ORIGINAL STUDIES
Emerging Macrolide Resistance in Mycoplasma pneumoniae
in Children
Detection and Characterization of Resistant Isolates
Xiao Li, PhD,* T. Prescott Atkinson, MD, PhD,† James Hagood, MD,† Chris Makris, MD,†
Lynn B. Duffy, MT (ASCP),* and Ken B. Waites, MD*
Background: Epidemiologic data from Asia have documented the rapid
and extensive emergence of macrolide resistance in Mycoplasma pneu-
moniae. This drug resistance has also been documented in Europe and
recently in the United States, but there is very little information currently
available on its prevalence. A rapid technique to identify macrolide-
resistant M. pneumoniae is needed to guide management of patients with
community-acquired respiratory infections.
Methods: Culture and Minimum Inhibitory Concentration testing identi-
fied macrolide-resistant M. pneumoniae infection in 2 seriously ill hospi-
talized children with community-acquired pneumonia. A portion of the 23S
ribosomal RNA gene from 2 macrolide-resistant M. pneumoniae isolates
from these children as well as 4 laboratory-induced macrolide-resistant
strains was amplified by PCR, and the PCR products were sequenced to
identify mutations associated with macrolide resistance. A real-time PCR
assay was designed to identify 3 known mutations in the 23S rRNA gene
associated with macrolide resistance and applied to the clinical specimens
from which these isolates were obtained and to the bacterial isolates.
Results: Macrolide-resistant M. pneumoniae from both children were
found to carry an A2063G transition in the 23S rRNA gene previously
identified in resistant isolates from China, Japan, France, and recently in an
encephalitis outbreak in Rhode Island. Three laboratory-induced mutant
strains had an A2064G mutation whereas the other one had an A2063G
mutation. A real-time PCR assay successfully detected the macrolide-
resistant M. pneumoniae directly in clinical specimens and discriminated
them from wild-type isolates.
Conclusions: Macrolide-resistant M. pneumoniae can be associated with
prolonged severe respiratory infection in children. Real-time PCR offers a
rapid method of diagnosing macrolide resistance in community-acquired
respiratory infections due to M. pneumoniae.
Key Words: Mycoplasma pneumoniae, respiratory infection, polymerase
chain reaction, macrolide resistance, children
(Pediatr Infect Dis J 2009;28: 693– 696)
M ycoplasma pneumoniae infections can involve both the up-
per and lower respiratory tract and occur both endemically
and epidemically worldwide in persons of all ages.1-3 Although
Accepted for publication January 27, 2009.
From the Departments of *Pathology and †Pediatrics, University of Alabama at
Birmingham, Birmingham, Alabama.
Supported by grant NHLBI P01 HL073907-04 (to T.P.A.).
Address for correspondence: Ken B. Waites, MD, Department of Pathology,
University of Alabama at Birmingham, WP 230, 619 19th Street South,
Birmingham, AL 35249. E-mail: waiteskb@uab.edu.
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions
of this article on the journal’s Web site (www.pidj.com).
Copyright © 2009 by Lippincott Williams & Wilkins
ISSN: 0891-3668/09/2808-0693
DOI: 10.1097/INF.0b013e31819e3f7a
The Pediatric Infectious Disease Journal • Volume 28, Number 8, August 2009
most cases are of only mild to moderate severity, serious infections
requiring hospitalization sometimes occur. Clinical manifestations
of mycoplasmal respiratory disease are similar to those of other
bacteria and are rarely sufficiently distinct to allow precise diag-
nosis on clinical grounds. Culture is seldom used for laboratory
diagnosis since M. pneumoniae grows so slowly and requires
special growth media and expertise that is not widely available.
Measurement of IgM, IgG, and IgA antibodies in paired sera may
provide accurate diagnostic information, but this approach can be
unreliable since IgM antibody may persist for variable periods
after infection has resolved and may not be detectable in adults
with recurrent mycoplasmal respiratory infections.1 The PCR as-
say is becoming more widely used for detection of mycoplasmal
respiratory infections, but no commercially manufactured kits are
approved by the Food and Drug Administration for use in the
United States.1 Laboratory diagnosis of M. pneumoniae infection
is most often sought when there a serious illness requiring hospi-
talization, failure to respond to treatment with the drugs of choice,
and/or an outbreak of several cases in the same community.
M. pneumoniae is generally considered to be susceptible to
macrolides, tetracyclines, and fluoroquinolones. Among these,
macrolides are drugs of choice in pediatrics due to potential
toxicities of the other agents. Thus far, no naturally occurring
resistance to tetracyclines and fluoroquinolones has been reported.
However, macrolide-resistant strains have been documented since
1970s,1 and resistant mutants have also been generated in vitro,
which contain point mutations in the peptidyl transferase loop of
domain V of 23S rRNA, which reduces affinity of these agents for
the ribosomes.1
Clinical resistance to macrolides in M. pneumoniae has been
reported in Europe4 and has become common in Asia since 2000,
coincident with more widespread usage of azithromycin.5–9 Re-
ports from Japan between 2001 and 2006 found macrolide resis-
tance in 10% to 33% of M. pneumoniae isolates, all of which had
mutations in the 23S rRNA gene.5-9 Liu et al10 reported that 44 of
53 (83%) M. pneumoniae isolates from the lower respiratory tracts
of children in Shanghai, China had high-level resistance to mac-
rolides. A recent report from the United States indicated that 3 of
11 (27%) isolates obtained during an investigation of an enceph-
alitis outbreak in Rhode Island were macrolide-resistant.11 Data
from Japanese studies indicate that the macrolide resistance in M.
pneumoniae may have clinical significance in terms of diminished
response to treatment with drugs in this class.7
In the present study, we have characterized macrolide
resistance in M. pneumoniae isolates from 2 children hospital-
ized with pneumonia in 2005 in Birmingham, Alabama and 4
strains with laboratory-induced mutations in 23S rRNA. We
also developed and validated a real-time PCR assay designed to
detect the 3 most common mutations associated with macrolide
resistance.
www.pidj.com | 693
Li et al The Pediatric Infectious Disease Journal • Volume 28, Number 8, August 2009
PATIENTS AND METHODS
Patients
Case 1
A 9-year-old previously healthy boy developed fever,
cough, myalgias, and intermittent rash. Although his brother had
asthma, he had wheezed only occasionally and was not considered
to have asthma. He was diagnosed with Fifth disease by his
pediatrician 2 days after onset of symptoms and treated symptom-
atically. As symptoms persisted a chest radiograph was obtained
on day 5 of the illness. A left perihilar infiltrate was noted and he
was given amoxicillin/clavulanate. Three days later, he developed
fever to 103°F, a productive cough with posttussive emesis,
increased work of breathing, poor fluid intake, and hypoxia. At this
point, he was referred for admission. During hospitalization, he
continued to receive amoxicillin/clavulanate. For persistent symp-
toms and continuing oxygen requirement, azithromycin was added
on day 2. He received a second 5 day course of azithromycin as
symptoms and oxygen requirement persisted. Hypoxia resolved on
day 10 and he was discharged home. Laboratory testing for
streptococcal pharyngitis and respiratory viruses were negative. A
diagnosis of M. pneumoniae pneumonia was made based on
markedly elevated IgM antibody 2 positive sputum cultures and a
positive sputum PCR assay that were reported following discharge.
Case 2
A 9-year-old previously healthy girl was admitted for a 2
week history of cough and fever up to 103°F accompanied by
lethargy and anorexia. She received no antibiotic therapy initially
but was started on a 5-day course of azithromycin 4 days before
admission without improvement. Fever was improving on the day
of admission (Tmax 100°F), but room air oximetry in her pediatri-
cian’s office revealed an oxygen saturation of 89% and a chest
radiograph showed an interstitial pattern. She was referred for
admission and required oxygen supplementation at 5 liter per
minute by facemask to maintain saturations above 90%. Due to
increasing hypoxia and worsening respiratory status, she was
transferred to the intensive care unit for monitoring and care. Fiber
optic bronchoscopy with bronchoalveolar lavage (BAL) was per-
formed in an effort to identify a causative pathogen. Bronchoscopy
revealed marked inflammation of the airway but no other abnor-
malities were identified. She was discharged after 4 days of
hospitalization. All microbiology tests including PCR for Borde-
tella pertussis, and a viral respiratory panel were negative except
for an elevated IgM antibody against M. pneumoniae and 3
cultures and PCR assays performed on lung washings obtained by
BAL, which were reported as positive after discharge.
Culture and Antimicrobial Susceptibility Testing
Cultures were performed and microbroth dilution Minimum
Inhibitory Concentrations (MICs) for azithromycin, erythromycin,
and telithromycin were determined for 4 M. pneumoniae isolates
(1 isolate from case 1 and 3 from case 2) using procedures
described by Waites et al.12
DNA Preparation
M. pneumoniae DNA was extracted from subcultures or
clinical specimens first by centrifuging samples at 14,000g for 20
minutes at 4°C in a refrigerated minicentrifuge and then digesting
the pellets with 200 ␮L proteinase K (1 mg/mL) lysis buffer for 1
hour at 60°C. Proteinase K was inactivated by incubation at 95°C
for 10 minutes.
694 | www.pidj.com
Sequence Analysis of 23S rRNA Gene
Amplification of the 2063/2064 region of domain V of the
23S rRNA gene was performed by PCR using primers described
by Wolff et al10 and by detection primers MpnMR2063F and
MpnMR2063R (mentioned in Table, Supplemental Digital Con-
tent 1, http://links.lww.com/A1302) under PCR conditions described
later. Amplification of the 2617 region of 23S rRNA gene was performed
using detection primers MpnMR2617F and MpnMR2617R
(mentioned in Table, Supplemental Digital Content 1,
http://links.lww.com/A1302). Amplicons were purified using Qiagen
QIAquick Gel Extraction Kit according to manufacturer’s instructions
and sequenced.
Real-Time PCR Assay for Macrolide Resistance
A multiplex real-time PCR assay was developed to detect mac-
rolide-resistant M. pneumoniae directly in clinical specimens. Primers
MpnMR2063F and MpnMR2063R defined a 224 bp amplicon con-
taining 2063/2064 position, which was recognized by probes
MpnMR2063P1 and MpnMR2063P2. Primers MpnMR2617F and
MpnMR2617R amplified a 151 bp sequence containing 2617 site that
was signaled by probes MpnMR2617P1 and MpnMR2617P2 (primer
and probe sequences are available in Table, Supplemental Digital
Content 1, http://links.lww.com/A1302). PCR was performed using
the Roche LightCycler 2.0 and PCR master mix contained: 0.2 ␮M of
MpnMR2063F and MpnMR2617F, 0.5 ␮M of MpnMR2063R and
0.6 ␮M of MpnMR2617R, 0.2 ␮M of each MpnMR2063 probe,
0.3 ␮M of each MpnMR2617 probe; 0.025 U/␮L of uracil-DNA
glycosylase, 2 mM of MgCl2, 4 ␮L of 5X Multiplex DNA Master
HybProbe buffer (Roche Diagnostics). To reach a total reaction
volume of 20 ␮L, 2 ␮L of the specimens was added. A touch-
down PCR program was designed to improve the specificity and
sensitivity: Preincubation 10 minutes at 40°C and 95°C, respec-
tively. First amplification was 10 cycles of annealing at 63°C for
6 seconds and extension at 72°C for 8 seconds. Secondary ampli-
fication included 45 cycles of denaturing at 95°C for 10 seconds;
annealing first at 63°C for 6 seconds,and stepped down to 56°C for
1°C step size, and 1 cycle step delay, single acquisition mode; and
extending at 72°C for 8 seconds. A melting curve was generated by
95°C for 0 seconds; 55°C, held for 15 seconds; 80°C for 0 seconds,
with slope ⫽ 0.05°C/s and continuous acquisition mode.
