Antibiotic Resistance: Implications for Global Health and Novel Intervention Strategies: Workshop Summary (2010)

Arturo Casadevall²

Albert Einstein College of Medicine

At the beginning of the twenty-first century, the treatment of microbial diseases is increasingly complicated by drug resistance, the emergence of new pathogenic microbes, the relatively inefficacy of antimicrobial therapy in immunocompromised hosts, and the reemergence of older diseases, often with drug-resistant microbes. Some of these problems can be traced to the switch between pathogen-specific antibacterial therapy and the nonspecific antibacterial therapy that followed the transition from serum therapy to modern antimicrobial chemotherapy. The widespread availability of cheap, effective, nontoxic wide-spectrum antibacterial therapy for almost 75 years fostered a culture of therapeutic empiricism that neglected diagnostic technologies. Despite unquestioned lifesaving efficacy for individuals with microbial diseases, the use of broad-spectrum antimicrobials was associated with fungal superinfections and antibiotic-associated

colitis, helped to catalyze the emergence of resistance, and is now tentatively associated in the pathogenesis of certain chronic diseases, including atopy, asthma and – perhaps – certain forms of cancer. This article briefly reviews these trends and suggests that the current strategy of nonspecific therapy is fundamentally unsound because it damages the microflora and – consequently – the human symbiont. The essay argues for the development of immunotherapy and pathogen-specific therapies, especially with regard to bacterial and fungal diseases, and suggests possible routes to that future.

Current antimicrobial therapy is largely pathogen-specific for viral diseases and nonpathogen-specific for bacterial, fungal, and parasitic diseases (Casadevall, 1996). Although some of the latter diseases are sometimes treated with pathogen-specific drugs, such as the use of isoniazid for tuberculosis, the overwhelming majority of compounds targeting bacteria, fungi, and parasitic diseases have activity against multiple microbes. Furthermore, these compounds target both pathogenic and nonpathogenic microbes. This current antimicrobial paradigm is currently in use at a time of significant upheaval in the therapy of microbial diseases, which is the only field of medicine in which one can argue that therapeutic options have declined over time. For example, in the 1950s Jawetz noted that the then currently available antimicrobial drugs were satisfactory for the treatment of bacterial diseases (Jawetz, 1956). However, in recent years the field of infectious diseases has seen dramatic increases in antimicrobial resistance, an increasing prevalence of bacterial and fungal superinfections in treated individuals, a relatively low therapeutic efficacy of antimicrobial therapy in individuals with impaired immunity, the emergence of new infectious diseases, and the reemergence of older microbial diseases, often with highly resistant microbes such as XDR-Tb. Given this status quo, it behooves us to ask the questions: How did we get here? What are the consequences of the choices made then and now? Can we do better and how do we get there?

Effective antimicrobial therapy can be dated to the introduction of serum therapy in the 1890s, which, for the first time, provided physicians with the ability to intervene and cause a favorable outcome for an infectious disease. Serum therapy was developed against numerous bacterial and viral diseases, including pneumococcal pneumonia, meningococcal meningitis, erysipelas, anthrax, and measles (for reviews, see refs Casadevall and Scharff, 1994; Casadevall and Scharff, 1995; Buchwald and Pirofski, 2003). The heyday of serum therapy was the 1930s, but the modality was rapidly abandoned because serum could not compete with small-molecule antimicrobial therapy, such as sulfonamides and

penicillin, with regard to price, stability, ease of use, and (low) toxicity. For some diseases such as meningococcal meningitis, small-molecule antimicrobial therapy was clearly more effective than serum therapy; however, for pneumococcal pneumonia the difference in efficacy was less clear. In addition to serum therapy, the few other therapies available (e.g., quinine for malaria, salvarsan for syphilis, optochin for pneumococcus, and phage therapy) were all pathogen specific. In a prior essay (Casadevall, 2006), I argued that the time of serum therapy and the subsequent era of therapy with small molecules constituted the two first ages of antimicrobial therapy. When viewed through the prism of microbial specificity, the greatest difference in the therapeutic approach between the first and second ages of antimicrobial therapy was a switch from pathogen-specific to nonspecific therapy with regard to antibacterial therapeutics. In this essay, I argue that this change was to have enormous implications, which are root causes for some of the problems we face today.

In evaluating the therapeutic paradigm for microbial diseases, it is worthwhile contrasting it with the therapy of cancer. Like therapy for infectious diseases, the treatment of tumors has relieved [sic] heavily on antibiotics made by microorganisms; adryamicin, actinomycin D, bleomycin etc. are all microbial products. Like antimicrobial antibiotics, these antimetabolite antibiotics are each nonspecific in the sense that they are cytotoxic to multiple tumors. However, unlike most antimicrobial antibiotics, these agents have tremendous toxicity for the host and, consequently, are never used empirically. Hence, oncology practice has placed great emphasis on diagnosis and in exploiting subtle pharmacological differences between these agents to enhance their therapeutic index.

In fairness to infectious diseases, it noteworthy that the temporal kinetics of microbial infections and tumorogenesis favored a more deliberate approach to diagnosis as tumors, which unlike microbes, seldom killed the host rapidly. Nevertheless, the analogy is relevant because it provides an inkling of how the practice of infectious diseases might have developed if early antimicrobials had more significant toxicity, as evidenced by the hesitant empiric use of amphotericin b and Ara-C for fungal and herpetic diseases, respectively, Consistent with this notion, the development of the relatively nontoxic antiherpetic drug acyclovir as a replacement for Ara-C was followed with significantly greater empiric use, especially in neonates and cases of encephalitis. Similarly, the introduction of low-toxicity azoles and echinochandins as replacements for the highly toxic amphotericin b has promoted the empirical use of antifungal therapy. Hence, the advantage of low toxicity has the perverse effect of promoting empirical and inappropriate use.

In comparing the ages of antimicrobial therapy, it is clear that the change in the specificity of therapeutic agents did not affect all types of antimicrobial therapy equally. Serum therapy for viral diseases was specific and current antiviral drugs remain largely pathogen-specific, with the caveat that some drugs like acyclovir have activity against multiple herperviruses [sic]. For mycobacterial

diseases, there was no effective therapy in the preantibiotic era and most drugs that were subsequently developed (isoniazid, ethambutol, and others) were used primarily for the therapy of tuberculosis. For fungal diseases, there was no effective therapy prior to the late 1950s when amphotericin B was introduced; a compound active against most fungal pathogens and antifungal therapy has always relied on nonpathogen-specific agents. For bacterial diseases, the change from serum to small-molecule therapeutics was a revolution, as therapeutic specificity was abandoned in favor of agents with increasingly greater spectrum of antimicrobial activity. However, what made the switch from pathogen-specific to non-pathogen-specific therapy so significant with regard to antibacterial therapy is that the human host is a symbiont, with microflora consisting mostly of desired commensal bacteria. By contrast, there are no known desirable commensal viruses and the known fungal flora is limited to a few fungal species where Candida spp predominate. Unlike bacteria, a beneficial function has not been demonstrated for the host-associated fungal microflora. Hence, the use of nonspecific bacterial therapy carried an inherent potentially detrimental effect in damaging the associated bacterial microflora, and thus the human symbiont.

The nonspecificity of antibacterial, and to a lesser extent antifungal, therapies was to have profound consequences on the practice and outcome of infectious diseases that reverberate to current times. The availability of nonspecific antibacterial therapies with broad spectrum and low toxicity allowed physicians to rapidly treat many infectious diseases without a need for a microbial diagnosis. For individuals with bacterial diseases, such therapy was often lifesaving. However, the ability to effectively treat many diseases safely without making a diagnosis deemphasized diagnostic clinical microbiology and fostered a culture of empiricism. For example, the diagnosis of pneumococcal pneumonia with the identification of the offending serotype took approximately 6 – 8 h in the 1930s and used the mouse peritoneal infection assay followed by typing with rabbit type-specific serum. This methodology was developed to rapidly ascertain the presence and serotype of pneumococcus in sputum because the efficacy of serum therapy depended on matching the bacterial serotype with the specificity of the antiserum. Despite the problems in unequivocally diagnosing pneumonia from sputum, this approach was successful for selecting therapeutic sera and supported the use of serum therapy. However, the introduction of penicillin and later antimicrobial drugs made the test much less relevant and it was abandoned as a diagnostic tool. Currently, a definitive diagnosis of pneumococcal pneumonia is possible only when accompanied by bacteremia, information that requires 48 h. For fungal diseases, a full embrace of empiric therapy was checked by the toxicity of amphotericin b, but by the late 1990s, the availability of relatively nontoxic azole and echinocandin-type drugs had ushered greater empiric use. By contrast,

for conditions that required specific therapy, such as viral and mycobacterial diseases, the practice ethos supported continued emphasis on diagnostic identification of the causative microbe.

For bacterial and later fungal diseases, the availability of relatively nontoxic broad-spectrum therapy contributed to the emergence of resistance among both targeted and nontargeted microbes. Although specific therapy can also elicit resistance, as witnessed by the emergence of isoniazid-resistant Mycobacterium tuberculosis, only nonspecific therapy can elicit resistance among nontargeted microbes such as common inhabitants of the microflora. Furthermore, only non-specific therapy can damage the microflora to create alterations that foster the emergence of usually commensal microbes such as Candida and Enteroccocus spp, first as major pathogenic microbes and then as drug-resistant pathogenic microbes. Consequently, the discipline of infectious diseases may be the only specialty of medicine where previously effective therapeutic options have to be abandoned because of drug resistance creates [sic] obsolescence.

Another consequence of nonspecific antibacterial and antifungal therapy was damage to the human symbiont. There is rapidly accumulating evidence that the human microflora is established early in life through complex steps and that there are individual differences in microbial species composition, a fact that could reflect differences in the timing of acquisition or modulation by the host immune system. The microbial flora is essential for development of the immune system, helps digestion, provides numerous nutrients including vitamins, and protects the human host by niche-denial to more pathogenic microbes. There is conclusive evidence that damage to the microflora by nonspecific antibacterial therapy can translate into antibiotic-associated colitis and fungal diseases such as oral thrush and candidal vaginitis. However, there are ominous signs that nonspecific antimicrobial use might translate into certain chronic diseases such as atopy (Kusel et al., 2008), asthma (Kozyrskyj et al., 2007), and even some types of cancer (Velicer et al., 2004), possibly by altering the development of the immune system in childhood and/or affecting metabolites produced by the microflora. In this regard, it is noteworthy that there is a temporal association between widespread antimicrobial use and the increase in immunoreactive diseases such as allergies and asthma, although it is premature to conclude causality as there may be confounding variables (Wickens et al., 2008). Nevertheless, the available evidence does provide reason for concern.

In summary, the development of effective, nontoxic, nonspecific antibacterial and antifungal therapy has had great consequences, some positive and some negative. Positive consequences include a significantly enhanced capacity to treat bacterial and fungal diseases early and effectively, which has translated to reduced mortality. Furthermore, the ability to treat early, safely, and without knowledge of the causative microbe has created a permissive environment for the development of complex surgeries, aggressive chemotherapy for tumors, and organ transplantation, procedures that would have unacceptable mortality without

such drugs. However, the same approach has also created a culture of empiricism that promoted antibiotic use, which in turn selected for resistance in targeted and nontargeted microbes, promoted the phenomenon of superinfection and damaged the symbiont with consequences that are only now beginning to be understood. In this regard, empiricism was a practice largely dictated by clinical findings and historical probability that essentially rejected causality in favor of associations.

Of course we can do better. Even for the short historical time that effective antimicrobial therapy has been available it is clear that the effectiveness of therapy and diagnosis has fluctuated with time. In a previous essay (Casadevall, 2006), I argued that we are in the throes of a major paradigm shift that will usher in the third age of antimicrobial therapy. This age can be envisioned as an equilateral triangle with pathogen-specific therapy, greatly improved diagnostics, and immunotherapy at each apex. Nonspecific therapy will always have a role for the treatment of polymicrobial diseases and to insure proper coverage in individuals with fulminant disease but its use could be limited by the combination of rapid diagnostics and pathogen-specific drugs. Even for such polymicrobial diseases as abdominal sepsis originating from a ruptured viscus there is evidence that damage is caused by only a few microbial species and their identification would permit employment of pathogen-specific drugs. In this age, immunotherapy, whether with large molecules, such as antibodies or small-molecular-weight immuno-modulators, would have co-equal status with therapies designed to directly kill or inhibit the microbe. Although this author believes that third-age therapeutics will arrive in the twenty-first century, significant scientific, economic, and behavioral hurdles must be overcome for the realization of this vision.

On the scientific front, drug discovery would have to move from trying to identify common therapeutic pathways among phylogenetically distant bacteria to exploiting differences in physiology and virulence mechanisms and/or to augmenting host mechanisms that promote microbial clearance, which, interestingly, are nonspecific. This formidable task is made even more difficult by the economics of antimicrobial drug discovery. As for other diseases, the economics of drug development is a function of the prevalence of the disease, which dictates market size. However, in antimicrobial drug discovery this formula is further modified by the fact that the market size is directly proportional to the width of the drug antimicrobial spectrum. Given the cost of drug development, the economics are stacked against pathogen-specific drugs in favor of broad-spectrum drugs. One caveat in this analysis is that drug resistance can disproportionately shorten the useful life of broad-spectrum drugs and that the emergence of resistant microbes can in itself create new market opportunities. For example, the emergence and spread of methicillin-resistant Staphylococcus aureus (MRSA) creates a niche such that a new staphylococcal-specific drug active against methicillin- and

possibly vancomycin-resistant isolates would probably be developed clinically if available. The use of pathogen-specific drugs would necessitate advances in diagnostics to provide rapid and accurate information to support their use, and this would require new investments in research and laboratory assays. Finally, physicians would have to change their approach to patients with presumed infectious diseases, emphasizing the need for diagnosis to select appropriate therapy in an echo to the practices of physicians in the age of serum therapy.

Perhaps the hurdles are so high that pathogen-specific therapy is only in the far horizon. If that is the case, there are concrete actions that can be taken in the present to slow the spread of drug resistance and damage to the human microbial flora. For example, educational campaigns aimed at physicians and the general public can promote more prudent use of antimicrobial drugs. At a political level, policy makers should be made aware of the economic and regulatory hurdles that slow the development of rapid diagnostic tests and pathogen-specific drugs. However, perhaps things can change more rapidly that one can anticipate. Certainly, if future research was to associate disturbances in the microflora with such chronic diseases as asthma, atopy, and cancer, this would create tremendous medical and legal disincentives in the use of nonspecific microbial therapy. Another powerful force could be the categorization of such complications of broad-spectrum therapy as C. difficile colitis and candidiasis as medical errors, which would be followed by aversion of third-party payers for hospital and physician reimbursements. At the same time, economic incentives for the development of pathogen-specific therapy by industry could be created by linking the patent protection time of antimicrobial drugs to the width of the antimicrobial spectrum and inclusion of narrow-spectrum drugs as orphan drugs. For example, patent policy could be amended such that narrow-spectrum drugs with small markets enjoy much longer patent protection than broad-spectrum drugs. Although in 2009 a revolution in the antimicrobial therapeutic paradigm seems distant, it is worth noting that only a generation ago smoking was widely permitted and accepted in most public places. For smoking, it was the realization that second-hand smoke was dangerous that catalyzed the creation of smoke-free environments in most public places. Perhaps increased awareness of the consequences of long-term damage to the human flora will have a similar catalytic effect in promoting pathogen-specific antimicrobial therapies.

The re-introduction of pathogen-specific therapy for bacterial diseases, and its extension to fungal diseases, would require a concerted effort and collaboration between intellectual leaders in the field, industry, and government to find mechanisms that would promote and encourage the development of such drugs. There are indications of movement in this direction. A recent report by the Institute of Medicine recommended ‘development of strategies that will selectively target pathogenic organisms while avoiding targeting the host and beneficial or benign organisms’, which in other words is pathogen-specific therapy.³ Several therapies

narrow-spectrum are currently in development, for example, the renewed interest in phage therapy, monoclonal antibody therapies, and drugs aimed primarily at targeting highly resistant bacteria. However, the task of refocusing anti-bacterial and antifungal therapy to pathogen specificity is too great for any individual party and cooperation from industry, government, and the medical community will be needed to effect change. There is an acute need for an economic model that would allow the development and use of pathogen-specific drugs. Despite these hurdles, it is clear that pathogen-specific therapy makes sense and, given that the current nonspecific strategies are increasingly bankrupt, it behooves all parties to begin a dialogue on how to get there, and get there sooner than later.

The author states no conflict of interest and has received no payment in preparation of this manuscript.

Buchwald UK, Pirofski L. Immune therapy for infectious diseases at the dawn of the 21st century: the past, present and future role of antibody therapy, therapeutic vaccination and biological response modifiers. Curr Pharm Des 2003;9(12):945-68

Casadevall A. Crisis in Infectious Diseases: Time for a new paradigm? Clin Infect Dis 1996;23:790-4

Casadevall A, Scharff MD. “Serum Therapy” revisited: Animal models of infection and the development of passive antibody therapy. Antimicrob Agents Chemother 1994;38:1695-702

Casadevall A, Scharff MD. Return to the past: the case for antibody-based therapies in infectious diseases. Clin Infect Dis 1995;21:150-61

Casadevall A. The third age of antimicrobial therapy. Clin Infect Dis 2006;42(10):1414-6

Jawetz E. Antimicrobial therapy. Ann Rev Microbiol 1956;10:85-114

Kozyrskyj AL, Ernst P, Becker AB. Increased risk of childhood asthma from antibiotic use in early life. Chest 2007;131(6):1753-9

Kusel MM, de KN, Holt PG, Sly PD. Antibiotic use in the first year of life and risk of atopic disease in early childhood. Clin Exp Allergy 2008;38(12):1921-8

Velicer CM, Heckbert SR, Lampe JW, et al. Antibiotic use in relation to the risk of breast cancer. JAMA 2004;291(7):827-35

Wickens K, Ingham T, Epton M, et al. The association of early life exposure to antibiotics and the development of asthma, eczema and atopy in a birth cohort: confounding or causality? Clin Exp Allergy 2008;38(8):1318-24

Henry F. Chambers⁵ and Frank R. DeLeo⁶

Staphylococcus aureus is notorious for its ability to become resistant to antibiotics. Infections that are caused by antibiotic-resistant strains often occur in epidemic waves that are initiated by one or a few successful clones. Methicillin-resistant S. aureus (MRSA) features prominently in these epidemics. Historically associated with hospitals and other health care settings, MRSA has now emerged as a widespread cause of community infections. Community or community-associated MRSA (CA-MRSA) can spread rapidly among healthy individuals. Outbreaks of CA-MRSA infections have been reported worldwide, and CA-MRSA strains are now epidemic in the United States. Here, we review the molecular epidemiology of the epidemic waves of penicillin- and methicillin-resistant strains of S. aureus that have occurred since 1940, with a focus on the clinical and molecular epidemiology of CA-MRSA.

Staphylococcus aureus is naturally susceptible to virtually every antibiotic that has ever been developed. Resistance to antibiotics is often acquired by the horizontal transfer of genes from outside sources, although chromosomal mutation and antibiotic selection are also important. This exquisite susceptibility of S. aureus led to Alexander Fleming’s discovery of penicillin, which ushered in the ‘antibiotic era’. Penicillin was truly a miracle drug: uniformly fatal infections could now be cured. However, by the mid 1940s, only a few years after its introduction into clinical practice, penicillin resistance was encountered in hospitals, and within a decade it had become a notable problem in the community.

A fundamental biological property of S. aureus is its ability to asymptomatically colonize healthy individuals. Approximately 30% of humans are asymptomatic nasal carriers of S. aureus (Kluytmans and Verbaugh, 1997; Gorwitz et al., 2008) such that in these individuals S. aureus is part of the normal flora. S. aureus

carriers are at higher risk of infection and they are presumed to be an important source of the S. aureus strains that spread among individuals. The primary mode of transmission of S. aureus is by direct contact, usually skin-to-skin contact with a colonized or infected individual, although contact with contaminated objects and surfaces might also have a role (Miller and Diep, 2008; Kazakova et al., 2005; Lowy, 1998; Muto et al., 2003). Various host factors can predispose individuals to infection, including the loss of the normal skin barrier, the presence of underlying diseases such as diabetes or AIDS and defects in neutrophil function.

Infections that are caused by antibiotic-resistant strains of S. aureus have reached epidemic proportions globally (Tiemersma, 2006). The overall burden of staphylococcal disease, particularly disease caused by methicillin-resistant S. aureus (MRSA) strains, is increasing in many countries in both health care and community settings (Kaplan et al., 2005; Hersh et al., 2008; Klevens et al., 2007; Hope et al., 2008; Laupland et al., 2008; European Antimicrobial resistance Surveillance System, 2008). In the United States, the emergence of community associated MRSA (CA-MRSA) strains accounts for much of this increase, as it is a major cause of skin and soft-tissue infections (Moran et al., 2006; Fridkin et al., 2005). The rapidity and extent of the spread of CA-MRSA strains has been remarkable. In addition to the United States, CA-MRSA strains have been reported in Canada, Asia, South America and Australia as well as throughout Europe, including in countries that historically have a low prevalence of MRSA, such as Norway, the Netherlands, Denmark and Finland (Laupland et al., 2008; Larsen et al., 2007; Larsen et al., 2008; Wannet et al., 2005; Deurenberg et al., 2009; Vandenesch et al., 2003; Stam-Bolink et al., 2007; Huang et al., 2007; Nimmo and Coombs, 2008; Kanerva et al., 2009; Park et al., 2009; Gardella et al., 2008; Francois et al., 2008; Fang et al., 2008; Conly and Johnston, 2003). Globally, CA-MRSA strains have shown considerable diversity in the number of different clones that have been identified.

In addition to their increasing prevalence and incidence, CA-MRSA strains seem to be particularly virulent. Overwhelming and tissue-destructive infections, such as necrotizing fasciitis and fulminant, necrotizing pneumonia (Francis et al., 2005; Gonzalez et al., 2005; Kallen et al., 2009), were rarely seen before the emergence of CA-MRSA strains. The factor (or factors) that is responsible for this hypervirulent behaviour is not known, but Panton–Valentine leukocidin (PVL), which has been epidemiologically associated with severe skin infections and pneumonia that are caused by methicillin-susceptible S. aureus (MSSA) strains (Lina et al., 1999), is a leading candidate.

Antibiotics arguably constitute the most concentrated selective pressure on S. aureus in its long coevolutionary history with mankind. The consequences of this selective pressure, in conjunction with horizontal and vertical gene transfer, are discussed in this Review. Given their crucial importance as therapeutic agents, we focus on resistance to penicillins and the structurally related β-lactam antibiotics.

The emergence of antibiotic resistance in S. aureus can be visualized as a series of waves (Figure A2-1). The first wave began in the mid 1940s as the proportion of infections caused by penicillin-resistant strains of S. aureus increased in hospitals (Kirby, 1944; Barber and Rozwadowska-Dowzenko, 1948). These strains produced a plasmid-encoded penicillinase, which hydrolyses the β-lactam ring of penicillin that is essential for its antimicrobial activity. Penicillin-resistant strains soon began to cause community infections, and by the early 1950s they had become pandemic (Roundtree and Freeman, 1956). These infections, both in hospitals and in the community, were frequently caused by an S. aureus clone known as phage type 80/81 (Roundtree and Freeman, 1956; Blair and Carr, 1960; Bynoe et al., 1956; Roundtree and Beard, 1958). Pandemic phage type 80/81 S. aureus infections largely disappeared after the introduction of methicillin (Jevons and Parker, 1964), but the prevalence of penicillinase-producing strains from other S. aureus lineages has remained high.

