cDNA Libraries 

In contrast to genomic libraries, which contain raw DNA sequences harvested from an organism's chromosomes, cDNA libraries are composed of processed nucleic acid sequences harvested from an organism's RNA pools (see Complementary DNA (cDNA)). cDNA can be prepared from any RNA source, including in RNA and noncoding RNA, isolated from single cell organisms, cultured metazoan cells, and isolated tissues. cDNA libraries provide a powerful means of examining cell- and tissue-specific gene expression. For example, mammalion cDNA libraries contain only the fraction of sequences that are expressed in the subject tissues at the time of harvest, typically in the range of 10% of the genes carried by the genome. Also, cDNA is typically prepared enzymatically from polyadenylated mRNA, a population of RNA that has previously undergone post-transcriptional editing and removal of any intervening sequences (introns). Whereas genes in genomic DNA are often interrupted by numerous introns, the copy of a gene in a cDNA or mRNA commonly contains an uninterrupted sequence encoding the gene product. Thus, cDNA inserts are directly expressible as both RNA and protein gene products. Vectors used in cDNA libraries often contain expression elements that mediate transcription and translation of the cloned cDNAs.

In constructing cDNA libraries, the goal is usually to represent full-length protein coding sequences for all of the mRNAs expressed in the subject cells or tissue. Thus, it is essential to preserve the integrity of cellular mRNA during extraction and purification protocols. Libraries made from fragmented RNA will not preserve the full open reading frame, rendering the resultant partial clones less useful. Unfortunately, RNA tends to be a transient species by design, as cells have developed complex post-transcriptional mechanisms to regulate gene expression through mRNA decay. The problem is worse in certain tissues, such as pancreas, that secrete large amounts of degradative enzymes as part of their normal function. Though troublesome, the RNA decay problem can be overcome through specialized procedures that seek to inactivate ribonuclease enzymes concomitantly with RNA extraction from intact tissue.

 Once total RNA is extracted from the subject tissue, the mRNA fraction is purified by annealing to oligodeoxythymidine (oligo-(dT)) cellulose. Oligo-(dT) is a short polymer that hybridizes specifically to polyadenylate. Because the majority of cellular mRNAs are polyadenylated, affinity chromatography over oligo-(dT) cellulose provides a simple means to remove contaminating RNA species, such as ribosomal and transfer RNAs, which do not contain poly A tails. The purified mRNA is then annealed to oligo-(dT), which serves as a primer that can be extended by the enzyme reverse transcriptase (RT). RT extension thus generates antisense DNA copies of the purified mRNA species. Under appropriate conditions, RT can be induced to complete a second-strand DNA synthesis, using the first DNA strand as a template. The result is a collection of double-stranded DNA copies of the parental mRNA population (hence the “c” in “cDNA”). The double-stranded DNA population can then be cloned into an appropriate plasmid or bacteriophage vector to constitute a cDNA library.

cDNA libraries may be screened either by hybridization to the cloned inserts or by detection of RNA or protein expression products encoded by the cDNA inserts. Hybridization screening is similar in principle for both genomic and cDNA libraries and uses synthetic oligonucleotide probes or gene probes derived from existing clones. The probes are labeled with radioactivity or a fluorescent tag and then annealed to the total nucleic acids contained within a population of bacterial colonies or viral plaques. Colonies or plaques that contain nucleic acids complementary to the probe can then be detected and purified. Protein expression libraries can be screened based on affinity to antibodies, proteins, nucleic acids, or small-molecule ligands. In addition, functional properties can be screened for, such as catalysis or induction of transcription of a reporter gene.

 Useful information regarding tissue-specific patterns of gene expression can be ascertained through screening cDNA libraries. By definition, a cDNA population is derived from the fraction of genes expressed in a given tissue through sampling the mRNA pool. These expressed sequences can be sampled and characterized in brute force fashion by high-throughput random sequencing of cDNA clones isolated from a library. It is not necessary to determine the full-length cDNA sequence in order to derive useful information from this approach. The partial sequences obtained are entered into an Expressed Sequence Tag (EST) database where the frequency of each partial sequence (known as a “tag”) can be scored (1). In this manner, it is possible to determine the expression level of individual genes for any tissue. Furthermore, investigators that identify a gene product of interest in the EST database are already one step ahead on their sequencing efforts, because much of the open reading frame may already be in the public database. Finally, it is possible to request many EST clones from the I.M.A.G.E. Consortium (Integrated Molecular Analysis of Genomes and their Expression; http://www-bio.llnl.gov/bbrp/image/image.html), providing an easy source of genetic material for subcloning projects.

The concept of ESTs as quantitative indicators of gene expression levels has been further refined through a new method termed Serial Analysis of Gene Expression (SAGE) (2). SAGE provides a means of tagging every mRNA in the cell with a small, intrinsic nine-nucleotide sequence that is essentially unique to each transcript. In addition, SAGE provides a means to eliminate artifacts arising through processing and polymerase chain reaction (PCR) amplification of the original cDNA, thereby rendering the method highly quantitative. Finally, SAGE enables high throughput sequencing of the nine-nucleotide tags on a scale sufficient to quantify several hundred thousand cDNA species for a given tissue (3). Although the sequence information obtained from a single SAGE tag is minimal, it is usually sufficient to identify the transcript via hybridization or database searching methods. SAGE is being conducted on a large scale through the Cancer Genome Anatomy Project (CGAP; (site currently unavailable)), with the goal of rendering in detail the molecular differences between various normal tissues and tumors, in order to understand better the molecular basis of cancer. Ultimately the CGAP database, among others, will provide the capability of performing “virtual” Northern blots, in which any sequence segment can be screened through WWW-based databases to ascertain all known sequence, genetic, and biological information associated with mRNAs containing the sequence segment.

Tissue- or cell-specific gene expression can also be studied through specialized cDNA libraries derived from the subset of genes that are differentially expressed between two preparations of mRNA (4). For example, a cell culture line can be treated with a hormone, such as dexamethasone, and RNA can be prepared from samples of the cell culture before and after treatment. cDNAs prepared from the pretreatment population can then be used to remove selectively any complementary mRNAs from the post treatment preparation through a process known as subtractive hybridization. The majority of the mRNA species remaining in the post treatment pool should be those transcripts that were synthesized as a direct result of hormone treatment. These specific mRNAs can then be converted into cDNA and cloned into a suitable library vector, to constitute a differential cDNA library. Screening or sequence tagging a subtractive library greatly increases the efficiency of characterizing tissue-specific or stimulus-specific differences in gene expression.

References

1. M. D. Adams, M. Dubnick, A. R. Kerlavage, R. Moreno, J. M. Kelley, T. R. Utterback, J. W. Nagle, C. Fields, and J. C. Venter (1992) Nature 355, 632–634. 

2. V. E. Velculescu, L. Zhang, B. Vogelstein, and K. W. Kinzler (1995) Science 270, 484–487. 

3. L. Zhang, W. Zhou, V. E. Velculescu, S. E. Kern, R. H. Hruban, S. R. Hamilton, B. Vogelstein, and K. W. Kinzler (1997) Science 276, 1268–1272. 

4. J. S. Wan, S. J. Sharp, G. M. Poirier, P. C. Wagaman, J. Chambers, J. Pyati, Y. L. Hom, J. E. Galindo, A. Huvar, P. A. Peterson, M. R. Jackson, and M. G. Erlander (1996) Nat Biotechnol 14, 1685–1691.