OUP user menu

Molecular basis of bacterial resistance to chloramphenicol and florfenicol

Stefan Schwarz, Corinna Kehrenberg, Benoît Doublet, Axel Cloeckaert
DOI: http://dx.doi.org/10.1016/j.femsre.2004.04.001 519-542 First published online: 1 November 2004


Chloramphenicol (Cm) and its fluorinated derivative florfenicol (Ff) represent highly potent inhibitors of bacterial protein biosynthesis. As a consequence of the use of Cm in human and veterinary medicine, bacterial pathogens of various species and genera have developed and/or acquired Cm resistance. Ff is solely used in veterinary medicine and has been introduced into clinical use in the mid-1990s. Of the Cm resistance genes known to date, only a small number also mediates resistance to Ff. In this review, we present an overview of the different mechanisms responsible for resistance to Cm and Ff with particular focus on the two different types of chloramphenicol acetyltransferases (CATs), specific exporters and multidrug transporters. Phylogenetic trees of the different CAT proteins and exporter proteins were constructed on the basis of a multisequence alignment. Moreover, information is provided on the mobile genetic elements carrying Cm or Cm/Ff resistance genes to provide a basis for the understanding of the distribution and the spread of Cm resistance – even in the absence of a selective pressure imposed by the use of Cm or Ff.

  • Chloramphenicol acetyltransferase
  • Cat and cml gene nomenclature
  • Enzymatic inactivation
  • Florfenicol resistance
  • Specific exporter
  • Multidrug transporter
  • Co-selection

1 Chemical structures and properties of chloramphenicol and florfenicol

Chloramphenicol (Cm), originally referred to as chloromycetin, was isolated from Streptomyces venezuelae in 1947 [1] and shown to be a broad spectrum antibiotic with a novel structure (Fig. 1), remarkable both for a p-nitrophenyl group (at C-1) and an N-dichloroacetyl substituent (at C-2) attached to a 1,3-propanediol with two chiral centers (C-1 and C-2) [2]. Cm was the first naturally occurring substance described to contain a nitro group. The relative simplicity of Cm made it the first antibiotic to be marketed as the product of chemical synthesis and Cm has been produced exclusively this way since 1950. Only one (d-threo) of the four possible diastereoisomers possesses antibiotic activity. The C-3 primary hydroxyl group, initially thought to be essential for inhibition of protein synthesis through its affinity for the peptidyltransferase of 50S ribosomes, can be replaced with fluorine [2]. Besides the fluoro substitution at C-3 (in florfenicol), very few other substitutions are tolerated without adverse effects on antimicrobial activity [2]. Among them, the substitution of the nitro group (–NO2), which was considered to be responsible for the dose-unrelated aplastic anemia, by a sulfomethyl group (–SO2CH3) at the para position of the 1-phenyl moiety became effective in thiamphenicol and florfenicol (Fig. 1).

Figure 1

Structure of chloramphenicol and related substances.

Cm is a highly stable antibiotic which can be stored for prolonged times at room temperature. It is amphiphilic and unionized at physiological pH [2]. Cm can pass biological membranes to reach intracellular bacteria and is able to readily traverse the blood–brain barrier [24]. The fluorinated Cm derivative florfenicol (Ff) is a synthetic drug, of which also only the d-threo stereoisomer is antibiotically active. Ff is unionized in a pH range from 3 to 9 [5] and also poorly soluble in aqueous solutions. Due to its lipophilicity Ff shows a good tissue penetration.

2 Use in human and veterinary medicine

Cm and some derivatives, such as thiamphenicol and azidamfenicol (Fig. 1), have been used over the years in human medicine. Certain esters of Cm, such as Cm palmitate or Cm succinate, have been produced for therapeutic applications. They do not exhibit antimicrobial activity until Cm is released after hydrolysis by esterases [3]. Cm succinate shows a good solubility in water and therefore is used for parenteral applications. The water soluble azidamphenicol is only used in eye drops [3]. In the early years after its introduction into clinical use, Cm was considered as a promising broad spectrum antibiotic [2]. However, a number of adverse effects have been observed since the mid-1960s in connection with the application of Cm [6]. These side-effects include a dose-unrelated irreversible aplastic anemia which occurs at frequencies of 1:10,000–1:40,000 [3] or 1:20,000–1:600,000 [7], a dose-related reversible bone-marrow suppression, or the Gray syndrome in neonates and infants [2, 4]. Occasionally, hypersensitivity to Cm ranging from skin rashes to anaphylaxis has been observed, too [4]. Based on these adverse effects and on the availability of less toxic antimicrobial agents with a similar spectrum of activity, the use of Cm in humans is nowadays limited to the therapy of a small number of life threatening infections [2, 4]. Since Cm readily crosses the blood–brain barrier, it remains an alternative therapeutic agent for the treatment of meningitis caused by susceptible strains of Haemophilus influenzae, Neisseria meningitidis or Streptococcus pneumoniae when no other antimicrobial agents can be used, e.g. in penicillin-allergic patients [8].

The use of Cm in veterinary medicine in the European Union (EU) is currently limited to pets and non-food-producing animals. It was banned in 1994 from use in any food-producing animals in the EU. The main reason for this ban was protection of the consumer from potential adverse effects arising from Cm residues in carcasses of food animals. Because of the dose-independent irreversible aplastic anemia in humans, the “non-observed effect level” (NOEL) could not be determined for Cm. In toxicological studies, NOEL represents the dose at and below which adverse effects do not occur [9]. The calculation of the “maximum residue level” (MRL), which represents the maximum level of antibiotic residues acceptable in carcasses at slaughter without any adverse effect on public health, is based on the NOEL and, therefore, could not be determined. As a consequence, EU legislation banned Cm along with several antibiotics, including nitroimidazoles and furazolidinones, from use in food-producing animals.

Since adverse side-effects, in particular the dose-independent irreversible aplastic anemia, have not been observed in animals, the fluorinated Cm derivative Ff has been licensed for the control of bacterial respiratory tract infections in cattle and pigs. Ff was approved in the EU for the use in cattle and in pigs in 1995 and in 2000, respectively. Moreover, Ff is also approved for the treatment of infectious pododermatitis (interdigital phlegmon) in cattle due to Fusobacterium necrophorum and Bacteroides melaninogenicus in the USA. In commercial salmon farming, a Ff premix is used for the treatment of furunculosis in salmons caused by Aeromonas salmonicida. Fluorinated Cm derivatives are currently not used in human medicine.

3 Mode of action and spectrum of activity

In procaryotes, Cm is a highly specific and potent inhibitor of protein biosynthesis. Cm-dependent inhibition of bacterial protein biosynthesis is mainly due to the prevention of peptide chain elongation. Its bacteriostatic activity is based on a reversible binding to the peptidyltransferase centre at the 50S ribosomal subunit of 70S ribosomes [10]. 80S ribosomes of eucaryotic cells are not targets of Cm and its derivatives. However, it has been assumed that Cm may interact with mitochondrial ribosomes which are similar in their structure to 70S ribosomes rather than to 80S ribosomes. As a possible consequence, the mitochondrial function of stem cells in bone marrow may be impaired resulting in a suppression of the bone marrow function [6, 11].

The substrate spectrum of Cm includes Gram-positive and Gram-negative, aerobic and anaerobic bacteria, but also chlamydiae, mycoplasmas, and rickettsiae [2, 4]. Cm analogs including the fluorinated derivative Ff have a similar spectrum of antimicrobial activity as Cm [4]. Intrinsic resistance to Cm and Ff has not been observed although members of different bacterial species and genera may differ in their basic levels of susceptibility to both drugs as confirmed by the determination of minimum inhibitory concentrations (MICs) [4, 12].

