SIR1, an Upstream Component in Auxin Signaling Identified by Chemical Genetics

See allHide authors and affiliations

Science  22 Aug 2003:
Vol. 301, Issue 5636, pp. 1107-1110
DOI: 10.1126/science.1084161


Auxin is a plant hormone that regulates many aspects of plant growth and development. We used a chemical genetics approach to identify SIR1, a regulator of many auxin-inducible genes. The sir1 mutant was resistant to sirtinol, a small molecule that activates many auxin-inducible genes and promotes auxin-related developmental phenotypes. SIR1 is predicted to encode a protein composed of a ubiquitin-activating enzyme E1–like domain and a Rhodanese-like domain homologous to that of prolyl isomerase. We suggest a molecular context for how the auxin signal is propagated to exert its biological effects.

Auxin has been implicated in almost every aspect of plant growth and development. Molecular genetics studies on auxin-resistant Arabidopsis mutants (1, 2) and biochemical analyses of early auxin-inducible genes (3, 4) have elucidated many aspects of auxin signaling. The current model of auxin signaling suggests that negative regulators such as the auxin/indole-3-acetic acid (Aux/IAA) proteins are targeted for degradation in an auxin-dependent manner through the ubiquitin-related protein degradation machinery, thereby derepressing a network of genes to guide proper growth and development (5).

Analysis of auxin signaling by classical genetics is complicated by auxin polar transport, a process whereby an auxin-concentration gradient is maintained among neighboring cells (6). This process limits the accessibility of certain cells and tissues to exogenous auxin, and therefore certain auxin signaling components might well be missed from genetic screens for mutants that are resistant to exogenous auxin. Moreover, auxin polar transport and auxin signaling are well-coordinated, interdependent processes, making it difficult to isolate and specifically analyze either process. Here we identify a tool to modulate auxin signaling exclusively, without affecting auxin polar transport, in order to analyze the molecular mechanisms of auxin signaling.

We used a chemical genetics (7) approach to identify an auxin signal transduction component in Arabidopsis. Sirtinol (Fig. 1A), an inhibitor of the Sirtuin family of nicotinamide adenine dinucleotide (NAD)–dependent deacetylases in Saccharomyces cerevisiae, affects root and vascular tissue development in Arabidopsis (8). Sirtinol specifically activates many auxin-inducible genes, thereby promoting several auxin-related developmental phenotypes. With this approach, we identified the gene SIR1 as a key regulator of many auxininducible genes.

Fig. 1.

Regulation of auxin-inducible genes by sirtinol. (A) The chemical structure of sirtinol. (B) Activation of the auxin-reporter gene DR5-GUS. Seedlings grown on half-strength (0.5×) MS (left) and on 25 μM sirtinol (right) for 5 days were visualized by staining with X-glucaronide for β-glucaronidase activity. (C) Degradation of the AXR3-NT-GUS fusion induced by sirtinol treatment. Expression of AXR3-NT-GUS was induced by heat shock at 37°C for 2 hours. Twenty min after the heat-shock, seedlings were further treated for 20 min with either water (the left two seedlings) or 25 μM sirtinol (the right two seedlings).

We screened for compounds that could alter the expression pattern and/or levels of an auxin reporter line, DR5-GUS (9). Sirtinol caused up-regulation and ectopic expression of the auxin reporter gene (10) that was induced throughout the plant, with the highest expression levels along the vascular tissues (Fig. 1B). We next performed a microarray analysis using the Arabidopsis “whole genome” chip (Affymetrix) to determine whether sirtinol activated the expression of auxin-inducible genes at the transcriptional level (10). Of the 23,000 genes on the chip, about 16,000 genes gave signals that were significantly above background levels in all samples. Sirtinol treatment induced 138 genes that were expressed at least 2.5 times as high in the sirtinol-treated samples as in the water-treated seedlings. Many of the known auxin-inducible genes, including Aux/IAA genes, were represented among the 138 sirtinol-induced genes (Table 1). The gene expression profile of sirtinol-treated plants was similar to that of auxin-treated plants: More than 65% of the genes induced by sirtinol were also induced by auxin (table S1).

Table 1.

Microarray analysis of gene expression induced by sirtinol and auxin treatment. a, absent, indicating that the microarray signal is below background levels; nc, no change in expression level. The Gene Access no. is the GenBank accession number. The numbers in the sirtinol and auxin columns refer to the fold-change from untreated samples.

