Point-of-care molecular diagnosis of Mycoplasma pneumoniae including macrolide sensitivity using quenching probe polymerase chain reaction

Objectives Macrolides are generally considered to be the drugs of choice for treatment of patients with Mycoplasma pneumoniae infection. However, macrolide-resistant M. pneumoniae has been emerging since about 2000. The Smart Gene® system (MIZUHO MEDY Co., Ltd., Tosu, Japan) is a novel fully automated system for detection of pathogens using the method of quantitative polymerase chain reaction (qPCR) with QProbe (QProbe PCR). The entire procedure is completed within 50 min and the size of the instrument is small (15 x 34 x 30 cm). The purpose of this study was to evaluate the usefulness of the Smart Gene® system for detection of M. pneumoniae and detection of a point mutation at domain V of the 23S rRNA gene of M. pneumoniae. Materials Pharyngeal swab samples were collected from 154 patients who were suspected of having respiratory tract infections associated with M. pneumoniae. Results Compared with the results of qPCR, the sensitivity and specificity of the Smart Gene® system were 98.7% (78/79) and 100.0% (75/75), respectively. A point mutation at domain V of the 23S rRNA gene was detected from 7 (9.0%) of 78 M. pneumoniae-positive samples by the Smart Gene® system and these results were confirmed by direct sequencing. The minimum inhibitory concentrations of clarithromycin among the 5 isolates of M. pneumoniae with a point mutation at domain V of the 23S rRNA gene were >64 μg/ml and those among the 33 isolates without a mutation in the 23S rRNA gene were <0.0625 μg/ml. Conclusion The Smart Gene® system is a rapid and accurate assay for detection of the existence of M. pneumoniae and a point mutation at domain V of the 23S rRNA gene of M. pneumoniae at the same time. The Smart Gene® system is suitable for point-of-care testing in both hospital and outpatient settings.


Introduction
The Smart Gene 1 system (MIZUHO MEDY Co., Ltd., Tosu, Japan) was developed on the basis of the concept of point-of-care testing for genetic testing of pathogens. This system does not require any special technique. The Smart Gene 1 system automatically performs nucleic acid extraction and gene detection using the method of qPCR with QProbe (QProbe PCR), and the use of the system was approved in Japan in 2018 [25][26][27]. The Smart Gene 1 system consists of a fully automatic gene analyzer (Smart Gene 1 ) and a dedicated cartridge (Smart Gene 1 Myco). All of the reagents necessary for extraction of nucleic acids, amplification and detection of the targeted sequences of M. pneumoniae are included in the dedicated cartridge (Smart Gene 1 Myco) [28]. After setting the dedicated cartridge (Smart Gene 1 Myco) into the gene analyzer (Smart Gene 1 ), nucleic acids are extracted from samples by a silica solid phase extraction method [29,30] and the genomic region encoding 23S ribosomal RNA (23S rRNA) is amplified by QProbe PCR. The existence of the genome of M. pneumoniae and point mutations at domain V of the 23S rRNA gene of M. pneumoniae are analyzed simultaneously with melting curve analysis using the quenching probe method [26,27]. A macrolide inhibits protein synthesis mainly by binding to domain V of the 23S rRNA at nucleotide positions 2063 and 2064; these positions appear to be essential for binding [31,32]. Therefore, mutations at A2063 or A2064 confer the highest resistance to these antimicrobials [31]. The entire operation from setting of the cartridge to obtaining the result proceeds automatically and the entire procedure is completed within 50 min. Contamination due to handling should be minimal. The gene analyzer (Smart Gene 1 ) is small (152 mm (W), 343 mm (D) and 300 mm (H)) and inexpensive ($5,000).
It has been reported that NAATs were used in only 3.8% of patients who were suspected of having M. pneumoniae infection in Japan [33]. For the majority of patients, antibody testing (65.6% of patients) and antigen testing including an immunochromatographic assay (30.6% of patients) were used [33]. The Smart Gene 1 system is expected to be useful in a medical setting for point-of-care testing, especially in countries such as Japan in which NAATs are not widely used.
The purposes of this study were (1) to determine the sensitivity and specificity of the Smart Gene 1 system for detection of M. pneumoniae from pharyngeal swab samples compared with the results of qPCR, (2) to confirm the results for point mutations at domain V of the 23S rRNA gene of M. pneumoniae detectable in the Smart Gene 1 system by direct sequencing and (3) to confirm the in vitro anti-mycoplasma activities of antibiotics against isolates of M. pneumoniae with or without a point mutation at domain V of the 23S rRNA gene detected by the Smart Gene 1 system.

