Data generated with the ICEPlex system using the ViraQuant assay was compared with that using the TaqMan method and the 7500 Real-Time PCR system (Applied Biosystems, Foster City, CA, USA), with primer/probe kits either from Applied Biosystems or from Argene (North Massapequa, NY, USA). Target genes for CMV and EBV were cloned into the pUC19 vector (New England Biolabs). When viral particles were assayed, they were spiked into negative human plasma from 1 × 102 to 5.5 × 105 particles/mL and extracted using the NucliSENS easyMAG system (bioMérieux, Durham, NC, USA).Tests for potential cross-contamination
The ViraQuant assay was used to assess the potential for cross-contamination between samples. A checkerboard arrangement was used with alternating negative and positive controls in 48 sample wells using a 48-capillary CE cartridge. Upon completion of the run, the ICEPlex cartridge was automatically decontaminated by immersing the cannulae into the onboard decontamination solution in the idle tray for 10 min with periodic swirling. This was followed by a run of 48 negative reactions to confirm the absence of potential well-to-well and run-to-run carryover contamination.Results and discussion Performance characteristics of the prototype ViraQuant assay on ICEPlex
STAR technology is based on monitoring PCR amplification by sampling the PCR at sequential cycles and detecting the fluorescent signals during real-time CE separation (8). Using the ICEPlex system, CE resolves and identifies PCR amplicons designed to be of different sizes, thus allowing high numbers of amplicons to be detected simultaneously, while at the same time increasing confidence in result interpretation considering expected versus detected amplicon sizes. Images are continuously and regularly collected at the detection window to monitor the varying peak signals generated from fluorescently labeled migrating amplicons. As the areas of these peaks increased over consecutive PCR cycles, they are used to construct amplification curves for each amplicon. Similar to real-time PCR amplification, a Ct value is determined from an amplification curve for each of the targets while in the exponential phase of PCR amplification. Multiple calibrator controls of predetermined copy numbers are included in each assay to quantify the initial abundance of all targets through a comparison of Ct values between targets and calibrators.
This process of multiplex amplification and quantification is illustrated in Figure 1 for the prototype multiplex assay, ViraQuant, implemented on the ICEPlex platform. It depicts the digital image electropherogram (Figure 1A) showing the increase in fluorescent signal for each of the five viral targets, along with sensitivity and calibrator controls, as a function of sequential sampling and CE separations over increasing cycle number. From these fluorescent intensity values, amplification curves were generated for each target (Figure 1B), from which Ct values were determined. The number of copies for each of the five viral targets initially present in the reaction was then quantified according to the formula presented in the Materials and methods section (Figure 1C).ICEPlex optical detection range
Figure 2 depicts the optical operating range of the ICEPlex system, which demonstrates its ability to detect both FAM and TYE signals inexcess of three orders of magnitude of relative fluorescence units (RFU). This broad range of log-linear sensitivity permits quantitative detection of a wide range of target concentrations over many PCR cycles. The log-log linear regression slope for the FAM signal was 0.80, and greater than 0.91 for TYE. The difference in slopes suggests minor discrepancies in signal acquisition or filtering that is specific for the two wavelengths in use.ICEPlex target concentration operating range
A 10-fold dilution series of the pBK plasmid clone was used to assess the target operating range (Figure 3). The assay exhibited linear performance across all concentrations, with an overall slope of 0.9865, and a DNA target detection range extending at least seven orders of magnitude (1 × 108 to 10 copies/reaction). When the polynomial fit method was applied to these log transformed data (9), the pBK targets were shown to be log-linear over five orders of magnitude, from 102 to 107 copies/reaction [slope (s.e.) = 0.9865 (0.0143), intercept (s.e.) =-0.0164 (0.0699), r2 = 0.9971, s.e. residuals = 0.1009].