Epidemiology and Virulence Insights From MRSA and MSSA Genome Analysis

Vladimir Lazarevic; Marie Beaume; Anna Corvaglia; David Hernandez; Jacques Schrenzel; Patrice François


Future Microbiol. 2011;6(5):513-532. 

In This Article

Can We Predict the Virulence of Staphylococcus aureus Based on its Genome/Transcriptome?

Microbial Toxins & Adhesins

Several lines of evidence demonstrate that acute infections are related to the capacity of the bacterium to secrete exotoxins[15,16] and hydrolytic enzymes.[17,18] The production of these toxins involves numerous and complex regulation pathways.[42] Many staphylococcal toxins are pore-forming molecules able to contribute to cellular damage through their insertion in host cytoplasmic membranes. The α-, β-, δ- and γ-toxins are hemolysins shown to damage the host cellular integrity.[43–45] PVL is also an important pore-forming toxin targeting neutrophils and often found in CA-MRSA isolates.[46,47] The exfoliative toxins, including epidermolytic toxins A and B, cause skin erythema and separation, as evidenced in the staphylococcal scalded skin syndrome.[48] The toxic shock syndrome toxin (TSST-1) is highly virulent and is involved in severe S. aureus infections[19,49,50] and is harbored by 20% of S. aureus strains.[17]

In addition, the species S. aureus is able to produce more than 20 enterotoxins that display superantigenic properties, corrupting immune defenses.[51] Some of them cause severe food-poisoning.[52,53] The contribution of most of these toxins has been documented in various experimental models of acute infections.[20,21,43,54]

Besides toxins, S. aureus is able to produce several cell-wall-associated proteins allowing bacterial interactions with host plasma or extracellular proteins such as fibronectin, fibrinogen, collagen, vitronectin, laminin or bone sialoprotein.[55] The contribution of these bacterial compounds to the virulence of S. aureus and the regulation of their expression remain to be studied. For example, the contribution of fibrinogen- or fibronectin-binding proteins in infective endocarditis[56,57] and the role of the collagen-binding protein in septic arthritis[58] are only partly known, as is the case with the ability of S. aureus to persist intracellularly.[59–62] Most of the epidemiological or clinical studies associated with highly diverse strains are not fully understood since the studies conducted so far have relied on single-gene mutant strains. These genetically engineered mutants can show the impact of a single loss of function; however, compensatory mechanisms that can be triggered at the organism level are not readily recognized.

The host response appears essential in the severity of symptoms triggered by S. aureus toxins. Recent studies dealing with the mechanisms of action of these potent molecules, particularly PVL, clearly demonstrated that innate immunity was involved through binding and/or interaction of the toxin with TLR2 and CD14[63] in alveolar macrophages. In the context of infection, the role of bacterial peptidoglycan and lipoproteins present on bacterial envelopes also appears crucial in mediating TLR2-dependent activation of the NF-κB pathway and subsequent cytokine production. In addition, these natural S. aureus-derived cell-wall fragments are potent stimulators of the innate immune system.[64]

Defence of Genomic Integrity & Acquisition of Resistance Determinants

Restriction-modification systems are important defenses against interspecies and even intraspecies uptake of exogenous and potentially harmful (e.g., bacteriophage) DNA. To date, four types of restriction-modification system have been described: type I, type II, type III and type IV, named following their order of discovery. Types II and IV are composed of a methyltransferase, which, by modifying specific DNA sequence, protects 'self' DNA from cleavage, and an endonuclease allowing digestion of nonprotected DNA. In the type I system, the specificity subunit 'HsdS' is responsible for the recognition of the target sequence, which has to be methylated by HsdM. In S. aureus, the SauI type I system is the only chromosomally encoded system and it is widely distributed among recovered clinical isolates.[65] Type II restriction enzymes cut DNA at specific recognition motifs and are therefore widely used as molecular biology tools. The presence of restriction-modification systems limits the acquisition of foreign DNA in natural and laboratory conditions, and thus the possibility to construct mutants of S. aureus clinical isolates. For this reason, strain RN4220 has become an essential intermediate for laboratory manipulations of S. aureus. This strain, obtained by mutagenesis including chemical and UV exposition,[66] contains a nonsense mutation in the hsdR gene encoding the endonuclease of the type I restriction-modification system,[65] as recently documented in the complete genome sequence.[67] This type I system also includes the hsdS and hsdM genes that encode proteins involved in the specificity of the sequence recognition and the methylase, respectively. Integrity in the functionality of these systems appears to be protection against genetic evolution. Conversely, mutations affecting the biological integrity of these restriction systems could confer an increased susceptibility to the acquisition of genes influencing the virulence or epidemicity of strains.

