key: cord-0951659-f52tfsmh
authors: Parker, J.K.; Chang, T.‐Y.; Meschke, J.S.
title: Amplification of viral RNA from drinking water using TransPlex™ whole‐transcriptome amplification
date: 2011-05-05
journal: J Appl Microbiol
DOI: 10.1111/j.1365-2672.2011.05029.x
sha: 760e8b61a8b0e40bae371055d836210b1f911f86
doc_id: 951659
cord_uid: f52tfsmh

Aims: Viral pathogens in environmental media are generally highly diffuse, yet small quantities of pathogens may pose a health risk. This study evaluates the ability of TransPlex™ whole transcriptome amplification (WTA) to amplify small quantities of RNA viruses from complex environmental matrices containing background nucleic acids. Methods and Results: DNA extracts from mock drinking water samples containing mixed microbial populations were spiked with small quantities of echovirus type 13 (EV) RNA. Samples were amplified using a Transplex™ WTA kit, and EV‐specific quantitative reverse transcription polymerase chain reaction (qRT‐PCR) was used to quantify target pathogens before and after application of WTA. Samples amplified by WTA demonstrated a decreased limit of detection. The log‐linear relationship between serial dilutions was maintained following amplification by WTA. Conclusions: WTA is able to increase the quantity of target organism RNA in mixed populations, while maintaining log linearity of amplification across different target concentrations. Significance and Impact of the Study: WTA may serve as an effective preamplification step to increase the levels of RNA prior to detection by other molecular methods such as PCR, microarrays and sequencing.

To date, application of WTA to detection or characterization of microbial RNA has been limited because of the relative novelty of WTA technologies and difficulty of adapting existing methods to microbial RNA. IVT was successfully adapted to artificial and environmental mixed bacterial community samples by attaching random hexamers to the T7 RNA promoter sequence in lieu of oligo(dT) (Gao et al. 2007 ). However, microarray analysis of WTA product from these mixed bacterial communities showed representative detection was dependent upon the amount of starting RNA. In another study, Rift Valley fever virus from brain biopsies was successfully amplified by WTA in the presence of Staphylococcus aureus DNA (Berthet et al. 2008) , but starting amounts of RNA were high, ranging from 10 4 to 10 6 copies. More recently, TransPlex TM WTA has been used on cultures of coronavirus to generate sufficient nucleic acid for sequencing new human viruses when propagation in cell culture is inefficient (Banach et al. 2009) .

A potential new application of WTA is to improve detection of RNA viruses in environmental media where they may pose a risk to human health. RNA viruses may occur in very small concentrations in drinking water and other environmental media, making them difficult to detect with molecular methods. Additionally, there are numerous types of RNA viruses (e.g. enteroviruses, noroviruses, astroviruses) that may be present in environmental media, further complicating detection because they do not share a common gene, analogous to the conserved 16S gene in bacteria, which may be targeted for detection by a singleplex or easily multiplexed PCR. To determine whether WTA can improve molecular detection by amplifying viral RNA from environmental media, this study assessed the ability of WTA to maintain log-linear amplification over a range of initial RNA concentrations, to efficiently amplify RNA from environmental matrices and to overcome inhibition by substances in these matrices.

Echovirus 13 was chosen as a representative RNA virus for this study, based on the listing of echoviruses on the EPA's Drinking Water Candidate Contaminant List 2 (CCL2) (USEPA 2005) . Although echovirus is not often associated with known outbreaks, it is a putative agent of waterborne disease (Leclerc et al. 2002) . One concern with the application of WTA to low numbers of pathogens is the requirement for a specific amount of starting RNA, as quantities suggested for WTA are higher than the amount of target RNA which would be anticipated in a typical environmental sample. This study is one of the first to assess the ability of WTA to improve the detection of environmentally relevant quantities of human pathogens in environmental media and to test TransPlex TM WTA on environmental pathogens.