The currently valid MIC breakpoints approved by the National Committee for Clinical Laboratory Standards (NCCLS) [13] indicate that streptococci, including S. pneumoniae, are considered as susceptible to Cm when they exhibit MICs of >4 mg l−1 whereas the breakpoint for susceptibility is at 8 mg l−1 in bacteria other than streptococci. Approved breakpoints for Ff are currently available for Pasteurella multocida, Mannheimia haemolytica and H. somnus from respiratory diseases in cattle as well as for P. multocida, Actinobacillus pleuropneumoniae, Bordetella bronchiseptica and S. suis from respiratory diseases in pigs and Salmonella enterica serovar Choleraesuis from infections in pigs. Isolates of all aforementioned bacterial species, except S. Choleraesuis, are considered as susceptible when their MIC of Ff is >2 mg l−1. For S. Choleraesuis, the Ff-specific breakpoint for susceptibility is 4 mg l−1.

4 Bacterial resistance to chloramphenicol and florfenicol

Over the years, bacteria have developed a number of mechanisms which enable them to circumvent the inhibitory effects of Cm. The first and still most frequently encountered mechanism of bacterial resistance to Cm is enzymatic inactivation by acetylation of the drug via different types of chloramphenicol acetyltransferases (CATs) [14]. However, there are also reports on other mechanisms of Cm resistance, such as efflux systems, inactivation by phosphotransferases, mutations of the target site and permeability barriers [2, 14]. As the number of genes associated with resistance to Cm and related drugs increases, inconsistencies of their nomenclature were observed. These included the assignment of identical designations for different Cm resistance genes and that of different designations for virtually the same resistance gene (Tables 13). As previously shown for the nomenclature of tetracycline resistance [15] and macrolide–lincosamide–streptogramin resistance genes [16], there is also an urgent need for a generally accepted and unified nomenclature for genes involved in resistance to Cm and Ff.

View this table:
Table 1

Type A chloramphenicol acetyltransferases

GroupGene designation(s)Bacterial source(s)Plasmid/transposon/chromosome% IdentityDatabase accession no(s).References
A-1catIEscherichia coliTn9,R42998.3–10097.3–100V00622[27]
catIAcinetobacter baumanniiChromosome (Tn2670)M62822[28]
catAcinetobacter calcoaceticusTn2670-likeM37690[28]
pp-catPhotobacterium damselae subsp. piscicidapSP9351D16171[29]
Pseudomonas putidaUnknownE02706[30]
catSalmonella TyphipHCM1AL513383, NC_003384[31]
catSerratia marcescensR478NC_005211, BX664015[32]
catShigella flexneriChromosomeAF326777[33]
A-2catIIHaemophilus influenzaepRI234, pMR37589.3–99.792.0–99.5X53797[24]
catIIAgrobacterium tumefaciens, Escherichia colipSaX53796[24]
catPhotobacterium damselae subsp. piscicidaPlasmidbAB082569[35]
A-3catIIIShigella flexneriR38799.8–10099.5–100X07848[36]
catA3Mannheimia taxon 10pMHSCS1AJ249249[38]
catA3Mannheimia varigenapMVSCS1AJ319822[39]
catUncultured eubacteriumpIE1130NC_004973, AJ271879[40]
A-4catProteus mirabilisChromosomeM11587[41]
A-5catStreptomyces acrimyciniChromosomeP20074[42]
A-6cat86Bacillus pumilusChromosomeK00544[43]
A-7cat(pC221), catCStaphylococcus aureuspC22196.6–10095.3–100X02529[45, 46]
catStaphylococcus aureuspKH7U38429[53]
catStaphylococcus aureuspUB112X02872[47]
catStaphylococcus intermediuspSCS1M64281[48]
catStaphylococcus aureuspSCS6X60827[52]
catBacillus subtilispTZ12M16192[63]
catStreptococcus agalactiaepGB354U83488[58]
catStreptococcus agalactiaepIP501X65462[57]
catEnterococcus faecalispRE25X92945[61]
A-8cat(pC223)Staphylococcus aureuspC22389.2–10085.2–100NC_005243, AY355285[54]
catStaphylococcus aureuspSCS7M58516[50]
catListeria monocytogenespWDB100X68412[64]
catStaphylococcus aureuspSBK203M90091[51]
catLactococcus lactispK214X92946[67]
catStaphylococcus haemolyticuspSCS5M58515[49]
catEnterococcus faeciumpRUMNC_005000, AF507977[62]
A-9cat(pC194)Staphylococcus aureuspC19493.9–99.887.0–94.9V01277[44]
cat-TCLactobacillus reuteripTC82U75299[65]
catStreptococcus suisTnSs1AB080798[60]
catStaphylococcus aureuspMC524-MBMAJ312056[55]
A-10catBacillus clausiiChromosomeAY238971[74]
A-11catPClostridium perfringenspIP401:Tn4451100100U15027[75]
catPNeisseria meningitidisChromosomeAF031037[81]
catDClostridium difficileChromosome (Tn4453)X15100, AF226276[83]
A-12catSStreptococcus pyogenesChromosomeX74948[84]
A-13catCampylobacter colipNR9589M35190[85]
A-14catListonella anguillarumpJA7324S48276[86]
A-15catBClostridum butyricumChromosomeM93113[78]
A-16catQClostridium perfringensChromosomeM55620[76]
  • aAmino acid.

  • bNo plasmid designation available.

View this table:
Table 2

Type B chloramphenicol acetyltransferases

GroupGene designation(s)Bacterial source(s)Plasmid/transposon/chromosome% IdentityDatabase accession no(s).References
B-1cat, catB1Agrobacterium tumefaciensChromosome100100M58472[98]
catAgrobacterium tumefaciensChromosomeNC_003063[99]
B-2catB2Escherichia colipNR79:Tn242499.5–99.899.0–100AF047479[100]
catB2Salmonella EnteritidisPlasmidAJ487034[101]
catB2Uncultured eubacteriumpSp39AY139601[103]
catB2Pasteurella multocidapJR1NC_004771, AY232670[102]
B-3catB3Salmonella TyphimuriumpWBH30184.3–10084.8–100AJ009818[105]
catB3Salmonella TyphimuriumIncF1 plasmidAJ310778[106]
catB3Acinetobacter baumanniiChromosomeAF445082[107]
catB3Escherichia colipHSH2AY259086[108]
catB4Enterobacter aerogenespWBH301U13880[105]
catB4Klebsiella pneumoniaepEKP0787-1AF322577[109]
cat, catB5Morganella morganiiTn840X82455[104]
catB6Pseudomonas aeruginosapPAM-101AJ223604[112]
catB8Klebsiella pneumoniaepKB42AF227506[110]
catB8Salmonella TyphipST2301AY123251[111]
catB8Pseudomonas aeruginosaUnknownAF418284[113]
B-4catB7Pseudomonas aeruginosaChromosome100100AF036933[114]
catB7Pseudomonas aeruginosaChromosomeAE004506[115]
B-5catB9Vibrio choleraeChromosome100100AF462019[116]
catB9Vibrio choleraeChromosomeNC_002506[117]
View this table:
Table 3

Specific exporters mediating resistance to chloramphenicol or chloramphenicol/florfenicol

