP) Upregulation of efflux pumps (some species)	Polymyxins Aminoglycosides	Enterobacter cloacae P. aeruginosa S. enterica K. pneumoniae Yersinia pestis	[151–172]
PmrAB	Modification of lipid A by 4-aminoarabinose or phosphoethanolamine Mutations can give rise to SCVs in P. aeruginosa	Polymyxins Aminoglycosides	A. baumannii S. enterica P. aeruginosa K. pneumoniae	[39,153,157,161–164,167,168,171,173]
RcsBCD	Increases biofilm formation Regulates a gene required for 4-aminoarabinose modification of lipid A	Daptomycin Cationic antibiotics Intrinsic resistance	E. coli S. enterica Proteus mirabilis	[158,174–186]
RetS-GacSA	Increases formation of biofilm	Intrinsic resistance	P. aeruginosa	[187–189]
SagS	Upregulation of MDR efflux pumps Increases formation of biofilm via BfiSR TCS	Tobramycin Intrinsic resistance	P. aeruginosa	[190–197]
SmeRySy	Upregulation of SmeZ efflux pump	Aminoglycosides	Stenotrophomonas maltophilia	[198]
VanSR	Alters vancomycin target through several actions	Vancomycin	S. aureus	[199–204]
VbrKR	Activation of β-lactamase genes	β-lactams	Vibrio parahaemolyticus	[6,205]
VicKR	Increases formation of biofilm	Intrinsic resistance	Streptococcus species	[137,206–208]
VirRS	Upregulation of ABC transporter AnrAB	Nisin	L. monocytogenes	[60,209,210]
VraSR	Increases peptidoglycan synthesis and expression of PBP2	Methicillin Vancomycin Daptomycin Oxacillin β-lactams	S. aureus	[60,211–217]
WalKR (YycFG)	Mechanism unknown; hypothesized increased permeability or decreased efflux Increases formation of biofilm	Macrolides Lincosamides Intrinsic resistance	S. aureus	[218–220]
YvcPQ	Upregulation of MDR efflux pumps	Lantibiotics Bacitracin	Bacillus thuringiensis	[221]

The purpose of this table is to provide the reader with a list of two-component regulatory systems that can increase antibiotic resistance, the mechanisms by which resistance occurs, examples of antibiotics that can be affected, and examples of bacterial species in which one or more of these resistances can occur. Table entries that occupy the same row are not intended to be exclusively linked.

MDR: Multidrug resistance; SCV: Small colony variant; TCS: Two-component system.

Modification of cell surfaces

Different classes of antibiotics work through a variety of mechanisms that cause bacterial lysis or stasis. However, all antibiotics must first interact with the outermost portion of the cell. In Gram-negative bacteria, this is the outer membrane (OM), and in Gram-positive bacteria, it is the cell wall composed of peptidoglycan. Many antibiotics directly target the OM, peptidoglycan or their biogenesis as their destabilization is lethal to cells [222].

Gram-negative bacteria

Strongly positively charged antibiotics such as polymyxin B, colistin, aminoglycosides, as well as host cationic antimicrobial peptides (CAMPs) take advantage of the net negative charge of the OM in Gram-negative cells. Polymyxin B, colistin and host CAMPs utilize a self-promoted uptake system in which they interact with the OM to form neutral patches that lead to membrane breakage, thereby permitting drug or peptide passage to the periplasm. Here, the amphipathic portion of the cationic molecules insert into the cytoplasmic membrane to form pores, leading to the breakdown of the membrane and cell death. It is also possible for this insertion to occur at the OM [223,224]. Aminoglycosides use the difference in charge to cross the membrane and reach their target – the bacterial ribosome [225–227]. The anionic nature of the OM is due to the prevalence of lipopolysaccharide, which contains a negatively charged lipid A moiety. Bacteria can reverse this state by covalent modification of the lipid A through which the OM becomes positively charged, resulting in decreased or abolished antibiotic action. The three most common modifications to lipid A are the addition of 4-aminoarabinose (4AA), phosphoethanolamine (PEtN) or palmitic acid, the latter of which reduces membrane fluidity but does not affect charge [151,152].

