Rapid Analysis of Gene Expression (RAGE)
Overview
The basic strategy of the RAGE technique is to use PCR methods to amplify fragments of the cDNAs present in two populations, e.g. experimental and control treatments of a single cell type, or wild type and transgenic cells, or tumor and normal cells. The amplifications are done under conditions where each cDNA gives rise to a unique amplification product (amplimer) and the amount of product formed is directly proportional to the concentration of template. We then simply quantitate the amount of amplimer produced in reactions from paired populations and compare them. This gives a measure of the relative expression of the chosen gene in the two cell populations.
The important features of the RAGE technique that allow us to do this are:
- We use a defined fragment of each cDNA as template rather than the entire molecule;
- We use a limited number of different primers, all of which work well under a defined set of conditions.
To prepare the templates for RAGE amplification, a unique fragment of each cDNA, the gene-specific target, is prepared by digestion with restriction nucleases and poly(A) selection, and primer binding sites are ligated onto the cut ends of the fragments. The amplifications are done with primers that contain the sequence of the primer binding site, but extend 3-4 nucleotides further in the 3' direction, the specificity region. Because the primers share common primer binding sequences, the optimum conditions for amplification by different pairs of primers are fairly similar, and all amplifications can be conducted efficiently under a single set or a narrow range of conditions. The specificity regions of the primers limit the number of different cDNA fragments that can act as template for any given pair of primers to two-three, on average. The amplimers produced in any given reaction will generally be distinguishable by size, and can be separated and quantitated by standard gel electrophoresis methods.
These principles are explained in more detail in the illustrations by following the cDNAs for three genes through the procedure. These three genes, ribosomal protein S5, heat shock protein HSP27, and the oncogene MLN62 are all expressed in both normal human mammary epithelial cells and the breast tumor line SKBR3.
![[Fig. 1 Definition of RAGE Targets in Three cDNAs]](Images/RAGE_tech_Fig_1.jpg)
Figure 1. Definition of RAGE Targets in Three cDNAs
The mRNA sequences for S5, HSP27 and MLN62 from GenBank were analyzed for the positions of Nla III and Dpn II restriction sites as shown in the Figure. The gene-specific target for each gene is shown in yellow; note that the cDNA is prepared with a biotinylated oligo(dT) primer. The gene-specific target is defined as the sequence between the last Nla III site (i.e. the site closest to the biotin label) and the closest Dpn II site moving in the 3' direction from the last Nla III site. This "on-paper" definition will be made into an experimental definition by the steps shown in Figure 2 below.
![[Fig. 2 Preparation of Bitags]](Images/RAGE_tech_Fig_2.jpg)
Figure 2. Preparation of Bitags
The template that is ultimately used in the specific PCR reactions, called a bitag, is prepared from the cDNA as follows. The cDNA is cut with Nla III, leaving a 3' overhang of 5'-CATG; by convention we term this the "A-end". A common linker sequence with a matching sticky end, linker "A", is ligated onto the cut ends, and the fragment corresponding to the 3' end of the cDNA is isolated by binding to streptavidin-coated magnetic beads through the biotin label. The fragments of the cDNAs that lack the biotin label, and therefore are not the 3'-ends, are washed from the beads and discarded. The fragments of cDNA attached to the beads are then cut with Dpn II, and subfragments that lack the biotin label are washed from the column and purified; the fragments remaining on the beads are discarded. The Dpn II restriction enzyme leaves a 5'-GATC overhang (the "B"-end), and a second common linker (linker "B") with a matching sticky end is ligated onto these cut ends. The resulting preparation contains the desired bitags, consisting of the desired gene-specific targets with an A linker at one end and a B linker at the other end; because we added the A linker first, this preparation is called A/B bitags. It also contains products that failed one or both ligation steps and fragments that have two B-ends and hence two B-linkers. However, the A/B bitags are the only species that will amplify with a pair of primers specific for the A and the B linkers. A description of these primers is given in Figure 3 below.
![[Fig. 3 RAGE Primers]](Images/RAGE_tech_Fig_3.jpg)
Figure 3. RAGE Primers
Two sets of RAGE primers are constructed as follows. The A-end primers contain the sequence introduced by the A-end linker, followed by the Nla III restriction recognition sequence. The 3' end of these primers contains a 4-nucleotide "specificity region" with all possible combinations; there are 256 unique A-end primers. The B-end primers contain the sequence introduced by the B-end linker, followed by the Dpn II restriction recognition sequence. The 3' end of these primers contains a 3-nucleotide "specificity region" with all possible combinations; there are 64 unique B-end primers. The arrows in the Figure indicate primers that are used to amplify the bitags from the three model genes as shown in Figure 4 below.
![[Fig. 4 Bitags]](Images/RAGE_tech_Fig_4.jpg)
Figure 4. Bitags
The partial structures of the bitags for the three model genes, S5, HSP27 and MLN62, are shown. From the known sequence of each mRNA, one can infer the exact sequences of the RAGE primers which match the sequences at the A- and B-ends, and which will therefore amplify a particular bitag; the length of the amplimer produced can also be predicted. In the example chosen, two A-end primers and two B-end primers can be utilized in three unique combinations to amplify the bitags derived from each of the three genes. Actual results with these primers are shown in Figure 5 below.
![[Fig. 5 RAGE detection of S5, HSP27, and MLN62]](Images/RAGE_tech_Fig_5.jpg)
Figure 5. RAGE detection of S5, HSP27 and MLN62
A/B bitags were prepared from normal human mammary epithelial cells. PCR reactions were run with primer pairs specific for each of the genes, as indicated at the top of the Figure, and increasing amounts of the bitags as template (indicated by the wedges, 0.2 - 3.8 ng cDNA equivalent). PCR conditions were 1x "D" buffer (EpiCentre Technologies, Madison, WI), 27 cycles, 62° annealing temperature. Aliquots of the reactions were separated on an 8% polyacrylamide gel, stained with Vistra Green, and the resulting bands were visualized and quantitated using a FluorImager; standard DNA ladders were run in the same gel. The expected sizes of the amplimers (arrows at the left of the Figure) were: HSP27, 60 bp; S5, 41 bp; MLN62, 153 bp. For each gene, an amplimer of the expected size was produced in reactions with the predicted pair of RAGE primers, but not with the other pairs of primers. For MLN62, the amount of amplimer produced appeared to increase linearly with the amount of bitag used as template up to at least the equivalent of 3.8 ng cDNA per 25 µL PCR reaction. For HSP27 and S5, the amount of amplimer produced reached a plateau between 1 and 2 ng template; the linear region is confined to less than about 0.4 ng template per reaction. We assume that this is due to higher abundance of transcripts for these genes compared to the MLN62 oncogene in normal cells.