INTRODUCTION

A lower respiratory tract infection (LRTI) is an acute illness usually presenting with cough as the main symptom, and with at least one other LRT symptom such as fever, sputum production, dyspnea, wheezing, or chest discomfort/pain and no alternative explanation^[1]. LRTIs have a high mortality and morbidity worldwide, and The Global Burden of Disease Study reported that LRTIs are the second leading cause of deaths in 2013^[2]. The age-standardized mortality rate of LRTIs has been reported to be 41.7 in 100, 000 (95% CI 37.1-44.1)^[2].

A large and growing number of microbial pathogens, including bacteria, atypical agents, viruses, and fungi, have been reported to be responsible for LRTI^[3-5]. The clinical presentations of LRTI caused by different pathogens are very similar, making differentiation by clinical symptoms and pathology alone difficult.

Traditional methods used to identify respiratory pathogens include Gram stain of the sputum, culture of respiratory secretions and blood serological tests^[6]. Although these traditional methods have their advantages in identifying respiratory pathogens to some extent, the disadvantages cannot be ignored. Such as time-consuming, labor-intensive and poor sensitivity. Molecular methods such as polymerase chain reaction (PCR) and real-time PCR have been developed for the rapid detection and identification for respiratory pathogens^[7-10]. These methods have high sensitivity and specificity, and can be used to analyze clinical specimens, including blood, sputum, and urine. However, most molecular methods cannot detect more than five pathogens per reaction^[7-10]. Therefore, designing a rapid, labor-saving, and high-throughput method to detect and differentiate respiratory pathogens in a single reaction would be greatly useful in enhancing rapid response for the prompt treatment and control for LRTI.

The Applied Biosystems 3730-XL 96-capillary DNA analyzer, which has been designed by Applied Biosystems (ABI, USA) Company, is based on the technology of capillary electrophoresis separation. It allows for multiplexed detection of molecules, and various molecular targets can be detected by using four types of fluorescent tags in 96 samples within a single 96-well plate. Additionally, fourteen 96-well plates could be placed in this analyzer at the same time, and the analyzer can automatically detect the different plates one by one. Compared with other methods, this is a high-throughput, rapid detection method with high specificity and sensitivity, and products differing by as little as 7-10 bp in size can be separated by this system^[11].

In this study, a ABI 3730-XL DNA analyzer-based multiplex PCR and capillary electrophoresis (MPCE) assay was developed and applied for the simultaneous detection of 13 respiratory bacteria, including Streptococcus pneumoniae, Haemophilus influenzae, Moraxella catarrhalis, Pseudomonas aeruginosa, Klebsiella pneumoniae, Escherichia coli, Staphylococcus aureus, Mycoplasma pneumoniae, Legionella spp., Bordetella pertussis, Mycobacterium tuberculosis complex, Corynebacterium diphtheriae, and Streptococcus pyogenes. The sensitivity and specificity of the MPCE assay were tested, and 152 clinical samples collected from patients with LRTI were assayed for bacterial detection using this MPCE assay. The results using this method were then compared with results using the single real-time PCR assays mentioned below.

METHODS Bacterial Strains, Simulative Samples, Clinical Specimens, and Controls

The 13 respiratory bacteria detected in this study are listed in Table S1 available in www.besjournal.com.

Simulative samples were composed by mixing sputum from individuals who didn't have respiratory infection symptoms and 11 target bacteria (from the 13 except for M. tuberculosis and M. pneumoniae). The sputum from these individuals were confirmed negative for the 13 target bacteria using 13 single real-time PCR assays. Serial tenfold dilutions (10⁸-10⁰CFU/mL) of 11 bacteria (from the 13 target bacteria, except M. tuberculosis and M. pneumoniae) were mixed with the sputum samples (100 μL) separately. And the plate count method was used to measure bacteria in solution mentioned above (CFU/mL).

