Occurrence of macrolide-resistant Mycoplasma pneumoniae strains in Germany


Corresponding author and reprint requests: R. Dumke, Dresden University of Technology, Medical Faculty Carl Gustav Carus, Institute of Medical Microbiology and Hygiene, Fetscherstrasse 74, D – 01307 Dresden, Germany
E-mail: roger.dumke@tu-dresden.de


Clin Microbiol Infect 2010; 16: 613–616


In a total of 167 respiratory tract specimens from adult outpatients with confirmed Mycoplasma pneumoniae pneumonia, sampled between 2003 and 2008, and a further 99 isolates obtained from patients between 1991 and 2009 in Germany, M. pneumoniae was tested for macrolide resistance. Using PCR, real-time PCR and sequencing of the 23S rRNA gene, 1.2% of M. pneumoniae in the respiratory tract samples and 3.0% of the isolates were found to be resistant. The results indicate a limited but not negligible importance of macrolide-resistant M. pneumoniae in the population investigated, which requires the monitoring of macrolide susceptibility of isolates or the testing of respiratory samples by molecular methods.


Mycoplasma pneumoniae is one of the most common causes of community-acquired infections of the human upper and lower respiratory tract and can lead to severe and long-lasting interstitial pneumonia [1]. Owing to the fact that mycoplasmas are cell wall-less bacteria, macrolides are in most cases the first choice for antibiotic treatment of M. pneumoniae infections [1]. For young children especially, the alternatives, tetracyclines and new fluoroquinolones, are not recommended. Resistance of isolated strains to macrolides is based on point mutations in domain V of the 23S rRNA gene of M. pneumoniae. An A to G/C transition at position 2063 or 2064 of the gene (M. pneumoniae numbering) resulted in a high-level macrolide resistance of the M. pneumoniae strain, whereas a C to G or C to A mutation at position 2617 was associated with a lower resistance to macrolide antibiotics [2,3].

Recent occurrence of macrolide resistance was shown in studies from Japan in which a significant increase in resistant M. pneumoniae strains of up to more than 30% in the year 2006 was reported [4,5]. Results of two current studies from China showed rates of resistant isolates of 83% and 92%, respectively [6,7]. In contrast, in an investigation of macrolide resistance among 155 M. pneumoniae isolates obtained from patients between 1994 and 2006 in France only two resistant isolates were found [8]. Furthermore, macrolide resistance was detected in a relatively low number of M. pneumoniae-positive respiratory tract samples (five/100 isolates) in the United States [9]. A more comprehensive overview of world-wide macrolide resistance is hampered by the difficulty of cultivating M. pneumoniae strains, which has been successful in reference laboratories only. The use of molecular methods such as real-time PCR offers the possibility of a culture-independent characterization of the circulating M. pneumoniae strains, including their resistance [4,9]. Furthermore, previous reports on macrolide-resistant M. pneumoniae have mainly been concerned with respiratory tract specimens and isolates from children who were in most cases hospitalized. Although systematic investigations are not available, differences in the occurrence of macrolide-resistant M. pneumoniae strains between adult and paediatric patients cannot be excluded as preliminary results have suggested [5].

This is the first study to focus on investigating respiratory tract samples from adult pneumonia outpatients using molecular methods in order to detect the actual macrolide resistance in M. pneumoniae found in these specimens. In addition, the investigation of 100 isolates from patients of different ages was included to obtain a more comprehensive survey of macrolide-resistant M. pneumoniae over time in Germany.

Materials and Methods

Respiratory tract samples (bronchoalveolar lavage fluids, nasopharyngeal or pharyngeal swabs, sputa) were a German network (Community-Aquired Pneumonia NETwork, CAPNET) established to investigate the aetiology of community-acquired pneumonia. Briefly, 12 local clinical centres distributed over different regions of Germany cooperated with a number of general practitioners to systematically collect samples from patients with symptoms of pneumonia. Between 2003 and 2008, 167 respiratory tract samples of adult pneumonia outpatients tested positive for M. pneumoniae [10]. Independent of the specimens sampled within the CAPNET programme, M. pneumoniae isolates from clinical material obtained mainly from hospitalized pneumonia patients of different ages were included in the study. Isolation and propagation of 100 M. pneumoniae strains isolated between 1991 and 2009 from 99 patients (two strains were isolated from different locations of the bronchial system of one patient, see below) were carried out as described [11]. DNA from patient samples and from M. pneumoniae strains grown in Plenrophenmonia-like organism (PPLO) broth cultures (200 μL each) was extracted using the QIAamp DNA mini kit (QIAGEN, Hilden, Germany) according to the manufacturer’s instructions (the protocol for blood and body fluids called for elution volumes of 50 and 200 μL, respectively). DNA from patient samples was pretested using a real-time PCR approach targeting a conserved part of the gene coding for the main P1 adhesin of M. pneumoniae [12]. Isolated M. pneumoniae strains and strains in respiratory specimens confirmed as positive were grouped into the known subtypes and variants by sequencing a variable part of the repetitive element RepMP2/3 in the P1 gene of M. pneumoniae as described recently [10].

For partial amplification of the 23S rRNA gene in M. pneumoniae-positive respiratory tract samples, the primers (Biomers, Ulm, Germany) MN23SDVF and MN23SDVR (first amplification) were used as described [4]. Nested PCR with the primers MN23SDVFn (5′-GAC TGT TTA ACT AAA ACA CAA CTC TAT G-3′, position: 1777-1804) and MN23SDVRn (5′-CTA GAA GCA ACA CTC TTC AAT CTT C-3′, position: 2661-2637) was carried out according to standard procedures (annealing temperature: 60°C, 30 cycles) and resulted in an 885-bp product. The 23S rRNA gene region from DNA from patient isolates was amplified with the same primer combination. Specificity of the PCR was tested using DNA of the following reference strains: Klebsiella pneumoniae (ATCC 13883), Staphylococcus aureus (ATCC 25923), Staphylococcus epidermidis (ATCC 12228), Escherichia coli (ATCC 43895), Pseudomonas aeruginosa (ATCC 27853), Haemophilus influenzae (ATCC 49247), Chlamydophila pneumoniae (strain TW-183), Legionella pneumophila (ATCC 33152), Streptococcus pneumoniae (ATCC 6305), Mycoplasma genitalium (ATCC 33530), Mycoplasma salivarium (ATCC 23064), Mycoplasma orale (ATCC 23714), Mycoplasma hominis (ATCC 23114), Ureaplasma urealyticum (ATCC 27818), Ureaplasma parvum (ATCC 27815) and Mycoplasma penetrans (ATCC 55252), respectively. Furthermore, the DNA of 50 respiratory tract samples of M. pneumoniae-negative pneumonia patients was included.

An aliquot (1 μL) of the amplification product of the nested PCR (respiratory tract samples) and of the first amplification (M. pneumoniae strains) was used as template for a pre-screening real-time PCR for the detection of mutations at positions 2063 and 2064 of the 23S rRNA gene of M. pneumoniae. Amplification was carried out using primers MN23SDVFn and MN23SLCR (5′-GTA GTA TTC CAC CTT TCG CAT C-3′, position: 2183-2162) and probes MN23SLCP1 (5′-GTG AAG ACA CCC GTT AGG CGC AAC-FL, position: 2032-2055) and MN23SLCP2 (5′-LC640-GGA CGG AAA GAC CCC GTG-PH, position: 2057-2074). With a LightCycler 1.5 instrument (Roche, Mannheim, Germany), real-time PCR was performed in a final volume of 20.0 μL containing 6.6 μL of water (PCR grade; Roche), 2.4 μL of MgCl2 (25 mM, Roche), 2.0 μL of LightCycler FastStart DNA Master HybProbe mix (Roche), 2.0 μL of each primer (5 pmol), 2.0 μL of probes (2 pmol) and 1.0 μL of template. The capillaries were incubated under the following cycling conditions: pre-incubation at 95°C (10 min), 40 cycles of denaturation at 95°C (8 s), hybridization at 57°C (8 s) and elongation at 72°C (10 s). Melting curve analysis was carried out after heating from 55°C to 85°C (temperature transition rate: 0.1°C/s). Finally, the reactions were cooled to 40°C for 30 s. Data were analysed using the LightCycler software version 3.5 (Roche). Putative mutations at position 2063/2064 of the 23S rRNA gene of M. pneumoniae, as suggested by melting peak analysis, were confirmed by sequence analysis.

Antibiotic susceptibility to erythromycin (Sigma, St Louis, MO, USA) of M. pneumoniae isolates which were regarded as macrolide-resistant according to the real-time PCR and sequencing results was determined by microdilution with PPLO broth [13]. The MIC was defined as the lowest concentration of erythromycin preventing a colour change of the medium after an incubation time of 5–7 days.

Results and Discussion

The specificity of the nested PCR used to amplify the part of the region of the 23S rRNA gene carrying the known mutations associated with macrolide resistance in M. pneumoniae was tested with the DNA of seven mycoplasma species which can be found in humans and nine bacterial species that commonly cause pneumonia. Amplification products were obtained from M. penetrans and M. genitalium only, which can be expected from the high identity of the 23S rRNA genes of the closely related species M. pneumoniae and M. genitalium (96.7%). None of the 50 M. pneumoniae-negative DNA samples from pneumonia patients tested gave an amplification product. Despite the fact that M. penetrans and M. genitalium do not belong to the spectrum of frequent agents of community-aquired pneumonia, it should be noted that the nested PCR used was not developed for the primary detection of M. pneumoniae but is intended as a diagnostic approach for the further characterization of M. pneumoniae-positive samples, pre-screened by a sensitive method showing no cross-reaction to other relevant microorganisms.

In contrast, an amplification product from the nested PCR targeting the 23S rRNA was obtained from all 167 investigated respiratory tract DNA samples pretested as M. pneumoniae-positive by real-time PCR, confirming the sensitivity of the approach. Using an aliquot of the PCR product as template in the real-time PCR, the melting curve analysis of the LC Red 640-labelled detection probe demonstrated clear differences between amplicons carrying a mutation at position 2063 or 2064 and PCR products with an unchanged 23S rRNA gene of M. pneumoniae (Fig. 1). Whereas amplicons from macrolide-susceptible isolates showed a mean Tm of 64.5 ± 0.3°C (range, 63.6-64.8°C), an A2063G (n = 3) or A2064G (n = 1) mutation resulted in a mean Tm of the amplicon of 58.9 ± 0.2°C. Furthermore, the A2063C (n = 1) and A2064C transition (n = 1) can be discriminated by an increased Tm of 61.1 and 60.3°C, respectively. Overall, real-time PCR products suggesting a resistance-relevant mutation in the part of the 23S rRNA gene described, were obtained from three (3.0%) of the 99 clinical isolates of M. pneumoniae and from two (1.2%) of 167 M. pneumoniae-positive respiratory tract samples (Table 1).

Figure 1.

 Example of the detection of macrolide-resistant Mycoplasma pneumoniae isolates using real-time PCR amplifying part of the 23S rRNA gene, followed by melting peak analysis.

Table 1.   Occurrence of macrolide-resistant Mycoplasma pneumoniae isolates in respiratory tract samples of pneumonia patients in Germany (without multiple isolations from one patient)
Type of sampleNumber of samplesMacrolide-resistant M. pneumoniae isolates (%)Mutation in 23S rRNA geneIsolation/ sampling year
M. pneumoniae isolates993 (3.0)A2063G (2x)
1998, 2006
M. pneumoniae-positive respiratory tract samples1672 (1.2)A2063C

In all cases, the assumed transitions at positions 2063 and 2064 of the 23S rRNA gene were confirmed by sequencing. Interestingly, in two isolates obtained from the bronchoalveolar fluids from different segments of the bronchi of a hospitalized male 35-year-old pneumonia patient, an A2064G change was detected in one isolate and an A2064C change in the other. According to the typing scheme based on the sequence differences in the repetitive elements located in the main P1 adhesin [10], both strains were classified as subtype 1. This fact indicates that M. pneumoniae strains with different mutations in the 23S rRNA gene can be isolated from a single patient.

In previous studies, the MICs of erythromycin, the most widely used reference macrolide for the detection of M. pneumoniae mutants with base changes at position 2063 or 2064 of the 23S rRNA gene, ranged from >400 to 32 mg of erythromycin/L [4–9,14,15]. In the present study, erythromycin susceptibility testing using microdilution demonstrated that the four strains with the mutations at these positions are resistant, with MIC values between >200 and 100 mg of erythromycin/L (independent of an A2063/2064G or A2064C transition). In parallel, a limited number (n = 6) of strains without a change at position 2063 or 2064 were tested and MIC values between 0.006 and 0.012 mg of erythromycin/L were found which is in agreement with the data in other reports [4–9,14]. In contrast, a mutation at position 2617 has been shown to result in moderate resistance, with MIC values between 1 and 8 mg of erythromycin/L [3,4]. To assess the occurrence of this mutation, we sequenced the amplicons from all of the 167 M. pneumoniae-positive respiratory tract specimens included in this study. A mutation at position 2617 was not found in any of these cases, confirming the results of other reports which did not detect any, or only detected a small portion, of macrolide-resistant M. pneumoniae isolates with a mutation at position 2617 in comparison with strains with a change at position 2063 or 2064 [5,7,15]. The sequenced region of the 23S rRNA gene in all cases shows 100% identity with the corresponding region in M. pneumoniae (GenBank, BLAST algorithm) and can be clearly differentiated from the 23S rRNA gene of the most closely related species, M. genitalium (identity, 97%). This observation further confirms the specificity of the amplification described.

No clear association was observed between macrolide-resistant isolates or isolates in clinical materials and the subtype of M. pneumoniae (two resistant strains were classified as subtype 1 and three strains as subtype 2) which is in accordance with previous results from Japan and China [4,7,15] and revealed the occurrence of resistant strains in the two most common genotypes of M. pneumoniae.

In conclusion, the results of the present study demonstrate the occurrence, in limited numbers, of macrolide-resistant M. pneumoniae strains in recently sampled respiratory tract specimens from adult pneumonia patients and in isolates mainly from hospitalized patients, in Germany. The real-time PCR-based method developed is a reliable tool to detect the most common molecular cause of macrolide resistance in M. pneumoniae. This approach circumvents the difficult, insensitive and time-consuming isolation of M. pneumoniae from clinical materials and allows resistance screening in a high number of M. pneumoniae-positive respiratory tract samples. At the moment, the percentage of resistant strains in Germany does not justify a re-evaluation of the standard recommendations for treating community-acquired pneumonia [16]. However, monitoring isolated M. pneumoniae strains as well as M. pneumoniae-positive clinical materials from different patient groups seems to be necessary in order to recognize early changes in the antibiotic resistance pattern of this important agent of human respiratory tract infections.

Transparency Declaration

The work was supported by the BMBF network of competence (CAPNET). The authors declare that they have no conflicts of interest in relation to this work.