GroupGene designation(s)Bacterial source(s)Plasmid/transposon/chromosome% IdentityDatabase accession no(s).References
E-1cmlB, cmlA2Enterobacter aerogenespIP83383.1–10083.9–100AF034958[135]
cmlASalmonella TyphimuriumPlasmidAJ487033[127]
cmlA5Escherichia coliR751 (Tn2000)AF205943[126]
cmlA1Klebsiella pneumoniaepILT-3AF458080[129]
cmlA1Pseudomonas aeruginosaRPL11 (Tn1403)AF313472[130]
cmlA4Klebsiella pneumoniaepTK1AF156486[128]
cmlA5Uncultured bacteriumpSp1AY115475[103]
cmlA6Pseudomonas aeruginosaPlasmidAF294653[131]
cmlA7Pseudomonas aeruginosaChromosomeAJ511268[132]
cmlA, cmlA1Pseudomonas aeruginosapR1033:Tn1696U12338, M64556, AF078527[123, 124]
E-2cmlEscherichia coliR26M22614[136]
E-3cmlA-likeSalmonella Typhimurium DT104Chromosome95.6–10087.7–100AF071555[140]
floRSalmonella Typhimurium DT104ChromosomeAF118107[139]
floSalmonella Typhimurium DT104ChromosomeAJ251806[143]
floRSalmonella Typhimurium DT104ChromosomeAF261825[145]
floRSalmonella Typhimurium DT104ChromosomeAY339985[144]
floEscherichia coliPlasmidAF252855[133]
floREscherichia coliPlasmidAF231986[152]
floREscherichia colipMBSF1AJ518835[154]
floRKlebsiella pneumoniaeR55AF332662[155]
floRVibrio choleraeChromosome (SXT element)AY034138[157]
floRVibrio choleraeChromosome (SXT element)AY055428[158]
pp-floPhotobacterium damselae subsp. piscicidapSP92088D37826[138]
E-4fexAStaphylococcus lentuspSCFS2AJ549214[160]
E-5cmlStreptomyces lividansChromosomeX59968[161]
E-6cmlvStreptomyces venezuelaeChromosomeU09991[162]
E-7cmrARhodococcus rhodochrousTn556177.586.2AF015087[164]
cmrRhodococcus fascianspRF2Z12001[163]
E-8cmrCorynebacterium glutamicumpXZ1014599.999.7U85507[166]
cmxCorynebacterium striatumpTP10:Tn5564AF024666[165]

4.1 Chloramphenicol acetyltransferases

Cm acetyltransferases (CATs) are able to inactivate Cm as well as thiamphenicol and azidamfenicol. Due to the replacement of the hydroxyl group at C-3 by a fluor residue, the acceptor site for acetyl groups was structurally altered in Ff. This modification rendered Ff resistant to inactivation by CAT enzymes, and consequently, Cm-resistant strains, in which resistance is exclusively based on the activity of CAT, are susceptible to Ff [17]. There are two defined types of CATs which distinctly differ in their structure: the classical CATs, referred to in this review as type A CATs and the novel CATs, also known as xenobiotic CATs [14], but referred to in this review as type B CATs. In addition, annotations of cat genes were found in the whole genome sequences of Rhodobacter capsulatus [18], Mesorhizobium loti [19], S. agalactiae strain 2306 [20], Bacillus cereus [21], and Brucella melitensis [22]. The potential CAT variants encoded by these five presumed cat genes do not exhibit structural features that allow their assignment to either type A or type B. As long as functional activity has not been confirmed, assignment of these CAT-like proteins to further novel subtypes has to be postponed.

4.1.1 Type A chloramphenicol acetyltransferases

Type A CATs have been detected in a wide variety of bacteria [2, 14]. Despite the differences in their amino acid sequences, the type A CATs share some common properties. The native CAT is usually composed of three identical polypeptides each ranging in size between 207 and 238 amino acids (aa) [2, 14]. In cells in which two different, but related, CATs are present, functionally active heterotrimers may also occur [14]. The cat gene codes for the CAT monomer. In all currently known type A CATs, some amino acids, which are involved in substrate binding, catalytic activities, folding of the monomers, or assembly of the monomers to a trimer, appear to be conserved (for a review, see [14]). Some of the type A CATs have specific properties, such as the capability to mediate resistance also to fusidic acid [2, 23] or sensitivity to inhibition by thiol-reactive reagents [24]. One type A CAT enzyme, CATIII from Shigella flexneri, has been studied by X-ray crystallography [25, 26] and the data derived have been the basis for the understanding of the catalytic activities and the assembly of the CAT monomers.

There are at least 16 distinct groups, A-1–A-16, of catA genes. The corresponding type A CAT proteins – assigned to the same group – exhibit amino acid sequence identities of more than 80%. The different groups and their representatives are listed in Table 1. Phylogenetic relationships of the different type A CAT proteins are displayed in Fig. 2. All phylogenetic trees shown in this review are based on multisequence alignments and were produced using the DNAMAN software (Lynnon-BioSoft, Ont., Canada).

Figure 2

Phylogenetic tree of the class A CATs. Branch lengths are scaled according to amino acid exchanges observed in a multisequence alignment. The numbers at the major branch points refer to the percentage of times that a particular node was found in 1000 bootstrap replications. The bacterial source and the database accession number are given for each CAT protein. Moreover, the groups according to Table 1 are indicated on the right-hand side.

The prototype cat gene of group A-1, catI, was originally identified as part of transposon Tn9 [27] in Escherichia coli and has been detected on a variety of resistance plasmids of Gram-negative bacteria, such as Acinetobacter spp. [28], Photobacterium damselae subsp. piscicida, formerly known as P. piscicida [29], and Pseudomonas putida [30]. More recently cat genes of this group have also been identified in S. Typhi [31], Serratia marcescens [32] and S. flexneri [33]. Enzymes resembling that encoded by catI have also been detected in B. ochraceus [2]. The cat genes of group A-2 are mainly found on plasmids of H. influenzae which usually carry at least one more resistance gene [24, 34]. Genes closely related to catII have also been detected in E. coli and Agrobacterium tumefaciens [24] and P. damselae subsp. piscicida [35]. An CAT enzyme similar to that encoded by catII was also reported to be present in B. fragilis [2]. The members of group A-3 are commonly found on plasmids in Enterobacteriaceae [36] and Pasteurellaceae [3739], but a cat gene of this group has also been detected on a plasmid from an uncultured bacterium [40]. Most of these plasmids carry one or more additional resistance genes. The groups A-4–A-6 are represented by unique cat genes, all of which are located in the chromosomal DNA of either Proteus mirabilis [41], Streptomyces acrimycini [42], or B. pumilus [43].

The cat genes commonly found in staphylococci [4455], streptococci [5660] and enterococci [61, 62], but also in B. subtilis [63], Listeria monocytogenes [64], Lactobacillus reuteri [65] and Lactococcus lactis [66, 67] represent the three groups of cat genes A-7–A-9. The prototype plasmids from which sequences of the respective cat genes had first been deposited in the databases are pC221 [45, 46], pSCS7 [50], and pC194 [44]. These three groups of cat genes are commonly located on small multicopy plasmids which carry either the cat gene alone or in combination with a streptomycin resistance [6870] or a macrolide resistance gene [57]. In rare cases, such genes were found to be part of multiresistance plasmids [67, 71, 72] or conjugative transposons [73]. The only representative of group A-10 was found in the chromosomal DNA of the probiotic B. clausii [74]. Group A-11 comprises two cat genes, so far reported in the literature as catP and catD. The gene catP from Clostridium perfringens was identified as part of the Cm resistance transposon Tn4451 [75] which is able to integrate into plasmids as well as into the chromosomal DNA. The gene catP has been detected not only in clostridia [7580], but also in N. meningitidis [81, 82]. An identical gene, designated catD, was identified in C. difficile [83]. The gene catS from S. pyogenes represents group A-12 and has been detected in single streptococcal isolates of serogroups A, B and G [84]. Even though only part of this gene is deposited in the database, analysis of the amino acid sequence deduced from this internal segment shows approximately 77% identity to the catP and catD gene products.