The introduction of methicillin marks the onset of the second wave of resistance (Figure A2-1). The first reports of a S. aureus strain that was resistant to methicillin were published in 1961 (Barber, 1961; Jevons, 1961). Although the specific gene responsible for methicillin resistance (mecA, which encodes the low-affinity penicillin-binding protein PbP2a (also known as PbP2′)) was not identified until over 20 years later, it was appreciated early on that the resistance mechanism involved was different from penicillinase-mediated resistance because drug inactivation did not occur. Unlike penicillinase-mediated resistance, which is narrow in its spectrum of activity, methicillin resistance is broad, conferring resistance to the entire β-lactam class of antibiotics, which include penicillins, cephalosporins and carbapenems. Among the earliest MRSA clinical isolates was the archetypal MRSA strain COL, a member of the ‘archaic’ clone of MRSA and perhaps the most studied MRSA strain, which was isolated from a patient in Colindale, UK, in 1960 (Jevons, 1961). COL is a member of the most successful MRSA lineage, which includes both hospital and community-associated strains.

Archaic MRSA strains circulated in hospitals throughout Europe until the 1970s (Crisostomo et al., 2001). There were also isolated reports of MRSA in hospitals in the United States (Barrett et al., 1968; Bran et al., 1972), but the rest of the world was largely unaffected, and these early MRSA strains never gained a foothold in the community. By the 1980s, for reasons that remain unclear, the archaic MRSA clone had largely disappeared from European hospitals, marking the end of the second and the beginning of the third wave of antibiotic resistance. Descendants of the archaic MRSA clone (for example, the Iberian and Rome clones (Mato et al., 2004) and other, highly successful MRSA lineages emerged (Enright et al., 2002; Robinson and Enright, 2003; Deurenberg and Stobberingh, 2008) (Table A2-1). Outbreaks of infections caused by MRSA strains were reported in hospitals in the United States in the late 1970s, and by the mid 1980s

these strains were endemic (Crossley et al., 1979; Peacock et al., 1980), leading to the worldwide pandemic of MRSA in hospitals that continues to the present time. Although global in its distribution and impact, MRSA was still confined mainly to hospitals and other institutional health care settings, such as long-term care facilities. The ever-increasing burden of MRSA infections in hospitals led to the increased use of vancomycin, the last remaining antibiotic to which MRSA strains were reliably susceptible. This intensive selective pressure resulted in the emergence of vancomycin-intermediate S. aureus (VISA) strains, which are not inhibited in vitro at vancomycin concentrations below 4–8 μg ml^–1 (Hiramatsu et al., 1997), and vancomcyin-resistant S. aureus (VRSA) strains, which are inhibited only at concentrations of 16 μg ml^–1 or more (Weigel et al., 2003).

The MRSA invasion of the community constitutes the fourth and most recent wave of antibiotic resistance (Figure A2-1). Some of the earliest cases of CA-MRSA infection occurred in indigenous populations in Western Australia in the early 1990s (O’Brien et al., 2004; Coombs et al., 2004; Udo et al., 1993). These MRSA strains were distinguishable from the contemporary clones or genotypes that were circulating in Australian hospitals by their pulsed-field gel electrophoresis patterns and their susceptibility to most antibiotics other than β-lactams, suggesting that they were either remote, ‘feral’ descendants of hospital strains or community strains that had acquired mecA by horizontal gene transfer. In the United States, the first well-documented cases of MRSA infection that were truly community associated occurred in otherwise healthy children from 1997 to 1999 (CDC, 1999). These children had no risk factors for developing MRSA and all died with overwhelming infection, suggesting that these CA-MRSA strains were especially virulent. Like their Australian counterparts, these CA-MRSA isolates were unrelated to hospital associated clones and were susceptible to most antibiotics. The CA-MRSA epidemic in the United States can be traced back to the early 1990s on the basis of retrospective data from 1993 to 1995, which show a dramatic increase in MRSA infections in Chicago among children who lacked risk factors for hospital-associated MRSA exposure (Herold et al., 1998). CA-MRSA has since been reported in numerous populations, including American Indians and Alaskan natives (Baggett et al., 2004), Pacific Islanders (CDC, 2004), athletes (Kazakova et al., 2005), jail and prison inmates (Aiello et al., 2006), men who have sex with men (Diep et al., 2008), contacts of patients with CA-MRSA infection (Johansson et al., 2007), military personnel (Aiello et al., 2006), adult emergency room patients (Moran et al., 2006) and children in day care centres (Adcock et al., 1998). CA-MRSA clones have also gained a foothold in hospitals and are increasingly being identified as a cause of hospital-onset and heath care-associated infections (Klevens et al., 2007; Laupland et al., 2008; Park et al., 2009; Liu et al., 2008; Seybold et al., 2006).

The epidemic wave of CA-MRSA in the United States and Canada (Gilbert et al., 2006; Mulvey et al., 2005) is actually two overlapping epidemics. The USA400 clone, which was isolated from the paediatric cases described above,

was most prevalent before 2001 (Lowy, 1998; CDC, 1999; Stemper et al., 2004) and remains a common cause of community-onset disease among indigenous populations in Alaska and the Pacific Northwest (David et al., 2008). A second epidemic clone, MRSA strain USA300, which is unrelated to USA400 and has largely displaced it in most other locations, emerged between 1999 and 2001 and now causes most of the CA-MRSA infections in the United States (Lowy, 1998; Kazakova et al., 2005; Pan et al., 2003; Pannaraj et al., 2006; Diep et al., 2004; Chavez-Bueno et al., 2005).

Outbreaks and epidemics of CA-MRSA now occur worldwide and have a similar epidemiology, although the specific clones that have emerged vary with geographical location. CA-MRSA strains are not merely escapees from health care facilities; their genotypes indicate that they are not closely related to endemic hospital clones and they are susceptible to numerous antibiotics to which hospital strains are routinely resistant. Two molecular markers that are not found in typical hospital MRSA strains are strongly associated with the emergence of CA-MRSA regardless of geographical origin: a specific cassette element encoding mecA and genes encoding PVL. These markers are discussed in detail below.

Robust, sequence-based molecular methods for genotyping strains of S. aureus, and multilocus sequence typing (MLST) (Enright et al., 2000) in particular, have made it possible to study the evolutionary history of this pathogen (Box A2-1). MLST is carried out by sequence analysis of ~450 bp internal fragments of seven housekeeping genes (Figure A2-2). Isolates that have identical sequences at all seven loci are considered to be a clone and are assigned a unique sequence type (ST). STs that differ by single nucleotide polymorphisms (SNPs) at fewer than three loci are thought to be closely related and are grouped into clonal complexes (CCs). This grouping is accomplished by the eBURST algorithm, which uses MLST data to group closely related strains into a CC. It also predicts the probable founding clone, or ST, of each group and the recent evolutionary descent of all other strains in the CC from the founder (Feil et al., 2004; Turner et al., 2007). The analysis can be further refined to identify specific subclones by the addition of other methods, such as spa typing (Shopsin et al., 1999) or pulsed-field gel electrophoresis of genomic DNA (Box A2-1), or by the presence of other genetic markers (for example, toxin genes or specific plasmids).

Studies of MSSA strains, carriage isolates and hospital and community isolates causing disease that were collected worldwide between 1961 and 2004 show that 88% of the collected strains can be assigned to one of 11 clonal complexes (CC1, CC5, CC8, CC9, CC12, CC15, CC22, CC25, CC30, CC45 and CC51/121) (Enright et al., 2002, 2000; Feil et al., 2004, 2003; Tenover et al., 2008; Goering

et al., 2008; Hallin et al., 2007; Feng et al., 2008; Feil and Enright, 2004; Lindsay et al., 2006) (Figure A2-3a). For ten of these CCs, the percentage of isolates in each complex ranges from 2% to 9%; CC30 is an outlier, accounting for 21% of isolates.

The CCs for contemporary isolates are almost certainly the same as those of strains that were circulating before 1940. For example, the ST5 lineage (the founder of CC5) is estimated to have existed for over 2,000 years (Nubel et al., 2008). Gomes and colleagues (Gomes et al., 2006) genotyped 22 penicillin-susceptible and 67 penicillin-resistant MSSA blood culture isolates that were collected between 1957 and 1973 by the Statens Serum Institute in Copenhagen, which has collected and maintained every blood culture isolate from patients in Denmark from 1957 to the present. They found that 86% of the isolates fell into

seven CCs, the most common being CC8 and CC30, which together accounted for 46% of the isolates (Figure A2-3b). The distributions of penicillin-sensitive and penicillin-resistant isolates were similar. In this analysis, only a few isolates were tested and they all originated from a single country, which probably accounts for the absence of isolates from CC9, CC12, CC15 and CC22.

CC8 and CC30 have given rise to epidemics during each of the four waves of antibiotic resistance. The first well-characterized pandemic of antibiotic-resistant S. aureus that is attributable to a single clone was caused by phage type 80/81 strains, which belong to CC30 (Robinson et al., 2005). Phage type 80/81 strains were originally isolated in Australia in 1953 (Roundtree and Beard, 1958). They are penicillin resistant and have caused both hospital and community outbreaks on a global scale (Robinson et al., 2005). These strains are prevalent in collections that date back to 1927; they were thought to be highly transmissible and particularly virulent and were also among the first to be identified as penicillin resistant (Blair and Carr, 1960). Almost all of the phage type 80/81 isolates in a collection dating to the 1950s and 1960s encode PVL88, which is reminiscent of the association between PVL and resistance to methicillin in the contemporary epidemic CA-MRSA strains. For unknown reasons, phage type 80/81 strains virtually disappeared in the early 1960s, and this coincided with the first use of semisynthetic penicillins, which are resistant to penicillinase. Modern descendents of the CC30 lineage include the PVL-positive southwest Pacific (SWP) clone of CA-MRSA in Australia and the hospital-associated ST36 EMRSA-16 clone, a major cause of nosocomial infections and bacteraemia in both Australia and the United Kingdom (Robinson et al., 2005; Cox et al., 1995; Johnson et al., 2001).

The first MRSA clinical isolates, of which COL is an example, were ST250 and members of CC8. ST250 MRSA strains circulated in the United Kingdom and the rest of Europe before the 1970s but did not become established in the United States and had largely disappeared by the 1980s. However, other highly successful clones emerged, including the ST247 Iberian or EMRSA-5 clone, which is closely related to ST250. No fewer than nine other endemic nosocomial clones are descendants of the ST8 founder of this lineage. The CA-MRSA strain USA300 (which is PVL positive) that is prevalent in the United States is also ST8 (McDougal et al., 2003). MRSA strains have generally been found to be members of a subset of S. aureus CCs, including CC1, CC5, CC8, CC22, CC30 and CC45, although CA-MRSA strains have exhibited some diversity (discussed below). These CCs were widespread before the emergence of methicillin resistance (Crisostomo et al., 2001; Gomes et al., 2006), indicating that superior epidemicity preceded the acquisition of drug resistance and that the adaptations and innovations that make S. aureus clones successful can also favour their adaptation to antibiotic selective pressures.

The discovery by Hiramatsu and colleagues (Ito et al., 2001) that mecA is always found in a mobile cassette element was a great advance for our understanding of the biology of methicillin resistance and provided an additional tool for determining the evolutionary relationships among MRSA strains. Staphylococcal chromosome cassette mec (SCCmec) is integrated into orfX, an S. aureus gene of unknown function (Figure A2-4). To date, eight SCCmec allotypes, designated SCCmecI–SCCmecVIII (Deurenberg and Stobberingh, 2008; Ito et al., 2001; Ma et al., 2002; Oliveira et al., 2006; Higuchi et al., 2008; Zhang et al., 2009), have been described (Table A2-2), along with numerous subtypes, and more will probably be identified as sequence data become available for more MRSA strains (see the SCCmec website for additional descriptions and information). Similar elements are present in coagulase-negative staphylococci, which are commensal organisms that are part of the normal skin flora of humans and other mammals (Ruppe et al., 2009). Two gene complexes, mec and ccr (the recombination and excision locus encoding the gene or genes that mediate the integration and excision of the whole cassette into and out of orfX), are used to classify the SCCmec allotypes (Table A2-2). There are also other differences among the various SCCmec allotypes, particularly in terms of insertion sequences and antimicrobial resistance genes. However, as these are themselves mobile elements, they have not proved useful for the classification of the main allotypes, although they are useful for defining subtypes.

The class A mec gene complex is the prototype complex and is found in SCCmecII (Figure A2-4a), SCCmecIII and SCCmecVIII. It contains mecA, the complete mecR1 and mecI regulatory genes upstream of mecA, and the hypervariable region (HVR) and insertion sequence 431 (IS431) downstream of mecA. The class b mec gene complex is found in SCCmecI, SCCmecIV (Figure A2-4b) and SCCmecVI and is composed of mecA, a truncated mecR1 (resulting from the insertion of IS1272) upstream of mecA, and the HVR and IS431 downstream of mecA. There are two distinct class C mec gene complexes, both of which contain mecA, a truncated mecR1 (resulting from the insertion of IS431) upstream of mecA, and the HVR and IS431 downstream of mecA. In the class C1 mec gene complex, the IS431 elements upstream and downstream of mecA are in the same orientation, whereas in the class C2 mec gene complex, which is found in SCCmecV and SCCmecVII, the orientation of the IS431 upstream of mecA is reversed. C1 and C2 are regarded as different mec gene complexes, as they have probably evolved independently. The mecA, mecR1 and mecI sequences are highly conserved, with >99% nucleotide sequence identity.

The ccr gene complex consists of two adjacent genes, ccrA and ccrB, in SCCmecI–SCCmecIV, SCCmecVI and SCCmecVIII and one gene, ccrC, in SCCmecV and SCCmecVII. MRSA strains that were isolated before 1990, which were all nosocomial isolates, contained predominantly SCCmecI–SCCmecIII.

FIGURE A2-4 Comparison of the methicillin resistance cassettes that are typical of hospital- or community-acquired methicillin-resistant Staphylococcus aureus. Staphylococcal chromosome cassette mecII (SCCmecII) is most abundant in hospitals, whereas SCCmecIV is present in the most abundant community-acquired methicillin-resistant Staphylococcus aureus strains. The mecR1 gene (R1) in SCCmecIV is truncated, whereas the copy in SCCmecII is full-length. Transposon Tn554, which is present in SCCmecII but not in SCCmecIV, encodes resistance to macrolide-lincosomide-streptogramin B antibiotics and spectinmycin. pUB110 is an integrated plasmid that encodes a tobramycin resistance gene. SCCmecII therefore encodes resistance to multiple antibiotics, whereas SCCmecIV encodes resistance to methicillin alone. A, mecA; I mecl; IS431, insertion sequence 431.

CA-MRSA isolates most frequently contain variants of the SCCmecIV or SCCmecIV allotypes; less commonly, they contain SCCmecV (Francois et al., 2008; Okuma et al., 2002). SCCmecIV is also increasingly identified in contemporary hospital MRSA strains.

The three epidemic waves of MRSA correspond to evolutionary changes in SCCmec. The early MRSA strains (COL and other CC8 strains that circulated in the United Kingdom and Denmark in the early 1960s) all carried SCCmecI. They were replaced in the 1980s by new and arguably more successful lineages that eventually became established in hospitals throughout the world. These clones, which were predominantly CC5 and CC8, carried SCCmecII or SCCmecIII (for example, New York/Japan EMRSA, EMRSA-16 in Australia and the United Kingdom, the Brazilian clone and the Hungarian clone), or the type IA variant of the archaic SCCmecI (the Iberian clone). Why SCCmecII and SCCmecIII were more successful than SCCmecI is not known, but it could be that the recombinase genes, which are defective in SCCmecI but functional in SCCmecII and

SCCmecIII92, limited the potential for horizontal gene transfer of SCCmecI into new genomes.

SCCmecIV, which seems to have evolved from SCCmecI (although it has the ccrA and ccrB genes of SCCmecII (Lina et al., 2006)), gave rise to the most recent worldwide epidemic wave of CA-MRSA. Originally identified in the community-associated USA400 strain, MRSA strain MW2, the first occurrence of SCCmecIV in S. aureus might have been in the ST5 ‘paediatric’ clone that was circulating in hospitals in the late 1980s and the 1990s (Oliveira et al., 2001). The ultimate origins of mecA and SCCmec elements might never be known, but there is good evidence suggesting that coagulase negative staphylococci are the sources (Hanssen et al., 2004; Hanssen and Ericson Sollid, 2006; Wu et al., 1996).

The success of SCCmecIV is borne out by two observations. First, it is the most widely distributed SCCmec among S. aureus isolates. It has been found in nine distinct MRSA CCs or STs, whereas there are only two such lineages for SCCmecI, three for SCCmecII and two for SCCmecIII (Lina et al., 2006). Second, CA-MRSA strains containing SCCmecIV have faster growth rates than hospital MRSA strains carrying other SCCmec allotypes, and these growth rates are no different from MSSA isolates. In a rabbit bacteraemia model the fitness and virulence of USA300, which carries SCCmecIVA, were indistinguishable from those

of its isogenic MSSA variant (Diep et al., 2008). Thus, the SCCmecIV seems to confer little or no cost in fitness on the organism.

As mentioned above, the earliest reported cases of CA-MRSA infection in the United States were caused by a USA400 strain, MW2 (CDC, 1999). MW2 is closely related to the PVL-negative clone WA-1, which is an important CA-MRSA clone in Australia, and to the MSSA476 strain in the United Kingdom (Coombs et al., 2004). USA400 has been supplanted by USA300, which is currently by far the most frequent cause of CA-MRSA infections in the United States (Kennedy et al., 2008). The USA300 clone seems to be well adapted to the community, and there are reports of CA-MRSA infections caused by USA300 or its close relatives in Australia, Denmark and Colombia (Bartels et al., 2007; Gottlieb et al., 2008; Arias et al., 2008). USA300 strains can also cause health care-associated infections (Liu et al., 2008; Seybold et al., 2006; Maree et al., 2007; Gonzalez et al., 2006). Although there is evidence for the international spread of USA300 and USA400 (Wannet et al., 2005; Nimmo and Coombs, 2008; Tristan et al., 2007; Larsen et al., 2009), CA-MRSA strains that are not related to either USA300 or USA400 have been responsible for infections outside of the United States. ST80 is the predominant clone circulating in Europe, ST59 is the main clone in Taiwan and ST30 is the most frequent in Eastern Australia, demonstrating that CA-MRSA strains have evolved in separate geographical regions (Stam-Bolink et al., 2007; Huang et al., 2007; Nimmo and Coombs, 2008). There can also be considerable diversity in CA-MRSA strains from country to country. For example, in Australia 45 distinct clones of CA-MRSA have been identified; many of these are related to well-known MRSA lineages, but others seem to be new. The diversity of CA-MRSA isolates has also been noted by other studies (Wannet et al., 2005; Francois et al., 2008; Bartels et al., 2007; Tristan et al., 2007; Larsen et al., 2009). In the United Kingdom, most CA-MRSA infections are caused by EMRSA-15 (ST22) and EMRSA-16 (ST36), which are also important hospital-acquired clones (Rollason et al., 2008); ST80 is also present, but accounts for only a small proportion of isolates (Holmes et al., 2005). A CA-MRSA strain of swine origin that is transmissible to humans, ST398, has also been described (Huijsdens et al., 2006; Loeffler et al., 2009).

The epidemiology of CA-MRSA is similar regardless of the country of origin. Isolates tend not to be resistant to multiple drugs, SCCmecIV or SCCmecV is typically present, and infections of skin and soft tissue are the most common. The presence of PVL among CA-MRSA isolates is more variable. For example, in Australia and the United Kingdom most CA-MRSA clones do not produce PVL (Nimmo and Coombs, 2008; Rollason et al., 2008), and the prevalence of PVL among the more common CA-MRSA isolates from Denmark ranges from 17%

to 100% (Larsen et al., 2009). Conversely, isolates of clones that typically do not carry PVL genes (for example EMRSA-15 and EMRSA-16) have occasionally been found to be PVL-positive.

Nasal carriage of MRSA has increased in parallel with the emergence of MRSA as a community pathogen, which is not unexpected given that approximately 30% of individuals are asymptomatic nasal carriers of S. aureus. Between 2001 and 2004, carriage of MRSA strains in a US population-based study approximately doubled from 0.8% to 1.5% (Gorwitz et al., 2008), and the percentage of CA-MRSA genotypes increased from 7% to 24.2% (Tenover et al., 2008). Although the sites of carriage (for example, nares versus groin versus other sites) and the relationship between the carriage of CA-MRSA strains and disease are not entirely clear, CA-MRSA strains, especially USA300, seem to be more easily transmitted than other strains (Crum et al., 2006), which could account for the increasing carriage rates in the community. Thus, no individual or group can be considered not to be at risk for CA-MRSA infection.

CA-MRSA infections have been associated with fulminant and lethal infections and worse clinical outcomes than are seen with infections caused by health care-associated MRSA strains and community MSSA (Francis et al., 2005; Turner et al., 2007; Davis et al., 2007), giving rise to the impression that CA-MRSA strains, especially USA300, are more virulent than other strains. Much of our understanding of the unique virulence properties of CA-MRSA is based on studies of USA300 strains, the most extensively investigated clone. The USA300 core genome (the chromosome, excluding any mobile genetic elements) is similar to that of the early MRSA strain COL (Diep et al., 2006). However, studies in animal models indicate that USA300 is more virulent than COL (Voyich et al., 2005; Li et al., 2009). The expression of virulence factors by USA300 is high, and this and other closely related strains are more lethal than their more distant relatives and cause more extensive disease in animal models of infection (Li et al., 2009; Montgomery et al., 2008; Wang et al., 2007). The main difference between the COL and USA300 genomes is in their mobile genetic elements, which include prophages, plasmids, pathogenicity islands and transposons that have been acquired through horizontal gene transfer. These elements encode factors that can affect transmission, antibiotic resistance and virulence. Prophages ΦSA2 and ΦSA3, which are present in USA300 strains but not in COL, could contribute to the noted differences in virulence between these two lineages. Prophage ΦSA2 contains lukS–PV and lukF–PV, which encode PVL. Prophage ΦSA3 is present in strains other than CA-MRSA and encodes staphylokinase, staphylococcal complement inhibitor (SCIN) and S. aureus chemotaxis inhibitory protein (CHIPS), all of which are modulators of the innate immune system (Rooijakkers et al., 2006; van Wamel et al., 2006). In addition, USA300 contains

the pathogenicity island SaPI5, which is similar to the island that is present in COL. SaPI5 encodes two superantigens that are not present in COL, staphylococcal enterotoxin Q (SEQ) and staphylococcal enterotoxin K (SEK), which are also found in other MRSA and MSSA lineages. S. aureus produces many other molecules that promote host colonization, facilitate evasion of the innate immune system and alter immune responses (Wang et al., 2007; Deleo et al., 2009; Li et al., 2007) (see Supplementary information S1 [Table A2-S1]). Most of these molecules are not unique to CA-MRSA. The virulence factors that are found more commonly in CA-MRSA than in other strains, that are linked by epidemiology to CA-MRSA infections or that have been studied in animal models of CA-MRSA infection are discussed below.

