Staphylococcus aureus has an incredible ability to survive, either by adapting to environmental conditions or defending against exogenous stress. Although there are certainly important genetic traits, in part this ability is provided by the breadth of modes of growth S. aureus can adopt. It has been proposed that while within their host, S. aureus survives host-generated and therapeutic antimicrobial stress via alternative lifestyles: a persister sub-population, through biofilm growth on host tissue or by growing as small colony variants (SCVs). Key to an understanding of chronic and relapsing S. aureus infections is determining the molecular basis for its switch to these quasi-dormant lifestyles. In a multicellular biofilm, the metabolically quiescent bacterial community additionally produces a highly protective extracellular polymeric substance (EPS). Furthermore, there are bacteria within a biofilm community that have an altered physiology potentially equivalent to persister cells. Recent studies have directly linked the cellular ATP production by persister cells as their key feature and the basis for their tolerance of a range of antibiotics. In clinical settings, SCVs of S. aureus have been observed for many years; when cultured, these cells form non-pigmented colonies and are approximately ten times smaller than their counterparts. Various genotypic factors have been identified in attempts to characterize S. aureus SCVs and different environmental stresses have been implicated as important inducers.
Antimicrobial medicines, centrally this refers to antibiotics, are an essential component in the armoury we possess against infectious diseases. However, the increasing failure of antibiotic treatment of bacterial infections has lead governments, professional bodies and institutions to recognize that antimicrobial resistance (AMR) is a significant global health priority. Failure of antibiotics is not simply due to the genetically encoded mechanisms of resistance but also to the more complex nature of bacterial tolerance of antibiotics. The distinction between resistance and tolerance has been recently well reviewed and discussed .
There is considered to be a small portion (sub-population) of a bacterial culture (this is true of an antibiotic-susceptible population) that can tolerate high concentrations of antibiotics and thereby they remain present after a seemingly successful treatment; this population of cells is referred to as persister cells (Figure 1). Even at the dawning of the antibiotic era there was research that highlighted a less than clear picture of antibiotic killing of bacteria. Harry Eagle, together with Musselman, in 1948 did studies that paradoxically showed that some bacteria returned to viable growth at concentrations well beyond its calculated minimum inhibitory concentration (MIC). This has been termed the Eagle effect or paradoxical effect . Prior to this, in 1944, Joseph Warwick Bigger and then similar research in 1945 by Edward Stephens Duthie, showed that Staphylococcus pyogenes (aureus) was not fully killed by penicillin; there remained approximately 1% of the population that was not killed. This sub-population (the persister cells) had not developed a resistance but through an iterative, repetition of these experiments it was shown that progeny of these cells continued to remain largely antibiotic susceptible – albeit with a sub-population of persister cells.
Bacterial persister cells and antibiotic tolerance
In addition to these, persister cells, but potentially overlapping, are bacterial biofilms and then the generation of small colony variant (SCV) cells. The development of these novel lifestyles has been clearly associated with chronicity and relapse in S. aureus infections [3–5]. This has been shown to be the case in various pathologies: chronic osteomyelitis [6,7], arthritis , chronic rhinosinusitis [5,9], cystic fibrosis (CF) [3,10,11], soft tissue infections , sepsis , endocarditis [3,14] and medical devices associated infections [3,15,16]. The metabolic and biochemical pathways that are involved in S. aureus SCV cell development have been well reviewed [17,18] and are not the focus of this present study.
Infections that involve either biofilm or SCVs are rarely resolved by host defences because these bacteria exist within a dormant state, hidden from the usual immune responses . These cells are furthermore inherently tolerant of antibiotics. When suitable conditions arise, the infections can recur.
The estimated frequency of occurrence of human S. aureus SCVs varies between 1 and 30% of clinical samples . S. aureus SCV was found in 29% of patients with osteomyelitis  and 17–46% of patients with CF who were chronically colonized with S. aureus [11,19,20]. The molecular evolution of these S. aureus cells has been investigated and to some degree there is an understanding of their biochemical and metabolic features. However, various questions are still open, in particular about the systems driving the development of SCV cells. The key difficulty in the study of SCVs is that when they are cultured in the laboratory they mostly revert to their parental growth-type. Antibiotics have been used to induce SCVs (and in some cases these are stable) . It has also been shown that antibiotics in a clinical setting select for SCVs, specifically trimethoprim selected for thymidine-dependent SCVs .
