Changes in Auxin Response from Mutations in an AUX/IAA Gene

See allHide authors and affiliations

Science  27 Feb 1998:
Vol. 279, Issue 5355, pp. 1371-1373
DOI: 10.1126/science.279.5355.1371


Transcription of the AUX/IAA family of genes is rapidly induced by the plant hormone auxin, but evidence thatAUX/IAA genes mediate further responses to auxin has been elusive. Changes in diverse auxin responses result from mutations in the Arabidopsis AXR3 gene. AXR3 was shown to be a member of the AUX/IAA family, providing direct evidence thatAUX/IAA genes are central in auxin signaling. Molecular characterization of axr3 gain-of-function and loss-of-function mutations established the functional importance of domains conserved among AUX/IAA proteins.

Plant growth and development are regulated by the hormone auxin. Mutational analysis inArabidopsis has identified genes important in auxin action (1). One such gene, AXR3, is defined by three semidominant mutations that result in increased auxin responses (2). Here, we describe intragenic suppressors of theaxr3-1 phenotype and the sequence of the AXR3gene. AXR3 encodes a member of the auxin-inducible AUX/IAA family of short-lived nuclear proteins (3-5), demonstrating that AUX/IAA genes are central in auxin signaling.

Seeds homozygous for the axr3-1 mutation and a second marker, gl1 (6), were subjected to mutagenesis, and the resultant M2 population was screened for revertants on the basis of shoot morphology (7). Five independent revertant lines were recovered, all of which contained thegl1 mutation, indicating that they were not wild-type contaminants. When the revertants were crossed with the wild type, noaxr3-1 plants segregated in the resulting F2populations. When the revertants were crossed with axr3-1mutants, the F1 plants were phenotypically similar toaxr3-1 heterozygotes, and axr3-1 and revertant plants segregated in the F2 populations in a ratio not significantly different from 3:1 (8). Because the data suggested that the revertant phenotypes resulted fromaxr3-1 intragenic mutations, the corresponding lines were named axr3-1R1 to axr3-1R5.

None of the revertants is completely wild type (Fig.1). With primary root length as a measure of allelic strength, the rank order of reversion from strongest to weakest was axr3-1R4 > axr3-1R3 > axr3-1R2 = axr3-1R5 > axr3-1R1 (Fig. 1B). With the exception ofaxr3-1R4, all the revertants have agravitropic roots. The roots of axr3-1R4 plants grow downward but are abnormally straight, indicating that the root wave response is defective (9). Shoot phenotypes such as leaf curling persist only inaxr3-1R1 and axr3-1R5.

Figure 1

The phenotype of axr3-1 intragenic revertants. (A) Ten-day-old plants (18) homozygous for axr3-1 intragenic revertant mutations. From left to right, the plants are AXR3,axr3-1R4, axr3-1R3, axr3-1R2,axr3-1R5, axr3-1R1, and axr3-1. (B) Mean root lengths for each genotype after growth for 7 days under the same conditions. Data are means ± SEM (n = 15).

The AXR3 gene maps to chromosome 1, ∼1.5 centimorgans distal to AXR1 (Fig. 2) (2). The flanking markers dis1, ga4, and cer1 (6) in the Landsberg genetic background were used to identify lines with recombinational break points flankingAXR3. With the use of these lines, AXR3 was mapped with respect to DNA polymorphisms in the region (10).AXR3 maps immediately distal to the genomic clone 0846A, which was then used to probe bacterial artificial chromosome (BAC) and yeast artificial chromosome (YAC) libraries. The ends of the 0846A-positive BACs and YACs were used to probe the existing 0846A-positive BACs and YACs, as well as the BAC and YAC libraries (10), and contiguous clones extending distal to 0846A were collected (Fig. 2). YAC and BAC end clones were tested for polymorphisms between the Landsberg and Columbia genetic backgrounds. The distal end of BAC IGF20G19 was shown to include a polymorphic Tsp5091 site that maps distal to AXR3, delimiting a 30-kb region that includes AXR3 (Fig. 2).

Figure 2

Map-based cloning of theAXR3 gene. Six clones encompassing theAXR3 region; two YAC clones (prefixed y) and four BAC clones (prefixed IGF), are shown (open boxes). The position of theAXR3 locus is shown relative to the visible markers axr1, ga4, dis1, andcer1 (2, 6) and to DNA markers (ATEAT1, 0846A, IGF20G19 distal end, and an IGF21M11 subclone) (10). Of ∼1760 F2 chromosomes, 17 showed recombination between AXR3 and ATEAT1; only one of these 17 showed recombination between AXR3and 0846A. Of ∼750 F2 chromosomes, two showed recombination between AXR3 and the IGF21M11 subclone; only one of these two showed recombination betweenAXR3 and the IGF20G19 distal end. The delimited region includes two members of the AUX/IAA gene family (11), IAA17 andIAA3, which are shown with arrows indicating the direction of their transcription. cM, centimorgan.