PVL has been studied extensively since its discovery by Panton and Valentine 70 years ago (Wright, 1936). The role of PVL in the marked epidemicity and enhanced virulence of CA-MRSA is a subject of debate. PVL is composed of two subunits, LukS-PV and LukF-PV (Woodin, 1960), which are encoded by the horizontally acquired prophage ΦSA2 (Kaneko et al., 1998) and are secreted by the bacterium. These subunits bind to specific membrane receptors, which have yet to be identified, and associate to form pores in the membrane of host leukocytes (Meyer et al., 2009; Colin et al., 1994). At high concentrations (for example, 200 nM) PVL causes lytic cell death, but at sublytic concentrations (for example, 5 nM) it seems to partially activate neutrophils in a phenomenon known as priming, as they secrete potent mediators of inflammation, such as leukotriene b4 and interleukin 8, and also cause the release of neutrophil granule contents through exocytosis (Konig et al., 1995; Woodin and Wieneke, 1964; Genestier et al., 2005). In addition, PVL primes neutrophils for the enhanced production of reactive oxygen species on stimulation with the widely used neutrophil agonist fMLP (N-formyl-methionylleucyl-phenylalanine) (Colin and Monteil, 2003). Therefore, PVL could contribute to pathogenesis by causing an exaggerated inflammatory response and injury to the host. Several lines of evidence that are largely circumstantial indicate that PVL is associated with severe skin infections and severe necrotic haemorrhagic pneumonia (Lima et al., 1999; Gillet et al., 2002, 2007). Both USA300, which is now the leading cause of skin and soft tissue infections in the United States and a cause of extremely severe infections, and the penicillin-resistant phage type 80/81 strains that were associated with numerous outbreaks and severe disease in the 1950s produce PVL. The epidemiological association between PVL and the emergence of genetically unrelated CA-MRSA strains (that is, different and unrelated STs) that are geographically dispersed is striking.

There are other observations that call into question the presumption that PVL is driving the CA-MRSA epidemic. First, PVL is found infrequently in

other common, successful community strains. For example, the genes encoding PVL are present in only ~1–10% of MSSA clinical isolates (Goering et al., 2008; Kuehnert et al., 2006; Ellington et al., 2007). Second, although both USA300 and USA400 express PVL, USA300 has become the predominant CA-MRSA clone in the United States. This suggests that factors other than PVL are important for the recent emergence of CA-MRSA.

The experimental evidence does not provide a clear picture either. Voyich et al. (2006) found that USA300 and USA400 wild-type and isogenic PVL-deficient (∆pvl) strains caused virtually identical courses of infection in mouse abscess and sepsis models. Furthermore, there was no difference in neutrophil phagocytosis or lysis after uptake of the bacteria. However, because these experiments were carried out using culture supernatants, the results could reflect the action of multiple lytic factors. Similar results from a rat pneumonia model were reported by Montgomery and Daum (2009). Bubeck Wardenburg et al. (2007, 2008) also showed that USA300 and USA400 wild-type and isogenic ∆pvl strains were equally virulent in mouse abscess and pneumonia models. Diep et al. (2008) used two rabbit bacteraemia models to compare the haematogenous dissemination of wild-type and ∆pvl CA-MRSA strains to major organs: although PVL did not promote seeding of lungs, spleen or blood by USA300, there was a modest, transient contribution of PVL to colonization of the kidneys. In a series of experiments that used the same USA300 wild-type and mutant (∆pvl) strain pair as Voyich et al. (2006), Brown et al. (2008) found that the parent strain was more virulent than the ∆pvl mutant in mouse pneumonia and abscess models and that the disease caused by the wild-type strain was attenuated by immunization with recombinant LukF-PV or LukS-PV. In addition, Labandeira-Rey et al. (2007) found evidence to suggest that PVL might have a role in disease development in a mouse model of staphylococcal pneumonia: direct instillation of high doses of purified toxin provoked an inflammatory response in the lung and reduced survival. The authors used a laboratory strain of S. aureus that had been transduced with PVL-encoding bacteriophage to establish infection, and reported more severe disease in mice infected with this PVL-producing variant than in those infected with the PVL-negative parent. However, in addition to the presence of PVL, this transduced laboratory strain has substantial alterations in global gene expression that confounded the interpretation of the data. As PVL has no impact on protein or gene expression in USA300 or USA400 (Diep et al., 2008), it is possible that factors other than PVL accounted for the experimental results. Taken together, the data suggest that the contribution of PVL to CA-MRSA pathogenesis could be minor or perhaps dependent on an as-yet-unidentified bacterial factor or host susceptibility component.

The pore-forming toxin α-haemolysin (also known as Hla or α-toxin) causes the destruction of a wide range of host cells, including epithelial cells, erythrocytes, fibroblasts and monocytes, and is lethal in animal models when injected in purified form (Bhakdi and Tranum-Jensen, 1991). α-haemolysin is ubiquitous among clinical isolates, although some strains lack an active α-toxin. Recent studies by Bubeck Wardenburg et al. (2007) showed that α-haemolysin is essential for USA300 and USA400 to cause lethal pneumonia in a mouse model of the disease. The amount of this toxin that is produced by these strains in vitro correlates with the severity of the resultant lung disease (Montgomery et al., 2008; Bubeck Wardenburg et al., 2007; Burlak et al., 2007).

α-type phenol-soluble modulins (PSMαs) are a newly discovered group of peptides in S. aureus that are similar to the PSMs of Staphylococcus epidermidis (Wang et al., 2007). High expression of PSMαs might contribute to the enhanced virulence of CA-MRSA; PSMs are produced at higher levels in vitro by prominent CA-MRSA strains, including USA300 and USA400, than by hospital-acquired MRSA strains (Wang et al., 2007). PSMα peptides recruit, activate and ultimately lyse human neutrophils, thereby promoting S. aureus pathogenesis, and greatly contribute to the virulence of USA300 and USA400 in mouse abscess and sepsis models. The study by Wang et al. (2007) was the first to identify molecules from CA-MRSA that could account at least in part for the enhanced virulence of USA300 and USA400.

The arginine catabolic mobile element (ACME) is a 30.9 kb segment of DNA that seems to be unique to USA300 (Diep et al., 2008). This element is adjacent to SCCmecIV and is mobilized by the recombinases that are encoded by SCCmec. It contains two potential virulence factors, a cluster of arginine catabolism (arc) genes that encode an arginine deiminase pathway and opp3, which encodes an oligopeptide permease (Coulter et al., 1998; Degnan et al., 1998). Deletion of ACME but not SCCmec has been shown to decrease the fitness of USA300 in a rabbit bacteraemia model (Diep et al., 2008). Therefore, ACME might contribute to the fitness and epidemic spread of USA300.

Although mobile genetic elements such as ACME are likely to play a part in the transmission of CA-MRSA, there are differences in virulence potential and human disease manifestation even among similar USA300 isolates. For example, Kennedy et al. (2008) used comparative whole-genome sequencing to determine whether USA300 arose by convergent evolution towards a

hypervirulent phenotype or from a recent common ancestor of high virulence potential. Ten USA300 isolates, including some from a wide range of clinical syndromes and from different geographical locations in the United States, were examined. The strains differed from the USA300 reference strain FPR3757 genome by only a few SNPs, ranging from 11 to 408 in number. Phylogenetic analysis indicated that 8 of the strains, differing on average by 32 SNPs from the reference strain and 50 SNPs from each other, clustered with the reference strain and had descended from a recent common ancestor. These nine closely related isolates constitute the epidemic USA300 clone. Eight of the nine strains were ACME positive and all nine contained the same SCCmecIVA subtype. The two other strains were outliers, both lacking ACME and carrying a different SCCmec subtype, type IVB. Unexpectedly, the virulence of the more closely related isolates was variable in animal infection models. Some of these isolates had caused dramatically different disease syndromes in humans (for example, necrotizing pneumonia versus abscesses were caused by isolates that differed by only 23 SNPs), which serves to highlight the importance of host factors in disease presentation and severity.

CA-MRSA has had a marked impact on empirical therapy of suspected staphylococcal infection. Most β-lactam antibiotics, including all orally available agents, can no longer be assumed to be effective for a range of common staphylococcal infections, in particular for skin and soft-tissue infections. In regions where CA-MRSA is prevalent, antimicrobial therapy should be active against MRSA strains. However, there are few clinical data to support the use of agents other than vancomycin, daptomycin or linezolid. Despite a lack of rigorous clinical studies, the oral agents that are recommended for the treatment of CA-MRSA skin and soft-tissue infections include clindamycin, long-acting tetracyclines (doxycycline and minocycline) and trimethoprim–sulphamethoxazole, as well as rifampin and fusidic acid as adjunctive agents to be used in combination (Gorwitz et al., 2006; Barton et al., 2006; Nathwani et al., 2008).

Surgical incision and drainage is the treatment of choice for cutaneous abscesses; adjunctive antimicrobial therapy is of little or no benefit in most of these cases (Moran et al., 2006; Fridkin et al., 2005; Llera and Levy, 1985; Lee et al., 2004). Antibiotic therapy after drainage of CA-MRSA abscesses is not routinely recommended unless the patient has severe or extensive disease, has rapid progression in the presence of associated cellulitis, has symptoms of systemic illness, is very old or very young, has another illness or immune suppression (for example, type I diabetes, HIV infection or neoplastic disease), has an abscess in an area that is difficult to drain or has an abscess that is associated with septic phlebitis (Gorwitz et al., 2006).

Vancomycin is still the preferred drug for the treatment of serious MRSA infections. However, its effectiveness is limited by prolonged, persistent or recurrent bacteraemia during therapy (Khatib et al., 2009; Hawkins et al., 2007), high rates of microbiological and clinical failures (Dombrowski and Winston, 2008), nephrotoxicity (Lodise et al., 2008) and the increasing prevalence of non-susceptible strains (Steimkraus et al., 2007; Wang et al., 2006). Randomized clinical trials of alternative agents, such as linezolid and daptomycin, show that they are comparable or, more precisely, neither inferior nor superior to standard therapy (Arbeit et al., 2004; Shorr et al., 2005; Wunderink et al., 2003; Weigelt et al., 2005; Kaplan et al., 2003; Fowler et al., 2006). Resistance and drug toxicity will remain concerns regardless of the choice of agent.

One or more new compounds that are currently being developed are likely to become available for the treatment of MRSA infections in the near future (Lentino et al., 2008; Pan et al., 2008). Telavancin, dalbavancin and oritavancin are vancomycin derivatives that rapidly kill S. aureus in a concentration-dependent manner in vitro. Whether more rapid killing will translate into an improved efficacy over vancomycin for more serious infections, such as endocarditis or bacteraemia, remains to be determined. Carbapenems and cephalosporins that bind PBP2a, the penicillin-binding protein that mediates methicillin resistance, with much higher affinity than the currently available β-lactams have been developed (Koga et al., 2005). Two cephalosporins, ceftobiprole and ceftaroline, were shown to be clinically effective for the treatment of MRSA skin and soft-tissue infections (Parish and Scheinfeld, 2008; Anderson and Gums, 2008). One drawback with these and the other anti-MRSA β-lactams under development is that they are broad-spectrum antibiotics and are therefore not narrowly targeted treatments of MRSA infection. Further studies are needed to define their eventual role in the therapy of MRSA infections. Moreover, the vancomycin derivatives and anti-MRSA β-lactams, which can only be administered intravenously, do not address the need for orally administered agents. Orally bioavailable oxazolidinones that are active against MRSA are in the early stages of development (Shaw et al., 2008).

Several non-traditional approaches to the treatment and prevention of MRSA infections have been or are still being investigated. These include lysostaphin (Dajcs et al., 2001), antimicrobial peptides (Lawton et al., 2007) and other natural products (for example, tea tree oil) (Stapleton et al., 2007), as well as anti-staphylococcal vaccines (Bubeck Wardenburg and Schneewind, 2008). There are considerable challenges to be faced in the development of these agents, including prohibitively expensive costs, the potential for patient hypersensitivity (caused by the repeated administration of protein products), the short half-lives that are associated with systemic administration and the short-lived or only partially protective immunity that is gained from vaccines, as was the case with an anti-capsular vaccine that proved to be ineffective (Shinefield et al., 2002). These approaches are years away from being available in the clinic, if they make it at all. Prudent

use of the agents that are now available is essential to avoid further erosion of the antimicrobial armamentarium.

S. aureus is an extraordinarily adaptable pathogen with a proven ability to develop resistance. The steady erosion of the effectiveness of β-lactam antibiotics since their first use only 60 years ago is particularly worrying. As we have described, there have been four waves of resistance over the past 60 years. Although the details vary, the basic themes of each successive wave of antibiotic resistance are similar. Often occurring as a consequence of horizontal gene transfer, resistance is initially encountered in hospitals and health care institutions, where the selective pressures for resistance are greatest. Resistant strains are temporarily contained in hospitals but eventually, through a series of modifications and adjustments, they find their way into or arise from within the community to emerge as fully fit and virulent pathogens. Our understanding of the forces that direct the evolution of virulent and drug-resistant organisms is not perfect, but the overuse and misuse of antibiotics is clearly a contributing factor. The discovery and development of new antimicrobials, although necessary, is unlikely to solve the problem of drug resistance for long. New technologies that lead to improved and more rapid diagnostics, a better understanding of the pathogenesis of staphylococcal disease and non-antimicrobial approaches to the prevention and treatment of infection will also be needed to forestall the coming of the post-antibiotic era.

Adcock, P. M., Pastor, P., Medley, F., Patterson, J. E. & Murphy, T. V. Methicillin-resistant Staphylococcus aureus in two child care centers. J. Infect. Dis. 178, 577–580 (1998).

Aiello, A. E., Lowy, F. D., Wright, L. N. & Larson, E. L. Meticillin-resistant Staphylococcus aureus among US prisoners and military personnel: review and recommendations for future studies. Lancet Infect. Dis. 6, 335–341 (2006).

Anderson, S. D. & Gums, J. G. Ceftobiprole: an extended-spectrum anti-methicillin-resistant Staphylococcus aureus cephalosporin. Ann. Pharmacother. 42, 806–816 (2008).

Arbeit, R. D., Maki, D., Tally, F. P., Campanaro, E. & Eisenstein, B. I. The safety and efficacy of daptomycin for the treatment of complicated skin and skin-structure infections. Clin. Infect. Dis. 38, 1673–1681 (2004).

Arias, C. A. et al. MRSA USA300 clone and VREF—a US–Colombian connection? N. Engl. J. Med. 359, 2177–2179 (2008).

Baggett, H. C. et al. Community-onset methicillin-resistant Staphylococcus aureus associated with antibiotic use and the cytotoxin Panton-Valentine leukocidin during a furunculosis outbreak in rural Alaska. J. Infect. Dis. 189, 1565–1573 (2004).

Barrett, F. F., McGehee, R. F. Jr & Finland, M. Methicillin-resistant Staphylococcus aureus at Boston City Hospital. Bacteriologic and epidemiologic observations. N. Engl J. Med. 279, 441–448 (1968).

Barber, M. Methicillin-resistant staphylococci. J. Clin. Pathol. 14, 385–393 (1961).

Barber, M. & Rozwadowska-Dowzenko, M. Infection by penicillin-resistant staphylococci. Lancet 1, 641–644 (1948).

Bartels, M. D., Boye, K., Rhod Larsen, A., Skov, R. & Westh, H. Rapid increase of genetically diverse methicillin-resistant Staphylococcus aureus, Copenhagen, Denmark. Emerg. Infect. Dis. 13, 1533–1540 (2007).

Barton, M. et al. Guidelines for the prevention and management of community-acquired methicillinresistant Staphylococcus aureus: a perspective for Canadian health care practitioners. Can. J. Infect. Dis. Med. Microbiol. 17 (Suppl. C), 4–24 (2006).

Bhakdi, S. & Tranum-Jensen, J. Alpha-toxin of Staphylococcus aureus. Microbiol. Rev. 55, 733–751 (1991).

Blair, J. E. & Carr, M. Distribution of phage groups of Staphylococcus aureus in the years 1927 through 1947. Science 132, 1247–1248 (1960).

Bran, J. L., Levison, M. E. & Kaye, D. Survey for methicillin-resistant staphylococci. Antimicrob. Agents Chemother. 1, 235–236 (1972).

Brown, E. L. et al. The Panton-Valentine leukocidin vaccine protects mice against lung and skin infections caused by Staphylococcus aureus USA300. Clin. Microbiol. Infect. 15, 156–164 (2008).

Bubeck Wardenburg, J., Bae, T., Otto, M., Deleo, F. R. & Schneewind, O. Poring over pores: α-hemolysin and Panton-Valentine leukocidin in Staphylococcus aureus pneumonia. Nature Med. 13, 1405–1406 (2007).

Bubeck Wardenburg, J. & Schneewind, O. Vaccine protection against Staphylococcus aureus pneumonia. J. Exp. Med. 205, 287–294 (2008).

Bubeck Wardenburg, J., Palazzolo-Ballance, A. M., Otto, M., Schneewind, O. & DeLeo, F. R. Panton-Valentine leukocidin is not a virulence determinant in murine models of community-associated methicillin-resistant Staphylococcus aureus disease. J. Infect. Dis. 198, 1166–1170 (2008).

Burlak, C. et al. Global analysis of community-associated methicillin-resistant Staphylococcus aureus exoproteins reveals molecules produced in vitro and during infection. Cell. Microbiol. 9, 1172–1190 (2007).

Bynoe, E. T., Elder, R. H. & Comtois, R. D. Phagetyping and antibiotic-resistance of staphylococci isolated in a general hospital. Can. J. Microbiol. 2, 346–358 (1956).

CDC. Community-associated methicillin-resistant Staphylococcus aureus infections in Pacific Islanders—Hawaii, 2001–2003. MMWR Morb. Mortal. Wkly Rep. 53, 767–770 (2004).

CDC. Four pediatric deaths from community-acquired methicillin-resistant Staphylococcus aureus— Minnesota and North Dakota, 1997–1999. MMWR Morb. Mortal. Wkly Rep. 48, 707–710 (1999).

Chavez-Bueno, S. et al. Inducible clindamycin resistance and molecular epidemiologic trends of pediatric community-acquired methicillin-resistant Staphylococcus aureus in Dallas, Texas. Antimicrob. Agents Chemother. 49, 2283–2288 (2005).

Colin, D. A. & Monteil, H. Control of the oxidative burst of human neutrophils by staphylococcal leukotoxins. Infect. Immun. 71, 3724–3729 (2003).

Colin, D. A., Mazurier, I., Sire, S. & Finck-Barbancon, V. Interaction of the two components of leukocidin from Staphylococcus aureus with human polymorphonuclear leukocyte membranes: sequential binding and subsequent activation. Infect. Immun. 62, 3184–3188 (1994).

Conly, J. M. & Johnston, B. L. The emergence of methicillin-resistant Staphylococcus aureus as a community-acquired pathogen in Canada. Can. J. Infect. Dis. 14, 249–251 (2003).

Coombs, G. W. et al. Genetic diversity among community methicillin-resistant Staphylococcus aureus strains causing outpatient infections in Australia. J. Clin. Microbiol. 42, 4735–4743 (2004).

Coulter, S. N. et al. Staphylococcus aureus genetic loci impacting growth and survival in multiple infection environments. Mol. Microbiol. 30, 393–404 (1998).

Cox, R. A., Conquest, C., Mallaghan, C. & Marples, R. R. A major outbreak of methicillin-resistant Staphylococcus aureus caused by a new phage-type (EMRSA-16). J. Hosp. Infect. 29, 87–106 (1995).

Crisostomo, M. I. et al. The evolution of methicillin resistance in Staphylococcus aureus: similarity of genetic backgrounds in historically early methicillin susceptible and -resistant isolates and contemporary epidemic clones. Proc. Natl Acad. Sci. USA 98, 9865–9870 (2001).

Crossley, K., Landesman, B. & Zaske, D. An outbreak of infections caused by strains of Staphylococcus aureus resistant to methicillin and aminoglycosides. II.Epidemiologic studies. J. Infect. Dis. 139, 280–287(1979).

Crum, N. F. et al. Fifteen-year study of the changing epidemiology of methicillin-resistant Staphylococcus aureus. Am. J. Med. 119, 943–951 (2006).

Dajcs, J. J. et al. Lysostaphin is effective in treating methicillin-resistant Staphylococcus aureus endophthalmitis in the rabbit. Curr. Eye Res. 22, 451–457 (2001).

David, M. Z., Rudolph, K. M., Hennessy, T. W., Boyle- Vavra, S. & Daum, R. S. Molecular epidemiology of methicillin-resistant Staphylococcus aureus, rural southwestern Alaska. Emerg. Infect. Dis. 14, 1693–1699 (2008).

Davis, S. L. et al. Epidemiology and outcomes of community-associated methicillin-resistant Staphylococcus aureus infection. J. Clin. Microbiol. 45, 1705–1711 (2007).

Degnan, B. A. et al. Inhibition of human peripheral blood mononuclear cell proliferation by Streptococcus pyogenes cell extract is associated with arginine deiminase activity. Infect. Immun. 66, 3050–3058 (1998).

Deleo, F. R., Diep, B. A. & Otto, M. Host defense and pathogenesis in Staphylococcus aureus infections. Infect. Dis. Clin. North Am. 23, 17–34 (2009). Review of the virulence factors found in S. aureus.

Deurenberg, R. H. & Stobberingh, E. E. The evolution of Staphylococcus aureus. Infect. Genet. Evol. 8, 747–763 (2008).

Deurenberg, R. H. et al. Cross-border dissemination of methicillin-resistant Staphylococcus aureus, Euregio Meuse-Rhin region. Emerg. Infect. Dis. 15, 727–734 (2009).

Diep, B. A. et al. The arginine catabolic mobile element and staphylococcal chromosomal cassette mec linkage: convergence of virulence and resistance in the USA300 clone of methicillin-resistant Staphylococcus aureus. J. Infect. Dis. 197, 1523–1530 (2008).

Diep, B. A., Sensabaugh, G. F., Somboona, N. S., Carleton, H. A. & Perdreau-Remington, F. Widespread skin and soft-tissue infections due to two methicillinresistant Staphylococcus aureus strains harboring the genes for Panton-Valentine leucocidin. J. Clin. Microbiol. 42, 2080–2084 (2004).

Diep, B. A. et al. Emergence of multidrug-resistant, community-associated, methicillin-resistant Staphylococcus aureus clone USA300 in men who have sex with men. Ann. Intern. Med. 148, 249–257 (2008).

Diep, B. A. et al. Complete genome sequence of USA300, an epidemic clone of community-acquired meticillin-resistant Staphylococcus aureus. Lancet 367, 731–739 (2006). Comparative genomics of USA300 and other MRSA strains.

Diep, B. A. et al. Contribution of Panton-Valentine leukocidin in community-associated methicillin-resistant Staphylococcus aureus pathogenesis. PLoS ONE 3, e3198 (2008).

Dombrowski, J. C. & Winston, L. G. Clinical failures of appropriately-treated methicillin-resistant Staphylococcus aureus infections. J. Infect. 57, 110–115 (2008).

Enright, M. C. et al. The evolutionary history of methicillin-resistant Staphylococcus aureus (MRSA). Proc. Natl Acad. Sci. USA 99, 7687–7692 (2002). Description of the MRSA clones and SCCmec allotypes present in a worldwide collection of mainly nosocomial isolates.

Enright, M. C., Day, N. P., Davies, C. E., Peacock, S. J. & Spratt, B. G. Multilocus sequence typing for characterization of methicillin-resistant and methicillin susceptible clones of Staphylococcus aureus. J. Clin. Microbiol. 38, 1008–1015 (2000). Description of the MLST method and how it can be applied to elucidate the population structure of S. aureus.

Ellington, M. J. et al. Is Panton-Valentine leucocidin associated with the pathogenesis of Staphylococcus aureus bacteraemia in the UK? J. Antimicrob. Chemother. 60, 402–405 (2007).

European Antimicrobial Resistance Surveillance System. Annual Report 2007. (EARSS, Bilthoven, 2008).

Fang, H., Hedin, G., Li, G. & Nord, C. E. Genetic diversity of community-associated methicillin-resistant Staphylococcus aureus in southern Stockholm, 2000–2005. Clin. Microbiol. Infect. 14, 370–376 (2008).

Feil, E. J. & Enright, M. C. Analyses of clonality and the evolution of bacterial pathogens. Curr. Opin. Microbiol. 7, 308–313 (2004).

Feil, E. J., Li, B. C., Aanensen, D. M., Hanage, W. P. & Spratt, B. G. eBURST: inferring patterns of evolutionary descent among clusters of related bacterial genotypes from multilocus sequence typing data. J. Bacteriol. 186, 1518–1530 (2004).

