USDA-ARS and University of Missouri
Quantitative genetic analysis of loci controlling synthesis of maysin, a corn earworm resistance factor, in maize silks
--P.F. Byrne, M.D. McMullen, M.E. Snook, T. Musket, N.W. Widstrom, B.R. Wiseman

The antibiosis type of resistance to the corn earworm (CEW) Helicoverpa zea (Boddie) in maize is due primarily to the concentration of maysin and related flavone glycosides in the silk. Although there are a number of clues about the genetic control of CEW resistance and maysin synthesis (Byrne et al., MNL68:35, 1994; Widstrom and Snook, Plant Breeding 112:120-126, 1994; Wiseman and Bondari, J. Econ. Entomol. 85:293-298, 1992), understanding of the inheritance of maysin content is incomplete.

Because maysin synthesis occurs as a branch of the flavonoid metabolic pathway, our study sought to relate variation in maysin concentration to loci of that well-characterized pathway. One major section of the pathway, which controls the synthesis of C-glycosyl flavones (including maysin) and phlobaphenes (responsible for red cob and pericarp pigments), is regulated by the p1 locus (Styles and Ceska, Maydica 34:227-237, 1989). That locus also affects the silk-browning trait, whereby silks of some genotypes turn brown after wounding and others do not change color (Levings and Stuber, Genetics 69:491-198, 1971). The other section of the pathway leads to flavonols and 3-hydroxy anthocyanins and is regulated by the coordinate action of either r1 + c1 or b1 + pl1 in a tissue-specific manner (Coe et al., In G.F. Sprague and J.W. Dudley, Corn and Corn improvement, Am. Soc. Agron., 1988; Dooner et al., Ann. Rev. Genet. 189:136-141, 1991). Although the two parts of the pathway require common intermediates up to the flavanone branch point, each portion appears to be independently regulated, at least in pericarp tissue (Grotewold et al., Cell 76: 543-553, 1994; Styles and Ceska, Maydica 34:227-237, 1989).

Our study sought to identify and estimate the contribution of loci that affect maysin synthesis in the population (GT114 x GT119 )F2, derived from a cross of a high- by a low-maysin parent.

Maysin concentrations in silks of 285 F2 plants were determined by reversed-phase HPLC (Snook et al., J. Chromatogr. 477:439-447, 1989) at the USDA-ARS Phytochemical Research Unit, Athens, Georgia. RFLP genotypes at loci encoding flavonoid pathway enzymes or linked marker loci were determined for the same plants. Single-factor analysis of variance was first used to detect significant associations between maysin concentration and genotypic classes at individual RFLP loci, based on a comparison-wise error rate of 0.05. Then, significant loci from the individual locus analyses and significant digenic epistatic interaction terms were included in multiple-locus models. The "best" model was determined to be that which explained the greatest proportion of the phenotypic variance (i.e., had the highest R2 value) and in which all terms were significant at the 0.05 significance level.

The p1 locus had by far the largest effect on maysin concentration, accounting for 58% of the total phenotypic variance (Table 1). Significant loci unlinked to p1 included umc207 and umc113 (both closely linked to c1 and bz1), r1, and a1, which was marginally significant. We are uncertain which of the genes c1 or bz1 influenced maysin concentration. We plan to differentiate between these loci by testcrossing the parental lines to c1 and bz1 tester stocks and observing aleurone color development. We are not able to test the c1 locus directly because the parental lines were not polymorphic for the c1 probe with the eight enzymes tested.

The "best" multiple-locus model accounted for 71.1% of the phenotypic variance for maysin concentration. In addition to p1, the model included the c1 - bz1 region and epistatic interactions of p1 with c1 - bz1, umc166b , and r1. umc166b was scored as a second locus which appeared in the analysis of umc166a, a marker flanking a2 on chromosome 5. Our data place umc166b on chromosome 1 about 40 cM from p1 toward the distal end of the long arm, but we are uncertain what function that locus or a linked locus might have.

