Comparison of Variation at Molecular Markers And Quantitative Traits In Alfalfa.
Sergey Ivashuta, Kazuhiro Uchiyama, Mitsuru Gau and Sachiko Isobe
Hokkaido National Agricultural Experimental Station, Toyohira-ku, Sapporo 062-8555, Japan
RAPD analysis combined with AMOVA (Analysis of Molecular Variance) were used for assessment of genetic variation within and between two alfalfa populations of contrasting freezing tolerance and fall dormancy (var.Moapa 67 and var.Rambler). The analysis revealed high level of variation within populations and relatively low level of differentiation between populations. Of the total genetic diversity about 80% was attributed to the individual differences within a populations and less than 20% - to the between-populations variance components. Meanwhile, there were notable differentiation of these populations in the freezing tolerance and fall dormancy as revealed by nested ANOVA. A comparison showed that there is little congruence between patterns of variation at molecular marker loci and in quantitative traits under study.
Generally, the similar level of RAPD polymorphism was revealed between any two alfalfa plants drawn from same population (similar freezing tolerance and fall dormancy) or from different populations (differing significantly in freezing tolerance and fall dormancy). We did not find population-specific RAPD markers and only several bands were found that had significant differences in frequency between two populations. Another class of molecular markers, namely AFLP, also showed high level of genomic DNA polymorphism between any two alfalfa plants. On the contrary, when ds cDNA was used as a template for AFLP reaction level of "polymorphism" was significantly lower. cDNA was synthesized from mRNA isolated from genotypes of contrasting freezing tolerance grown under normal and stress conditions (cold treatment). The fragment was counted as "polymorphic" if it present in at least one mRNA samples (non-hardened or hardened plants) of one genotype and absented in both mRNA samples of another genotype. In our preliminary experiment, on average less than 5% of the total number of bands were different between plants from different populations. This is several times lower as compared with level of polymorphism that was revealed for genomic DNA. In addition, it is evident that only some of the "polymorphic" bands displayed in the cDNA-AFLP experiment are truly polymorphic on genomic level (rather than differentially expressed genes) as could be proven by Southern hybridization.
We suppose that DNA polymorphism of the expressed sequences detected with cDNA-based markers (used as a probe or after converting to DNA-based PCR markers) can more clear reflect the pattern of variation in quantitative traits in outcrossing species. These markers would be also useful for genetic map construction and MAS.