Testing of SYBR Green based Real-Time PCR for analysis of conditional gene deletion

To allow analysis of gene deletion using SYBR Green based Real-Time PCR we designed primers allowing specific amplification of the 1lox and 2lox allele repectively. As illustrated in Figure 1A positioning of the 2lox specific primers is comparatively easy because they can be placed anywhere within the potentially deleted region. However it has to be considered that in some cases the excised gene fragment can persist, particularly in non-proliferating tissue after initiation of Cre expression [8]. Therefore in such cases - in tissues with low mitotic activity and shortly after induction of Cre expression - 2lox primer placement in analogy to Figure 2A might be preferable. When designing 1lox specific primers, care has to be taken to place them within unique genomic sequences flanking the remaining loxP site of the truncated 1lox allele (Figure 1A). However, when following this approach to quantify conditional Dnmt3a deletion in our mice we noted the continuous presence of a double peak in the melting curve of the 1lox reaction (Figure 1B). This double peak persisted even when choosing different annealing temperatures, multiple alternative primer pairs for the 1lox reaction, different primer concentrations and different DNA concentrations. Since the presence of two peaks in the melting curve of SYBR Green based Real-Time PCR assays frequently indicates the presence of two different amplicons we analyzed the resulting PCR product using both gel electrophoresis (Figure 1C) and capillary electrophoresis (data not shown). Surprisingly, although we had observed two peaks in the melting curve analysis, we could identify only one distinctive PCR product. Furthermore when we size selected the PCR product using gel purification to exclude primers and alternative products and repeated the melting curve analysis we reproduced the same double peak (data not shown), suggesting that the amplified sequence itself was responsible for this aberrant signal. A possible explanation for this observation is that the palindromic loxP sequence motif could serve as a core motif that promotes annealing of sense/sense and antisense/antisense strands or possibly the formation of hairpins. Such hybridization products/hairpins would have a lower melting temperature than antisense/sense hybrids and therefore contribute to a second peak in melting curve analysis. Since the melting curve is critical for quality control in SYBR Green assays, data with an aberrant double peak are frequently excluded from further analysis. Our example illustrates that SYBR Green based assays are not only limited by possible amplification and measurement of nonspecific PCR products but are also potentially affected by alternate re-annealing or secondary structures of the specific PCR product itself. In agreement with our own data, difficulties with SYBR Green based analysis of 1lox/2lox allele ratios have also been reported independently [8].

To resolve such potential difficulties of SYBR Green based analysis of loxP containing sequences we therefore developed an alternative assay to quantify 1lox/2lox allele ratios using a probe based Real-Time PCR approach. Our aim was to establish a universal probe that could be used for all reactions and models of Cre-lox mediated gene deletion such that only the flanking primer sequences need to be adapted. Such an assay would combine the convenience of a SYBR Green based assay with the specificity of a probe based assay.

Testing of probe based Real-Time PCR for analysis of conditional gene deletion

To establish a probe based assay for the 1lox and 2lox reaction respectively we selected a hydrolysis probe (UP #69, Roche Applied Science) that specifically binds to the loxP site itself and designed primers flanking the loxP site as illustrated in Figure 2A. The use of a probe covering the loxP site (UP #69) allowed us to use the same probe for both the 1lox and 2 lox reaction. To generate allele specific standards that would allow us to validate the assay and test its accuracy we cloned 1lox and 2lox specific PCR products into a TA vector and verified resulting clones by sequencing. To generate standards for Real-Time PCR we then released allele specific fragments by restriction digest and mixed the fragments at variable molar ratios. The resulting 1lox/2lox standards were verified using capillary electrophoresis, which confirmed the expected ratios (Figure 2B and Figure 3A).

To determine the relative efficiency of the 1lox and 2lox specific Real-Time PCR reaction and to detect possible interference between the two alleles we conducted Real-Time PCR using 1:100 dilutions of the above mentioned fragment mixtures. As illustrated in Figure 2C we observed a linear correlation between the Crossing point (Cp) and Log concentration data for both reactions, with all data points close to the curve (r² > 0.99), indicating that the efficiency of both reactions was not affected by variable admixtures of the other allele. To further test the accuracy of the Real-Time PCR assay we conducted quantitive PCR, using incremental standards ranging from 9/1 to 1/9 and calculated allele ratios from the measured data. As summarized in Figure 3A the 1lox/2lox allele ratios as determined by the probe based Real-Time PCR assay closely reflected allele ratios measured by capillary electrophoresis (Figure 3B) and the expected allele ratios (Figure 3C).

Testing of probe based Real-Time PCR for analysis of Cre-lox recombination in Fabpl^{4× at -132}-Cre transgenic mice

After establishing the accuracy of the method we measured conditional deletion of Dnmt3a in intestinal epithelial cells of Fabpl^{4× at -132}-Cre transgenic mice. For this we harvested intestinal epithelial cells from the proximal, middle and distal third of the small intestine of Dnmt3a^(2lox/2lox), Fabpl^{4× at -132}-Cre^(+/-) mice and measured 1lox/2lox allele ratios using Real-Time PCR. As shown in Figure 3D we detected a low deletion efficiency in the proximal small intestine and increasing deletion efficiency in the distal small intestine, with the 1lox/2lox allele ratio varying from 0 to approximately 0.5 respectively. Importantly the data of the two Cre⁺ mice were highly similar and the variation within the sample replicates was very low, confirming the reproducibility of the method (Figure 3D). Also, these results are in good agreement with published data on the same Fabpl^{4× at -132}-Cre strain using Northern Blotting and Immunohistochemistry [10], which further confirms the utility of the assay.

Preparation of the 1lox/2lox standard

1lox and 2lox specific PCR products (1lox-forward primer: GCCGGCTTTTCCTCA GACAGTGGAGATAGC, 1lox-reverse primer: CCTGTGTGCAGCAGACACTTCT TTGGCGTC; 2lox-forward primer: CCTCTGGGGATTAAACTCTTGGCCAGCCC, 2lox-reverse primer = 1lox-reverse primer) were generated using genomic DNA from Dnmt3a^(2lox/2lox), Fabpl^{4× at -132}-Cre^(+/-), APC^(Min/+) transgenic mice. PCR products were then purified using agarose gel-electrophoresis followed by gel extraction (QIAGEN Minielute Gelextraction Kit) and cloned into TA vectors (TOPO-cloning Kit, Invitrogen). Minipreps were prepared from individual colonies resulting from both 1lox and 2lox ligations and verified by sequencing. Plasmids with verified sequences were digested with EcoRI to release fragments containing the 1lox and 2lox specific inserts, respectively. Following gel electrophoresis and gel extraction the concentration of resulting 1lox and 2lox fragments were measured using capillary electrophoresis (Agilent, High sensitivity DNA Assay) and 1lox-2lox mixtures were prepared at variable molar ratios (1/9, 2/8, 3/7, 4/6, 5/5, 6/4, 7/3, 2/8, 1/9). The 1lox/2lox ratios of these final mixtures were verified again using capillary electrophoresis.

Probe based Real-Time PCR

The hydrolysis probe based Real-Time PCR reaction was performed as follows: 1lox- and 2lox reactions were conducted separately in 10 μl reactions containing 0.25 μM each of forward and reverse primers (1lox-forward primer: TAATCCCAGCACTGC ACTCA, 1lox-reverse primer: TTCTTTGGCGTCAATCATCA; 2lox-forward primer: CCTCTGGGGATTAAACTCTTGGCCAGCCC, 2lox-reverse primer: CCTGTGTGC AGCAGACACTTCTTTGGCGTC), 0.05 μl Universal Probe #69 (Roche Applied Science), LightCycler 480 Probes Master Mix (Roche Applied Science) and 50ng genomic DNA or 4.3pg of the 1lox/2lox standard. The cycling conditions were as follows: Preincubation at 95°C for 10min, then 45 cycles of 95°C for 10sec, 58°C for 45sec and 72°C for 1sec. Each reaction was measured in triplicates. For data analysis we used "LightCycler(r) 480 Software release 1.5.0 SP3" (Roche Applied Science) and the "Advanced Relative Quantification" module where the crossing points (Cp = Ct) were determined using the "2nd Derivative Maximum Method". The mentioned module quantifies 1lox signal (T, Target) relative to 2lox signal (R, Reference), considering the respective amplification efficiencies (E) using the following formula: E _T ^CpT /E _R ^CpR . The efficiencies were determined using the slope of the respective standard curve (E = 10^-1/slope) with Cp values plotted against the logarithm of different dilutions. As dilution standards we used the incremental 1lox/2lox standard mixtures (10 standards, 10% dilution steps), in other words the samples themselves were used for calculation of reaction efficiencies. As a final control the PCR product was analyzed using agarose gel electrophoresis.

SYBR Green based Real-Time PCR

Real-Time PCR was performed using equipment of Roche Applied Science (LightCycler 480 SYBR Green I Master, Light Cycler 480). 1lox and 2lox-reactions were conducted separately in 10 μl volume containing 50ng DNA and 0.25 μM each of forward and reverse primers (1lox-forward primer: TAATCCCAGCACTGCACTCA, 1lox-reverse primer: TTCTTTGGCGTCAATCATCA; 2lox-forward primer: CCGAT GCAGACAGCCTCAGC, 2lox-reverse primer: CTTGTCACTAACGCCCATGG CCA). The following cycling conditions were chosen: Preincubation at 95°C for 10min, then 45 cycles of 95°C for 10sec, 60°C for 15sec and 72°C for 30sec. Each reaction was measured in triplicates. Data analysis was executed in analogy to Probe based Real-Time PCR. As a final control the PCR product was analyzed using agarose gel electrophoresis.

Animals

The Dnmt3a^(2lox/2lox), Fabpl^{4× at -132}-Cre^(+/-), APC^(Min/+) transgenic mice were generated as previously described for Dnmt3b^(2lox/2lox), Fabpl^{4× at -132}-Cre^(+/-), APC^(Min/+) mice [5, 6]. Briefly the conditional Dnmt3a allele contains loxP sites located in intron 17 and intron 20 of the endogenous Dnmt3a allele. Cre mediated excision results in deletion of exons 18-20, resulting in deletion of the catalytic domain. For a detailed description of the Dnmt3a ^(2lox/2lox) mice see [9], for details on Fabpl^{4× at -132}-Cre^(+/-) mice see [5].

No animal experiments were conducted with the intention of generating material for the present study. All mouse tissues and DNA analyzed in the present study were harvested for the purpose of an alternative, ongoing project analyzing the role of DNA Methyltransferase 3a in mouse tumorigenesis. All animal experiments of that project were approved by the MIT Department of Comparative Medicine, Boston, USA and were executed according to the institutional guidelines.

Harvesting of epithelial cells of the small intestine

Intestinal epithelial cells were harvested as described previously [13]. Briefly, the small intestine was dissected from mice, divided into three thirds of equal length (proximal, middle, distal), rinsed and incubated in 20 ml harvesting buffer (3 mmol/L ethylenediaminetetraacetic acid (EDTA) plus 0.5 mmol/L dithiothreitol in PBS) at 37°C, rotating at 250 rpm for 30 min. Following incubation the intestinal tube was removed and floating epithelial cells centrifuged at 200×g and 4°C for 4 min. After removal of the supernatant, cell pellets were frozen at -80°C.

DNA-Isolation

DNA of the harvested small intestine epithelial cells was isolated using the QIAGEN DNeasy Kit using the manufacturer's protocol.