Update on macrolide–lincosamide–streptogramin, ketolide, and oxazolidinone resistance genes
Correspondence: Marilyn C. Roberts, Department Environmental & Occupational Health Sciences, Box 357234, School of Public Health and Community Medicine, University of Washington, Seattle WA 98195, USA. Tel.: +1 206 543 8001; fax: +1 206 543 3873;
This Minireview summarizes the changes in the field of bacterial resistance to macrolide, lincosamide, streptogramin, ketolide, and oxazolidinone (MLSKO) antibiotics since the nomenclature review in 1999. A total of 66 genes conferring resistance to this group of antibiotics has now been identified and includes 13 new rRNA methylase genes, four ATP-binding transporter genes coding for efflux proteins, and five new inactivating enzymes. During this same time period, 73 new genera carrying known rRNA methylase genes and 87 new genera carrying known efflux and/or inactivating genes have been recognized. The number of bacteria with mutations in the genes for 23S rRNA, L4 and L22 ribosomal proteins, resulting in reduced susceptibility to some members of the group of MLSKO antibiotics has also increased and now includes nine different Gram-positive and 10 different Gram-negative genera. New conjugative transposons carrying different MLSKO genes along with an increased number of antibiotics and/or heavy metal resistance genes have been identified. These mobile elements may play a role in the continued spread of the MLSKO resistance genes into new species, genera, and ecosystems.
The first macrolide, erythromycin, was discovered in 1952 and since then macrolides have had an important role in treating infectious diseases (Kirst, 2002). Erythromycin has a moderate spectrum of activity which has been enhanced with the production of the newer semi-synthetic derivatives such as azithromycin, clarithromycin, and more recently, the ketolides (Kirst, 2002). Macrolides are used to treat acute upper and lower respiratory tract infections due to a variety of Gram-positive bacteria, Mycoplasma pneumoniae, Chlamydia pneumoniae, Legionella pneumophila, Haemophilus influenzae, Moraxella catarrhalis, skin and soft tissue infections, sexually transmitted diseases caused by Chlamydia trachomatis, Treponema pallidum or Ureaplasma urealyticum, chronic pulmonary infections in cystic fibrosis patients, and diseases caused by Bordetella pertusis, Borrelia burgdorferi, Bartonella henselae, Campylobacter spp., Listeria monocytogenes, Rickettsia spp., Rhodococcus equi, as well as Mycobacterium avium-intracellulare and other atypical mycobacterium. Macrolides have also been used to treat diseases due to single cell eukaryotes such as Babesia microti, Cryptosporidium parvum, Entamoeba histolytica, Pneumocystis carinii, Plasmodium species, and Toxoplasmagondii (Iacoviello & Zinner, 2002). Macrolides influence bacterial virulence factors and cause reduction of adherence by Moraxella catarrhalis to ciliated epithelial cells, decreased adherence of Neisseria meningitidis, decreased binding of fibronectin to Staphylococcus aureus, and a variety of effects on Pseudomonas aeruginosa infecting cystic fibrosis patients. Macrolides also have anti-inflammatory characteristics which provide improvement in severe steroid-dependent asthma, diffuse pan-bronchiolitis, atheroma, cancer, and arthritis (Iacoviello & Zinner, 2002). A detailed review on macrolides as potent immunomodulators is now available (Giamarellos-Bourboulis, 2008).
Macrolides, lincosamides, streptogramins, ketolides (semi-synthetic derivatives of erythromycin A), and oxazolidinones (MLSKO), though chemically distinct, are usually considered together (Vester & Douthwaite, 2001; Zhanel et al., 2001; Sutcliffe & Leclercq, 2003). The MLSK antibiotics share overlapping binding sites on the 50S subunit of the ribosome, while linezolid binds to the 50S ribosomal subunit but is not affected by methylation of the erm genes (Franceschi et al., 2004; Poehlsgaard & Douthwaite, 2005; Tu et al., 2005). MLSKO antibiotics inhibit protein synthesis by binding to the 50S ribosomal subunit and blocking peptide bond formation and/or translation. More details on the interaction between these antibiotics and the ribosome can be found in a number of recent papers some of which are referenced in this review (Vester & Douthwaite, 2001; Lui & Douthwaite, 2002; Douthwaite et al., 2005; Poehlsgaard & Douthwaite, 2005; Tu et al., 2005). Shortly after the introduction of erythromycin in the 1950s into clinical