S. aureus infection and antibiotic treatment
S. aureus causes a wide range of illnesses throughout the body due to its capacity to colonize and grow in different host tissues. Anterior nares and skin are the frequent sites that S. aureus colonizes as a harmless commensal organism [23,24] but it can be also found in several body sites such as axillae , vagina  and the gastrointestinal tract . The minor cutaneous infections caused by S. aureus include carbuncles, boils, impetigo, burn and surgical site infections and wound infections. The more severe infections such as sinusitis, tonsillitis, osteomyelitis, pneumonitis, endocarditis, meningitis and bacteraemia generally occur when S. aureus enters the body via an opening cut or wound . Additionally, by releasing specific toxins within food or in the bloodstream, S. aureus can cause food poisoning, scalded skin syndrome and toxic shock syndrome [29,30].
When penicillin was no longer efficient to control Staphylococcal infections, the antibiotic methicillin was introduced but within a short time (from its introduction in 1960 until 1962), it became less efficient because of repeated cases of resistance . The bacteria evolving resistance to methicillin was first detected in hospitals and these were called methicillin-resistant S. aureus (MRSA). Afterwards, MRSA strains were known to be resistant to not only methicillin but also a range of penicillin-like antibiotics (β-lactams) such as amoxicillin, oxacillin methicillin, cephalosporins and other agents such as erythromycin and aminoglycosides . From genomics studies, the acquisition of resistance in MRSA has been assigned to the presence of the gene ‘mecA’ located on the Staphylococcal chromosome cassette mec (SCCmec); a novel, mobile resistance element . The gene mecA encodes the 78-kDa penicillin-binding protein 2A (PBP2A) that has a low affinity for β-lactam antibiotics leading to the inhibition of cell wall synthesis by inactivating transpeptidase [33,34]. The regulation of mecA is controlled by the repressor MecI and the transmembrane β-lactam-sensing signal-transducer MecR1, which are divergently transcribed . The integration of SCCmec into the S. aureus genome is carried out with the cassette chromosome recombinases (ccr) genes including ccrA, ccrB and ccrC that are located on all SCCmec elements at a specific site. These genes excise and integrate at the SCCmec attachment site (attBscc) at the 3′ end of an ORF (orfX) [32,35]. There are at least 11 subtypes of SCCmec from I to XI and six classes (A, B, C1, C2, D and E) performing different resistance patterns based on the arrangement of mec complex genes including mecA, regulatory genes mecI and mecR1 and with the insertion sequences .
The glycopeptide vancomycin became the last line of defence against S. aureus infections. In 2002, there was the first report of MRSA strains with reduced susceptibility to vancomycin, vancomycin intermediate-resistance S. aureus (VISA) . The vancomycin-resistant S. aureus strain (VRSA) was first isolated in U.S.  and then several countries including France, South Africa, Brazil and Korea also reported the presence of VRSA . In both the MRSA and VRSA, there are genetic cassettes that harbour the genes encoding the resistance mechanisms and thereby provide an antibiotic resistance. The resistant strains are quite separate from VISA isolates. The common features among these strains and the molecular basis for their response to vancomycin have recently been well reviewed .
Alternative lifestyles of
S. aureus as a stress response
S. aureus notably adapts to different environmental conditions outside and inside the host by switching from their normal mode of growth (metabolic and physical characteristics) to form quiescent phenotypes such as within a biofilm and SCVs [3,41,42]. These quasi-dormant lifestyles have been characterized by a slower growth rate but with a prolonged survival capacity in the host both extracellularly and intracellularly when compared with planktonic cells [41,43,44]. According to the current definition, biofilms are characterized by heterogeneous multilayers of sessile single cells and microcolonies, which are encased in a matrix of extracellular polymeric substance (EPS) [3,45]. This matrix is possibly composed of polysaccharide intercellular adhesin (PIA) and proteins along with extracellular genetic materials (such as extracellular DNA, eDNA) [45,46]. Within biofilms, they are classically defined to be four distinct metabolic states of growing cells: aerobically (often located in the outer most layers exposed to oxygen and nutrients), fermentatively, dormant and dead, in which the dormant cells are dominant and lodge in the anoxic layers, potentially these are persister cells  (Figure 2). The changes in cell-type related to these phenotypic changes, importantly include an altered growth rate and an accompanied global change in expression of genes related to the cell’s metabolic pathways, surface structures and virulence factors [45,48–50].
Biofilm formation and antibiotic tolerance
SCVs were first reported in 1910 in Salmonella enterica serovar Typhi, and then widely found in many species including S. aureus, S. epidermidis, S. capitis, Pseudomonas aeruginosa, Vibrio cholera, Escherichia coli, Shigella spp., Lactobacillus acidophilus, Serratia marcescen, Burkholderia cepacia, Brucella melitensis and Neisseria gonorrhoeae . S. aureus SCVs have a slow growth rate and atypical morphological and biochemical properties compared with the parental phenotype and can appear only after incubation for up to 48 or 72 h. These colonies can have significantly reduced pigment or indeed they are usually non-pigmented; their haemolysis is also greatly reduced or negative [4,5,43].