Table 1. Results of single-factor analysis of variance examining the association of maysin concentration with genotype classes at 16 flavonoid pathway loci or closely linked markers.
Marker (locus) Linkage group SignificanceÝ R2§ N Source of higher allele
Significant loci:          
umc185 (p1) 1 <0.001 0.580 284 GT114
npi286 (near p1) 1 <0.001 0.574 276 GT114
umc166bý 1 <0.001 0.148 283 GT114
umc207 (near c1,bz1) 9 <0.001 0.067 278 GT119
umc113 (near c1,bz1) 9 <0.001 0.063 284 GT119
umc182 (r1) 10 0.005 0.038 275 GT114
umc44 (near r1) 10 0.002 0.044 282 GT114
umc189 (a1) 3 0.025 0.026 278 GT114
umc63 (near a1) 3 0.039 0.023 285 GT114
Non-significant loci:          
umc181 (bz2) 1 0.402 0.007 282  
csu164 (near bz2) 1 0.337 0.008 285  
chi (chi1) 1 0.134 0.014 285  
umc84 (near chi1) 1 0.207 0.011 281  
umc198 (whp1) 2 0.552 0.007 175  
csu64 (near whp1) 2 0.419 0.006 273  
umc6 (near b1) 2 0.200 0.011 285  
umc34 (near b1) 2 0.146 0.014 283  
umc4 (near chi2) 2 0.604 0.004 284  
npi239 (near F3H¶) 2 0.800 0.002 284  
umc198 (c2) 4 0.211 0.017 185  
npi270 (near c2) 4 0.522 0.005 284  
umc166a (near a2) 5 0.251 0.010 284  
chi (chi3) 5 0.649 0.005 173  
npi409 (near chi3) 5 0.162 0.013 273  
umc21 (near sm1) 6 0.568 0.004 278  
csu13 (near in1) 7 0.742 0.002 284  
c1 9 non-polymorphic      
Ý F-test P-value.
§ Portion of the phenotypic variance accounted for by genotypic classes at a locus.
ý Approximately 40 cM to the right of p1.
¶ Locus encoding flavanone-3-hydroxylase (L. Taylor, personal communication).

Mean maysin concentrations for genotypic classes indicate a variety of types of gene action (data not shown). The effect of p1 is strikingly additive, while the GT114 allele at umc207 appears dominant for low maysin concentration. In their interactions with p1, both umc166b and r1 affect maysin concentration only when p1 is homozygous dominant for the high-maysin allele; however, they differ in that the effect of umc166b is significant in the heterozygote, and the effect of r1 is in the GT114 homozygous class.

The importance of p1 in our analyses was predicted, as the concentration of maysin and its analogs is known to be associated with the p1-controlled silk-browning trait (Byrne et al., MNL68:35, 1994), and the parents used in our study have contrasting silk-browning phenotypes. The two findings of this study we find most interesting are:

(1) The involvement of loci from the r1/c1 controlled portion of the pathway in the expression of a trait predominantly controlled by p1. Although the contribution of r1 to maysin concentration was small, that of the c1 - bz1 region was highly significant, accounting for an additional 9% of the phenotypic variation (as an individual locus and in interaction with p1) after p1 was already in the model. This contrasts with indications of independent regulation of the two parts of the pathway in pericarp tissue (Grotewold et al., Cell 76: 543-553, 1994; Styles and Ceska, Maydica 34:227-237, 1989). However, in coleoptile tissue, Styles and Ceska (Can. J. Genet. Cytol. 23:691-704, 1981) reported that p1 affected concentrations of some 3-glucosylated flavonols (in the r1/c1 portion), and that alleles at bz1 (also in the r1/c1 portion) affected flavone concentrations, although the significance of that difference was not clear.

(2) The importance of regulatory loci (p1, r1, possibly c1), as opposed to structural loci, in the expression of these chemical concentration traits. The only known structural loci having significant associations were a1 (which was marginally significant in the maysin single-factor analysis and eliminated in the multiple-locus model) and possibly bz1. One explanation for this result is that the parental lines may not have segregated for functionally distinct alleles at the structural loci evaluated, while they did segregate at p1 and r1 (based on our observations of cob and brace root color). Another, more speculative explanation for our results is that changes at regulatory loci are inherently more important than changes at structural loci in determining end-product concentration. Biochemical flux-enzyme theory offers support for this explanation; in complex pathways, substantial reductions in activity of any one structural enzyme are likely to have only small effects on flux through the system (Kacser and Burns, Genetics 97: 639-666, 1980).

Will regulatory loci be equally important in other populations and for other quantitative traits? We plan to test that hypothesis for maysin concentration and for CEW larval growth response by analyzing several additional populations, including one segregating for suspected functional differences at a structural locus (c2), and another segregating for an apparent dominant inhibitor allele at a regulatory locus (c1).

Please Note: Notes submitted to the Maize Genetics Cooperation Newsletter may be cited only with consent of the authors

Return to the MNL 69 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page