Development of a multiplex real-time PCR assay for detection of Mycoplasma pneumoniae, Chlamydiapneumoniae and mutations associated with macrolide resistance in Mycoplasma pneumoniae from respiratory clinical specimens

The aim of this study was to improve detection of Mycoplasma pneumoniae and Chlamydia pneumoniae in clinical specimens by developing a multiplex real-time PCR assay that includes identification of macrolide-resistant M. pneumoniae. Novel assays targeting a M. pneumoniae conserved hypothetical protein gene, M. pneumoniae 23S rRNA gene mutations associated with macrolide resistance and human β-globin gene (an endogenous internal control) were designed and combined with a previously published C. pneumoniae PCR targeting ompA gene. The resulting quadraplex PCR was validated with a panel of clinical specimens supplemented with external quality assessment specimens, simulated specimens and various bacterial and viral strains. The obtained results were compared to those obtained by reference PCRs or confirmed by sequencing (typing of macrolide resistance). The novel multiplex PCR assay was in 100 % agreement with reference PCRs. Four M. pneumoniae strains with macrolide resistance-associated mutations were identified among 42 strains, which comprises 9.5 % of the study material. Amplification of an internal control excluded sample-derived inhibition possibly leading to false-negative reporting. In conclusion, we have developed a resources conserving multiplex real-time PCR assay for simultaneous detection of M. pneumoniae, C. pneumoniae and the most common mutations leading to macrolide resistance in M. pneumoniae. The assay is a widely useful tool for detection of these respiratory pathogens and will also shed light on the occurrence of macrolide resistance in M. pneumoniae. Electronic supplementary material The online version of this article (doi:10.1186/s40064-015-1457-x) contains supplementary material, which is available to authorized users.


Background
Mycoplasma pneumoniae and Chlamydia pneumoniae are common respiratory pathogens causing approximately 10-30 % and 5-10 %, respectively, of community acquired pneumonias (Atkinson et al. 2008;Kuo et al. 1995). M. pneumoniae infections can also present with extrapulmonary symptoms, including mucocutaneous and central nervous system manifestations (Atkinson et al. 2008), and the clinical spectrum of C. pneumoniae infection includes reactive arthritis, worsening of asthma and cardiovascular disease (Orrskog et al. 2013).
Mycoplasma pneumoniae infection can occur in all age groups, but the severity of the disease tends to increase with the age of the patient. For immunocompromised patients, M. pneumoniae infection can lead to a severe pneumonia (Waites and Talkington 2004). C. pneumoniae IgG antibody prevalence bounces between 5 and Open Access *Correspondence: laura.mannonen@hus.fi Virology and Immunology, University of Helsinki and Helsinki University Hospital, P.O. Box 400, 00029 HUS Helsinki, Finland 15 years of age and then increases steadily suggesting that most primary infections take place in school-age children and repeated infections occur later in life. Antimicrobial therapy is available for M. pneumoniae and C. pneumoniae infection. In case antimicrobial treatment is indicated, specific and rapid laboratory diagnosis is important as the diagnosis cannot be based on clinical symptoms only. M. pneumoniae pneumonias are often treated with macrolides but also tetracyclines and fluoroquinolones are effective (Waites and Talkington 2004). Macrolides are often used for treatment due to their safety especially for children and pregnant women. However, in the last few years macrolide-resistant M. pneumoniae strains have emerged, with high rates in China and Japan, and lower rates in Europe and Canada (Principi and Esposito 2013). Resistance to macrolides is caused by point mutations in the bacterial 23S rRNA gene which reduces the affinity of the antibiotic for the ribosomes (Wolff et al. 2008). The most common point mutations leading to the resistance are located in the 23S rRNA gene at positions 2063 and 2064, and include transition A to G or C (Waites and Talkington 2004;Lucier et al. 1995).
In recent years, nucleic acid-based assays targeting M. pneumoniae and C. pneumoniae have been developed Jacobs 2009, 2014;Winchell et al. 2008). PCR-based methods have high sensitivity, short turnaround time and the testing can be completed in a single working day. Moreover, simultaneous amplification of more than one target in the same reaction (multiplex PCR) is becoming a rapid, economical and convenient means in a diagnostic laboratory to detect multiple pathogens causing similar symptoms and their genetic properties at the same. A variety of multiplex PCRs have been developed for detection of M. pneumoniae and C. pneumoniae (Thurman et al. 2011;Welti et al. 2003;Gullsby et al. 2008;Miyashita et al. 2004;McDonough et al. 2005). However, also molecular diagnosis has limitations since the mere presence of nucleic acid can denote either acute infection, persistence of nucleic acid after infection or asymptomatic carriage.
The aim of this study was to improve detection of M. pneumoniae and C. pneumoniae in clinical specimens by developing a time-saving and inexpensive multiplex real-time PCR assay for detection of C. pneumoniae, M. pneumoniae and the most common mutations in the M. pneumoniae 23S rRNA gene leading to resistance for macrolide antibiotics. To our knowledge, this is the first report of a multiplex real-time PCR assay that in addition to the detection of M. pneumoniae, C. pneumoniae and internal control includes detection of macrolide resistance-associated mutations in M. pneumoniae.

Sensitivity of co-amplification
To assess the sensitivity of co-amplification of C. pneumoniae, M. pneumoniae and human beta-globin by multiplex real-time PCR, a range of concentrations of M. pneumoniae M129 and C. pneumoniae K-6 DNA were co-amplified. In addition, low concentrations of M. pneumoniae and C. pneumoniae were amplified in presence of high concentration of human DNA. Crossing threshold (Ct) values remained unchanged for the DNA that was in a lower concentration even when a high concentration of DNA from another organism was co-amplified (data not shown).

Internal control (IC) human beta-globin
The human β-globin was co-amplified in the multiplex reaction as IC. Neither sample matrix-derived inhibition in the PCR reaction, nor specimen processing errors potentially leading to false negative reporting were observed in this study. The Ct values in nose and/ or throat swabs (n = 72) varied between 18.9 and 28.1 (mean 23.2, median 23) and for BALs (n = 13) the Ct range was 21.6-29.6 (mean 25.3, median 26).

Analytical sensitivity
The analytical sensitivity, the limit of detection (LOD) with 95 % probability, was determined for M. pneumoniae and C. pneumoniae components of the multiplex real-time PCR using Probit analysis of the Validation Manager software (Finbiosoft, Espoo, Finland). The analytical sensitivity for M. pneumoniae was 2.9 copies/PCRreaction and for C. pneumoniae (K-6) the LOD was 0.13 IFU/PCR reaction. To confirm that the C. pneumoniae primers and probe are able to detect different strains, DNA from nine available C. pneumoniae isolates were tested: The strains isolated in Finland K6, K8, K66, FIL, HELS-12, MS and Parola, CV6 from Germany (a kind gift from Prof. M Maass), and TW-183 (ATCC VR-2282) and CWL-029 (ATCC VR-1310). The multiplex PCR detected these strains.

Analytical specificity
The specificity of the primers and probes was evaluated using BLAST analysis, and analyzing nucleic acid from various bacterial and viral strains in the multiplex real-time PCR (Additional file 1, Table S1). In silico and in vitro analyses revealed that the M. pneumoniae primers and probe designed for detection of the most common resistance-associated mutations could also amplify DNA from Mycoplasma genitalium.