The cat gene of group A-13 was located on plasmid pNR9589 of Campylobacter coli which also harboured an aphA-3 gene for kanamycin resistance [85] whereas that of group A-14 was found as part of a multiresistance plasmid which also conferred resistance to tetracycline, sulfonamides and streptomycin from Vibrio anguillarum, meanwhile reclassified as Listonella anguillarum [86]. The representatives of the remaining groups of cat genes A-15 and A-16, catB from C. butyricum [78] and catQ from C. perfringens [76] are located in the chromosomal DNA and have so far been identified only in these clostridial hosts.

Database searches also revealed the presence of putative catA genes in Bacillus anthracis strain Ames [87], Deinococcus radiodurans R1 [88], C. acetobutylicum ATCC824 [89], C. tetani E88 [90], Zymomonas mobilis [91], and B. thetaiomicron VPI-5482 [92]. These genes were detected by whole genome sequencing. If CAT activity of their gene products is confirmed, they may be considered as the representatives of another six individual groups of catA genes.

Of the known catA genes, only cat86 and the cat genes similar to those located on plasmids pC221, pC223/pSCS7, or pC194 are inducibly expressed via translational attenuation, with Cm itself acting as an inducer [93]. The translational attenuators which are located immediately upstream of the cat structural genes consist of a single pair of inverted repeated sequences IR1 and IR2 as well as a reading frame for a short peptide of 6–9 aa. Since a common mRNA transcript is produced from the cat gene and its regulatory region, IR1 and IR2 are able to form a stable stem-loop structure with the cat-associated ribosome binding site located within IR2 (Fig. 3). The codons 2–5 of the short peptide are highly conserved and include a ribosome stall sequence which is complementary to sequences at the 3 terminus of the 16S rRNA. A ribosome stalled at this position overlaps at least in part the IR1 sequence, thereby preventing formation of a mRNA secondary structure and rendering IR2 accessible to a second ribosome. It is believed that binding of Cm to ribosomes triggers conformational changes that eventually put the relevant area of the 16S rRNA in an exposed position which then allows base pairing with the complementary part of the cat transcripts. Inducible catA genes mediate high level resistance to Cm with MICs of <128 mg l−1. Expression of the catA gene found in P. mirabilis strain PM13 was obviously based on a flip-flop control mechanism involving an invertible promoter and a trans-acting product [41]. As far as information is available, the remaining catA genes are expressed constitutively and their MICs of Cm differ with regard to the strength of the catA-associated promoter and the copy number of the catA gene.

Figure 3

Presentation of the regulatory region of the catA gene located on the S. intermedius plasmid pSCS1 [48]. The ORF9- and catA-associated ribosome binding sites (RBS) are boxed. The start codons of ORF9 and catA are underlined and the corresponding coding sequences are displayed in bold type letters. The inverted repeated sequences IR1 and IR2 are marked by arrows and a stable mRNA secondary structure formed by these IR sequences is shown.

4.1.2 Type B chloramphenicol acetyltransferases

Type B CATs, occasionally referred to as xenobiotic acetyltransferases, also inactivate Cm by acetylation. Type B CATs share some common properties with the type A CATs: native type B CATs are also homotrimers composed of monomers which are in the range of 209–212 aa [14]. However, on the basis of their amino acid sequences, type B CATs differ distinctly in their structure from type A CATs and appear to be related to other acetylating enzymes of staphylococci and enterococci involved in resistance to A compounds of the streptogramins, such as Vat(D) (formerly known as SatA) [94], Vat(E) (formerly known as SatG; [95]), Vat(A) (formerly known as Vat) [96], or Vat(B) [97]. The structural relationships of some of the type B CATs to one another and to other acetylating enzymes are described in detail in [14].

There are at least five different groups of type B cat genes: B-1–B-5 (Table 2, Fig. 4). The first type B cat gene described, catB1, was cloned from the chromosome of A. tumefaciens [98, 99]. The catB2 gene was initially found on the multiresistance transposon Tn2424 from E. coli [100]. Closely related catB2 genes – all assigned to group B-2 – were also reported to occur on plasmids isolated from S. Enteritidis [101], P. multocida [102], or from an uncultured eubacterium [103]. The group B-3 comprises a number of genes so far known as catB3catB6 and catB8 (Table 2). These genes have often been associated with either multiresistance transposons such as Tn840 from Morganella morganii [104] or plasmid-borne multiresistance integrons and have been detected in a variety of enterobacterial species [105111] as well as in P. aeruginosa [112, 113] (Table 2). The catB7 genes – representing group B-4 – were found in the chromosome of P. aeruginosa PAO222 [114] and PAO1 [115] whereas catB9 genes – representing group B-5 – were found in the chromosomal DNA of Vibrio cholerae [116, 117] as part of a super-integron [116].

Figure 4

Phylogenetic tree of the class B CATs. Branch lengths are scaled according to amino acid exchanges observed in a multisequence alignment. The numbers at the major branch points refer to the percentage of times that a particular node was found in 1000 bootstrap replications. The bacterial source and the database accession number are given for each CAT protein. Moreover, the groups according to Table 2 are indicated on the right-hand side.

During whole genome sequencing, a gene related to the cat genes of group B-2 was found in the chromosome of Shewanella oneidensis strain MR-1 [118] and a putative catB gene was detected in the chromosome of V. parahaemolyticus [119]. CAT activity of the gene products of these two genes, however, has not yet been confirmed. In addition, the incomplete amino acid sequences of another four type B CAT proteins from S. marcescens, P. aeruginosa, B. sphaericus and Staphylococcus aureus have been reported [14]. These observations suggest a wider distribution of type B cat genes among Gram-negative and Gram-positive bacteria than initially assumed.

Translational attenuation has also been proposed as the regulatory mechanism for the Cm-inducible catB1 gene from A. tumefaciens [120]. In contrast to the single pair of inverted repeated sequences seen in the regulatory regions of the staphylococcal catA genes, four different pairs of inverted repeats were detected immediately upstream of catB1 resulting in a more complex mRNA secondary structure [120]. This gene has been reported to confer only low level Cm resistance with 5 to <20 mg l−1 [14]. Several catB genes are part of gene cassettes and thus are transcribed from a promoter located in the integron. Cassette-borne genes located closest to the promoter are more highly expressed than distal cassettes [116]. Rowe-Magnus et al. investigated the level of Cm resistance mediated by the cassette-borne gene catB9 in relation to its position within a multiresistance integron from V. cholerae consisting of seven gene cassettes. When placed in the most distal seventh position, the MIC of Cm was <1 mg l−1 which corresponded to a Cm susceptible phenotype. However, when the catB9 cassette was placed in the first four positions, Cm resistance at levels of <25 mg l−1 were observed [116].

4.2 Chloramphenicol exporters

The export of Cm or Ff from the bacterial cell can be mediated by either specific transporters and/or multidrug transporters. Specific transporters have a substrate spectrum which is commonly limited to a small number of structurally closely related compounds whereas that of the multidrug transporters often includes a wide range of unrelated substances. Specific transporters commonly mediate distinctly higher levels of resistance as compared to those of multidrug transporters. While specific transporters involved in the export of Cm or Ff have no known function in the physiological cell metabolism, multidrug transporters play an important role in the excretion of toxic compounds, occasionally also including specific antimicrobial agents such as Cm and Ff, from the bacterial cell.

4.2.1 Specific exporters

Genes associated with the export of Cm or Cm/Ff are found in a wide variety of clinically relevant and environmental bacteria. A short description of mobile genes coding for specific efflux proteins, including those mediating the export of Cm or Cm/Ff from the bacterial cell, was recently published by Butaye et al. [121]. At least eight different groups of specific exporters, E-1–E-8, are currently known (Table 3). Their phylogenetic relationships are shown in Fig. 5. Only the exporters assigned to groups E-3 and E-4 have been reported to mediate resistance to both, Cm and Ff.

