ABSTRACT
The Parechovirus genus of the Picornaviridae family contains two species, Human parechovirus (HPeV) and Ljungan virus (LV). The HPeVs (including the former echoviruses 22 and 23, now HPeV type 1 (HPeV1) and HPeV2, respectively) cause a wide spectrum of disease, including aseptic meningitis, gastroenteritis, encephalitis, acute respiratory illness, and neonatal sepsis-like disease. The LVs were isolated from bank voles in Sweden during a search for an infectious agent linked to fatal myocarditis cases in humans. Because of the decline in use of cell culture and neutralization to investigate enterovirus-like disease, very few laboratories currently have the capability to test for parechoviruses. We have developed a real-time reverse transcription-PCR (RT-PCR) assay for detection of all known members of the genus Parechovirus. The assay targets the conserved regions in the 5′ nontranslated region (5′NTR) of the parechovirus genome and can detect both HPeVs and LVs, unlike other published parechovirus 5′ NTR assays, which only detect known HPeVs or only LVs. HPeV and LV can be differentiated by sequencing the 5′NTR real-time RT-PCR amplicon, when needed. The assay is approximately 100 times more sensitive than cell culture and may be used to test original clinical specimens. The availability of a broad-specificity PCR method should facilitate the detection of new human parechoviruses, as well as new parechoviruses in other mammalian species, and provide an opportunity to investigate the role of these viruses in human and animal disease.
The genus Parechovirus (family Picornaviridae) contains two species, Human parechovirus (HPeV) and Ljungan virus (LV). The parechovirus genome is a positive-sense, single-stranded RNA of ∼7,300 bases for HPeV and ∼7,600 bases for LV. Six HPeV serotypes have been described. HPeV type 1 (HPeV1) (formerly echovirus 22) and HPeV2 (formerly echovirus 23) were first isolated in 1956 and originally classified as enteroviruses (32), while HPeV3 to -6 were described only recently (1, 2, 4, 11, 31). The LVs were isolated from bank voles (Clethrionomys glareolus) in Sweden (21). Molecular and physical characterization identified LV as a novel parechovirus that is related to but distinct from the HPeVs (13). LV has also been isolated from North American rodent species (12, 18).
The spectrum of clinical syndromes associated with the HPeVs mirrors that of the human enteroviruses and includes respiratory and gastrointestinal diseases, aseptic meningitis, myocarditis, encephalitis, acute flaccid paralysis, and neonatal sepsis (29). A WHO study of 581 cases of HPeV1 infection showed that 29% of patients had gastrointestinal symptoms, 26% had respiratory symptoms, 12% had central nervous system symptoms, and 6% had cardiac, skin/mucosa, or muscular symptoms; the remaining 27% were classified as having other predominant signs or symptoms (10). Two seroepidemiologic studies, in Japan and Finland, indicated that children are infected with HPeVs early in life, reaching a high rate of seroprevalence by the time they reach school age (15, 30). The apparent high level of infection with HPeVs in young children indicates that, like enterovirus infections, many HPeV infections are probably asymptomatic or result in only mild illness.
It has been suggested that LV may be involved in the etiology of a variety of human diseases, including type 1 diabetes mellitus, myocarditis, and Guillain-Barre syndrome, because the incidence of these diseases in northern Sweden rises and falls in parallel with the cyclic population changes observed in the small rodent species from which LV has been isolated (20). Up to one-third of captured bank voles develop type 1 diabetes, and diabetes in voles can be induced by LV infection (19). A recent study used immunohistochemistry to detect LV antigen in brain and placental tissue in cases of intrauterine fetal death in humans (22). The possibility of zoonotic transmission of LV to humans and association with human disease remains under investigation.
HPeV1 and HPeV2 have traditionally been identified by neutralization assay following virus isolation in cell culture, using standardized antiserum pools (17), but the parechoviruses often grow poorly in culture and antigenic typing reagents are not widely available for HPeV3 to -6 and LV. Increasingly, parechoviruses are being detected by nucleic acid amplification (2, 3, 5, 7, 8, 14-16, 24-26, 28).
