INTRODUCTION
On 20 September 2012, a report appeared on ProMED-mail (
http://www.promedmail.org/direct.php?id=20120920.1302733) of a novel human coronavirus (CoV) isolated several months earlier from a hospitalized patient in Saudi Arabia who had died of severe respiratory complications (
1). Like the severe acute respiratory syndrome (SARS)-CoV, this new virus was most closely related to known bat coronaviruses but was genetically distinct, being classified phylogenetically in the group 2C coronavirus clade (
2).
This virus was subsequently named the Middle East respiratory syndrome (MERS)-CoV because of its geographic predilection (
3), and the genomic sequence obtained from this isolate was used to develop real-time reverse transcription (rRT)-PCR assays that were released on the Eurosurveillance website on 27 September 2012 (
4). These assays, targeting regions upstream of the envelope gene (upE) for specimen screening and open reading frames (ORFs) 1b and later 1a (
5) for test confirmation, have been used extensively to investigate the emergence of this new virus. As of 4 October 2013, 136 laboratory-confirmed cases of MERS-CoV infection, including 58 deaths, have been reported from 8 countries in the Middle East and Europe, primarily using these assays (
http://www.who.int/csr/don/2013_10_04/en/index.html).
On 25 September 2012, Christian Drosten at the University of Bonn Medical Center kindly provided the U.S. Centers for Disease Control and Prevention (CDC) with sequence data for the MERS-CoV nucleocapsid (N) protein gene in advance of publication. Based on this sequence, the CDC quickly developed several rRT-PCR assays targeting the N gene to support the public health response to MERS-CoV. This report describes the validation of these assays and presents comprehensive data on the performance of the published upE assay using multiple specimen types.
(Some data from this study were presented at the 29th Clinical Virology Symposium, Daytona Beach, FL, 28 April to 1 May 2013.)
MATERIALS AND METHODS
Viruses and clinical specimens.
MERS-CoV strain Jordan-N3/NCV (2012905864/VeroP1) was kindly provided by U.S. Naval Medical Research Unit 3 (NAMRU-3) (Cairo, Egypt), with permission from the Jordan Ministry of Health (MOH). Other high-titer respiratory virus stocks and virus-positive and -negative clinical specimens used for assay specificity studies were available from CDC collections. Extracts from pooled nasal wash specimens predicted to contain diverse human microbiological flora from 20 consenting healthy new military recruits were kindly provided by Lisa Lott, Eagle Applied Sciences (San Antonio, TX).
A total of 336 diverse fresh or frozen clinical specimens collected between April 2011 and April 2013 from 321 persons who had severe acute respiratory illness (SARI) and either were resident in or had a history of travel to the Middle East were available for testing. Of these, 280 were combined nasopharyngeal (NP)/oropharyngeal (OP) swab specimens collected in viral transport medium from hospitalized Jordanian children <2 years of age (
15), with most of the remaining specimens being from adults. A bronchoalveolar lavage fluid sample and a serum specimen collected by the Jordan MOH Central Public Health Laboratory staff from two fatal SARI cases from a MERS-CoV pneumonia outbreak cluster at a Jordanian hospital in April 2012, and independently confirmed as positive for MERS-CoV by culture and/or sequencing by NAMRU-3, were also available for testing.
MERS-CoV culture.
On receipt of the virus at the CDC, Vero E6 cell monolayers were inoculated and observed daily for cytopathic effect (CPE). At 3 to 4+ CPE, the cell culture lysate was recovered, divided into aliquots of small volumes, and stored at −70°C or below. The titers of this stock virus were determined, and the 50% tissue culture infective dose (TCID50) was calculated using standard methods (stock titer, 1.3 × 104 TCID50/ml). Stock virus used in the spiking experiments described below was inactivated by gamma irradiation, and the sequence was confirmed over the rRT-PCR signature regions.
Sample processing and nucleic acid extraction.
For sputum samples or other lower respiratory tract specimens too viscous for downstream nucleic acid extraction, the sample was added to an equal volume of 500 mM freshly prepared No-Weigh dithiothreitol (catalog no. 20291; Pierce) and incubated at room temperature, with intermittent mixing, for 30 min or until the sample was sufficiently liquefied for processing. For stool specimens, 10% suspensions were prepared by adding 100 μl of liquid stool or a pea-sized amount of solid stool to 900 μl of phosphate-buffered saline (pH 7.4; Gibco), pulse vortex mixing the mixture for 30 s, and centrifuging the mixture at 4,000 × g for 10 min at 4°C. The clarified supernatant was then carefully removed for extraction. Total nucleic acid extractions were performed on 200 μl of sample using the NucliSens easyMAG system (bioMérieux, Durham, NC), following the manufacturer's default instrument settings, and 100-μl elution volumes were collected. For some comparison studies (see below), simultaneous extractions were also performed with the MagNA Pure Compact system, using nucleic acid isolation kit I (Roche Applied Science). Extracts were either tested immediately or stored at −70°C or below until use.
