Rapid determination of the complexity of cDNA bands extracted from DDRT-PCR polyacrylamide gels
Rapid determination of the complexity of cDNA bands extracted from DDRT-PCR polyacrylamide gelsNeil R. Smith1,2,*, Alan Li1, Mark Aldersley1, Alec S. High2, Alexander F. Markham1 and Philip A. Robinson1,2
1Molecular Medicine Unit and 2Leeds Dental Institute, Clinical Sciences Building, St James's University Hospital, University of Leeds, Leeds LS9 7TF, UK
Received March 27, 1997;Revised and Accepted July 8, 1997
ABSTRACT
A band extracted from a differential display polyacrylamide gel often represents a composite of heterogeneous products. We have developed a non- radioactive method to simply and rapidly analyse its complexity. A fluorescent restriction enzyme fingerprint of the composite mixture is generated. The number of individual bands observed in this fingerprint indicates the complexity of the re-amplified cDNA mixture. Restriction fingerprints of the inserts of cDNA subclones derived from the re-amplified cDNA mixture are compared to the composite fingerprint to select those representing the most intense bands in the composite. This dramatically reduces the number of clones required for further characterisation.
Differential display reverse transcription polymerase chain reaction (DDRT-PCR) is one of the most widely employed techniques to identify differential gene expression (1 ,2 ). However, a major problem is often encountered when extracting PCR products of interest from denaturing polyacrylamide gels for further characterisation; an apparently single band on a DDRT-PCR gel is frequently a composite of identically sized but distinct cDNA sequences. These `contaminating cDNAs' are co-purified with the differentially transcribed cDNA and are likely to be responsible for the large frequency of false positives associated with DDRT-PCR. It is essential, therefore, that the moiety of interest is separated from the contaminating cDNAs before further characterisation is undertaken.
Generally, in order to overcome this problem the gel eluted material is subcloned and the insert corresponding to the differentially transcribed gene identified by northern blot analysis. At this stage, however, the complexity of the eluted cDNA is not known. Consequently, neither is the number of colonies that need to be screened before the clone containing the differentially expressed moiety is identified. Individual clones can be distinguished by sequencing, the most abundant species can then be taken for northern analysis. But again, as the complexity of the eluted cDNA is unknown then the number of clones that need to be sequenced before the most frequently occuring cDNA is identified cannot be predicted. These approaches are, therefore, time consuming and require large amounts of RNA.
One strategy to assess the complexity of DDRT-PCR bands employs single strand conformation polymorphism (SSCP) gels to resolve the individual cDNAs constituting the mixture (3 ). However, as the size and complexity of the DDRT-PCR band increases it becomes increasingly difficult to clearly establish the number and abundance of individual cDNAs in the composite due to products occupying the same positions on the SSCP gel. Alternative approaches employ radiolabelled cDNA from the excised band (4 ) or DDRT-PCR reaction (5 ) as probes to screen for the differentially amplified moiety. Only one differential can be screened at a time by these approaches. In addition, isolated bacterial colonies can be screened with labelled cDNA probe prepared from total RNA extracted from the two populations of cells under investigation (6 ). The major drawback of this approach is that it does not detect transcripts of low abundance and is time consuming, requiring the screening of duplicate colony lifts.
We have developed a non-radioactive strategy to circumvent these problems based on our previously described method of fluorescent DDRT-PCR or FDDRT-PCR (7 ). In FDDRT-PCR, modified 3'-anchoring oligo (dT) primers are employed to reverse transcribe total RNA such that a common 20mer sequence (dTGGTCTCACGGATCCGTCGA) is introduced at the 5'-end of every cDNA. A fluorescently-labelled (FAM-5-carboxyfluoroscein- blue) universal primer (FAM-dCTCACGGATCCGTCGATTTT) is then used in every PCR together with an arbitrary 10mer to generate 3' fluorescently-labelled cDNA which is analysed on an ABI Automated Sequencer (Perkin-Elmer). When a differential is identified by FDDRT-PCR the appropriate PCR is repeated in the presence of radioactivity and the segment of denaturing polyacrylamide gel containing the band of interest is cut out and cDNA eluted.
The complexity is assessed by first re-amplifying the eluted cDNA using a combination of the fluorescently-labelled universal primer and random 10mer used in the original amplification procedure. The amplified product is then divided such that one aliquot is taken for restriction enzyme analysis and the other is subcloned into an appropriate vector. Individual colonies are selected at random from the subcloned material and the inserts amplified using the same combination of primers used for the first round of amplification. Both the non-cloned and cloned amplicons are then digested with a series of restriction enzymes.
Restriction enzyme digestion of a fluorescently-labelled cDNA moiety will yield a single fluorescently-labelled product corresponding to the 3'-end of that mRNA. Thus, a heterogeneous mixture of end-labelled cDNAs of this type, produced by the re-amplification of cDNA eluted from a DDRT-PCR band, will yield a mixture of different sized fluorescently-labelled digestion products. These are separated and visualised by electrophoresis using an ABI Automated Sequencer thereby generating a restriction enzyme fingerprint of the gel eluted cDNA mixture. Each band size generated with a restriction enzyme will be characteristic of an individual cDNA. Hence, the number of different bands observed in the fingerprint will indicate the maximum number of cDNAs originally present in the eluate. Employing 4 bp cutting restriction enzymes will increase the chance of cutting a small differential display band. Upon cloning of the re-amplified cDNA mixture the same number of different clones would be expected. Their individual fingerprints can then be aligned with the composite fingerprint produced from the uncloned material. In this way the maximum number of different clones that need to be analysed can be immediately established. This procedure vastly decreases the number of amplified false positive inserts that need to be taken as likely candidates for confirmatory RNA studies such as northern blotting.
REFERENCES
1 Liang, P. and Pardee, A.B. (1992) Science257, 967-971.MEDLINE Abstract
2 Bauer, D., Miller, H., Reich, J., Rieder, H., Ahrenkiel, V., Warthue, P. and Strauss, M. (1993) Nucleic Acids Res.18, 4272-4280.
3 Mathieu-Daud, F., Cheng, R., Welsh, J. and McClelland, M. (1996) Nucleic Acids Res. 24, 1504-1507.
4 Callard, D., Lescure, B. and Mazzolini, L. (1994) BioTechniques16, 1096-1103.MEDLINE Abstract
5 Vsgeli-Lange, R., Burckert, N., Boller, T. and Wiemken, A. (1996) Nucleic Acids Res. 24, 1385-1386.
6 Zhang, H., Zhang, R. and Liang, P. (1996) Nucleic Acids Res. 24, 2454-2455.MEDLINE Abstract
7 Smith, N.R., Aldersley, M., Li, A., High, A.S., Moynihan, T.P., Markham, A.F. and Robinson, P.A. (1997) BioTechniques, 23, in press.
RELATED PAPERS RECENTLY PUBLISHED IN NUCLEIC ACIDS RESEARCH