Inactivation of the hsdR gene in the clinical isolate UAMS-1 allowed efficient transduction with phages as well as transformation with plasmids, prepared from strain RN4220. However, transformation with plasmid DNA prepared from Escherichia coli was still impossible suggesting the presence of an additional barrier for horizontal gene transfer. This barrier was finally removed by inactivation of a gene showing 98% identity to ORF2790 of strain NTC8325.

Sequencing of the type I- and type III-like restriction system genes in an endemic MRSA strain of sequence type 228 (ST228), easily transformable with plasmid DNA prepared from E. coli, revealed a deletion in the ORF2790 homolog. It is interesting to note that the bovine pathogen lineage ST151 is highly susceptible for the acquisition of the vanA-encoding transposon, naturally harbored by enterococci, and that this lineage has mutations in both hsdS1 and hsdS2 genes.[68] Importantly, the sequenced genome of strain RF122 of the ST151 lineage does not contain any homolog of the newly discovered restriction endonuclease, which, according to recent biochemical characterization, belongs to the type IV modification-restriction system.[69]

It has also been shown that the S. aureus type IV endonuclease gene, originally described as a type III-like gene, plays an important barrier function for the transfer of the vancomycin resistance transposon from Enterococcus faecalis.[70] The expression of this novel restriction system potentially limits the acquisition of exogenous DNA such as vancomycin resistance determinants and the taphylococcal cassette chromosome mec (SCCmec) element, by clinical strains. The identification of the specificity determinants of this type IV restriction endonuclease and the distribution analysis of this barrier to horizontal gene transfer will be crucial for further understanding of resistance and virulence genes spreading in S. aureus. Disruption of genes encoding restriction endonuclease systems would render many clinical strains of this versatile pathogen amenable to genetic manipulations, providing the possibility to better study its physiology and virulence.[70]

The emergence of bacterial resistance in S. aureus has limited the efficacy of most antimicrobial drug classes. The most dangerous threat is MRSA strains that harbor additional antibiotic resistance determinants, thus warranting the utilization of last-line drugs such as the glycopeptides. More recently, the emergence of MRSA with reduced susceptibility to vancomycin has raised a particularly serious concern.[4,34] Control of HA infections caused by such strains requires expensive surveillance programs including patient isolation and contact precautions.[6]

Staphylococcal resistance was reported shortly after the introduction of penicillin, with Kirby's first description of penicillinase-producing strains in 1944.[71] Penicillin is inactivated by penicillinase (β-lactamase), a serine protease that hydrolyzes the β-lactam ring of the antibiotic. The prevalence of penicillinase-producing strains of S. aureus within hospitals began to rise as soon as penicillin became available after World War II and within approximately 6 years, 25% of hospital strains were resistant.[35] To date, less than 5% of isolates remain susceptible to penicillin G.[17]

While the first S. aureus isolates resistant to methicillin, a semi-synthetic penicillin derivative, were reported in the early 1960s,[72] endemic MRSA strains carrying multiple additional resistance determinants only became a worldwide nosocomial problem in the early 1980s.[73,74] The presence of MRSA in a healthcare setting is clearly paralleled by an increased rate in bacteremia or other severe MRSA infections,[75,76] and it has a major impact on the length of hospital stay.[77]

The spread of MRSA in healthcare centers is difficult to control and requires elaborate infection control guidelines[76,78–83] including: large-scale screening of suspected carriers, automated computerized alerts, specific recommendations for at-risk patients, such as contact isolation and significant improvement of hand hygiene compliance. These data, together with successful containment effort programs,[76,78–83] prompt for screening high-risk patients even in a highly endemic setting.[84] However, despite intensive efforts in the application of such guidelines, MRSA spread remains difficult to control. The difficulty to eradicate nosocomial MRSA infections may be explained by the presence of an unknown hidden reservoir of MRSA carriers, or the emergence of novel highly epidemic S. aureus clonotypes[35] such as EMRSA-15 and EMRSA-16.[85–87] Some authors reported that the screening of high-risk patients for MRSA colonization was a cost-effective measure for limiting the spread of the organism in hospitals.[79] Early detection of MRSA carriers is also crucial for therapeutic decision to use last-line antibiotics against MRSA, for example glycopeptides and oxazolidinones.[88] In the clinical routine laboratory, S. aureus is isolated on poorly selective culture media and then presumptively identified before definitive overnight characterization,[89] which is a time-consuming process.