The analytical sensitivity was validated by testing serial
dilutions of genomic DNA of M. pneumoniae strain M129 (ATCC
29342) and one of the laboratory-derived mutants at concentra-
tions ranging from 7 ⫻ 104 to 7 molecules/␮L. Specificity was
validated by testing 36 M. pneumoniae reference strains, 25 strains
of other mycoplasmal and ureaplasmal species of human and
nonhuman origin, 26 bacterial and viral species commonly asso-
ciated with respiratory tract infections, and human genomic DNA.
RESULTS
MICs of Clinical Isolates
MICs for the 3 drugs tested were: azithromycin (8 to 32
␮g/mL); erythromycin (ⱖ32 ␮g/mL) and telithromycin (ⱖ32
␮g/mL) indicating high level resistance for each of the isolates.
MICs for these drugs for are typically ⱕ0.125 ␮g/mL for macrol-
ide-susceptible M. pneumoniae.1
Sequence Analysis of 23S rRNA Gene From
M. pneumoniae Isolates
Sequence of whole PCR products of portions of the 23S
rRNA gene revealed that M. pneumoniae isolates from the 2
children carried the A2063G mutation in the rRNA gene as shown
in Table 2 (Table, Supplemental Digital Content 2,
http://links.lww.com/A1303). Only 1 isolate was tested from case
© 2009 Lippincott Williams & Wilkins
The Pediatric Infectious Disease Journal • Volume 28, Number 8, August 2009
2, but it is assumed that the other 2 macrolide-resistant isolates
grown in culture from specimens collected from this child at the
same time but from different anatomic locations in the lung carried
the same mutation. No mutations were found at the 2617 position
in any of these isolates.
Real-Time PCR Testing for Mutations Associated
With Macrolide Resistance
Four clinical specimens from the 2 cases described above (1
from case 1 and 3 from case 2) were tested along with the WT control
M. pneumoniae strain M129 (ATCC 29342) for mutations in the 23S
rRNA gene. The melting curve analysis clearly discriminated the WT
M129 strain from the specimens containing the A2063G
mutation (mentioned Fig., Supplemental Digital Content 3,
http://links.lww.com/A1304). The Tms of A2063G mutants were
about 4°C lower than the WT Tm, a difference that was greater
than had been predicted. The WT melting peak was 67.31°C, while
the Tms of A2063G mutants were 63.25 ⫾ 0.04°C. The
C2617A/G mutations were tested simultaneously. Since the organ-
isms in the clinical specimens available to us did not mutate at this
position, they showed similar WT Tms of about 68°C. The
C2617A/G mutant Tm was predicted to be about 10°C lower than
that of WT.
We also evaluated 4 macrolide-resistant M. pneumoniae
mutants (erythromycin MICs ⱖ32 ␮g/mL) that had been induced
in vitro by selective incubation in subinhibitory concentrations of
erythromycin (mentioned in Fig., Supplemental Digital Content 4,
http://links.lww.com/A1305). In reactions for detecting point mu-
tations in positions 2063/2064, all 4 strains had Tms of approxi-
mately 63.4°C, which is similar to the Tm of the A2063G mutant
control, indicating these strains most likely harbored mutations in
position 2063 or 2064. Sequencing data revealed that 1 mutant
(strain 20822) carried an A2063G transition while the other 3
(strains 20823, 20824, and 20825) had A2064G transitions. Thus,
this reaction was able to detect 2 point mutations at positions 2063
and 2064, although it was unable to discriminate them from one
another. No mutations were detected in position 2617, which was
confirmed by sequencing (data not shown).
PCR Sensitivity and Specificity Assessment
This assay successfully amplified serial dilutions of genomic
DNA of strain M129 using concentrations ranging from 7 ⫻ 104 down
to 7 molecules/␮L (mentioned in Fig., Supplemental Digital Content
5, http://links.lww.com/A1306). Melting curve analysis indicated
that it needed a 10-fold higher concentration to get an effective
melting peak. Thus, converting to the concentration of genomic
DNA molecules in the PCR reaction mixture, the testing limits for
both reactions were 7 molecules/␮L. According to the standard
curve generated from the amplification of serial DNA dilutions, the
PCR efficiency for 2063/2064 mutations was 1.902 and 1.775 for
2617. When testing the serial dilutions of genomic DNA of the
A2063G mutants, similar results were obtained.
All 36 M. pneumoniae reference strains were successfully
recognized, including representatives for both major subtypes.
Among other Mycoplasma and non-Mycoplasma species, only
Mycoplasma genitalium and Mycoplasma pirum were amplified by
both tests (data not shown). The Tms of M. genitalium in the 2
assays were different from those of mutant and WT M. pneu-
moniae and enabled us to discriminate the real M. pneumoniae
mutants from the M. genitalium. The cross-reaction with M. pirum
was weak although the TMs were similar to the M. pneumoniae
mutants. Ureaplasmas were also weakly amplified but were unable
to produce a melting peak. When tested with DNA concentration
comparable to clinical specimens, cross-reactions with M. pirum
and Ureaplasma species were not detectable.
© 2009 Lippincott Williams & Wilkins
Macrolide-Resistant M. pneumoniae
DISCUSSION
The 2 cases presented in this report illustrate the dilemma in
which the clinician is placed when confronted with patients who
have relatively severe pulmonary symptoms, suspicion that M.
pneumoniae infection is present, and no information regarding
antibiotic susceptibility. In both of these cases, the patients expe-
rienced a prolonged, severe lower respiratory infection that was
poorly responsive to macrolide therapy. Good clinical practice
includes submission of laboratory specimens for diagnosis of M.
pneumoniae infections in the appropriate clinical setting. Labora-
tory diagnosis of M. pneumoniae infection may include orders for
serology and/or PCR where available, but very rarely for culture.
Susceptibility testing has never been recommended, even when
cultures are obtained because of the assumption that M. pneu-
moniae is uniformly susceptible to macrolides, the drugs of choice.
The only reason these 2 macrolide-resistant infections were de-
tected was because the microbiology laboratory was performing an
in vitro susceptibility study of an investigational antimicrobial
agent several months after the initial isolations. Physicians should
now have an appreciation that M. pneumoniae can sometimes
cause clinically severe respiratory infections and that macrolide
susceptibility can no longer be assumed.
Point mutations in 3 positions (2063, 2064, 2617) of the 23S
rRNA gene have been reported to be related to development of
macrolide resistance in M. pneumoniae.4-10 Similar mutations in
the corresponding positions have been observed in other species
such as Helicobacter pylori,13 indicating these positions are in or
near the active sites for macrolides. To rapidly detect these
mutations, real-time PCR methods have been developed in several
species.13,14 Recently, real-time PCR methods based on high
resolution melting curve analysis for detection of point mutations
in M. pneumoniae at position 2063 or 2064 were described by
Wolff et al.11 However, those methods may not be optimum for
clinical diagnosis because of the reduced sensitivity in testing
direct patient specimens, especially when culture is not successful.
To overcome these limitations, we have developed a multiplex
real-time PCR assay to detect point mutations in all 3 positions of
domain V of the 23S rRNA gene in original clinical specimens as
well as in cultured bacterial isolates. This assay was designed to
use FRET hybridization probes and the Roche LightCycler 2.0
instrument. The detection limit is as low as 7 mutant molecules/␮L
in the PCR mixture. Our new method tested a broad range of
species and showed satisfactory specificity. Thus, this method can
be used clinically for detecting macrolide-resistant M. pneu-
moniae. The assay was successful in original clinical specimens
and in M. pneumoniae isolates. Based on the predicted Tms and
what we have observed thus far in evaluation of a limited number
of specimens, the assay can be expected to distinguish the 2617
mutation in M. pneumoniae from the more common 2063/2064
mutations. It may not be able to distinguish between the latter 2, but
such distinction is probably not important for patient management
since both mutations can confer high level macrolide resistance.6,8-10
The emergence of clinically significant macrolide resistance
in M. pneumoniae increases the need for an accurate rapid diag-
nostic test for detection of this organism. The ability to perform
our assay on the same day as a specimen is obtained enables the
clinician to be informed of the presence of a resistant organism in
time to initiate treatment with another drug class. The widespread
use of macrolides in the United States, particularly azithromycin,
in suspected or confirmed M. pneumoniae infections and the fact
that a single base mutation in the 23S rRNA gene confers virtually
complete resistance ensures that macrolide resistance in the United
States is certain to continue rising as it has in Asia.
www.pidj.com | 695
Li et al The Pediatric Infectious Disease Journal • Volume 28, Number 8, August 2009
REFERENCES
1. Waites KB, Talkington DF. Mycoplasma pneumoniae and its role as a
human pathogen. Clin Microbiol Rev. 2004;17:697–728.
2. Korppi M, Heiskanen-Kosma T, Kleemola M. Incidence of community-
acquired pneumonia in children caused by Mycoplasma pneumoniae: se-
rological results of a prospective, population-based study in primary health
care. Respirology. 2004;9:109 –114.
3. Layani-Milon MP, Gras I, Valette M, et al. Incidence of upper respiratory
tract Mycoplasma pneumoniae infections among outpatients in Rhone-
Alpes, France, during five successive winter periods. J Clin Microbiol.
1999;37:1721–1726.
4. Pereyre S, Charron A, Renaudin H, et al. First report of macrolide-
resistant strains and description of a novel nucleotide sequence varia-
tion in the P1 adhesin gene in Mycoplasma pneumoniae clinical strains
isolated in France over 12 years. J Clin Microbiol. 2007;45:3534 –3539.
5. Okazaki N, Narita M, Yamada S, et al. Characteristics of macrolide-
resistant Mycoplasma pneumoniae strains isolated from patients and in-
duced with erythromycin in vitro. Microbiol Immunol. 2001;45:617– 620.
6. Matsuoka M, Narita M, Okazaki N, et al. Characterization and molecular
analysis of macrolide-resistant Mycoplasma pneumoniae clinical isolates ob-
tained in Japan. Antimicrob Agents Chemother. 2004;48:4624–4630.
7. Suzuki S, Yamazaki T, Narita M, et al. Clinical evaluation of macrolide-
resistant Mycoplasma pneumoniae. Antimicrob Agents Chemother. 2006;
50:709 –712.
696 | www.pidj.com
8. Morozumi M, Iwata S, Hasegawa K, et al. Increased macrolide
resistance of Mycoplasma pneumoniae in pediatric patients with com-
munity-acquired pneumonia. Antimicrob Agents Chemother. 2008;52:
348 –350.
9. Morozumi M, Hasegawa K, Kobayashi R, et al. Emergence of macrolide-
resistant Mycoplasma pneumoniae with a 23S rRNA gene mutation. Anti-
microb Agents Chemother. 2005;49:2302–2306.
10. Liu Y, Ye X, Zhang H, et al. Antimicrobial suscceptibility of Myco-
plasma pneumoniae isolates and molecular analysis of macrolide-resis-
tant strains from Shanghai, China. Antimicrob Agents Chemother. 2009;
53:2160 –2162.
11. Wolff BJ, Thacker WL, Schwartz SB, et al. Detection of macrolide
resistance in Mycoplasma pneumoniae by real-time PCR and high resolu-
tion melt analysis. Antimicrob Agents Chemother. 2008;52:3542-3549.
12. Waites KB, Duffy LB, Schwartz S. Mycoplasma and Ureaplasma. In:
Isenberg H, ed. Clinical Microbiology Procedure Handbook. 2nd ed.
Washington, DC: American Society for Microbiology Press; 2004.
13. Oleastro M, Menard A, Santos A, et al. Real-time PCR assay for rapid
and accurate detection of point mutations conferring resistance to
clarithromycin in Helicobacter pylori. J Clin Microbiol. 2003;41:397–
402.
14. Pandori MW, Gordones C, Castro L, et al. Detection of azithromycin
resistance in Treponema pallidum by real-time PCR. Antimicrob Agents
Chemother. 2007;51:3425–3430.
© 2009 Lippincott Williams & Wilkins