Rapid and accurate diagnosis of infection and control measures are critical for preventing the spread of Middle East respiratory syndrome coronavirus (MERS-CoV) [1]. For laboratory diagnosis of MERS-CoV infection, lower respiratory samples are preferred irrespective of the length of time between symptom onset and sample collection [2]. Real-time reverse transcription PCR (rRT-PCR) with sputum samples is commonly used to diagnose MERS-CoV infection.

If a negative result is obtained for patients with a high index of suspicion for MERS-CoV infection, additional samples should be collected and tested [3]. A series of negative results does not rule out the possibility of MERS-CoV infection [13]. Possible explanations for false-negative MERS-CoV test results include low viral load very early or late in the illness, inadequate sputum samples due to poor sample collection from patients with dry cough or no cough, inappropriate sample handling or shipping, and technical issues, such as viral genome mutation or PCR inhibition [3]. Sufficient nucleic acid extraction and the removal of substances that inhibit amplification are critical factors influencing the detection of MERS-CoV by rRT-PCR [4]. However, rRT-PCR of sputum samples yields conflicting results owing to the difficulty of extracting RNA from sputum containing mucus. Many reports have indicated that lower respiratory samples such as bronchoalveolar lavage (BAL) fluid and sputum contain the highest viral loads [3567]. However, one report indicated that no significant difference occurred in viral load cycle threshold (Ct) values between sputa and nasopharyngeal aspirates (NPAs) []. For this reason, sputum sample homogenization is desirable prior to isolating nucleic acids for MERS-CoV rRT-PCR. In this study, we compared three different sputum homogenization methods and identified an optimal homogenization method prior to the nucleic acid isolation from sputum samples.

According to Murray and Washington's grading system [9], among the 48 sputum samples tested, 18 (37.5%) belonged to group 1, 6 (12.5%) belonged to group 3, 2 (4.2%) belonged to group 4, 5 (10.4%) belonged to group 5, and 17 (35.4%) belonged to group 6. Thirty-two samples (66.7%), including 20 from groups 1–3, harbored a detectable quantity of ciliated columnar epithelial cells. Seventeen samples (35.4%) harbored identifiable proteinaceous mucous threads (viewed via light microscopy).

The efficacies of the three homogenization methods tested with NPA samples are summarized in Table 1. For DNA respiratory viruses (such as adenovirus and human bocavirus), PK-DNase treatment caused a slight increase in Ct values compared with PBS or NALC treatment. When we additionally tested six adenovirus-positive NPAs and five bocavirus-positive NPAs, the Ct values obtained following PBS preparation were 1.50–6.18 (median 5.42) and 1.60–7.61 (median 3.4) lower, respectively, than those obtained following PK-DNase preparation. Among the 10 samples derived from RNA respiratory viruses, the minimum Ct values resulted from PK-DNase treatment of three samples, PBS treatment of five samples, and NALC treatment of two samples.

Using identical volumes of sputum and inactivated viral culture diluted 1:10 or 1:1,000, the effects of each homogenization method on the MERS-CoV rRT-PCR results were evaluated (Table 2). Using a 1:10 dilution of inactivated viral culture, all samples treated with PK-DNase exhibited lower Ct values than those subjected to PBS or NALC treatment. The differences in Ct values between samples treated with PK-DNase or PBS were 5.4–7.8 for upE and 4.4–6.8 for ORF1a, while those for samples treated with PK-DNase or NALC were 4.7–5.8 for upE and 4.4–5.6 for ORF1a. Using the same volume of a 1:1,000 dilution of inactivated viral culture, one purulent sputum sample showed a false-negative result after PBS treatment. The Ct values of the other two sputum samples were similar following PK-DNase, PBS, and NALC treatments.

Using a one-tenth volume of a 1:1,000 dilution of inactivated viral culture (to simulate conditions comparable to those of real clinical samples), the effects of each method on the MERS-CoV rRT-PCR results were evaluated (Table 3). Among the three sputum samples, false-negative results were detected in two samples following PBS treatment and in three samples after NALC treatment (Table 3). Although the Ct values after PK-DNase treatment tended to be lower than those following PBS or NALC treatment, the values for BAL samples did not differ markedly depending on the homogenization method used.

To assess statistical significance, we tested 15 additional spiked samples, which were prepared by adding a one-tenth volume of 1:1,000 diluted viral culture (Table 4). Among the 15 spiked sputum samples prepared in duplicate (30 total samples), false-negative results were detected in five samples (16.7%) using the PBS method and in four samples (13.3%) using the NALC method. The ranges of Ct values for upE from sputum samples were 31.1–35.4 using the PK-DNase method, 34.7–39.0 using the PBS method, and 33.9–38.6 using the NALC method. Compared with the control, which were prepared by adding a one-tenth volume of 1:1,000 diluted viral culture to PBS solution, the ranges of Ct values following the PBS and NALC methods significantly differed from the mean Ct value (33.2) of the control (both P<0.0001).