Gene access no. Gene product Sirtinol Auxin
At2g23170 GH3-like protein 200.0 1096.0
At3g03840 Putative auxin-induced protein similar to SAUR 36.6 2697.0
At1g59500 Auxin-regulated protein GH3 33.1 99.5
At4g32280 Similar to Aux/IAA proteins 27.1 60.3
At3g03830 Putative auxin-induced protein similar to SAUR 27.1 66.7
At3g58190 Putative protein 24.5 99.5
At4g38850 Small auxin up RNA (SAUR-AC1) protein 20.0 44.7
At5g03350 Putative protein 18.2 a
At1g29440 Auxin-induced protein 16.4 33.1
At2g15490 Putative glucosyltransferase 16.4 2.7
At3g56710 SigA binding protein 14.9 1.6
At1g29490 Unknown protein 14.9 493.0
At1g29460 Auxin-induced protein 13.5 27.1
At2g36770 Putative glucosyl transferase 13.5 2.0
At1g29500 Auxin-induced protein 12.2 44.7
At1g29450 Auxin-induced protein 12.2 60.0
At4g14560 Auxin-responsive protein IAA1 11.0 74.0
At5g47370 Homeobox-leucine zipper protein-like 11.0 30.0
At4g36110 Putative auxin-induced protein 11.0 270.0
At1g17170 Putative glutathione transferase 11.0 2.7
At5g18060 Auxin-induced protein-like 10.0 40.5
At3g56400 DNA-binding protein-like 10.0 nc
At3g50340 Putative protein 10.0 55.0
At1g29510 Auxin-induced protein 10.0 16.4
At2g21200 Putative auxin-regulated protein 10.0 7.4
At4g16515 Expressed protein 9.0 41.0
At1g05680 Putative indole-3-acetate beta-glucosyltransferase 9.0 a
At3g15540 Early auxin-induced protein IAA19 7.4 67.0
At1g29430 Auxin-induced protein 7.4 30.0
At1g26770 Expansin 10 7.4 2.0
At2g29490 Putative glutathione S-transferase 7.4 2.2
At5g54510 Auxin-responsive-like protein 6.7 33.0
At5g52900 Unknown protein 6.0 20.0
At5g02760 Protein phosphatase 6.0 37.0

The rapid degradation of negative regulators, such as Aux/IAA proteins induced by auxin to activate the expression of many auxin-inducible genes, is a hallmark of auxin signaling (11, 12). Using IAA17/AXR3-GUS transgenic Arabidopsis plants (1), we found that sirtinol treatment, like auxin treatment, led to a rapid degradation of the AXR3-GUS fusion protein (Fig. 1C), indicating that sirtinol-induced gene expression could also arise from regulated-protein degradation.

We next asked whether sirtinol promoted auxin-related developmental changes. Sirtinol inhibits primary root elongation and hypocotyl development in light-grown seedlings (8) (Fig. 2A). Seedlings grown in total darkness on sirtinol-containing media had short hypocotyls and lacked an apical hook, which is characteristic of auxin-treated or auxin-overproducing dark-grown seedlings (Fig. 2B) (13, 14). Sirtinol also stimulated adventitious root formation (Fig. 2C). Defoliation was often observed at the bottom of the hypocotyls, where adventitious root growth was initiated (Fig. 2C). Transferring sirtinol-treated plants to unsupplemented Murashige and Skoog (MS) medium accelerated the formation of adventitious roots (Fig. 2C). Defoliation and adventitious root formation are phenotypes commonly associated with auxin-overproducing mutants in Arabidopsis (13), and exogenous auxin treatment is known to lead to the inhibition of primary root elongation and promotion of adventitious root development (15). The similarities between the phenotypes of sirtinol-treated plants and auxin-treated or auxin-overproducing plants are consistent with our observation that sirtinol treatment activates auxin-inducible genes and thus amplifies auxin signal output.

Fig. 2.

The effects of sirtinol treatment on Arabidopsis growth and development. (A) Wild-type (WT) Arabidopsis grown on 0.5× MS media (left) and 25 μM sirtinol (right) for 5 days. (B) Arabidopsis seedlings grown in total darkness on MS media (left), 10 μM sirtinol (middle), and 1 μM 2,4-dichlorophenoxyacetic acid (2,4-D) (right) for 3 days. (C) Stimulation of adventitious root growth by sirtinol. The left two seedlings are wild-type, grown on 25 μM sirtinol for 8 days. The third plant from the left is a wild-type seedling grown on 25 μM sirtinol for 5 days and then transferred to 0.5× MS media for 1 day (only the hypocotyl structure is shown). The right seedling is wild-type, grown on 25 μM sirtinol for 5 days and then grown on 0.5× MS for 3 days. (D) Development of a tubelike true leaf in a sirtinol-treated Arabidopsis seedling (left). The tube-like true leaf can elongate and develop into a trumpet-like structure when transferred to MS media (middle), and normal true leaves developed when the sirtinol-treated plants continued to grow on MS media (right). (E) Effects of sirtinol on known auxin mutants. Seedlings were grown in the dark for 3 days and hypocotyl lengths were measured. For clarity, the axr2 and tir3 data are not shown in the graph.