Ethical approval
Ethical approval for this study was obtained from the Institutional Review Board of Hokkaido University Hospital for Clinical Research (018-0051).

Study population and procedures
Pharyngeal swab samples were collected from patients who were suspected of having respiratory tract infections associated with M. pneumoniae from July 1, 2018 to December 31, 2019 at the Sapporo Cough, Asthma, and Allergy Center and the Department of Pediatrics in NTT-East Sapporo Hospital, Sapporo, Japan. Two swabs were used for collection of specimens: one for QProbe PCR and the other for isolation of M. pneumoniae. A swab provided with the Smart Gene 1 Myco was slowly inserted from the oral cavity into the pharynx, and mucous membrane was collected by rubbing the posterior pharyngeal wall or faucial tonsil several times. The other swab for isolation of M. pnuemoniae was suspended in Universal Transport Medium (UTM, Copan Diagnotics Inc, Italy) and then stored in the deep freezer.
The age and sex of each patient, the time of onset (the first time that the patient had a fever of more than 37.5˚C), symptoms of the respiratory system (cough, nasal discharge, sore throat, wheezing, dyspnea, headache and fatigue), radiographic findings, antibiotics taken within one week if any and diagnosis (pneumonitis, bronchitis or upper respiratory tract infection) were recorded by the physicians. Written informed consent was obtained from all patients or guardians.
Detection of M. pneumoniae with the Smart Gene 1 system Smart Gene 1 Myco is a cartridge in which nucleic acid extraction and QProbe PCR are carried out. The tip of a swab containing a specimen is dipped into the extraction reagent containing guanidine thiocyanate and Triton-X100, and the nucleic acids are extracted from the specimen. Four drops of the extraction reagent containing nucleic acids are placed onto the sample spot of the cartridge (S1 Fig). The nucleic acids in the sample are trapped on the surface of the silica particles in the membrane filter. After washing with the washing buffer, the nucleic acids on silica particles are used for the following QProbe PCR reaction (S2 Fig).
The Smart Gene 1 system uses the PCR method with QProbe [26,27]. QProbe is a hybridization probe in which the terminal cytosine is labeled with a fluorescent dye that quenches upon binding to the target sequence, and it is able to differentiate single-nucleotide polymorphisms using melting curve analysis [26,27]. QProbe emits upon dissociation to the target sequence ( Fig 1A, left) and quenches upon binding to the target sequence (Fig 1A, right). Resistance to macrolides results from a mutation at any one of the positions 2063, 2064, 2067 and 2617 in domain V of the 23S rRNA gene within the mycoplasma cell, which causes a subsequent conformational change and precludes macrolide binding [4,34]. The specific primers of PCR were designed to amplify the genomic region of domain V of the 23S rRNA gene containing positions 2063, 2064 and 2067 at domain V of the 23S rRNA gene. The specific QProbe for M. pneumoniae (MCR QP) was designed to bind to the sequences containing positions 2063, 2064 and 2067 at domain V of the 23S rRNA gene, and the sequences of MCR QP are identical to those of the wild-type strain. The melting temperature (Tm) of the specific primers and the MCR QP are set to around 66˚C. If the genome sequence of M. pneumoniae is homologous to the wild type, the MCR QP anneals to the PCR product and QProbe fluorescence is quenched at either 55˚C or 66˚C (Fig 1B). If the genome sequence of M. pneumoniae has a mutation at any one of the positions 2063, 2064 and 2067 in domain V of 23S rRNA, the MCR QP does not anneal to the PCR product at 66˚C and QProbe fluorescence is quenched at 55˚C but not at 66˚C ( Fig  1B). If the genome sequence of M. pneumoniae is not detected, the MCR QP does not anneal to the PCR product and QProbe fluorescence is not quenched at either 55˚C or 66˚C.
The PCR protocol involves denaturation at 98˚C for 120 sec followed by 46 cycles of annealing and extension at 55˚C or 66˚C for 20 sec and denaturation at 95˚C for 10 sec. The results of the Smart Gene 1 system are displayed as follows. (1) If the genome sequence of M. pneumoniae is identical to the wild type, the MCR QP anneals to the PCR product and QProbe fluorescence is quenched at either 55˚C or 66˚C (Figs 1B and 2A). The Smart Gene 1 system reports "MP-positive and no mutation". (2) If the genome sequence of M. pneumoniae has a mutation at any one of the positions 2063, 2064 and 2067 in domain V of the 23S rRNA gene, the MCR QP does not anneal to the PCR product at 66˚C and QProbe fluorescence is quenched at 55˚C but not at 66˚C (Figs 1B and 2B). The Smart Gene 1 system reports "MP-positive and mutation". (3) If the genome sequence of M. pneumoniae is not detected, the MCR QP does not anneal to the PCR product and QProbe fluorescence is not quenched at either 55˚C or 66˚C ( Fig 2C). The Smart Gene 1 system reports "MP-negative".
The Smart Gene 1 Myco includes a negative control plasmid that contains identical sequences to the specific primers of M. pneumoniae but does not contain sequences homologous to the MCR QP. The internal control QProbe (IC QP) was designed to bind to the intervening sequences between annealing sites of the specific primers of M. pneumoniae in the negative control plasmid. Because the Tm of IC QP is set to around 38˚C, the IC QP does not anneal to the PCR product and QProbe fluorescence is not quenched during the PCR test (The temperature of the rection mixture is over 55˚C during the PCR reaction.). After completion of the PCR reaction, the temperature of rection mixture declines to 38˚C, the control QProbe anneals to the PCR product and QProbe fluorescence is quenched (Fig 1C). If QProbe fluorescence is quenched at 38˚C along with the decrease in the temperature of the rection mixture, the PCR reaction is shown to be successful ( Fig 2C). The positive control plasmid for the Smart Gene 1 Myco is prepared separately. Because the Smart Gene 1 system uses one reaction tube per sample, the PCR reaction cannot be verified by the positive control plasmid in one reaction tube. The PCR reaction should be verified by the positive control plasmid in another reaction tube.