practice, bacterial resistance to erythromycin was reported in Staphylococci (Zhanel et al., 2001). Since then a large number of bacteria have been identified that are resistant to macrolides due to the presence of a number of different genes which included 13 identical or nearly identical genes (98–100% amino acid identity) which coded for the erm(B) rRNA methylase gene but had different names because there was no standardized nomenclature or criteria used (Roberts et al., 1999). Much of this information including individual GenBank numbers, DNA, and amino acid homology within the gene class and association with particular plasmids or transposons can be found at http://faculty.washington.edu/marilynr/. To remedy this situation, colleagues and I wrote a review standardizing the nomenclature for genes which conferred resistance to macrolide, lincosamide, and streptogramin antibiotics which required that the new gene must be functional and have <80% identity at the amino acid level with all previously characterized MLSKO genes (Roberts et al., 1999; http://faculty.washington.edu/marilynr/). In addition, a nomenclature center was established that provides a central place to assign names to new genes and a website was developed for the dissemination of information on resistance genes which encode for rRNA methylases, efflux and inactivating enzymes that confer resistance to MLSKO antibiotics. The site is modified twice a year or as needed, to reflect the ongoing changes in the field (http://faculty.washington.edu/marilynr/). The aim of this Minireview will be to focus on information available after the 1999 review was published (Roberts et al., 1999). In the current review I have also included a section on mutations in the genes coding for 23S rRNA domain, L4 and L22 proteins since the majority of isolates with resistance to ketolide and/or oxazolidinones have been shown to carry mutations in one or more of these three genes. In addition, there is a section on new transposons and linkages between MLSKO resistance genes and other antibiotic and/or heavy metal resistance genes and their association with integrons or particular elements. The literature since 1999 is extensive and cannot be adequately referenced in this current Minireview. A more complete list of MLSKO references can be found with the following reviews and papers (Roberts, 1997, 2007; Vester & Douthwaite, 2001; Zhanel et al., 2001; Wang et al., 2003; Webber & Piddock, 2003; Sutcliffe & Leclercq, 2003; Roberts & Sutcliffe, 2005) and with updates at http://faculty.washington.edu/marilynr/.
New resistance genes
There have been 22 new MLSKO-resistance genes identified since the 1999 publication which include 13 new genes coding for rRNA methylases, four genes coding for efflux proteins, four found coding for transferases, and one coding for a phosphorylase (Table 1).
Table 1. Mechanisms of MLS resistance genes
|Listed in 1999*||n=20|
erm(A), erm(B), erm(C), erm (D), erm(E)
erm(F), erm(G), erm(H), erm(I), erm(N)
erm(O), erm(Q), erm(R), erm(S), erm(T)
erm(U), erm(V), erm(W), erm (X), erm(Y)
srm(B), tlr(C), vga(A)†, vga(B)
|Not listed 1999||n=13|
erm(Z), erm(30), erm(31), erm(32), erm(33)
erm(34), erm(35), erm(36), erm(37), erm(38)
erm(39), erm(40), erm(41)
Currently there are 66 MLSKO resistance genes and 33 are erm genes that code for rRNA methylases which add one or two methyl groups to a single adenine (A2058 in Escherichia coli) in the 23S rRNA moiety and generally conferred resistance to macrolides, lincosamides, and streptogramin B antibiotics (MLSB) (Table 1) (Poehlsgaard & Douthwaite, 2005). Out of the 33 erm genes, 13 are new including four which have been renamed; erm(Z) formerly pikR1, erm(31) formerly pik2, erm(32) formerly tlr(B), and erm(30) all from Streptomyces, and five innate methylase genes [erm(37), erm(38), erm(39), erm(40), and erm(41)] identified from different species of Mycobacterium which confer resistance to macrolides and lincosamides but not to streptogramin B (Nash et al., 2006). The erm(33) gene, from Staphylococcus has 62% DNA and 58% amino acid homology to both an erm(A) and an erm(C) gene, respectively. The DNA sequence of erm(33) indicates that it was likely created by a recombination event between these two genes (Schwarz et al., 2002). The remaining three genes, erm(34) from Bacillus, erm(35) from Bacteriodes, and erm(36) from Micrococcus may be found in other genera but currently have not been used extensively in screening resistant bacteria (Tables 1 and 2).
