Quantification of alternatively spliced FGFR2 RNAs using the RNA invasive cleavage assay

Abstract

The regulated splicing of fibroblast growth factor receptor-2 (FGFR2) transcripts leads to tissue-specific expression of distinct receptor isoforms. These isoforms contain two different versions of the ligand binding Ig-like domain III, which are encoded by exon IIIb or exon IIIc. The mutually exclusive use of exon IIIb and exon IIIc can be recapitulated in tissue culture using DT3 and AT3 rat prostate carcinoma cells. We used this well-characterized system to evaluate the precision and accuracy of the RNA invasive cleavage assay to specifically measure FGFR2 alternative splicing outcomes. Experiments presented here demonstrated that the RNA invasive cleavage assay could specifically detect isoforms with discrimination levels that ranged from 1 in 5 x 10(3) to 1 in 10(5). Moreover the assay could detect close to 0.01 amole of FGFR2 RNAs. The assay detected the expected levels of transcripts containing either exon IIIb or IIIc, but, surprisingly, it detected high levels of IIIb-IIIc double inclusion transcripts. This finding, which has important implications for the role of exon silencing and of mRNA surveillance mechanisms, had been missed by RT-PCR. Additionally, we used the RNA invasive cleavage assay to demonstrate a novel function for the regulatory element IAS2 in repressing exon IIIc inclusion. We also show here that purification of RNA is not necessary for the invasive cleavage assay, because crude cell lysates could be used to accurately measure alternative transcripts. The data presented here indicate that the RNA invasive cleavage assay is an important addition to the repertoire of techniques available for the study of alternative splicing.

FIGURE 1.

Multiple layers of combinatorial interactions result in tissue-specific alternative splicing of FGFR2 transcripts. A schematic of exons 7–10 and introns 7–9 of FGFR2 is shown indicating that silencing of exon IIIb dominates in mesenchymal cells. In epithelial cells, however, a layer of regulation combines activation of exon 8 (IIIb) with repression of exon 9 (IIIc). These activities are mediated by two cis-elements: IAS2 and ISAR.

FIGURE 2.

Schematics of the Invader RNA assay probe sets and the FGFR2 RNA variants derived from the alternative splicing of exons IIIb and IIIc. (A) Splice variants generated from the minigene constructs used in this report contain U and D exons, which are derived from Adenovirus2 L1 and L2 exons. Variants, depicted in the 5′–3′ orientation, can also contain IIIb and/or IIIc exons, which are derived from the rat FGFR2 gene (see Materials and Methods). Five Invader probe sets (see Table 1 ▶ for oligonucleotide sequences) were used to uniquely detect the splicing products. It should be noted that the IIIc-D probe set could also recognize this junction in the U-IIIb-IIIc-D RNA (not shown in the figure). Each probe set consists of, from left to right, a stacking oligonucleotide, the probe (with noncomplementary 5′ flap), and an Invader oligonucleotide (see B for more details). Two probe sets were run per Invader reaction using the biplex detection format (see Materials and Methods for the probe set combinations). Invader probe sets are shown in the 3′–5′ orientation as black lines above the splice variants except for the probe 5′ flaps, which are color-coded green or red to signify the fluorophore (F₁ or F₂) used for detection (right side of figure). Vertical arrows indicate Cleavase enzyme cleavage sites, which results in the release of multiple 5′ flaps (left side) or unquenched fluorophores (right side). Curved arrows indicate use of the cleaved 5′ flaps produced in the first reaction (left side) in the second reaction (right side). See text for further description of the Invader assay. (B) Detailed view of the U-D Invader probe set. The stacking oligonucleotide, probe, and Invader oligonucleotide are shown aligned on the U-D transcript from left to right. A dashed line box indicates the overlap structure and the vertical arrow indicates the probe cleavage site for the Cleavase enzyme. Transcript sequences for the IIIb, IIIc, and IIIb-IIIC splice variants are also shown to illustrate the inability of the probe and Invader oligonucleotide to form the overlap structure required for cleavage of the 5′ flap. A separate duplex structure between the probe and arrestor oligonucleotide, which is added in the secondary reaction to enhance signal generation, is depicted. Underlined sequences for the arrestor and stacking oligonucleotides indicate use of 2`-O-methylated nucleotides (see text). For further details see Eis et al. (2001) and de Arruda et al. (2002).

FIGURE 3.

The Invader RNA assay discriminates between alternatively spliced FGFR2 transcripts. (A, B) The Invader RNA assay, using the IIIb-D (A) or the U-IIIc probe set (B), was carried out with increasing levels (0.01–10,000 amole) of the four variant RNAs (Fig. 2 ▶). Linear regression (R²) values for the fitted regions of the data are 0.992 and 0.998 for IIIb-D and U-IIIc, respectively. Error bars represent one standard deviation. (C) Table of the discrimination capability of the five probe sets on the four RNA variants. The discrimination capability for each probe set was calculated by dividing the highest target level in which no significant signal was detected for another RNA variant by the limit of detection for the probe set being tested. Values listed represent the inverse of the discrimination capability (e.g., the IIIb-D probe set has a limit of detection of 0.02 amole and it detects background signal from the U-D transcript at 5,000 amole: 5,000/0.02 = 250,000 discrimination level).

FIGURE 4.

Quantification of alternatively spliced FGFR2 transcripts. (A) Schematic of double-exon minigene constructs used to study the cell-type-specific inclusion of exons IIIb and/or IIIc. The presence or absence of IAS2 and ISAR is indicated. These minigenes can direct the expression of four RNAs produced by alternative splicing: skipping of IIIb and IIIc yields U-D RNA, inclusion of either IIIb or IIIc yields U-IIIb-D or U-IIIc-D, respectively, and double inclusion yields U-IIIb-IIIc-D. (B) DT3 cells and (C) AT3 cells were transfected with the minigene constructs in A. Total RNA was extracted and analyzed with either the Invader RNA assay (gray bars) or semiquantitative RT-PCR (white bars) for the presence of the four RNA variants. Error bars for the Invader RNA assay data represent one standard deviation. In B and C the percent inclusion of a product (e.g., U-IIIb-D) among transcripts was calculated as follows: percent product (e.g., U-IIIb-D) inclusion = {[product (e.g., U-IIIb-D)]/([U-D] + [U-IIIb-D)] + [U-IIIc-D] + [U-IIIb-IIIc-D])} × 100.

FIGURE 5.

Comparison of the Invader RNA assay, RNase protection assay, and RT-PCR assay in the quantification of FGFR2 transcripts. pI12DE:Wt, which is described in Figure 4A ▶, was transfected into DT3 cells. Total RNA was extracted and analyzed with the Invader RNA assay, RNase protection assay, or RT-PCR to quantify U-IIIb-IIIc-D RNA (gray bars), U-IIIb-D RNA (white bars), or U-IIIc-D RNA (black bars). Error bars represent one standard deviation. The percent inclusion of a product (e.g., U-IIIb-D) among transcripts was calculated as follows: percent product (e.g., U-IIIb-D) inclusion = {[product (e.g., U-IIIb-D)]/([U-IIIb-D)] + [U-IIIc-D] + [U-IIIb-IIIc-D])} × 100. This RPA cannot accurately quantify the skipped product (U-D) because the probe products protected by this splice variant were also generated from the protection of other spliced products (i.e., U-IIIb-D and U-IIIc-D).

FIGURE 6.

IAS2 represses the IIIc exon in epithelial cells. (A) Schematic of single-exon minigene constructs used to study the inclusion of exon IIIb and the repression of exon IIIc in DT3 cells. The presence or absence of IAS2 and ISAR is indicated. The IIIb minigenes can direct the expression of two RNAs produced by alternative splicing: skipping of IIIb leads to the U-D RNA and the inclusion of IIIb leads to the U-IIIb-D RNA. The IIIc minigenes can direct the expression of two RNAs produced by alternative splicing: skipping of IIIc leads to the U-D RNA and the inclusion of IIIc leads to the U-IIIc-D. (B) DT3 cells were transfected with the minigene constructs in A, and total RNA was extracted and analyzed with either the Invader RNA assay (gray bars) or semiquantitative RT-PCR (white bars) for the presence of the RNA variants. Error bars represent one standard deviation. In B the percent inclusion of a product (e.g., U-IIIb-D) among transcripts was calculated as follows: percent product (e.g., U-IIIb-D) inclusion = {[product (e.g., U-IIIb-D)]/([U-D] + [U-IIIb-D])} × 100.

FIGURE 7.

The Invader RNA assay accurately reports levels of FGFR2 alternatively spliced RNAs in crude cell extracts. DT3 cells were transfected with the single-exon minigene constructs described in Figure 6 ▶, and either purified total RNA (gray bars) or crude cell lysates (striped bars) were analyzed for the presence of U-D, U-IIIb-D (left panel), or U-IIIc-D RNAs (right panel). The percent inclusion of a product among transcripts was calculated as in Figure 6 ▶.