Then, 215 adult patients who were hospitalized in the Department of Respiratory Medicine, Shengjing Hospital, China Medical University, Shenyang, China, from May 2015 to January 2016, and were diagnosed with LRTI, including 180 cases of pneumonia, 14 cases of acute bronchitis, 17 cases of acute exacerbation of chronic obstructive pulmonary disease, and 4 cases of pulmonary abscess, were recruited to this study. A total of 215 clinical specimens were gathered from these 215 patients. The following criteria were used to diagnose LRTI: fever and/or an increased blood leukocyte count (≥ 11 × 10⁹/L), increased focal symptoms from the lower airways, with at least one of the following three newly developed symptoms: increased dyspnea, increased coughing, and/or increased purulent sputum^[12]. The patients were required to gargle with water in the morning before they spat. The second mouth sputum samples from deep lower respiratory tract were collected in sterile containers and immediately used for Gram staining and microscopic analysis. Samples with fewer than 10 epithelial cells and more than 25 leukocytes in each field at 100 × magnification were included in this study as available sputum samples^[13]. And finally, 152 sputum samples from 152 patients were eligible and tested in this study.

For the testing of clinical samples, the targeting 13 respiratory bacteria were used as positive controls, and the sputum from individuals which were confirmed negative for the 13 target bacteria using 13 single real-time PCR assays were used as negative controls. All these samples mentioned above were stored at -80 ℃ until tested.

DNA Extraction

The target DNA of 11 bacteria were extracted using a QIAamp DNA Mini Kit (Qiagen, Germany), according to the manufacturer's instructions. The 11 bacterial specimens were serially diluted tenfold from 10⁸ to 10⁰ CFU/mL. The clinical specimens were diluted 1:1 with an equal volume of phosphate-buffered saline and homogenized by vortexing for 1 min. Homogenized sputum was transferred into a microcentrifuge tube, and centrifuged at 8, 000 ×g for 5 min. Finally, 200 µL of homogenized sputum on the bottom including the pellet was used for DNA extraction. DNA extraction was carried out using a QIAmp DNA Mini Kit (Qiagen) according to the recommended protocol (Blood and Body Fluids Spin Protocol). The DNA was eluted in a final volume of 100 μL. The DNA of the simulative samples was extracted in the same manner as that described for the clinical specimens.

All the purified DNA, and the DNAs of M. tuberculosis and M. pneumoniae, which had been supplied by the National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention were stored at -20 ℃ in 1.5 mL microcentrifuge tubes until analysis. And DNA concentrations were measured by NanoDrop 2000 (Thermo Fisher Scientific, Shanghai, China)

Primer Design

Gene sequences for S. aureus were obtained through the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/). The S. aureus primers were designed in a highly conserved region for increased specificity and evaluated by the NCBI Primer-Blast and Primer Premier 5.0 software (Table S1). The primer sequences, their target genes, and the size of the resulting amplicons are shown in Table S1. All the forward primers were fluorescently labeled at the 5'-end with carboxyfluorescein (FAM).

Multiplex PCR for Bacterial Strains and Simulative Samples

First, primers for the target bacterial strains were confirmed by simplex PCR reactions. The mPCR was affected by many factors such as annealing time, annealing temperature, and primer concentrations. These conditions were optimized for the mPCR to ensure specific amplification of the selected targets. Finally, three mPCR reactions were built. The three mPCR systems are listed in Table S2 available in www.besjournal.com, and the three mPCR reactions were performed in a Thermo Cycler (SensoQuest LabCycler, Senso, Germany). For the first mPCR, cycling parameters were as follows: denaturation at 95 ℃ for 10 min; followed by 35 cycles of 95 ℃ for 30 s, 54 ℃ for 45 s, and 72 ℃ for 1 min; and a final extension at 72 ℃ for 10 min. For the second and the third mPCR reactions, cycling parameters were as follows: denaturation at 95 ℃ for 10 min; followed by 35 cycles of 95 ℃ for 30 s, 60 ℃ for 45 s, and 72 ℃ for 1 min; and a final extension at 72 ℃ for 10 min. The reaction products were resolved on a 2.0% agarose gel stained with GoldView, visualized under UV light, and analyzed using a Gel Doc system (BioRad, CA, USA).

Nested mPCR for Clinical Specimens

Nested mPCR was used to obtain the target DNA data from clinical specimens. After optimization, three nested mPCR reactions were built. All the primers used for the first round and the second round of these nested mPCR were just the same ones listed in Table S1. The three nested mPCR systems are listed in Table S4 available in www.besjournal.com, and the cycling parameters can be seen in Table S3 available in www.besjournal.com. And then the nested mPCR reactions were performed and analyzed as that described above.