Despite the development of molecular detection assays, very few clinical laboratories currently have the capability to test for parechoviruses. In addition, the published molecular methods are unable to detect all known parechoviruses in a single assay. As a result, the true burden of HPeV disease is probably underestimated. The aim of this work was to provide a flexible, reliable, and sensitive PCR-based assay to test for the presence of all known parechoviruses, both HPeVs and LVs, in primary clinical specimens. To that end, we have developed a real-time PCR assay targeting the conserved parechovirus 5′ nontranslated region (5′NTR), allowing the sensitive detection of HPeV1 to -6 and LVs in a variety of original specimen types.
MATERIALS AND METHODS
Virus strains.The virus strains used to test primer/probe specificity are listed in Table 1. The strains include reference strains for HPeV1, HPeV2, HPeV4, HPeV5, and LV and representative clinical isolates for HPeV3 and HPeV6. An additional 48 HPeV isolates from the CDC virus collection were tested to further demonstrate the specificity of the primers/probe (Table 1). Other picornavirus genera tested included the cardioviruses, encephalomyocarditis virus (Rueckert strain), and Theilers' murine encephalomyelitis virus (strain GDVII) and representative enterovirus strains from species A (coxsackievirus A16), species B (echovirus 30), species C (coxsackievirus A24), and species D (enterovirus 68).
Reference strains and representative clinical isolates used to test assay primers and probe specificity for HPeVs and LVs
Standardized virus stocks of HPeV1-Harris and LV-SWE87-012 were grown in 25-cm2 flasks (RD cells for HPeV1 and Vero cells for LV) for assay sensitivity experiments. The viruses were harvested at 60 to 70% cytopathic effect by freeze-thawing the flasks two times, collecting the supernatants, and then clarifying the supernatants by centrifugation at 10,000 × g for 10 min. The clarified virus stocks were aliquoted and frozen at −70°C. Both virus stocks were serially diluted 10-fold and titrated on their respective cell lines to determine end point titers, which were expressed in units of 50% cell culture infectious dose (CCID50) (27). RNA was extracted from the stocks and serially diluted to contain from 103 CCID50 to 10−4 CCID50 per 5 μl.
Clinical specimens.HPeV-positive cerebrospinal fluid (CSF), stool (stool suspensions [StSp]), rectal swabs (RS), nasopharyngeal swabs (NP), and lung and spleen tissues were tested to show assay efficacy with clinical specimens (Table 2). One hundred thirty-two CSF specimens were tested to demonstrate specificity, including 34 specimens that were enterovirus-positive by reverse transcription-PCR (RT-PCR). These included coxsackieviruses A9 (n = 2), B4 (n = 2), and B5 (n = 6); echoviruses 6 (n = 12), 7 (n = 2), 9 (n = 3), 11 (n = 1), 13 (n = 1), 15 (n = 1), 18 (n = 7), and 30 (n = 1); and enterovirus 71 (n = 1) and were identified by sequencing a portion of the VP1 gene (23).
HPeV and enterovirus clinical specimens
RNA extraction.Virus isolates and NP supernatants were extracted with the QIAamp viral RNA minikit (Qiagen, Inc., Valencia, CA), which was used according to the manufacturer's instructions. StSp and RS supernatants were extracted with an equal volume of Vertrel XF (DuPont Fluorochemicals, Wilmington, DE) prior to using the QIAamp kit. The RNA was eluted from the QIAamp column with 60 μl of sterile nuclease-free water. Tissues were homogenized in 500 μl Tri-Reagent (Sigma, St. Louis, MO) and then extracted with 200 μl chloroform. The emulsion was centrifuged and the aqueous phase transferred to a fresh tube. The RNA was precipitated with 3 volumes of isopropanol and 2 μl Pellet Paint coprecipitant (Novagen, Madison, WI) at −20°C overnight, pelleted by centrifugation at 20,000 × g at 4°C for 15 min, and washed with 70% ethanol. Pelleted RNA was air dried for 15 min in a biosafety cabinet and resuspended in 35 μl nuclease-free water.
