The advent of high density RFLP maps in crop plants and the development of specific methods for targetting DNA markers (e.g. RFLPs and RAPDs) to particular regions of a chromosome (Young et al. 1988, Martin et al 1991, Michelmore person. comm) opens up the possibility of using chromosome walking techniques to isolate genes of interest. However, before initiating a chromosome walk from a DNA marker to a linked gene of interest, it is imperative to know with certainty between which two DNA markers the target gene resides. This is relatively straightfoward when dealing with markers that are only loosely linked. However, if the DNA markers and the target gene are tightly linked (<1 cM), examination of hundreds or even thousands of segregating progenies may be required to determine the position of the target gene relative to its closest flanking DNA markers.
Recently we have encountered this problem in orienting DNA markers tightly linked to several disease resistance genes in tomato. In all cases, the resistance genes were linked to a number of DNA markers at distances < 0.1 cM. Faced with the prospect of having to extract DNA from potentially thousands of segregating individuals, we began looking for an alternative strategy for high resolution mapping. We have settled on a pooled mapping technique that potentially reduces the amount of time and labor required for high resolution mapping by a factor of 10 or more.
When mapping tightly linked markers relative to a target gene, only a very small proportion of the plants in a segregating population are informative in determining the order of the markers (ie. those that contain a crossover between the markers and the target). For example, if several DNA markers, along with a target gene, are all located within 1 cM interval, only 1% of the plants in a backcross or 2% of the plants in the F2 would be likely to contain a recombinant chromosome and thus to be helpful in deciding the linear order of the RFLP markers relative to the target gene. The remainder of the plants would provide no information about order of the markers. Unfortunately, with traditional mapping methods, one needs to extract DNA separately from all of the plants and to probe them with the markers in order to identify those rare individuals with informative crossovers in the interval.
The solution to this dilema is to pool individuals homozygous for the target gene in a segregating population (e.g. F2) before DNA extraction. DNA from the pools can then be extracted and examined for the rare crossovers. Using a dominant disease resistance gene as an example, a large F2 population of say 1000 plants is grown out and innoculated with the pathogen. Leaf samples from only the susceptible (and thus homozygous) individuals is pooled into groups of 10. Thus if 250 susceptible individuals were identified from the original 1000 plants, there would be 25 pools corresponding to leaf tissue from 10 plants each. DNA is then extracted from these 25 pools and probed with the DNA markers in the vicinity of the gene. Since most plants would not have a crossover between the marker and the target gene, most pools would reveal only the cis allele of the linked marker. If an individual with a crossover individual did occur in the pool, the trans allele could still be identified against the background of the cis allele. In the example given, 500 gametes (2 for each homozygous plant) would have been assayed for informative crossovers and this would be accomplished with only 25 DNA extractions and assays. We are currently using this method in our laboratory and find that leaf samples from at least 10 plants can be pooled while retaining the ability to observe a single crossover allele in the pool using RFLP markers. Larger pools may be possible using PCR or other more sensitive probing methods and would allow even finer mapping with a reduced amount of effort.
References:
Martin GB, Williams J, Tanksley SD. 1991. Proc Natl Acad Sci 88:2336-2340
Young ND, Zamir D, Ganal MW, Tanksley SD. 1988. Genetics 120:579-585