Diagnostic performance of the multiplex PCR
To evaluate the reliability of the developed real-time multiplex PCR for clinical testing the performance of the assay was compared to those of the reference PCR assays. A total of 42 M. pneumoniae positive specimens and 24 M. pneumoniae simulated BALs were studied with the multiplex assay. The multiplex realtime PCR detected all simulated BALs correctly and the assay was in 100 % agreement with the reference PCRs (Table 1).
The macrolide resistance-associated mutations in M. pneumoniae strains were identified by dissociation curve analysis of the amplicon. The analysis distinguishes mutations in 23S rRNA sequence at positions 2063G/C or 2064G/C. The melting temperature of the wild type is 64-65 °C, for the mutation C 60-61 °C and for the mutation G 57.5-58.5 °C (Fig. 1). The dissociation curve analysis does not differentiate whether the mutation is in the base 2063 or 2064. In our study material, 9.5 % (4/42) of the M. pneumoniae positive specimens contained a M. pneumoniae strain with macrolide resistance-associated mutation. The PCR results were confirmed by sequence analysis. Two BALs and two sputum specimens contained M. pneumoniae with a mutation associated with macrolide resistance. Three M. pneumoniae strains had a point mutation at 2063 A->G and one sputum specimen contained both wild type DNA and DNA with a point mutation at 2064 A->G.
The C. pneumoniae component in the multiplex PCR uses same oligonucleotides as the previously validated reference PCR and hence the performance of the C. pneumoniae component was verified with a smaller amount of specimens. The clinical sensitivity was verified by testing 28 C. pneumoniae positive or simulated positive upper and lower respiratory tract specimens. In addition, a total of 43 C. pneumoniae negative respiratory tract specimens were assessed. The multiplex PCR assay was in 100 % agreement with the reference method (Table 1).

External quality assessment panel
EQA panel C. pneumoniae and M. pneumoniae 2014 (QCMD, Glasgow, Scotland, UK) was analyzed with the multiplex real-time PCR assay. The multiplex PCR identified all core proficiency samples correctly. Only two educational, infrequently detected M. pneumoniae positive specimens were not detected by the multiplex PCR (CP.MP14-03 and CP.MP14-12) and one educational infrequently detected C. pneumoniae positive specimen CP.MP14-06 gave a negative result (Table 2).

Discussion
Mycoplasma pneumoniae and C. pneumoniae cause atypical pneumonia and have clinically similar respiratory manifestations. Hence it is logical to detect them simultaneously. To improve laboratory diagnosis and to reduce the cost and time for detecting M. pneumoniae and C. pneumoniae, we designed and validated a multiplex realtime PCR that can also detect the most common mutations in the 23S rRNA gene of M. pneumoniae leading to macrolide resistance in a single reaction. Human betaglobin gene was included as an internal control to allow assessment of the specimen quality and the nucleic acid extraction as well as the amplification processes. No inhibition or specimen processing errors were identified in this study. The variation of the detected IC Ct values for swab specimens and BAL specimens follows normal distribution (mean and median are almost identical) and is therefore more likely to reflect variation in the sampling procedure.
All specimens were analysed both in singleplex and in multiplex reactions. Multiplex PCR may lack sensitivity when compared to singleplex PCR (Mannonen et al. 2012), but here the carefully optimized multiplex PCR had identical sensitivity in comparison to the singleplex PCR (data not shown). The obtained LOD 2.9 cp/reaction for M. pneumoniae can be considered functional and is comparable to those reported by others Jacobs 2009, 2014;Welti et al. 2003;Schmitt et al. 2013). For the C. pneumoniae component of the multiplex PCR, the LOD of 0.13 IFU/PCR reaction was obtained. The sensitivities of previously published PCR assays are challenging to compare since no standard preparations for measuring analytical sensitivity are available (Dowell et al. 2001). Until international molecular standards become available, molecular EQA panels may be the best way to estimate sensitivity of a molecular test. Here we analysed a QCMD 2014 proficiency panel for C. pneumoniae and M. pneumoniae with the developed multiplex PCR method. The multiplex PCR method identified all core samples correctly indicating acceptable level of proficiency. In addition, 3/6 educational samples were identified correctly. The remaining three educational samples    (Bébéar et al. 2011;Principi and Esposito 2013). Mutations in the 23S rRNA gene can confer resistance to macrolides and the most commonly occurring mutations are A2063G and A2064G. In our study, 9.5 % of M. pneumoniae strains were shown to have either of these mutations. The observed rate is similar to those reported in other European countries and USA (0-26 %), whereas in Japan and China, the occurrence of macrolide resistance is much higher (>90 %) (Bébéar et al. 2011;Principi and Esposito 2013). A2063G is the most commonly detected mutation worldwide (>90 % of all reported point mutations), and this was also the case in our study (three out of four cases had A2063G mutation). In silico and in vitro studies revealed that the primers and the probe for macrolide resistance mutations also detected M. genitalium. The 23S rRNA genes of M. pneumoniae and M. genitalium are closely related (96.7 % identity) and therefore the observed cross-reactivity is not surprising (Dumke et al. 2010). M. genitalium, however, is not frequently detected in respiratory specimens. Moreover, the presence of M. pneumoniae is first identified by specific oligonucleotides and the presence of mutation is evaluated only in M. pneumoniae positive specimens, so the risk of incorrect identification was considered minimal.
As antimicrobial therapy effective against M. pneumoniae and C. pneumoniae infections is available and recommended in lower respiratory tract infections, molecular testing can aid in patient management by providing a microbiological diagnosis and guiding antibiotic selection. Sensitive nucleic acid based detection techniques also contribute to a better understanding of the epidemiology and manifestations of these infections.

Conclusions
We have developed a resources conserving multiplex real-time PCR assay which has been successfully implemented in a clinical diagnostic laboratory for simultaneous detection of M. pneumoniae, C. pneumoniae and the most common mutations leading to macrolide resistance in M. pneumoniae. The assay is a widely useful tool for detection of these respiratory pathogens and will also shed light on the occurrence of macrolide resistance in M. pneumoniae.

Design of primers and probes
Primer Express software v 3.0 (Applied Biosystems, Foster City, CA) was used to design primers and probes targeting a conserved hypothetical protein C985_0367 of M. pneumoniae (MP) (MP M129-B7) and human β-globin (BG) gene. The C. pneumoniae (CP) primers and probe have previously been described by Mannonen et al. 2004. The primers and probe for detection of macrolide resistance-associated mutations in M. pneumoniae (MR-MP) were designed by TIB MOLBIOL (Berlin, Germany). All primers and probes were tested for specificity by basic local alignment search tool (BLAST). The primer and probe sequences are listed in Table 3.

Multiplex real-time PCR design
Endogenous IC human β-globin PCR assay was optimized as a primer limited reaction to avoid competition between other amplifications. The multiplex PCR reac- Multiplex PCRs were performed using Stratagene Mx3005p instrument (Agilent Technologies Inc, Santa Clara, CA, United States). Positive and negative controls as well as a no template control (NTC) were included in each run. Cycling conditions were 1 cycle 95 °C for 10 min followed by 45 cycles of 95 °C for 15 s and 60 °C for 60 s. After amplification a dissociation curve analysis was performed with a thermal profile: 95 °C for 60 s, 55 °C for 30 s followed by a slow rise in the temperature (0.2 °C/s) to 95 °C with continuous acquisition of fluorescence signal.

Reference PCRs
The M. pneumoniae results by the new multiplex PCR were compared to those obtained by the diagnostic M. pneumoniae PCR test in use at the time of the study at HUSLAB (adapted from van Kuppeveld et al. 1992) or to M. pneumoniae PCR targeting CARDS toxin (CARDStx) gene adapted from Winchell et al. 2008. For the diagnostic PCR, 5 µl of nucleic acid extract was amplified in a 50 µl reaction volume with 0.5 µM of primers (Table 3). Amplified products were detected by microplate hybridization (Vesanen et al. 1996). PCR products were detected with digoxigenin labelled probe (Table 3), anti-digoxigenin antibody conjugated with alkaline phosphatase and LumiPhos Plus substrate (Lumigen, Souhfield, MI, USA). For the CARDStx PCR 5 µl of nucleic acid extract was amplified in 25 µl PCR reaction with 0.25 µM of primers and 0.125 µM probe (Table 3). The real-time PCR reaction was carried out in Stratagene Mx3005p instrument.
The C. pneumoniae results obtained by the new multiplex PCR were compared to those of the diagnostic C. pneumoniae PCR test in use at the time of the study at HUSLAB (Mannonen et al. 2004) (with minor modifications). The primers and probe were the same as in the new multiplex test. Five µl of nucleic acid extract was amplified in 50 µl reaction mixture with 0.3 µM forward and reverse primer and 0.225 µM probe (Table 3).

Specimens
Respiratory tract specimens referred to HUSLAB, Department of Virology and Immunology for testing of M. pneumoniae, C. pneumoniae and/or respiratory viruses between years 1999-2014, were analyzed in this study. The specimens included 42 M. pneumoniae positive specimens assessed by diagnostic PCR of that time (29) (van Kuppeveld et al. 1992) or by CARDS Tx PCR (13) . The specimens were nasopharyngeal swabs (24), nasopharyngeal aspirates (5), sputum samples (6), tracheal aspirates (5) and bronchoalveolar lavages (BAL; 2). Due to the limited amount of M. pneumoniae positive BAL specimens, 24 negative BALs were spiked with M. pneumoniae positive respiratory specimens in dilutions 1:10-1:1000.
The C. pneumoniae positive specimens included two sputum samples and three nose and/or throat swabs, 12 simulated BALs and 11 simulated respiratory tract specimens. The simulated specimens were prepared by spiking negative BALs and respiratory tract specimens with C. pneumoniae strain K6 yielding 1.25 inclusion forming unit (IFU)/PCR reaction or 250 IFU/PCR reaction. Altogether 43 M. pneumoniae and C. pneumoniae negative upper and lower respiratory tract specimens were Specimens tested by the reference M. pneumoniae and C. pneumoniae PCRs had been extracted by either using MagNA Pure LC, EasyMAG instrument (bioMerieux, Marcy l'Etoile, France) or phenol-chloroform extraction. Starting volumes were 1800 µl for BALs, 200 µl for nasopharyngeal aspirates and sputum specimens, and 100 µl for nasopharyngeal swabs. Nucleic acids extracted by EasyMAG were eluted in 25 µl and those extracted by phenol-chloroform protocol were eluted either in 50 µl (nasopharyngeal swabs) or 20 µl (other specimens).

Sequencing analysis of PCR products
Nucleotide sequence analysis of M. pneumoniae 23S rRNA gene (GeneBank. X68422) was performed to all specimens containing mutations associated with macrolide resistance in M. pneumoniae according to the developed multiplex real-time PCR. The primer sequences have been previously described (Table 3). The PCR products were sequenced in the Institute of Biotechnology at the University of Helsinki.

Analytical sensitivity
The analytical sensitivity of the multiplex real-time PCR assay was assessed by amplifying serial dilutions of M. pneumoniae M129 DNA control (ZeptoMetrix Corporation, Franklin, MA) and DNA from C. pneumoniae K-6 (quantitated at Mirja Puolakkainen's research laboratory). Four dilutions with ten replicates of each were amplified with the multiplex real-time PCR carried out in Mx3005p instrument. The tested M. pneumoniae M 129 DNA dilutions were 6.210, 3.105, 1.552 and 0.776 genome copies (cp)/PCR reaction, and C. pneumoniae K-6 DNA dilutions were 0.250, 0.125, 0.063 and 0.032 IFU/PCR reaction.

Analytical specificity
To assess the analytical specificity of the multiplex PCR assay various bacterial and viral samples were studied (Additional file 1, Table S1).