cDNA selection/capture

The techniques described have been popular when dealing with inserts from cosmid dones or subdones. Increasingly, however, large-scale physical mapping procedures rely on building YAC contigs. For any YAC, cognate cosmids (cosmid dones containing sequences from the same lod as those present in the YAC) can be identified by various procedures, induding using (small) YACs as hybridisation probes to screen chromosome-spedfic cosmid libraries (Davies et al, 1981) or preparing cosmid libraries from the DNA of a purified YAC. An alternative is to try to identify expressed sequences directly from the YACs.

Direct cDNA selection techniques (also called direct selection and cDNA selection) involve forming genomic DNA/cDNA heteroduplexes by hybridising a complex doned DNA, such as the insert of a YAC, to a complex mixture of cDNAs, such as the inserts from all cDNA clones in a cDNA library (Lovett, 1994). The prindple underlying the technique is that cognate cDNAs corresponding to genes found within the YAC will bind preferentially to the YAC DNA; several rounds of hybridisation should lead to a huge enrichment of the desired cDNA sequences, enabling the identification of the corresponding genes. The first successful hybridisation selections described (Parimoo et al, 1991; Lovett et al, 1991) were essentially inverted screening of cDNAs with YACs. In these schemes, the YAC DNA was bound to a filter and a library of cDNAs was hybridised to this filter. After hybridisation, the filter was washed

extensively and the specifically hybridising cDNAs were eluted. Because the yield from this type of selection is low, a PCR amplification step was incorporated to generate sufficient eluted material for additional steps or for molecular cloning. In both initial reports, purified YACs and cosmids were used as genomic targets and cDNAs were derived from conventional libraries. However, cDNA inserts were amplified using the PCR and flanking vector primers to yield a population of inserts essentially devoid of contaminating vector sequences. Repetitive elements were suppressed by blocking either the cloned genomic DNA (Parimoo et al., 1991), or the population of cDNAs (Lovett et al.,

1991). It is a testimony to the robustness of the technique that botii strategies worked well, and enriched the desired cDNAs a few thousand-fold after two rounds of selection.

Filter hybridisation is however not as easily quantitated or as sensitive as solution hybridisation (Young and Anderson, 1987). Due to the lack of control over the hybridisation it is likely to be irreprodudble (Lovett, 1994). This is attributed to the pseudo-first-order hybridisation kinetics of filter hybridisations (Britten and Davidson, 1985) in contrast to the more easily controlled second-order kinetics in a solution hybridisation reaction. To better quantitate the repeat suppression and hybridisation steps, the selection system was modified to include biotin labelling and subsequent capture steps (Kom et al., 1992; Morgan et al., 1992). This biotin-streptavidin system (Figure 1.3) was tested and applied successfully to the isolation of cDNAs encoded by the Xq28 region in one study (Kom et al., 1992). When the same scheme (Figure 1.3) was applied to the regions surrounding the human interleukin 4 and 5 genes there was successful isolation of cDNAs in the study by Morgan et al., (1992), with enrichments achieved in the range of a few thousand-fold to more than 100,000 fold.

The direct cDNA selection method described here is a very simple, rapid and effective tool for the generation of a regional transcription map. It screens the genomic region studied as a whole and does not require its precise analysis, selection of single probes, subcloning into special vectors and the screening of large conventional cDNA libraries, if the full cDNA is not requested. cDNAs or cDNA libraries from multiple tissues can easily be mixed or used in parallel, thus increasing the probability of detection of tissue-specific transcripts. Furthermore, no specific demands such as evolutionary conservation, presence of a CpG island or favourable distribution of exon/ intron boundaries must be fulfilled to allow the isolation of the gene. The analysis of the sub-library is very rapid as many identical sub-library filters can be spotted and hybridised in parallel with many probes and the resulting short or medium size cDNA clones are good substrates for DNA sequencing.

Direct selection can be affected by factors associated with nucleic add hybridisations. One problem is that short exons fail to hybridise to their genomic locus effidently. Another possible disadvantage is the isolation of transcripts from other regions of the genome based on the presence of a pseudogene in the region studied, in

( f à

Mixture of com p lex amplified cD N A p o o ls

(random prim ed and dT sy n th e sis )

Block highly repetitive s e q u e n c e s to C o t= 2 0 Elute, PCR amplify, r e se le c t and clo n e IB h B |B h| B ■ i B B i B M C loned g en o m ic DNA labeled with biotin Hybridize to Cot >20

Capture gen om ic DNA plus hybridized cD N A S on streptavidin-coated param agnetic b e a d s B Biotin S Streptavidin I I R epetitive s e q u e n c e ele m e n t

■— cD N A in serts with am plification c a s s e t te s

Figure 1.3 A biotin-streptavidin capture system for direct cDNA selection. Lovett, M. (1994)

contrast to the CpG island approach (non-transcribed pseudogenes loose their CpG islands) or exon trapping (processed pseudogenes are not isolated). Conversely, the region-specific gene locus can serve as a template for the isolation of a transcribed pseudogene from another region of the genome in parallel with the isolation of the region-specific gene transcript. Complications arise due to the bias of the PCR amplification steps for shorter fragments (Kom et ah, 1992; Sedlacek et ah, 1993), leading often to a preponderance of relatively short clones in the sub-libraries, as well as the possibility of introducing sequence changes during the PCR amplification. Both problems can be overcome by the use of the probes to screen conventional cDNA libraries. Sub-libraries with large insert sizes can, however, be constructed, if appropriate precautions are taken. A summary of some advantages and disadvantages of direct selection is given in Table 1.2.

Table 1.2 Summary of some advantages and disadvantages