The genetic basis of methicillin-resistance in S. aureus isolates is related to the presence of the mecA gene, which is part of the SCCmec, a mobile genetic element that is invariably inserted in orfX.[90] To date, eight differently organized SCCmec elements have been characterized.[91–94] However, numerous recent works report strains showing nontypable SCCmec cassettes and new types or subtypes of organization, thus prompting for a new classification.[90,94,95] Bacterial strains responsible for CA infections depend on the local epidemiology and were recently described as polyclonal by several research groups,[96,97] whereas others reported more uniformity in CA-MRSA lineages.[98] As previously suspected, the epidemiology of MRSA is evolving[35] and CA-MRSA has recently been shown to be responsible for nosocomial infections.[99,100]

Genotyping Methods for S. aureus

The ability to rapidly and reliably identify relatedness between clinical isolates is crucial for the investigation of outbreaks and also for epidemiological surveillance. Both phenotypic and genotypic methods have been used to characterize S. aureus strains (see review[101]). Phenotyping relies on: antimicrobial susceptibility panels, use of specific antibodies directed against bacterial surface components, or phage pattern determination.[86] These techniques are generally considered poorly discriminative, time consuming and slow.[102] PCR-based techniques appear advantageous to study slow growing or difficult to grow organism.[103]

An array of genotyping methods relying on characteristics of specific genes or parts of S. aureus genomes were developed and used for epidemiological purposes (Table 1).

Pulse-field Gel Electrophoresis Pulse-field gel electrophoresis (PFGE) is a method that is based on the enzymatic digestion of the bacterial genome. For genome fingerprinting, the bacterial chromosome is cleaved with a restriction endonuclease that generates gel-resolvable fragments. Typically, five to a few dozen fragments are obtained and resolved on the electrophoresis gel. Restriction fingerprinting allows the identification of bacterial strains and the distinction between related and unrelated strains.[104] PFGE represents the gold standard method for outbreak analysis.[102]

Multi-locus Sequence Typing Multi-locus sequence typing (MLST) is a typing technique for strain characterization that indexes variation in multiple housekeeping genes. The method relies on the sequencing of PCR-amplified genomic fragments of approximately 400–600 bp. This amplicon length was chosen because it could be reliably read in a single run on the gel-based automated sequencing instruments of the mid-1990s. Following sequencing, a database containing all allele types[105] is queried with each fragment. MLST stands as the method of choice to evaluate strain relatedness.[105]

Spa-typingSpa-typing is based on the polymorphism of the variable region of protein A,[106] a well known species-specific marker. Both copy number and sequence of the 24-bp tandem repeats vary between S. aureus strain genotypes. Similarly to MLST, spa-typing facilitates interlaboratory data comparisons as it consists of PCR amplification and sequence analysis of a single target gene. The web database provides a standardized international nomenclature of spa types.[302]

Other Genotyping Methods Molecular strategies with lower cost and reduced turn-around times have been reported to study strain relatedness or outbreaks, such as random amplification of polymorphic DNA (RAPD),[107] amplified fragment length polymorphism (AFLP)[108] or multiple-locus variable-number tandem-repeat analysis (MLVA).[109] Genotyping methods compatible with high-throughput analysis allow evaluation of population dynamics, as recently shown by Melles and co-workers. A similar approach was used to study the population dynamics of nasal carriage isolates and their potential origin, either related to community or hospital. Another important contribution was provided by Kuhn et al. who developed an original sequence-based analysis of genes belonging either to the core adhesin or accessory gene categories.[110] The limited number of parameters evaluated by these methods results in a rather moderate resolution power.

