Modifications of cellulose synthase confer resistance to isoxaben and thiazolidinone herbicides in Arabidopsis Ixr1 mutants

Abstract

In many higher plants, cellulose synthesis is inhibited by isoxaben and thiazolidinone herbicides such as 5-tert-butyl-carbamoyloxy-3-(3-trifluromethyl) phenyl-4-thiazolidinone. Semidominant mutations at the IXR1 and IXR2 loci of Arabidopsis confer isoxaben and thiazolidinone resistance. Isolation of the IXR1 gene by map-based cloning revealed that it encodes the AtCESA3 isoform of cellulose synthase. The two known mutant alleles contain point mutations that replace glycine 998 with aspartic acid, and threonine 942 with isoleucine, respectively. The mutations occur in a highly conserved region of the enzyme near the carboxyl terminus that is well separated from the proposed active site. Although the IXR1 gene is expressed in the same cells as the structurally related RSW1 (AtCESA1) cellulose synthase gene, these two CESA genes are not functionally redundant.

Figure 1

Map-based cloning of the IXR1 gene. (A) Representation of a region of Arabidopsis chromosome 5 showing a megabase-scale (Upper), the position of RFLP markers g3715 and m217, SSLP markers CIW13, 14, 15, and SSLP marker nga158. Chromosome 5 P1 clones (http://www.kazusa.or.jp/arabi/) containing the SSLP markers are represented by white boxes. The number of recombinant chromosomes (meiotic breakpoints) in a total of 2,112 examined chromosomes, found for each marker and their calculated genetic distance to the IXR1 locus [in centimorgans (cM)], is given. (B) The region containing the flanking SSLP-markers CIW15 and nga158, showing the position and the number of recombinants found for CAPS (CIW17, PAI2) and SSLP (CIW16, 18) markers and the nonoverlapping parts of P1 and TAC clones spanning the region. (C) An 8-kb segment of TAC clone K2A11 (nucleotides 17, 500–9, 500) showing the intron/exon structure of the IXR1 gene (note crossover of the dotted lines). The start codon and the point mutations in the ixr1–1 and ixr1–2 alleles in the last exon and the predicted amino acid exchanges in the mutant gene products are indicated.

Figure 2

Representation of the positions and sizes of various cosmid clones (C-S) tested for ability to complement an ixr1 mutation. The genomic DNA contained in TAC clones K2A11 and K18I23 (Fig. 1B) is represented as bold black and gray lines, respectively. Predicted and known genes (white boxes) were annotated by Kaneko et al. (23). HinDIII restriction sites (vertical thin lines) and the position of hybridization probes X, Y, and Z (small black boxes) used for cosmid library screening are shown.

Figure 3

Transformation of the ixr1–2 mutant with a wild-type IXR1 gene results in isoxaben sensitive progeny. (A) Segregating T₂ progeny of an ixr1–2 mutant transformed with cosmid I (Fig. 2). Approximately 75% of the progeny, representing heterozygous and homozygous ixr1–2 transformants, are herbicide-sensitive when grown on vertical MS plates containing 600 nM isoxaben. (B) Progeny of an ixr1–2 mutant transformed with cosmid F were 100% isoxaben-resistant.

Figure 4

Northern analysis of cellulose synthase genes (IXR1, RSW1, AtCESA2, and IRX3) in (1) stems, (2) rosette leaves, (3) cauline leaves, and (4) flowers of wild-type and ixr1–1 mutant.

Figure 5

Northern analysis of cellulose synthase genes (IXR1, RSW1, and AtCESA2) in shoots (S) and roots (R) of wild-type, ixr1–1, and ixr1–2 mutants grown in hydroponic culture and treated with (+) or without (−) 0.5 μM isoxaben for 2 days. No signal was detectable for IRX3.

Figure 6

Medial longitudinal sections of roots from transgenic plants containing the RSW1:GUS promoter fusion (A, B), IXR1:GUS promoter fusion (C, D) or nontransformed (E, F). A, C, and E show sections made just above the boundary of the root meristem and the elongation zone. Green fluorescence indicates GUS activity.

Figure 7

A heteromeric model for the structure of cellulose synthase. In this model, the complex is composed of 36 CesA subunits that are distinguished by the number of subunit binding sites. The α subunits have binding sites for three CesA subunits (one α and two β), whereas the β subunits have only two binding sites.