Fig. 1. The location of marker loci associated with the segregation of atom % 15N excess in an F`2` population from Palawan/IR 42.
International Rice Research Institute, P. 0. Box 933, Manila, Philippines
To understand the genetic nature of rice varietal ability to stimulate N`2` fixation in the rhizosphere, an F`2` population (231 individuals) derived from a cross between Palawan and IR42 was used in an 15 N-dilution pot experiment in the greenhouse from May to September 1993. The 15N labeled soil was incubated under flooded condition in a concrete tank until the stabilization of atom % 15N excess (Rarivonson and Ladha 1992) was achieved (0.292+/-0.003). Normal segregation of the atom % 15N excess was obtained in the F`2` population with an average of 0.219 with 10% of plants below IR42 (0.188) and 15% of plants above Palawan (0.248). DNA was extracted from plants developed from tillers for RFLP analysis. Ninety restriction fragment length polymorphic marker loci (RFLP marker loci) were scored in the F`2` progenies. Individual marker loci were tested for linkage to QTLs affecting atom % N excess by orthogonal contrasts between the means of atom % 15N excess within three genotypic marker classes (Edwards et al. 1987) and the marker loci were considered to be significantly associated with a QTL when the F-test for a contrast exceeded the F-value with a probability of less than 0.01. All possible digenic epistatic interactions were tested between pairs of significant markers detected. Multiple regression analysis using a stepwise procedure was performed to develop the predicted model (Martin et al. 1989).
Two marker loci, RZ588 and RG944 located on chromosomes 6 and 3 (Fig. 1) were significantly (P<0.01) associated with the segregation of atom % 15N excess. Two other marker loci, RZ409 and RG213 linked with RG944 and RZ588, respectively, were also detected at the 5% significant level (Table 1). Only additive gene action was detected for each signifcant marker. A significant digenic epistatic interaction (Add. by Add.) was found between RG944 and RZ588 (Table 1). The group of lines containing marker alleles inherited from IR42 at these 4 loci was superior to the group containing Palawan alleles (data not shown). A multiple regression model including the four marker loci (additive
Fig. 1. The location of marker loci associated with the segregation of atom %
15N excess in an F`2` population from Palawan/IR 42.
Table 1. Orthogonal contrast for means of atom % 15N excess within
marker genotypic classes. Only the significant marker loci, and
the significant interaction are shown
_______________________________________________________________________________
Source df F-value
______________________________________________
RZ409 RZ588 RG944 RG213
_______________________________________________________________________________
Linear (Add.) 1 4.82* 8.19** 8.26** 6.21*
Quadratic(Dom.) 1 0.30 0.19 0.37 0.02
Interaction
(RG944 by RZ588)
Add. by Add. 1 2.72**
_______________________________________________________________________________
*,and ** represent significance at 0.05, 0.01 and 0. 001 levels, respecively.
Table 2. Multiple regression analysis for prediction of atom % 15N excess
developed by using stepwise regression procedure. only marker loci
which were significant (P<0.05) after being added to the model
remained in the model
_______________________________________________________________________________
Parameter Effect Esimate SE T-statistic Prob > t
_______________________________________________________________________________
Atom % 15N excess (R=0.27)
Constant 0.217 0.001 156.40 0.000
RZ588 Add. -0.024 0.002 -2.075 0.039
RG944 Add. -0.005 0.002 -3.083 0.002
_______________________________________________________________________________
effects plus the interaction effect between RG944 and RZ588) account or of
the total phenotypic variation. A model including only the two most
significant marker loci, RZ588 and RG944, accounted for 27% of phenotypic
variation. The interaction term contributed very little to the model. The
individual marker loci both explained about 19% of the total phenotypic
variation for atom % 15N excess. Our results indicate that the varietal
ability of rice to enhance N2 fixation in the rhizosphere is controlled by
mutiple genes. This is consistent with previous report by Iyama et al.
(1983). The identification of putative QTLs underlying the ability to
stimulate N`2` fixation provides the first real evidence of the presence of
genetic factors which interact with diazotrophs in the rice rhizosphere and
lays the foundation for increasing N`2` fixation in rice fields through genetic
manipulation of rice plants.References
Edwards, M., W. Stuber and J. F. Wendel, 1987. Molecular-marker-facilitated investigations of quantitative trait loci in maize. I. Number, genome and types of gene action. Genetics III: 113-125.
Iyama, S., Y. Sano and T. Fujii, 1983. Diallel analysis of nitrogen fixation in the rhizosphere of rice. Plant Sciences Letters 30: 427-135.
Martin, B., J. Nienhuis, G. King and A. Schaefer, 1989. Restriction fragment length polymorphisms associated with water use efficiency in tomato. Science 243: 1725-1728.
Rarivonson, G and J. K. Ladha, 1992. Isotype 15N enrichment of soil ammonium N as a reference to estimate N`2` fixation by stem-and-root inoculated S. rostrate. Biol. Fertil. Soil 13: 74-78.