Separation by Capillary Electrophoresis with the 3730-XL DNA Analyzer and Fragment Analysis

After amplification, 3 µL of each amplified product was mixed and homogenized by vortexing for 10 min. Then, 1 µL of the mixture was mixed with 8.5 µL of Hi-Di™ Formamide (Applied Biosystems) and 0.5 µL of GeneScan™ 1200 LIZ^® Size Standard (Applied Biosystems). This mixture was denatured for 5 min at 95 ℃ and cooled for 4 min at 4 ℃. The denatured DNA fragments were then separated by capillary electrophoresis using an ABI 3730-XL DNA analyzer (Applied Biosystems) and the results were examined using Genemapper software v. 4.0 (Applied Biosystems). The peak height of each gene was determined from the electropherogram.

Sensitivity of the MPCE Assay

To determine the relative sensitivities of the MPCE assay, serial tenfold dilutions (10²-10^-6ng/μL) of the 13 target DNA samples were extracted. Besides, simulative samples were mixed with different concentrations of bacteria (10⁸-10⁰ CFU/mL), respectively, and later total DNA was extracted, respectively. Moreover, different concentrations of 13 target DNA and simulative samples for 11 bacteria of the same order of magnitude were prepared separately and tested by the MPCE assay to ascertain the ability of this assay to co-amplify multiple gene targets and gene targets present at different concentrations in simulative samples.

Specificity of the MPCE Assay

The species used to assess the specificity of the assay are listed in Table S1. For each pathogen, monoplex PCR assays were performed with its target DNA, and negative controls (other DNA except the target template) were simultaneously tested for each organism. To determine the specificity of the MPCE assay, DNA of the 13 target bacteria were analyzed in the 3 mPCR systems, and after amplification, the 13 mPCR products (5 μL each) were mixed and analyzed again in the capillary electrophoresis system.

Detection in Clinical Samples

A total of 152 sputum samples were collected from Shengjing Hospital, China Medical University in Shenyang, China. DNA was extracted from the samples and tested by both the MPCE (nested mPCR) assay and single real-time PCR assays.

Real-time PCR for Bacterial Detection

The 13 pairs of specific primers and probes and amplification conditions used in 13 single real-time PCR reactions are shown in Table S5 available in www.besjournal.com.

We used a Stratagene Mx3000P (Stratagene, La Jolla, CA) for the real-time PCR assay. The reaction system consisted of 12.5 μL of Premix Ex Taq (Takara) and 1.25 μL of each primer and fluorescent probe. The final volume of the system was adjusted to 25 μL with 0.5 μL of ROX Reference Dye II (Takara) and distilled water. The specificities of the 13 real-time PCR reactions have been proven in previous studies^[25-33], and the relevant references have been listed in Table S5 available in www.besjournal.com. The analytical sensitivities for these 13 pathogens were measured by 10-fold serial dilutions of each target DNA from 10⁸-10⁰ CFU/mL. The threshold cycle (CT) values of all the dilutions for these 13 pathogens were confirmed in triplicate.

Statistical Analyses

Consistency between the results of the MPCE assay and real-time PCR assays was verified using Cohen's kappa test in SPSS 17.0 software (IBM SPSS, United States). The kappa value was graded as follows: < 0, no agreement; 0-0.20, slight; 0.21-0.40, fair; 0.41-0.60, moderate; 0.61-0.80, substantial; and 0.81-1, almost perfect agreement.

RESULTS Establishment of the Three mPCR Assays

The three mPCR reactions were optimized by evaluating the primer concentrations, annealing time, and temperature, and no cross-amplification was observed (Figure 1).

Specificity of the MPCE Assay

The target DNA samples of all the 13 pathogens were used to evaluate the specificity of the primers. The results revealed that each primer pair could detect the targeted pathogen without cross-amplification. First, all bacterial targets were individually tested in the MPCE assay, and the results are presented in Figure 2.

Second, to further determine the specificity of the MPCE assay, DNA of the 13 target bacteria were analyzed in three mPCR systems, and the resultant amplified mPCR products were mixed and analyzed in the capillary electrophoresis system, and the results are presented in Figure 3. As shown in Figure 3, the resultant amplicon sizes for the 13 pathogens in the mixture were similar to their corresponding target sequences (Table S1), verifying the specificity of this method.

Sensitivity of the MPCE Assay

The sensitivity of the MPCE assay was examined. The mPCR reaction products obtained from three mPCR assays were mixed with the same order of magnitude of each sample and tested in an ABI 3730-XL DNA analyzer. The detection limits in the MPCE assay can be seen in Table 1 and Figure 4. The detection limits for these pathogens ranged from 10^-7 to 10^-2 ng/μL. The lowest detection limit was 10^-7 ng/μL for S. pneumoniae.