Primers and probes.
Multiple primer/probe sets targeting regions in the 3′, middle, and 5′ regions of the N gene sequence (GenBank accession no. JX869059.2) were designed using Primer Express software, version 3.0 (Applied Biosystems, Foster City, CA). Primer/probe sets were predicted to specifically amplify MERS-CoV with no major combined homologies with other coronaviruses or human microflora on BLASTn analysis that would potentially yield false-positive test results. All primers and probes were synthesized by standard phosphoramidite chemical techniques at the CDC Biotechnology Core Facility. Hydrolysis probes were labeled at the 5′ end with 6-carboxyfluorescein (6-FAM) and at the 3′ end with Black Hole Quencher 1 (BHQ1) (Biosearch Technologies, Inc., Novato, CA). Optimal primer/probe concentrations were determined by checkerboard titrations. Primers/probes with the highest amplification efficiencies with RNA transcripts (see below) were retained for further study (
Table 1).
In vitro RNA transcripts and viral template control.
Single-stranded DNA oligonucleotides covering the amplified region of each rRT-PCR signature and containing a 5′ T7 RNA polymerase promoter sequence (TAATACGACTCACTATAGGG) were synthesized. The oligonucleotides were amplified using the 5′ T7 promoter sequence as the forward primer, with the corresponding rRT-PCR reverse primer for each signature (
Table 1). Amplification products were transcribed using a MEGAshortscript high-yield transcription kit (Invitrogen/Life Technologies). The RNA transcripts were purified using a MEGAclear kit (Invitrogen/Life Technologies) and were quantified by UV spectroscopy. A MERS-CoV viral template control (VTC) was prepared by combining the 3 signature templates with human genomic DNA (Promega) and then drying the mixture into a visible pellet with Pellet Paint Co-Precipitate (EMD Millipore) to create a thermostable product.
Real-time RT-PCR assay.
The rRT-PCR assay was performed using the Invitrogen SuperScript III Platinum One-Step quantitative RT-PCR system (Life Technologies). Each 25-μl reaction mixture contained 12.5 μl of 2× master mix, 0.5 μl of SuperScript III reverse transcriptase/Platinum Taq DNA polymerase, 0.5 μl of probe, 0.5 μl each of the forward and reverse primers, 5.5 μl of nuclease-free water, and 5 μl of nucleic acid extract. Amplification was carried out in 96-well plates on an Applied Biosystems 7500 Fast Dx real-time PCR instrument (Life Technologies). Thermocycling conditions consisted of 30 min at 50°C for reverse transcription, 2 min at 95°C for activation of the Platinum Taq DNA polymerase, and 45 cycles of 15 s at 95°C and 1 min at 55°C. Each run included one viral template control and at least two no-template controls (NTCs) for the sample extraction and reaction set-up steps. A positive test result was defined as a well-defined exponential fluorescence curve that crossed the threshold within 45 cycles. Positive viral template control (VTC) and no-template control (NTC) samples were included in all runs to monitor assay performance. All specimens were tested for the human RNase P (RP) gene by rRT-PCR to monitor nucleic acid extraction efficiency and the presence of PCR inhibitors.
DISCUSSION
In response to the emergence of MERS-CoV in the Middle East and its spread to several European countries, the U.S. Health and Human Services announced on 29 May 2013 that the virus posed a significant public health threat to U.S. citizens. On 5 June 2013, the U.S. Food and Drug Administration authorized emergency use of the CDC rRT-PCR assay as an
in vitro diagnostic test for the presumptive detection of MERS-CoV in patients with clinical signs and symptoms of MERS-CoV infection, in conjunction with clinical and epidemiological risk factors (
http://www.fda.gov/MedicalDevices/Safety/EmergencySituations/ucm161496.htm). Reagent kits were distributed by the CDC Laboratory Response Network to state public health departments and to select U.S. Department of Defense surveillance laboratories equipped to perform assays. The assay was also distributed to international public health partners in the affected region and to countries with extensive travel to and from the Middle East.
