Previous Article | Next Article ![]()
Journal of Clinical Microbiology, January 2005, p. 168-173, Vol. 43, No. 1
0095-1137/05/$08.00+0 doi:10.1128/JCM.43.1.168-173.2005
Copyright © 2005, American Society for Microbiology. All Rights Reserved.
Department of Environmental Infectious Disease, Graduate School of Medical Science,1 Department of Microbiology, School of Medicine, Kitasato University, Sagamihara, Kanagawa,4 Department of Otolaryngology, Tohoku University School of Medicine, Sendai, Miyagi,2 Department of Otolaryngology, Nagasaki University School of Medicine, Nagasaki, Japan3
Received 14 August 2003/ Returned for modification 8 October 2003/ Accepted 24 September 2004
|
|
|---|
|
|
|---|
In S. pneumoniae, macrolide resistance is commonly caused by two major mechanisms: the efflux pump and target modification. The efflux pump, encoded by mefA, is associated only with resistance to 14- and 15-membered macrolides (19). On the other hand, target modification due to methylase, encoded by ermB, confers macrolide-lincosamide-streptogramin B (MLSB) resistance (10). Among Japanese clinical isolates of S. pneumoniae, 42.7, 52.7, and 3.3% of strains have mefA, ermB, or both genes, respectively (6).
It has been reported that two transposons, Tn1545 (22) and Tn917 (17), carry the ermB gene in S. pneumoniae. Tn1545 is almost identical to pAM77 from Streptococcus sanguinis (98% identity) and also mediates resistance to tetracycline via tetM (12) and resistance to kanamycin via aphA-3 (4). In contrast, Tn917 has no resistance genes for tetracycline or kanamycin and was identified on the nonconjugative multiple-resistance plasmid pAD2 in Enterococcus faecalis DS16 (17).
In this study, we investigated macrolide resistance among S. pneumoniae isolates in Japan with respect to (i) the relationship between transposon Tn1545 or Tn917 and resistance to macrolides, (ii) the question of whether macrolide resistance is inducible or constitutive, and (i) regional differences in the prevalence of the transposons in Japan.
|
|
|---|
Antimicrobial agents. Reference powders of known potency of the following antimicrobial agents were used: penicillin G (Meiji Seika Kaisha, Ltd., Tokyo, Japan), erythromycin (Shionogi Pharmaceutical, Osaka, Japan), azithromycin (Pfizer Laboratories, Gorton, Conn.), rokitamycin (Asahi Kasei, Tokyo, Japan), and clindamycin (Upjohn, Tokyo, Japan). All of these antimicrobial agents were kind gifts from the respective manufacturers. Telithromycin (Nippon Hoechst Marion Roussel, Tokyo, Japan) was also used.
Determination of MICs. MICs were tested by using the six antibiotics mentioned above at concentrations between 0.06 and 128 mg/liter. MICs were determined by the twofold agar dilution method using sensitivity test agar (Mueller-Hinton agar medium; Eiken Chemicals, Tokyo, Japan) with 8% Strepto Haemo supplement (SHS; Eiken Chemicals). Bacteria were cultured overnight at 35°C in sensitivity test broth (Eiken Chemicals) supplemented with 8% SHS, after which the culture was diluted to a final concentration of 5 x 107 CFU/ml with buffered saline containing gelatin. The bacterial suspensions were then plated with an inoculator (Sakuma Seisaku, Tokyo, Japan) at an inoculum size of 5 x 104 CFU/spot onto agar plates containing various concentrations of a test drug. The plates were incubated for 18 h at 35°C, and the MIC was defined as the lowest drug concentration that prevented visible growth of bacteria.
