Rapid diagnosis of Mycoplasma pneumonia infection by denaturation bubble-mediated strand exchange amplification: comparison with LAMP and real-time PCR

M. pneumoniae infection is often ignored due to its similar clinical symptom with respiratory tract infections caused by bacteria or viruses, and thus leading to misdiagnosis and delayed treatment. It is critical to develop a rapid, sensitive and specific diagnosis method. Denaturation Bubble-mediated Strand Exchange Amplification (SEA) was established, which is an isothermal method with only a primer pair and one Bst DNA polymerase. Notably, colorimetric SEA assay was developed with simple visual readout, making instrument-independent in detection step. The method could detect as low as 1.0 × 104 copies/mL genomic DNA within 60 min. Considering that more than 80% infected patients have 1.0 × 105−1.0 × 107 copies/mL M. pneumonia DNA, SEA is available for the practical diagnosis of M. pneumoniae in clinical specimens. Through comparing 224 sputum specimens, excellent performance of SEA assay with 90.48% sensitivity and 100% specificity relative to real-time PCR was observed. Compared with LAMP, a comparable sensitivity and low false positive rate was observed for SEA method. Therefore, SEA is a promising method for detecting M. pneumoniae directly from clinical specimens, which is especially suitable for point-of-care testing in primary care facilities and resource-limited settings with minimal equipment and technological expertises.

Scientific REPoRTS | (2019) 9:896 | DOI: 10.1038/s41598-018-36751-z methods 7 . Real-time PCR has been widely applied to detect M. pneumoniae clinically, which is more sensitive and do not require agarose gel electrophoresis; However, this method requires the complex operation, specialized instruments and trained personnel, and thus is time-consuming and high cost 8 . In recent years, a series of isothermal nucleic acid amplification techniques become more powerful alternative for point of care diagnosis, which can amplify nucleic acids at a single temperature and read out results using either real time fluorescent probes or colorimetric methods 9 . Among them, loop-mediated isothermal amplification (LAMP) is proven to have nearly the same sensitivity and specificity for detecting M. pneumoniae as PCR 10 , and thus is a useful technique for the rapid diagnosis of M. pneumoniae infection in clinical practice 11 . However, LAMP assay requires at least four primers targeting six specific regions, making primer design more complex; it is also highly susceptible to carryover contamination, leading to false positive results 12 . Recently, a novel isothermal amplification method, denaturation bubble-mediated strand exchange amplification (SEA) has been developed, which requires only a simple primer pair and utilizes successive natural strand "breathing" to form unwound double-stranded DNA rather than heat denaturation 13 . In general, dsDNA can dynamically dissociate due to ambient thermal fluctuations, even without the variation of temperature or pH, producing a single-stranded denaturation bubble 14 . Then, a short oligonucleotide primer could invade the denaturation bubble, triggering the extension by DNA polymerase to generate the amplification products. Considering that denaturation step is not required, the amplification can be performed at a single temperature below the melting temperature of dsDNA, eliminating the dependence on thermal cycler. More importantly, one pair of primers and one Bst DNA polymerase could trigger SEA reaction, simplifying reaction system. Considering the above advantages, SEA technology became another promising option for simple, rapid and sensitive diagnostic detection. In this study, we established a SEA method for rapid detection of M. pneumoniae and compared it with LAMP method and real-time PCR. The objective is to provide a rapid and simple detection method for early diagnosis of M. pneumoniae infection.

Results and Discussion
SEA assay for M. pneumonia. SEA assay has been accepted as a simple and rapid isothermal detection method since the first report in 2016 13,15 . SEA reaction requires only a pair of primers and targets a short sequence with 40-60 bp length, and thus it is easy to find a unique sequence as amplification target for a variety of pathogenic bacteria. Generally speaking, SEA assay has a promising application for clinical, food and environmental analysis. The workflow for the detecting of M. pneumoniae by SEA was shown in Fig. 1. The extracted genomic DNA from bacteria strains or sputum samples was used as the template for SEA reaction. The amplification was either fluorescently monitored or visually observed by colorimetric analysis. During the colorimetric detection, the color of positive reactions changed from orange to purple red, while negative reaction remains orange. Therefore, SEA assay provides a simple, rapid and specific method for detecting M. pneumoniae, which was suitable for on-site and rapid diagnosis.
