The case of chicken attachment region binding protein reinforces the importance of pre-screening of several libraries for cDNA cloning.

cDNA cloning very often involves screening of a library by hybridization with a short fragment of the desired cDNA as a probe. Commonly the choice of the library is made on the basis of the properties of the vector, the extent of expression in the tissue from which the library is made, the ‘quality ’ of the library, and possibly other parameters. However, even when these choices are optimally made, several caveats have to be considered. For example, a high expression rate is not a guarantee that a corresponding high number of positive clones will be obtained, and cases have been reported of successful screening of a library prepared from a weakly expressing tissue [1]. Furthermore, particular mRNAs are difficult to reverse-transcribe, possibly owing to a high CG content or base-paired structures, leading to an under-representation of the desired cDNA clone in the library, CG-rich sequences generally behave problematically in polymerase reactions, even at elevated temperatures (72 °C), and several attempts have been made to solve these problems [2,3]. In the severe case reported here, we attempted to clone the cDNA of the chicken nuclear-matrix protein ARBP (attachment region binding protein) [4]. We first synthesized a single-stranded cDNA by reverse transcription of chicken brain polyadenylated RNA and produced a 200 bp fragment by PCR amplification using oligonucleotide primers derived from partial peptide sequences of the purified protein [5]. In a long-lasting screening project employing this fragment as a hybridization probe, we were unable to find any positive clone among a million phages each from a λgt10 brain library (Clontech), a λZAP library from MSB-1 cells (W. C. Earnshaw), or a λZAP liver library (Stratagene). We therefore decided to use a second PCR amplification that generates a 107 bp fragment as a pre-screening analysis to evaluate several alternative libraries for the presence of ARBP cDNA inserts. As differential decreases in phage titre would lead to erroneous results, we first amplified these libraries. A total of 1¬10' plaque-forming units (pfu) was plated on 15 cm plates, the phages were rinsed with λ dilution buffer, and titres were again determined (approx. 1¬10"" pfu}ml) [5]. A total of 1¬10) pfu of each amplified library was then incubated at 70 °C for 30 min in PCR buffer [10 mM Tris}HCl (pH 9±0)}50 mM KCl}1±5 mM MgCl # ] and subjected to PCR reactions containing (in a total volume of 50 μl) 0±5 μl of Taq polymerase (Pharmacia; 5000 units}ml), 0±2 mM each dATP, dCTP, dGTP and TTP, and 1 μM of each oligonucleotide primer JW29 and JW28 (5«GAAGCTCAAACAGCGCAAG-3« and 5«-AGTACGCGATCAGTTCCACC-3« respectively). These primers were located at nucleotides 366–384 and 453–472 respectively of the ARBP cDNA sequence. PCR was performed in a thermal cycler (OmniGene; Hybaid) for 40 cycles, with each cycle consisting of 95 °C for 30 s, 57 °C for 30 s and 72 °C for 30 s. At the beginning of the cycles, the samples were heated for 5 min at 95 °C and, at the end, for 10 min at 72 °C. A control sample lacked phage DNA template. In pioneering experiments, the combination of oligonucleotide primers, the optimal number of cycles and optimal annealing temperature were established to make the PCR reaction under the described conditions extremely reproducible, yielding large amounts of the correct 107 bp product and minimal amounts of non-specific products. PCR reaction products were electrophoresed in a 2% agarose}0±04 M Tris}acetate}0.001 M EDTA (pH 8±0) gel and revealed with ethidium bromide (Figure 1A). DNA fragments were then Southern-blotted on to a nitrocellulose filter and hybridized to