Russell Malberg suggested that the community assemble a collection of
"kanamycin resistant transformants with constitutive reporter alleles"
(e.g. 35S-GUS fusions). This collection would consist of strains having
such markers distributes across the genome. I agree that this would prove
most useful to genetic manipulations. We have been using such a tool in
our chromosome walking experiments aimed at the det1 (de-etiolated 1) locus.
I will summarize the strategy below.
In any walking experiment, one needs to establish orientation toward the
gene sought, and later determine an interval within which the gene resides.
Both of these objectives may be achieved through the mapping of RFLP markers
that have been identified in the course of the walk, and then determining
the genotype of mutant plants in the F2 of a cross between your mutant of
interest and another Arabidopsis ecotype. As one walks toward the gene,
the RFLP profiles will always display the RFLP pattern of the mutant parent,
as the gene is approached. Walks away from the gene can show the profile of
either parent. At higher resolution, one can map the RFLP pattern of the
rare plants in such an F2, that have undergone a crossover between the gene
of interest and the start point of your walk. These plants are valuable since
they contain a junction between the DNAs of both the mutant parent and the
other parent in the cross. RFLP analysis across this interval can show the
genotype at each of the RFLP sites examined, and thus help to delineate a
region in the walk that can contain the gene sought. The resolution of
this approach is a function of two factors:
-having sufficient plants that have undergone recombination between
the start point of the walk, and the gene of interest, and
-having the RFLP markers in this interval that permit genotypic
Limitation of either of these resources will hinder delimiting the region
of the gene, using this method. Obtaining the RFLP markers is not discussed
here. However identifying the rare recombinants that have undergone
recombination within a centimorgan of the gene is fascilitated through using
plant lines carrying insertions such as those mentioned by Russell Malberg.
We are using this method to identify plants recombinant between one of our
genes of interest (de-etiolated-1) and a closely linked T-DNA insertion line
(agamous). The F2 from a cross of these lines will segregate det1/kan-res
progeny, which necessarily contain at least one recombinant chromosome
between each of these lines. The value of this approach is that we can plate
the F2 out on selective medium, killing 25% of the progeny, and then screen
out det1 segregants. It is necessary to identify polymorphisms between the
det1 parent (Col-O) and the insertion line (Ws-O). We have had some success
at this, but it is early to generalize. The bottom line is that we have a
means for identifying dozens (if not more) crossovers between the start
point of our walk and DET1. In comparison with screening random det1
individuals from an F2, I estimate that this approach increases the
yield of desired recombinants about fourfold (i.e. hundreds of fewer DNA
preps). This approach relies on the dominant selectable kan-res marker
carried in many transformation vectors as well as on the T-DNA insertions.
In this example, the selectable marker was about 24 map units from the gene
of interest. Closer markers would of course be even more useful.
As people seem to be walking all over the map, or at least plan to,
a resource of Arabidopsis lines carrying single insertions of this kind
would prove extremely useful. Ideally such a collection would contain
the popular mapping ecotypes (Col-O, La-O, Nd-O), obviating the need
to identify new RFLPs, at least for the standard RFLP probes from
E. Meyerowitz' and H. Goodman's laboratories.
(DELANEY at SALK)
(Alan Pepper has contributed to the work discussed here. Both myself
(TD) and AP are postdoc's in Dr. Joanne Chory's laboratory at the