Echovirus 13 strain Del Carmen (ATCC # VR-43) (EV) was produced in buffalo green monkey kidney (BGMK) cells grown in minimum essential medium (MEM) with 10% foetal bovine serum. When CPE was complete (c. 5 days), infected cells were freeze-thawed once at )80°C, virus was purified with a chloroform extraction, and the EV viral stock was stored at )80°C. Titre of viral lysate was assayed in 24-well plates of confluent cells. Wells were scored as positive or negative for the evidence of CPE, and titre was calculated (Reed and Muench 1938) in units of tissue culture infectious dose (TCID 50 ). RNA extractions for EV were performed on sample volumes of 140 ll using the QIAamp Viral RNA Mini kit (QIAGEN, Valencia, CA, USA) and eluted into final volumes of 60 ll.

WTA was conducted using the TransPlex TM Whole Transcriptome Amplification kit (Sigma-Aldrich), following the specified protocol. Sample volumes of 10 ll were incubated with reverse transcriptase and primers in a strand displacement reaction to create the TransPlex TM cDNA library of target fragments with a universal end sequence. The library was then amplified with universal primers for 25 cycles. Incubations and amplifications for WTA were performed on a DNA Engine Dyad Thermal Cycler (Bio-Rad Laboratories, Hercules, CA, USA).

To determine the optimal amplification cycle number for WTA, initial runs were monitored with SYBR Green qRT-PCR on a Chromo4 Real-Time PCR Detection System (Bio-Rad Laboratories) to visualize the amplification plateau. An average optimal cycle number of 25 cycles was selected.

Previously described enterovirus primers and probe (Schwab et al. 1995) were used to target EV: Pan-E5¢ (CCT CCG GCC CCT GAA TG), Pan-E3¢ (ACC GGA TGG CCA ATC CAA) and fluorescently labelled TaqMan Pan-EP (6FAM -TAC TTT GGG TGT CCG TGT TTC -BHQ1). These primers target the conserved 5¢untranslated region of enteroviruses. Initially, during optimization of WTA reactions, qRT-PCRs for EV were performed using iScript One-Step RT-PCR kit with SYBR Green (Bio-Rad Laboratories). SYBR Green-based reagents were later replaced with the iScript One-Step RT-PCR kit for Probes (Bio-Rad Laboratories) for use with target-specific TaqMan probes, which were used for all final qRT-PCR results. All qRT-PCRs were conducted in 25 ll volumes containing 0AE2 lmol l )1 primers and 0AE15 lmol l )1 probe. Either 1 or 2 ll template RNA was added to each reaction. Cycling parameters were as follows: 50°C for 10 min, 95°C for 5 min, 40 cycles of 95°C for 10 s and 55 for 30 s. Each set of qRT-PCRs included an EV 4-point standard curve for quantification of sample concentrations and a no-template negative control (NC) reaction. Standards were quantified relative to EV tissue culture titres in units of TCID 50 .

The background (BG) population of micro-organisms was composed of equal amounts of DNA extracted from cultures of Escherichia coli (EC) and Bacillus subtilis (BS). The surface water (SW) population of micro-organisms was composed of DNA extracted from SW collected from the north side of Portage Bay, Seattle, Washington off of NE Boat Street. Both BG and SW populations also included small seeded quantities of the pathogens Adenovirus 41 (AdV), Aeromonas hydrophila (AH) and Mycobacterium avium (MAC) (3AE6 ng ml )1 total pathogen DNA in 'EV A' sample described below, data not shown).

DNA extractions were performed using the PowerSoil DNA Isolation kit (MO BIO Laboratories, Carlsbad, CA, USA). For the BG population, 1 ml aliquots of overnight EC and BS cultures was centrifuged for 10 min at 9000 g. The pellet was resuspended in 100 ll PBS and extracted. For the SW population, 2 l of water was filtered through replicate 47-mm-glass fibre prefilters, Type APFB, 1 lm pore size (Millipore, Billerica, MA, USA), and frozen at )80°C in sterile microtubes. Frozen SW filters were torn into pieces using sterile forceps and divided to yield two extractions per filter.

The Quant-iT PicoGreen dsDNA kit (Invitrogen, Carlsbad, CA, USA) was used to quantify dsDNA in the extracted samples using the NanoDrop ND-3300 Fluorospectrometer (NanoDrop Technologies, Wilmington, DE, USA). Total EV RNA concentrations were too small to be quantified by Nanodrop fluorometric measurements but are estimated to be orders of magnitude <5 ng, based on the mass of the EV genome and an assumed particle to infectivity ratio of 1 : 1 (Table 1) .

RNA concentrations ranging from c. 0.01 · 10 )6 to 3 · 10 )6 lg were spiked into DNA from different backgrounds of mixed populations of micro-organisms (laboratory-grown and environmental micro-organisms) and WTA was applied (Fig. 1 ). After WTA, side-by-side qRT-PCR detection of pre-WTA and post-WTA samples was performed to quantify the impact of WTA on target nucleic acid concentration and WTA:pre-WTA ratios were determined. Triplicate experiments were conducted with low concentrations of EV spiked into the BG population and the SW population. Pre-WTA samples created for each experiment included four EV samples (EV), two BG or SW samples and eight samples containing EV spiked into BG or SW (Mix). Dilutions for the four EV samples were chosen so that two (EV A and EV B) were above and 2 (EV C and EV D) were below the estimated EV qRT-PCR limit of detection (LOD) ( Table 1) . Final concentrations of BG ⁄ SW in Mix samples were BG 10 3 · ⁄ 10 4 · or SW 10 4 · ⁄ 10 5 · (values relate to EV A concentration, e.g. DNA in 10 3 · is c. 1000 times the concentration of target in EV A). Exact BG ⁄ SW DNA concentrations were 10, 100 and 1000 ng ml )1 for 10 3 ·, 10 4 · and 10 5 ·, respectively. WTA samples were produced from the 14 pre-WTA samples. EV and Mix samples were amplified with WTA in triplicate and BG ⁄ SW samples (WTA negative controls) in singlicate. All samples (pre-WTA and WTA samples) were analysed with qRT-PCR in triplicate to quantify target pathogens before and after application of WTA. WTA products were diluted 1 ⁄ 100 prior to qRT-PCR, which was factored into final calculations of WTA sample concentration.

Linear regressions of pre-WTA and WTA cycle threshold (C t ) values were performed for each set of experiments (BG and SW) using EV A and EV B samples (samples with reliably detectable pre-WTA EV concentrations). Linear regression equations were generated, with slope and R 2 values reported as measures of efficiency and goodness of fit ⁄ variance, respectively. For EV C and EV D samples (samples near ⁄ below the estimated qRT-PCR LOD prior to WTA), Fisher's exact test was used to test for a significant difference between numbers of positive pre-WTA samples vs positive WTA samples.

Equivalent TCID 50 values were estimated using the EV standard curves for each qRT-PCR run. Data were log transformed, and average and median TCID 50 values were calculated for data points. For reactions resulting in nondetects, values equal to the approximate LOD were imputed for statistical analyses. Descriptive statistics and linear regressions for all data were conducted using Microsoft Ò Office Excel 2003.

According to the Transplex TM WTA protocol, optimal amplification, estimated to be 17 cycles, is reached 2-3 cycles into the amplification plateau, at which point the reaction can be stopped. Optimization of the WTA library amplification step by real-time visualization with SYBR Green showed that, for samples containing RNA at low concentrations of interest, most samples had reached the amplification plateau by 25 amplification cycles. Therefore, 25 cycles were selected as an appropriate number of amplification cycles for all future WTA reactions.

Melting curve analysis of SYBR Green-based qRT-PCR product revealed multiple nonspecific product peaks for WTA samples, indicating detection of products other than the target EV amplicon (data not shown). TaqMan qRT-PCR of the same samples eliminated detection of false positives and was selected for final qRT-PCR analyses. However, TaqMan qRT-PCR analysis of WTA product initially resulted in fewer positive results than anticipated based on pre-WTA sample concentrations, and positive WTA samples had low fluorescence (£0AE2) and a large amount of probe signal drift, where fluorescence gradually increases over time but is not exponential. It was hypothesized that reagents were being utilized prior to successful amplification because of the presence of excess WTA product. Therefore, 1 ⁄ 100 dilution of WTA product was required prior to analysis to eliminate false negatives and maintain efficient exponential amplification.

Linear regressions of pre-WTA and WTA C t values demonstrate the effective amplification of WTA over several orders of magnitude at low initial copy number concentrations (Figs 2 and 3 ). Regressions are highly linear, with R 2 values ranging from 0AE89 to 0AE99 despite the variability in the highly sensitive qRT-PCR assays and potential RNA degradation ⁄ variation between experiments. As C t values correspond to the logarithm of the initial copy number, results show that log linearity of amplification is preserved over a range of environmentally significant concentrations. This suggests the ability to accurately quantify initial copy number, given a sample amplified by WTA.

Slope is indicative of the average ratio of WTA:pre-WTA target over all of the samples (Figs 2 and 3 ). If slope = 1, amplification efficiency is the same across concentrations. Variation in the slope (variation in amplification efficiency) is seen depending on the amount and type of mixed population addition. Slopes are close to 1, but decrease somewhat with increasing addition of BG ⁄ SW, showing that amplification is less efficient for the smallest concentrations when inhibiting substances are present. Position of samples along the y-axis is also indicative of amplification efficiency. Compared with the 1 : 1 reference regression line, all samples are positioned lower on the y-axis, so concentrations are improved by application of WTA. Samples with the highest concentrations of BG ⁄ SW are positioned higher on the y-axis, indicating decreased amplification because of the mixed DNA addition. On average, EV was amplified c. 2-4 logs by WTA.

Ratios of WTA:pre-WTA concentrations decrease with decreasing initial target concentration, based on both medians and averages (Table 2 ). For nine replicates of low target concentration, there is a high probability that ‡½ of WTA samples will be ND, resulting in medians that are also ND, while averages may be detectable. WTA:pre-WTA ratios also show that increasing addition of BG or SW decreases efficiency of amplification.

The number of EV C and EV D samples detected as positive before and after application of WTA was tabulated (Table 3) . Results show that there was an increase in the Figure 3 Linear regressions of average echovirus pre-WTA and post-WTA C t values for surface water experiments. A 1 : 1 C t ratio line is shown for reference. 4 SW 10 5 ·; r SW 10 4 ·; h Control (no competing nucleic acid). EV D SW 10 5 · ND >1AE1 BG 10 3 · ND ND SW 10 4 · ND ND BG 10 4 · ND ND SW 10 5 · ND ND LOD, limit of detection. *Ratios calculated using estimated qRT-PCR LOD of 1AE1 TCID 50 (as imputed from standard curve) in place of nondetects (ND).

number of EV C and EV D samples positive for EV after WTA. Results from Fisher's exact test showed that the difference between the number of samples positive before vs after WTA was significant (two-tailed, P = 0AE0280).

The ultimate goal of this study was to evaluate WTA for successful amplification of environmentally relevant amounts of RNA, which are less than the amounts of RNA previously suggested for successful WTA. Trans-Plex TM WTA, which had not been applied to microbial RNA in the presence of background nucleic acids prior to this study, successfully amplified viral RNA from lower levels of starting material than suggested for the kit.

In previous studies, other WTA methods have not been able to efficiently amplify low copy number samples. For IVT of bacterial mRNA, 50-100 ng RNA is needed for representative detection (Gao et al. 2007) , and amplification fidelity decreases when <100 ng RNA is used, particularly if the target transcripts are underrepresented in the sample (Wang 2005) . This suggests that the sensitivity of IVT would be inadequate for detection of viruses in environmental samples. TransPlex TM WTA recommends input of 50 ng RNA, but notes that <5 ng is usable. Total EV RNA concentrations in this study are estimated to be orders of magnitude <5 ng ( Table 1) .

Modification of the WTA protocol was required for successful WTA under the conditions of this study. TransPlex TM WTA was designed to amplify human RNA and had not previously been optimized for the amplification of microbial nucleic acids or low copy number samples. Increasing the number of amplification cycles greatly improved amplification and was necessary to maximize product formation.

Transplex TM WTA combines MDA and universally primed PCR. MDA typically has higher product yield than previously described universally primed methods (Bergen et al. 2005; Sorensen et al. 2007; Uda et al. 2007 ), but hyperbranching products may obscure the target site (Vora et al. 2004) . In TransPlex TM WTA, MDA creates random, overlapping strands of variable size that are flanked by universal binding regions. These strands are then amplified by universally primed PCR. This takes advantage of the benefits of each technology, yielding superior results. Independent researchers have previously combined MDA and universally primed PCR for DNA amplification, yielding improvements in detection (Vora et al. 2004; Breitbart and Rohwer 2005; Panelli et al. 2005) .

Increases in sensitivity and total sample volume with WTA Using WTA as a preamplification step can successfully increase the sensitivity of downstream molecular analyses. In this study, increases in EV target copy number of c. 2-4 logs after application of WTA resulted in significant improvements in detection by qRT-PCR, as evidenced by the statistically significant increase in the total number of positively detected low copy number (at or near the LOD) samples. However, it is difficult to accurately quantify improvements in sensitivity from application of WTA (other than positive ⁄ negative detection), because consistently obtaining an aliquot positive for virus when the sample concentration is below a certain limit is improbable (Chen et al. 2007) . In this study, when target concentrations are at or below the estimated LOD, the probability of successful WTA is decreased because of decreased probability of target being added to the WTA reaction. Additionally, as formation of WTA template is a random process, metatranscriptomes produced from replicate amplifications can differ, particularly for low copy number samples. Although WTA of low copy number samples may result in infrequent amplification because of these factors, when a target is successfully amplified, improvements in concentration are significant, as demonstrated by pre-WTA:WTA ratios that are higher for averages than medians (Table 1) .

Another benefit of WTA is to increase the total volume of sample available for analysis. Improvements in detection from WTA have been described on a concentration basis, but it is also useful to describe gains in terms of volume. TransPlex TM WTA increases an initial sample of 10-375 ll of WTA product. Access to a larger amount of product makes it easier to run multiple reactions for multiple pathogens or conduct assays where a larger volume of high-quality sample is needed such as microarray analysis. There is more room for assay error or to run the assay multiple times if the quantity after amplification is near the LOD. Volume may no longer be a limiting factor for some types of molecular assays, which is beneficial even if no significant gains are seen in terms of target concentration. The ability to quantify the original target number of samples amplified by WTA is another important feature of this study. Because the dilution series stayed consistent post-WTA application, resulting in log-linear amplification, environmental samples amplified by WTA may not be limited to a presence ⁄ absence interpretation of data. Samples can instead be quantified based on known standard curves. This demonstrates that increases in sensitivity and total sample volume gained using WTA are not at the expense of sample quantitation.

The second major goal of this research was to determine whether WTA can be successfully applied to environmental samples containing mixed populations of microorganisms. WTA was able to amplify target RNA despite competition from background nucleic acids. Moderate concentrations of nucleic acids had a minor effect on amplification efficiency, while higher concentrations had a greater impact. As background nucleic acids increase, there is likely a decreased chance of amplification of the target as nontarget nucleic acids out-compete low concentration targets during the random amplification. Although EV WTA:pre-WTA ratios were lower for WTA of EV added to mixed nucleic acids (c. 1-3 log increase in target) compared with WTA of EV only (c. 2-4 log increase in target), the ratios still show significant amplification of EV targets were gained despite the presence of competing nucleic acids.

Experiments with mixed microbial communities were also used to determine the ability of WTA to overcome inhibition of qRT-PCR from background nucleic acids and other contaminating substances. There is evidence that whole-genome amplification (WGA) can improve detection in the presence of inhibitors by amplifying nucleic acids to levels which dilute the inhibiting substances (Gonzalez et al. 2005) . This effect could also overcome decreased sensitivity caused by the background metagenome. Although inhibition was difficult to quantify, WTA successfully amplified targets of low concentration despite the presence of large of amounts of competing nucleic acids from a natural water sample that probably contained substances that typically interfere with PCR. This suggests that WTA would be a good replacement for the initial round of nested PCR, which has been used to improve detection of low copy number samples, particularly those with inhibiting substances.

SYBR green for qRT-PCR of WTA product SYBR Green may not be able to effectively serve as a fluorescent label for qRT-PCR analysis of TransPlex TM WTA product. Although SYBR Green qRT-PCR detection of pre-WTA samples with the Schwab et al. (1995) EV primer set was effective, it was not effective for analysis of WTA product. Detection of multiple product peaks in melting curve analyses of SYBR Green-based qRT-PCR indicates that multiple products were formed during WTA, a by-product of template shearing during the nonspecific, nonprimer-directed amplification. The multiple nontarget products ultimately interfered with detection of the target EV amplicon. Also, TransPlex TM WTA creates both singlestranded and double-stranded cDNA product. Because SYBR Green is an intercalating dye, which indiscriminately binds between all dsDNA molecules, the high background fluorescence can probably be attributed to the large quantities of nontarget dsDNA produced during WTA. TaqMan probes increase the specificity of qRT-PCR by binding only with the target amplicon. However, product information from Sigma-Aldrich now states that cDNA from WTA should be purified prior to use because some components may interfere with downstream applications. This information was not provided in the pre-release version of the kit used in these experiments, but could perhaps improve the use of SYBR Green in this situation.