Sirtinol also affected leaf development in Arabidopsis. Sirtinol-treated seedlings often developed cup-shaped organs (Fig. 2D). When sirtinol-treated seedlings were transferred to MS media, the bottom part of the cup turned green and the whole tube-like leaves continued to elongate to form a trumpet-like structure (Fig. 2D). The shoot meristem of the sirtinol-treated plants was buried inside the cup. When the cup-like plants were transferred to MS media or soil, normal true leaves developed (Fig. 2D). The ability to regulate plant development in such a timed and reversible manner suggests the utility of small molecules in dissecting a variety of developmental processes in any model organism.

Although sirtinol activates auxin-inducible genes and partially phenocopies auxin-treated plants, it caused some additional phenotypes. For example, cup-shaped true leaves were not observed in auxin-treated plants. It may be that sirtinol is more effectively transported to cells that exogenous auxin cannot reach because of auxin polar transport. This is consistent with the observation that auxin polar transport inhibitors affected leaf initiation and patterning when applied locally (16). We tested the responses of known auxin mutants to sirtinol treatment. All the tested auxin-signaling mutants [axr1 (12), axr2 (17), tir1 (18), and nph4 (19)] were less sensitive to sirtinol than the wild type (Fig. 2E), whereas the tested auxin-transport mutants [aux1 (20), pin2 (21), and tir3 (22)] responded similarly to the wild type (Fig. 2E), indicating that sirtinol affects only auxin signaling and probably is not transported via the auxin polar transport system.

We then undertook a genetic screen for mutants that were resistant to the effects of sirtinol, in an attempt to isolate sirtinol targets or downstream components. After screening 60,000 ethylmethane sulfonate–mutagenized M2 Arabidopsis seeds, we identified a single mutant, sir1 (for sirtinol resistant 1), which displayed elongated primary roots, elongated hypocotyls, and normal cotyledons in the presence of 25 μM sirtinol (Figs. 2E and 3A). The sir1 mutation also suppressed the ectopic expression of the auxin reporter genes that were induced by sirtinol treatment (Fig. 3B), suggesting that SIR1 regulates auxin-inducible genes.

Fig. 3.

sir1 was resistant to sirtinol treatment. (A) sir1 seedlings developed normally in the presence of 25 μM sirtinol. (B) Suppression of sirtinol-induced ectopic expression of DR5-GUS by the sir1 mutation. Left, DR5-GUS grown on 25 μM sirtinol; right, sir1/DR5-GUS grown on 25 μM sirtinol. (C) In the absence of sirtinol, sir1 plants are smaller in stature and have shorter primary roots and fewer lateral roots than wild-type plants. (D) Mature sir1 plants were pale green and had delayed development. (E) sir1 is hypersensitive to auxin. Root elongation was measured with 5-day-old seedlings that were transferred for 2.5 days to plates that contained various concentrations of auxin.

When grown on unsupplemented MS media, sir1 resulted in additional phenotypes, including smaller cotyledons and shorter primary roots than wild-type seedlings. In addition, sir1 plants were a paler green than the wild type and had far fewer lateral roots (Fig. 3C). When sir1 plants were grown in soil in a greenhouse, their development was delayed (Fig. 3D). Finally, sir1 was hypersensitive to exogenous auxin in a root elongation assay (Fig. 3E), indicating that SIR1 may negatively regulate auxin signaling.

sir1 was recessive and it segregated from a back-cross F2 population at a frequency of 25%, indicating that the associated phenotypes arise from a mutation in a single gene. The sir1 gene was mapped to the bottom of chromosome V within an interval of 65 kb (Fig. 4). We sequenced all the open reading frames (ORFs) in the 65-kb interval and found a point mutation that resulted in a substitution of a highly conserved serine with a phenylalanine residue in ORF At5g55130 (GenBank accession no. NP 200324) (fig. S1A). Transformation of sir1 with a 5-kb genomic fragment that contained only the ORF of At5g55130 and its regulatory sequences restored the sirtinol sensitivity and reversed other sir1 phenotypes, providing proof that At5g55130 is in fact SIR1 (Fig. 4). SIR1 was expressed throughout all stages of Arabidopsis development (fig. S2).

Fig. 4.