Application to Simulative Samples

Simulative DNA samples for 11 bacteria were extracted and analyzed in three mPCR assays, respectively. And the mPCR reaction products obtained from three mPCR assays were mixed with the same order of magnitude of each sample and tested in an ABI 3730-XL DNA analyzer. As shown in Table 1 and Figure 5, the detection limits for these simulative samples ranged from 10^-7 to 10^-2 ng/μL. The lowest detection limits was 10^-7 ng/μL for S. pneumoniae.

Application to Clinical Samples

A total of 152 clinical samples were assayed using the three nested mPCR assays, and tested in an ABI 3730-XL DNA analyzer. And the results were confirmed by the simplex real-time PCR for all the 13 species listed in Table S5. The MPCE (nested mPCR) assay successfully detected and differentiated all 13 target respiratory pathogens. Compared with the results of the simplex real-time PCR, the consistency of MPCE assay exceeded 96.1% for all the 13 target pathogens (Table 2).

Of the 152 clinical samples, the MPCE assay dete-cted S. pneumoniae, M. pneumoniae, M. tuberculosis, C. diphtheriae, and S. pyogenes in more samples than the real-time PCR assay. Furthermore, the MPCE assay detected B. pertussis in fewer samples than the real-time PCR assay. The MPCE assay and the simplex real-time PCR assay both identified the same number of positive samples for H. influenzae, M. catarrhalis, P. aeruginosa, K. pneumoniae, E. coli, Legionella spp. and S. aureus. In this study, some samples were detected positive by the MPCE assay but negative by the real-time PCR assay. These samples were then confirmed as true positives by combination of simplex PCR and capillary electrophoresis with the primers designed for the target genes. Besides, these samples were also confirmed as true positives by 16SrRNA sequence analysis. A total of 152 clinical samples were assayed using the MPCE assay, and coinfections were identified in 17 samples.

The obtained sensitivity and specificity of this MPCE assay relative to those of real-time PCR were as follows: 100.0% and 96.5% for S. pneumoniae, 100.0% and 99.3% for M. pneumoniae, 100.0% and 99.3% for M. tuberculosis, 100.0% and 96.6% for C. diphtheriae, 100.0% and 96.9% for S. pyogenes, 100.0% and 100.0% for H. influenzae, M. catarrhalis, P. aeruginosa, K. pneumoniae, E. coli, S. aureus, Legionella spp., and B. pertussis, respectively.

The detection results of our MPCE assay were also compared with conventional diagnostic methods used in our hospital, such as microscopy, blood serological tests, and urinary antigen tests. For M. tuberculosis, nine of the 152 clinical samples detected positive by MPCE assay were also confirmed positive by Ziehl-Neelsen stain. For S. pneumoniae, 41 clinical samples detected positive by MPCE assay were also confirmed positive by urinary antigen test (Binax). For Legionella spp., three clinical samples detected positive by MPCE assay were also confirmed positive by urinary antigen test (Binax). For M. pneumoniae, seven clinical samples detected positive by MPCE assay were also confirmed IgM positive by paired blood serological tests (SERODIA-MYCOII, Tokyo, Japan). These results demonstrated that the MPCE assay was comparable to the conventional diagnostic methods used in the hospital.

DISCUSSION

S. pyogenes, M. tuberculosis, M. catarrhalis, M. pneumoniae, S. pneumoniae, S. aureus, C. diphtheriae, H. influenzae, K. pneumoniae, P. aeruginosa, Legionella spp., B. pertussis, and E. coli are the most commonly reported causative pathogens in LRTI worldwide^{[1-2, 4-5]}. All these bacteria have similar clinical manifestations, and may present as multiple infections, further complicating the LRTI and increasing its morbidity. Conventional diagnostic methods such as isolation and serological investigations for identifying these causative bacteria are laborious and time-consuming. Furthermore, in the case of multiple infections, conventional diagnostic methods need to be carried out separately for each pathogen. Many molecular tests, including single PCR, multiplex PCR, single real-time PCR, and multiplex real-time PCR, for the detection of several of the above mentioned LRTI agents have been developed^{[10, 16, 27, 29-30]}. However, these methods can detect only a limited number of bacterial species at a time^[7-10]. Therefore, there is an urgent need for a high-throughput and rapid method that can detect multiple causative agents in LRTI for effective disease surveillance and disease management.