Table 2. Distribution of MLS genes in new bacterial genera
| Confers resistance to macrolides, lincosamides, streptogramin B|
| erm(A)||7||4|| ||Bacteriodes, Helcococcus, Peptostreptococcus, Prevotella|
| erm(B)||33||23|| ||Acinetobacter, Aerococcus, Arcanobacterium, Bacillus, Bacteriodes, Citrobacter, Corynebacterium, Enterobacter, Eubacterium, Fusobacterium, Gemella, Haemophilus Lactobacillus, Micrococcus, Pantoeae Peptostreptococcus, Porphyromonas, Proteus, Pseudomonas, Ruminococcus, Rothia, Serratia, Treponema|
| erm(C)||16||8|| ||Actinomyces, Bacteriodes, Corynebacterium, Enterococcus, Haemophilus, Micrococcus, Prevotella, Peptostreptococcus|
| erm(D)||2||1|| ||Salmonella|
| erm(E)||6||5|| ||Bacteroides, Eubacterium, Fusobacterium, Ruminococcus, Shigella|
| erm(F)||24||8|| ||Corynebacterium, Enterococcus, Lactobacillus, Mobiluncus, Peptostreptococcus, Ruminococcus, Shigella, Staphylococcus|
| erm(G)||7||5|| ||Catenibacterium, Lactobacillus, Prevotella Porphyromonas, Staphylococcus|
| erm(H)*||1||0|| || |
| erm(I)*||1||0|| || |
| erm(N)*||1||0|| || |
| erm(O)*||1||0|| || |
| erm(Q)||6||2||L42817||Bacteroides, Staphylococcus|
| erm(R)*||1||0|| || |
| erm(S)*||1||0|| || |
| erm(U)*||1||0|| || |
| erm(V)||3||2|| ||Eubacterium, Fusobacterium|
| erm(W)*||1||0|| || |
| erm(X)||4||3||NC_005206||Arcanobacterium, Bifidobacterium Propionibacterium|
| erm(Y)*||1||0|| || |
| Confers resistance to lincomycin, erythromycin or erythromycin + streptogramin B, olenadomycin, spiramycin, tylosin streptogramin A, or lincomycin + streptogramin A|
| ATP-binding Transporters|
| car(A)*||1||0|| || |
| msr(A)||7||6|| ||Corynebacterium, Enterobacter, Enterococcus, Gemella, Pseudomonas, Streptococcus|
| msr(C)||1||1||AY004350, AF313494AJ243209||Enterococcus|
| msr(D)¶||20||20||AF227521, SA318993, AF274302||Acinetobacter, Bacteroides§, Citrobacter, Corynebacterium, Enterobacter, Enterococcus, Escherichia, Fusobacterium, Gemella, Klebsiella, Morganella, Neisseria, Proteus, Providencia, Pseudomonas, Ralstonia, Serratia, Staphylococcus, Streptococcus, Stenotrophomonas,|
| ole(A)*||1||0|| || |
| ole(B)*||1||0|| || |
| ole(C)*||1||0|| || |
| srm(B)*||1||0|| || |
| tlr(C)||1||0|| || |
| vga(A)||1||0|| || |
| vga(B)||2||1|| ||Enterococcus|
| Major facilitators|| || || || |
| mef(A)||24||19|| ||Acinetobacter, Bacteroides, Citrobacter, Enterobacter, Escherichia, Fusobacterium, Gemella, Klebsiella, Lactobacillus, Morganella, Neisseria, Pantoeae, Providencia, Proteus, Ralstonia, Pseudomonas, Salmonella, Serratia, Stenotrophomona|
| Confers resistance (usually high level) to erythromycin|
| ere(A)||11||7|| ||Pantoeae, Providencia, Pseudomonas, Serratia, Staphylococcus, Stenotrophomonas, Vibrio|
| ere(B)||8||5|| ||Acinetobacter, Citrobacter, Enterobacter, Pseudomonas, Staphylococcus|
| Confers resistance to streptogramin B, currently only in Gram-positive genera|
| vgb(A)||2||0|| || |
| vgb(B)||1||0|| || |
| Confers resistance to streptogramin A|
| lnu(A)||2||1|| ||Clostridium|
| lnu(B)||3||2||AJ238249||Staphylococcus, Clostridium|
| lnu(F)||2||2||AJ561197||Escherichia, Salmonella|
| vat(A)||1||0|| || |
| vat(B)||2||1|| ||Enterococcus|
| vat(C)||1||0|| || |
| vat(D)||1||0|| || |
| vat(E)||2||1||AJ488494, NC_004566||Lactobacillus|
| Confers resistance to macrolides|
| mph(A)||10||9|| ||Aeromonas, Citrobacter, Enterobacter, Klebsiella, Pantoeae, Pseudomonas, Proteus, Serratia, Stenotrophomonas|
| mph(B)||4||3|| ||Enterobacter, Pseudomonas, Proteus|
| mph(C)||2||1|| ||Stenotrophomonas|
| mph(D)||6||6||AB048591 only partially sequenced||Escherichia, Klebsiella, Pantoeae, Proteus, Pseudomonas, Stenotrophomonas|
There are now 14 different genes that code for either ATP-transporters or Major Facilitator Transporters and the genes produce proteins that pump one or more of the MLSKO antibiotics out of the cell (Table 1). Of the four new genes identified since 1999, one is an innate enterococcal gene [msr(C)], while the msr(D) gene is always found linked to the newly described efflux gene mef(A) and two genes, lsa(A) and lsa(B), coding for lincosamide efflux proteins (Tables 1 and 2) (Singh et al., 2002; Kehrenberg et al., 2004; Reynolds & Cove, 2005; Roberts & Sutcliffe, 2005). In addition, a variant vga(A)lc of the previously described vga(A) gene is listed in Tables 1 and 2 because unlike the vga(A) gene which confers resistance to streptogramin A, the vga(A)lc variant confers resistance to streptogramin A and lincomycin and thus merits a variant status (Novotna & Janata, 2006).
