12 16 S rRNA Methyltransferases : An Emerging Resistance Mechanism Against Aminoglycosides in Salmonella

Infections with non-typhoidal Salmonella are the second most common cause of bacterial gastroenteritis in developed countries, with the incidence of multidrug-resistant non-typhoidal Salmonella increasing considerably in the last two decades. Non-typhoidal salmonellosis in otherwise healthy individuals usually results in mild, self-limiting diarrhoea, but treatment with an appropriate antimicrobial can be life-saving in vulnerable patient groups such as the elderly, immunocompromised patients or those with underlying disease, and in cases of invasive disease such as Salmonella bacteraemia, osteomyelitis and meningitis.


Introduction
Infections with non-typhoidal Salmonella are the second most common cause of bacterial gastroenteritis in developed countries, with the incidence of multidrug-resistant non-typhoidal Salmonella increasing considerably in the last two decades. Non-typhoidal salmonellosis in otherwise healthy individuals usually results in mild, self-limiting diarrhoea, but treatment with an appropriate antimicrobial can be life-saving in vulnerable patient groups such as the elderly, immunocompromised patients or those with underlying disease, and in cases of invasive disease such as Salmonella bacteraemia, osteomyelitis and meningitis.
Aminoglycosides are often used in combination with broad-spectrum ß-lactams for the treatment of life-threatening infections due to both Gram-positive and Gram-negative bacteria due to their potent concentration-dependent bactericidal activity and postantibiotic effect, and their ability to act synergistically with many other antimicrobials (Lacy et al., 1998). They have been classified as critically important antimicrobials in human medicine (World Health Organisation, 2007). By binding irreversibly to the highly-conserved aminoacyl (A-site) of the 16S ribosomal subunit they inhibit bacterial protein synthesis, thereby leading to cell death (Kotra et al., 2000). The most common mechanism for resistance to aminoglycosides results from production of aminoglycoside-modifying enzymes (phosphotransferases, nucleotidyltranferases and acetyltransferases), which compromise the binding of the aminoglycoside to the target site. Resistance may also result from reduced intracellular drug uptake and accumulation (likely to be due to changes in membrane permeability) or mutation of ribosomal proteins or rRNA (Mingeot-Leclercq et al., 1999). However, recent years have seen the emergence of several 16S rRNA methyltransferases in clinical isolates of Gram-negative bacteria in Europe, the Far East, and North and South America.

Aminoglycoside resistance mediated by 16S rRNA methylation
Aminoglycoside-producing actinomycetes are intrinsically resistant to very high levels (MIC >512 mg/L) of the aminoglycoside they produce due to rRNA methyltransferase enzymes, Please use Adobe Acrobat Reader to read this book chapter for free.
Just open this same document with Adobe Reader. If you do not have it, you can download it here. You can freely access the chapter at the Web Viewer here. but until recently no clinical isolate that was resistant to aminoglycosides as a result of an rRNA methyltransferase had been identified (Davies & Wright, 1997). The first reported 16S rRNA methyltransferase gene armA ( aminoglycoside resistance methyltransferase) was identified in a multiresistant Klebsiella pneumoniae isolated in France (Galimand et al., 2003). A further seven enzymes have been associated with 16S rRNA methylation: rmtA (Yokoyama et al., 2003), rmtB (Doi et al., 2004), rmtC (Wachino et al., 2006a), rmtD (now renamed as rmtD1) , rmtD2 (shares 96.4% amino acid identity with rmtD1) (Tijet et al., 2011), rmtE (Davis et al., 2010) and npmA (Wachino et al., 2007). The deduced amino acid sequences show homology with 16S rRNA methyltransferases produced by Actinomycetales, including Streptomyces and Micromonospora spp. (Fig. 1). However, the degree of identity shared overall between the 16S rRNA methyltransferase genes is less than 30%, which suggests that this is not a recent transfer event.

Clinical relevance of 16S rRNA methyltransferases
Unlike the aminoglycoside-modifying enzymes, which vary in their substrate specificity, 16S rRNA methyltransferases armA and rmtA-E confer high-level resistance to all clinicallyimportant aminoglycosides used for systemic therapy except streptomycin; npmA confers high-level resistance mainly to 4,5, disubstituted deoxystreptamines like apramycin. These genes are linked with mobile genetic elements and usually confirmed as located on large conjugative plasmids, allowing potential spread among bacterial populations. In addition, they have been commonly associated with other genes encoding resistance to clinically relevant antimicrobials such as ß-lactams (bla SHV , bla CTX-M , plasmid-mediated AmpC), carbapenems (bla SPM-1 , bla NDM-1 , bla KPC-2 ) and fluoroquinolones (plasmid-mediated qepA, aac(6')-Ib-cr and qnr family), thereby allowing potential co-selection and maintenance of

Phenotypic detection
armA and rmtA-E confer resistance to 4,6-disubstituted deoxystreptamines such as amikacin, kanamycin, tobramycin and gentamicin by methylation at position A1405 within the A-site of the 16S rRNA, whilst npmA confers panaminoglycoside resistance due to methylation at position A1408 (Liou et al., 2006;Wachino et al., 2007). 16S rRNA methyltransferase producers characteristically demonstrate aminoglycoside MICs 256 mg/L and are unique in expressing high-level resistance to the semisynthetic aminoglycoside arbekacin (Doi & Arakawa, 2007). Production of a 16S rRNA methyltransferase should be expected in enterobacteria exhibiting resistance to multiple aminoglycosides according to Clinical and Laboratory Standards Institute (CLSI) guidelines and demonstrating little or no inhibitory zones when using gentamicin, amikacin and arbekacin disks. Although arbekacin may not be readily available Doi & Arakawa (2007) noted that inclusion of an arbekacin disk was preferable as this increased the positive predictive value of this method to 90%, compared to only ca. 60% when using amikacin alone (Lee et al., 2006). Alternatively, using aminoglycoside MICs 256 mg/L as a screen has been reported to have an excellent positive predictive value (Lee et al., 2006;2007).

Molecular detection and characterization
With the exception of npmA, phenotypic resistance traits cannot be used to differentiate between the 16S rRNA methyltransferase variants. Instead PCR amplification and comparison of the amplicon sequence to previously reported gene sequences must be used for confirmation. The published literature contains many PCR protocols for detection of these genes (Doi & Arakawa, 2007;Fritsche et al., 2008;Granier et al., 2011).

16S rRNA methyltransferases in Salmonella enterica
Although all eight 16S rRNA methyltransferase genes have now been identified in enterobacteria; armA and rmtB are the most commonly identified and have spread worldwide. The first reported 16S rRNA methyltransferase gene in Salmonella enterica was armA in S. enterica serovar Enteritidis isolated in a hospital in Bulgaria (Galimand et al., 2005). Transfer of high-level aminoglycoside resistance to an Escherichia coli recipient led to transconjugants expressing resistance to 4,6-disubstituted deoxystreptamines due to armA, to ß-lactams due to acquisition of bla TEM-1 and bla CTX-M-3 , to streptomycin-spectinomycin because of ant"9, to sulphonamides because of sul1 and to trimethoprim because of dfrXII. All these resistance genes were located on a ca. 90-kb plasmid of incompatibility group IncL/M previously identified in Poland in Citrobacter freundii as pCTX-M3 (GenBank accession number AF550415; Gobiewski et al., 2007). Plasmid pCTX-M3 is responsible for the extensive spread of the extended-spectrum ß-lactamase (ESBL) bla CTX-M-3 in enterobacteria in Poland (including within S. enterica serovars Typhimurium, Enteritidis, Mbandaka and Oranienburg expressing very high-level resistance to aminoglycosides) Please use Adobe Acrobat Reader to read this book chapter for free.
Just open this same document with Adobe Reader. If you do not have it, you can download it here. You can freely access the chapter at the Web Viewer here. (Baraniak et al., 2002;Gierczynski et al., 2003a;Gierczynski et al., 2003b). The ant"9, sul1, dfrXII and armA genes were part of a 16.6-kb composite element flanked by two direct copies of IS6, which was designated Tn1548. Tn1548 has been identified on plasmids of other incompatibility groups and in enterobacteria of human and animal origin from several countries, thereby indicating its importance in the dissemination of armA (Galimand et al., 2005;González-Zorn et al., 2005a;González-Zorn et al., 2005b).
armA was also identified on a plasmid identical, or very similar to pCTX-M3 in a single isolate of S. enterica serovar Virchow among 1,078 non-typhoidal Salmonella isolates from patients hospitalized with gastroenteritis in Saint Petersburg, Russia between 2002(Egorova et al., 2007. The armA gene associated with Tn1548 was also identified in one isolate of S. enterica serovar Stanley submitted in 1999 among a collection of 18,261 nontyphoidal Salmonella isolated between 1996-2006 at the Centers for Disease Control and Prevention, USA (Folster et al., 2009).
In 2004 an outbreak of diarrhoea occurred in the neonatology unit of the Hospital of Constantine in Algeria. Stool cultures yielded S. enterica serovar Senftenberg, which were resistant to amoxicillin, ticarcillin, piperacillin, cefalothin, extended-spectrum cephalosporins, trimethoprim-sulfamethoxazole, amikacin, gentamicin, netilmicin, tobramycin and streptomycin (Naas et al., 2005). Whilst Salmonella infections in hospitals are usually food-associated, in this instance spread of infection was likely to be due to horizontal transmission as the commercially prepared milk was fed to other babies that did not become infected. A subsequent screening of 12 representative ESBL-producing serovar Senftenberg isolates isolated from the ward between 1982-2005 that expressed resistance to aminoglycosides was performed (Naas et al., 2009). Only one isolate from 1998 that expressed resistance to all clinically-relevant aminoglycosides was found positive for armA; all other isolates were negative for 16S rRNA methyltransferase genes and showed variable levels of resistance to aminoglycosides. As before, armA was located within Tn1548 on a bla CTX-M-3 -encoding IncL/M plasmid.
Another outbreak occurred on the neonatology ward of the same hospital between September 2008 to January 2009 (Naas et al., 2011). S. enterica serovar Infantis isolates were obtained from 138 patients (mostly from stool cultures but two neonates had positive blood cultures and another had positive gastric fluid). All but two isolates were resistant to all ßlactams (except cephamycins and carbapenems), kanamycin, netilmicin, tobramycin, amikacin, gentamicin, rifampicin and trimethoprim/sulfamethoxazole. Molecular typing by pulsed-field gel electrophoresis (PFGE) of representative isolates indicated that they were genetically related. The presence of armA associated with Tn1548, together with bla CTX-M-15 and bla TEM-1 was confirmed, with the 16S rRNA methyltransferase and ß-lactamase determinants co-located on a 140-kb self-transferable IncL/M plasmid. This suggests that 12 years on, a plasmid similar to that identified in serovar Senftenberg was still present on the ward and may have transferred to serovar Infantis with the bla CTX-M-3 evolving to bla CTX-M-15 via a single amino acid substitution (Asp-240→Gly) that has been associated with increased activity to ceftazidime (Poirel et al., 2002). Other enterobacterial species harbouring both armA, and bla CTX-M-3 or bla CTX-M-15 were isolated from Algerian patients transferred to Belgium (Bogaerts et al., 2007), and have been identified in Klebsiella spp. in Taiwan and China (Ma et al., 2009;Zhang et al., 2008), indicating that the Infantis plasmid may also have been acquired from a different bacterial species. Implementation and strengthening of Please use Adobe Acrobat Reader to read this book chapter for free.
Just open this same document with Adobe Reader. If you do not have it, you can download it here. You can freely access the chapter at the Web Viewer here.
hygiene and infection control measures brought the outbreak under control but was insufficient to prevent further sporadic cases in 2009. These may have occurred after rehospitalisation of colonised patients or from re-introduction of the strain onto the ward by colonised members of staff.
The continuing spread of pCTX-M3-like plasmids in the Salmonella population was highlighted in the recent identification of serovar Gambia strains harbouring a ca. 80-kb IncL/M plasmid bearing armA and CTX-M-3 together with sulphonamide and trimethoprim resistance determinants in France (Moissenet et al., 2011). The strains isolated from two babies in the intensive care unit of a Paris hospital in 2005 shared identical resistance profiles (resistance to cefotaxime, ceftazidime, gentamicin, tobramycin, amikacin and cotrimoxazole) and indistinguishable PFGE patterns. Given the rarity of serovar Gambia it was concluded that the second baby acquired the strain via cross-infection. Both babies had underlying disease and whilst the first baby was successfully treated with imipenem and ciprofloxacin, the second baby developed further complications and died despite broadspectrum therapy with imipenem, gentamicin, metronidazole and vancomycin. This was the first report of a fatality associated with non-typhoidal Salmonella harbouring a 16S rRNA methyltransferase gene.

Enhanced surveillance for 16S rRNA methyltransferases in Salmonella enterica
Between March 2006 and December 2009 a retrospective screen to identify 16S rRNA methyltransferase-producing enterobacteria among the strain collections of veterinary, medical and food science research institutes in ten European Union (EU) countries was performed under the auspices of the European-funded MED-VET-NET Network of Excellence. Isolates were selected for further study from nearly one million bacterial isolates from all sources along the food chain, including the environment, food-production animals, food products and humans based on expression of high-level resistance (MIC >256 mg/L) to amikacin or gentamicin. PCR using a harmonized protocol was used to screen selected isolates for the presence of armA.
Among 46 S. enterica isolates selected from the culture collection of the Health Protection Agency (HPA), UK, five serovar Oranienburg isolates received in 2002 from Poland were positive for armA (Hopkins et al., 2007). A ca. 90-kb IncL/M conjugative plasmid, which harboured both armA and bla CTX-M-3 was isolated, with armA borne by Tn1548 as previously described. Transposon mapping using PCR producing overlapping amplicons along the length of the transposon indicated that the transposon structure was identical to that of pCTX-M3 (Fig. 2).
Please use Adobe Acrobat Reader to read this book chapter for free.
Just open this same document with Adobe Reader. If you do not have it, you can download it here. You can freely access the chapter at the Web Viewer here. S. enterica serovar Oranienburg harbouring bla CTX-M-3 and expressing resistance to aminoglycosides has previously been isolated in Poland, where armA is estimated to be present in ca. 1.6% of human clinical isolates of enterobacteria (Gierczynski et al., 2003a;Gierczynski et al., 2007). These five isolates were part of an outbreak of enteritis associated with an orphanage for infants and children below two years of age in central Poland. The orphanage was reported to provide poor living conditions (due to a high concentration of children) and sanitary standards (i.e. overuse of antimicrobials). Infected children were hospitalised in three separate hospitals in central Poland (Table 1). In total, 52 related isolates were obtained as reported by a local sanitary and epidemiological unit. Until recently this was the only report of involvement of an armA-positive Salmonella strain in a hospital outbreak (Naas et al., 2009 (Hopkins et al., 2010). All isolates belonged to serovar Virchow of phage types 25 (n=6), 30 (n=5) or 31 (n=1). PCR was negative for armA; however screening for rmtA, rmtB, rmtC and rmtD identified rmtC (Doi & Arakawa, 2007). RmtC has previously been identified in Proteus mirabilis strain ARS68 from an inpatient in Japan (Wachino et al., 2006a) and P. mirabilis strain JIE273 from a patient recently returned from India to Australia (Zong et al., 2008), therefore this was the first report of rmtC in Europe. Twelve of the 13 strains were isolated from patients between 2005-2008, of which seven had reported recent travel abroad (four to India). S. enterica serovar Virchow expressing rmtC has also been reported in the United States in a child with a history of travel to India (Folster et al., 2009), suggesting that 16S rRNA methyltransferases may be relatively common in India. The remaining isolate was obtained from frozen produce. This was the first report of a 16S rRNA methyltransferase gene being identified in a bacterial strain isolated from food. Interestingly, a survey in Northern India identified S. enterica in 3.6% of vegetable samples; however >30% of isolates were resistant to amikacin (Singh et al., 2007). PFGE revealed the 13 isolates were highly related, showing only one to two band differences; the isolate from frozen produce was indistinguishable from those isolated from patients. Attempts to isolate rmtC by conjugation and electroporation experiments were unsuccessful and the gene was eventually discovered located on the chromosome in association with an ISEcp1-like element, which has previously been shown to play a role in the expression and transposition of rmtC (Wachino et al., 2006b). The rmtC gene was localised on a ca. 100-kb non-conjugative plasmid in P. mirabilis ARS68, but attempts to transfer rmtC from P. mirabilis JIE273 to a donor were unsuccessful. This contrasts with armA and rmtB, which are commonly localised on plasmids (Doi & Arakawa, 2007) and may explain at least in part the higher prevalence and increased dissemination of these genes in comparison to rmtC. rmtC-positive strains of S. enterica serovar Virchow belonging to the same phage types are still being received at the HPA from patients reporting recent return from India, suggesting persistence of this multidrugresistant clone (Hopkins, unpublished observation).
Further evidence for food being a possible vehicle of infection for bacterial strains harbouring 16S rRNA methyltransferase genes was provided by a recent study reporting the identification of armA in a S. enterica serovar 4,12:i:-isolate recovered from chicken meat in La Réunion, a French island in the Indian Ocean . This was the first report of armA in a bacterial isolate originating from food. As well as armA, the isolate harboured bla CTX-M-3 , bla TEM , the Salmonella Genomic Island-1 (as previously associated with multidrug-resistant Salmonella enterica serovar Typhimurium definitive phage type (DT) 104 (Boyd et al., 2000)), and acquired AmpC ß-lactamase bla CMY-2 . Conjugation experiments indicated that armA, bla CTX-M-3 and bla TEM were located on the same plasmid. However, further characterisation revealed that, in contrast to previous studies, these genes were located on a ca. 110-kb incompatibility group IncP plasmid. IncP plasmids have a broad host range, including Pseudomonas spp. and Gram-positive bacteria, therefore association with an IncP plasmid may further broaden the dissemination of armA. Transposon mapping using PCR also suggested a deletion event downstream of armA and insertion of an IS26 element, which knocked out genes encoding a macrolide resistance efflux pump and macrolide phosphotransferase gene (Fig. 2 and Fig. 3). Investigation of other S. enterica serovar 4,12:i:isolates recovered from broiler chickens failed to identify any expressing high-level resistance to aminoglycosides, therefore it is likely that this strain was present in the chicken meat sample as a consequence of cross-contamination by a food-handler.
Please use Adobe Acrobat Reader to read this book chapter for free.
Just open this same document with Adobe Reader. If you do not have it, you can download it here. You can freely access the chapter at the Web Viewer here.

Conclusion
The worldwide prevalence of S. enterica harbouring 16S rRNA methyltransferase genes is very low; however within the EU, Poland appears to represent a major reservoir of these genes. The majority of S. enterica isolates have been from nosocomial diarrhoeal outbreaks, with the occurrence of these genes in aminoglycoside-resistant strains from community patients relatively rare. Interestingly, veterinary use of aminoglycosides in food-production animals does not appear to be involved greatly in the spread of 16S rRNA methyltransferase genes as had previously been hypothesized in the literature (González-Zorn et al., 2005a). Food products have recently been identified as a source of S. enterica harbouring 16S rRNA methyltransferase genes (Hopkins et al., 2010;Granier et al., 2011), though whether this is an indication of an animal source of these genes, or cross contamination by food handlers has yet to be elucidated. Further evolution and dissemination of these genes may be limited to environments such as hospitals, and countries in which antimicrobial usage is unregulated where a strong selective pressure results from high-level use of a diverse range of aminoglycosides. This is compounded by their common association with other resistance genes, leading to potential co-selection and maintenance of resistance by use of extendedspectrum ß-lactams, carbapenems and fluoroquinolones (Cantón, 2009). Spread of 16S rRNA methyltransferase genes in association with these resistance genes would seriously compromise use of aminoglycosides for treatment of life-threatening infections caused by Gram-negative bacteria such as S. enterica and raises the possibility of untreatable Salmonella infections. Ongoing surveillance of aminoglycoside resistance in S. enterica isolated from humans, animals and food products is therefore crucial to delay the spread of resistance to these classes of antimicrobial agents.
Please use Adobe Acrobat Reader to read this book chapter for free.
Just open this same document with Adobe Reader. If you do not have it, you can download it here. You can freely access the chapter at the Web Viewer here.