10. Gene flow and population structure of Oryza glumaepatula distributed in the Amazon basin

M. Akimoto1, Y. Shimamoto1 and H. Morishima2

1. Fac. Agr., Hokkaido Univ.. Sapporo, 060 Japan

2. National Institute of Genetics. Mishima. 411 Japan

     Oryza glumaepatula is a diploid wild-rice species with AA genome and is distributed in the tropical region of Central and South America. One of the ecotypes of this species growing in the Amazon basin seems to have unique life-history traits under the environment, in which annual oscillation of water level is about 10m. Namely, at their certain growth stage their culms are frequently broken and their plant bodies released from the ground are floating on water. With rapid and vigorous development of adventitious roots and shoots from each node, they become a floating meadow and flow down the river by stream and wind (Akimoto et al. 1994, Rubin 1994).
    To learn about the allozyme variability and genetic structure of natural populations of 0. glumaepatula found in the Amazon basin, we examined allozyme variability at 30 isozyme loci of 16 enzymes for 34 natural populations collected during our trip in 1992-93. The distribution area of the populations was divided geographicaly into 5 regions and allozyme variability in each region was estimated in terms of Fixation index and Nei's genetic diversity.
    Along Rio Solimoes, main stream of the Amazon, proportion of polymorphic loci (P). mean number of alleles per locus (A) and "expected heterozygosity" as given by Hexp=1- S x2i (x,:allele frequency) gradually became higher from upper (Rio Solimoes-3) to lower region (Rio Solomoes-l)(Table 1). This suggests that gene flow proceeds from

Table 1. Summary of allozyme variation for 30 loci within 5 regions: proportion
of polymorphic loci (P). mean number ofalleles per locus (A), observed
heterozygosity (Hobs.), expected heterozygosity (Hexp.) and fixation index (F).
 

No. of populations P A Hobs. Hexp. F
Hobs. 3 0.20 1.27 0.002 0.059 0.957
Rio Negro-2 6 0.27 1.27 0.001 0.054 0.989
Rio Solimoes-1 11 0.67 1.77 0.006 0.119 0.953
Rio Solimoes-2 8 0.53 1.63 0.002 0.091 0.974
Rio Solimoes-3 6 0.23 1.23 0.002 0.009 0.802
Total 34 0.73 1.93 0.003 0.109 0.961
Hexp= 1- S x2i, where xI stands for allele frequency
Hobs =1- S x2ii, stands for homozygote frequency.
For instance, when the frequencies of AA. Aa and aa are
0.1. 0.7 and 0.2, 1- S x2i, =1-(0.452 +0.552 )=0.495;
S x2ii, =1-(0.12 +0.22 )=0.95.
F=1-h/2pq , Wright's fixation index: h=observed frequency of heterozygotes
[h=1- S x2ii, p+q=1]. upper to lower basin of the Amazon in one direction. In Rio Negro, the largest tributary of the Amazon, however, we could not find a clear difference between the upper (Rio Negro-2) and lower region (Rio Negro-2). Day and Davies (1986) reported that the water of Rio Negro is poor in nutrition and characterized by low pH as compared with other rivers, and is not favorable for plant growth.
    Observed heterozygosity as given by Hobs =1- Sum of x2ii(xiihomozygote frequency) was lower than "expected heterozygosity" and fixation index was near ' 1 ' in each region (Table 1). This suggests that 0. glumaepatula populations in the Amazon have developed a self-pollination system, which enables them to produce seeed under unstable conditions.
 
Table 2. Nei's genetic diversity calculated for 5 regions
No. of populations HT Hs dst gst
Rio Negro-1 3 0.059 0.040 0.018 0.310
Rio Negro-2 6 0.054 0.026 0.028 0.521
Rio Solimoes-1 11 0.119 0.071 0.048 0.403
Rio Solimoes-2 8 0.091 0.052 0.039 0.430
Rio Solimoes-3 6 0.009 0.008 0.001 0.109
Total 34 0.109 0.043 0.067 0.610
Inter region 5 0.118 0.068 0.050 0.421
Ht: average gene diversity for all populations.
Hs: within-population gene diversity.
Dst: Between-populations gene diversity (HTHs).
Gst: gene differentiation between populations (Dst/H-t). l00km
Fig. 1. Geographical distribution of 5 regions in the Amazon basin.


    In general, the annual populations of Asian wild rice 0. rufipogon are predominantly selfed, and their inter-populational gene diversity (Dst) is higher than intra-populational gene diversity (Hs). 0. glumaepatula populations do not show such a tendency (Table 2). Probably, the frequency of gene exchange among populations is higher in 0. glumaepatula than in Asian 0. rufipogon.
    Although we used a number of populations collected from a large area of nearly 20000km^2, allozymes were not so variable as compared with those in Asian 0. rufipogon (Barbier 1989). This is probably because frequent gene flow proceeds in one direction and self-pollination hampers to develop allozyme variability. (Pascal 1989). This is probably because frequent gene flow proceeds in one direction and self-pollination hampers to develop allozyme variability.

References Akimoto, M., M. Ohara, Y. Shimamoto and H. Morishima, 1994. Genetic structure of natural populations of
        Oryza glumaepatula distributed in the Amazon basin. RGN 11: 74-76.
Barbier, P., 1989. Genetic variation and ecotypic differentiation in the wild rice species Oryza rufipogon.
        II. Influence of the mating system and life-history traits on the genetic structure of populations.
        Jpn. J. Genet. 64:273-285.
Day, J. A. and B. R. Davies, 1986. The Amazon river system. In The ecology of river systems, p.289-318., Dr.
        W. Junk Publishers, Dordrecht, The Netherlands.
Rubim, M. A. L., 1994. A case study on life-history of wild rice: From germination to emergence of inflorescence.
        In Investigation of plant genetic resources in the Amazon basin with the emphasis on the genus 0ryza.
        p.38-42, Special Rep. from Nat. Inst. Genet., Japan.