Currently there are 19 inactivating enzymes including two esterases, two lyases, 11 transferases, and four phosphorylases. Of these 19, four genes coding for transferases and one gene coding for a phosphorylase were not in the 1999 review (Table 1) (Roberts et al., 1999). Unfortunately the mph(D) phosphorylase gene had not been fully sequenced.
Another 23S rRNA methyltransferase, encoded by cfr gene, and first identified on staphylococcal plasmids from animals should also be mentioned (Long et al., 2006). This gene conferred resistance to lincosamides, oxazolidinones, streptogramin A, phenicols and pleuromutilins, but not macrolides, and thus differs from erm rRNA methylase genes and has not been included in Tables 1 and 2.
Over the past 10 years, an increasing number of isolates that are resistant to MLSKO antibiotics have been identified which contain mutations in the V domain of the 23S rRNA genes, and/or the genes coding the ribosomal proteins L4 and L22 (Franceschi et al., 2004). The majority of telithromycin (ketolide) and/or linezolid (oxazolidinone) resistant bacteria carry mutations in one of these three genes (Garza-Ramos et al., 2001; Jones et al., 2002; Farrell et al., 2004; Douthwaite et al., 2005; Roberts & Sutcliffe, 2005; Toh et al., 2007). These mutational changes have been described in both Gram-positive and Gram-negative bacteria (Table 3) and alter the function of the 23S rRNA and/or proteins resulting in moderately decreased susceptibility to one or more of the MLSKO antibiotics (Garza-Ramos et al., 2001; Vester & Douthwaite, 2001; Jones et al., 2002; Ng et al., 2002; Sutcliffe & Leclercq, 2003; Farrell et al., 2004; Lukehart et al., 2004; Douthwaite et al., 2005; Morozumi et al., 2005; Tu et al., 2005; Luthje & Schwarz, 2006; Binet & Maurelli, 2007; Florez et al., 2007; Hong et al., 2007; Mayrhofer et al., 2007). These changes are normally passed to daughter cells during replication and generally not passed between strains or between different genera. In vitro selection experiments often obtained the same mutants as found in clinical isolates, while clinical strains carrying both mutations and one or more acquired MLSKO resistance genes has also been identified. Isolates with both mutations and acquired genes are often labeled as having acquired MLSKO resistance genes but not mutations which may lead to an under estimation of the number of resistant mutations in some studies.
Table 3. Mutations in 23S rRNA, L4, and/or L4 gene which increase MLSKO resistance
|Gram-positive and related genera|
| Arcanobacterium pyogenes||Yes|| || |
| Bacillus subtilis|| ||Yes||Yes|
| Enterococcus faecalis||Yes*||Yes||Yes|
| Lactobacillus rhamonosus||Yes|| || |
| Mycobacterium spp.†||Yes|| || |
| Mycoplasma spp.‡||Yes|| || |
| Propionibacterium species§||Yes|| || |
| Staphylococcus aureus||Yes|| ||Yes|
| Streptococcus pneumoniae||Yes||Yes¶||Yes|
| Streptococcus pyogenes||Yes||Yes||Yes|
| Turicella||Yes|| || |
| Brachyspira hyodysenteriae||Yes|| || |
| Bordetella pertusis||Yes|| || |
| Campylobacter jejuni||Yes||Yes∥||Yes∥|
| Campylobacter coli||Yes||Yes∥||Yes∥|
| Chlamydia trachomatis||Yes||Yes**|| |
| Escherichia coli||Yes||Yes||Yes|
| Haemophilus influenzae||Yes||Yes||Yes|
| Helicobacter pylori||Yes|| || |
| Neisseria gonorrhoeae||Yes|| || |
| Rickettsia spp.|| || ||Yes|
| Treponema palladium||Yes|| || |
| Treponema denticola||Yes|| || |
Mutations 23S rRNA
Mutations in the 23S rRNA can result in increased resistance to macrolides, lincosamides, streptogramin B, telithromycin and/or linezolid in both Gram-positive and Gram-negative bacteria (Sutcliffe & Leclercq, 2003; Poehlsgaard & Douthwaite, 2005). Eight Gram-positive and nine Gram-negative genera, including the intracellular pathogen C. trachomatis and two Treponema species, have been identified with 23S rRNA mutations (Table 3). In addition, laboratory derived mutants have been selected in other species and genera (Pereyre et al., 2007). Various mutations have been identified but the most common are in domain V with mutations at positions A2058 or A2059 the most frequently identified (E. coli numbering) (Vester & Douthwaite, 2001). Mutations in other positions within the domain V of the 23S rRNA have been associated with increased linezolid resistance. The first bacteria identified with mutations in their 23S rRNA had one or two copies of these genes such as Mycobacterium or Helicobacter. However, soon it was clear that other bacteria carrying more than two gene copies were also acquiring mutations and there was a correlation between the number of rrl genes that carried mutations and level of resistance to MLSKO antibiotics. Detailed reviews on specific mutations have been published (Vester & Douthwaite, 2001; Sutcliffe & Leclercq, 2003; Franceschi et al., 2004; Roberts & Sutcliffe, 2005). It is very likely that more species and genera carry mutations in the 23S rRNA than given in Table 3.