Feil, E. J. et al. How clonal is Staphylococcus aureus? J. Bacteriol. 185, 3307–3316 (2003).

Feng, Y. et al. Evolution and pathogenesis of Staphylococcus aureus: lessons learned from genotyping and comparative genomics. FEMS Microbiol. Rev. 32, 23–37 (2008).

Fowler, V. G. Jr et al. Daptomycin versus standard therapy for bacteremia and endocarditis caused by Staphylococcus aureus. N. Engl. J. Med. 355, 653–665 (2006).

Francis, J. S. et al. Severe community-onset pneumonia in healthy adults caused by methicillin-resistant Staphylococcus aureus carrying the Panton-Valentine leukocidin genes. Clin. Infect. Dis. 40, 100–107 (2005).

Francois, P. et al. Methicillin-resistant Staphylococcus aureus, Geneva, Switzerland, 1993–2005. Emerg. Infect. Dis. 14, 304–307 (2008).

Fridkin, S. K. et al. Methicillin-resistant Staphylococcus aureus disease in three communities. N. Engl. J. Med. 352, 1436–1444 (2005). First large study characterizing the outbreak of CA-MRSA that was caused by USA300 in the United States.

Gardella, N. et al. Community-associated methicillin-resistant Staphylococcus aureus, eastern Argentina. Diagn. Microbiol. Infect. Dis. 62, 343–347 (2008).

Genestier, A. L. et al. Staphylococcus aureus Panton- Valentine leukocidin directly targets mitochondria and induces Bax-independent apoptosis of human neutrophils. J. Clin. Invest. 115, 3117–3127 (2005).

Gilbert, M. et al. Outbreak in Alberta of community acquired (USA300) methicillin-resistant Staphylococcus aureus in people with a history of drug use, homelessness or incarceration. Can. Med. Assoc. J. 175, 149–154 (2006).

Gillet, Y. et al. Association between Staphylococcus aureus strains carrying gene for Panton-Valentine leukocidin and highly lethal necrotising pneumonia in young immunocompetent patients. Lancet 359, 753–759 (2002).

Gillet, Y. et al. Factors predicting mortality in necrotizing community-acquired pneumonia caused by Staphylococcus aureus containing Panton-Valentine leukocidin. Clin. Infect. Dis. 45, 315–321 (2007).

Goering, R. V. et al. Molecular epidemiology of methicillin-resistant and methicillin-susceptible Staphylococcus aureus isolates from global clinical trials. J. Clin. Microbiol. 46, 2842–2847 (2008).

Gomes, A. R., Westh, H. & de Lencastre, H. Origins and evolution of methicillin-resistant Staphylococcus aureus clonal lineages. Antimicrob. Agents Chemother. 50, 3237–3244 (2006). Analysis of penicillin-susceptible and penicillin resistant genotypes of S. aureus, carried out before the emergence of MRSA.

Gonzalez, B. E. et al. Community-associated strains of methicillin-resistant Staphylococcus aureus as the cause of healthcare-associated infection. Infect. Control Hosp. Epidemiol. 27, 1051–1056 (2006).

Gonzalez, B. E. et al. Pulmonary manifestations in children with invasive community-acquired Staphylococcus aureus infection. Clin. Infect. Dis. 41, 583–590 (2005).

Gorwitz, R. J. et al. Strategies for clinical management of MRSA in the community: summary of an expert’s meeting convened by the Centers for Disease Control and Prevention. CDC [online], http://www.cdc.gov/ncidod/dhqp/ar_mrsa_ca.html (2006).

Gottlieb, T., Su, W. Y., Merlino, J. & Cheong, E. Y. Recognition of USA300 isolates of community-acquired methicillin-resistant Staphylococcus aureus in Australia. Med. J. Aust. 189, 179–180 (2008).

Gorwitz, R. J. et al. Changes in the prevalence of nasal colonization with Staphylococcus aureus in the United States, 2001–2004. J. Infect. Dis. 197, 1226–1234 (2008).

Hallin, M. et al. Genetic relatedness between methicillin-susceptible and methicillin-resistant Staphylococcus aureus: results of a national survey. J. Antimicrob. Chemother. 59, 465–472 (2007).

Hanssen, A. M., Kjeldsen, G. & Sollid, J. U. Local variants of staphylococcal cassette chromosome mec in sporadic methicillin-resistant Staphylococcus aureus and methicillin-resistant coagulase-negative staphylococci: evidence of horizontal gene transfer? Antimicrob. Agents Chemother. 48, 285–296 (2004).

Hanssen, A. M. & Ericson Sollid, J. U. SCCmec in staphylococci: genes on the move. FEMS Immunol. Med. Microbiol. 46, 8–20 (2006).

Hawkins, C. et al. Persistent Staphylococcus aureus bacteremia: an analysis of risk factors and outcomes. Arch. Intern. Med. 167, 1861–1867 (2007).

Herold, B. C. et al. Community-acquired methicillinresistant Staphylococcus aureus in children with no identified predisposing risk. JAMA 279, 593–598 (1998). A report of CA-MRSA in children in Chicago, which stimulated an awareness of the scope of the epidemic.

Hersh, A. L., Chambers, H. F., Maselli, J. H. & Gonzales, R. National trends in ambulatory visits and antibiotic prescribing for skin and soft-tissue infections. Arch. Intern. Med. 168, 1585–1591 (2008).

Higuchi, W., Takano, T., Teng, L. J. & Yamamoto, T. Structure and specific detection of staphylococcal cassette chromosome mec type VII. Biochem. Biophys. Res. Commun. 377, 752–756 (2008).

Hiramatsu, K. et al. Dissemination in Japanese hospitals of strains of Staphylococcus aureus heterogeneously resistant to vancomycin. Lancet 350, 1670–1673 (1997).

Holmes, A. et al. Staphylococcus aureus isolates carrying Panton-Valentine leucocidin genes in England and Wales: frequency, characterization, and association with clinical disease. J. Clin. Microbiol. 43, 2384–2390 (2005).

Hope, R., Livermore, D. M., Brick, G., Lillie, M. & Reynolds, R. Non-susceptibility trends among staphylococci from bacteraemias in the UK and Ireland, 2001–2006. J. Antimicrobiol. Chemother. 62 (Suppl. 2), 65–74 (2008).

Huang, Y. C., Hwang, K. P., Chen, P. Y., Chen, C. J. & Lin, T. Y. Prevalence of methicillin-resistant Staphylococcus aureus nasal colonization among Taiwanese children in 2005 and 2006. J. Clin. Microbiol. 45, 3992–3995 (2007).

Huijsdens, X. W. et al. Community-acquired MRSA and pig-farming. Ann. Clin. Microbiol. Antimicrob. 5, 26 (2006).

Ito, T. et al. Structural comparison of three types of staphylococcal cassette chromosome mec integrated in the chromosome in methicillin-resistant Staphylococcus aureus. Antimicrob. Agents Chemother. 45, 1323–1336 (2001). Comparison of the genetic structure and organization of SCCmecI, SCCmecII and SCCmecIII.

Jevons, M. P. & Parker, M. T. The evolution of new hospital strains of Staphylococcus aureus. J. Clin. Pathol. 17, 243–250 (1964).

Jevons, M. “Celbenin”-resistant staphylococci. BMJ 1, 124–125 (1961).

Johansson, P. J., Gustafsson, E. B. & Ringberg, H. High prevalence of MRSA in household contacts. Scand. J. Infect. Dis. 39, 764–768 (2007).

Johnson, A. P. et al. Dominance of EMRSA-15 and -16 among MRSA causing nosocomial bacteraemia in the UK: analysis of isolates from the European Antimicrobial Resistance Surveillance System (EARSS). J. Antimicrob. Chemother. 48, 143–144 (2001).

Kallen, A. J. et al. Staphylococcus aureus community acquired pneumonia during the 2006 to 2007 influenza season. Ann. Emerg. Med. 53, 358–365 (2009).

Kaneko, J., Kimura, T., Narita, S., Tomita, T. & Kamio, Y. Complete nucleotide sequence and molecular characterization of the temperate staphylococcal bacteriophage ΦPVL carrying Panton-Valentine leukocidin genes. Gene 215, 57–67 (1998).

Kanerva, M. et al. Community-associated methicillinresistant Staphylococcus aureus, isolated in Finland in 2004 to 2006. J. Clin. Microbiol. 7, 2655–2657 (2009).

Kaplan, S. L. et al. Linezolid versus vancomycin for treatment of resistant Gram-positive infections in children. Pediatr. Infect. Dis. J. 22, 677–686 (2003).

Kaplan, S. L. et al. Three-year surveillance of community-acquired Staphylococcus aureus infections in children. Clin. Infect. Dis. 40, 1785–1791 (2005).

Kazakova, S. V. et al. A clone of methicillin-resistant Staphylococcus aureus among professional football players. N. Engl. J. Med. 352, 468–475 (2005).

Kennedy, A. D. et al. Epidemic community-associated methicillin-resistant Staphylococcus aureus: recent clonal expansion and diversification. Proc. Natl Acad. Sci. USA 105, 1327–1332 (2008). Deep sequence analysis of closely related USA300 strains and a comparison of their virulence in a mouse model.

Khatib, R. et al. Persistent Staphylococcus aureus bacteremia: incidence and outcome trends over time. Scand. J. Infect. Dis. 41, 4–9 (2009).

Kirby, W. Extraction of a highly potent penicillin inactivator from penicillin resistant staphylococci. Science 99, 452–453 (1944).

Klevens, R. M. et al. Invasive methicillin-resistant Staphylococcus aureus infections in the United States. JAMA 298, 1763–1771 (2007).

Kluytmans, J., van Belkum, A. & Verbrugh, H. Nasal carriage of Staphylococcus aureus: epidemiology, underlying mechanisms, and associated risks. Clin. Microbiol. Rev. 10, 505–520 (1997). Review of S. aureus colonization of humans.

Koga, T. et al. In vitro and in vivo antibacterial activities of CS-023 (RO4908463), a novel parenteral carbapenem. Antimicrob. Agents Chemother. 49, 3239–3250 (2005).

Konig, B., Prevost, G., Piemont, Y. & Konig, W. Effects of Staphylococcus aureus leukocidins on inflammatory mediator release from human granulocytes. J. Infect. Dis. 171, 607–613 (1995).

Kuehnert, M. J. et al. Prevalence of Staphylococcus aureus nasal colonization in the United States, 2001–2002. J. Infect. Dis. 193, 172–179 (2006).

Labandeira-Rey, M. et al. Staphylococcus aureus Panton-Valentine leukocidin causes necrotizing pneumonia. Science 315, 1130–1133 (2007).

Larsen, A. R. et al. Emergence and characterization of community-associated methicillin-resistant Staphyloccocus aureus infections in Denmark, 1999 to 2006. J. Clin. Microbiol. 47, 73–78 (2009).

Larsen, A., Stegger, M., Goering, R., Sorum, M. & Skov, R. Emergence and dissemination of the methicillin resistant Staphylococcus aureus USA300 clone in Denmark (2000–2005). Euro. Surveill. 12, 22–24 (2007).

Larsen, A. R. et al. Epidemiology of European community-associated methicillin-resistant Staphylococcus aureus clonal complex 80 type IV strains isolated in Denmark from 1993 to 2004. J. Clin. Microbiol. 46, 62–68 (2008).

Laupland, K. B., Ross, T. & Gregson, D. B. Staphylococcus aureus bloodstream infections: risk factors, outcomes, and the influence of methicillin resistance in Calgary, Canada, 2000–2006. J. Infect. Dis. 198, 336–343 (2008).

Lawton, E. M., Ross, R. P., Hill, C. & Cotter, P. D. Two-peptide lantibiotics: a medical perspective. Mini Rev. Med. Chem. 7, 1236–1247 (2007).

Lee, M. C. et al. Management and outcome of children with skin and soft tissue abscesses caused by community-acquired methicillin-resistant Staphylococcus aureus. Pediatr. Infect. Dis. J. 23, 123–127 (2004).

Lentino, J. R., Narita, M. & Yu, V. L. New antimicrobial agents as therapy for resistant gram-positive cocci. Eur. J. Clin. Microbiol. Infect. Dis. 27, 3–15 (2008).

Li, M. et al. The antimicrobial peptide-sensing system aps of Staphylococcus aureus. Mol. Microbiol. 66, 1136–1147 (2007).

Li, M. et al. Evolution of virulence in epidemic community-associated MRSA. Proc. Natl Acad. Sci. USA 106, 5883–5888 (2009).

Lina, G. et al. Staphylococcal chromosome cassette evolution in Staphylococcus aureus inferred from ccr gene complex sequence typing analysis. Clin. Microbiol. Infect. 12, 1175–1184 (2006). Sequence typing of SCCmec allotypes to define possible origins and evolution.

Lina, G. et al. Involvement of Panton-Valentine leukocidin-producing Staphylococcus aureus in primary skin infections and pneumonia. Clin. Infect. Dis. 29, 1128–1132 (1999). Epidemiological study suggesting that PVL is an important virulence factor in severe pneumonia.

Lindsay, J. A. et al. Microarrays reveal that each of the ten dominant lineages of Staphylococcus aureus has a unique combination of surface-associated and regulatory genes. J. Bacteriol. 188, 669–676 (2006).

Liu, C. et al. A population-based study of the incidence and molecular epidemiology of methicillin-resistant Staphylococcus aureus disease in San Francisco, 2004–2005. Clin. Infect. Dis. 46, 1637–1646 (2008). A population-based study of the USA300 epidemic in San Francisco, a city with a high prevalence of CA-MRSA.

Llera, J. L. & Levy, R. C. Treatment of cutaneous abscess: a double-blind clinical study. Ann. Emerg. Med. 14, 15–19 (1985).

Lodise, T. P., Lomaestro, B., Graves, J. & Drusano, G. L. Larger vancomycin doses (at least four grams per day) are associated with an increased incidence of nephrotoxicity. Antimicrob. Agents Chemother. 52, 1330–1336 (2008).

Loeffler, A. et al. First isolation of MRSA ST398 from UK animals: a new challenge for infection control teams? J. Hosp. Infect. 72, 269–271 (2009).

Lowy, F. D. Staphylococcus aureus infections. N. Engl. J. Med. 339, 520–532 (1998).

Ma, X. X. et al. Novel type of staphylococcal cassette chromosome mec identified in community-acquired methicillin-resistant Staphylococcus aureus strains. Antimicrob. Agents Chemother. 46, 1147–1152 (2002). Genetic structure and organization of SCCmecIV.

Maree, C. L., Daum, R. S., Boyle-Vavra, S., Matayoshi, K. & Miller, L. G. Community-associated methicillin-resistant Staphylococcus aureus isolates causing healthcare-associated infections. Emerg. Infect. Dis. 13, 236–242 (2007).

Mato, R. et al. Clonal types and multidrug resistance patterns of methicillin-resistant Staphylococcus aureus (MRSA) recovered in Italy during the 1990s. Microb. Drug Resist. 10, 106–113 (2004).

McDougal, L. K. et al. Pulsed-field gel electrophoresis typing of oxacillin-resistant Staphylococcus aureus isolates from the United States: establishing a national database. J. Clin. Microbiol. 41, 5113–5120 (2003).

Meyer, F., Girardot, R., Piemont, Y., Prevost, G. & Colin, D. A. Analysis of the specificity of Panton-Valentine leucocidin and gamma-hemolysin F component binding. Infect. Immun. 77, 266–273 (2009).

Miller, L. G. & Diep, B. A. Clinical practice: colonization, fomites, and virulence: rethinking the pathogenesis of community-associated methicillin-resistant Staphylococcus aureus infection. Clin. Infect. Dis. 46, 752–760 (2008).

Montgomery, C. P. & Daum, R. S. Transcription of inflammatory genes in the lung after infection with community-associated methicillin-resistant Staphylococcus aureus: a role for Panton-Valentine leukocidin? Infect. Immun. 77, 2159–2167 (2009).

Montgomery, C. P. et al. Comparison of virulence in community-associated methicillin-resistant Staphylococcus aureus pulsotypes USA300 and USA400 in a rat model of pneumonia. J. Infect. Dis. 198, 561–570 (2008).

Moran, G. J. et al. Methicillin-resistant S. aureus infections among patients in the emergency department. N. Engl. J. Med. 355, 666–674 (2006).

Mulvey, M. R. et al. Community-associated methicillin-resistant Staphylococcus aureus, Canada. Emerg. Infect. Dis. 11, 844–850 (2005).

Muto, C. A. et al. SHEA guideline for preventing nosocomial transmission of multidrug-resistant strains of Staphylococcus aureus and Enterococcus. Infect. Control Hosp. Epidemiol. 24, 362–386 (2003).

Nathwani, D. et al. Guidelines for UK practice for the diagnosis and management of methicillin-resistant Staphylococcus aureus (MRSA) infections presenting in the community. J. Antimicrob. Chemother. 61, 976–994 (2008).

Nimmo, G. R. & Coombs, G. W. Community-associated methicillin-resistant Staphylococcus aureus (MRSA) in Australia. Int. J. Antimicrob. Agents 31, 401–410 (2008).

Nubel, U. et al. Frequent emergence and limited geographic dispersal of methicillin-resistant Staphylococcus aureus. Proc. Natl Acad. Sci. USA 105, 14130–14135 (2008). Evidence that MRSA infections are locally derived as opposed to internationally translocated, and that SCCmec has entered S. aureus strains on numerous occasions.

O’Brien, F. G. et al. Diversity among community isolates of methicillin-resistant Staphylococcus aureus in Australia. J. Clin. Microbiol. 42, 3185–3190 (2004).

Okuma, K. et al. Dissemination of new methicillin-resistant Staphylococcus aureus clones in the community. J. Clin. Microbiol. 40, 4289–4294 (2002).

Oliveira, D. C., Milheirico, C. & de Lencastre, H. Redefining a structural variant of staphylococcal cassette chromosome mec, SCCmec type VI. Antimicrob. Agents Chemother. 50, 3457–3459 (2006).

Oliveira, D. C., Tomasz, A. & de Lencastre, H. The evolution of pandemic clones of methicillin-resistant Staphylococcus aureus: identification of two ancestral genetic backgrounds and the associated mec elements. Microb. Drug Resist. 7, 349–361 (2001).

Pan, E. S. et al. Increasing prevalence of methicillin-resistant Staphylococcus aureus infection in California jails. Clin. Infect. Dis. 37, 1384–1388 (2003). The first description of USA300.

Pan, A., Lorenzotti, S. & Zoncada, A. Registered and investigational drugs for the treatment of methicillinresistant Staphylococcus aureus infection. Recent Pat. Antiinfect. Drug Discov. 3, 10–33 (2008).

Pannaraj, P. S., Hulten, K. G., Gonzalez, B. E., Mason, E. O. Jr & Kaplan, S. L. Infective pyomyositis and myositis in children in the era of community-acquired, methicillin-resistant Staphylococcus aureus infection. Clin. Infect. Dis. 43, 953–960 (2006).

Parish, D. & Scheinfeld, N. Ceftaroline fosamil, a cephalosporin derivative for the potential treatment of MRSA infection. Curr. Opin. Investig. Drugs 9, 201–209 (2008).

Park, S. H. et al. Emergence of community-associated methicillin-resistant Staphylococcus aureus strains as a cause of healthcare-associated bloodstream infections in Korea. Infect. Control Hosp. Epidemiol. 30, 146–155 (2009).

Peacock, J. E. Jr, Marsik, F. J. & Wenzel, R. P. Methicillin-resistant Staphylococcus aureus: introduction and spread within a hospital. Ann. Intern. Med. 93, 526–532 (1980).

Robinson, D. A. et al. Re-emergence of early pandemic Staphylococcus aureus as a community-acquired meticillin-resistant clone. Lancet 365, 1256–1258 (2005).

Robinson, D. A. & Enright, M. C. Evolutionary models of the emergence of methicillin-resistant Staphylococcus aureus. Antimicrob. Agents Chemother. 47, 3926–3934 (2003).

Rollason, J. et al. Epidemiology of community-acquired meticillin-resistant Staphylococcus aureus obtained from the UK West Midlands region. J. Hosp. Infect. 70, 314–320 (2008).

Rooijakkers, S. H. et al. Early expression of SCIN and CHIPS drives instant immune evasion by Staphylococcus aureus. Cell. Microbiol. 8, 1282–1293 (2006).

Roundtree, P. & Freeman, V. Infections caused by a particular phage type of Staphylococcus aureus. Med. J. Aust. 42, 157–161 (1956).

Roundtree, P. & Beard, M. Further observations on infections with phage type 80 staphylococci in Australia. Med. J. Aust. 2, 789–795 (1958).

Ruppe, E. et al. Diversity of staphylococcal cassette chromosome mec structures in methicillin-resistant Staphylococcus epidermidis and Staphylococcus haemolyticus strains among outpatients from four countries. Antimicrob. Agents Chemother. 53, 442–449 (2009).

Seybold, U. et al. Emergence of community-associated methicillin-resistant Staphylococcus aureus USA300 genotype as a major cause of health care-associated blood stream infections. Clin. Infect. Dis. 42, 647–656 (2006).

Shaw, K. J. et al. In vitro activity of TR-700, the antibacterial moiety of the prodrug TR-701, against linezolid-resistant strains. Antimicrob. Agents Chemother. 52, 4442–4447 (2008).

Shinefield, H. et al. Use of a Staphylococcus aureus conjugate vaccine in patients receiving hemodialysis. N. Engl. J. Med. 346, 491–496 (2002).

Shopsin, B. et al. Evaluation of protein A gene polymorphic region DNA sequencing for typing of Staphylococcus aureus strains. J. Clin. Microbiol. 37, 3556–3563 (1999).

Shorr, A. F., Kunkel, M. J. & Kollef, M. Linezolid versus vancomycin for Staphylococcus aureus bacteraemia: pooled analysis of randomized studies. J. Antimicrobiol. Chemother. 56, 923–929 (2005).

Stam-Bolink, E. M., Mithoe, D., Baas, W. H., Arends, J. P. & Moller, A. V. Spread of a methicillin-resistant Staphylococcus aureus ST80 strain in the community of the northern Netherlands. Eur. J. Clin. Microbiol. Infect. Dis. 26, 723–727 (2007).

Stapleton, P. D., Shah, S., Ehlert, K., Hara, Y. & Taylor, P. W. The β-lactam-resistance modifier (-)-epicatechin gallate alters the architecture of the cell wall of Staphylococcus aureus. Microbiology 153, 2093–2103 (2007).

Steinkraus, G., White, R. & Friedrich, L. Vancomycin MIC creep in non-vancomycin-intermediate Staphylococcus aureus (VISA), vancomycin-susceptible clinical methicillin-resistant S. aureus (MRSA) blood isolates from 2001–2005. J. Antimicrob. Chemother. 60, 788–794 (2007).

Stemper, M. E., Shukla, S. K. & Reed, K. D. Emergence and spread of community-associated methicillin resistant Staphylococcus aureus in rural Wisconsin, 1989 to 1999. J. Clin. Microbiol. 42, 5673–5680 (2004).

Tenover, F. C. et al. Characterization of Staphylococcus aureus isolates from nasal cultures collected from individuals in the United States in 2001 to 2004. J. Clin. Microbiol. 46, 2837–2841 (2008).

Tiemersma, E. Emergence and resurgence of meticillin-resistant Staphylococcus aureus as a public-health threat. Lancet 368, 874–885 (2006).

Tristan, A. et al. Global distribution of Panton-Valentine leukocidin positive methicillin-resistant Staphylococcus aureus, 2006. Emerg. Infect. Dis. 13, 594–600 (2007).