Primer and probe design.The available parechovirus 5′NTR sequences were aligned by using PileUp (Wisconsin sequence analysis package, version 11.1; Accelrys, San Diego, CA). The average similarity across the entire alignment was determined, using PlotSimilarity (Wisconsin package) with an analysis window of 5 nucleotides. A portion of this similarity plot and primer and probe design are summarized in Fig. 1; primers and probes are listed in Table 3. Primer and probe sequences were screened by NCBI nucleotide BLAST to detect any possible cross-reaction with nonparechovirus targets.
(A) Schematic representation of the parechovirus genome, showing the location of the 5′NTR. (B) 5′NTR nucleotide sequences of HPeV1 to -6 and four LVs were aligned with Pileup (Wisconsin sequence analysis package, version 11.1; Accelrys, San Diego, CA). The average similarity across the entire alignment was plotted using PlotSimilarity (Wisconsin Package) with a window of 5 nucleotides. A part of this similarity plot is shown. Similarity scores of 100% indicate nucleotide sequence identity among all 14 aligned parechovirus sequences. (C) The minimal sequence heterogeneity among the aligned HPeVs and LVs is shown along with the locations of the primer and probe sites. The numbering of the primer/probe site nucleotide positions is relative to the published sequence of HPeV1-Harris and is for orientation only. (D) The degenerate primer and probe sequences are shown with primer orientations and probe labeling. Ambiguity codes: R, A or G; Y, C or T; W, A or T; and S, C or G. GenBank sequences aligned for primer and probe design included the following: HPeV1, S45208 and EF051629; HPeV2, AJ005695; HPeV3, AB084913 and AJ889918; HPeV4, AM235750 and DQ315670; HPeV5, AM235749 and AF055846; HPeV6, AB252582; and LV, AF327920, AF327921, AF327922, and AF538689.
Primers
TaqMan real-time PCR.Real-time PCR experiments were run on a Stratagene Mx4000 multiplex quantitative PCR system with version 4.20 analysis software. The Superscript III Platinum one-step quantitative RT-PCR system (Invitrogen, Carlsbad, CA) was used according to the manufacturer's instructions with 0.4 μM primers AN345 and AN344, 0.2 μM AN257 TaqMan probe, 5 mM MgSO4, 50 nM ROX (a reference dye included in each reaction mixture to normalize the fluorescent reporter signal), and 5 μl specimen RNA in a total reaction volume of 50 μl. The thermocycling profile for cDNA synthesis and real-time PCR was 50°C for 30 min and 95°C for 10 min, followed by 50 cycles of 95°C for 15 s, 58°C for 30 s, and 72°C for 10 s, with probe detection during the 58°C annealing step. All threshold cycle (CT) values were determined by using the Mx4000 analysis software with the moving average and amplification-based threshold algorithm enhancements turned off. Standards were run in triplicate and averaged before standard curve plotting, using the least-mean-squares curve-fitting algorithm.