Mutations in ribosomal proteins L4 and/or L22
There have been a number of mutations identified in the L4 and L22 ribosomal proteins and include single amino acid changes, insertion and/or deletion which add or delete one or more amino acids to these proteins (Table 3). Three Gram-positive and five Gram-negative genera with mutations in the L4 gene and four Gram-positive and five Gram-negative genera with mutations in the L22 gene have been described. Details of various mutations that confer macrolide resistance in Streptococcus pneumoniae, Streptococcus pyogenes, and H. influenzae can be found in Franceschi et al. (2004). It is very likely that Table 3 underestimates the number of species and genera which carry mutations in the L4 and/or L22.
Mutations in intrinsic genes
Virtually all bacteria have multiple efflux pumps which import and export a variety of different compounds. As the understanding of intrinsic efflux pumps has increased it is clear that some of these pumps have roles in mediating low-moderate levels of antibiotic resistances in Gram-positive and Gram-negative bacteria (Webber & Piddock, 2003; Li & Nikaido, 2004). These efflux pumps use the proton motive force or ATP as energy sources for transporting multiple antibiotics, dyes, detergents, and disinfectants across the cell membrane. Exposure to any number of chemicals may lead to mutational changes that result in increased resistance to a variety of antibiotics. One example is triclosan, this antibacterial compound can be found in soap, toothpaste, and other products and can select for bacteria which have mutations in their efflux pumps and/or in genes that regulate expression of the genes coding for the efflux proteins. The mtrR gene encodes for a regulator protein that regulates transcription of the efflux proteins. Missense, deletions, and/or insertions within the mtrR gene, of Neisseria gonorrhoeae, increases resistance to multiple antibiotic resistant classes including macrolides (Zarantonelli et al., 2001). Upstream of the mtrR a one base pair deletion of an adenine results in the loss of expression of the mtrR and leads to a four-fold increase in resistance to erythromycin, penicillin, and tetracycline (Zarantonelli et al., 2001; Cousin et al., 2003). This resistance is clinically relevant and can lead to treatment failures. In the 2005 surveillance, coordinated by the Centers for Disease Control, 5.7% of the isolates had acquired antibiotic-resistance genes while 13.9% had chromosomal mediated mutational resistance in the United States (http://www.cdc.gov/std/Gisp2005/GISPSurvSupp2005short.pdf). In contrast, acquired resistance is much more prevalent in Asia (Tapsall, 2005). Thus illustrating that on occasion, mutational resistance may be more common for some bacteria in particular geographical areas than acquired resistance.
Spread of acquired MLSKO genes and/or mutations
Currently 66 genes are listed in Table 1, 22 of which have been described in the past 9 years (Table 1). During this same time period, 73 new genera carrying known rRNA methylase genes and 87 new genera carrying known efflux and/or inactivating genes have been identified (Table 2). Twenty-three (35%) of the genes are now in multiple genera (2–33 genera) with 18 of the genes found in both Gram-positive and Gram-negative bacteria (Table 2). It is not clear why only some of the genes are found in multiple species and others are not. For example, the Mycobacterium and Streptomycetes erm genes have a high >70% G+C content, while the rest of the erm genes <40% G+C content. It is possible that some of the Streptomycetes genes are in Mycobacterium and/or Norcardia spp. similar to the tetracycline resistance Streptomycetes otr(A), otr(B), otr(C) genes or vice versa, however, it is unlikely that the high % G+C erm genes will be found in other genera though the experiments have not been done (Chopra & Roberts, 2001). In other cases the genes have not been used in surveillance studies and may be found in more species than listed in Table 2. These genes are marked with an asterisk in Table 2.
The rRNA methylases are the largest group of acquired MLSKO genes and include innate genes found in Streptomycetes, where they provide a self-protective mechanism for these macrolide producers, and innate genes in naturally macrolide resistant Mycobacterium species (Nash et al., 2006). Each species of Mycobacterium carries a unique erm gene and together with the Streptomycetes genes have a >70% G+C content, while the rest of the erm genes <40% G+C content (Chopra & Roberts, 2001).
