Increased sensitivity of one-tube, quantitative RT-PCR.

With the introduction of real-time, fluorescence-based 5′ nuclease PCR (3,4) and instruments such as the ABI PRISM 7700 (TaqMan) sequence detector (Applied Biosystems, Foster City, CA, USA), quantitative RT-PCR is now a widely accepted method for measuring gene expression levels. Quantitative RTPCR is a sensitive technique and is particularly useful for the analysis of samples containing limited amounts of nucleic acids, such as in clinical tissues (2). When quantitating these small amounts of RNA and/or very low-abundance mRNA species, obtaining maximum sensitivity from a quantitative RTPCR is extremely important. While consecutive rounds of nested PCR are often used to obtain maximum sensitivity, this is difficult to achieve and still maintain accurate quantitation. Furthermore, multiple rounds of PCR increase the risk of contamination, a serious problem when working at the desired sensitivity levels. One-tube RT-PCR (RT and PCR in the same tube using the reverse PCR primer for the RT) reduces the risk of contamination when using the ABI PRISM 7700 because the reaction tubes are never opened. Theoretically, a one-tube procedure should have the same sensitivity as a two-step approach (separate RT followed by PCR), but in practice this is not the case (1). We have found that the sensitivity of one-tube RT-PCR is limited by the relative nonspecificity of the RT step. This nonspecificity arises from the fact that the RT is carried out at relatively low temperature and without a hot start, thus allowing nonspecific priming by both the desired RT “reverse” primer and also from the “forward” PCR primer. As the amount of target decreases in the input RNA sample, priming artifacts from the coldstart RT process can compete with, and reduce the efficiency of, PCR amplification of the desired target sequence. Thus, as RNA input decreases in a onetube procedure, nonspecific side reactions eventually out-compete the desired reaction, and no specific product is generated. In a two-step or nested RT-PCR procedure, specificity can be achieved with the use of hot-start PCR and a primer set 5′ upstream from the RT primer. However, this is not the case in a one-tube procedure unless one is willing to open the reaction tube to add new primers (thus making it a one-tube but two-step procedure). We hypothesized that by using an external RT primer and keeping the RT and PCR primers separated during the RT step, we should be able to maintain PCR specificity and therefore sensitivity in a one-tube RT-PCR. Here, we report a modified one-tube RT-PCR assay that greatly increases sensitivity and can be used for quantitative RT-PCR on the ABI PRISM7700. Standard one-tube reactions were set up for β-glucuronidase (β-gus) mRNA in 50-μL volumes with the following final concentrations: 20 nM β-gus RT primer (5′-TTTGGTTGTCTCTGCCGAGT-3′), 100 nM each β-gus PCR primer (GUS-F, 5′-CTCATTTGGAATTTTGCCGATT-3′; GUS-R, 5′-CCGAGTGAAGATCCCCTTTTTA-3′), 100 nM β-gus probe (5′-6-fam-TGAACAGTCACCGACGAGAGTGCTGG-tamra-3′), 1× TaqMan reaction buffer (Applied Biosystems), 5.5 mM MgCl2, 300 μM each dNTP, 20 U RNase inhibitor, 62.5 U SUPERSCRIPT II reverse transcriptase (Life Technologies, Rockville, MD, USA) and 1.25 U AmpliTaq Gold (Applied Biosystems). In the modified procedure, physical separation between the RT reaction mixture and the PCR primers was achieved by the use of AmpliWax PCR gem 50 (Applied Biosystems). First, the β-gus PCR primers were pipetted into the PCR plate in a final 5.0-μL volume. One PCR gem 50 was placed in each well, the wells were capped and the plate was centrifuged briefly to avoid the adherence of reagents to the tube wall above the wax barrier. The plate was then heated to 80oC for 2 min and cooled to 4oC to produce a wax barrier. A 45-μL upper layer was then pipetted into each well. This mixture contained the β-gus RT primer, the RNA, RNase inhibitor and SUPERSCRIPT II reverse transcriptase. Both layers were formulated to contain all of the remaining reaction components (buffer, nucleotides, MgCl2) at the concentrations described above. The presence of AmpliTaq Gold in the RT layer is inconsequential because this enzyme is inactive until heated to 95oC. All reactions were carried out on the ABI PRISM 7700 with the following thermocycler conditions: 48oC hold for 30 min, 95oC hold for 12 min, followed by 40 cycles of 95oC for 20 s and 60oC for 1 min. The wax layer remained intact for the RT step at 48oC but was melted during the 12-min, 95oC AmpliTaq Gold activation step, thus allowing the two layers to mix before the PCR begins. Data were analyzed with Applied Biosystems’ sequence detection software. First, we evaluated the effect of the wax layer on the fluorescence detection in the TaqMan assay to determine the extent of fluorescence quenching by the wax. Using randomly primed cDNA from a lung adenocarcinoma cell line (A549), we performed 20 replicates of PCR for β-gus with and without the wax layer. The results showed no decrease in the overall fluorescence (P = 0.935) and no change in the cycle threshold value (P = 0.55) between the two groups when compared by the independent samples t-test. To compare the sensitivity of the one-tube RT-PCR with and without the wax, we used serial dilutions of spleen total RNA (Clontech Laboratories, Palo Alto, CA, USA) from 5 ng to 10 pg. The results of the RT-PCR without the wax layer showed that the fluorescence (∆Rn) was weak even at 5 ng RNA input and decreased further by an average factor of 75% every dilution (Figure 1A). As a result, the 200-pg sample fell below the threshold for detection. However, with the use of the wax layer, we saw that the ∆Rn remained essentially the same down to the 40-pg dilution, and only at 10 pg was there a 60% drop in the ∆Rn (Figure 1B). Thus, our modified procedure resulted in at least a 20fold increase in sensitivity. The efficiency (E) of the RT-PCR (as calculated by the formula E = 10(1/-s)-1, where “s” is the slope of the standard curve from the dilutions) (5) was also improved by the use of the wax (67% without wax and 77% with wax). A 10-μL aliquot of each reaction was run out on a 10% Benchmarks