In vitro-transcribed sRNAs.Synthetic RNA standards (sRNAs) were made for HPeV1-Harris and LV-SWE87-012 to test the absolute sensitivity of the assay and to construct standard curves for evaluation of the real-time assay. Strain-specific primers designed for HPeV1 (AN287 and AN288) and LV-SWE87-012 (AN289 and 290) are shown in Table 3. The forward primers for each virus include an upstream phage T3 RNA polymerase promoter. Synthesis of cDNA was carried out in a 20-μl reaction mixture containing 4 μl RNA, 0.5 mM each deoxynucleoside triphosphate, 4 μl 5× reaction buffer (Invitrogen), 10 mM dithiothreitol, 0.25 μM of reverse primer, 40 U of RNasin (Promega), and 200 U of Superscript II reverse transcriptase (Invitrogen), with incubation at 42°C for 10 min, 50°C for 30 min, and 95°C for 5 min. The PCR used 4 μl cDNA in a 100-μl reaction mixture containing 10 μl 10× FastStart Taq buffer (Roche), 200 μM each deoxynucleoside triphosphate, 0.4 μM of forward and reverse primers, and 5 U of FastStart Taq polymerase (Roche). Thermocycler conditions consisted of a hold for 10 min at 95°C, followed by 40 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for 60 s. The PCR products were resolved on 1.5% agarose gels, the bands of the correct size were cut, and the DNA was purified using the QIAquick gel extraction kit (Qiagen) according to the manufacturer's instructions. The purified DNA was quantitated by measuring absorbance at 260 nm, dried, and resuspended in sterile nuclease-free water to contain 0.1 μg DNA per μl. RNA was transcribed in vitro from 0.2 μg template DNA in a total volume of 20 μl, using the MEGAscript high-yield transcription kit (Ambion, Austin, TX) according to the manufacturer's protocol. After incubation for 5 h at 37°C, 2 U of DNase I was added to the reaction mixture and incubation at 37°C was continued for 30 min. RNA was purified from the transcription reaction mixture with the MEGAclear kit (Ambion), using the manufacturer's microcentrifuge protocol. The resulting single-stranded, positive-sense standard RNA products (HPeV1-Harris, 860 bases; LV-SWE87-012, 890 bases) were quantitated as described above, and the RNA concentration was converted to units of RNA molecules per microliter. A stock solution of each in vitro-transcribed RNA was made to contain 1 × 1012 RNA copies per 5 μl in sterile, nuclease-free water and stored in siliconized tubes at −70°C. Tenfold serial dilutions (106 to 10−1 RNA copies per 5 μl) were made from these stocks for sensitivity measurements. The assay's limit of detection was more stringently defined by determining the sRNA copy number that gave 100% positivity in each of 30 replicates, using half-log dilutions of HPeV1 and LV sRNAs, containing 100, 30, 10, 3, or 1 copies per 5 μl.
RESULTS
The similarity plot for the alignment of HPeV and LV 5′NTR sequences revealed three areas of high sequence conservation (Fig. 1). Degenerate PCR primers (AN345 and AN344) were designed from the flanking conserved regions, and a probe (AN257) was designed from the conserved internal region. To account for variation among the available parechovirus sequences, degeneracy was introduced at four positions in AN345, at three positions in AN344, and at four positions in AN257. The flanking primers AN345 and AN344 produce a PCR product of 203 to 204 bp for LV and 194 to 195 bp for HPeVs. The antisense probe AN257 binds 11 to 13 bp upstream from the antisense primer AN344. NCBI nucleotide BLAST analyses of the parechovirus primers and probe showed no assay cross-reactivity with human cellular sequences, other picornavirus genera, or any other virus family.
The specificity of the primers and probe for HPeV 1 to -6 and LV was determined by amplifying 62 reference strains and clinical isolates (Table 1). The real-time PCR successfully amplified all of the HPeV and LV isolates and produced no detectable amplification product from nonparechovirus RNA templates (i.e., those from picornaviruses in the genera Enterovirus and Cardiovirus) (Table 1). The CT values for HPeV templates ranged from 16 to 28, while those for LV templates were 18 to 38. When needed, parechoviruses can be differentiated as HPeV or LV by sequencing the parechovirus 5′ NTR real-time RT-PCR amplicon, since HPeV shares only 53.9 to 61.7% nucleotide identity to LV within the 5′NTR amplicon region.
Analytical sensitivity was determined using two methods. Sensitivity relative to cell culture infectivity was measured using diluted RNA extracted from stocks of prototype strains HPeV1 Harris and LV SWE87-012 whose titers had been determined. PCRs were run in triplicate. The real-time PCR method was linear over the entire detectable range, from 103 to 10−3 CCID50 (approximately 1.03 × 108 to 103 genome copies) per 5 μl for HPeV1 and from 102 to 10−2 CCID50 (approximately 2.30 × 106 to 230 genome copies) per 5 μl for LV-SWE87-012 (Fig. 2), suggesting that the PCR assay is approximately 100- to 1,000-fold more sensitive than cell culture for detecting both HPeV and LV. Sensitivity was also measured by using in vitro-transcribed synthetic single-stranded positive sense sRNAs derived from prototype strains HPeV1-Harris and LV-SWE87-012. The detection limit for the parechovirus 5′NTR real-time assay was approximately one copy for both HPeV1 Harris and LV-SWE87-012 sRNAs (Fig. 2). The CT values were 44 for one copy of the HPeV1-Harris sRNA (Fig. 2) and 42 for one copy of the LV-SWE87-012 sRNA.
Amplification of dilution series for HPeV1 and LV viral RNAs and synthetic RNAs, run in triplicate. The insets show the standard curves for each of the standard dilution series. Positive reaction CT values were averaged before standard curve plotting, using the least-mean-squares curve-fitting algorithm of the Stratagene Mx4000 software (version 4.2). (A) Real-time PCR amplification of 10-fold serial dilutions of HPeV1 viral RNA, containing the equivalent of 103 to 10−3 CCID50 per reaction mixture. (B) Real-time PCR amplification of 10-fold serial dilutions of LV viral RNA, containing the equivalent of 102 to 10−2 CCID50 per reaction mixture. (C) Real-time PCR amplification of 10-fold serial dilutions of HPeV1 sRNA, containing the equivalent of 106 to 100 RNA copies per reaction mixture. (D) Real-time PCR amplification of 10-fold serial dilutions of LV sRNA, containing the equivalent of 106 to 100 RNA copies per reaction mixture.
The limit of detection and CT cutoff value for sRNA were determined by using low-copy-number, half-log-diluted standards with high replicate numbers. For the HPeV1 sRNA, 100% positivity for 30 replicates was obtained at 30 copies (mean CT value = 37.03; range = 35.95 to 38.80; percent coefficient of variation [CV]) = 1.99). At 10 copies of HPeV1 sRNA, 24/30 (80%) were positive (mean CT value = 37.69; range = 36.56 to 39.52; CV = 2.22), and at 3 copies 11/30 (36.7%) were positive (mean CT value = 38.97; range = 37.55 to 39.66; CV = 1.39). For LV sRNA, 100% positivity for 30 replicates was obtained at a minimum of 100 copies (mean CT value = 38.04; range = 36.52 to 39.58; CV = 2.22). At 30 copies of LV-SWE87-012 sRNA, 17/30 (56.7%) were positive (mean CT value = 38.70; range = 37.80 to 39.68; CV = 1.45), and at 10 copies 4/30 (13.3%) were positive (mean CT value = 40.55; range = 39.15 to 41.45; CV = 2.43).
Efficacy of detection and the range of CT values from clinical specimens were demonstrated by amplification from CSF, StSp, NP, RS, and lung and spleen tissues (Table 2). The CT values for the CSF specimens ranged from 27 to 40 (mean, 32.5; n = 11), those for the StSp ranged from 20 to 40 (mean, 32.8; n = 16), those for the NP and RS ranged from 31 to 40 (mean, 37.3; n = 4), and those for the lung and spleen tissues ranged from 35 to 41 (mean, 38.8; n = 4). Among the 132 CSF specimens tested, 11 were parechovirus positive, while 34 were enterovirus positive. The parechovirus-positive specimens were confirmed and the virus identified as HPeV3 by partial sequencing of the VP1 region. All of the enterovirus-positive specimens were negative in the parechovirus real-time assay.
A reasonable cutoff value for the real-time assay, based on the 30 replicate experiments with 100% positivity (HPeV1, 30 copies; LV87-012, 100 copies) for both pure sRNAs, would be 40 cycles. However, clinical specimens, such as the bloody lung and spleen tissues assayed, sometimes contain nonspecific inhibitors of PCR that delay amplification. The lung and spleen tissues described above, when run in replicate from different RNA extracts, gave CT values as high as 44. At the limit of detection of one sRNA copy, the HPeV1 sRNA gave a CT value of 44 and the LV87-012 sRNA gave a CT value of 42. As a result of the replicate experiments with pure sRNA and the efficacy experiments, using common clinical specimens and difficult-to-amplify bloody tissues, 45 cycles of PCR was chosen as the cutoff for a positive result.