Application of WTA to environmental pathogens WTA has not previously been assessed as a tool for improving molecular detection of low levels of pathogens in environmental samples. Results from these experiments demonstrate that WTA improves molecular detection of small quantities of RNA from small quantities of viral pathogens. WTA can be effectively applied to mixed microbial communities, such as those that might be found in drinking water. WTA also increases the quantity of nucleic acids so that volume is not the limiting factor to multiple detection assays for multiple pathogens. The potential applications for WTA as a preamplification step are diverse and include any molecular analyses which could benefit from an increase in target copy number (e.g. qRT-PCR), increase in sample volume (e.g. microarray analysis), dilution of inhibiting substances, a replacement for more inferior preamplification methods (e.g. nested PCR), immediate preservation of unstable samples (i.e. conversion to cDNA for future use) and ⁄ or a generally higher-quality, higher-concentration nucleic acid sample (e.g. for sequencing, microsatellite analysis). For some downstream applications, WTA product may need additional processing steps prior to analysis (e.g. whole-genome sequencing might require an additional concentration step, such as magnetic bead-based separation, prior to sequencing).

Analyses of samples before and after WTA indicate that linear regressions of these concentrations remain relatively stable despite the addition of different types and concentrations of microbial community DNA. This suggests that standard curves may be used to estimate starting copy numbers of target RNA and quantify unknown amounts of pathogens. Based on the results of this study, it is recommended that, for samples where low quantities of pathogens are anticipated, replicate reactions containing larger amounts of template should be performed to increase the chances of detecting low copy number pathogens. The implications for improving detection of pathogens in drinking water are promising. It is essential to test a variety of other human pathogens present in actual environmental samples such as this to further validate WTA.

Human airway epithelial cell culture to identify new respiratory viruses: coronavirus NL63 as a model

Comparison of yield and genotyping performance of multiple displacement amplification and OmniPlex whole genome amplified DNA generated from multiple DNA sources

Phi29 polymerase based random amplification of viral RNA as an alternative to random RT-PCR

Method for discovering novel DNA viruses in blood using viral particle selection and shotgun sequencing

Effect of sample aliquot size on the limit of detection and reproducibility of clinical assays

Microarray-based analysis of microbial community RNAs by whole-community RNA amplification

Multiple displacement amplification as a pre-polymerase chain reaction (pre-PCR) to process difficult to amplify samples and low copy number sequences from natural environments

Comparison of RNA amplification techniques meeting the demands for the expression profiling of clinical cancer samples

Microbial agents associated with waterborne diseases

Ligation overcomes terminal underrepresentation in multiple displacement amplification of linear DNA

A simple method of estimating fifty per cent endpoints

Concentration and purification of beef extract mock eluates from water samples for the detection of enteroviruses, hepatitis A virus

Whole genome amplification on DNA from filter paper blood spot samples: an evaluation of selected systems

Whole transcriptome amplification for gene expression profiling and development of molecular archives

Comparison of whole genome amplification methods for detecting pathogenic bacterial genomic DNA using microarray

Fact Sheet: The Drinking Water Candidate Contaminant List -The Source of Priority Contaminants for the Drinking Water Program

Amplified RNA synthesized from limited quantities of heterogeneous cDNA

Nucleic acid amplification strategies for DNA microarray-based pathogen detection

RNA amplification for successful gene profiling analysis

This publication was developed under STAR research assistance agreement R833011 awarded by the U.S. Environmental Protection Agency (EPA). This publication has not been formally reviewed by the EPA. The views expressed herein are solely ours, and the EPA does not endorse any products or commercial services mentioned in this publication. Thanks to Megan Parker for help with figures.