The relevant characteristic of SCVs in clinical cases is that most of the patients had been treated unsuccessfully with long courses of antibiotics . The model that has been proposed (based largely on laboratory generated mutants) is a greatly reduced production of cellular ATP that results in limited cell wall associated features (capsule, pigment), a reduced growth rate and reduced membrane potential. It is intriguing to postulate the overlap in our current understanding of the physiology of persister cells and SCV – predominantly based around ATP production (discussed further below) .
Recent work has uniquely been able to develop an SCV-dominated population of stable SCV cells from a non-SCV starting inoculum; this was after prolonged growth in continuous culture (a chemostat) and was without antibiotic [52,53]. The changes during the transition to an SCV population were reflected by a change in cell morphology as well as in metabolic pathways and the expression of genes related to the expression of surface proteins, cell lysis, lantibiotics, staphylokinase, capsule biosynthesis, leucocidins and carotenoids. In addition, genomic changes may contribute to the lifestyle alteration of these S. aureus SCVs . There were two important single nucleotide polymorphisms (SNPs), the first leading to a change in the protein sequence of RsbU, a serine phosphatase that is part of the regulatory pathway controlling sigma B (SigB); the second SNP was in mgrA that encodes the transcriptional regulator MgrA that regulates autolytic activity but also has global effects on the cell.
Mutations in gene rplF (encoding ribosomal L6 required for cell growth) have been previously detected in S. aureus SCV formation . Mutations in the rsh, a RelA homologue is believed to result from a defective ppGpp hydrolase activity that is also required for bacterial growth, thereby leading to an increase in the stringent response for SCVs . Intriguingly, this exact pathway (the stringent response and ppGpp levels) is involved in the development of E. coli persister cells.
A number of studies have reported that S. aureus SCVs are formed in vitro when exposing the cell under certain conditions, such as sub-inhibitory concentration of antibiotics including gentamicin [56–58], vancomycin and penicillin  and quinolones [60,61]. Strains that are known to switch to an SCV form as well as model laboratory SCV strains (mutants) display an elevated antibiotic tolerance .
While in vivo there are further complex issues such as a combination of antibiotics within the local environment, such as in the case of gentamicin bead placement for osteomyelitis treatment, may exactly be selecting for SCVs . Additionally, in the case of polymicrobial niches (such as has been shown in cystic fibrosis (CF) lungs) the exposure to the exoproducts from other bacteria have an influence [63,64]. It has been observed that P. aeruginosa releases signal molecules including 4-hydroxy-2-heptylquinoline-N-oxide and pyocyanin that affect the S. aureus energy pathways and induce the formation of SCVs [64,65]. In contrast, P. aeruginosa seemingly also produces a compound (cis-2-decenoic acid, cis-DA) that awakens S. aureus persister cells, and while the original S. aureus persister cells display little virulence (and subsequently avoid engulfment by macrophages), these awakened cells are actually more virulent but still remain protected from macrophage attack .
A combination of therapies has been employed as a means to target persisters or SCV. Daptomycin is effective against non-growing S. aureus and its efficiency has been further enhanced by adding other compounds or antibiotics [67,68]. A stationary phase S. aureus culture treated with high doses of daptomycin removes all the cells except for the persister cells. Interestingly, the addition of glucose to this treatment of stationary phase cells increases the daptomycin killing 5-fold . Further to this, rifampicin together with acyldepsipeptide-4 (ADEP4) antibiotic lead to an activation and non-specificity of ClpP protease and even the removal of persister cells .
The likely targets against SCV cells are based around the suspected pathways required for SCV development. The early studies have observed a link between specific auxotrophic phenotypes (menadione, haemin and thiamine) with the slow growth rate of SCVs. The addition of these compounds to the growth medium could restore the size of colonies . Menadione and haemin auxotrophic SCVs have often been found in clinical patients suffering chronic respiratory infections and being treated with aminoglycosides [10,58]. Thymidine auxotrophic SCVs have been initially isolated from patients (especially those with CF) who have had long-term trimethoprim sulfamethoxazole (SXT) treatment . However, these have also been found in non-pulmonary sites in patients such as endocarditis, osteomyelitis, soft tissue infections, bacteraemia and sinusitis who undergo different antibiotic regimes and device-related infections [72–74].
Menadione and haemin are required for the biosynthesis of cytochromes and menaquinone in the electron transport system responsible for the generation of ATP. The production of ATP is required for cell wall biosynthesis, pigmentation and membrane potential to form the normal colonies. The membrane potential is known to facilitate the uptake of antibiotics, in particular, aminoglycosides. Some SCVs have been identified with the additional mutation in the rfl (or named fusE) that encodes ribosomal protein L6, required for protein synthesis. This affects the protein synthesis rate and also increases membrane fluidity; this increases aminoglycoside resistance.
