University of Arizona

Thoughts on future efforts for developing the maize genetics linkage map using RFLPs

--Tim Helentjaris

The development of molecular marker systems (such as isozymes, RFLPs, etc.) has given geneticists an exciting new tool and profoundly changed the way many of us now approach questions in both basic and applied genetic research. Those of us involved in the development of these tools over the last several years have been in discussion recently as to how this technology can best be advanced for the benefit of the general maize research community. Better correlation of these markers with previous marker types is one obvious direction and the Univ. of Mo. has been steadily working towards this goal with support from NSF. Better correlation of all of the available marker sets is also in progress utilizing the recombinant inbred lines developed by B. Burr.

Increased use of this technology is resulting in a rapid growth in the identification of the genomic positions of loci affecting plant phenotypes, particularly those that influence economically important traits. Meanwhile, interest has also mounted in utilizing this information to isolate the causative genes. A variation of chromosome walking has already been successfully used in arabidopsis to clone one such gene and these types of efforts will certainly continue in this species. However, one must recognize that the initial arabidopsis efforts have been much more challenging than originally anticipated and extrapolation of this approach into other species may be impractical. In maize, a tight positional linkage of 1cM would consist on average of greater than two million base pairs (5 X 106bp / 2300cM per genome), compared with a similar value for arabidopsis of about 140,000bp. The former represents an almost insurmountable distance to both traverse and search through for coding sequences. The presence of numerous repetitive elements in the maize genome will also severely complicate use of such a strategy.

The application of such approaches would require the development of a very high density marker map (approx. 1cM resolution) that is tightly correlated to the desired phenotypic loci. Before embarking on any such development, I have questioned whether this is the only practical strategy or if others should also be considered. Instead of using random PstI-generated genomic fragments as most of us have to date, I propose that it might be more productive in the long run to use randomly-selected cDNA clones as the source for marker sequences. Most of these are simple in their genomic organization, making them quite suitable for RFLP-mapping. The added benefit is that since they directly represent expressed genic sequences, every time one maps one of these sequences, one has also determined the genomic location of a gene which can be compared with the locations of loci affecting plant phenotypes. Where correlations are found, further efforts could in theory associate a cloned genic sequence with a phenotypic alteration. If this idea could be carried to the logical extreme, whereby one has mapped all of the expressed genic sequences and all of the loci that cause phenotypic alterations, this would represent a powerful tool for maize geneticists. Anyone interested in cloning a locus, whose genomic location could be established, could simply contact a stock center and request all of the expressed sequences within that region. Final confirmation to establish which of the candidate sequences actually represented the desired gene would necessitate the use of other strategies (i.e. DNA sequencing, anti-sense or complementation transformation) not based upon positional information. This confirmation would also be required of any chromosome walking strategy; what one gains here is not having to traverse a large segment of chromosome and subsequently searching through it for all of the expressed sequences.

Given such a scenario, the first question might be: is it practical? As this type of strategy is �processive� and not directed, it will require the development of both a �complete� and �normalized� cDNA library. I believe a relatively complete cDNA library can be prepared by using mRNA isolated from several tissues and subsequently combined. At least three normalization strategies have also been proposed that might be applicable in this case. Given some of the intrinsic properties of maize (very high polymorphism rates) and improvements in technology (such as PCR), it is already practical for a group to analyze 100-200 clones per week. By using 100 RIs or F2s as our mapping population, we could achieve an average resolution of less than half a centimorgan in the final map. This would represent an average resolution (enrichment) for any cDNA of about 5500-fold (i.e. 5456 �bins� are created in 100 RIs or F2s). I will use an estimate for the number of genes in a higher plant as approximately 25,000 (arabidopsis genome size of 0.7 X 108bp / an average-sized gene of 3000bp). At a rate of only 100 clones analyzed per week and 40 active weeks per year, one could have potentially mapped almost half of the genes in the maize genome in a four-year program. Pushing this only by a factor of two will mean that one could map a majority of the maize genes. There are a number of other considerations, such as duplicated chromosomal segments, nonpolymorphic sequences, non-poly-A mRNAs, etc., that one should also recognize, but I do not believe they significantly burden the overall concept.

E. Coe, D. Weber, and myself are currently proposing to the USDA as part of the Plant Genome Mapping Initiative that this strategy be considered for maize. The Univ. of Az. would be responsible for most of the molecular mapping of random cDNA clones by recombinational methods. Illinois State Univ. would use cytogenetic landmarks to correlate the genetic and physical maps. Coe and Chao at the Univ. of Mo. would attempt to improve our information on the genomic positions of previously mapped Naked Eye Polymorphisms. Additionally the Univ. of Mo. will try to dramatically expand this number to several thousand loci using interval mapping strategies and the wealth of mutants collected and maintained by the maize research community, such as the one produced at Columbia by M. G. Neuffer totaling 6000 members. Our intention would be to release all biologicals and information to the general community as soon as we felt their analysis was satisfactory.

While this is certainly an ambitious goal, in many ways it simply represents a logical extension of previous efforts which have been enhanced by considering new ideas and technologies. Even if we fail in the final goal of mapping enough genic sequences to allow a researcher to possess a reasonable chance of finding their desired sequence amongst them, we should still obtain a very high density marker map (several thousand loci) that would be very amenable to other uses including chromosome walking. Meanwhile we should be able to include additional studies into the basic screening process that will reveal much more detail about the structure of the maize genome, the conservation of expressed sequences within the Gramineae, the regulation of individual sequences in different maize tissues or under different environmental conditions, etc. We have also committed ourselves to making a major effort to map maize sequences homologous to those isolated from and identified in other organisms. For instance the mapping of maize sequences homologous to Ti-tagged arabidopsis loci could provide a tremendous source of cloned genic sequences that could be associated with similar maize mutant phenotypes. Our reason for publicizing this idea now is to gain comments from the greater maize research community as to its practicality and utility. We also believe that there are many points in this process where other individuals could participate, by supplying either clones or mutants for analysis that would benefit them directly. Accordingly I would appreciate suggestions and criticisms on any of these topics.

Return to the MNL 65 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage