Differential screening of gene expression difference enriched by differential display.

Since its original description in 1992 (1), differential display has become a widely used methodology for cloning differentially expressed genes. Much effort has since been made to further optimize and streamline the technique (2,3). However, one major bottle neck remaining for the method lies at the screening step for the verification of the cDNA fragments identified by differential display. The standard protocol recommends the use of reamplified cDNA directly as a probe for Northern blot verification (4). The other approach is to clone the reamplified cDNA and then use it as a probe. Both approaches have their pros and cons. It is known that the reamplified cDNA probes from differential display sometimes contain a mixture rather than a single cDNA species, due to the contamination of other co-migrating cDNA fragments. As a result, direct use of the cDNA reamplified may be problematic because multiple messenger RNA species may be detected, making the identification of the truly differentially expressed gene difficult. On the other hand, if the cDNA probe is cloned, the choice of which clone to use as a probe becomes a matter of luck depending on the extent of cDNA contamination. Compounding the problem is the laborious screening by Northern blot for each cDNA fragment identified or purified. It is especially difficult under circumstances where the amount of RNA samples is limited. To this end, various approaches have been tried to improve the method (5). Here we describe the use of differential display followed by differential screening, using ‘Reverse Northern’ and colony hybridization to screen for cDNA fragments that truly represent differentially expressed mRNAs. The method takes advantage of the high cloning efficiency of the pCR-TRAP cloning system, which is based on positive-selection for inserts. Reamplified cDNA products from differential display are ligated with the pCR-TRAP vector and only recombinant plasmids confer antibiotic resistance. Antibiotic resistant colonies are transferred onto duplicate filters, which are then probed with cDNAs labeled by reverse transcription from the ‘plus’ and ‘minus’ RNA samples. As part of our ongoing effort to apply differential display to identify genes that are regulated by the Ras signaling pathway, the immortalized rat embryo fibroblast cell line Rat1 and its H-ras transformed derivative, T101-4 (6) were compared. Differential display was performed using the RNAimage kit (GenHunter, Nashville, TN) essentially as instructed. The cDNA fragments of interest (Fig. 1A) were retrieved from the denaturing polyacrylamide gels and reamplified with the same sets of primers as in the initial differential display PCR reactions. Five microliters of the PCR products were directly ligated with 2 μl pCR-TRAP vector (GenHunter, TN) as instructed. The ligated plasmids were transformed into GH competent cells (GenHunter, TN) and plated on LB plates containing 20 μg/ml tetracycline. After an overnight incubation at 37 C, 50–200 tetracycline resistant colonies were obtained with each cDNA insert, while the vector self-ligation gave no colonies (data not shown). Reverse Northern dot-blot was performed initially to test the feasibility of differential screening. Specifically, four tetracyclineresistant colonies were randomly picked from each plate and lysed by boiling in 50 μl lysis buffer consisting of 0.1% Tween 20 in TE buffer, pH 8.0. Using the primers that flank the cloning site of the vector, the cloned cDNA fragments were amplified from the colony lysates (Fig. 1B). As can be seen, most of the colonies analyzed contained inserts of the expected sizes. To prepare for blotting, 30 μl of each PCR-amplified cDNA insert from Figure 1B was mixed with 10 μl 2 N sodium hydroxide, denatured by boiling for 5 min, and then neutralized with 10 μl 3 M sodium acetate, pH 5.0. After bringing the total volume to 110 μl with dH2O, 50 μl of each sample was dot blotted onto duplicate Nylon membranes using the Bio-Dot microfiltration system (BioRad, CA). The membranes were UV cross-linked and rinsed in 6× SSC before being used for hybridization (4). The cDNA probes were prepared from 10–50 μg of each of the two RNA samples by reverse transcription in a 50 μl reaction which consists of 25 mM Tris–HCl, pH 8.3, 38 mM KCl, 1.5 mM MgCl2, 5 mM DTT, 30 μM dNTP (without dCTP), 1 μM T20 primer and 25–100 μCi [α-32P]dCTP (3000 Ci/mmol; NEN, MA). After a 5 min incubation at 65 C, the samples were shifted to 37 C and 1000 U MMLV reverse transcriptase (GenHunter, TN) was added, followed by continued incubation for 1 h. After the reverse transcription, Quick Spin column (Boehringer Mannheim, IN) was used to remove the unincorperated 32P. Equal counts (5–10 × 106 c.p.m.) of the cDNA probes from Rat1 and T101-4 were heat denatured and used to probe the duplicate blots. The results from the dot-blot experiment are shown in Figure 1C. Mob-5, a previously cloned gene expressed only in the transformed cells (Liang et al., unpublished result), was used as a positive control. Both clones of mob-5 showed signals with the cDNA probe from T101-4 cells but not from the Rat1 cells, as expected. As a negative control, two clones from a band equally expressed from the differential display (Fig. 1A) gave equal signals on dot blots for both cDNA probes, indicating equal probe labeling. Three of the Mob-36 clones showed quantitative difference between the two cell lines, with the forth clone (Mob-36d) being a false positive. Mob-37, Mob-38 and Mob-39 each had 3, 2 and 4 clones scored positive, respectively. Mob-37d (no insert) and