DISCUSSION
Several investigators have published parechovirus-specific PCR methods that permit detection of HPeV1 to -3 (5, 7, 14, 25), or even HPeV1 to -6 (2), but not LV. Other published methods are even more limited, detecting only HPeV1 but not HPeV2, HPeV3, or LV (8, 15, 16). Using any of these methods, multiple primer sets would be required to detect all known parechoviruses. The method described here permits detection of all parechoviruses, including the HPeVs and LVs, using one pair of primers and one probe. Differentiation of HPeV and LV can be done by sequencing the 5′ NTR real-time RT-PCR amplicon. Several of the published methods have been used to detect HPeV in a very limited number of clinical specimens (<3), but their sensitivity has not been established. LVs have not yet been definitively associated with human infection or disease; however, since the full range of HPeV diversity is unknown, it is important that any assay amplify the broadest possible range of viruses within the genus. Finding common primer and probe sites between the HPeVs and LVs increases the probability that the test will amplify yet-to-be-described members of the genus.
The real-time PCR assay was 100% specific, as it amplified all available strains of the genus Parechovirus but none of the nonparechovirus strains tested. Like the assay of Baumgarte et al. (2), our assay is approximately 100- to 1,000-fold more sensitive than virus isolation in cell culture (Fig. 2). One reason for this difference is that the ratio of total virus particles to infectious particles is generally at least 100 for enteroviruses, such that there are many more genome copies than infectious virions; in addition, PCR assays can also detect free RNA genomes. We have also tested a seminested version of the assay, using a nested primer similar in sequence to the AN257 probe (data not shown). Absolute sensitivity is approximately equal for the two versions of the test (both can detect as little as one copy of a parechovirus sRNA) but the real-time assay format is simpler and faster (2.5 h, versus 8 h for the seminested PCR assay) and is less prone to cross-contamination. The real-time assay was linear over the tested range of template concentrations, using either viral RNAs or sRNAs (Fig. 2). The efficacy of the real-time method was demonstrated over a broad range of clinical specimens, including difficult human tissue specimens. Although the parechovirus real-time assay is presented here as a qualitative diagnostic assay, the test could be easily adapted to quantitation of virus loads in clinical specimens.
We chose to use sRNAs for our real-time standard curves rather than purified virus RNA. In vitro-synthesized RNA can be made in large lots, is easily quantitated, and is noninfectious. Additionally, there is no absolute correlation between the amount of viral RNA present (detectable by RT-PCR) and the number of viable virions (detectable by cell culture), since RT-PCR can detect free, damaged, or nonviable viral RNA molecules, the proportions of which may vary depending on virus strain, growth conditions, and other factors.
Parechoviruses have been associated with a wide range of diseases affecting the gastrointestinal tract, the respiratory tract, and the central nervous system. It has also been proposed that maternal infection by LV may be associated with intrauterine fetal death in humans. A number of studies have identified and characterized four new HPeV types and the LV species (1, 2, 4, 11-13, 21, 31). This situation is reminiscent of the rapid discovery of dozens of new enteroviruses in the 1950s, following the introduction of cell culture as the method of choice to isolate viruses (6, 9). This comparison suggests that many more new parechoviruses may be awaiting discovery. Cell culture is relatively nonspecific; that is, a wide range of viruses can generally be isolated in a given cell line. Since PCR is highly specific, it is important for virus discovery efforts that the PCR method used is capable of amplifying the broadest possible range of viruses (all members of a genus or family, for example) while remaining specific enough to avoid false-positive results. Since our parechovirus PCR method amplifies all known members of the genus, including highly divergent viruses infecting divergent hosts (humans and rodents), we would expect it to also amplify the widest possible range of currently unknown parechoviruses.
ACKNOWLEDGMENTS
We thank David Schnurr, California Department of Health Services, for providing HPeV4 and HPeV5 reference strains; Gerald Sedmak, Milwaukee Public Health Laboratory, for providing HPeV3 and HPeV6 strains; and Elisabeth Winterburg for technical assistance with some of the initial experiments.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention.
FOOTNOTES
- Received 11 February 2008.
- Returned for modification 28 March 2008.
- Accepted 22 May 2008.
- Copyright © 2008 American Society for Microbiology