Figure 5

Phylogenetic tree of the specific exporter proteins involved in Cm or Cm/Ff resistance. Branch lengths are scaled according to amino acid exchanges observed in a multisequence alignment. The numbers at the major branch points refer to the percentage of times that a particular node was found in 1000 bootstrap replications. The bacterial source and the database accession number are given for each exporter protein. Moreover, the groups according to Table 3 are indicated on the right-hand side.

Resistance to Cm not due to enzymatic inactivation was first detected in 1979 in P. aeruginosa [122] and later on shown to be based on the presence of the transposon Tn1696. Sequence analysis of the Cm resistance gene of Tn1696, cmlA, revealed that the corresponding protein of 419 aa had 12 transmembrane domains and thus resembled closely other transmembrane transport proteins of the major facilitator superfamily [123125]. The cmlA gene proved to be part of a gene cassette. However, in contrast to other cassette-borne resistance genes, the cmlA gene had its own promoter and regulation of cmlA expression was inducibly regulated via translational attenuation. An attenuator-like structure – similar to that of inducibly expressed catA genes – was detected upstream of the cmlA gene [123]. During the last decade, a number of genes closely related to or indistinguishable from cmlA– and all assigned to group E-1 – have been identified in a wide variety of Gram-negative bacteria, including E. coli [126], S. Typhimurium [127], Klebsiella pneumoniae [128, 129], and P. aeruginosa [130132], but also from an uncultured eubacterium [103] (Table 3). E. coli strains carrying the gene cmlA were reported to exhibit MICs of Cm of <32 to >256 mg l−1 and MICs of Ff of <8 to 64 mg l−1 [133, 134].

As compared to the other members of group E-1, the gene cmlA2 (also referred to as cmlB) from Enterobacter aerogenes [135] showed only 84% nucleotide sequence identity and 85% identity in the amino acid sequence. Database search also identified a Cm resistance gene, designated cml, which was located on plasmid R26 [136]. The Cml protein, which does not mediate resistance to fluorinated Cm analogs [137], consists of 302 amino acids and exhibits only five transmembrane segments. It is in part similar to the distinctly larger CmlA protein of Tn1696 and represents group E-2.

A number of genes, referred to in the published literature as pp-flo, cmlA-like, floSt, flo, or floR, mediate combined resistance to Cm and Ff and were grouped together in group E-3. Despite their varying designations, these genes are closely related and show 96–100% identity in their nucleotide sequences and 88–100% identity in the amino acid sequences of their products. The first member of this group, pp-flo, was detected on plasmids in the fish pathogen P. damselae subsp. piscicida in 1996 [138]. More recently, genes of group E-3 were identified in a chromosomal multiresistance gene cluster of S. Typhimurium DT104 [139144]. This antibiotic resistance gene cluster of about 13 kb is included in a chromosomal genomic island called Salmonella Genomic Island 1 (SGI1) [145]. The 43-kb SGI1 is located at the 3 end of the thdF gene in the chromosome of S. Typhimurium DT104. SGI1 or variants of SGI1 have also been identified at the same chromosomal location as in S. Typhimurium DT104 in another S. Typhimurium phage type, i.e. DT120, and in other S. enterica serovars, i.e. S. Agona, S. Paratyphi B, S. Albany and recently in S. Newport, indicating the horizontal transfer potential of SGI1 [146150]. In one of these SGI1 variants, the floR gene was interrupted by an IS6100 element (see SGI1-E in Fig. 6). Transposition of IS6100, found at the 3 end of the various SGI1 associated antibiotic resistance gene clusters, resulted also in inversion of part of this antibiotic resistance gene cluster. The floR gene was also identified on plasmids and in the chromosome of E. coli from cattle, poultry and pigs [133, 134, 151154], on IncC plasmid R55 from K. pneumoniae [155] and on a closely related plasmid from S. Newport [156], and in the chromosomal SXT elements of V. cholerae [157, 158]. Analysis of the DNA sequences upstream and downstream of the floR gene in plasmids and in the SXT element showed regions with large homology upstream and downstream of the floR gene (Fig. 7). In addition, DNA sequence analysis revealed the presence of open reading frames (orfA, orfA5Δ3, orfA3Δ5) whose gene products showed considerable homology to transposase proteins (Fig. 7), suggesting that the floR gene could be located within a yet unidentified transposable element. Since the 374 amino acid protein from Photobacterium was shorter than the corresponding 404 amino acid proteins from Enterobacteriaceae and V. cholerae, prediction of transmembrane segments yielded only 10 transmembrane segments in the protein from Photobacterium in comparison to 12 transmembrane segments in those from Salmonella, E. coli, K. pneumoniae, and V. cholerae. However, it was speculated that the pp-flo sequence may be incomplete and/or contains sequencing errors [159]. Bovine E. coli strains from Northern America carrying the gene floR showed MICs of Cm of 16 to >256 mg l−1 and MICs of Ff of 16 to >256 mg l−1 [133, 134]. In contrast, European E. coli strains from bovine and porcine origin exhibited MICs of Ff of <128 mg l−1 [152154].

Figure 6

Genetic organization of the antibiotic resistance gene clusters containing the floR gene of Salmonella Genomic Island 1 (SGI1) in different Salmonella enterica serovars. SGI1, SGI1-A, SGI1-E, SGI1-F, and SGI1-H are schematized. SGI1 and variants are always located between the thdF and yidY chromosomal genes. DR-L and DR-R are the left and right direct repeats, respectively, bracketing SGI1. The first SGI1 gene int codes for a putative integrase probably involved in site-directed integration of SGI1 at the 3 end of thdF. The floR gene and other antibiotic resistance genes are represented as black and grey arrows, respectively. Besides floR conferring cross-resistance to Cm and Ff, the other antibiotic resistance genes mediate resistance to ampicillin (pse-1), gentamicin (aac(3)-Id), streptomycin and spectinomycin (aadA2, aadA7), sulfonamides (sul1), tetracyclines (tet(G)), and trimethoprim (dfrA1, dfrA10).

Figure 7

Organization of the floR gene loci of different plasmids from E. coli, K. pneumoniae, P. damselae subsp. piscicida and of the chromosomal SXT element from V. cholerae. The floR (or pp-flo) gene is shown as a black arrow while the putative transposase genes, orfA or parts of orfA, are displayed as grey arrows. The regions of homology with the floR locus of the plasmid from E. coli 10660-1 are indicated by thick black bars. The genes strA–strB, dfr18 and sul2 genes confer resistance to streptomycin, trimethoprim, and sulfonamides, respectively.

A novel type of exporter gene, designated fexA, represents group E-4. This gene has recently been identified on the 34-kb plasmid pSCFS2 from S. lentus [160]. The FexA protein consists of 475 amino acids and exhibits 14 transmembrane domains. It mediates resistance to Cm and Ff. Inducible expression of fexA is obviously due to translational attenuation. An attenuator-like structure was identified immediately upstream of the fexA gene [160]. The S. lentus isolate carrying fexA showed basic MICs of Cm of 64 mg l−1 and Ff of 32 mg l−1 which after induction with either Cm or Ff were increased to 128 mg l−1. This gene is the first exporter gene identified in Gram-positive cocci which mediates combined resistance to Cm and Ff. The gene fexA shows no similarity to any other genes involved in the export of Cm or Ff.

Several Cm exporters – distinctly different from the aforementioned ones found in clinically relevant bacteria from humans and animals – were identified in soil and environmental bacteria of the genera Streptomyces (groups E-6 and E-7), Rhodococcus (group E-8) or Corynebacterium (group E-9). The cml gene identified in the chromosome of S. lividans 1326 [161] encodes a protein with 12 transmembrane segments while the chromosomal gene cmlv of S. venezuelae ISP5230, a chloramphenicol producer, encodes a protein of 10 transmembrane segments assumed to be involved in the export of Cm [162]. Cm exporter proteins of 12 transmembrane domains were encoded by the genes cmr and cmrA from Rhodococcus spp. The gene cmr was found on the conjugative plasmid pRF2 from Rhodococcus fascians [163] while the gene cmrA appeared to be associated with transposon Tn5561 in R. rhodochrous [164]. As part of the transposon Tn5564 the gene cmx was detected on a multiresistance plasmid in Corynebacterium striatum [165]. A transposon-like element closely related to Tn5564 and carrying a Cm resistance gene, designated cmr, was found on plasmid pXZ10145 in C. glutamicum [166]. The Cmx and Cmr proteins from Corynebacterium spp. appear to have 10 transmembrane segments.

As previously seen with the catA and catB genes, annotations of florfenicol resistance genes were also detected in whole genome sequences of C. tetani strain E88 [90], B. melitensis [22], Yersinia pestis [167], F. nucleatum subsp. nucleatum [168], Legionella pneumophila [169], Heliobacillus mobilis [170], and the photosynthetic green-sulfur bacterium Chlorobium tepidum TLS [171]. The functional activity of the corresponding proteins in florfenicol resistance, however, has not been confirmed yet. Moreover, they only exhibit low levels of amino acid identity to the proteins FloR (8.5–29.2%) and FexA (9.1–21.9%) known to export Ff from the bacterial cell. In addition, whole genome sequencing revealed the presence of two genes coding for a 399 amino acid protein from Sinorhizobium meliloti [172] and a 385 amino acid protein from Mesorhizobium sp. strain BNC1 [173] which exhibited identities of 60% and 63%, respectively, and homologies of 75% to the FloR protein. But also for these two proteins, it is unknown whether they are involved in the efflux of Ff and Cm.

4.2.2 Multidrug transporters

In addition to specific exporters, a number of multidrug transporter systems have been identified whose substrate spectrum includes Cm and/or Ff. In general, the levels of Cm and Ff resistance mediated by multidrug transporters are lower than those mediated by specific exporters.

The AcrAB-TolC multidrug efflux system is able to export Cm and Ff at low levels (MICs of Cm and Ff of 4 mg l−1). Overproduction of this system, due to mutations at regulator loci, however, leads to clinical levels of resistance to Cm (MIC 16–32 mg l−1), Ff (MIC 32 mg l−1) and other antimicrobials by active efflux [174177]. Another multidrug transporter, MdfA, which also exports Cm has been identified in E. coli [178]. It shows 96% amino acid identity to the E. coli protein Cmr, a 12 TMS protein of 411 aa which specifies a Cm efflux pump [179]. Multidrug transporters whose substrate spectrum include Cm have also been described in P. aeruginosa. Similarly to the AcrAB–TolC system in E. coli, these multidrug transporters are also composed of three components, a protein of the resistance/nodulation/cell division family (MexB, MexD or MexF), a membrane fusion protein (MexA, MexC or MexE) and an outer membrane protein (OmpM, OmpJ or OprN), which interact cooperatively to enable export of the drugs [159, 180]. Multidrug transporters of a similar structure which can also export Cm have been identified in Burkholderia cepacia (CeoAB–OpcM) and P. putida (ArpAB–ArpC; TtgAB–TtgC) [180]. Overexpression of most of these multidrug transporter systems led to a distinct increase of the MIC of Cm [174, 177, 181] whereas functional deletion resulted in a distinctly more susceptible phenotype [182]. It should be noted that several types of multidrug transporters may be present in the same bacterial strain and that specific transporters may occur side-by-side with multidrug transporters. Lee et al. investigated the effects of simultaneous expression of several efflux pumps, including specific exporters such as CmlA and multidrug transporters such as MdfA, AcrAB–TolC or MexAB–OprM, and observed additive as well as multiplicative effects on Cm resistance of E. coli and P. aeruginosa [177].

Some multidrug transporters from Gram-positive bacteria, such as NorA from S. aureus [183] or Blt from B. subtilis [184] were reported to be able to export Cm. However, studies on strains expressing the gene norA at elevated levels showed that their MICs of Cm and Ff were in the same low range as those of strains not carrying the gene norA [Schwarz, S. and Kehrenberg, C., unpublished data]. This observation suggested that carriage of the gene norA is most probably not a relevant factor in Cm and Ff resistance in staphylococci.

4.3 Other resistance mechanisms

Besides inactivation by acetylation, there are other ways to inactivate Cm. Some of which, such as dehalogination, glucuronidation, and reduction of the nitro group, are usually seen during biotransformation in hepatocytes of humans and animals [4], but have not been identified in bacteria. Other mechanisms such as O-phosphorylation [162] and hydrolytic degradation of Cm to p-nitrophenylserinol [185] are seen in the Cm producer S. venezuelae ISP5230. These latter mechanisms seem to have a self-defense function in the antibiotic producer. Recently, the 3-O-phosphotransferase was crystallized and its X-ray structure was determined [186].

Non-enzymatic Cm resistance mechanisms based on permeability barriers have been described in various bacteria. The loss of an outer membrane protein was considered to play a role in Cm resistance of H. influenzae strains which did not exhibit CAT activity [187]. Cm resistance due to decreased permeability of the outer membrane was also observed in B. cepacia [188]. The absence of a 50 kDa porin in Tn1696-carrying strains may also enhance Cm resistance [189]. In S. Typhi, the lack of the OmpF protein, which is required for the entry of Cm into the bacterial cell was found to result in high level resistance to Cm [190]. The mar locus which is present in bacteria of many enterobacterial genera, has also been reported to contribute to Cm resistance of E. coli. The transcriptional activator MarA is able to activate the gene micF which produces an antisense RNA that effectively inhibits ompF translation [191].

Mutations in the major ribosomal protein gene cluster of E. coli [192] and B. subtilis [193] as well as in the 23S rRNA gene of E. coli [194] are known to confer resistance to Cm. However, in contrast to resistance to other protein biosynthesis inhibitors, e.g. macrolide–lincosamide–streptogramin antibiotics [195], Cm resistance as a consequence of target site mutation/modification is rarely seen. One plausible suggestion for this observation is that structural changes at the peptidyl transferase center that might prevent Cm binding are incompatible with satisfactory ribosome function [196].

Finally, a novel gene, cfr, which mediates resistance to Cm and Ff by a yet unidentified mechanism has recently been detected on plasmid pSCFS1 from S. sciuri [197]. The corresponding gene product shows no homology to any of the so far known Cm resistance proteins, does not inactivate Cm or Ff, and also does not display transmembrane topology. Structural comparisons revealed a certain degree of similarity with Fe–S-binding oxidoreductases of the MoaA/NifB/PqqE family. Two domains were detectable: the N-terminal domain represented a putative Cys-rich Fe-S binding sequence (CISSQCGCNFGCKFC), whereas the C-terminal domain might contain a NAD-binding Rossman fold. However, the target site of Cfr as well as the Cfr-mediated mechanism of resistance remain to be determined. The MICs of Cm and Ff of the S. sciuri strain carrying pSCFS1 were 32 and 64 mg l−1 and could be increased after induction to 64 and 512 mg l−1, respectively. A potential regulatory region which resembled a translational attenuator was detected immediately upstream of the cfr reading frame [197].