For the extraction of MERS-CoV RNA from sputum sample, homogenization of sputum by PK-DNase I treatment was superior to treatment with PBS or NALC and sodium citrate. Proteinase K is a serine protease that exhibits broad cleavage specificity [11]. The enzyme cleaves peptide bonds adjacent to carboxylic group of aliphatic and aromatic amino acids and is useful for the general digestion of proteins in biological samples. Moreover, proteinase K degrades any RNases present in samples, preventing RNA degradation. Viral nucleic acids are protected from DNase I degradation by the viral protein capsids, although some enveloped viruses may be lost [12]. The genomes of DNA viruses such as adenovirus and bocavirus can be affected by proteinase K and DNase I, possibly owing to the digestion of viral capsid proteins by proteinase K and subsequent degradation of genomic DNA by DNase I.

Confirmation of probable or suspected MERS-CoV infection based on clinical and epidemiological criteria can only be accomplished through laboratory testing. However, other viral respiratory pathogens and bacterial pathogens should be considered during diagnosis [3]. When using lower respiratory samples originally earmarked for MERS-CoV testing, other viral respiratory pathogens should be tested for routine diagnostic processes. Because the ability to detect DNA respiratory viruses can be reduced by DNase I treatment, proteinase K treatment alone is worth considering for the routine detection of respiratory viruses from sputum samples.

The United States Centers for Disease Control and Prevention recommends that respiratory samples be extracted by using either the NucliSENS easyMAG system or MagNA Pure Compact Nucleic Acid Isolation kit [2]. Sputum samples categorized into groups 4–5 according to Murray and Washington's grading system [9] usually cannot undergo automated extraction owing to high viscosity and, thus, require a pre-homogenization process. Treatment with PBS or NALC and sodium citrate occasionally produces false-negative results with purulent sputum. Adding PBS and mixing via vortexing are necessary to culture respiratory viruses from sputum, and this procedure offers the advantage of involving minimum sample handling. However, thoroughly homogenizing thick or purulent sputum using this method is problematic. Because viral isolation by cell culture is not recommended for the routine confirmation of cases of MERS-CoV [23], the PBS plus vortexing method for processing sputum samples is not suitable for detecting MERS-CoV by rRT-PCR. Previously, it was reported that NALC and sodium citrate treatment (without NaOH) did not damage mycobacterial mRNA or rRNA in sputum samples [13]. NALC and sodium citrate solutions are readily available in clinical microbiology laboratories. Treatment with NALC and sodium citrate solution resulted in complete homogenization of sputum samples, including purulent sputum. However, this method is not suitable for detecting RNA targets in purulent sputum. Because proteinase K treatment requires heat inactivation of potential pathogens (incubation for 15 min at 55℃), this method offers additional laboratory safety.

Only 10% of sputum samples sent for MERS-CoV rRT-PCR testing were clinically relevant in terms of bacterial culture (group 5 according to Murray and Washington's grading system [9]). Among patients infected with MERS-CoV, a dry cough is more common than a productive cough [1617]. Therefore, optimum sample collection is not feasible for patients unable to produce sputum. However, two thirds of sputum samples in our study contained identifiable ciliated columnar epithelial cells. Dipeptidyl peptidase 4, the human receptor for MERS-CoV [16], is expressed in type I and II alveolar cells, ciliated or non-ciliated bronchial epithelial cells, and bronchial submucosal glands [1819]. Although it is difficult to obtain high-quality sputum from patients with no cough or a dry cough, obtaining and testing sputum samples containing ciliated columnar epithelial cells is recommended, if possible.

Our study has several limitations. First, it was performed by using spiked sputum samples; thus, our results may differ from those obtained by using authentic MERS-CoV-positive samples. However, the spiked sputum samples were similar to real sputum samples, as they were produced by adding a one-tenth volume of spiked viral solution to total sputum samples. Second, we tested a relatively small number of sputum samples under each test condition. Therefore, our findings should be confirmed by additional, larger studies.

In conclusion, the PK-DNase I treatment is suitable for the homogenization of sputum prior to RNA extraction for MERS-CoV rRT-PCR detection. Furthermore, adequate sputum homogenization prior to the isolation of nucleic acids is critical for accurate diagnosis of MERS-CoV infection.