TCSs play a major role in OM lipid A modification, and the two of the best known and most well-characterized TCSs, PhoPQ and PmrAB are found in many Gram-negative bacterial species including but not limited to: Salmonella enterica, Enterobacter cloacae, Pseudomonas aeruginosa, Klebsiella pneumoniae, and Yersinia pestis [151–170]. PhoQ is a HK located in the inner membrane (IM) which activates in response to low concentrations of Mg²⁺ or to the presence of polycationic peptides. The SK PhoQ then activates the RR PhoP, which can bring about modifications to the OM via genetic regulation. The path leading to lipid A modification differs from species to species. In S. enterica, PhoP can further activate pmrAB, which can lead to PEtN and 4AA modifications to lipid A by activation of the pmrCAB and arnBCADTEF-pmrE operons, respectively [38,164,173,228–232]. In P. aeruginosa, the arnBCADTEF-pmrE operon can be directly induced by PhoP or else via the PmrAB TCS [154]. Furthermore, PmrAB is capable of activating independently of PhoPQ through sensing of low pH or high Fe³⁺ levels [39,163,171,231], and a recent study of clinical isolates of Acinetobacter baumannii, a multidrug resistant Gram-negative nosocomial pathogen, found mutations in PmrAB could invoke up to a 30-fold increase in transcription of pmrC, which encodes the lipid A phosphoethanolamine transferase [233]. Finally, PhoP can also activate pagB, a gene encoding a palmitoyltransferase which modifies lipid A by addition of a palmitic acid [158].

Two additional TCSs in P. aeruginosa, ParRS and CprRS can activate the arnBCADTEF operon, which regulates 4-AA modification of lipid A in response to the detection of subinhibitory concentrations of polymyxins and colistin [83–87,147]. A recent study further demonstrated the critical need to understand these systems, as the researchers found that 4-AA lipid A modifications are a prerequisite to the evolution of colistin resistance in P. aeruginosa [234].

Finally, in several Gram-negative organisms, including E. coli and Salmonella species, the Rcs TCS, also known as the Rcs phosphorelay, regulates expression of ugd, which is essential to the production and incorporation of 4-AA into the lipopolysaccharide [158,174–184]. Further, in E. coli the Rcs TCS has been shown to increase PagP expression in mature biofilms, leading to increased lipid A palmitoylation and consequently enhanced resistance to treatment with positively charged antibiotics [184,235].

Gram-positive bacteria

The peptidoglycan of bacterial cell walls is formed by alternating sugar residues N-acetylglucosamine (NAG) and N-acetylmuramic acid (NAM), polymers of which are linked together by cross-linked peptide side chains bound to NAM. During cell wall biosynthesis, the peptide side chains begin as pentapeptides bound to a Lipid II precursor molecule, and in many species the pentapeptide contains a terminal D-Ala D-Ala moiety. The cross-linkage of peptide side chains on opposite strands of the NAM-NAG polymer is catalyzed by a transpeptidase enzyme that recognizes the D-Ala D-Ala signature. The transpeptidase cleaves the distal D-Ala from each side chain being cross-linked and forms a bond between the two side chains.

Both the glycopeptide and β-lactam classes of antibiotics attempt to shut down cell wall synthesis by preventing the final cross-linking step of peptidoglycan synthesis. In the case of glycopeptides, such as vancomycin, inhibition is accomplished by binding of the drug to the D-Ala peptides, which simultaneously blocks polymerization of Lipid II through transglycosylation as well as the transpeptidase action [236]. β-lactams, such as penicillin, mimic the D-Ala-D-Ala moiety and bind to the active site of the transpeptidase enzyme, impeding it from binding its true substrate. For this reason, transpeptidase enzymes that cross-link peptidoglycan are called penicillin binding proteins (PBPs) [237].

The TCSs in many bacterial species can sense antimicrobials or antimicrobial-induced cell wall damage and initiate a response. Vancomycin-resistant Enterococcus and vancomycin-resistant Staphylococcus aureus and several Streptomyces species utilize a chromosomal or plasmid encoded VanSR TCS to counteract vancomycin. The VanS, the HK, is located in the plasma membrane and autophosphorylates on exposure to vancomycin. Subsequent activation of the RR VanR results in the activation of several genes – vanA/B, vanH, vanX and vanY – whose products collectively negate vancomycin’s mechanism of action. The VanA/B ligates D-Lac to D-Ala, lowering the concentration of D-Ala pentapeptides targetable by vancomycin. The VanH reduces pyruvate to D-Lac, further supporting the action of VanA/B. The VanX and VanY work together to deplete the cell of the normal peptidoglycan precursors displaying the D-Ala moiety required for vancomycin recognition [199–204].

In addition to VanSR defenses, S. aureus encodes VraSR and GraRS, both of which prevent antibiotic targeting of cell surfaces. The VraSR acts to mitigate antibiotic degradation of the cell wall specifically during the early and late stages of peptidoglycan synthesis. The system is induced by the action of β-lactams or glycopeptides, increasing resistance to vancomycin, daptomycin and oxacillin [60,211–216]. The VraSR has also been shown to increase expression of PBP2, which plays a role in methicillin and vancomycin resistance [217]. The GraRS, also known as the APS system (antimicrobial peptide sensing), reverses the surface bacterial charge via D-alanylation of teichoic acids to repel positively charged antibiotics and host cAMPs [129,130,238,239].