5 Dissemination, co-selection and persistence of chloramphenicol resistance genes

There are few studies which describe the distribution of specific cat or exporter genes within bacterial populations. A study on 28 Cm resistant E. coli isolates from cattle, swine and poultry identified not further specified catA genes in 68% and cmlA1-like genes in 36% of the isolates whereas floR genes were not detected at all [198]. A study on the prevalence of floR and cmlA genes among 48 bovine E. coli isolates showed that floR genes were present alone in 71%, cmlA genes alone in 5% and a combination of floR and cmlA in 10% of the isolates [133]. In contrast, 89.6% of E. coli strains from swine carried the gene cmlA alone and another 8.3% in combination with a cat gene of group A-2, while the gene floR was detected only in a single isolate (2.1%) [134]. Another study identified the cassette-borne Cm resistance gene catB2 in 31 (10%) of 313 motile aeromonads from a fish farming environment. In all cases the catB2 cassette was located in a class 1 integron that also contained cassettes for resistance to streptomycin and trimethoprim [199]. A catA gene similar to that of plasmid pC194 was detected in 44 of 100 erythromycin resistant S. pneumoniae isolates [200], whereas this gene could only be detected in single streptococcal isolates of serogroups A or B as well as in a single Enterococcus faecium isolate [84]. In contrast, catA genes of group A-7 were detected in streptococci of serogroups A, B or G and in E. faecium, catA genes of group A-8 in a single serogroup B Streptococcus and in four E. faecalis isolates and catP as well as catQ genes in single serogroup B streptococci [84]. Such a heterogeneity of cat genes has not been detected in staphylococci. Plasmid-borne catA genes of groups A-7 and A-8 have been detected at almost similar frequencies among S. aureus isolates from bovine mastitis [201], S. lentus isolates from mink [202] and staphylococci from pigeons [203], whereas catA genes of group A-7 genes were the dominant genes in S. hyicus from pigs [204] and S. intermedius from dogs [205, 206].

Many of the genes coding for any of the two types of cat genes or specific exporters are located on mobile genetic elements, such as plasmids, transposons or gene cassettes (Tables 13). The location of the resistance gene on a mobile element is an important prerequisite for fast and efficient distribution among bacteria of the same or different genera and species. Other requirements are an appropriate selective pressure imposed by the use of antimicrobial agents as well as the close contact between donor and recipient bacteria which is commonly available in the mixed populations on the skin or the mucosal surfaces of the intestinal, respiratory or genito-urinary tract of humans and animals. In this regard, the selective pressure does not have to be based only on the use of Cm or Ff, but may also arise from the application of various other antimicrobials. The reason for this is that plasmids carrying cat genes of type A (e.g. those assigned to groups A-1–A-3) frequently also carry several additional genes which code for resistance not only to other antimicrobial agents [31, 32, 34, 3840], but also to heavy metals such as mercury [31] or tellurium [32]. In several plasmids from Pasteurella and Mannheimia strains [38, 39] and an uncultured bacterium [40] the catA3 gene was found to be integrated between the sulfonamide resistance gene sul2 and the streptomycin resistance gene strA to form a multiresistance gene cluster. Based on sequence comparisons, this integration event was considered as an illegitimate recombination between an R387-like plasmid carrying catA3 and an RSF1010-like plasmid carrying sul2 and strA [39]. Since there were no promoters upstream of catA3 and strA, it was assumed that all three genes were transcribed from a common promoter upstream of sul2 [38, 39]. Persistence of a catA3 gene in a multiresistance gene cluster represents a plausible explanation for the persistence of a Cm resistance gene for more than eight years after the ban of Cm use in cattle.

Some of the aforementioned plasmids are conjugative [31, 32] and thus can arrange their self-transfer to new host cells. Smaller plasmids like pMHSCS1 and pMVSCS1 from Mannheimia spp. [38, 39] and pIE1130 from uncultured eubacterium [40] are mobilizable. Transformation experiments also showed that the Mannheimia plasmids pMHSCS1 and pMVSCS1 can replicate and express their resistance properties even in distantly related hosts such as E. coli [38, 39]. When plasmids mediating resistance to Cm or Cm/Ff are transferred from one bacterial host to another, they are not always able to replicate in the new host. Besides a general inability to replicate in the new host, restriction-modification systems present in the new host, but also incompatibilities with plasmids already resident in the new host may prevent efficient replication of the new plasmid. Recombination between the new plasmid and the plasmids resident in the new host is an effective way to circumvent these replication problems. Such recombinations may lead to the formation of novel resistance plasmids which carry the resistance genes of both parental plasmids and are well adapted with regard to their replication in the respective host. Such recombination events have not only been observed in the plasmids from Mannheimia spp., but are also considered to play a role in the formation of small staphylococcal plasmids that carried a pC221-like catA gene and a pS194-like streptomycin resistance gene [6870, 207]. The widespread occurrence of pC221-like catA genes in large conjugative multiresistance plasmids such as pIP501 [57] is most probably also based on recombinational events. The fact that broad host range plasmids such as pIP501 can replicate in a wide variety of Gram-positive bacteria [71] furthers the spread of the respective catA gene. The host range of pIP501 was reported to include bacteria of the genera Streptococcus, Enterococcus, Staphylococcus, Clostridium, Listeria, and Pediococcus [71]. The spread of Tn4451– or Tn4453-associated catP genes from Clostridium spp. to N. meningitidis [8183] represents a serious health threat since Cm is still the standard therapy for meningococcal meningitis in developing countries [81]. This observation illustrates that transposon-borne cat genes can also be spread across species and genus borders.

Since catB genes have often been identified to be part of gene cassettes which are located on chromosomal or plasmidic multiresistance integrons, the co-selection issue is also of major importance. The same is true for cmlA genes which are also cassette-borne genes and are most frequently found within plasmidic multiresistance integrons. In this regard, a large number of multiresistance integrons have been identified in a variety of Gram-negative bacteria. The observation that virtually the same cat gene (e.g. those of groups B-2 or B-3) or exporter gene (e.g. those of group E-1) have been identified in different bacteria such as Salmonella, Escherichia, Klebsiella and Pseudomonas underlines the efficient transfer of these cassette-borne genes. Other exporter genes, e.g. those of group E-3, have only been identified either on plasmids [150, 152, 154, 155] or in the chromosomal DNA where they represented part of a multiresistance gene cluster of SGI1 in S. enterica serovars [145149] and part of the SXT element in V. cholerae [158, 159]. The SXT element of V. cholerae is a conjugative self-transmissible chromosomally integrating element of about 100 kb, and has been related to integrative conjugative elements (ICEs) [208, 209]. Thus, SXT might contribute to the spread of the Cm/Ff resistance gene floR by conjugal transfer to a variety of Gram-negative bacteria. The identification of SGI1 or variants of SGI1 carrying the floR gene reported for S. enterica serovars Typhimurium DT104 and DT120, Agona, Paratyphi B, Albany and Newport (Fig. 6) indicates the horizontal transfer potential of SGI1 [145150]. Moreover, the ant(4)-IIb gene of P. aeruginosa strain BM4492 has been identified recently as part of a chromosomal multiresistance gene cluster homologous to that of SGI1 which comprises the resistance genes floR and tetR-tet(G) [210]. The horizontal transfer of SGI1 remains to be experimentally confirmed, but could probably occur in a similar way as the conjugal transfer of SXT. Thus, the floR gene spreads together with other antibiotic resistance genes by means of large mobile genomic islands such as SXT and SGI1 that integrate into the chromosome and may remain stable afterwards. A transposable element has also been assumed to be involved in the spread of floR since the floR genes detected on plasmids of E. coli and K. pneumoniae, but also on the SXT element, were flanked by transposase-like reading frames [152, 155] and/or transposon relics [154, 158, 159] (Fig. 7). Again, all these floR-carrying plasmids and chromosomal elements also carried at least one additional resistance gene. Recently, conjugative plasmids carrying a floR gene and the extended spectrum cephalosporin resistance gene blaCMY−2 were detected in canine E. coli strains [211], S. Typhimurium and S. Newport strains from humans and retail foods [212] and S. Newport from various animal and human sources [213].