The MPCE assay based on the technology of capillary electrophoresis separation is a high-throughput, rapid detection method with high specificity and sensitivity, and products differing by as little as 7-10 bp in size can be separated by this system^[11]. Additionally, this assay is very useful because it enables the simultaneous detection of a broad range of bacterial agents related to respiratory infections, which is an optional feature, meaning that it can be used as a single-target assay for each pathogen or organized into different multiplex assays for different diagnostic purposes, as required. Moreover, this method can also be expanded to include more pathogens, further increasing its abilities. For example, one way to achieve this is to use different fluorescent dyes as well, so the different pathogens could be separated by both size and fluorescence, similar to the strategy that was described before^[34]. Previously, capillary electrophoresis was successfully applied to investigate ethnic variations of polymorphism and to analyze the differences in genetic polymor-phism^[35-36]. Recently, capillary electrophoresis has been used for the rapid identification of Candida fungal infections and for the multiplex detection of nine foodborne pathogens^[37-38]. To our knowledge, this is the first study to investigate the microbial pathogens responsible for LRTI using this method.

In this study, 13 specific primer pairs were used to develop a MPCE assay based on the ABI 3730-XL analyzer for the simultaneous detection of 13 common LRTI pathogens, namely, S. pyogenes, M. tuberculosis, M. catarrhalis, M. pneumoniae, S. pneumoniae, S. aureus, C. diphtheriae, H. influenzae, K. pneumoniae, P. aeruginosa, L. pneumophila, B. pertussis, and E. coli. Compared with the results of the real-time PCR reactions (Table 2), the consistency of the MPCE assay exceeded 96.1% for all 13 target pathogens, which demonstrated that the MPCE assay had comparable detection to real-time PCR. In addition, the detection limits for most of the pathogens in our assay are very low. For example, 10^-7 ng/μL for S. pneumoniae; 10^-6 ng/μL for M. pneumoniae and S. pyogenes; 10^-5 ng/μL for M. tuberculosis; 10^-4 ng/μL for H. influenzae, M. catarrhalis, S. aureus, and Legionella spp.

Another advantage of the MPCE assay is that it is a rapid and high-throughput method. The entire reaction can be completed within 4.5 h. And the analyzer can detect fourteen 96-well plates with 96 samples one by one automatically.

However, some of the samples that were found to be positive by the single real-time PCR assay tested negative by the MPCE assay. The reason for the low detection rate of B. pertussis in our assay may be related to the PCR primers used, which targeted the B. pertussis S1 subunit, which is important for the detection of toxin-producing B. pertussis, while the primers in the real-time PCR assay targeted IS481, which is the pertussis toxin promotor region specific to B. pertussis. To improve the specificity of our assay, new primers with high specificity and sensitivity should be designed in the future. In this study, we used sputum samples to verify the validity of the MPCE assay, and we tried our best to avoid contamination in the process of collecting the samples. Although this method has its limitation in etiological diagnosis using sputum samples, the MPCE assay is still a newly built method with high-throughput and is easy to operate with high reliability. Based on the findings of the present study, this is a promising new method with potential for identifying respiratory pathogens in LRTI.

CONCLUSIONS

In conclusion, this study illustrates that the MPCE assay is a rapid, reliable, labor-saving, and high-throughput method with high specificity and sensitivity for the identification of 13 common respiratory pathogens responsible for LRTI. This assay has great potential in the molecular epidemiological survey of respiratory pathogens.

ETHICS APPROVAL

The research protocol for the current study has been approved by The Ethics Committee of the Shengjing Hospital of China Medical University, Shenyang, Liaoning, China (2016PS238K).

CONFLICT OF INTEREST

All the authors declare that there are no conflicts of interests.

AUTHORS' CONTRIBUTIONS

JLX carried out the experiments of this study, except those documented in Figure 1 which were performed by RHY, and took part in writing the manuscript. ZHJ participated in data analysis and took part in writing the manuscript. CY and QT participated in conceiving and coordinating experimental work, and took part in writing the manuscript. And they also provided materials for this study. ZSH and HBY collected all the sputum samples for the study, and wrote the manuscript. All authors read and approved the final manuscript.