Enterococcus faecalis and Enterococcus faecium utilize the CroRS TCS to resist β-lactam antibiotics including ampicillin and cephalosporins. In this case the CroS HK activates in response to antimicrobial-induced cell wall damage, and in response, CroR upregulates PBP5, an alternative PBP which can still carry out transpeptidase activity while having a low-binding affinity for β-lactams [100–106].

Several Streptococcus species including S. pneumoniae, S. mutans, S. agalactiae (group B Streptococcus) utilize LiaSR to respond to cell-envelope stressors such as vancomycin, bacitracin and polymyxin B, resulting in activation of genes involved in cell wall maintenance [34,71,133–136].

TCS regulation of drug influx & efflux

Bacteria regulate entry and exit of many molecules via the expression of porins and efflux pumps. Porins, also called outer membrane proteins (OMPs), are β-barrel proteins embedded in the OM of Gram-negative bacteria that allow passive diffusion of molecules. Hydrophilic antibiotics such as β-lactams, aminoglycosides and fluoroquinolones can enter cells via porins, and therefore, downregulation of porins can reduce permeation and decrease susceptibility to a wide variety of antibiotics [240–244].

Efflux pumps are active transport proteins essential to bacterial homeostasis through expulsion of toxic substances and are found in all types of bacteria. Naturally, bacteria also use them to expel antibiotics, and many drug resistances arise from increased expression or activity of efflux pumps [242,243,245]. Some types of pumps transport a wide variety of structurally dissimilar compounds, and it is these that give rise to multiple drug resistances (MDR) and are therefore termed MDR efflux pumps. Increased expression of efflux pumps is often a first step to high levels of resistance as it allows the bacteria to cope with low to intermediate levels of an antibiotic, thereby giving it a chance to gain additional mutations [246]. It is therefore unsurprising that there are many instances of TCS-mediated upregulation of efflux pumps found across many species of MDR bacteria. Here, we will consider a few examples to appreciate the prevalence and diversity of these resistance mechanisms.

Gram-negative bacteria

One of the best known TCSs in Gram-negatives is the EnvZ/OmpR, which is found in a wide variety of species including E. coli, A. baumannii, Yersinia species and S. enterica among others [89,90,108–121]. EnvZ is a HK that primarily senses osmolarity changes and modulates the RR OmpR to regulate expression levels of the OM porins OmpC and OmpF according to levels of certain chemicals in the environment. Reduced expression levels of these OMPs has been seen to upregulate resistance to β-lactams in E. coli and S. enterica Typhimurium [38,89,90,108–110,112], and furthermore, OmpR has been seen to act in additional regulatory pathways in Yersinia, activating the AcrAB-TolC MDR efflux pump [122].

In P. aeruginosa, the TCSs CzcRS and CopRS respond to the presence of metals in the environment. The HK CzcS senses zinc, cobalt, cadmium and copper, while the CopS HK senses copper. Both RRs, CzcR and CopR, repress expression of the OrpD porin. As a result, carbapenem antibiotics, a subclass of β-lactams used to treat highly problematic MDR infections, can no longer access their target PBPs due to the lack of an entry route [81,82,107]. Another P. aeruginosa sensor kinase SagS, a TCS hybrid, activates the RR BlrR via cyclic-di-GMP to turn on MDR ABC transporters [190–197]. Additionally, the AmgRS TCS responds to membrane stress and damage induced by aminoglycosides by upregulating the mexAB-oprM MDR efflux pump [45].

Similarly, K. pneumoniae encodes two TCSs, CpxAR and PhoBR, that repress the KpnO porin, decreasing susceptibility to several antibiotics including chloramphenicol, amikacin, nalidixic acid, and tetracycline. Both TCSs activate due to OM stress [91,150]. Furthermore, CpxAR simultaneously upregulates three MDR RND-family efflux pumps in response to OM stress in K. pneumoniae [91]. A CpxAR homolog in E. coli acts identically, repressing expression of the porins OmpF and OmpC [92,93] and simultaneously activating the mar operon which increases expression of the MDR pump AcrAB-TolC [122].

A. baumannii has at least two TCSs that upregulate MDR efflux pumps. The first, AdeRS, activates the AdeABC MDR pump, and the second, BaeSR, also activates AdeABC as well as two other pumps: AdeIJK and MacAB-TolC [38–44,49]. In E. coli, BaeSR acts similarly, activating the mdtABCD gene cluster encoding a multidrug transporter [50,51]. Finally, BaeSR acts in S. enterica Typhimurium to upregulate two OMPs, resulting in increased resistance to ceftriaxone [52].