The SCVs that are deficient in thymidine biosynthesis are not related to the electron transport system. Thymidine biosynthesis is performed with the presence of thymidylate synthase (thyA), thus the deletion of the gene thyA is believed to cause the deficiency of thymidine and form thymidine-dependent SCVs . This hypothesis has been clarified through the finding of a clinical thymidine–auxotrophic strain that produces a typical SCV can be complemented with thyA [75,76]. However, thymidine auxotrophs show a phenotype that is almost identical with electron transport variants, thus it is suggested to be linked with the electron transport system through the hypothesized demand for a correct membrane potential. Directed studies (using a thyA mutant) have shown that thyA has a role in virulence and mutations are involved in clinically isolated SCV .
S. aureus persisters and energy
Among the population of cells that make up an infecting strain, there are some cells predisposed to switching into an alternative lifestyle or some that are already actually in an alternative phenotypic state and furthermore, as a direct response to the various environmental stresses they encounter, there are some that then switch to one of these alternative states: specifically biofilm, SCV or persister cells (Figure 3). Persister cells are defined as phenotypic variants capable of tolerating high levels of bactericidal antibiotics. The molecular pathways that underpin the continued survival of persister cells have been studied to some detail in several bacteria, especially E. coli where toxin–antitoxin systems, the stringent response and phosphate metabolism (and other pathways) are all known to play a role (there is a recent mini-review addressing the key areas of the metabolism of bacterial persister cells, ). Similar attempts to define such pathways have been made in S. aureus using molecular mechanisms . Recently, it was determined that S. aureus persister cells tolerate antibiotic challenge due to their low ATP levels. Bactericidal antibiotics work by targeting active mechanisms in cells, resulting in the production of corrupted products and cell death. Fluoroquinolones act by converting DNA gyrase and topoisomerase into endonucleases; aminoglycosides cause mistranslation, which produces toxic misfolded peptides; and β-lactams kill cells by forcing a futile cycle of peptidoglycan synthesis. A decrease in cellular ATP will result in less active antibiotic targets and subsequently, less damage will occur in the presence of an antibiotic. This leads to increased tolerance and survival of these low ATP cells. Hence, heterogeneity in ATP levels in cells during infection may lead to the survival of sub-populations of cells. Furthermore, ATP levels of cells during infection is presumably lower than under rich culture conditions, particularly in anaerobic infection environments such as the bone or the CF lung [80–82], which would result in higher antibiotic tolerance. It may be that ATP levels represent a unifying determinant of antibiotic tolerance in persister cells, biofilm associated cells and SCVs. It was previously shown that ADEP4, an antibiotic that dysregulates the ClpP peptidase resulting in cell death was capable of killing persister cells and eradicating a deep-seated biofilm infection in mice. Interestingly, ADEP4 kills independently of ATP, which may explain its activity against persisters and biofilm and ability to eradicate infection. It will be interesting if future studies are able to determine if other compounds that kill independently of target activity or cellular ATP levels are also capable of eradicating chronic infection.
A bacterial population contains a cell type diversity
Bacterial infections that are recalcitrant to antibiotic treatment are known to involve a sub-population of cells that are quasi-dormant. The altered metabolism and cell processes provide an infection with a phenotypic tolerance of a wide range of antibiotics. These cells fall into different groups; for S. aureus the important cell types are persister cells, biofilms cells and SCV cells. S. aureus causes various infectious diseases and importantly many are associated with chronic and relapsing infections (chronic osteomyelitis, chronic rhinosinusitis, endocarditis and the lungs of CF patients). Research in the last couple of decades has attempted to study the biochemical pathways that are functional during cell survival in these states. In recent years, there have been major advances that have provided a real understanding of the unique nature of the metabolism of dormant/quasi-dormant S. aureus cells. This opens avenues to prevent this pathogenic bacterium switching to these evasive life styles and in doing so avoid chronic and relapsing pathologies.
This article summarizes the clinical relevance of the alternative lifestyles (persister cells, biofilms and SCVs) of S. aureus during chronic and relapsing pathologies.
The recent advances in the understanding of the metabolism and energy generation of these cells are described.
The identification of ATP production by S. aureus as a core element in persister cells, and this as a therapeutic target is discussed.
The role of SCVs in antibiotic tolerance is introduced and the prolonged growth of S. aureus in continous culture as a method to evolve a population of cell types, including stable SCVs and thereby maintain SCV cells for accurate, molecular analysis is described.
The authors declare that there are no competing interests associated with this manuscript.
cassette chromosome recombinase
extracellular polymeric substance
minimum inhibitory concentration
methicillin-resistant S. aureus
polysaccharide intercellular adhesin
small colony variant
single nucleotide polymorphism
vancomycin intermediate-resistance S. aureus
vancomycin-resistant S. aureus