The identification of additional DNA markers is vital for the generation o f more detailed genetic and physical m ^ s . The localisation of these markers to the region of interest by performing linkage studies can only be achieved if the markers are polymorphic. However, alternative resources known as somatic cell hybrids, are available
which allow all types o f DNA markers to be rapidly localised. These hybrid cell lines are rodent cells containing either single human chromosomes or subchromosomal fragments o f human DNA. Southern blot hybridisation o f probes to DNA from these cell lines can be used to m ^ markers but, the analysis of hybrid DNA using PCR is the most efficient way to localise them.
The earliest cell hybrids were generated in the 1960's. Littlefield (1964) was the first to use a hybrid selection system, known as HAT selection, to identify rare spontaneous fusion events between fibroblasts. This selection involves growing cells in HAT (hypoxanthine, aminopterin and thymidine) medium. Aminopterin blocks the de novo pathway, by which normal cells manufacture nucleotides from sugars and amino acids, and forces the cells to use the alternative pathway o f nucleotide synthesis (the salvage pathway). In the salvage pathway, the enzymes thymidine kinase (TK) and hypoxanthine phosphoribosyl transferase (HPRT) utilise exogenous hypoxanthine and thymidine, respectively for the production of nucleotides. In the experiments performed by Littlefield (1964), fusion events between TK-deficient cells and HPRT-deficient cells were selected for by culturing in HAT medium as only the hybrid cells were able to grow. This system has been adapted so that only the faster growing parental line is under selection, and relies on the hybrid cells out-growing the other parent (Davidson and Ephrussi, 1965). Harris and Watkins (1965) reported that inactivated Sendai virus could be used to encourage fusion between cells of widely different species. They demonstrated that hybrids obtained by fusing human and mouse tumour cells were viable and underwent at least one or two mitotic divisions. The first established cell line of a hybrid containing human components was generated in 1967 (Weiss and Green, 1967).
subsequent cell divisions. Certain observations concerning the loss o f these chromosomes have been consistently made. Firstly, hybrids generated from the same species, i.e. intraspecific hybrids, tend to lose chromosomes very slowly, whereas interspecific hybrids lose chromosomes very rapidly, and usually the chromosomes which are lost are all o f one parental type. Secondly, most of the chromosome loss occurs in the first few cell divisions and often, hybrid cells become stable after a short period o f growth, allowing them to be cultured without any further loss of human material.
Hybrids have been generated which contain single human chromosomes with known translocations resulting in subchromosomal fragments from more than one chromosome. In addition, single-chromosome hybrids subjected to mutagenic agents have been used to generate new hybrids deleted for parts of a human chromosome (Kao et al., 1976). Hybrids containing only portions of a particular chromosome are more useful for mapping studies as they allow markers to be localised more precisely.
Somatic cell hybrids contain relatively large chromosomal fragments which limits their use for more precise chromosomal localisations. Cox et al. (1990) developed an alternative technique for generating hybrid cells which contain much smaller sub chromosomal fragments of DNA. These hybrids, known as "radiation hybrids" are generated by using a somatic cell hybrid donor cell containing a single human chromosome which is subjected to y-irradiation. This was an adaptation o f the original ^ p ro a c h devised by Goss and Harris (1975) which used a diploid human cell. The irradiation causes chromosomal breakage; the degree o f breakage being dependent on the dose of x-rays given. However, irradiation also results in the death of the donor cell and therefore, the chromosomal fragments have to be rescued by fusion to rodent recipient cells. The irradiated cells carry an endogenous or integrated selectable marker e.g. the
neomycin resistance gene {ned). The recipient cells are deficient for the marker and so, in the appropriate selection media only fused cells can survive. The advantage o f using a somatic cell hybrid as the donor cell is that all the human DNA fragments retained in the radiation hybrids are derived from a known chromosome.
A single hybrid cell line is not a particularly useful mapping tool whereas, the use o f a panel of hybrids, each containing a well-defined fragment o f a particular chromosome, can be used together to finely localise a particular marker. In addition, radiation hybrids can be used specifically to estimate the distances between markers and also their order, in a manner analogous to meiotic m o p in g (Cox et al., 1990). Approximately a hundred hybrid clones containing fragments of a particular chromosome are analysed for the presence or absence of specific DNA markers. The further apart two markers are on the chromosome, the more likely a given dose o f X-rays will break the chromosome between them, placing markers on two separate fragments. Therefore the frequency of breakage between two markers can be used as a measure of distance. However, because radiation hybrids often retain more than one fragment, the calculation o f the frequency of breakage has to allow for this. Unlike a meiotic recombination frequency, which can vary from 0 to 0.5, the breakage frequency varies from 0 to 1.0; where 0 indicates two markers are never broken ^ a r t and 1.0 indicates they are always broken apart. Although the breakage frequency is a good measure of the distance between markers that are close together, it can underestimate the distance between markers which are far apart. The mapping function (D), which assumes no interference gives a more accurate estimate:
D = -ln (l-0 )
quoting distances in cR, information about the X-ray dose is given e.g., 1 cRgooo corresponds to a 1% frequency of breakage after exposure to 8000 rad o f X-rays. To allow comparison of distances between markers generated using different doses o f X-rays, cR distances have to be scaled to physical distances using methods such as pulsed-field gel electrophoresis (section 1.1 4) or generating DNA contigs (see below).