In Stenotrophomonas maltophilia, a nosocomial pathogen, the TCS PhoPQ not only contributes to polymyxin resistance through lipid A modification, but also causes resistance to gentamicin, kanamycin, and streptomycin (aminoglycosides) through activation of SmeZ, an efflux protein [172], which is also regulated by an additional S. maltophilia TCS called SmeRySy [198].

Gram-positive bacteria

S. aureus encodes two TCSs that increase resistance by upregulation of various ABC transporters in response to low concentrations of antibiotics: GraRS, which has been implicated in vancomycin resistance through upregulation of the VraFG ABC transporter [61,130,131] and BraRS, also called BceRS or NsaRS, which activates the BraDE and VraDE ABC transporters, increasing resistance to nisin and bacitracin [60–64].

Bacillus subtilis, a soil-dwelling model organism, increases expression of BceAB, an ABC-family efflux pump, via its TCS BceRS in response to exposure to the antibiotic bacitracin [65,66]. Similarly, the YvcPQ TCS in Bacillus thuringiensis, a member of the Bacillus cereus group, modulates efflux via two ABC transporters in response to lantibiotics and bacitracin [221].

Mycobacteria

Finally, Mycobacterium tuberculosis encodes an essential TCS called MtrAB that primarily functions in replication and cell wall biosynthesis. However, MtrA also binds and activates the iniBAC promoter, an operon encoding an efflux pump that increases resistance to isoniazid and ethambutol, two first-line antituberculosis drugs [143–145,247–253]. Additionally, mutations to mtrA in M. smegmatis resulted in decreased resistance to rifampicin and vancomycin, albeit with increased resistance to isoniazid [254]. In the opportunistic pathogen M. avium, MtrAB is required for conversion to a virulent, MDR colony morphotype, presumably by decreasing cell permeability [146].

TCS regulation of antibiotic-modifying enzymes

A discussion of antibiotic resistance would not be complete without mention of the first resistance mechanism clinicians encountered after the introduction of penicillin: the production of enzymes that target the antibiotic itself. As their name suggests, β-lactamases are enzymes that degrade β-lactam antibiotics, which they accomplish through hydrolyzation of the β-lactam ring. They are comprised of many subclasses including but not limited to carbapenemases, cephalosporinases and penicillinases. Other classes of enzymes inactivate antibiotics by direct modification, such as aminoglycoside and chloramphenicol acetyltransferases [255].

The CreBC TCS in P. aeruginosa not only affects biofilm formation, but also activates the chromosomal ampC gene encoding a β-lactamase. The HK CreB directly senses β-lactam activity, making this system particularly effective: it detects a specific threat and responds with a specific countermeasure [96–99].

Another example is the BlrAB TCS found in bacteria of the Aeromonas genus (Gram-negative). Analogous to the CreBC TCS, the BlrB HK autophosphorylates in response to accumulation of pentapeptides in the periplasm, the result of β-lactam inhibition of PBP4. Subsequent phosphorylation of BlrA results in activation of three β-lactamases: a cephalosporinase, a penicillinase and a carbapenemase [57–59].

A final example in the MDR nosocomial pathogen Providencia stuartii, the AarG sensor kinase has been implicated in aminoglycoside resistance through regulation of the aac(2’)-Ia gene encoding an aminoglycoside acetyltransferase. Furthermore, a mutation in aarG gave rise to several additional intrinsic resistances to tetracycline, chloramphenicol and ciprofloxacin, a phenotype that was hypothesized to be mediated via derepression of aarP, encoding a global transcriptional regulator that targets various MDR-associated genes [37].

Alternative forms of resistance

Upregulation of biofilm formation

Biofilms are structurally complex bacterial communities that present a major obstacle to the treatment of many infections. While a thorough discussion of the issues of antibiotic efficacy in eradication of bacteria contained in biofilms is beyond the scope of this review, it is important to convey that biofilm formation generally results in increased tolerance or resistance to antibiotics in several ways including differential gene expression in biofilms, decreased drug penetration through the matrix and the presence of dormant populations of cells or persisters [256–264].

Several Streptococcus species including S. pyogenes, S. mutans and S. pneumoniae utilize the TCS VicKR, which upregulates biofilm formation in response to an as-yet-undetermined environmental signal [137,206–208]. In S. mutans, LiaSR also increases biofilm formation in response to cell envelope stress and can directly react to the antibiotic bacitracin [135,137]. At least two other TCSs, ComDE and hk11/rr11 in S. mutans, also appear to function in the formation of biofilm [79,80,132].