Our assay design and validation strategy were guided by several principles. First we chose to retain the upE signature designed by Corman et al. (
4) due to its wide and successful use in MERS-CoV surveillance. A second signature developed by those authors to confirm positive upE test results, targeting the MERS-CoV 1b open reading frame (ORF), proved less sensitive than upE in comparison studies (
4) and was not adopted; another assay signature, targeting ORF 1a, which was claimed to be as sensitive as upE, was later introduced (
5). As an alternate testing strategy, we introduced two new signatures targeting the MERS-CoV nucleocapsid (N) gene; one assay (N2) was combined with upE testing to enhance sensitivity for specimen screening, and the second assay (N3) was reserved for positive test confirmation. Theoretically, rRT-PCR assays targeting the MERS-CoV N gene should offer enhanced diagnostic sensitivity due to the relative abundance of N gene subgenomic mRNA produced during virus replication, although we found no clear evidence of this in our study and this was not shown in practice for clinical diagnosis of SARS-CoV (
6). Validation of all assay signatures was conducted with multiple specimen types, including upper and lower respiratory tract specimens, serum samples, and stool specimens, all shown to be diagnostically valuable for SARS-CoV (see below). Finally, we chose to validate the assay using instruments and reagents in common use by U.S. state and international public health laboratories, to minimize the occurrence of off-protocol use of the test.
Although the MERS-CoV rRT-PCR assay panel proved both sensitive and specific, the study was subject to several limitations. First, only two authentic specimens from patients with independently confirmed MRS-CoV infection were available for testing. Most data were derived from mock specimens spiked with cultured virus, which may not accurately replicate specimens obtained during natural virus infections. Also, spiked mock specimens were not subjected to the same collection, handling, and storage conditions to which authentic specimens would be subjected, which might negatively affect virus detection. Moreover, we cannot be certain that the specimens collected from other suspected MERS-CoV cases were truly negative for the virus. However, patient demographic and clinical features, evidence of infection with other respiratory pathogens, confirmed MERS-CoV seronegativity in some cases, and the self-validating negative test results obtained with all three assay signatures support this assumption. Finally, assay validation was necessarily limited to the use of specific instrumentation, reagents, and procedures. Use of different assay platforms or modifications in methodology could negatively affect assay performance.
The choice of appropriate specimen type, collection technique, and timing after the onset of symptoms is also critical for diagnostic success. Among some patients, MERS-CoV appears to be detected more often and with higher viral loads in lower respiratory tract specimens than in specimens from the upper respiratory tract (
7,
8); consequently, lower respiratory tract specimens have been prioritized for collection by the WHO and the CDC (
http://www.who.int/csr/disease/coronavirus_infections/update_20121221/en/). Other specimen types, such as serum/blood samples and stool specimens, may also prove valuable. Studies performed during the SARS epidemic found that SARS-CoV could be detected in serum/blood samples during the early prodromic phase of infection (
9,
10) and was shed for prolonged periods at high titers in stool, facilitating detection later in the course of illness (
11,
12). MERS-CoV RNA was reported to be detected in stool and urine specimens from one infected immunosuppressed patient (
7), and the virus has been identified in serum samples from other patients (this study). Testing of more samples from additional patients infected with MERS-CoV, at all stages of illness, is essential to guide testing strategies. When respiratory specimens are collected late or are not available for molecular testing, serological testing may be an effective diagnostic alternative for MERS-CoV (
13).
Even when optimal specimens and molecular tests are available, accurate diagnosis of MERS-CoV infection can still be challenging. Although rRT-PCR tests are less susceptible to amplicon contamination than are conventional RT-PCR assays, false-positive rRT-PCR results can still occur if practices designed to minimize the risk of contamination are not stringently followed. Access to multiple rRT-PCR assays targeting different regions of the MERS-CoV genome, with some assays kept in reserve for positive test result confirmation, is essential to prevent misidentifying MERS-CoV cases. It is also recommended that laboratories conducting MERS-CoV surveillance partner with the CDC or another qualified reference laboratory that can independently confirm rRT-PCR positive results by sequencing. While rRT-PCR assays are relevant for rapid diagnosis and patient management, genomic sequencing can provide public health authorities with needed confidence for response planning, can help avoid false alarms, and can provide data essential for monitoring both virus evolution and rRT-PCR assay signature integrity.
ACKNOWLEDGMENTS
We thank Christian Drosten from the University of Bonn Medical Center for providing the MERS-CoV nucleocapsid gene sequence. We also thank others at the CDC who contributed substantially to this effort, including Kanwar Bedi, Willie Betts, Michael Farrell, Matthew Marcum, Eileen Schneider, Yvonne Stifel, Nicky Sulaiman, and Azaibi Tamin.
The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the CDC or the U.S. Department of Health and Human Services (DHHS). Names of specific vendors, manufacturers, or products are included for public health and informational purposes; inclusion does not imply endorsement of the vendors, manufacturers, or products by the CDC or the DHHS.