Detection of resistance genes and analysis of the regulatory regions of the ermB gene. The presence of macrolide resistance genes was investigated by PCR using a commercially available kit (Gene Amp PCR kit with AmpliTaq DNA polymerase; Takara, Kyoto, Japan) and a model PH2000 DNA thermal cycler (Perkin-Elmer Cetus Instruments, Emeryville, Calif.). The primer set for the internal region of ermB comprised 5'-GAAAAGGTACTCAACCAAATA-3' and 5'-AGTAACGGTACTTAAATTGTTTAC-3', while that for the internal region of mefA comprised 5'-AGTATCATTAATCACTAGTGC-3' and 5'-TTCTTCTGGTACTAAAAGTGG-3' (18). Primers 5'-ATCTGACGGTGACATCTCTC-3' and 5'-GGTTGAGTACCTTTTCATTCGTTAA-3' for Tn917 (17), and primers 5'-CTTAGAAGCAAACTTAAGAGTGTGT-3' and 5'-GGTTGAGTACCTTTTCATTCGTTAA-3 for Tn1545 (22), were designed to amplify a fragment including the regulatory region and 48 bp of the 5' end of ermB. PCR was carried out by 35 cycles of amplification consisting of 1 min of denaturation at 94°C, 1 min of annealing at 52°C, and 1 min of elongation at 72°C, followed by heating at 72°C for 7 min. PCR products were analyzed by using a DNA sequencing kit (ABI PRISM Big Dye Terminator cycle sequencing, FS; Perkin-Elmer Applied Biosystems, Tokyo, Japan). S. pneumoniae BM4200 containing Tn1545, kindly provided by P. Courvalin (Institut Pasteur, Paris, France) (22), and E. faecalis DS16 containing Tn917 (17), kindly provided D. B. Clewell (University of Michigan, Ann Arbor, Mich.), were used as positive controls. Reaction products were precipitated with ethanol-potassium acetate, dissolved in template suppression reagent, and run on an ABI PRISM 310 sequencer. As a result, the organisms were separated into two groups: a Tn917 group and a Tn1545 group.
Analysis of reduction of telithromycin susceptibility by disk diffusion. The disk diffusion test was performed (9) using six strains selected from both groups based on the MICs for erythromycin and rokitamycin. Against the six strains, the MICs of erythromycin and rokitamycin were both >128 mg/liter (two strains), 4 and 8 mg/liter (two strains), and 4 and 1 mg/liter (two strains), respectively. Two milliliters of the bacterial suspension (approximately 108 CFU/ml) was spread over the surface of 10 ml of sensitivity test agar containing 8% SHS. After excess bacterial suspension was removed, paper disks (diameter, 8 mm; Tokyo Roshi Kaisha, Tokyo, Japan) impregnated with erythromycin (20 µg/disk), azithromycin (20 µg/disk), rokitamycin (20 and 100 µg/disk), or telithromycin (10 µg/disk) were placed on the surface of each agar plate. Then the plates were incubated overnight at 35°C. The presence of induced telithromycin resistance was assessed based on the shape of the zone of inhibition around the telithromycin disk (whether the so-called D-shape was observed).
Serogroup and genotyping analysis. The isolates were cultured in Mueller-Hinton broth (Eiken Chemical Co.) with 8% SHS at 35°C for 18 to 24 h; they were then harvested by centrifugation at 3,000 x g for 10 min. The serogroup of each strain was determined by using antisera for S. pneumoniae (Seiken Antisera; Denkaseiken, Tokyo, Japan). The isolates were genotyped by pulsed-field gel electrophoresis (PFGE), as reported previously (23), after strains from both groups were selected based on the serogroup and the geographical location of isolation in addition to the MIC for the macrolides.
Nucleotide sequence accession number. The nucleotide sequence data obtained in this study are available in the GenBank/EMBL nucleotide database under accession number AB111455.
|
|
|---|
![]() View larger version (11K): [in a new window] |
FIG. 1. Schematic representation of the reported regulatory region of the ermB gene on Tn917 or Tn1545, and the regions found in macrolide-resistant S. pneumoniae in this study. SD1 and SD2, ribosome-binding sites for the leader peptide and ermB gene, respectively; 10 and 35, putative promoter regions according to references 7 and 8. Deletions are indicated by dashes.
|
Susceptibility.
The MIC data for the macrolide-resistant S. pneumoniae strains with different transposons are displayed in Table 1. The MICs of erythromycin and rokitamycin were higher for strains from the Tn917 group than for strains from the Tn1545 group. In the Tn1545 group, the MICs of erythromycin and clindamycin for two strains with the ermB gene were
0.25 mg/liter, but other strains exhibited the MLSB phenotype. All of the isolates in the Tn917 group had the MLSB phenotype. Both groups showed resistance to azithromycin but were sensitive to telithromycin.
|
View this table: [in a new window] |
TABLE 1. In vitro activities of macrolides, clindamycin, telithromycin, and penicillin G against macrolide-resistant S. pneumoniae strains with different transposons
|
![]() View larger version (102K): [in a new window] |
FIG. 2. Phenotype of macrolide-resistant S. pneumoniae strains determined by the disk diffusion method. Abbreviations: E, erythromycin (20 µg/disk); A, azithromycin (20 µg/disk), R1, rokitamycin (20 µg/disk); R2, rokitamycin (100 µg/disk); T, telithromycin (10 µg/disk). (A) MICs (in milligrams per liter) were >128 for E, A, and clindamycin; 128 for R; and 0.25 for T. (B) MICs (in milligrams per liter) were 4 for E, >128 for A and clindamycin, 8 for R, and 0.13 for T. (C) MICs (in milligrams per liter) were >128 for E, A, and clindamycin; 128 for R; and 0.25 for T. (D) MICs (in milligrams per liter) were 4 for E, >128 for A, 1 for R, 128 for clindamycin, and 0.06 for T.
|
|
View this table: [in a new window] |
TABLE 2. The five most frequent serogroups in each group of macrolide-resistant S. pneumoniae strains
|
![]() View larger version (84K): [in a new window] |
FIG. 3. PFGE of S. pneumoniae chromosomal DNA after digestion with the SmaI restriction enzyme. (A) Strains with Tn1545. (B) Strains with Tn917. Lanes M, lambda ladder. For details, see Table 3.
|
|
View this table: [in a new window] |
TABLE 3. PFGE resultsa for macrolide-resistant S. pneumoniae
|
|
|
|---|
In this study, the major macrolide resistance element in Japanese isolates was Tn917 (or a Tn917-like element). Although it is not clear why Tn917 is spreading among Japanese S. pneumoniae strains, it is possible that strains with Tn917 are more resistant to 16-membered-ring macrolides than isolates with Tn1545, since macrolides with a 16-membered-ring (such as rokitamycin) are widely used in Japan.
We found the deletion of 182 bp of the Tn917 sequence, including the 35 region of the ermB promoter, in all strains with Tn917 isolated in Japan. Although the existence of a new promoter is conceivable, it could not be determined by computer simulation because we did not find a new promoter that was stronger than the putative original promoter. It is thought that these mutations have contributed to the increase in the production of the methylase, but those mechanisms have not been clarified. Although Oh et al. found a relationship between macrolide resistance and two mutations of the leader peptide in the ermB gene on Tn917 from E. faecalis by a reporter gene assay (14), they concluded that TAAA duplication at the T591 (T421 in our data) site (which was also detected in this study) generated a translation stop codon and shortened the leader peptide by 9 amino acids so that reporter gene expression was dramatically elevated, i.e., TAAA duplication increased expression of the methylase and resulted in strong resistance to rokitamycin.
Strains that had the ermB gene on either transposon (or on a transposon-like sequence) showed reductions in telithromycin susceptibility in the disk diffusion test. However, for strains with Tn1545 for which the rokitamycin MIC was lower, less telithromycin resistance was reduced by the 100-µg rokitamycin disk than for strains having Tn917 or Tn1545 for which rokitamycin MICs were high. In contrast, the isolates with Tn917 showed reductions in telithromycin susceptibility induced by the 100-µg rokitamycin disk irrespective of the rokitamycin MICs. Such a result cannot be explained clearly. It has been reported that the basal level of ribosomal methylation differs from strain to strain and that the nucleotide sequence of the regulatory region shows variation in each isolate, resulting in different levels of constitutive ermB gene expression (24). However, we found that all strains had the same leader sequence of the ermB gene, even if the macrolide MICs were different. Such differences may be specific to each strain and could be based on complex mechanisms (11).
Although all strains were initially sensitive to telithromycin, the susceptibility of telithromycin was decreased by erythromycin, azithromycin, and rokitamycin (9). Because these macrolides are not used together clinically, this effect may not be a problem. In addition, there were no strains with constitutive macrolide resistance.
The PFGE patterns and serogroup distribution showed that Tn917 spread both horizontally and clonally, while Tn1545 spread clonally. The results of this study suggest that Tn917 may have been transmitted from E. faecalis or other organisms, so that the dominant macrolide resistance gene is ermB on Tn917 in Japanese pneumococci.
Recently, other mechanisms of macrolide resistance (amino acid substitutions in domain V of 23S rRNA or mutations of L4 and L22 proteins) in vitro (3, 20) and in clinical isolates (2, 15, 21) have been reported often. Kaieda et al. reported an S. pneumoniae strain with an ermB mutation that showed in vitro constitutive macrolide resistance (S. Kaieda, S., H. Yano, N. Okitsu, Y. Hosaka, R. Okamoto, H. Takahashi, and M. Inoue, Abstr. 42nd Intersci. Conf. Antimicrob. Agents Chemother., abstr. C1-1580, p. 72, 2002). It was isolated along with inducible macrolide resistance strains by culture with a disk containing 0.5 mg of telithromycin/liter, and the MIC for telithromycin was 8 mg/liter. However, the isolation rate was very low (1010), so it is thought that the appearance of mutants with constitutive resistance to macrolides, including telithromycin, is very uncommon in the clinical setting.
It is important to continue the surveillance of antimicrobial susceptibility in S. pneumoniae and to carefully monitor the emergence of constitutive macrolide resistance and resistance due to ribosomal mutations in S. pneumoniae.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»