The BAC IGF19P19 extends 16 kb into the proximal portion of the delimited region and has been sequenced (11) as part of theArabidopsis genome initiative. This 16-kb sequence includes two members of the AUX/IAA gene family, IAA3(4) and IAA17 (12), both of which were then subcloned and sequenced from axr3-1 DNA (13). The IAA3 gene of axr3-1 was identical to that of the wild type, whereas IAA17 contained a single nucleotide difference, predicted to convert the proline at position 88 to leucine (Fig. 3).

Figure 3

The IAA17/AXR3 protein andaxr3 mutations. The positions of the nuclear localization sequences (NLS) (open boxes), the βαα dimerization and DNA binding domain (shaded boxes), exon boundaries (*), and conserved domains I to IV are shown (4). The domain consensus sequences are shown with uppercase letters indicating complete conservation, lowercase letters indicating partial conservation, and dots indicating nonconserved residues. The amino acid sequences for domains I to IV of IAA17/AXR3 are shown under the consensus (12). Amino acid substitutions in theaxr3 alleles are highlighted (14). Amino acid insertions resulting from the splice site mutations are shown after the relevant asterisk (14). Abbreviations for the amino acid residues are as follows: A, Ala; C, Cys; D, Asp; E, Glu; F, Phe; G, Gly; I, Ile; K, Lys; L, Leu; M, Met; N, Asn; P, Pro; Q, Gln; R, Arg; S, Ser; T, Thr; V, Val; W, Trp; and Y, Tyr.

To confirm that IAA17 is AXR3, we amplified fragments encompassing the gene by the polymerase chain reaction (PCR) and sequenced other axr3 alleles (14). Theaxr3-4 and axr3-1 mutations are identical, and the axr3-3 mutation affects the adjacent valine (Fig. 3). The five revertant alleles retain the axr3-1 mutation and include an additional point mutation that in three alleles (axr3-1R1, axr3-1R2, and axr3-1R3) results in a single amino acid change, and in two alleles (axr3-1R4 and axr3-1R5) affects a splice site. RNA was extracted from axr3-1R4 and axr3-1R5plants and subjected to reverse transcription and PCR to amplify fragments corresponding to the affected exon boundaries (14). Sequence analysis of the PCR products revealed that, in both instances, splicing occurs to a cryptic site within the affected intron. For axr3-1R5, such splicing results in the insertion of 33 nucleotides and, hence, 11 amino acids (Fig. 3). Foraxr3-1R4, four nucleotides are inserted, resulting in a shift in the reading frame and deletion of half of domain IV of the protein (Fig. 3). Because AXR3 and IAA17 map to the same 30-kb region and eight independent axr3 alleles all have mutations in IAA17, we conclude that IAA17and AXR3 are the same gene.

Members of the auxin-inducible AUX/IAA gene family have been isolated from several species (3, 4). They vary with respect to tissue specificity of expression, the kinetics of auxin induction, and the auxin dose-response relation (4). They encode short-lived nuclear proteins that contain four highly conserved domains (Fig. 3). Domain III shows similarity to βαα dimerization and DNA binding domains (4, 5). AUX/IAA proteins form homodimers and heterodimers through domains III and IV (12). Furthermore, these proteins interact with the auxin response factors (ARFs) ARF1 and IAA24 (12, 15), which bind to the auxin response element present in the promoters of many auxin-inducible genes (15). DNA binding by ARFs is mediated through the NH2-terminus, whereas interaction with AUX/IAA proteins is mediated by the COOH-terminus, which shows sequence similarity to domains III and IV of AUX/IAA proteins.

Semidominant mutations in the IAA17/AXR3 gene result in a wide range of auxin-related phenotypes, consistent with an increase in the amplitude of auxin responses and including ectopic expression from the SAUR-AC1 promoter (2). This promoter contains the auxin response element to which ARFs bind (16). The ectopic SAUR-AC1 expression thus supports the hypothesis that AUX/IAA proteins interact with ARFs to regulate gene expression directly.

The axr3 mutations affect the four conserved AUX/IAA protein domains, confirming their functional significance. The tight clustering of the semidominant mutations contrasts with the scattered distribution of the intragenic revertant mutations, indicating that the revertant mutations cause loss of or a reduction in gene function, negating the gain of function conferred by the axr3-1 mutation. Proteins containing the semidominant mutations may act in a dominant negative manner—for example, by forming nonfunctional complexes with AUX/IAA or ARF proteins. However, the increased auxin responses of axr3plants suggest that the mutations are hypermorphic, resulting, for example, in increased IAA17/AXR3 stability.

Arabidopsis contains at least 25 AUX/IAA genes (12). Presumably, each encoded protein is capable of interacting with other family members and with ARFs. Furthermore, each may bind DNA directly as heterodimers or homodimers through the βαα domain. This complex web of interactions links auxin to its downstream responses. The molecular characterization of the diverseaxr3 mutations offers an opportunity to understand better the molecular basis of AUX/IAA-mediated auxin signaling.

  • * To whom correspondence should be addressed. E-mail: hmol1{at}


View Abstract

Navigate This Article