Detection of M. pneumoniae with qPCR
In order to evaluate the sensitivity and specificity of the Smart Gene 1 system for detection of M. pneumoniae in comparison with those of qPCR, DNA was extracted with a DNA extraction

PLOS ONE
Point-of care molecular diagnosis of M. pneumoniae using QProbe PCR DNA solution was quantified by qPCR using Mp181-F and Mp181-R primers and an Mp181-P probe as described elsewhere [13,35,36].

Detection of macrolide-resistant point mutations at domain V of the 23S rRNA gene
In order to determine whether the Smart Gene 1 system can detect macrolide-resistant strains of M. pneumoniae, mutations associated with resistance to macrolides at sites 2063, 2064, 2067 and 2617 in domain V of the 23S rRNA gene in M. pneumoniae were detected by a sequencing method described elsewhere [37]. M. pneumoniae possessing a point mutation at domain V of the 23S rRNA gene was defined as MRMP.

Isolation of M. pneumoniae and determination of minimum inhibitory concentrations (MICs)
Isolation of M. pneumoniae and determination of MICs were performed according to a previous report [34].

Statistical analyses
The sensitivity, specificity, positive predictive value, and negative predictive value of Smart Gene 1 Myco were calculated using the Clopper and Pearson method with 95% confidence intervals (CIs). Standard deviation at age was calculated using the Wilcoxon rank sum test. Categorical variables were compared by Fisher's exact test. A p-value of less than 0.05 was considered statistically significant. All calculations were conducted using the R 4.1.0 software program (The R Foundation, Vienna, Austria).  Table) are shown in supplements. Although statistical analysis could not be performed due to the small number of samples, the positive rate by qPCR did not differ significantly according to the period from onset of fever to sampling (S4 Fig).

Detection of M. pneumoniae by the Smart Gene 1 system
Compared with the results of qPCR, the sensitivity and specificity of the Smart Gene 1 system were 98.7% (78/79) and 100.0% (75/75), respectively. The positive and negative predictive values of the Smart Gene 1 system were 100.0% (78/78) and 98.7% (75/76), respectively. The diagnostic accuracy was 99.4% (153/154) ( Table 1). In the sample for which the result was negative by the Smart Gene 1 system, the estimated copy number of M. pneumoniae DNA measured by qPCR was 6.7 copies per reaction of the Smart Gene 1 system. The threshold cycle (Ct) values of the Smart Gene 1 system were correlated with the copy numbers of M. pneumoniae DNA measured by qPCR (correlation coefficient = -0.92, p<0.0001) (Fig 4).