In summary, the spread of cat genes or exporter genes mainly depends on the genetic element on which the respective gene is located. Based on the transfer abilities of this element, a distribution of the Cm or Cm/Ff resistance gene across species and genus borders is mediated by either conjugation, mobilization, transduction or transformation. Another important aspect is the presence of additional resistance genes on the same mobile genetic element. Since other resistance genes are commonly found on plasmids, transposons or integrons harbouring cat genes or exporter genes, co-selection of Cm or Cm/Ff resistance genes needs to be taken into account when discussing the spread and the persistence of cat genes or genes coding for specific exporters.

6 Summary and conclusions

Even though the use of Cm has been reduced widely in human medicine in the industrial countries, it is extensively used in developing countries throughout the world for the treatment of severe diarrhea and pneumonia [214]. This practice is continued since Cm is inexpensive, still effective against a wide range of bacteria and does not require specific conditions for storage [214]. In developing countries, antimicrobials, including Cm, are commonly available to consumers by over-the-counter sale in pharmacies without prescription by medical doctors. Self-medication and misuse of these drugs favours the development of resistance [215]. Thus, the major force to select for or maintain Cm resistance in bacteria from humans, namely the widespread use of Cm, is still present in those countries. Another important aspect, which, however, is not only restricted to developing countries, is the use of Cm for prophylaxis or therapy in food-producing animals including aquaculture. The regulation of Cm application to food animals in other parts of the world – if available at all – may differ distinctly from the strict limitations of the use of Cm in food-producing animals in the EU or Northern America [216]. Import controls, done on the basis of random samples, usually refer to Cm residues still present in carcasses, but not to the presence of Cm resistant bacteria. Thus, travel abroad, global trade or open markets may lead to the “import” of Cm resistant bacteria resident in or on humans, animals or food of animal origin. A recent study from Denmark [217] showed striking differences in the percentages of Cm resistance in S. Typhimurium isolated in Danish pork (22%) versus imported pork (61%), but also from human cases acquired domestically (14%) versus associated with travel abroad (42%). This observation underlined the role of imported food and travelling abroad as relevant factors in the spread of Cm resistant strains. A third aspect is the use of Cm in pets and other non-food-producing animals which live in close contact to humans in most industrial countries. Pet and companion animals may represent a reservoir of resistant strains. Transmission of Staphylococcus strains between pets and humans has been demonstrated [218, 219] and in one case indistinguishable catA-carrying plasmids have been identified from S. epidermidis strains of human and canine origins suggesting the exchange of such plasmids [220]. Since many of the Cm resistance determinants reside on mobile genetic elements (Tables 13), which most often carry additional resistance genes, co-selection and persistence of Cm or Ff resistance genes may occur, even if there is no selective pressure imposed by the use of Cm or Ff. Furthermore, the linkage of resistance genes with virulence genes may also occur and explain why resistance genes persist in the absence of selection pressure [221].

New Cm derivatives, which are active against bacterial strains that harbour any of the known resistance mechanisms, are not currently under development, either for use in human or in veterinary medicine. Ff is one of the most recent antimicrobial agents exclusively licensed for use in animals. For use in cattle and pigs, Ff is only approved as an injectible drug for individual animal therapy. This minimizes the risk of underdosing – as compared with medication via feed and water – and thus eliminates a major factor that contributes to the resistance development in the veterinary field. However, the use of Ff also resulted in the detection of new resistance genes, such as floR, cfr or fexA. Since these genes also mediate Cm resistance, it is most likely that they have been developed prior to the introduction of Ff into veterinary use. Nevertheless, increasing veterinary use of Ff bears the possibility that all those determinants which also mediate resistance to fluorinated Cm analogs will spread or may give rise to new Cm/Ff or Ff resistance determinants possibly specifying new mechanisms. Therefore, prudent use of highly potent antimicrobials, such as Ff, in veterinary medicine is strongly required to pertain the efficacy of Ff for the future.

All these aspects show that the problem of bacterial resistance to Cm and Ff is multifaceted in its origins and manifestations. Knowledge of (i) the resistance genes present in the different bacteria, (ii) the mobile elements on which these resistance genes are located, (iii) the transfer capacities and host range of these mobile elements and (iv) the organisation of the Cm and Cm/Ff resistance genes in multiresistance integrons and gene clusters as summarized in this review will provide a substantial database for the understanding of the spread and the persistence of Cm and Cm/Ff resistance genes, but also the co-selection of genes conferring resistance to other classes of antimicrobial agents, such as extended spectrum cephalosporins, which are even more relevant than Cm or Ff for therapeutic interventions in human and veterinary medicine.


S.S. and C.K. were supported by grants of the Deutsche Forschungsgemeinschaft (SCHW 382/6-1, SCHW 382/6-2). The authors would like to thank William V. Shaw for helpful discussions.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
  77. [77].
  78. [78].
  79. [79].
  80. [80].
  81. [81].
  82. [82].
  83. [83].
  84. [84].
  85. [85].
  86. [86].
  87. [87].
  88. [88].
  89. [89].
  90. [90].
  91. [91].
  92. [92].
  93. [93].
  94. [94].
  95. [95].
  96. [96].
  97. [97].
  98. [98].
  99. [99].
  100. [100].
  101. [101].
  102. [102].
  103. [103].
  104. [104].
  105. [105].
  106. [106].
  107. [107].
  108. [108].
  109. [109].
  110. [110].
  111. [111].
  112. [112].
  113. [113].
  114. [114].
  115. [115].
  116. [116].
  117. [117].
  118. [118].
  119. [119].
  120. [120].
  121. [121].
  122. [122].
  123. [123].
  124. [124].
  125. [125].
  126. [126].
  127. [127].
  128. [128].
  129. [129].
  130. [130].
  131. [131].
  132. [132].
  133. [133].
  134. [134].
  135. [135].
  136. [136].
  137. [137].
  138. [138].
  139. [139].
  140. [140].
  141. [141].
  142. [142].
  143. [143].
  144. [144].
  145. [145].
  146. [146].
  147. [147].
  148. [148].
  149. [149].
  150. [150].
  151. [151].
  152. [152].
  153. [153].
  154. [154].
  155. [155].
  156. [156].
  157. [157].
  158. [158].
  159. [159].
  160. [160].
  161. [161].
  162. [162].
  163. [163].
  164. [164].
  165. [165].
  166. [166].
  167. [167].
  168. [168].
  169. [169].
  170. [170].
  171. [171].
  172. [172].
  173. [173].
  174. [174].
  175. [175].
  176. [176].
  177. [177].
  178. [178].
  179. [179].
  180. [180].
  181. [181].
  182. [182].
  183. [183].
  184. [184].
  185. [185].
  186. [186].
  187. [187].
  188. [188].
  189. [189].
  190. [190].
  191. [191].
  192. [192].
  193. [193].
  194. [194].
  195. [195].
  196. [196].
  197. [197].
  198. [198].
  199. [199].
  200. [200].
  201. [201].
  202. [202].
  203. [203].
  204. [204].
  205. [205].
  206. [206].
  207. [207].
  208. [208].
  209. [209].
  210. [210].
  211. [211].
  212. [212].
  213. [213].
  214. [214].
  215. [215].
  216. [216].
  217. [217].
  218. [218].
  219. [219].
  220. [220].
  221. [221].
View Abstract