The feasibility of SEA was firstly evaluated by detecting a vector plasmid DNA containing 16S rDNA sequence of M. pneumoniae M129 which served as a positive control and a sputum specimen infected by M. pneumoniae (Fig. 2). As expected, the significant fluorescence signals in the reactions of the positive control and the sputum specimen compared with the no-template control (NTC) indicated that SEA could effectively detect M. pneumoniae ( Fig. 2A), which was further verified by the expected 43 bp amplification products in native PAGE (Fig. 2B). These results demonstrated the feasibility and suitability of SEA for M. pneumoniae detection from clinical specimens. More importantly, SEA assay results could be directly observed by pre-adding colorimetric visual dye, by which positive reaction would be confirmed by color change from orange to purple red (Fig. 2C). As shown in Fig. 2C, the positive reactions developed a purple red color while reaction for NTC retained its orange color. Obviously, colorimetric SEA assay enabled result readout to realize the direct observation by the naked eye,  Fig. S1). Therefore, no cross-reactivity was found between M. pneumoniae and those tested pathogens, suggesting the good specificity of M. pneumoniae detection, which is particularly important for the differentiation of M. pneumoniae from other pathogens leading to respiratory tract infection.
To determine the sensitivity of SEA for detection of M. pneumoniae, serial 10-fold dilutions of plasmid DNA with target sequences ranging from 1.0 × 10 4 to 1.0 × 10 8 copies/mL were used as a template. As shown in Fig. 3A, the significant fluorescence signal increase was observed for the reaction with different concentration of target sequence. The minimum amount of DNA with real time fluorescence signal was 1.0 × 10 4 copies/mL within 60 min and colorimetric result were shown in Supplementary Fig. S5. The linear relationship was obtained between the threshold time value (T t ) and the concentration of target DNA ranging from 1.0 × 10 4 to 1.0 × 10 8 copies/mL (Fig. 3B), yielding a corresponding correlation coefficient (R 2 ) of 0.9875, which demonstrated the great stability of SEA assay to detect M. pneumoniae. Compared with the reported 2.2 × 10 3 copies/mL for LAMP assay and 1.1 × 10 3 copies/mL for real-time PCR 16,17 , a lower sensitivity was obtained for SEA reaction of M. pneumoniae detection. However, high sensitivity of LAMP assay often means high possibilities in false positive problem due to aerosol pollution 18 . As reported previously, LAMP assay was highly susceptibility to aerosol pollution, thus leading to false positive problem 18 . The false positive problem for LAMP assay was also found in the M. pneumoniae detection of our clinical specimens as discussed in section 3.4. In this study, SEA assay of M. pneumoniae could be completed within 60 min, showing that it is a time-efficient method.
Previous studies reported that M. pneumoniae carrier rates of 0.1-13.5% were detected in healthy individuals 3 . Asymptomatic carriage of M. pneumoniae must be considered when detecting M. pneumoniae infection in clinical specimens using amplification-based methods. Thus, it is crucial to establish a reliable diagnosis approach for M. pneumoniae infection which can distinguish real pathogens from asymptomatic carriage. A threshold of 10 4 copies/mL genomic DNA was proposed to distinguish clinical infection for M. pneumoniae from carriage 19 .   Fig. S4). In contrast, all 182 specimens with negative real-time PCR results were identified as M. pneumoniae negative with SEA assay, which corresponded to a specificity of 100.0%. To the best of our knowledge, this is the first study to evaluate SEA assay as a rapid diagnostic method for M. pneumoniae infection.
