Functional Variation for DIMBOA Content in Maize

Butrón A1, Chen Y-C2, Rottinghaus GE2, Guill K3, McMullen MD3, 1 Misión Biológica de Galicia (CSIC), Apdo. 28, 36080 Pontevedra, Spain; 2Veterinary

Medical Diagnostic Laboratory, University of Missouri; 3 USDA-ARS and the University of Missouri, Curtis Hall 302, Columbia 65211 MO, USA

The main hydroxamic acid in maize (Zea mays L.) is 2-4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one (DIMBOA). DIMBOA confers resistance to leaf-feeding by several corn borers including the European corn borer. Most genes involved in the DIMBOA metabolic pathway (Bx1, Bx2, Bx3, Bx4, Bx5, Bx6, Bx7 and Bx8) are located on the short arm of chromosome 4, and QTLs involved in maize resistance to leaf-feeding by European corn borer have been localized to that region. However, the low resolution of QTL linkage mapping does not allow convincing proof that genetic variation at bx loci was responsible for the variability for resistance. Therefore, a sub-sample of the Nested Association Mapping (NAM) population (Yu et al. 2008, Genetics 178:539-551) was used to finely map QTLs for DIMBOA content. A model including eight markers (on chromosomes 1, 2, 4, 5, 6, and 8) explained approximately 34% of phenotypic variability within eight RIL families. The closest marker to the major QTL detected in chromosome 4.01 at 7.2 cM explained 15.8% of phenotypic variation across the eight families (Table 1).

The position of the largest QTL co-localizes with the majority of structural genes of the DIMBOA pathway. Therefore association analysis was used to determine if genetic variation at bx1 affects variation of DIMBOA content. General (GLM) and Mixed Linear Models (MLM) revealed associations between variability for DIMBOA content and sequence polymorphisms at bx1 locus. After adjusting for population structure, a model including two polymorphisms in Bx1 (X76713.5_620 and Bx1_1.9_143) explained 12% of phenotypic variation in a population of 282 diverse lines. This is close to the 15% effect predicted by the GLMSELECT analysis for the chromosome 4 QTL across 8 families of RILs. These results suggest that genetic variation for bx1 is likely behind the QTL detected in the region 4.01, but the specific causal polymorphisms responsible for that variation were not identified.

Table 1. Effects of the marker identified as significant in the region 4.01 for DIMBOA (ppm) content among eight RIL families (from crosses of B73 to CML322, CML52, IL14H, M37W, MS71, NC350, Oh43 and Tx303) with their corresponding probability levels.

Negative effects mean that the allele for higher amount came from B73

Marker Contig Parameter CML322 CML52 IL14H M37W MS71 NC350 Oh43 Tx303

PHM1184 155 Effectp

-489.9<0.001

-1062.0<0.001

300.00.020

325.10.001

435.5<0.001

-338.70.001

1042.9<0.001

47.00.660

0

0.2

0.4

0.6

0.8

1

1.2

82000 83000 84000 85000 86000 87000 88000 89000

Distance (bp)

r2

0

1

2

3

4

5

6

7

8

- Log P

Fig. 1 Association of DNA polymorphisms with DIMBOA content across genomic portions of genes Bx1 (82911-85155 bp) and Bx2 (87362-89517 bp). The positions correspond to BAC sequence AC200309.3. Blue diamonds indicate association with DIMBOA content. Level of statistical association for each SNP and INDEL is expressed as –Log P. Pink squares indicate r2 LD scores for all marker pairs involving X76713.5_620 and yellow triangles r2 LD scores for all marker pairs involving Bx1_1.9_143.

Damage made by borers at mid-Whorl stage when DIMBOA is still present

Indole-3-glycerole phosphate

BX1

Indole

BX2, BX3, BX4, BX5

DIBOA

BX8/BX9

DIBOA-glucoside

BX6

TRIBOA-glucoside

BX7

DIMBOA-glucoside

Fig 1. Enzymes and intermediates of DIMBOA-glc biosynthesis in maize as described by Jonczyk et al. (2008, Plant Physiology 146:1053-1063)