Detection of a point mutation at domain V of the 23S rRNA gene of M. pneumoniae
A point mutation at domain V of the 23S rRNA gene of M. pneumoniae was detected by the Smart Gene 1 system in 7 (9.0%) of 78 samples in which M. pneumoniae was detected by the Smart Gene 1 system. The A2063G mutation at domain V of the 23S rRNA gene was detected in 6 samples and the A2064C mutation at domain V of the 23S rRNA gene was detected in one sample by direct sequencing. In the remaining 71 M. pneumoniae-positive samples, mutations  Table). Therefore, the sensitivity and specificity of the Smart Gene 1 system for detection of a point mutation at domain V of the  23S rRNA gene were 100.0% (7/7) and 100.0% (71/71), respectively. Statistical significance was observed for the differences in the prevalence of point mutations at domain V of the 23S rRNA gene between patients with and those without macrolide pre-administration: 71.4% (5 of 7) and 9.7% (7 of 72), respectively (p = 0.0003) (S3 Table).

Antibiotic susceptibility
M. pneumoniae was isolated from 38 pharyngeal swab samples and in vitro anti-mycoplasma activities of three antibiotics (clarithromycin, minocycline and tosufloxacin) against isolates of M. pneumoniae with or without a point mutation at domain V of the 23S rRNA gene were measured ( Table 2). Five strains of M. pneumoniae were isolated from samples with the A2063G mutation at domain V of the 23S rRNA gene, and 33 strains of M. pneumoniae were isolated from samples without a mutation at domain V of the 23S rRNA gene. In Japan, the most frequently administered macrolide antibiotic in 2016 was clarithromycin (3.28 DID: the defined daily dose per 1000 inhabitants per day) followed by azithromycin (0.65 DID) [38]. Clarithromycin was therefore chosen for MIC assays. The MICs of clarithromycin among the 5 isolates with the A2063G mutation at domain V of the 23S rRNA gene were >64 μg/ml, and those among the 33 isolates without a mutation at domain V of the 23S rRNA gene were <0.0625 μg/ml ( Table 2). The MICs of minocycline and tosufloxacin among the 5 isolates with the A2063G mutation at domain V of the 23S rRNA gene were 2-4 and 0.5 μg/ml, respectively, and those among 33 isolates without a mutation at domain V of the 23S rRNA gene were 0.25-4 and 0.125-0.5 μg/ml, respectively ( Table 2). The results for point mutations at domain V of the 23S rRNA gene of M. pneumoniae detected by the Smart Gene 1 system were consistent with the antibiotic susceptibility of clinical isolates.

Discussion
In the present study, the Smart Gene 1 system showed sensitivity and specificity of 98.7% (78/ 79) and 100.0% (75/75), respectively, for the detection of M. pneumoniae from pharyngeal swab samples compared with the results of qPCR. Point mutations at domain V of the 23S rRNA gene of M. pneumoniae in 7 samples detected by the Smart Gene 1 system were confirmed by direct sequencing. The MICs of clarithromycin among the 5 isolates with a point mutation at domain V of the 23S rRNA gene evaluated by the Smart Gene 1 system were >64 μg/ml, and those among the 33 isolates without a mutation at domain V of the 23S rRNA gene were <0.0625 μg/ml. Commercial molecular assays for detection of M. pneumoniae are listed in Table 3  and copy numbers. The limit of detection of LAMP (Eiken Chemical Co., Ltd., Japan) was reported to be 6 copies per test [50]. The multiplex Lightmix 1 RT-PCR (TIBMolbiol, Berlin Germany) showed a limit of detection between 5 and 10 copies per test [51]. To investigate the limit of detection of the Smart Gene 1 system, the strain M. pneumoniae MHN8-014 harboring A at the positions 2063, 2064 and 2067 (wild type) and the strain M. pneumoniae MHN8-016 harboring one transposition A2063G (mutant type) were added to the extraction reagent solution at final concentrations ranging between 1.25 and 80 copies/uL. The extraction reagent solutions containing different concentrations of M. pneumoniae were measured by the Smart Gene1 system in triplicate (S5 Table). The Smart Gene1 system could detect wild-type and