High-throughput Sequencing

The molecular basis of S. aureus virulence in mammals appears highly complex and involves a large number of bacterial compounds, the presence of numerous molecular pathways that control the expression of virulence factors and that are probably highly interdependent and very elaborate, the ability of S. aureus to persist in suboptimal growth conditions through a network of stress response pathways not fully elucidated[111–113] and capability to acquire foreign genetic material. Molecular identification of S. aureus virulence factors was first based on standard genetic screens of single virulence-encoding gene mutants and then, following the advent of the genomic era, on whole genome sequencing. The description of the whole 2.8-Mb genome sequence of the two MRSA strains Mu50 and N315 in 2001[114] was the starting point for more global, genome-wide studies of molecular pathogenesis and molecular epidemiology. Subsequently, the whole genome sequence of a highly virulent CA-MRSA,[115] followed by those of three additional S. aureus strains from the USA and Europe[116,117] became available. These sequencing efforts coupled to the development of high-density microarrays have provided powerful new tools for studying the evolutionary genomics of S. aureus.[118–121]

Today, over 1000 complete prokaryotic genomes are publicly available.[303] Exploitation of this huge amount of sequence information has provided new insights on the genetic diversity of bacterial pathogens. The availability of bacterial genome sequences has opened the way to comparative genomics techniques[122] for evaluating relatedness and diversity in gene composition across and within species. In addition, the genome comparisons allowed identification of acquisition of novel genes[123] and genome reduction.[124] These events contribute to the bacterium's ability to survive in a variety of environments.

Harris and colleagues reported the whole genome sequences of 69 'clonal' S. aureus isolates, from various geographical origins. This approach, providing a high-resolution view of the epidemiology, revealed the global geographic structure within the MRSA ST239 lineage, and its intercontinental transmission over four decades. In addition, they were able to trace the spread of the isolates within hospitals[125] and to detect point mutations, particularly in vital genes belonging to the core genome, specifying targets of some antimicrobial compounds. Harris and colleagues revealed that local usage of drugs has a direct molecular impact on identified target genes. Selection pressure is the driving force of the evolution of the bacterial genome. This observation potentially drives local epidemiology and provides the 'chance' to observe the sudden spreading of a new clone in any country, which, although rare, is possible. A strain will probably not be able to acquire mutations or exogenous genetic material in a period of time short enough to allow its survival in an environment that is different from its 'usual one'. Displacement of a prevalent clone has only been rarely reported. In that context, the risk of large-scale spreading of CA-MRSA, generally susceptible to many antibiotics, appears limited. As described in the past by Edward Feil and co-workers, MRSA exists in a relatively clonal population,[126] showing propensity to favor genetic material exchange between 'sisters or brothers' but not between 'cousins or uncles'. Conceptually, limiting the capacity to exchange genetic material when one is potentially exposed to varied environmental conditions appears questionable. S. aureus and MRSA are able to limit genetic material exchange through the utilization of efficient restriction systems or other incompatibilities in transfer systems.

Molecular methods with a moderate resolution used for decades for the characterization of clinical isolates were adequate to detect outbreaks in the hospital setting. However, using these standard methods, many genetic changes may remain unseen and genetically heterogeneous populations may be wrongly considered as clonal. Nowadays, a full bacterial genome can be obtained in a few days with a resolution at the level of the single nucleotide. Comparison of genomic sequences of S. aureus isolates will likely challenge the current concept of 'clonality'.

Use of Molecular Genotyping: Association of Virulence with Genome Content?

Classic genotyping strategies allow identification of bacterial outbreaks[102,127] and to establish phylogenetic links between isolates.[105,126] Alternative molecular methods with lower costs and reduced turn-around times such as RAPD,[107] AFLP[108] and MLVA[109] have been used to study strain relatedness or outbreaks. These methods, amenable to high-throughput analysis, allow the evaluation of population dynamics in clinical isolates of S. aureus or MRSA. Melles and co-workers used AFLP to analyze a collection of over 900 clinical isolates identified as carriage or infecting strains in well-characterized epidemiological situations.[108] The authors described the expansion of multiresistant clones and clones associated with skin disease, suggesting the epidemic potential of pathogenic strains of S. aureus. Subpopulations harboring specific toxin patterns were found in S. aureus strains responsible for abscesses and arthritis. In contrast, carriage strains were rarely found in the category of toxin producer. Finally, the most important observation was that almost any S. aureus clone carried by humans could evolve into a serious and occasionally life-threatening pathogen.

This discovery suggests that any S. aureus strain can potentially evolve to become a pathogen adapted to cause different types of infection. However, it should be noted that the resolution of the genotyping methods can impact the reliability of such discoveries. We explored the possibility that the host initiates bacterial genomic changes that modulate the behavior of the infecting strain by converting a carriage isolate into a dangerous one. The observation that severe systemic infections are usually caused by the strain previously carried by the patient [SCHERRER S, UNPUBLISHED DATA] favors this hypothesis and appears to be in accordance with pioneering work in the field.[128]

More recently, using a sequence-based analysis of core adhesin genes and accessory adhesin genes,[110] Kuhn and colleagues revealed high polymorphism in sporadic or epidemic clones and documented recombination events between related isolates belonging to the same clonal complexes. Such recombination events in a bacterium described as clonal[126] might have contributed to limit S. aureus evolution. The integrity of restriction systems appears to be a barrier to genetic exchanges between distant strains.[65,129]

A method showing higher resolution power based on microarrays has also been used to identify markers of epidemicity or virulence. Fitzgerald and colleagues reported the use of microarrays to evaluate genetic variation in a collection of S. aureus strains isolated from human and bovine infections and representing a subset of the most abundant S. aureus lineages worldwide.[130] Based on the genome of strain COL, the authors developed a whole-genome DNA microarray and analyzed the genomic composition of 36 unrelated S. aureus strains.[130] Representative strains from well-defined human clinical diseases were analyzed, including toxin-producing isolates and MRSA strains, and compared with representative strains from bovine and ovine intramammary infections. 78% of the chromosomal genes were common to all strains examined, identifying 'core genes' essential for cell maintenance and growth. Conversely, 22% of genes were not present in every strain, but were assumed to play a role in the adaptation to specialized niches, or other contingency functions. A total of 18 large chromosomal regions of difference (RD) ranging from 3 to 50 kb were identified, and these were variably present among the strains examined. Several RDs showed extensive variation in size and gene content, suggesting that they may represent mobile genomic regions with increased recombination activity. S. aureus clinical isolates are generally considered as a highly clonal population[126,131] but the equivalent of the 400 kb of the strain COL genome, showing divergence between various strains, contains large polymorphic regions that encode many putative virulence factors, toxins and determinants of host and disease specificity.[130,132]

Several of these RDs contain genes encoding gene-mobility proteins such as integrases or transposases, or code for virulence determinants or proteins mediating antibiotic resistance. These data strongly suggest that horizontal gene transfer and recombination may play a fundamental role in the genetic evolution of pathogenic S. aureus strains. This study also provided important insights into the evolution of MRSA strains, suggesting that the mec element encoding the methicillin-resistance phenotype has been transferred several times by horizontal gene transfer into different methicillin-susceptible S. aureus (MSSA) genetic backgrounds. Overall, this comparative genomics study provided important hints to understanding of genetic variation and evolution in natural populations of S. aureus and demonstrated the power of DNA microarray technology.

Moore and colleagues used microarrays covering whole S. aureus genomes to document epidemiologic features of HA-MSSA.[133] They identified variable regions subject to frequent horizontal transfers, which also contain important virulence factors. Another study by the same research group described the genome content of the ten major S. aureus lineages involved in human infections.[134] The authors were able to identify 'core variable' genes, a gene category composed mainly of regulators of virulence genes or cell-wall-associated targets involved in the interaction with the host during colonization and infection. For each lineage, the isolates showed frequent variation in the carriage of mobile genetic elements indicating frequent horizontal transfers. A part of these elements also contained virulence factors. The situation is similar for other pathogenic bacteria such as E. coli. A recent report showed that E. coli strains responsible for bacteremia are distributed over the entire spectrum of E. coli phylogenetic diversity whereas clonal complexes represent important units for pathogenesis.[135]

The above mentioned studies identified common features within each S. aureus lineage but failed to detect any association between lineages or gene content and clinical outcomes. This suggests either that the host itself contributes to the severity of the infection or, more probably, that the evaluation of the presence/absence of coding sequences in bacteria could not adequately assess this interaction. Utilization of various genotyping methods failed to identify common epidemiological traits or virulence features among MRSA lineages. Thus, the question of "what makes a successful pathogen?"[136] remains open.

The experimental procedures and designs used so far have not been able to assess subtle differences within bacterial populations. In addition, microarrays used for epidemiological purposes suffered from several biases: they were generally built with long PCR products (thus, using suboptimal probes in terms of specificity), they contained a limited number of sequenced genomes (thus, moderately representing the diversity observed in clinical isolates), and they assessed intergenic regions poorly.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: