Drug Resistance Mechanisms of Mycoplasma pneumoniae to Macrolide Antibiotics

Throat swabs from children with suspected Mycoplasma pneumoniae (M. pneumoniae) infection were cultured for the presence of M. pneumoniae and its species specificity using the 16S rRNA gene. Seventy-six M. pneumoniae strains isolated from 580 swabs showed that 70 were erythromycin resistant with minimum inhibitory concentrations (MIC) around 32–512 mg/L. Fifty M. pneumoniae strains (46 resistant, 4 sensitive) were tested for sensitivity to tetracycline, ciprofloxacin, and gentamicin. Tetracycline and ciprofloxacin had some effect, and gentamicin had an effect on the majority of M. pneumoniae strains. Domains II and V of the 23S rRNA gene and the ribosomal protein L4 and L22 genes, both of which are considered to be associated with macrolide resistance, were sequenced and the sequences were compared with the corresponding sequences in M129 registered with NCBI and the FH strain. The 70 resistant strains all showed a 2063 or 2064 site mutation in domain V of the 23S rRNA but no mutations in domain II. Site mutations of L4 or L22 can be observed in either resistant or sensitive strains, although it is not known whether this is associated with drug resistance.


Introduction
Mycoplasma pneumoniae is a common pathogen of respiratory tract infection in children and adolescents and can cause serious pneumonia and external lung complications [1]. At present, the preferred treatment is macrolide antibiotics. In recent years, however, many countries have reported the isolation of clinically drug-resistant strains [2][3][4][5], the main mechanism of resistance being a mutation in the 23S rRNA gene which is the target of macrolide antibiotic action [6,7]. In China, there are a number of reports of clinically resistant M. pneumoniae strains [8,9], resulting in serious failure of effective antibacterial agents. In this study macrolideresistant M. pneumoniae strains isolated between 2003 and 2007 were screened for sensitivity to antibacterial activity of tetracycline, ciprofloxacin, and gentamicin. The strains were typed for domain II and V of 23S rRNA, and for the ribosomal protein L4 and L22 genes, which are also associated with macrolide resistance.

Culture and Identification.
The throat swab specimens were inoculated in M. pneumoniae liquid medium, mixed evenly, and placed in an incubator at 37 ∘ C, with 5% CO 2 for cultivation. Growth of M. pneumoniae causes decrease  Table 1). The PCR mixture for domain II of 23S rRNA with L4 and L22 contained 15 pmol in each of forward and reverse primers, 1U TaqDNA polymerase, 20 L total reaction volume, and 5 L DNA template. PCR conditions were initial denaturation for 2 min at 94 ∘ C, followed by 94 ∘ C for 45 s, 55 ∘ C for 1 min, 72 ∘ C for 30 cycles, and a final 5 min extension cycle at 72 ∘ C. The specific nested PCR method for domain V of 23S rRNA was as described in [2]. The amplified products were purified and subjected to full-auto DNA sequencing (ABI 3730XL sequenator from Shanghai Sangon Biological Technologies & Service Co., Ltd.). Resulting sequences were compared with the corresponding sequences of the standard strain M129 registered at NCBI.

PCR Amplification and DNA Sequencing.
Electrophoretograms of the four PCR products of genes targeted in this study are shown in Figures 1, 2, 3, and 4 (note: M represents the standard band, FH represents MP standard strain, N is the negative control, and digits refer to the strain number).
DNA sequencing results are shown in Table 2.   Domain V of 23S rRNA. 46 of the 50 previously cultured strains tested were resistant [11], and 40 of these strains showed the A2063G mutation, one strain showed an A2063C mutation, and five showed an A2064G mutation. Of the 26 recently cultured M. pneumoniae strains, 24 were resistant strains and they all showed the A2063G mutation. Six sensitive strains and the standard strain FH showed no mutations

Discussion
The emergence of drug resistant M. pneumoniae strains is seriously reducing the effectiveness of macrolide drugs and  affecting clinical outcome of M. pneumoniae infection in children [2,7]. Before the year 2000, very few macrolideresistant M. pneumoniae strains were isolated [7]. By contrast, since 2001, macrolide-resistant M. pneumoniae isolates were first reported by Japan [10], the frequency of macrolideresistant M. pneumoniae cases has increased annually in Japan: 5.0% in 2003, 30.6% in 2006, 59.1% in 2009, and 89.5% in 2011 [12]. Similarly, since our team firstly reported the appearance of resistant strains of M. pneumoniae in 2005 [13], the frequency of macrolide-resistant M. pneumoniae remained high in China, ranging from 84.4% to 100% (not published) in our lab as well as in other Chinese researchers [9,14]. In France, Pereyre et al. reported the emergence of M. pneumoniae drug resistant strains; only 2 of 155 showed resistance to macrolide isolated between 1994 and 2006 [11] but risen to 10% between 2005 and 2007 reported by Peuchant et al. [15]. In the other countries, the results were as follows: In USA, the frequency of macrolide-resistant M. pneumoniae was 8.2% during 2007 and 2010 [16], and in Italy that was 11 out of 43 in 2010 [17]; in Germany, 2 of 167 throat swabs were macrolide resistant collected between 2003 and 2008, reported by Dumke et al. in 2010, while during 1991 and 2009, only 3 of 99 isolation showed resistance [6], and so on. In conclusion, M. pneumoniae shows high resistance to macrolide antibiotics, especially in Asian, and it is rising annually, which should be taken into consideration by the world.
Macrolide antibiotics act by inhibiting bacterial protein synthesis. Studies have found that the target site for macrolide is the large (50S) subunit of the bacterial ribosome. Alterations in specific nucleotides within the 23S rRNA lead to decreased affinity between drug and ribosome. Many cases of macrolide resistance in clinical strains can be linked to mutations in the sites 2063, 2064, 2067, and 2617 in domain V of 23S rRNA [2,7,18]. In China, resistant strains have shown the 2063 and 2064 mutations but no mutations in any other sites [8,9]. Ribosomal proteins L4 and L22 are also associated with drug resistance. These large subunit proteins have an elongated "tentacle" structure extending to the core of the large subunits to form the partial inner wall of the peptide output channel. Mutations within this structure can obstruct the channel and affect the binding of macrolide antibiotics [19]. In 2004, Pereyre et al. reported that amino acid changes of ribosomal proteins L4 and L22 could be induced in vitro appearing as H70R or H70L replacement, and 1∼3 G insertion in site 60 in L4 as well as P112R and A114T replacement or 111 IPRA 114 deletion in L22 [20]. However, no mutations were found in domain II of 23S rRNA in this study and it is not known whether the mutations induced in vitro arise in clinical samples [20].
In this study, among 76 clinical M. pneumoniae strains, 70 resistant strains were found showing A2063G, A2063C, and A2064G in domain V of 23S rRNA. None of the strains had any mutations in domain II of 23S rRNA. In addition to these major mutations, some of the strains showed C58A, T66G, G81T, A140C, and A209T mutations of L4, and C62A and T65A mutations of L22, further causing the changes of encoded amino acids. Whether there was a relationship between drug resistance and these mutations will require further study. C62A site mutation of ribosomal protein L22 was also observed in the erythromycin-induced strains, suggesting that there would be the possibility of inducing resistant strains in vivo during the process of using macrolide antibiotics.
Compared with M129, some strains, including the standard FH strain, showed C162A and A430G mutations of L4 and T279C and T508C mutations of L22 (Table 2). These were associated with two types of M. pneumoniae strains classified according to the difference in P1 gene, M129 being typical of P1-I type, and FH typical of P1-II type [21]. These mutations were therefore considered to be independent of M. pneumoniae drug resistance to macrolide. However, all the strains including the FH showed the T508C mutation in L22, and the P1 gene is considered to be independent of M. pneumoniae resistance to macrolide, T66G mutation in the ribosomal protein L4 of strain number 221 was synonymous and considered to be unrelated to drug resistance. A140C mutation in L4 of the MEN30 strain causes a Q47P change, but as MEN30 is a sensitive strain, it is inferred that this change is independent of resistance to macrolide. The A209T mutation in the ribosomal protein L4 of number 371 strain causes a H70L change. Pereyre et al. reported in 2004 that the telithromycin-induced M. pneumoniae resistant strain (T32) also showed a H70L change in L4, the mutation of C2617A in 23S rRNA domain V, and a A114T change in L22 [20]. Meanwhile, strain No. 371, a sensitive strain, showed a 2617 site mutation in 23S rRNA domain's zone V which is directly related to drug resistance as well as to changes in L4. Therefore, it is thought that the H70L change in L4 is unlikely to be related to drug resistance, although this needs further investigation. The remaining strains showed C58A and G81T site mutations in L4 as well as C62A and T65A site mutations in L22. Although both cause amino acid changes, but they also have the 2063 or 2064 mutation in the 23S rRNA domain V which is directly related to macrolide resistance ( Table 2). Further study is needed to ascertain whether these mutations are associated with drug resistance. In this study, some of the clinically isolated strains showed a C62A mutation in the ribosomal protein L22 also found by Matsuoka et al. in Japan [22], suggesting that the use of macrolide drugs is likely to induce M. pneumoniae resistant strains.
In vitro drug sensitivity tests proved that the majority of children from whom the specimens were collected were harboring macrolide-resistant M. pneumoniae resistant strains.
Some of the infection in these children could be controlled with macrolide antibiotics, mainly because macrolide antibiotics have an antimicrobial effect in vivo and are also involved in immune regulation during the recovery process of the disease, giving some clinical effectiveness [22]. At the same time, M. pneumoniae pneumonia is considered to be a selflimiting disease. In this study, all specimens were collected from children in the wards, and most of them had been exposed to macrolide antibiotics before sampling. Sensitive bacteria would have been inhibited, and total bacterial load reduced. This obviously made it difficult to estimate the actual levels of resistance found in our patients. However, resistant M. pneumoniae was obviously present and suggests the necessary to be very vigilant in monitoring the phenomenon of M. pneumoniae resistance to macrolide antibiotics.
In this study, in vitro drug sensitivity tests were used to determine the antibacterial activities of tetracycline, ciprofloxacin, and gentamicin against M. pneumoniae. Tetracycline mainly inhibits bacterial protein synthesis, by binding the ribosomal 30S subunits and blocking the extension or hindering the release of the protein synthesized peptide chain. Ciprofloxacin acts on M. pneumoniae primarily through inhibition of its DNA gyrase, thus affecting DNA replication, transcription, and expression. Gentamicin is a type of aminoglycoside antibiotic that acts on ribosomal 30S subunits in bacteria, inhibiting the protein synthesis and damaging the integrity of the cell membrane, but it has a limited inhibitory effect on M. pneumoniae. In this study, we found that 2063 and 2064 mutations in domain V of 23S rRNA of 50S subunits did not affect the binding of tetracycline, ciprofloxacin, and gentamicin with M. pneumoniae and did not give rise to M. pneumoniae resistance to these drugs. The results showed that tetracycline and ciprofloxacin were generally effective on M. pneumoniae, and gentamicin was effective on the majority of M. pneumoniae strains, although it did have poor antibacterial activity against some of the M. pneumoniae strains.
Because there is a risk of adverse reactions with tetracycline, and the safety of ciprofloxacin patients below the age of 18 has not yet been established, neither of these drugs is suitable for use in children. Current research should therefore be focused on finding new tetracycline or quinolone drugs with a strong antibacterial activity but low side effects, so as to provide effective alternative choice of drugs for the treatment of resistant M. pneumoniae infections.
In summary, 76 M. pneumoniae strains were isolated and cultured, 70 of which were macrolide-resistant, being