PLOS ONE
Point-of care molecular diagnosis of M. pneumoniae using QProbe PCR mutant-type M. pneumoniae at a concentration of 2.5 copies/ uL in 3 of 3 measurements. In contrast, the Smart Gene1 system could detect wild-type and mutant-type M. pneumoniae at a concentration of 1.25 copies/ uL in 1 of 3 measurements. In the present study, two (66.7%) of three samples for which copy numbers were under 20 copies/ uL could be detected by the Smart Gene 1 system. Contaminants contained in clinical samples might inhibit trapping of the nucleic acids on the surface of the silica particles and might be responsible for decrease in the sensitivity. Further study using clinical samples is needed to know the limit of detection of M. pneumoniae by the Smart Gene 1 system. (6) Size and weight of the instrument: The instruments using LAMP are small in size and light in weight, but those using qPCR are large in size and heavy in weight except for the Smart Gene 1 system. The size of the Smart Gene 1 system is 152 mm (W) x 343 mm (D) x 300 mm (H) and the weight is 6 kg. The cost of the Smart Gene 1 system is $5,000 for the gene analyzer (Smart Gene 1 ) and $30 for the dedicated cartridge (Smart Gene 1 Myco) per sample. The price of a qPCR thermal cycler ranges from $15,000 to over $90,000 and the cost of qPCR is $5 per sample. The price of equipment for LAMP (Loopamp Realtime Turbidimeter) is $18,000 and the cost of the LAMP assay is $15 per sample.
The Smart Gene 1 system can detect macrolide resistance associated with mutations at the positions 2063, 2064 and 2067 in domain V of the 23S rRNA gene (S4 Table). However, the Smart Gene 1 system cannot detect macrolide resistance associated with mutations at the    [59]. Furthermore, mutation at position 2617 was associated with low-level macrolide resistance compared to mutation at position 2063 or 2064 [8].
Taken together, the results indicate that the clinical impact of missing MRMP associated with a mutation at position 2617 in domain V of the 23S rRNA gene is minimal.
Recently, a dedicated cartridge of the Smart Gene 1 system for detection of SARS-CoV-2 (Smart Gene 1 SARS-CoV-2) has been marketed in Japan [60]. Dedicated cartridges for detection of Clostridioides difficile, Helicobacter pylori, influenza viruses, Bordetella pertussis, Chlamydia trachomatis and Neisseria gonorrhoeae are in development. These assays will use stool, nasopharyngeal swab, nasal swab, saliva and urine as samples but will not use blood.
Limitations of this study should be recognized. Only 7 of 79 patients were infected with M. pneumoniae with a mutation in domain V of the 23S rRNA gene. This number might not be sufficient to evaluate the capacity for detection of a point mutation at domain V of the 23S rRNA gene of M. pneumoniae by the Smart Gene 1 system. Recently, Nagita et al. reported that mutations in domain V of the 23S rRNA gene of M. pneumoniae were detected in 24 of 38 patients infected with M. pneumoniae using the Smart Gene 1 system and the results were confirmed by direct sequencing, revealing all mutations as A2063G [61].
In conclusion, the present study showed high sensitivity and high specificity of the Smart Gene 1 system for detecting M. pneumoniae from pharyngeal swab samples. The Smart Gene 1 system can detect the existence of M. pneumoniae and a point mutation at domain V of the 23S rRNA gene of M. pneumoniae at the same time. The entire procedure of the Smart Gene 1 system is completed within 50 min and the instrument is small (152 mm (W), 343 mm (D) and 300 mm (H)). The Smart Gene 1 system can also be useful in countries or areas where the macrolide-resistant rate of M. pneumoniae is currently low for monitoring the trend of macrolide resistance of M. pneumoniae. Therefore, the Smart Gene 1 system is suitable for point-of-care testing not only in a hospital setting but also in an outpatient setting.  Table. Lower limit of detection by the Smart Gene 1 system. The strain M. pneumoniae MHN8-014 harboring A at positions 2063, 2064 and 2067 (wild type) and the strain M. pneumoniae MHN8-016 harboring one transposition A2063G (mutant type) were used. Each strain was diluted with PPLO medium and the copy numbers of M. pneumoniae in PPLO medium were measured by qPCR [13,35,36]. Each strain of M. pneumoniae with 80, 40, 20, 10, 5, 2.5 and 1.25 copies per one μl was prepared by twofold serial dilution using the extraction reagent solution of the Smart Gene 1 system. The extraction reagent solutions containing different concentrations of M. pneumoniae were measured by the Smart Gene 1 system. The measurements were done in triplicate. After measurement by the Smart Gene 1 system, aliquots of the extraction reagent were used for confirmation of the copy number of M. pneumoniae measured by qPCR [13,35,36].