Although SEA assay is only 90.5% as sensitive as real-time PCR, 100.0% specificity demonstrated the reliability of detection results. In this study, 4 specimens were tested with real-time PCR-positive but SEA-negative results, leading to a low sensitivity of SEA assay compared with "gold standard" real-time PCR. There were two possible reasons for this result. Firstly, this result may be attributed to low M. pneumoniae load in these sputum specimens, which cannot be successfully detected by SEA assay. In addition, we could not exclude the possibility that these real-time PCR-positive specimens were actually innocent bystanders rather than real pathogens. Although SEA assay only showed the 90.5% sensitivity compared with real-time PCR, 100.0% specificity demonstrated that this method supports for clinical diagnosis with M. pneumoniae infection. More importantly, SEA assay had a low possibility for false-positive amplification. Therefore, SEA assay had a high performance for M. pneumoniae detection in clinical specimens. The threshold time (T t ) values obtained for SEA positive samples were compared with that of real-time PCR (Fig. 4). For real-time PCR, T t value is the corresponding time of cycle threshold (C t ). The T t values obtained for SEA showed high variability with a wide distribution range of T t values ranging from 14.36 to 52.21 min. SEA  Comparison of SEA method with LAMP. SEA assay for M. pneumoniae was also compared with LAMP method, which has been reported to be a useful isothermal detection method for M. pneumoniae with high sensitivity 23 . Of 224 sputum samples, 41 (18.30%) were positive in the LAMP assay, which is higher than that of 38 (16.96%) positive samples detected by SEA assay (Table 1). However, two of positive samples detected by "gold standard" real-time PCR were still positive in SEA assay but not in LAMP, implying false negative results for LAMP assay. Additionally, five of the specimens were negative in SEA assay but positive in LAMP, showing a low sensitivity of SEA method than LAMP with the values of 90.5% and 95.23% compared to real-time PCR, respectively. Of these five, one positive sample in LAMP was negative in both SEA and real-time PCR, implying the possibility of false positive. Although a comparable or higher sensitivity of LAMP with real-time PCR was also reported in other studies for M. pneumoniae detection 16,24 , LAMP method failed to detect 2 real-time PCR positive samples, resulting in an analytical specificity of 98.91%. Thus, it seems SEA assay showed a higher specificity than LAMP method when detecting M. pneumoniae in real clinical specimens. LAMP reaction is widely accepted to have greater or comparable sensitivity than real-time PCR due to its amplification mode by four or six primers 25 . The LAMP assay has been reported to detect M. pneumoniae as low as 200 copies/mL genomic DNA 24 and has been commercialized for M. pneumoniae diagnostic in Japan and China considering its high sensitivity, cost-effective and thermal cycler independence 25 . In practice, a high proportion of false-positive amplification has been reported for LAMP method and must be extensively retested to confirm specificity 16 . In LAMP reaction, false positive result usually occurs due to carryover contamination, wherein high sensitivity makes re-amplification easy to occur using amplicons from previous LAMP reactions as templates 26 . Another reason for false positive problem is LAMP-like amplification independent of template, which is driven by self-priming of generated stem-loops after LAMP autocycling is triggered 26 . In a previous study, about 5.9% false positive rate was reported for LAMP detection of M. pneumoniae directly from respiratory clinical specimens 16 . In this study, one specimen with LAMP positive result but SEA and real-time PCR negative result were confirmed to be false positive by PAGE gel electrophoresis (Supplementary Fig. S2). Obviously, fluorescent signal and amplification product were observed for LAMP reaction but not for SEA assay. Colorimetric assay for LAMP and SEA reaction also confirmed the concordant result ( Supplementary Fig. S2). Therefore, LAMP method produced a 2.4% false positive rate herein while SEA assay showed no false positive result. When compared T t values, no significant difference was found between SEA and LAMP for M. pneumoniae diagnostic from clinical specimens (p = 0.121), showing the median T t values of 30.5 and 35.1 min, respectively. LAMP method completed the reaction in 14 min -65 min, corresponding to T t values ranging from 10.3 min to 58.7 min (Fig. 4). Thus, the run time of SEA was comparable for LAMP method when detecting M. pneumoniae from clinical specimens. In summary, SEA method was a promising isothermal method, which possesses not only LAMP-similar advantage, such as reaction at constant temperature and independence of sophisticated instruments, but also possesses low false positive and high specificity compared with LAMP method. Compared with real-time PCR and LAMP, the most prominent advantage of SEA is its low cost due to simple reaction system and low requirements for instruments and professional personnel. The total costs for real-time PCR, LAMP and SEA assay were estimated based on labor costs and prices for reagents, supplies, and equipment maintenance. The regent cost of each run for real-time PCR and LAMP was estimated to be €2.8 and €8.4 respectively, whereas it only costs 28 cents for SEA regents. Labor and instrument costs would be higher for real-time PCR because it requires an expensive thermal cycler and trained professionals. In summary, SEA method is more cost-effective compared with real-time PCR and LAMP.