Turner, K. M., Hanage, W. P., Fraser, C., Connor, T. R. & Spratt, B. G. Assessing the reliability of eBURST using simulated populations with known ancestry. BMC Microbiol. 7, 30 (2007).

Udo, E. E., Pearman, J. W. & Grubb, W. B. Genetic analysis of community isolates of methicillin-resistant Staphylococcus aureus in Western Australia. J. Hosp. Infect. 25, 97–108 (1993).

Vandenesch, F. et al. Community-acquired methicillin-resistant Staphylococcus aureus carrying Panton-Valentine leukocidin genes: worldwide emergence. Emerg. Infect. Dis. 9, 978–984 (2003).

van Wamel, W. J., Rooijakkers, S. H., Ruyken, M., van Kessel, K. P. & van Strijp, J. A. The innate immune modulators staphylococcal complement inhibitor and chemotaxis inhibitory protein of Staphylococcus aureus are located on β-hemolysin-converting bacteriophages. J. Bacteriol. 188, 1310–1315 (2006).

Voyich, J. M. et al. Is Panton-Valentine leukocidin the major virulence determinant in community-associated methicillin-resistant Staphylococcus aureus disease? J. Infect. Dis. 194, 1761–1770 (2006).

Voyich, J. M. et al. Insights into mechanisms used by Staphylococcus aureus to avoid destruction by human neutrophils. J. Immunol. 175, 3907–3919 (2005).

Wang, G., Hindler, J. F., Ward, K. W. & Bruckner, D. A. Increased vancomycin MICs for Staphylococcus aureus clinical isolates from a university hospital during a 5-year period. J. Clin. Microbiol. 44, 3883–3886 (2006).

Wang, R. et al. Identification of novel cytolytic peptides as key virulence determinants for community-associated MRSA. Nature Med. 13, 1510–1514 (2007).

Wannet, W. J. et al. Emergence of virulent methicillin-resistant Staphylococcus aureus strains carrying Panton-Valentine leucocidin genes in The Netherlands. J. Clin. Microbiol. 43, 3341–3345 (2005).

Weigel, L. M. et al. Genetic analysis of a high-level vancomycin-resistant isolate of Staphylococcus aureus. Science 302, 1569–1571 (2003).

Weigelt, J. et al. Linezolid versus vancomycin in treatment of complicated skin and soft tissue infections. Antimicrob. Agents Chemother. 49, 2260–2266 (2005).

Woodin, A. M. Purification of the two components of leucocidin from Staphylococcus aureus. Biochem. J. 75, 158–165 (1960).

Woodin, A. M. & Wieneke, A. A. The participation of calcium, adenosine triphosphate and adenosine triphosphatase in the extrusion of the granule proteins from the polymorphonuclear leucocyte. Biochem. J. 90, 498–509 (1964).

Wright, J. Staphylococcal leucocidin (Neisser-Wechsberg type) and antileucocidin. Lancet 227, 1002–1005 (1936).

Wu, S., Piscitelli, C., de Lencastre, H. & Tomasz, A. Tracking the evolutionary origin of the methicillin resistance gene: cloning and sequencing of a homologue of mecA from a methicillin susceptible strain of Staphylococcus sciuri. Microb. Drug Resist. 2, 435–441 (1996).

Wunderink, R. G., Cammarata, S. K., Oliphant, T. H. & Kollef, M. H. Continuation of a randomized, doubleblind, multicenter study of linezolid versus vancomycin in the treatment of patients with nosocomial pneumonia. Clin. Ther. 25, 980–992 (2003).

Zhang, K., McClure, J. A., Elsayed, S. & Conly, J. M. Novel staphylococcal cassette chromosome mec type, tentatively designated type VIII, harboring class A mec and type ccr gene complexes in a Canadian epidemic strain of methicillin-resistant Staphylococcus aureus. Antimicrob. Agents Chemother. 53, 531–540 (2009).

F.R.D. is supported by the Intramural Research Program of the National Institute of Allergy and Infectious Disease (NIAID) and the National Institutes of Health (NIH) and H.F.C. is supported by the NIH, NIAID grant number R01 AI070289.

Michael A. Kohanski,^8,9,10,11,12 Mark A. DePristo,^9,10,11,13 and James J. Collins^{8,9,10,11,12,14,*}

Antibiotic resistance arises through mechanisms such as selection of naturally occurring resistant mutants and horizontal gene transfer. Recently, oxidative stress has been implicated as one of the mechanisms whereby bactericidal antibiotics kill bacteria. Here, we show that sublethal levels of bactericidal antibiotics induce mutagenesis, resulting in heterogeneous increases in the minimum inhibitory concentration for a range of antibiotics, irrespective of the drug target. This increase in mutagenesis correlates with an increase in ROS and is prevented by the ROS scavenger thiourea and by anaerobic conditions, indicating that sublethal concentrations of antibiotics induce mutagenesis by stimulating the production of ROS. We demonstrate that these effects can lead to mutant strains that are sensitive to the applied antibiotic but resistant to other antibiotics. This work establishes a radical-based molecular mechanism whereby sublethal levels of antibiotics can lead to multidrug resistance, which has important implications for the widespread use and misuse of antibiotics.

There are a number of mechanisms whereby bacteria can develop antibiotic resistance (Dwyer et al., 2009; Hegreness et al., 2008; Livermore, 2003; McKenzie and Rosenberg, 2001), including horizontal transfer of resistance genes (Davies, 1994), drug-specific selection of naturally occurring resistant

variants within a population, and increased mutagenesis in hypermutator strains (Andersson, 2003; Chopra et al., 2003). Quinolones, which are DNA-damaging antibiotics, can stimulate the emergence of drug resistance via SOS-independent recombination (López et al., 2007) and through the induction of RecA-mediated processes, including homologous recombination (Drlica and Zhao, 1997) and SOS-regulated error-prone polymerases (Cirz et al., 2005). β-lactams can also induce the SOS response via RecA (Kohanski et al., 2007) and the DpiAB two-component system (Miller et al., 2004), and these drugs have been shown to induce DinB in an SOS-independent fashion, resulting in increased frameshift mutations (Pérez-Capilla et al., 2005).

Antibiotic treatment can also result in multidrug resistance (Cohen et al., 1989), which has been associated with mutations in multidrug efflux pumps, such as AcrAB (Ma et al., 1993). These drug efflux pumps can be regulated by a number of transcription factors, including the superoxide-responsive SoxRS system (Greenberg et al., 1990). In addition, there is evidence that low level antibiotic treatment can lead to mutations that cause resistance (Girgis et al., 2009); however, the mechanisms underlying this effect are not well understood.

Bactericidal antibiotics, including β-lactams, quinolones, and aminoglycosides, can stimulate bacteria to produce reactive oxygen species (ROS) (Dwyer et al., 2007; Kohanski et al., 2007, 2008), which are highly deleterious molecules that can interfere with the normal functions of oxygen-respiring organisms (Brumaghim et al., 2003; Fridovich, 1978; Imlay, 2006). Certain ROS, such as hydroxyl radicals, can directly damage DNA and lead to an accumulation of mutations (Demple and Harrison, 1994; Friedberg et al., 2006; Imlay et al., 1988). Oxidative DNA damage also activates the error-prone SOS response (Carlsson and Carpenter, 1980; Imlay and Linn, 1986, 1987) and error-correcting repair systems such as the “GO” repair system (Michaels and Miller, 1992; Miller, 1996). In this study, we hypothesized that ROS formation due to treatment with low levels of bactericidal antibiotics leads to an increase in mutation rates, which can result in the emergence of multidrug resistance. We thus consider a possible molecular mechanism whereby bactericidal antibiotics act as active, reactive mutagens.

To test the above hypothesis, we examined mutation rates in E. coli strain MG1655 following treatment with low levels of the bactericidal antibiotics norfloxacin (quinolone), ampicillin (β-lactam), and kanamycin (aminoglycoside), respectively. Mutation rates were determined by plating aliquots of treated cultures onto rifampicin plates, counting rifampicin-resistant colonies, and using the MSS maximum likelihood method (Rosche and Foster, 2000) to estimate the number of mutation events per culture (see Experimental Procedures for additional details). The mutation rate for untreated wild-type E. coli was approximately 1.5 × 10⁻⁸ mutations/cell/generation.

Treatment with 1 μg/ml ampicillin, 3 μg/ml kanamycin, 15 ng/ml norfloxacin, or 50 ng/ml norfloxacin resulted in significant increases in the mutation rate relative to an untreated control (Figure A3-1). Treatment with 1 μg/ml kanamycin resulted in a modest increase in mutation rate (Figure A3-1). The largest increases in mutation rate were seen following treatment with ampicillin or 50 ng/ml norfloxacin (Figure A3-1A). These changes were on par with the increase in mutation rate observed following treatment with 1mM hydrogen peroxide (Figure A3-1A), a concentration of hydrogen peroxide known to induce hydroxyl radical formation via Fenton chemistry (Imlay et al., 1988). To determine if there is a correlation between these changes in mutation rate and ROS formation, we examined radical levels using the radical-sensitive dye 3′-(p-hydroxyphenyl) fluorescein (HPF) (Setsukinai et al., 2003) (see Experimental Procedures for more details). We found a significant correlation (R² = 0.8455) between the fold change in mutation rate and peak HPF signal for the treatments described above (Figure A3-1B).

The strong correlation between ROS formation and fold change in mutation rate following treatment with bactericidal antibiotics suggests that ROS actively contribute to bactericidal drug-induced mutagenesis. To test if this is indeed the case, we added 100 mM thiourea to wild-type E. coli treated with antibiotics or hydrogen peroxide at the concentrations noted above (Figure A3-1C). Thiourea is a potent hydroxyl radical scavenger that mitigates the effects of hydroxyl radical damage in both prokaryotes and eukaryotes (Novogrodsky et al., 1982; Repine et al., 1981; Touati et al., 1995). We have previously shown that thiourea reduces hydroxyl radical formation and cell killing following treatment with bactericidal antibiotics (Kohanski et al., 2007).

The addition of thiourea significantly reduced the mutation rate to near untreated levels following the addition of 1 mM hydrogen peroxide, norfloxacin, or ampicillin (Figure A3-1). Interestingly, we were unable to detect any rifampicin-resistant colonies after plating up to 10⁹ cells following treatment with both 3 μg/ml kanamycin and thiourea (Figure A3-1C). However, we were able to detect rifampicin-resistant colonies after scaling up the system to 1 L flasks and plating up to 10¹⁰ cells following treatment with both 3 μg/ml kanamycin and thiourea (data not shown). These results suggest a role for kanamycin-mediated interference with ribosome function and translation, in the absence of oxidative stress, on significantly lowering mutation rate.

To further demonstrate that antibiotic-mediated ROS formation has a mutagenic component, we examined mutation rates under anaerobic growth conditions (see Experimental Procedures for additional details) following treatment of wild-type E. coli with antibiotics or hydrogen peroxide as described above (Figure A3-1D). We observed mutation rates near untreated levels for all antibiotic treatments tested (Figure A3-1D). Treatment with 1 mM hydrogen peroxide, which results in direct addition of an oxidant, led to an increase in mutation rate relative to the no drug control under anaerobic growth conditions (Figure A3-1D), but

this increase was considerably smaller than that exhibited under aerobic growth conditions (Figure A3-1A).

Antibiotic-resistant strains can arise via drug-mediated selection of preexisting antibiotic-resistant variants that occur naturally within a population (Livermore, 2003). Antibiotic-induced oxidative stress may be an additional mechanism that allows for the accumulation of mutations that increase resistance to drugs, irrespective of the drug target of the applied antibiotic. To test this, we measured changes in the minimum inhibitory concentration (MIC) of wild-type E. coli over a period of 5 days of selective growth for the following antibiotics: norfloxacin, kanamycin, ampicillin, tetracycline, and chloramphenicol. During the growth period, the cultures were exposed to no drug, norfloxacin, ampicillin, or kanamycin (see Experimental Procedures for more details). In all cases, growth in the absence of antibiotics did not change the MIC for any of the drugs tested (data not shown).

Treatment with 25 ng/ml norfloxacin led to an increase in the MIC for norfloxacin and kanamycin (Figure A3-S1A). The observed increases in MIC following treatment with norfloxacin were concentration dependent (see Supplemental Information for more details). Treatment of wild-type E. coli with 3 μg/ml kanamycin led to an increase in the MIC for kanamycin and minimal increases in the MIC for norfloxacin and ampicillin, respectively (Figure A3-S1C). The MIC for tetracycline and chloramphenicol did not change (Figure A3-S1C), indicating that kanamycin treatment may not lead to mutants resistant to other classes of ribosome inhibitors.

Treatment of wild-type E. coli with 1 μg/ml ampicillin for 5 days led to an increase in the MIC to different levels for ampicillin, norfloxacin, kanamycin, tetracycline, and chloramphenicol (Figure A3-2A). These results show that treatment with a β-lactam can stimulate formation of mutants that are potentially resistant to a wide range of antibiotics. Cultures that had been grown for 5 days in the presence of low levels of ampicillin were shifted to a drug-free environment and grown without any ampicillin for 2 additional days. The MICs, which were increased after 5 days of ampicillin treatment (Figure A3-2A), remained elevated and did not change significantly following 2 days of growth in the absence of ampicillin (Figure A3-S1D). These findings demonstrate that the observed increases in MIC are stable and not due to a transient adaptation to growth in the presence of ampicillin.

To determine if the observed increases in MIC were related to antibiotic-mediated ROS formation, we measured the MIC for ampicillin, norfloxacin, kanamycin, tetracycline, and chloramphenicol, respectively, following treatment with no drug or 1 μg/ml ampicillin under anaerobic growth conditions. Untreated anaerobic growth had no effect on MIC relative to untreated aerobic growth (data not shown). Following treatment with 1 μg/ml ampicillin under anaerobic conditions, we observed almost no increase in MIC for ampicillin, kanamycin, tetracycline, or chloramphenicol (Figure A3-2B). The MIC for norfloxacin exhib-

ited a small increase by day 5 (Figure A3-2B); however, this change in MIC was much smaller than the increase in MIC for norfloxacin following ampicillin treatment under aerobic growth conditions (Figure A3-2A). These results suggest that ROS formation due to treatment with low levels of bactericidal antibiotics can lead to mutagenesis and the emergence of bacteria resistant to a wide range of antibiotics.

Drug resistance may not always be uniform throughout a population. Some cells within a population may remain susceptible to the antibiotic, whereas other cells display varying degrees of drug resistance (de Lencastre et al., 1993), a phenomenon referred to as heteroresistance. Antibiotic-stimulated, ROS mediated mutagenesis could be a mechanism that stimulates the formation of a range of mutations that result in varying MICs within a population of cells. We sought to determine if the observed increases in population-level MIC for ampicillin following 5 days of treatment with 1 μg/ml ampicillin (Figure A3-2A) exhibited heterogeneity in MIC at the single-colony level.

We isolated individual colonies following ampicillin treatment and measured the MIC of each clone to ampicillin. We found that these isolates exhibited a range of resistance to ampicillin (>2.5–12.5 μg/ml), with some isolates remaining completely susceptible (≤2.5 μg/ml) to treatment with this drug (Figure A3-3A). We also found that the MICs for these isolates to norfloxacin ranged from <100 ng/ml (completely susceptible) to ≥1000 ng/ml (Figure A3-3B). Although levels of resistance from clinical isolates are typically quite high (with MICs in the range of 10–1000 μg/ml for norfloxacin [Becnel Boyd et al., 2009]), the upper ranges of the MICs for ampicillin or norfloxacin observed here (Figure A3-3) are near the peak serum concentrations for these drugs (Bryskier, 2005), indicating that these MICs might be near the limit for the amount of drug a human can tolerate. These data show that heterogeneous increases in MIC to ampicillin arise in E. coli following treatment with low levels of ampicillin, and treatment with one drug class can lead to heterogeneous increases in MIC against other classes of antibiotics.

Resistance to multiple antibiotics has been linked to mutations in drug-efflux systems such as the AcrAB multidrug (MDR) efflux pump (George and Levy, 1983; Ma et al., 1993) as well as mutations in transcription factors controlling these systems, such as MarA (Alekshun and Levy, 1997), Rob (Ariza et al., 1995), and SoxS (Greenberg et al., 1990). Our results suggest that ROS-mediated DNA damage induced by low levels of bactericidal antibiotics can result in mutations in a wider range of genes, potentially in some unrelated to the applied antibiotic and drug efflux systems. This implies that treatment with ampicillin, for example, may generate mutants that are not ampicillin resistant but are resistant to other antibiotics.

To determine if these types of resistant strains arise, we examined multidrug resistance following 5 days of treatment with 1 μg/ml ampicillin or no treatment. Mutants from ampicillin treated or untreated cultures were selected on plates

containing norfloxacin, ampicillin, kanamycin, tetracycline, and chloramphenicol, respectively. From this primary selection, we determined cross-resistance to the other four antibiotics via replica plating (see Experimental Procedures for additional details). We found substantially more primary resistant colonies and higher rates of cross-resistance following ampicillin treatment as compared to no treatment (Table A3-1). Ampicillin-selected mutants displayed a range of cross-resistance to the other classes of antibiotics and showed a strong correlation (89% cross-resistance) with norfloxacin resistance (Table A3-1). We also found that ampicillin-treated cells selected originally on the basis of norfloxacin or kanamycin resistance were only 75% and 63% cross-resistant to ampicillin, respectively (Table A3-1). Interestingly, primary resistance selection with the static drugs tetracycline and chloramphenicol yielded isolates that were always

(100%) cross-resistant to ampicillin (Table A3-1); this effect deserves further study. Also of note, ampicillin-treated, kanamycin-resistant strains were found to have very low cross-resistance to tetracycline (7%) and no cross-resistance with chloramphenicol (Table A3-1). This is consistent with previous work showing a lack of cross-resistance to tetracycline or chloramphenicol following selective treatment with aminoglycosides (Grassi, 1979). While the majority of these multidrug cross-resistant strains exhibit resistance against the treatment drug, ampicillin, our results demonstrate that treatment with ampicillin can also generate mutants that are not resistant to ampicillin yet are resistant to other classes of antibiotics.

We sought to determine if some of the ampicillin-treated, cross-resistant isolates had acquired mutations in specific antibiotic targets or in genes making up the common oxidative damage cell death pathway induced by bactericidal antibiotics (Kohanski et al., 2007, 2008), or if the observed cross-resistance (Table A3-1) was solely a function of altered drug efflux. We examined six norfloxacin-resistant isolates, six kanamycin-resistant isolates, and the untreated control strain. We sequenced the following genes where mutations could potentially lead to an increase in drug resistance: gyrA, gyrB, rpsL, ampC, icdA, arcA, cpxA, sdhB, iscR, tolC, marRA and its promoter region, and acrA and its promoter region. gyrA and gyrB code for the subunits of DNA Gyrase; the known target of quinolones, rpsL, encodes a component of the 30S subunit of the ribosome and has been associated with aminoglycoside resistance; ampC has been associated with ampicillin resistance; icdA, arcA, cpxA, sdhB, and iscR are genes involved in the common mechanism of cell death; and tolC, marRA and its promoter region, and acrA and its promoter region are involved in multidrug efflux.

We found that 3 of the 6 norfloxacin-resistant isolates contained point mutations in gyrA that resulted in a substitution of glycine for aspartic acid at amino acid 82 in one isolate and a substitution of tyrosine for aspartic acid at amino acid 87 in two other isolates (Figure A3-4A). We also found that 1 of these 6 norfloxacin-resistant isolates, which did not have a mutation in gyrA, had a point mutation resulting in the conversion of serine to phenylalanine at residue 464 of GyrB (Figure A3-4B). Interestingly, the point mutations we found in gyrA and gyrB are all in the quinolone resistance-determining regions of GyrA and GyrB, respectively, and these mutations have been observed in clinical isolates of Bacteroides fragilis (Oh et al., 2001), Salmonella enterica (Weill et al., 2006), and Pseudomonas aeruginosa (Mouneimné et al., 1999).

As noted above, mutations in rpsL have been associated with aminoglycoside resistance. We found that 2 of the 6 kanamycin-resistant isolates had point mutations in rpsL. These mutations led to a frameshift and truncated form of RpsL in both isolates (Figure A3-4D). It is possible that these mutations contribute to kanamycin resistance in these isolates.

Among the ampicillin-treated, drug-resistant mutants, we did not find any mutations in ampC (data not shown), a gene associated with ampicillin resistance.

We also did not find any mutations in icdA, cpxA, sdhB, or iscR (data not shown). Interestingly, we did find a single insertion mutation in arcA in one of the drug-resistant isolates. ArcA is a two-component system transcription factor containing a sensor domain and a DNA-binding domain, and the mutation we found results in a truncated ArcA protein that is missing the DNA-binding element of the protein (Figure A3-4E). We have previously shown that two-component systems are important elements in the common mechanism of cell death, and a knockout of arcA is more tolerant to treatment with ampicillin and kanamycin compared to norfloxacin (Kohanski et al., 2008). This isolate is resistant to ampicillin and kanamycin, but not to norfloxacin. This result suggests that mutations leading to low-level antibiotic resistance can occur in genes that are involved in the common mechanism of cell death.

We did not find any mutations in tolC, marRA, the marRA promoter, or acrA (data not shown); however, we did find a T-to-A conversion in the promoter upstream of acrA (Figure A3-4C) in one of the norfloxacin-resistant isolates that also had a mutation in gyrA (Figure A3-4A). This promoter mutation occurs within the annotated -35 site of the promoter and the binding site for the repressor transcription factors AcrR and EnvR (Keseler et al., 2005; Miller et al., 2002). The observed mutations could reduce the ability of these repressors to bind to the acrAB promoter which would result in increased pump expression and drug resistance. These sequencing results demonstrate that ampicillin treatment can lead to the formation of norfloxacin-resistant strains with mutations in DNA Gyrase and/or mutations that can affect drug efflux pump activity, which likely contribute to the emergence of multidrug resistance.

To demonstrate that sublethal levels of bactericidal antibiotics can lead to an increase in multidrug cross-resistance in Gram-positive as well as Gram-negative bacteria, we also examined multidrug cross-resistance in Staphylococcus aureus following treatment with low levels of ampicillin (35 ng/ml) for 5 days. Previously, we demonstrated antibiotic-mediated ROS formation in S. aureus (Kohanski et al., 2007). In the present study, we found substantially more primary resistant S. aureus colonies and higher rates of cross-resistance following ampicillin treatment as compared to no treatment (Table A3-2). Interestingly, we were unable to enrich for tetracycline- or chloramphenicol-resistant S. aureus isolates following treatment with low-level ampicillin as compared with the no-drug treatment. This may be due to the lower level of ROS formation we have observed with S. aureus (Kohanski et al., 2007).

To demonstrate that these effects are not limited to lab strains, we considered a clinical isolate of E. coli from a patient with diarrhea (NCDC C771). We examined multidrug cross-resistance in the clinical isolate following treatment with 1 μg/ml ampicillin (see Experimental Procedures for more details). As with the wild-type strains, we found substantially more primary resistant colonies and higher rates of cross-resistance in the clinical isolates following ampicillin treatment as compared to no treatment. We also found that ampicillin-treated cells

selected originally on the basis of norfloxacin or kanamycin resistance were only 11.5% and 21.5% cross-resistant to ampicillin, respectively (Table A3-3). This further affirms that treatment with ampicillin can generate mutants that are not resistant to ampicillin yet are resistant to other classes of antibiotics.

Here, we establish a radical-based molecular mechanism whereby sublethal levels of antibiotics can lead to multidrug resistance. This occurs via bactericidal antibiotic-mediated radical formation that results in the formation of mutations, some of which confer antibiotic resistance. Low-level resistance likely provides a first step toward clinically significant resistance (Goldstein, 2007), and the mechanism we propose and validate here establishes an antibiotic-stimulated mutagenic effect that likely works in conjunction with SOS-induced mutagenesis in the emergence of mutations that confer drug resistance.

Clinical situations where bacteria are exposed to low levels of antibiotics can occur with incomplete treatment of an infection, noncompliance with antibiotic treatment (e.g., a missed pill), and reduced or limited drug accessibility to certain tissues (e.g., bone or cerebrospinal fluid [Bryskier, 2005]). It is possible that mutations arising via antibiotic-mediated oxidative stress could be maintained in the normal bacterial flora of the body and transferred to virulent bacteria via horizontal gene transfer, a mechanism that can be induced by DNA damage (Beaber et al., 2004). Novel therapeutics targeting ROS-forming systems or error-prone DNA damage repair systems may help reduce and contain the spread of new antibiotic-resistant bacteria.

All experiments were performed with wild-type E. coli strain MG1655 (ATCC 700926) in Luria-Bertani (LB) medium (Fisher Scientific; Waltham, MA). For all treatment conditions, we used 1 mM hydrogen peroxide (VWR;

West Chester, PA) and the following bactericidal antibiotics: norfloxacin (Sigma; St. Louis), ampicillin, and kanamycin (Fisher Scientific). Bactericidal antibiotics were used at concentrations of 15 ng/ml norfloxacin, 50 ng/ml norfloxacin, 1 μg/ml ampicillin, 1 μg/ml kanamycin, or 3 μg/ml kanamycin. Tetracycline (MP Biomedical; Solon, OH) and chloramphenicol (Fluka; St. Louis) were used for MIC assays, rifampicin (Sigma) for determination of antibiotic resistant rates, and thiourea (Fluka) for radical-quenching experiments. Anaerobic media was made by heating LB in 17 ml Bellco glass hungate tubes (FisherScientific) under anaerobic conditions in a Coy anaerobic chamber (Coy Laboratory Products Inc.; Grass Lake, MI) to drive out dissolved oxygen (Norris and Ribbons, 1969). Resazurin (10 mM) (Sigma), which turns clear in the absence of oxygen, was used as an indicator for anaerobic conditions. Multidrug resistance was also determined in wild-type S. aureus (ATCC 25923) and the E. coli clinical isolate NCDC C771 (ATCC 23985).

Mutation rates were examined following 24 hr of growth in the presence of a bactericidal antibiotic. Drug levels were chosen such that there was an observable effect on growth or survival within the first 6 hr after drug addition (Figure A3-S2), followed by “recovery” of the culture to near untreated colony density 24 hr after treatment. All treatment conditions exhibited recovery to near untreated colony density levels, with the exception of 50 ng/ml norfloxacin. This allowed us to compare mutation frequencies for cultures of similar densities following treatment with an antibiotic.

Mutation rates were determined using a rifampicin-based selection method (Giraud et al., 2001). Briefly, an overnight culture of E. coli was diluted 1:10,000 into 50 ml LB in a 250 ml baffled flask and grown for 3.5 hr at 37°C and 300 rpm. Cultures were grown at high shaking speeds and in baffled flasks to maximize aeration and ROS formation. The culture was diluted 1:3 into fresh LB containing no drug, an antibiotic, or hydrogen peroxide at the concentrations described above. For experiments with thiourea, thiourea in solid form was added to each diluted culture for a final concentration of 100 mM. Aliquots (1 ml; ten replicates) of these diluted cultures were grown in 14 ml tubes for 24 hr at 37°C and 300 rpm. Aliquots of each treatment were serially diluted and plated on LB-agar plates for colony forming unit per milliliter (cfu/ml) determination. Aliquots of each treatment were also plated on LB-agar plates containing 100 μg/ml rifampicin and grown for 48 hr at 37°C. Colonies were counted at 24 and 48 hr, and the colony count from the 48 hr time point was used to estimate mutation rates. For experiments in anaerobic conditions, cells were diluted 1:10,000 into 15 ml anaerobic LB in sealed hungate tubes to minimize exposure to oxygen. Antibiotic treatments, growth temperature, shaking speed, and sample collection were as described above for the aerobic growth conditions.

The colony counts from the ten replicates were then used in the MSS maximum-likelihood method (Rosche and Foster, 2000; Sarkar et al., 1992) to estimate the number of mutational events per culture. The MSS maximum likelihood method is a recursive algorithm based on the Lea-Coulson function for solving the Luria-Delbruck distribution for a given number of mutational events (Sarkar et al., 1992); its utility has been demonstrated in vitro (Rosche and Foster, 2000). The mutation rate was determined by dividing the number of mutational events per culture by the total number of bacteria plated on the rifampicin plates (Rosche and Foster, 2000). Fold change in mutation rate was determined for all treatments and conditions relative to an untreated MG1655 control. Three biological replicates were run for each treatment condition, and the averages are shown in Figure A3-1.

To detect ROS formation, we used the fluorescent reporter dye HPF (Invitrogen; Carlsbad, CA) and flow cytometry as previously described (Kohanski et al., 2007). Average fluorescence was determined at 0 (baseline), 1, 3, and 6 hr (normalized to a no-dye control) following antibiotic treatment at the concentrations described above, and peak fluorescence levels were used to determine the change in mean fluorescence relative to baseline (Figure A3-1B).

For wild-type E. coli, MICs for norfloxacin, ampicillin, kanamycin, tetracycline, and chloramphenicol were measured over 5 days of treatment with no drug, 25 ng/ml norfloxacin, 50 ng/ml norfloxacin, 1 μg/ml ampicillin, or 3 μg/ml kanamycin. Briefly, an overnight culture of E. coli was diluted 1:10,000 into 50 ml LB in a 250 ml baffled flask and grown for 3.5 hr at 37°C and 300 rpm. The culture was diluted 1:3 into fresh LB containing no drug or antibiotics at the above concentrations. Aliquots (1 ml) of these diluted cultures were grown in 14 ml tubes for 24 hr at 37°C and 300 rpm. Each day thereafter for 5 days, in order to avoid mutations arising due to evolution during stationary phase (GASP mutants) (Zinser and Kolter, 2004), cells were diluted 1:1000 into 1 ml LB in a 14 ml tube containing the respective antibiotic and grown for 24 hr at 37°C and 300 rpm.

MICs were also measured for anaerobically grown E. coli over 5 days of treatment with no drug or 1 μg/ml ampicillin. Briefly, an overnight culture of E. coli was diluted 1:1000 into 15 ml anaerobic LB in sealed hungate tubes containing no drug or 1 μg/ml ampicillin. These cultures were grown in the sealed hungate tubes for 24 hr at 37°C and 300 rpm. Each day thereafter for 5 days, cells were diluted 1:1000 into 15 ml anaerobic LB in a sealed hungate tube containing the respective antibiotic and grown for 24 hr at 37°C and 300 rpm.

To determine the MIC on each day, an aliquot of cells from each treatment condition was diluted 1:10,000 into LB and dispensed into 96-well plates (100 μl

total volume per well) containing various concentrations (ten replicates per drug concentration) of norfloxacin, ampicillin, kanamycin, tetracycline, or chloramphenicol. Plates were incubated at 37°C and 300 rpm for 24 hr, after which time the optical density at 600 nm (OD₆₀₀) was measured using a SPECTRAFluor Plus (Tecan; Männedorf, Switzerland). The median OD₆₀₀ was calculated for each drug concentration, and the MIC was determined as the concentration that inhibited 90% of growth based on OD₆₀₀. Fold change in MIC was determined by dividing the treated MIC on each day by its respective MIC from day 0.

Wild-type E. coli were grown for 5 days in the presence of 1 μg/ml ampicillin or no drug (untreated) as described above. These long-term-treated cultures were diluted 1:1000 into 25 ml LB in 250 ml flasks and grown for 3 hr at 37°C and 300 rpm. Aliquots (1 ml) were plated onto LB-agar plates containing 300 ng/ml norfloxacin, 7.5 μg/ml ampicillin, 15 μg/ml kanamycin, 8 μg/ml tetracycline, and 25 μg/ml chloramphenicol, respectively, and grown for 24 hr at 37°C. Approximately 100 ampicillin-treated colonies from each primary drug selection were purified by streaking them onto LB-agar plates containing the same selective antibiotic. Double the volume of untreated control cells were plated for primary resistance selection as compared to the ampicillin-treated cells, and these untreated colonies were also purified as described above. Plates were placed at 37°C for 24 hr; these strains were then transferred via replica plating onto LB-agar plates containing norfloxacin, ampicillin, kanamycin, tetracycline, or chloramphenicol. Cross-resistance for each primary antibiotic selection following the 5 day ampicillin treatment or the no-drug treatment was determined after 24 hr of growth at 37°C by counting the colonies that displayed growth on the various drug-containing replicated plates.

The MIC of 44 of the above isolates and the MG1655 control strain were determined for ampicillin and norfloxacin, respectively. Overnight cultures of each strain were diluted 1:10,000 into 100 μl LB plus varying concentration of antibiotic (four replicates per strain per drug concentration) in 96-well plates. Plates were incubated at 37°C and 300 rpm for 24 hr, after which time the OD₆₀₀ was measured using a SPECTRAFluor Plus (Tecan). The median OD₆₀₀ was calculated for each drug concentration, and the MIC was determined as the concentration that inhibited 90% of growth based on OD₆₀₀.

Wild-type S. aureus were grown for 5 days in the presence of 35 ng/ml ampicillin or no drug (untreated) as described above. E. coli clinical isolate NCDC C771 was grown for 5 days in the presence of 1 mg/ml ampicillin or no drug (untreated) as described above. These long-term-treated cultures were diluted 1:1000 into 25 ml LB in 250 ml flasks and grown for 3 hr at 37°C and 300 rpm. For S. aureus, 1 ml aliquots were plated onto LB-agar plates containing 3 μg/ml

norfloxacin, 7.5 μg/ml ampicillin, 15 μg/ml kanamycin, 8 μg/ml tetracycline, and 25 μg/ml chloramphenicol, respectively, and grown for 24 hr at 37°C. For NCDC C771, 1 ml aliquots were plated onto LB-agar plates containing 400 ng/ml norfloxacin, 8.5 μg/ml ampicillin, 20 μg/ml kanamycin, and 15 μg/ml chloramphenicol. Tetracycline cross-resistance was not quantified for NCDC C771, as this strain is resistant to tetracycline (MIC >35 μg/ml). Approximately 100 ampicillin-treated colonies from each primary drug selection were purified by streaking them onto LB-agar plates containing the same selective antibiotic. An equal volume of untreated S. aureus or the E. coli clinical isolate cells were plated for primary resistance selection as compared to the ampicillin-treated cells, and these untreated colonies were also purified as described above. The remainder of the cell growth and cross-resistance determination was performed as described above for wild-type E. coli.

Six ampicillin-treated, norfloxacin-resistant isolates and six ampicillin-treated, kanamycin-resistant isolates from the cross-resistance experiment described above, as well as the untreated MG1655 control strain, were grown to a cell density of approximately 10⁹ cfu/ml. Genomic DNA was extracted from each sample using a QIAGEN genomic DNA extraction kit according to the manufacturer’s instructions. Primers from IDT (Coralville, IA) (Table A3-S1) were utilized to PCR amplify, using Phusion DNA Polymerase (Finnzyme; Espoo, Finland), the regions surrounding gyrA, gyrB, tolC, acrA, marRA, ampC, rpsL, icdA, iscR, sdhB, arcA, and cpxR. These samples were sequenced by Agencourt Bioscience Corporation (Beverly, MA) using primers from IDT (Table A3-S1). Sequences were analyzed using Clone Manager 7 (Scientific & Educational Software; Cary, NC) and Sequence Scanner v1.0 (Applied Biosystems; Foster City, CA).

Supplemental Information includes Supplemental Results, Supplemental References, two figures, and one table and can be found with this article online at doi:10.1016/j.molcel.2010.01.003.

We thank D. Dwyer, B. Davies, and K. Allison for helpful discussions during the course of this work. We thank Q. Beg for help running the anaerobic chamber. This work was supported by the National Institutes of Health through

the NIH Director’s Pioneer Award Program, grant number DP1 OD003644, and the Howard Hughes Medical Institute.

Received: June 15, 2009

Revised: October 18, 2009

Accepted: November 12, 2009

Published: February 11, 2010

Alekshun, M.N., and Levy, S.B. (1997). Regulation of chromosomally mediated multiple antibiotic resistance: the mar regulon. Antimicrob. Agents Chemother. 41, 2067–2075.

Andersson, D.I. (2003). Persistence of antibiotic resistant bacteria. Curr. Opin. Microbiol. 6, 452–456.

Ariza, R.R., Li, Z., Ringstad, N., and Demple, B. (1995). Activation of multiple antibiotic resistance and binding of stress-inducible promoters by Escherichia coli Rob protein. J. Bacteriol. 177, 1655–1661.

Beaber, J.W., Hochhut, B., and Waldor, M.K. (2004). SOS response promotes horizontal dissemination of antibiotic resistance genes. Nature 427, 72–74.

Becnel Boyd, L., Maynard, M.J., Morgan-Linnell, S.K., Horton, L.B., Sucgang, R., Hamill, R.J., Jimenez, J.R., Versalovic, J., Steffen, D., and Zechiedrich, L. (2009). Relationships among ciprofloxacin, gatifloxacin, levofloxacin, and norfloxacin MICs for fluoroquinolone-resistant Escherichia coli clinical isolates. Antimicrob. Agents Chemother. 53, 229–234.

Brumaghim, J.L., Li, Y., Henle, E., and Linn, S. (2003). Effects of hydrogen peroxide upon nicotinamide nucleotide metabolism in Escherichia coli: changes in enzyme levels and nicotinamide nucleotide pools and studies of the oxidation of NAD(P)H by Fe(III). J. Biol. Chem. 278, 42495–42504.

Bryskier, A. (2005). Antimicrobial Agents: Antibacterials and Antifungals (Washington, D.C.: ASM Press).

Carlsson, J., and Carpenter, V.S. (1980). The recA+ gene product is more important than catalase and superoxide dismutase in protecting Escherichia coli against hydrogen peroxide toxicity. J. Bacteriol. 142, 319–321.

Chopra, I., O’Neill, A.J., and Miller, K. (2003). The role of mutators in the emergence of antibiotic-resistant bacteria. Drug Resist. Updat. 6, 137–145.

Cirz, R.T., Chin, J.K., Andes, D.R., de Crécy-Lagard, V., Craig, W.A., and Romesberg, F.E. (2005). Inhibition of mutation and combating the evolution of antibiotic resistance. PLoS Biol. 3, e176.

Cohen, S.P., McMurry, L.M., Hooper, D.C., Wolfson, J.S., and Levy, S.B. (1989). Cross-resistance to fluoroquinolones in multiple-antibiotic-resistant (MAR) Escherichia coli selected by tetracycline or chloramphenicol: decreased drug accumulation associated with membrane changes in addition to OmpF reduction. Antimicrob. Agents Chemother. 33, 1318–1325.

Davies, J. (1994). Inactivation of antibiotics and the dissemination of resistance genes. Science 264, 375–382.

de Lencastre, H., Figueiredo, A.M., and Tomasz, A. (1993). Genetic control of population structure in heterogeneous strains of methicillin resistant Staphylococcus aureus. Eur. J. Clin. Microbiol. Infect. Dis. 12 (Suppl 1), S13–S18.

Demple, B., and Harrison, L. (1994). Repair of oxidative damage to DNA: enzymology and biology. Annu. Rev. Biochem. 63, 915–948.

Drlica, K., and Zhao, X. (1997). DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiol. Mol. Biol. Rev. 61, 377–392.

Dwyer, D.J., Kohanski, M.A., Hayete, B., and Collins, J.J. (2007). Gyrase inhibitors induce an oxidative damage cellular death pathway in Escherichia coli. Mol. Syst. Biol. 3, 91.

Dwyer, D.J., Kohanski, M.A., and Collins, J.J. (2009). Role of reactive oxygen species in antibiotic action and resistance. Curr. Opin. Microbiol. 12, 482–489.

Fridovich, I. (1978). The biology of oxygen radicals. Science 201, 875–880.

Friedberg, E.C., Walker, G.C., Siede, W., Wood, R.D., Schultz, R.A., and Ellenberger, T. (2006). DNA Repair and Mutagenesis, Second Edition (Washington, D.C.: ASM Press).

George, A.M., and Levy, S.B. (1983). Amplifiable resistance to tetracycline, chloramphenicol, and other antibiotics in Escherichia coli: involvement of a non-plasmid-determined efflux of tetracycline. J. Bacteriol. 155, 531–540.

Giraud, A., Matic, I., Tenaillon, O., Clara, A., Radman, M., Fons, M., and Taddei, F. (2001). Costs and benefits of high mutation rates: adaptive evolution of bacteria in the mouse gut. Science 291, 2606–2608.

Girgis, H.S., Hottes, A.K., and Tavazoie, S. (2009). Genetic architecture of intrinsic antibiotic susceptibility. PLoS ONE 4, e5629.

Goldstein, F. (2007). The potential clinical impact of low-level antibiotic resistance in Staphylococcus aureus. J. Antimicrob. Chemother. 59, 1–4.

Grassi, G.G. (1979). Drug-inactivating enzymes of bacteria grown in subminimal inhibitory concentrations of antibiotics. Rev. Infect. Dis. 1, 852–857.

Greenberg, J.T., Monach, P., Chou, J.H., Josephy, P.D., and Demple, B. (1990). Positive control of a global antioxidant defense regulon activated by superoxide-generating agents in Escherichia coli. Proc. Natl. Acad. Sci. USA 87, 6181–6185.

Hegreness, M., Shoresh, N., Damian, D., Hartl, D., and Kishony, R. (2008). Accelerated evolution of resistance in multidrug environments. Proc. Natl. Acad. Sci. USA 105, 13977–13981.

Imlay, J.A. (2006). Iron-sulphur clusters and the problem with oxygen. Mol. Microbiol. 59, 1073–1082.

Imlay, J.A., and Linn, S. (1986). Bimodal pattern of killing of DNA-repair-defective or anoxically grown Escherichia coli by hydrogen peroxide. J. Bacteriol. 166, 519–527.

Imlay, J.A., and Linn, S. (1987). Mutagenesis and stress responses induced in Escherichia coli by hydrogen peroxide. J. Bacteriol. 169, 2967–2976.

Imlay, J.A., Chin, S.M., and Linn, S. (1988). Toxic DNA damage by hydrogen peroxide through the Fenton reaction in vivo and in vitro. Science 240, 640–642.

Keseler, I.M., Collado-Vides, J., Gama-Castro, S., Ingraham, J., Paley, S., Paulsen, I.T., Peralta-Gil, M., and Karp, P.D. (2005). EcoCyc: a comprehensive database resource for Escherichia coli. Nucleic Acids Res. 33(Database issue), D334–D337.

Kohanski, M.A., Dwyer, D.J., Hayete, B., Lawrence, C.A., and Collins, J.J. (2007). A common mechanism of cellular death induced by bactericidal antibiotics. Cell 130, 797–810.

Kohanski, M.A., Dwyer, D.J., Wierzbowski, J., Cottarel, G., and Collins, J.J. (2008). Mistranslation of membrane proteins and two-component system activation trigger antibiotic-mediated cell death. Cell 135, 679–690.

Livermore, D.M. (2003). Bacterial resistance: origins, epidemiology, and impact. Clin. Infect. Dis. 36 (Suppl 1), S11–S23.

López, E., Elez, M., Matic, I., and Blázquez, J. (2007). Antibiotic-mediated recombination: ciprofloxacin stimulates SOS-independent recombination of divergent sequences in Escherichia coli. Mol. Microbiol. 64, 83–93.

Ma, D., Cook, D.N., Alberti, M., Pon, N.G., Nikaido, H., and Hearst, J.E. (1993). Molecular cloning and characterization of acrA and acrE genes of Escherichia coli. J. Bacteriol. 175, 6299–6313.

McKenzie, G.J., and Rosenberg, S.M. (2001). Adaptive mutations, mutator DNA polymerases and genetic change strategies of pathogens. Curr. Opin. Microbiol. 4, 586–594.

Michaels, M.L., and Miller, J.H. (1992). The GO system protects organisms from the mutagenic effect of the spontaneous lesion 8-hydroxyguanine (7,8-dihydro-8-oxoguanine). J. Bacteriol. 174, 6321–6325.

Miller, J.H. (1996). Spontaneous mutators in bacteria: insights into pathways of mutagenesis and repair. Annu. Rev. Microbiol. 50, 625–643.

Miller, K., O’Neill, A.J., and Chopra, I. (2002). Response of Escherichia coli hypermutators to selection pressure with antimicrobial agents from differentclasses. J. Antimicrob. Chemother. 49, 925–934.

Miller, C., Thomsen, L.E., Gaggero, C., Mosseri, R., Ingmer, H., and Cohen, S.N. (2004). SOS response induction by beta-lactams and bacterial defense against antibiotic lethality. Science 305, 1629–1631.

Mouneimné, H., Robert, J., Jarlier, V., and Cambau, E. (1999). Type II topoisomerase mutations in ciprofloxacin-resistant strains of Pseudomonas aeruginosa. Antimicrob. Agents Chemother. 43, 62–66.

Norris, J.R., and Ribbons, D.W. (1969). Methods in Microbiology (Orlando, FL: Academic Press Inc).

Novogrodsky, A., Ravid, A., Rubin, A.L., and Stenzel, K.H. (1982). Hydroxyl radical scavengers inhibit lymphocyte mitogenesis. Proc. Natl. Acad. Sci. USA 79, 1171–1174.

Oh, H., El Amin, N., Davies, T., Appelbaum, P.C., and Edlund, C. (2001). gyrA mutations associated with quinolone resistance in Bacteroides fragilis groupstrains. Antimicrob. Agents Chemother. 45, 1977–1981.

Pérez-Capilla, T., Baquero, M.R., Gómez-Gómez, J.M., Ionel, A., Martín, S., and Blázquez, J. (2005). SOS-independent induction of dinB transcription by beta-lactam-mediated inhibition of cell wall synthesis in Escherichia coli. J. Bacteriol. 187, 1515–1518.

Repine, J.E., Fox, R.B., and Berger, E.M. (1981). Hydrogen peroxide kills Staphylococcus aureus by reacting with staphylococcal iron to form hydroxylradical. J. Biol. Chem. 256, 7094–7096.

Rosche, W.A., and Foster, P.L. (2000). Determining mutation rates in bacterial populations. Methods 20, 4–17.

Sarkar, S., Ma, W.T., and Sandri, G.H. (1992). On fluctuation analysis: a new, simple and efficient method for computing the expected number of mutants. Genetica 85, 173–179.

Setsukinai, K., Urano, Y., Kakinuma, K., Majima, H.J., and Nagano, T. (2003). Development of novel fluorescence probes that can reliably detect reactive oxygen species and distinguish specific species. J. Biol. Chem. 278, 3170–3175.

Touati, D., Jacques, M., Tardat, B., Bouchard, L., and Despied, S. (1995). Lethal oxidative damage and mutagenesis are generated by iron in delta fur mutants of Escherichia coli: protective role of superoxide dismutase. J. Bacteriol. 177, 2305–2314.

Weill, F.X., Guesnier, F., Guibert, V., Timinouni, M., Demartin, M., Polomack, L., and Grimont, P.A. (2006). Multidrug resistance in Salmonella enterica serotype Typhimurium from humans in France (1993 to 2003). J. Clin. Microbiol. 44, 700–708.

Zinser, E.R., and Kolter, R. (2004). Escherichia coli evolution during stationary phase. Res. Microbiol. 155, 328–336.