Ten erm genes have been identified in multiple genera with the erm(B) gene found in the widest host range, the most genera and in both Gram-positive and Gram-negative bacteria, aerobic, and anaerobic genera and in most ecosystems that have been examined (Table 2). This may be due to erm(B) genes association with mobile elements and linkage to a number of different genes (Table 4). The erm(B) gene was first identified in Enterococcus spp. from the 1950s and from S. pneumoniae in 1967 and has been identified in 23 new genera over the last 8 years (Roberts & Sutcliffe, 2005; http://faculty.washington.edu/marilynr/). The erm(B) gene is found in 33 different genera (Table 2) and has the widest host range of any of the acquired resistance genes listed in Table 1. The erm(B) gene is associated with conjugative transposons located in chromosomes as well as on plasmids. The erm(B) gene has also been found on nonconjugative transposons such as Tn917 and Tn551. The next most common gene is erm(F) found in a total of 24 different genera many of which are anaerobes, including eight new genera. Interestingly, both the erm(B) and erm(F) genes have been identified in a 1955 N. gonorrhoeae isolate (Cousin et al., 2003). The erm(C) was identified in eight new genera and in a total of 16 different genera, the erm(G) gene in seven genera including five new genera, the erm(A) gene has been found in seven genera of which four are new, the erm(E) gene has been found in six genera including five new genera, the erm(X) gene has been found in four genera including three new genera, the erm(Q) gene in six genera including two new genera, the erm(V) gene in three genera including two new genera, while the erm(T) gene is in two genera with one new (Table 2).
Table 4. MLS antibiotic resistance genes linked to other genes or elements
|cfr||Lincosamide, linezolid, Streptogramin A|
|mef(A), Tn2009, Tn2010||Macrolide|
|rteA, rteB||Regulatory genes|
| erm(33)||cfr||Lincosamide, linezolid, Streptogramin A|
|Tn2009, Tn2010, Tn1207.1, Tn1207.3, MEGA†||Transposons|
|cfr||Lincosamide, linezolid, Streptogramin A|
| ||Chloramphenicol, florfenicol|
|Tn2009, Tn2010, Tn1207.1, Tn1207.3, MEGA†||Transposons|
| ere(A)||Class 1 integron‡|| |
|Class 2 integron‡|| |
| vgb(B)||vat(C)||Streptogramin A|
| vat(A)||vgb(A)||Streptogramin B|
| vat(B)||vga(B)||Streptogramin A|
| vat(C)||vgb(B)||Streptogramin B|
| vat(D)||vgb(A)||Streptogramin B|
| lnu(F)||Class 1 integron‡|| |
| mph(A)||mrx||Hydrophobic protein Unknown function|
|IS26, IS6100||Insertion sequences|
|Class 1 integron‡|| |
| mph(C)||cadD/cadA||Cadmium efflux|
In the last 8 years, telithromycin resistant S. pyogenes, S. aureus and S. epidemidis have been reported (Farrell et al., 2004). Each of these strains carried a constitutively expressed erm(A) or erm(B) gene. Douthwaite et al. (2005) found that there was variation in how well the erm(A) and erm(B) enzymes dimethylated the rRNA molecules within the host bacteria. The enzymes which had the greatest proportion of the rRNA molecules modified conferred the highest levels of telithromycin resistance with a range of MICs from 4 to >64 μg mL−1, which explains the association between constitutively expressed erm(A) or erm(B) genes and ketolide resistance.
Among the efflux new genes the most prevalent was the msr(D) gene, which has been identified in 19 different genera, however, this gene always is downstream of the mef(A) gene and has been identified in 24 genera, of which 19 are new (Ojo et al., 2006a). This suggests that every strain which carries a mef(A) gene should also carry the msr(D) gene and the presence of the msr(D) gene should be checked in all mef(A) positive Gram-positive and Gram-negative bacteria (Table 2). Among previously described efflux genes six new genera including Gram-positive and Gram-negative have been identified carrying msr(A) (Ojo et al., 2006b).
The new lincosamide transferases genes, lnu(C) and lnu(D) have both been characterized from Streptococcus spp. while lnu(F) gene has been found in Escherichia and Salmonella (Achard et al., 2005; Petinaki et al., 2008). The innate vat(F) gene codes for a streptogramin A acetyltransferase found in Yersinia, while the other vat genes are found in Gram-positive bacteria and normally associated with plasmids (Seoane & Lobo, 2000). The new macrolide phosphorylase gene, mph(D) and is the only gene listed in Table 1 which has not been completely sequenced. The two erythromycin esterases, ere(A) and ere(B) genes, have been identified in seven Gram-positive and five new Gram-negative genera. Nine new Gram-negative genera carrying the macrolide phosphotransferases mph(A) gene, and three new Gram-negative genera carrying the mph(B) gene have been found (Table 2). The mph(C) gene has recently been identified in Gram-negative Stenotrophomonas and this is the only macrolide phosphotransferases found in both Gram-positive and Gram-negative bacteria. It would therefore be of interest to determine if the other mph genes are also found in Gram-negative genera. The transferase coded by the vat(E) gene has recently been found in Lactobacillus fermentum linked to an erm(T) gene and represents the first species other than Streptococcus spp. and Enterococcus spp. to carry a vat gene (Gfeller et al., 2003).