Considering that extracted clinical specimens frequently contain substances that inhibit enzyme based nucleic acid amplification processes, negative amplification test results do not necessarily indicate absence of infection 27 . It is of significant to introduce an internal control during practical M. pneumoniae diagnosis of clinical specimens. An internal control can help to identify inhibitory specimens by monitoring amplification of a second target nucleic acid. The negative result for the primary target could be validated by successful amplification of the internal control. For real-time PCR and LAMP, an internal control could be introduced, either by multiplexing and distinguishing between target and internal control amplifications (e.g. using probes with different reporter dyes) or by normalizing target Cts to the Cts of reference sequences that are spiked into the sample. Theoretically, introducing an initial control in SEA assay could be achieved by distinguishing amplicon annealing temperature (Tm) or threshold time value (Tt) from target sequence. Unfortunately, it is not possible to distinguish reference and target sequences in colorimetric SEA reaction. There is still a challenge to introduce an internal control in colorimetric SEA reaction.
In conclusion, the SEA method for rapid detection of M. pneumoniae from clinical specimens was established and compared with real-time PCR and LAMP. The SEA assay completed the reaction within 60 min. More importantly, colorimetric SEA assay further simplified readout procedure and enabled it suitable for point-of-care testing. Additionally, SEA assays had lower levels of false positive contamination relative to the LAMP assays and are more likely to be effective in resource-limited settings for improved pathogen detection. In summary, the SEA assay will enable rapid, low-cost and visual detection of M. pneumonia and have promising application potential for earlier recognition of outbreaks, especially in resource-limited settings. Nucleic acid extraction. Genomic DNA from sputum samples or reference strains was extracted using High Pure Viral Nucleic Acid extraction kit (Roche Applied Science, Mannheim, Germany). In brief, 200 µL sputum samples were added to 200 µL binding buffer supplemented with poly (A) and 50 µL Proteinase K. After incubation for 10 min at 72 °C, the mixture was transferred to a High Pure filter tube combined with a collection tube. The filter tube was centrifuged at 8 000 × g for 1 min and then 500 µL inhibitor removal buffer was added. After centrifugation at 8 000 × g for 1 min, the filter was washed twice with 450 µL washing buffer (20 mM NaCl and 2 mM Tris-HCl [pH 7.5] in ethanol). Finally, the pellet of crude nucleic acids was obtained by centrifugation at 12 000 × g for 1 min, and suspended in 50 µL elution buffer. SEA reaction. The SEA reaction was performed using the SEA detection kit and colormetric kit purchased from Navid Biotechnology Co., Ltd. (Qingdao, China). Briefly, the reaction was performed in a 10 µL mixture containing 1 µL of template, 1.5 × 10 −6 M of primer P1 and P2, 5 µL of 2 × reaction mix, and 0.5 × Eva Green. The reaction mixture was incubated at 65 °C for 60 min, and SEA amplifications were monitored by CFX Connect ™ Real-Time PCR System (Bio-Rad, CA, USA) at 1-min intervals. Alternatively, amplification products could be confirmed by the pre-addition of a pH sensitive dye and the color change could be directly observed by the naked eyes.