Identification of SIR1 and analysis of its domain architecture. (A) Cloning of sir1 by map-based cloning. cM, centimorgan; BAC, bacterial artificial chromosome; N, north; S, south. (B) Complementation of sir1 with the genomic fragment of ORF At5g55130. Left panel: Seedlings were transferred from MS plates to a plate that contained 25 μM sirtinol. Right panel: Plants in the left panel grown on sirtinol plates for 3 days. (C) The domain architecture of SIR1. SIR1 is composed of a ubiquitin-activating enzyme E1–like domain and a Rhodanese-like domain that shares significant homology with that of Arabidopsis prolyl cis-trans isomerase.

At5g55130 was originally annotated as a molybdopterin synthase sulfurylase, because of its homology to Escherichia coli MoeB protein (23). However, PSI-BLAST searches (fig. S1A) indicate that SIR1 is the Arabidopsis homolog of Uba4 from S. cerevisiae, a ubiquitin-activating enzyme E1–like protein (24) (Fig. 4 and fig. S1A). S. cerevisiae does not use molybdopterin as a cofactor and lacks all other molybdopterin biosynthesis machinery, and this indicates that the Uba4 homolog in Arabidopsis could participate in processes other than just molybdopterin biosynthesis. Pfam searches find that SIR1 has two major functional domains (Fig. 4 and fig. S1A): an N-terminal domain that is conserved among molybdopterin synthases, vitamin B1 synthases, and ubiquitin-activating enzymes (E1), and a C-terminal Rhodanese-like domain that shares homology with the C-terminal domain of Arabidopsis prolyl cis-trans isomerase. Each member of the E1/MoeB/ThiF superfamily catalyzes the adenylation of a C-terminal carboxyl group of a small protein, with adenosine triphosphate (ATP) as a substrate. The nucleotide-binding site in MoeB is very similar to that of the NAD-binding Rossmann folds (25). Given that sirtinol inhibits the Sirtuin family of NAD-dependent deacetylases in S. cerevisiae (8), our results suggest that the putative ATP-binding site in SIR1 may be a binding site for sirtinol, which is also consistent with our findings that the sir1 mutation occurred in the vicinity of the putative ATP-binding site and that SIR1 was a negative regulator of auxin signaling (fig. S1A). The Rhodanese domain is an alpha/beta fold that was originally found in Rhodanese proteins and later in a variety of other proteins (26). Extensive BLAST searches indicate that Rhodanese domains are uncommon in Arabidopsis but do occur in SIR1, in one of three prolyl isomerases, and in a small family of proteins involved in senescence (fig. S1, B and C).

Given that sir1 is hypersensitive to auxin and that SIR1 appears to be a direct target for sirtinol, we propose that SIR1 normally functions as a negative regulator in auxin signaling by dampening the positive auxin signals. The negative role of SIR1 is inhibited when sirtinol binds to SIR1, and therefore, the positive auxin signal runs unchecked and gives rise to the high auxin phenotypes we observed (Fig. 2, B to D). SIR1 functions up-stream of the Aux/IAA genes and the corresponding protein degradation machinery, and thus, auxin-resistant mutants of these components are also resistant to sirtinol (Fig. 2E). Therefore, by carrying out a sirtinol-resistant instead of an auxin-resistant mutant screen, we can not only identify components in the negative regulatory loop that should also be auxin-hypersensitive, but also identify all of the positive auxin-signaling components downstream of SIR1.

Although the detailed molecular mechanisms of how SIR1 may negatively regulate auxin signaling are not fully understood at present, the domain architecture of SIR1 offers important clues. Auxin is believed to regulate the degradation of Aux/IAA proteins through a ubiquitin-related system (11, 12). Phenotypes resulting from dominant mutations in Aux/IAA proteins arise from alterations in the stability of those proteins, but exactly how this happens is not yet clear [as reviewed in (27)]. All such mutations occur at or in the vicinity of two apparently essential proline residues (27). Given that SIR1 and one of the Arabidopsis prolyl isomerases each contain similar single-copy Rhodanese domains, it is conceivable that these molecules associate via these domains. Formation of such a complex may provide both a mechanism to regulate the conformation of the critical proline residues in Aux/IAA proteins in an auxin-dependent manner and a route to translate such a conformational change to a signal for protein degradation, presumably through the E1-like enzyme activity of the N-terminal domain of SIR1. Although prolyl isomerases have been found in many organisms and have been shown to regulate the stability of important cellular proteins (28), BLAST results find that the active-site rotamase domains of prolyl isomerases occur in conjunction with Rhodanese domains only in Arabidopsis and the rice Oryza sativa, suggesting that these particular prolyl isomerases may regulate a plant-specific process.

Supporting Online Material

Materials and Methods

Figs. S1 and S2

Table S1

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article