The first bacterial isolate with a cfr gene was a bovine Staphylococcus sciuri isolated from a florfenicol surveillance study of animals (Schwarz et al., 2000). More recently a methicillin and linezolid resistant S. aureus was isolated in 2005 from Columbia, South America from a patient. In this isolate the cfr gene was located downstream of an erm(B) gene and the two genes were co-transcribed from the erm(B) promoter (Toh et al., 2007). This gene cluster was found in the chromosome flanked by an IS21-558 element, which has been implicated in the mobility of the cfr gene (Kehrenberg et al., 2007). This S. aureus is also the first human linezolid resistant isolate characterized which is resistant due to the acquisition of an acquired resistance gene rather than to mutations. The identification of this human isolate suggests that more human strains with the cfr gene should be anticipated in the future.
The current mutations to the 23S rRNA ribosomal proteins L4 and L22 which confer increased resistance to MLSKO antibiotics are listed in Table 3. Both Gram-positive and Gram-negative species have been identified with mutations, however, only some of the bacteria listed have mutations in all three genes. This is most likely due to lack of available data rather than the actual lack of mutations in the bacteria listed. Some of these mutations also confer resistance to ketolides and oxazolidinone (Garza-Ramos et al., 2001; Jones et al., 2002; Liu & Douthwaite, 2002; Franceschi et al., 2004; Farrell et al., 2004; Douthwaite et al., 2005; Hong et al., 2007).
Mobile elements and new linkages
Lateral gene exchange was demonstrated c. 50 years ago and it is hypothesized that gene exchange is the primary way most bacteria acquire and transfer antibiotic resistance genes between species, genera, and ecosystems (Watanabe, 1963). Mobile elements found in bacteria include plasmids, transposons, conjugative transposons, and integrons. All may carry one or more different genes which confer antibiotic resistances, resistance to one or more heavy metals, genes coding for enzymes which allow the host to utilize alternative energy sources, genes coding for the production of toxins, bacteriocins, and/or virulence factors, genes for self-replication, conjugative and/or mobilizable genes, and in some cases genes which confer susceptibility to specific bacteriophages. Some of the earliest plasmids were identified in Japanese Shigella spp. in the 1950s and conferred resistance to three or four different class of antibiotics. Plasmids differ in their bacterial host range with some such as those found in Haemophilus spp. having a very restrictive range that allows transfer between strains of a closely related species, to broad host range plasmids which are able to be transferred and survive in a variety of unrelated genera. In contrast, transposons and conjugative transposons tend to have broader host ranges, and less host restriction than plasmids (Roberts, 1989; Rice, 2007).
Conjugative transposons (CTns) are found in chromosomes and plasmids of a variety of bacterial species and in many ecosystems. They may carry multiple antibiotic resistance genes, multiple heavy metal resistance genes, and are widely disseminated in bacterial populations (Rice, 2007). The first CTn was Tn916, an 18 kb element that codes for tetracycline and minocycline resistance due to the tet(M) gene, and the first member of the Tn916-family of transposons. This and other CTns carry genes needed for their own conjugal transfer with integration into the new host relatively nonselectively in some species and highly site-specific in others (Rice, 2007). At about the same time a second transposon designated Tn1545 was described which carried multiple different antibiotic resistance genes including the rRNA methylase erm(B) gene and an aphA-3 gene coding for kanamycin resistance (Courvalin & Carlier, 1987). Both of these genes were integrated downstream of the tet(M). The Tn916-family of CTns has been identified in 27 different Gram-positive and 22 different Gram-negative genera. Many more genera have acquired the Tn916-like elements in vitro. More recently, composite transposons have been described where one or more transposons have been inserted within other transposons creating new multidrug resistant elements such as the composite element Tn5385. Tn5385 codes for tetracycline/minocycline, erythromycin, gentamicin, penicillin, streptomycin, and mercury resistance and shares genes with both staphylococcal and enterococcal elements. An interesting feature of the composite elements is their ability to transfer as a single unit to a recipient or to transfer the inserted transposon separately thus allowing for different combinations of genes to transfer depending on which element was received (Rice, 2007).
New CTns and plasmids have been described in the last few years which identify linkages between a variety of MLSKO and other genes. Some of the other genes and elements associated with acquired MLSKO genes are illustrated in Table 4. These include elements which link the tet(M), mef(A)-msr(D) genes(Figueiredo et al., 2006), the tetracycline resistant gene tet(O) linked upstream of the mef(A) and msr(D) genes (Brenciani et al., 2004), the linkage between the vat(E) and erm(B) genes (Jensen et al., 2000) and erm(B) and copper resistances genes (Hasman et al., 2006).