Molecular Cell, Volume 37

Michael A. Kohanski, Mark A. DePristo and James J. Collins

Treatment with 25 ng/ml norfloxacin led to significant increases in the MIC for norfloxacin and kanamycin as well as modest, low-level increases in the MIC for ampicillin, tetracycline and chloramphenicol (Figure A3-S1A). This increase in MIC for norfloxacin was concentration dependent. Treatment with 50 ng/ml norfloxacin led to a 6-fold increase in the MIC for norfloxacin (Figure A3-S1B); however, we were unable to observe an increase in the MIC for ampicillin, kanamycin, tetracycline, or chloramphenicol following treatment

FIGURE A3-S1 Bactericidal antibiotics can lead to broad-spectrum increases in MIC. (A-C) Fold change in MIC relative to a no-drug control for ampicillin, norfloxacin, kanamycin, tetracycline and chloramphenicol, following 5 days of growth in the presence of (A) 25 ng/ml norfloxacin, (B) 50 ng/ml norfloxacin, or (C) 1 μg/ml kanamycin. (D) Ampicillin-mediated increases in MIC are stable. Fold change in MIC relative to a no-drug control for ampicillin, norfloxacin, kanamycin, tetracycline and chloramphenicol, following 5 days of growth in the presence of 1 μg/ml ampicillin and an additional 2 days of growth in the absence of drug.

with 50 ng/ml norfloxacin for 5 days (Figure A3-S1B). Interestingly, selection of drug-resistant mutants following quinolone treatment is concentration dependent, with higher concentrations of quinolone selecting only quinolone-resistant strains and lower levels of quinolone selecting broadly for drugresistant mutants with mutations in a wide array of targets in E. coli (Drlica, 2003) and Mycobacterium tuberculosis (Zhou et al., 2000). It is possible that treatment with 50 ng/ml norfloxacin selects for naturally occurring quinolone-resistant mutants before the drug-induced mutagenesis has a chance to create mutants resistant to other drugs.

Drlica, K. (2003). The mutant selection window and antimicrobial resistance. J Antimicrob Chemother 52, 11-17.

Zhou, J., Dong, Y., Zhao, X., Lee, S., Amin, A., Ramaswamy, S., Domagala, J., Musser, J.M., and Drlica, K. (2000). Selection of antibiotic-resistant bacterial mutants: allelic diversity among fluoroquinolone-resistant mutations. J Infect Dis 182, 517-525.

Patrice Courvalin

Institut Pasteur, Unité des Agents Antibactériens, Paris, France

The evolution of bacteria toward antibiotic resistance is unavoidable since it represents a particular aspect of the general evolution of bacteria. It results from two independent steps, emergence and dissemination; however, as we consider, the mechanism of the first one can largely influence the success of the second one. Resistance to antibiotics in bacteria is secondary to mutations in resident (housekeeping) structural or regulatory genes or to horizontal acquisition of foreign genetic information (Perichon and Courvalin, 2009). In this review, we consider the relationship between low levels of antibiotics and dissemination of resistance.

The emergence of resistance, an event that occurs by pure chance, can be a rare, even transient, event if it does not provide a selective advantage against a molecule present in the environment of the bacterium. Resistance potentially exists in nature, not only before the clinical use but also before the discovery, or even the design, of a new antibiotic. This is obvious for natural antibiotics, because the producing microorganisms must protect themselves against suicide by the products of their secondary metabolism, but it also holds true for semisynthetic (e.g., amikacin) or entirely synthetic (e.g., fluoroquinolone) antibiotics.

The bacterial genome is composed of the chromosome and of accessory genetic elements, self-transferable or mobilizable plasmids, integrative conjugative elements (ICEs), transposons, insertion sequences, and bacteriophages. The chromosome contains all the genetic information required for the life cycle of the bacterium, whereas, as their name indicates, accessory genetic elements carry genes that are dispensable, although under certain circumstances, they can provide major advantages for the survival of the host, such as antibiotic resistance. The chromosome is inherited vertically by the progeny of the cell and is not transferable horizontally, whereas accessory genetic elements can also be transmitted to other bacteria. As a result, resistance can thus be endogenous or exogenous. Endogenous resistance results from chromosomal mutations and is generally not infectious from bacteria to bacteria. In contrast, exogenous resistance is due to horizontal (lateral) transfer of DNA among bacteria, resulting in acquisition of mobile genetic elements.

The occurrence of chromosomal mutations is an efficient pathway to resistance. Mutations are considered rare because they occur at low frequency, generally between 10⁻⁷ and 10⁻¹⁰ and were considered errors that occurred during DNA replication. However, this limitation is easily overcome because, during infections in humans, bacterial populations are often very large. Mutations in chromosomal genes clearly represent the only mechanism of antibiotic resistance in genera such as Mycobacterium or strictly intracellular pathogens (such as Chlamydia, Rickettsia, Coxiella, Ehrlichia), which are not known to exchange DNA under natural conditions.

Dissemination of resistance has, in numerous instances, been shown to be closely associated with antibiotic use (Malhotra-Kumar et al., 2007), which stresses the importance of the prudent use of these molecules. In addition, resistance is, if at all, slowly reversible (Andersson, 2003). There are three levels of resistance dissemination, depending upon the vector: bacteria (clonal spread), replicons (plasmid epidemics), or genes (conjugative transposon [ICE] epidemics). These various levels of dissemination, which coexist in nature and thus account for the extraordinary rise in antibiotic resistance among bacteria, are not only infectious but also exponential because each is associated with DNA duplication. Clonal dissemination is associated with chromosome replication, plasmid conjugation with replicative transfer, and gene migration with replicative transposition.

It also turns out that conjugation has a very broad host range—plasmids and ICEs can transfer efficiently between phylogenetically remote bacterial genera—and that there are limited barriers to heterologous gene expression (Courvalin, 1994); that is, resistance genes can be expressed in very diverse hosts.

These observations led to the notion of a bacterial gene pool, in particular with respect to resistance, which means that genes are loosely bound to their hosts and can easily disseminate under natural conditions. This concept has many practical consequences, for example, in the case of the use of antibiotics as animal feed additives. Rather than discussing endlessly (and often in a biased fashion) whether the enterococci from animals and humans are similar (Phillips, 1999) (i.e., whether vancomycin-resistant enterococci from animals can stably colonize the human gut), the true question should rather be the following: Are the resistance genes (to glycopeptides, in this example) the same among bacteria of these two ecosystems? Along this line, studies published long ago that examined the biochemistry and genetics of aminoglycoside resistance, as well

as molecular study of the bacterial hosts (with the techniques available at that time), elucidated how the exclusive use of apramycin in animals could select gentamicin-resistant bacteria that were later found in humans (Chaslus-Dancla et al., 1986a, 1986b, 1991). This notion has since been largely documented, using more powerful techniques, for resistance to other drug classes (Stobberingh and van den Bogaard, 2000).

It was shown recently that bactericidal antibiotics kill bacteria by inducing the formation of highly deleterious hydroxyl radicals, reactive oxygen species, which can damage DNA (Kohanski et al., 2007). This oxidative stress leads to a significant increase in mutation rate either directly or indirectly by activation of the SOS DNA damage response pathway (Kohanski et al., 2010) as well as an increase in recombination (Figure A4-1). Thus, certain classes of antibiotics may behave as mutagens, in particular at low concentrations, and may select for resistance to other drug classes, whereas the mutant derivatives remain susceptible to the applied antibiotic (Kohanski et al., 2010).

The major human pathogen Streptococcus pneumoniae may represent a particular case of this mechanism. Evidence has been recently provided that the stress caused by low concentrations of certain antibiotics induces genetic transformability in pneumococci (Prudhomme et al., 2006). Transformation is a process inherent in this bacterial species that allows the transient uptake and integration of exogenous DNA in the recipient genome as well as the capability to kill noncompetent cells, a phenomenon referred to as fratricide. Low concentrations of bactericidal antibiotics, such as quinolones and aminoglycosides, induce full competence for genetic transformation, thereby increasing the rate of genetic exchange in S. pneumoniae and making chromosomal mutations horizontally transferable (Figure A4-2). Competence appears thus as a general stress response, playing a role similar to that of the SOS response in Escherichia coli that lacks in S. pneumoniae (Claverys and Havarstein, 2002). Fratricide is the killing of cells from the same species and can be considered a mechanism that is used by competent bacteria to acquire DNA from noncompetent pneumococci (Claverys et al., 2007). Considering the high incidence of asymptomatic carriage of and co-colonization by this human pathogen, inappropriate antibiotic use could accelerate the emergence of resistant clones, promote evolution toward virulence, and enrich in capsular types that are not included in the current vaccines. The latter observation represents an additional argument for not prescribing fluoroquinolones to children.

FIGURE A4-1 Antibiotic induced increase mutation rate in S. pneumoniae. Subinhibitory concentrations of bactericidal antibiotics promote production of reactive oxygen species (ROS) by bacteria via the stress response. This leads to DNA damage which i) increases recombination frequency and ii) induces a competence state resulting in transformation which both cause mutations.

Antibiotics can enhance gene transfer: they provide selective pressure for resistant bacteria to maintain and disseminate, but they can also induce the transfer of resistance genes. For example, it has been reported that (1) the use of subinhibitory concentrations of penicillins increased the conjugal transfer of plasmid DNA from Escherichia coli to Staphylococcus aureus and Listeria monocytogenes (Trieu-Cuot et al., 1993), (2) oxacillin increased the frequency of in vitro transfer of Tn916, an enterococcal ICE, from Enterococcus faecalis to Bacillus anthracis (Ivins et al., 1988), (3) the transfer frequency of conjugative transposons belonging to the Tn916/Tn1545 family (Figure A4-3), which contain a tetracycline resistance determinant, was increased 10- to 100-fold in vitro and in vivo in the presence of low concentrations of tetracycline (Doucet-Populaire et al., 1991), and (4) tetracycline also increased dramatically the transfer of a

FIGURE A4-2 Antibiotic promotes evolution of resistance in S. pneumoniae. The presence of an antibiotic generates a bacterial stress responsible for competence. The competence state induces transformation and fratricidy in which both can lead to antibiotic resistance and capsular switch.

Bacteroides conjugative transposon (Li et al., 1995). In the two latter cases, the antibiotic has a triple activity: as an antibacterial agent, as an inducer of resistance to itself, and as an inducer of the dissemination of resistance determinants. It thus appears that several antibiotics can behave like pheromones: they are synthesized by specific cells (such as the Actinobacteria), and they act on another cell, at low concentrations, on very specific targets to promote DNA exchange.

It was also shown that mitomycin C and ciprofloxaxin de-repressed the expression of genes necessary for transfer of an ICE in Vibrio cholerae (Beaber et al., 2004). This resulted in an unpredictable horizontal dissemination of the genetic element which confers resistance to chloramphenicol, trimethoprim, sulphonamides, and streptomycin.

Another example of increased resistance gene mobility by antibiotics is represented by the integrons (Mazel, 2006). These compact structures act as genetic systems for in vivo capture and expression of genes in the form of circular cassettes. These genes are the most tightly linked because they are not only adjacent but coexpressed from the same promoter. Integrons are thus typically responsible for coresistance: the stable association in the same cell of various resistance determinants, each conferring resistance to a drug class. Similarly to cross-resistance, which results in cross-selection, coresistance implies coselection: the use of any antibiotic that is substrate for a mechanism encoded by the integron will coselect for the other resistances. This genetic organization renders the consequences of

FIGURE A4-3 Transfer of an integrative conjugative element (ICE). ICEs are mobile genetic elements that carry one or several resistance genes. They excise by site-specific recombination between their flanking attachment sites, attR and attL, leading to the formation of an episomal ICE carrying an attI site and an empty attB site in the chromosome. They replicate during their transfer by conjugation and integrate in the chromosome of the recipient. Dissemination of resistance by ICEs is thus infectious and exponential.

the use of a single drug unpredictable. Because there is transcriptional attenuation along the operons in integrons (Collis and Hall, 1995), the use of an antibiotic selects not only for the neighboring resistance determinants but for a higher level of resistance to itself as well. This is achieved by the movement (excision, circularization, and re-integration) of the corresponding cassette that ends up downstream from the strong common promoter. Integrons, therefore, allow quantitative (self) and qualitative (nonself) alteration of resistance. Most interestingly, it has been shown recently that certain antibiotics such as mitomycin C, trimethroprim, the quinolones, and the β-lactams stimulate the intracellular mobility of the gene cassettes (Guerin et al., 2009).

Genes from Gram-positive cocci can be transferred by conjugation (of plasmids or ICE) not only among these microorganisms but also to Gram-negative bacteria (Courvalin, 1994). The reverse is not true because of limitations in heterologous gene expression. This is due to the fact that the −35 and −10 sequences

and their spacing that constitute the promoters for expression of the genes, as well as the ribosome binding site, are conserved and are close to the consensus in Gram-positives; they are thus also functional in Gram-negatives. In contrast, these motifs are much more degenerate in Gram-negatives and cannot be accommodated by Gram-positives. Similarly, the promoters from Bacteroides fragilis and from E. coli are dissimilar, resulting in lack of gene expression from E. coli promoters (even strong promoters) in B. fragilis and the inactivity of B. fragilis promoters in other bacterial species (E. coli, Bacillus subtilis, and Clostridium perfringens) (Bayley et al., 2000; Smith et al., 1992). One can thus confidently predict that strains of B. fragilis will not, or will extremely inefficiently, act as intermediates in resistance gene transfer or represent a pool of origin of these genes for human pathogens.

Acquisition of resistance by bacteria corresponds to a gain of function and is, thus, generally associated with a biological cost. In other words, resistant derivatives have a lower degree of fitness than the parental strain lacking the resistance genes; that is, daughter cells are less competitive for growth in a given ecosystem and in the absence of antibiotic, than the mother cell. The proportion of resistant strains in a bacterial population depends on several factors, such as the concentration and type of antibiotic used, the biological cost of resistance to that antibiotic, and the ability of bacteria to compensate for the fitness cost of the resistance mechanism. Acquisition of antibiotic resistance is often associated with a biological cost because (1) bacteria acquire a new gene (or set of genes) responsible for new functions, (2) the resistance mutations occur in genes with essential functions, or (3) additional energy is required for replication and maintenance of plasmids that bear the resistance genes. The biological cost determines the stability and potential reversibility of resistance.

A compensatory evolution could occur to reduce the biological cost leading to stabilization of the resistant bacteria in a natural population. This process allows resistant strains to regain competitivity relative to their susceptible counterparts in an antibiotic-free environment (Hughes and Andersson, 2001).

However, it has been recently demonstrated that inducibility of resistance is a compensatory mechanism (Foucault et al., 2010). This accounts for the observation that the majority of horizontally acquired antibiotic resistance mechanisms is tightly regulated and that resistance evolves to become selectively neutral in the absence of antibiotics.

Regardless of the mechanism of action of a drug class, it must be realized that resistance already occurs in nature or will inevitably emerge. This is, perhaps, obvious for natural antibiotics, because the producing organisms must avoid self-destruction, but it also holds true for nonnatural drugs. It is thus clear that bacteria are able to resist every antibiotic, naturally or in an acquired

fashion, and selection of resistant bacteria can be regarded as the ultimate criterion for activity of an antibiotic. In addition, and by the mechanisms we have considered, resistance, either by mutation or after acquisition of foreign genetic information, can be drastically enhanced by low concentrations of antibiotics in the environment of the bacteria. Because dissemination of resistance is closely linked to the magnitude of the selective pressure, the only hope is to delay this dissemination. This leaves us with a single recommendation: antibiotics should be used cautiously.

Andersson, D. I. 2003. Persistence of antibiotic resistant bacteria. Curr. Opin. Microbiol. 6:452–6.

Bayley, D. P., E. R. Rocha, and C. J. Smith. 2000. Analysis of cepA and other Bacteroides fragilis genes reveals a unique promoter structure. FEMS Microbiol. Lett. 193:149–54.

Beaber, J. W., B. Hochhut, and M. K. Waldor. 2004. SOS response promotes horizontal dissemination of antibiotic resistance genes. Nature 427:72–4.

Chaslus-Dancla, E., G. Gerbaud, J. P. Lafont, J. L. Martel, and P. Courvalin. 1986a. Nucleic acid hybridization with a probe specific for 3-aminoglycoside acetyltransferase type IV: A survey of resistance to apramycin and gentamicin in animal strains of Escherichia coli. FEMS Microbiol. Lett. 34:265–8.

Chaslus-Dancla, E., J. L. Martel, C. Carlier, J. P. Lafont, and P. Courvalin. 1986b. Emergence of 3-aminoglycoside acetyltransferase IV in Escherichia coli and Salmonella typhimurium from animal in France. Antimicrob. Agents Chemother. 29:239–43.

Chaslus-Dancla, E., P. Pohl, M. Meurisse, M. Marin, and J. L. Lafont. 1991. High genetic homology between plasmids of human and animal origins conferring resistance to the aminoglycosides gentamicin and apramycin. Antimicrob. Agents Chemother. 35:590–3.

Claverys, J. P., and L. S. Havarstein. 2002. Extracellular-peptide control of competence for genetic transformation in Streptococcus pneumoniae. Front. Biosci. 7:1798–1814.

Claverys, J. P., B. Martin, and L. S. Havarstein. 2007. Competence induced fratricide in streptococci. Mol. Microbiol. 64:1423–33.

Collis, C. M., and R. M. Hall. 1995. Expression of antibiotic resistance genes in the integrated cassettes of integrons. Antimicrob. Agents Chemother. 39:155–62.

Courvalin, P. 1994. Transfer of antibiotic resistance genes between Gram-positive and Gram-negative bacteria. Antimicrob. Agents Chemother. 38:1447–51.

Doucet-Populaire, F., P. Trieu-Cuot, I. Dosbaa, A. Andremont, and P. Courvalin. 1991. Inducible transfer of conjugative transposon Tn1545 from Enterococcus faecalis to Listeria monocytogenes in the digestive tract of gnotobiotic mice.. Antimicrob. Agents Chemother. 35:185–7.35:185–7..

Foucault, M.-L., F. Depardieu, P. Courvalin, and C. Grillot-Courvalin. 2010 (in press). Inducible expression eliminates the fitness cost of vancomycin resistance in enterococci. Proc. Natl. Acad. Sci. USA.

Guerin, E., G. Cambray, N. Sanchez-Alberola, S. Campoy, I. Erill, S. Da Re, B. Gonzalez-Zorn, J. Barbé, M.-C. Ploy, and D. Mazel. 2009. The SOS response controls integron recombination Science 324:1034.

Hughes, D., and D. I. Andersson, eds. 2001. Antibiotic development and resistance. London, United Kingdom: Taylor and Francis.

Ivins, B. E., S. L. Welkos, G. B. Knudson, and D. J. Leblanc. 1988. Transposon Tn916 mutagenesis in Bacillus anthracis. Infect. Immun. 56:176–81.

Kohanski, M. A., D. J. Dwyer, B. Hayete, C. A. Lawrence, and J. J. Collins. 2007. A common mechanism of cellular death induced by bactericidal antibiotics. Cell 130:797–810.

Kohanski, M. A., M. A. DePristo, and J. J. Collins. 2010. Sublethal antibiotic treatment leads to multidrug resistance via radical-induced mutagenesis. Mol. Cell. 37:311–20.

Li, L. Y., N. B. Shoemaker, and A. A. Salyers. 1995. Location and characteristics of the transfer region of a Bacteroides conjugative transposon and regulation of transfer genes. J. Bacteriol. 177:4992–9.

Malhotra-Kumar, S., C. Lammens, S. Coenen, K. Van Herck, and H. Goossens. 2007. Effect of azithromycin and clarithromycin therapy on pharyngeal carriage of macrolide-resistant streptococci in healthy volunteers: A randomised, double-blind, placebo-controlled study. Lancet 369:482–90.

Mazel, D. 2006. Integrons: Agents of bacterial evolution. Nat. Rev. Microbiol. 8:608–20.

Perichon, B., and P. Courvalin. 2009. Antibiotic resistance. In Encyclopedia of microbiology, 3rd ed., edited by M. Schaechter. Oxford, United Kingdom: Elsevier. Pp. 193–204.

Phillips, I. 1999. The use of bacitracin as a growth promoter in animals produces no risk to human health. J. Antimicrob. Chemother. 44:725–8.

Prudhomme, M., L. Attaiech, G. Sanchez, B. Martin, and J. P. Claverys. 2006. Antibiotic stress induces genetic transformability in the human pathogen Streptococcus pneumoniae. Science 313:89–92.

Smith, C. J., M. B. Rogers, and M. L. McKee. 1992. Heterologous gene expression in Bacteroides fragilis. Plasmid 27:141–54.

Stobberingh, E. E., and A. E. van den Bogaard. 2000. Spread of antibiotic resistance from food animals to man. Acta Vet. Scand. 93(Suppl.):47–52.

Trieu-Cuot, P., E. Derlot, and P. Courvalin. 1993. Enhanced conjugative transfer of plasmid DNA from Escherichia coli to Staphylococcus aureus and Listeria monocytogenes. FEMS Microbiol. Lett. 109:19–24.109:19–24.

Vivan Miao and Julian Davies¹⁷

The actinobacteria are arguably the richest source of small molecule diversity on the planet. These compounds have an incredible variety of chemical structures and biological activities (in nature and in the laboratory). Their potential for the development of therapeutic applications cannot be underestimated. It is suggested that an improved understanding of the biological roles of low molecular weight compounds in nature will lead to

the discovery an inexhaustible supply of novel therapeutic agents in the next decade. To support this objective, a functional marriage of biochemistry, genomics, genetics, microbiology, and modern natural product chemistry will be essential.

The phylum actinobacteria, one of the largest groups in the domain Bacteria (Figure A5-1), largely consists of environmental bacteria and the denizens of many varied habitats: soils, the rhizosphere, marine and extreme arid environments. A number live in close association with higher organisms; for example, as components of different microbiomes they constitute more than a third of the healthy human microbiota. Members of the genus Frankia, on the other hand, can form symbiotic nodules in certain species of trees and shrubs, and fix atmospheric nitrogen to allow their hosts to survive in nutrient-limiting environments. Actinobacteria typically have elevated guanosine-cytosine contents (65-75% G + C) and their genome sizes range from the 2.5-Mb skin commensal Micrococcus luteus to the 9.7-Mb environmental strain Rhodococcus jostii. Since the discovery of antibiotics in the 1940s, the actinomycetes have received a great deal of attention, and Streptomyces species in particular have become renowned as the principal sources of therapeutic pharmaceuticals. There have been several good reviews on actinobacteria of late, notably that by Ventura et al. (2007) on evolutionary and genomic aspects, as well as occasional articles focusing on specific genera. Interest in the phylum in recent years is evidenced by the increasing number of citations; streptomycetes lead, of course, with the mycobacteria not far behind! However, other genera, including Rhodococcus, are beginning to excite more interest (Larkin et al., 2005, Kitagawa and Tamura, 2008) and who knows, Streptomyces may command less attention in the future.

Streptomycetes are demonstrably a rich source of compounds, but no more so than other members of the actinobacteria, also the Bacilli and bacterial genera such as the myxobacteria (Wenzel and Muller, 2009) and pseudomonads (Gross and Loper, 2009). Among the eukaryotes, fungal genomes are replete with biosynthetic gene clusters for encoding small molecule production. The ability to make bioactive small molecules is not exclusive to microbes. Plants are rich sources of a great variety of compounds that have been used as pharmaceuticals for millennia; this resource remains poorly understood and still largely untapped.

There is a global crisis in the treatment of infectious diseases; people are dying of infections that were previously treatable. Microbes are the source and the solution for the crisis, and for this reason it is imperative that the search for novel therapeutic agents be intensified. The constant moan of the pharmaceutical industry, that the natural reservoir of molecules with antibiotic activity is close to being exhausted and that they can no longer find useful bioactive compounds, is due in part to Waksman’s focus (see below) on the streptomycetes. It can also be explained by the inability to detect bioactive compounds when they are present only in low concentrations; the industry has found all the easily accessible

bioactives, the so-called “low hanging fruit” (Baltz, 2006). Presumably it was not considered essential to develop the technology to find compounds that were missed. In addition, actinobacteria as a whole have been ignored, even though they too possess the capacity to produce a huge number of bioactive small molecules; to date, only a small proportion have been examined for therapeutic purposes. We are now in the “genomic era” and in the case of streptomycetes, exciting new information coming from complete genome sequencing efforts reveals that most of these bacteria have the genetic capacity to produce many more structurally different bioactive compounds than suspected. As such, they represent an inexhaustible collection of hidden chemical and biochemical diversity. Moreover, creative techniques for generating some of these compounds are being developed and exploited (Baltz, 2008; Challis, 2008). We have no excuse for being short of compounds to screen (assuming that reliable screens are in place)! In this short article we present the case for a more extensive survey of the biology, properties and uses of natural molecules, especially those from other members of the actinobacteria. Now that we know the ubiquity and diversity of bioactive small molecules, the most important questions remain: “How do we find them?” and “What are their roles in nature?” When we have the answers to these questions, we will be far better equipped to harvest and exploit this vast chemical and biological wealth.

This characteristic refers mainly to the discovery and production of microbial small molecules with antibiotic activity that began with Waksman’s work on actinomycetes in the early 1940s. These seminal studies, together with the discovery of the fungal product penicillin by Fleming and co-workers and its characterization around the same time, were responsible for momentous and radical changes in medicine. A representative, but grossly incomplete, list is presented in Table A5-1; for more information see the review by Demain and Sanchez (2009). The availability of antimicrobial agents made possible, for the first time, the successful treatment of most types of infectious diseases. The discovery of antibiotics also presaged many other uses for microbial compounds in human and animal therapy and in agriculture. Recently these microbial sources have provided treatments for many non-infectious diseases including cancer and heart disease. Another role, often overlooked, is their use in prophylaxis and in immunosuppression prior to invasive surgery, which has been one of the most important factors in the development and success of organ transplantation. Hundreds of millions of dollars have been invested by pharmaceutical companies in saving the lives of millions of people—with profits of many billions of dollars! In addition, many actinobacterial strains have been developed for industrial applications such as bioremediation, the destruction of toxic xenobiotics, vitamins, fine chemical transformation and production and, more recently, for the development

of biofuel conversions. Novel uses of the extensive biosynthetic capacities of the Rhodococci are being discovered and one can predict their increasing importance as industrial microorganisms (Martinkova et al., 2009).

In spite of the numerous benefits accruing from these seemingly inexhaustible sources, the ecology and biology of actinobacteria and their roles in environmental communities are poorly understood and the functions of their myriad low-molecular-weight products in the environment are even less well studied. The development of these products as antibiotics led to the assumption that their primary (and only) function in nature was for use as molecular weaponry by their producers. The field was driven by the concept of antagonism: during the past half-century, their discovery and the proof of their biological activity relied solely on tests of their inhibition of the growth of other microbes under laboratory conditions. Given the number of bacterial genera and the inestimable number of compounds involved, this implies that the microbial world is nothing less than a constant theatre of war (Hibbing et al., 2009). There is very little sound evidence for this extreme concept and such an anthropocentric viewpoint needs to be discarded.

As with all biologically active compounds, the properties of bacterial products depend on the concentrations at which they are tested, immediately creating a dilemma: what exists and happens in nature is often quite distinct from what is found in a laboratory. Recent studies using sensitive promoter-reporter libraries or RNA microarrays have shown that at sub-inhibitory concentrations many microbial compounds modulate transcription patterns in a variety of bacterial and eukaryotic cells (Yim et al., 2006). Do the transcriptional effects provide the mechanistic basis for their wide range of biological activities? We believe so and have proposed that bioactive compounds act by binding to receptors in cells,

triggering cellular responses that are many and various; in other words they are cell-cell signaling agents (Davies et al., 2006; Fajardo and Martinez, 2008). In the past, studies of mechanisms of antibiotic activity in bacteria led to the identification of specific targets/receptor molecules and macromolecules. These include components of the cell wall, ribosomes, ribosomal RNA, DNA replication, RNA synthesis, as well as numerous enzymatic reactions, such as those involved in the synthesis of fatty acids. There is substantial genetic validation for these interactions and for the roles of single, specific targets in the cell.

Detailed studies with eukaryotic organisms are sparse, but recent results in studies of the different microbial populations that make up the human microbiome indicate that bioactive compounds play important roles in many aspects of human physiology that impact health and disease (Kaper and Sperandio, 2005). It can be predicted that studies of bacterial-mucosal and bacterial-tissue interactions and the role of bioactive small molecules in these processes will be pursued actively in the coming years. The native bacterial communities of humans and other organisms presumably use inter-cellular signaling mechanisms to modulate and control the activities of bacterial consortia and the essential interactions with their host. These interactions have significant implications on related issues such as the activity of probiotics and their roles in regulating immune responses. Similarly, the roles of small-molecule mediation in the operation of distributed metabolic networks in natural microbial communities is another topic that demands scrutiny in the future (Vallino, 2003). Will the next decade be the age of bioactive small molecules?

The evolutionary origins of the great diversity of bacterial products are poorly understood. How old are the actinomycetes? The photosynthetic cyanobacteria are associated with the appearance of oxygen in the earth’s atmosphere, and evidence of bacterial cells in fossilized stromatolites suggests that bacteria are as old as 2.7 Gya; the domain Bacteria includes the most ancient living organisms in the biosphere (Oren, 2004). Detection of hopanoids in ancient shales and also as cell membrane components that play a role in the structure of aerial hyphae in streptomycetes is another clue to the pathway of bacterial evolution (Taylor, 1984). Many actinomycetes, including streptomycetes and the Rhodococci, possess putative genes for gas vesicle production associated with the ability to survive in aqueous environments (as might be found under primordial conditions) (van Keulen et al., 2005).

Both the wide variety of amino acid derivatives found in meteorites and the seminal “primordial soup” experiments by visionaries such as Miller and Urey (Miller et al., 1976) provide chemical evidence for the presence of many types of non-protein amino acid derivatives in the prebiotic world. This leads to the probability that molecules similar to modern nonribosomal peptides are among the oldest bioactive molecules, as is borne out by their extant biological functions and their production by many types of microbes and plants. Their presence defined the evolutionary direction of the earliest forms of life. The evolution of

the biosynthetic pathways for nonribosomal peptides and other natural products such as the polyketides remains unclear, although tangible models for their being have been proposed (Nett et al., 2009; Ridley et al., 2008). The widespread use of similar classes of bioactive compounds in microbial and plant life, their coevolution and coexistence, are clearly of related interest.

From an historical point of view the first useful antibiotics to be discovered and used as such came not from actinomycetes but from members of the Bacilli! The peptide gramicidin was reported by Rene Dubos in 1938, and is still employed. (Check your local pharmacy if you don’t believe this.) There is every reason to believe that all bacteria have the capacity to make similar types of compounds; confirmation comes from the discovery of hybrid NRP-PK toxins produced by certain strains of E. coli (Putze et al., 2009). Recently, a novel non-ribosomal peptide derivative has been isolated from a strain of Staphylococcus (Magarvey, personal communication). This leads to the conclusion that the number of bioactive microbial compounds is, at a minimal estimate, equal to the number of microbial species; therefore, in terms of production of bioactives, all microbes are “good”. So much for suggestions that the supply is close to exhaustion!

We live in an occult universe of low-molecular-weight compounds. Suffice it to say, the antiquity of bioactive small molecules and their huge range of chemical space explains their ubiquity and enormous range of functions in cell biology. Paraphrasing the words of Douglas Adams, the author of The Hitchhikers Guide to the Galaxy: “Microbial chemical space is big. You just won’t believe how vastly, hugely, mind-bogglingly big it is!” As has been suggested on several occasions, the many roles of low-molecular-weight natural products justify their place as elements of the “central dogma” along with DNA, RNA, and protein (Schreiber, 2005). More focused efforts on their biology will reap many intellectual advances along with increasing medical and industrial applications.

However, a major stumbling block is the isolation and characterization of the organic compounds in this vast repertoire. Methods for the chemical identification of microbial products have improved significantly in recent years; however, it still requires an enormous effort to isolate, purify and determine the structures of natural products. Even with the most advanced instrumentation (nuclear magnetic resonance, mass spectroscopy, etc. and various combinations thereof), unraveling the structures of natural compounds remains a slow and highly specialized process. The throughput of current platforms does not in any way meet the needs for identifying thousands of diverse bioactive molecules with multiple biological roles. Until there is a revolutionary advancement in the structure determination process (akin to the effect of pyro-sequencing on genomics) studies of the world of small molecules will lag behind other fields. It is imperative that we decipher the language of small molecules in nature. This major undertaking would provide huge benefits, not the least, novel medicines and the identification of other bioactive molecules with applications in many areas of human and animal health and industry.

It is worth noting that an inventory of bioactive molecules will be only the prelude to developing the methodology required to systematically determine their functions, thereby effecting the metamorphosis of structural data into results and ultimately into molecular understanding. Precious little is known at present about the natural functions of these compounds.

One can identify antibiotics, siderophores, redox-active agents, transcription factors, transporters, and cell signals, etc. (Dietrich et al., 2008), but what are they actually doing in microbial population dynamics? The possibilities are many and the proposed functions must be confirmed under natural conditions (possibly using modern in situ imaging techniques). There is no doubt that working with well-characterized compounds will permit more sophisticated biochemical and genetic studies in target organisms, with the subsequent identification of unsuspected receptors and functions. If macromolecules such as the bacterial ribosome possess dozens of different receptor sites (Yassin and Mankin 2007) this will be a significant enterprise!

We refer of course to parasitic strains that cause disease in other living organisms. From an evolutionary point of view any synergistic relationship can potentially lead to negative interactions; synergy with one partner or host can easily be translated into pathogenic interactions with another. The total number of known human and animal microbial pathogens is currently limited to a few thousand or so (including viruses); this is but a small percentage of the Bacteriaceae (Taylor et al., 2001). One can predict that the number may be much larger for plant pathogens.

We hear much about emerging pathogens in clinical studies: there are two broad classes: those organisms to which humans are newly exposed as a result of anthropogenic activities (reclaiming land, forest destruction, or social practices) and those dedicated pathogens that have recently acquired antibiotic resistance by mutation or horizontal gene transfer and thereby overcome/bypass the prevailing therapeutic options. Relatively few actinobacteria fall into the category of professional pathogens (that we know of) (Table A5-2).

Historically, M. tuberculosis is the most important pathogen and remains the most widely disseminated; there is evidence of human infection for 9,000 years (Hershkovitz et al., 2008). The total number of human deaths due to TB throughout history is not known, but it is estimated that M. tuberculosis caused at least 200 million deaths in the twentieth century (Kaufman and Van Helden, 2008). The first streptomycete-derived antibiotic and the most successful, streptomycin, was developed for the purpose of combating TB, the “White Plague”. At the present time, it is difficult to appreciate the incredible importance of the discovery of streptomycin; we have become nonchalant about the control of infectious diseases. Other critical drug discovery events in the 1940s built the reputation of

the actinomycetes and established the bias towards this family of soil bacteria as producers of antibiotics. Although penicillin and several antibiotics from Bacillus spp. predated streptomycin in therapeutic use, they did not cure TB!

Other pathogenic mycobacteria are of significance, such as M. leprae (leprosy) and M. ulcerans (buruli ulcers). Among the actinomycetes, Rhodococcus equi has been recently identified as an equine infection that is an opportunistic pathogen for humans. Interestingly, the disease-causing actinomycetes evolved primarily by extensive genome reduction compared to their environmental precursors, rather than by the horizontal gene transfer of myriad pathogenicity islands associated with the Gram-negative pathogens such as E. coli, etc. M. tuberculosis is a notoriously difficult organism to work with due to its virulence, slow growth, and, until recently, lack of facile genetic manipulation; thus comparative studies of close relatives among the actinomycetes have provided important information on novel aspects of mycobacterial metabolism and mechanisms of virulence. For example, there is the important question of how M. tuberculosis survives in human macrophages. This has been revealed by comparative genomic analyses with the genome sequence of Rhodococcus jostii RHA1 that identified a gene cluster encoding a possible cholesterol degradation pathway (McLeod et al., 2006). The observation that this matched a closely related sequence in M. tuberculosis led to studies that have shown that the pathogen does indeed use cholesterol as a carbon source, providing critical information on its intracellular survival mechanisms and the possibility of novel targets for TB drug development (Van der Geize et al., 2007).

There are no ugly actinomycetes. However, for every silver lining there is a cloud, and this family is no different. It has been demonstrated that most of the common antibiotic resistance genes or their progenitors have their origins in environmental bacteria, and evidence suggests strongly that actinobacteria may be one of the main natural sources of clinically significant antibiotic resistance genes (Wright, 2007). On the other hand, the actinomycetes produce the

clavulanate-derived inhibitors of β-lactamases and also enzymes that degrade the acylhomoserine lactones, signal molecules that are responsible for the induction of virulence functions in a number of common pathogens. If we had been smart enough to recognize this fact earlier, it might have been possible to devise inhibitors of these resistance mechanisms and so defuse the pathogens prior to years of unrestricted antimicrobial therapy.

We have mentioned the “occult universe of small molecules” and will conclude with a few additional comments on this theme. The existence and the roles of low-molecular-weight organic compounds in biology have been all but ignored. Despite a century of studies of the chemistry, physiology, and critical roles of vitamins, neurotransmitters, pheromones, alkaloids, and other useful products of plants and animals, the chemical store of the microbial world remains a great mystery. Recent studies of the phenomenon of quorum sensing communication have taught microbiologists and chemists that small can be beautiful and meaningful (Atkinson and Williams, 2009; Winans and Bassler, 2008). However, quorum-sensing activities, like antibiotic effects, are still largely studied as laboratory phenomena that do not necessarily represent the environmental roles of organic compounds; it remains difficult to assess concentrations of the signaling compounds in the wild. What is needed is more science and much less anthropocentricity; the latter provides the substance of exciting movies but is bad science. (Admittedly, for convenience we have slipped into anthropomorphic mode by using the descriptors “good”, “bad” and “ugly”; this is almost as lamentable as saying that bacteria “decide”, or “make lifestyle choices”, phrases seen in many publications!)

Finally, the diversity and ubiquity of bioactive small molecules, their multitudinous sources, and their potential and critical roles in the functioning and interactions of all living things lead us to propose that there should be significant, targeted funding initiatives (and perhaps even institutes) devoted to their study: chemical (structural and synthetic), genetic, biological, physical, imaging, etc. Surely, the increasing interest in systems biology will benefit from a full understanding of small-molecule biology? We can do no better than to quote the proverb “from small beginnings come great things”.

We are grateful to Dorothy Davies for her patient editing assistance and Dr. Marco Ventura for permission to reproduce Figure A5-1. Funding has been provided by the National Science and Engineering Research Council, the Canadian Institutes for Health Research, Merck Research Laboratories, and the Tally Fund.

Atkinson S, Williams P (2009) Quorum sensing and social networking in the microbial world. J R Soc Interface 6:959–978

Baltz RH (2006) Marcel Faber roundtable: is our antibiotic pipeline unproductive because of starvation, constipation or lack of inspiration? J Ind Microbiol Biotechnol 33:507–513

Baltz RH (2008) Renaissance in antibacterial discovery from actinomycetes. Curr Opin Pharmacol 8:557–563

Challis GL (2008) Mining microbial genomes for new natural products and biosynthetic pathways. Microbiology 154:1555–1569

Davies J, Spiegelman GB, Yim G (2006) The world of subinhibitory antibiotic concentrations. Curr Opin Microbiol 9:445–453

Demain AL, Sanchez S (2009) Microbial drug discovery: 80 years of progress. J Antibiot (Tokyo) 62:5–16

Dietrich LEP, Teal TK, Price-Whelan A, Newman DK (2008) Redox-active antibiotics control gene expression and community behavior in divergent bacteria. Science 321:1203–1206

Fajardo A, Martinez JL (2008) Antibiotics as signals that trigger specific bacterial responses. Curr Opin Microbiol 11:161–167

Gross H, Loper JE (2009) Genomics of secondary metabolite production by Pseudomonas spp. Nat Prod Rep 26:1408–1446

Hershkovitz I, Donoghue HD, Minnikin DE, Besra GS, Lee OY, Gernaey AM, Galili E, Eshed V, Greenblatt CL, Lemma E, Bar-Gal GK, Spigelman M (2008) Detection and molecular characterization of 9000-year-old Mycobacterium tuberculosis from a neolithic settlement in the eastern Mediterranean. PLos One 3:e3426

Hibbing ME, Fuqua C, Parsek MR, Peterson SR (2009) Bacterial competition: surviving and thriving in the microbial jungle. Nat Rev Microbiol 8:15–25 Antonie van Leeuwenhoek

Kaper JB, Sperandio V (2005) Bacterial cell-to-cell signaling in the gastrointestinal tract. Infect Immun 73:3197–3209

Kaufman SHE, van Helden P (2008) Handbook of tuberculosis vol. 3: clinics, diagnostics, therapy and epidemiology. Wiley-VCH, Weinheim

Kitagawa W, Tamura T (2008) Three types of antibiotics produced from Rhodococcus erythropolis strains. Microbes Environ 23:163–171

Larkin MJ, Kulakov LA, Allen CC (2005) Biodegradation and Rhodococcus—masters of catabolic versatility. Curr Opin Biotechnol 16:282–290

Martinkova L, Uhnakova B, Patek M, Nesvera J, Kren V (2009) Biodegradation potential of the genus Rhodococcus. Environ Int 35:162–177

McLeod MP, Warren RL, Hsiao WW, Araki N, Myhre M, Fernandes C, Miyazawa D, Wong W, Lillquist AL, Wang D, Dosanjh M, Hara H, Petrescu A, Morin RD, Yang G, Stott JM, Schein JE, Shin H, Smailus D, Siddiqui AS, Marra MA, Jones SJM, Holt R, Brinkman FSL, Miyauchi K, Fukuda M, Davies JE, Mohn WW, Eltis LD (2006) The complete genome of Rhodococcus sp. RHA1 provides insights into a catabolic powerhouse. Proc Natl Acad Sci USA 103:15582–15587

Miller SL, Urey HC, Oro J (1976) Origin of organic compounds on the primitive earth and in meteorites. J Mol Evol 9:59–72

Nett M, Ikeda H, Moore BS (2009) Genomic basis for natural product biosynthetic diversity in the actinomycetes. Nat Prod Rep 26:1362–1384

Oren A (2004) Prokaryote diversity and taxonomy: current status and future challenges. Philos Trans R Soc B 359:623–638

Putze J, Hennequin C, Nougayre`de J-P, Zhang W, Homburg S, Karch H, Bringer M-A, Fayolle C, Carniel E, Rabsch W, Oelschlaeger TA, Oswald E, Forestier C, Hacker J, Dobrindt U (2009) Genetic structure and distribution of the colibactin genomic island among members of the family Enterobacteriaceae. Infect Immun 77:4696–4703

Ridley CP, Lee HY, Khosla C (2008) Evolution of polyketide synthases in bacteria. Proc Natl Acad Sci USA 105:4595–4600

Schreiber SL (2005) Small molecules: the missing link in the central dogma. Nat Chem Biol 1:64–66

Taylor RF (1984) Bacterial triterpenoids. Microbiol Mol Biol Rev 48:181–198

Taylor LH, Latham SM, Woolhouse MEJ (2001) Risk factors for human disease emergence. Philos Trans R Soc B 356:983–989

Vallino JJ (2003) Modeling microbial consortiums as distributed metabolic networks. Biol Bull 204:174–179

Van der Geize R, Yam K, Heuser T, Wilbrink MH, Hara H, Anderton MC, Sim S, Dijkhuizen L, Davies JE, Mohn WH, Eltis LE (2007) A gene cluster encoding cholesterol catabolism in a soil actinomycete provides insight into Mycobacterium tuberculosis survival in macrophages. Proc Natl Acad Sci USA 104:1947–1952

Van Keulen G, Hopwood DA, Dijkhuyizen L, Sawers RG (2005) Gas vesicles in actinomycetes: old buoys in novel habitats? Trends Microbiol 13:350–354

Ventura M, Canchaya C, Tauch A, Chandra G, Fitzgerald GF, Chater KF, van Sinderen D (2007) Genomics of actinobacteria: tracing the evolutionary history of an ancient phylum. Microbiol Mol Biol Rev 71:495–548

Wenzel SC, Muller R (2009) The impact of genomics on the exploitation of the myxobacterial secondary metabolome. Nat Prod Rep 26:1385–1407

Winans SC, Bassler BL (eds) (2008) Chemical communication among bacteria. ASM Press, Washington, DC

Wright GD (2007) The antibiotic resistome: the nexus of chemical and genetic diversity. Nat Rev Microbiol 5:175–186

Yassin A, Mankin AS (2007) Potential new antibiotic sites in the ribosome revealed by deleterious mutations in RNA of the large ribosomal subunit. J Biol Chem 282:24329–24342

Yim G, Wang HH, Davies J (2006) The truth about antibiotics. Int J Med Microbiol 296:163–170

Michael A. Fischbach¹⁹ and Christopher T. Walsh^20,21

Antibiotic-resistant strains of pathogenic bacteria are increasingly prevalent in hospitals and the community. New antibiotics are needed to combat these bacterial pathogens, but progress in developing them has been slow. Historically, most antibiotics have come from a small set of molecular

scaffolds whose functional lifetimes have been extended by generations of synthetic tailoring. The emergence of multidrug resistance among the latest generation of pathogens suggests that the discovery of new scaffolds should be a priority. Promising approaches to scaffold discovery are emerging; they include mining underexplored microbial niches for natural products, designing screens that avoid rediscovering old scaffolds, and repurposing libraries of synthetic molecules for use as antibiotics.

There is a perpetual need for new antibiotics: Whereas most drugs will be just as effective in the future as they are today, the inevitable rise of resistance will erode the utility of today’s antibiotics (Walsh, 2003). Two factors exacerbate this supply problem by creating unique disincentives for antibiotic development (Nathan et al., 2005). First, antibiotics are used in smaller quantities than other drugs. Prescriptions for chronic illnesses can last years or decades, whereas a standard course of antibiotics lasts only weeks; therefore, antibiotics yield lower revenues than most drugs. Second, whereas most newly approved drugs can be prescribed to all who would benefit, the use of a newly approved antibiotic may be restricted to the treatment of serious bacterial infections. The result is a quandary: Resistance is on the rise while antibiotic discovery and development are on the decline (Nathan, 2004; von Nussbaum et al., 2006).

The unfavorable economics of antibiotic development have had a chilling effect on industrial discovery programs, and policy-based efforts to reverse this decline deserve attention (Nathan, 2004). This perspective focuses on a different, yet no less formidable, challenge: finding new classes of antibiotics.

On the face of it, antibiotic discovery would seem to be straightforward. The goal is to kill an organism that is only distantly related to humans; unique, essential targets should be abundant, and novel antibiotics with low toxicity should be easy to find. Yet, the history of antibiotic development suggests otherwise. Since the early 1960s, only four new classes of antibiotics have been introduced, and none of these has made a major impact yet; the ~$30 billion global antibiotics market is still dominated by antibiotic classes discovered half a century ago. Since then, most “new” antibiotics have been chemically tailored derivatives of these well-worn scaffolds. In this review, we argue that the rise of resistant pathogens should redouble our focus on discovering not just new antibiotics, but new classes of antibiotics. We then highlight some promising approaches to scaffold discovery: mining under explored microbial niches for natural products, designing screens that avoid rediscovering old scaffolds, and repurposing libraries of synthetic molecules for use as antibiotics.

Three classes of antibiotic-resistant pathogens are emerging as major threats to public health (Figure A6-1). First, methicillin-resistant Staphylococcus aureus