P. aeruginosa is well known for its role in cystic fibrosis-associated lung infections and the tremendous difficulties encountered in treatment due to the impenetrable nature of its biofilm. Therefore, it is not surprising that P. aeruginosa encodes several TCSs that facilitate biofilm formation [187]. The two-component hybrid SagS directs the transition from a planktonic to surface-associated way of life, activating a second TCS, BfiSR, which is essential to the formation of biofilm. As previously noted, SagS simultaneously activates the BrlR RR to upregulate MDR efflux pumps within the biofilm, greatly increasing the biofilm drug tolerance [190–197]. CreBC directly senses β-lactam inhibition of PBP4 and responds by increasing biofilm formation [97]. The GacSA TCS in P. aeruginosa is also essential for biofilm formation and is unique among other GacSA systems in that the GacS sensor kinase cannot autophosphorylate. Instead GacS must be activated by the orphan sensor kinase RetS, which functions as both a kinase and phosphatase to GacS [187–189]. Several more TCSs, including the Roc1 system, Rcs/Pvr, PprAB and PilRS, regulate the expression of cup fimbriae and/or type IV pili required for surface adhesion and prerequisite to biofilm formation in P. aeruginosa [35,185,187,265–268]. Furthermore and finally, small colony variants [269] in this species are known to have increased biofilm production and multidrug resistance [270], and interestingly, a recent study on clinical isolates of P. aeruginosa revealed that mutations in amgRS as well as pmrAB culminated in a highly drug resistant SCV phenotype [46].

BfmRS, a TCS in A. baumannii, also upregulates biofilm production, possibly in response to sublethal concentrations of chloramphenicol, and in doing so, becomes increasingly recalcitrant to treatment [1,38,39,53–56]. Finally, the Rcs TCS is also essential to biofilm formation in E. coli, Proteus mirabilis and S. enterica Typhimurium as well as other species [174–182,186,187].

Stress response-associated antibiotic resistance

A number of TCSs induce cellular stress responses in reaction to environmental changes including poor nutrition, fluctuations in temperature, membrane integrity and oxidative stress. Stress responses often bring about global transcriptional changes, some of which alter the efficacy of antibiotic action. Many of the previously discussed TCSs are considered to function in this manner, including PhoPQ, CpxAR, LiaSR, BaeSR, BceRS, BraRS, VraSR, CroRS and ParRS [34,133,271–273]. Other examples include the AmgRS TCS in P. aeruginosa, which is believed to respond to membrane stress caused by aminoglycoside-induced accumulation of mistranslated peptides by upregulating certain proteases and otherwise implementing measures that protect the membrane [47,48].

Future perspective

TCSs are incredibly versatile and provide bacteria with an indispensable tool for sensing and reacting to their surroundings. Given the essential role of TCSs in bacterial homeostasis as well as the diverse antibiotic resistance mechanisms that can be activated through them, TCSs are an intriguing target for the development of new antimicrobial therapeutics [1–21]. The ability to use predictive software to identify putative compounds that bind one or more TCSs to render them ineffective would be an essential tool in the pharmaceutical targeting of TCSs [2]. Further, the conserved nature of certain domains of the various HKs and RRs both within and across bacterial species could be exploited to develop new broad-spectrum antimicrobials, although portions of these domains share similarity to eukaryotic proteins and thereby could lead to inhibition of host cell processes. Tiwari et al. recently published a review on this topic that covers the benefits and feasibility of medicinal targeting of TCSs and lists several predicted and validated TCS inhibitors [21].

While this review cannot hope to detail the myriad and often pleiotropic effects of TCSs, one can certainly begin to grasp the scope of their importance in antibiotic resistance from the research compiled herein. The numerous instances of TCS-induced resistance mechanisms across both Gram-negative and Gram-positive species highlights the need for further and continued investigation into targeting TCSs in the future development of antimicrobial therapeutics.

Executive summary.

• Bacteria utilize coupled sensory-response proteins known as two-component systems (TCSs) to sense environmental changes, including exposure to antibiotics and conditions inside a host, and to respond with activation of genes that increase antibiotic resistance.

• TCSs are ubiquitous in bacteria and can initiate a variety of antibiotic resistance mechanisms, including modification of the cell surface, increased efflux, decreased influx, biofilm formation and upregulation of antibiotic-degrading enzymes.

• The importance of TCSs to antibiotic resistance, in addition to their role in regulating virulence, has led to the current focus in drug discovery for TCS inhibitors.