CTns may also play a role in maintaining specific genes in the bacterial population without selective pressure or without functioning in the particular bacteria. A good example can be found in the anaerobic Bacteroides spp. which often carry the erm(F) gene which confers resistance to the bacterial host and linked to a tet(X) gene, which codes for an enzyme that breaks down tetracycline in the presence of oxygen, but is nonfunctional in its anaerobic host (Chopra & Roberts, 2001). The hypothesis has been that the erm(F)-tet(X) genes which are now part of the CTnDOT element originated in an aerobic bacteria and was transferred as a unit to Bacteroides where the tet(X) gene product is nonfunctional. Recent data on a tetracycline resistant Sphingobacterium which carries the tet(X) gene, indicates that the genes flanking this tet(X) are the same genes as found flanking the tet(X) gene on the Bacteroides CTnDOT element except the erm(F) gene is missing in the Sphingobacterium isolate (Ghosh & LaPara, 2007; S. Ghosh, pers. Commun.). Thus, linkage between genes may be one way to keep non-expressed genes or genes which are currently not under selective pressure in bacterial populations.
Mobile elements and MLSKO genes
Sixteen of the MLSKO resistance genes have been linked to other antibiotic or heavy metal resistance genes, class 1 or 2 integrons and/or specific transposons (Table 4). Linkage of various antibiotic resistance genes in one mobile element allows the recipient to be converted from a susceptible isolate to a multiple resistant one with the integration of mobile elements carrying multiple antibiotic resistance genes. Selection for transfer of these elements can be done using multiple different antibiotics or chemicals as was observed with the replacement of avoparcin with tylosin without changes in the level of avoparcin resistant enterococci. It was then determined that the same plasmid which carried the vanA gene cluster for avoparcin resistance also carried an erm(B) gene which made the isolates resistant to tylosin. Therefore, when tylosin replaced avoparcin for animal use, the vanA gene remained because of its linkage with the erm(B) gene (Garcia-Migura et al., 2007). However, some caution does need to be taken when thinking about co-selection and resulting carriage of antibiotic resistant bacteria as our study on children treated for caries illustrates (Roberts et al., accepted). In this study, 75 children were treated with amalgam fillings and 75 were treated with composite fillings. Oral and urine cultures were taken before treatment, after treatment, and for 7 years of follow-up and then analyzed to determine whether the amalgam treated group had increased levels of antibiotic and/or mercury resistant commensal flora over baseline and higher levels than the control group treated with composite material. We found that there was no cultural evidence that children with amalgam fillings had higher levels of antibiotic or mercury resistant commensal oral or urinary bacteria as compared with the children treated with composite material.
Erythromycin was discovered over 60 years ago, but its use has diminished over time due to increased bacterial resistance. However, with the addition of the newer macrolide derivatives in the 1980s and more recently the ketolides and oxazolidinones, this group of antibiotics remains an important class of drugs for the treatment of a variety of community and hospital infectious diseases caused by Gram-positive and Gram-negative bacteria, and Mycobacterium spp. Dr Cohen (1992), The Centers for Diseases Control in the USA, coined the phrase ‘Post-Antimicrobial Era’ over 10 years ago to mean a time when community and hospital acquired infections will primarily be due to multidrug resistant pathogens where few or no viable antibiotic therapies exist for treating infections that were once treatable. Disease due to multidrug resistant pathogens increases the morbidity, mortality, and cost associated with treatment of these infections and the number of these infections increases yearly. To illustrate this point a recent study on methicillin resistant S. aureus (MRSA) infections from 2005 in the United States estimated that >90 000 MRSA infections occurred with >18 000 associated deaths (Klevens et al., 2007). This coupled with the low number of novel antibiotic classes currently in development, suggests that the ‘postantimicrobial era’ is almost here for the industrialized world for some multidrug resistant pathogens and for practical purposes has already been reached in much of the developing world. In tune with this concern is the increased numbers of ketolide resistant streptococci, oxazolidinone resistant enterococci, and staphylococci which have been identified suggesting that these current new antibiotics will also lose their effectiveness as more bacteria become resistant. Multiple strategies are needed to preserve the use of the currently effective MLSKO antibiotics. These strategies include the development of newer MLSKO derivatives, development of new classes of antibiotics and nonantibiotic therapies. Changes in infection control practices, personal hygiene, sanitation improvements, in agricultural and animal husbandry practices are also needed. In addition, a much better understanding is needed on how antibiotic resistance develops in different bacterial populations, why some of these genes spread quickly, and why others stay more localized. However, most importantly more resources are needed to monitor changes in antibiotic resistance patterns and the development of ways to extend the life of our current antibiotics such as using two different antibiotics for therapy instead of one antibiotic.