|Home | About | Journals | Submit | Contact Us | Français|
There is abundant evidence of multiple biosynthesis pathways for the major naturally occurring auxin in plants, indole-3-acetic acid (IAA), and examples of differential use of two general routes of IAA synthesis, namely Trp-dependent and Trp-independent. Although none of these pathways has been completely defined, we now have examples of specific IAA biosynthetic pathways playing a role in developmental processes by way of localized IAA synthesis, causing us to rethink the interactions between IAA synthesis, transport, and signaling. Recent work also points to some IAA biosynthesis pathways being specific to families within the plant kingdom, whereas others appear to be more ubiquitous. An important advance within the past 5 years is our ability to monitor IAA biosynthesis and metabolism at increasingly higher resolution.
The topic of auxin biosynthesis and metabolism in plants was comprehensively reviewed in 2005 (Woodward and Bartel 2005). Since then, more genes involved in IAA biosynthesis and metabolism have been identified. A combination of numerous valuable mutants, the manipulation of IAA synthesis in specific cell types, and direct measurement of IAA levels at tissue and cellular resolution now point to localized IAA biosynthesis and metabolism as playing key roles in specific developmental events (reviewed in Cheng and Zhao 2007; Lau et al. 2008; Zhao 2008; Chandler 2009). With apologies to any authors who were not included because of space constraints, this review summarizes those recent findings that require us to rethink yet again, the role of IAA biosynthesis and metabolism in auxin biology.
Within the past decade, the long-held hypothesis that IAA is synthesized primarily in the apical region and transported throughout the plant to form morphogenic gradients has undergone revision and refinement, concomitant with the ability to monitor de novo IAA synthesis in specific tissues at high resolution. Deuterium oxide (2H2O) is readily taken up by seedlings, and deuterium will exchange with hydrogen in the noncyclized intermediates early in the shikimate pathway (Fig. 1). Once incorporated, deuterium in the 4, 5, 6, and 7 positions, and to a lesser extent the 2 position of the indole ring in IAA are stable and will not exchange with hydrogen, even under alkaline conditions (Magnus et al. 1980). Thus, measuring the incorporation of deuterium into the indole ring of IAA is a measure of de novo IAA synthesis, as opposed to hydrolysis of IAA from pre-existing IAA conjugates. This approach has revealed that all parts of young seedlings are capable of synthesizing IAA (Ljung et al. 2001a; Bhalerao et al. 2002) with newly fertilized embryos, young leaves, and roots exhibiting high IAA synthesis activity (Ljung et al. 2001a; Ljung et al. 2001b; Ribnicky et al. 2002). Removal of aerial tissues early in seedling development and inclusion of the auxin transport inhibitor naphthalene acetic acid (NPA) in 2H2O labeling experiments reveals that shoot-derived IAA is important for lateral root emergence early in development. Later in development, auxin synthesis in the primary root is sufficient to sustain lateral root initiation and can drive lateral root emergence independently of shoot derived IAA (Bhalerao et al. 2002; Ljung et al. 2005). A very recent study using multiple Arabidopsis lines expressing cell-type specific GFP constructs allowed fluorescence activated cell sorting (FACS) of roots into cell types. 2H2O labeling of these sorted cells reveals that virtually all cells in the root apex are capable of synthesizing IAA (Petersson et al. 2009). These and other experiments described later support the hypothesis that it is a combination of IAA transport and localized IAA synthesis that form and maintain IAA gradients throughout the plant.
The biosynthetic pathways for IAA (Fig. 2) can be classified as Trp-dependent if IAA is derived via metabolism of Trp, or as Trp-independent if IAA is derived from an early indolic precursor of Trp (reviewed in Normanly et al. 2004; Woodward and Bartel 2005), most likely indole-3-glycerol phosphate (IGP) (Ouyang et al. 2000). Stable isotope labeling studies conducted in a variety of plant species have consistently demonstrated the differential utilization of Trp-dependent and Trp-independent IAA synthesis pathways at critical times in plant development (Ljung et al. 2002), including embryogenesis (Michalczuk et al. 1992; Ribnicky et al. 2002), fruit ripening (Epstein et al. 2002), in response to changes in temperature (Rapparini et al. 2002), during germination and early seedling growth (Ljung et al. 2001b), and following wounding (Sztein et al. 2002).
Decades of investigation have established that there are multiple Trp-dependent biosynthetic routes to IAA in plants (Normanly et al. 2004; Woodward and Bartel 2005) and microbes (Glick et al. 1999). Microbial IAA biosynthesis pathways are well defined, but none of the proposed IAA biosynthetic pathways in plants have been completely defined (dashed arrows in Fig. 2 represent pathway steps that have not been confirmed). Substantial progress has been made in the identification of IAA biosynthesis genes. An update follows.
The IPA pathway (Fig. 2) has been well characterized in microbes (Koga et al. 1991; Koga et al. 1992; Koga 1995). IPA has been identified as a native compound in Arabidopsis (Tam and Normanly 1998). Recently, TAA1 (Trp aminotransferase of Arabidopsis) has been shown to catalyze the formation of IPA from L-Trp in vitro (Stepanova et al. 2008; Tao et al. 2008), and mutants of TAA1 accumulate less IAA in response to simulated shade (Tao et al. 2008), ethylene (Stepanova et al. 2008), and probably high temperature (Yamada et al. 2009). As discussed later, the phenotypes of TAA1 and related TAR mutants provide specific examples of the role of localized IAA biosynthesis in the response to environmental and developmental cues. Whether the subsequent steps of this IAA synthesis pathway are similar to those in the microbial IPA pathway remains to be determined. Sequence analysis, including motif identification, predicts five pyruvate decarboxylases that are candidates for an indole pyruvate decarboxylase (IPDC) in Arabidopsis. All five genes were expressed in Escherichia coli, and a sensitive in vitro assay failed to demonstrate IPDC activity for any of the genes. Additionally, IPDC activity was not detected in protein extracts from wild-type Arabidopsis (Ye and Cohen 2009), so the enzymatic conversion of IPA to IAA may be very different from what is seen in microbes.
Although the molecular genetic tools associated with Arabidopsis have played a large role in the identification of many of the proposed IAA biosynthesis genes, dissection of IAA biosynthesis pathways in Arabidopsis is complicated by the way in which Trp metabolism and secondary metabolism intersect in this species. Specifically, some plants in the order Capparales, to which Arabidopsis belongs, uniquely make indole glucosinolates (IGs) and other indole-derived defense compounds (e.g., camalexin) (Kjaer 1974) from IAOx, a metabolite of Trp (Fig. 2). From mutant analysis, it appears that indolic secondary metabolites such as IGs and camalexin are a major metabolic sink for Trp (Bender and Celenza 2009). IAOx can be diverted to IAA synthesis from pathways that (a) primarily function in defense against biotic stress (Hull et al. 2000; Bak and Feyereisen 2001; Bak et al. 2001; Zhao et al. 2002; Bednarek et al. 2009; Clay et al. 2009; Mikkelsen et al. 2009), and (b) are not ubiquitous in the plant kingdom (Sugawara et al. 2009). IAOx has been measured in tobacco, rice, maize (Sugawara et al. 2009), peas (Quittenden et al. 2009), Brachypodium distachyon, and the aliphatic glucosinolate-producing watercress (J. Normanly, unpubl.), and found to be below the limits of detection (5–10pg). Based upon these observations, the IAOx pathway, which appears to have both indole-3-acetonitrile (IAN) and indole-3-acetamide (IAM) as parallel intermediates (Fig. 2), has been proposed to be specific to the IG-producing plant species (Sugawara et al. 2009).
Members of the YUCCA family of flavin monooxygenases have been implicated in IAA biosynthesis in maize (Gallavotti et al. 2008), rice (Yamamoto et al. 2007), petunia (Tobeña-Santamaria et al. 2002), and tomato (Exposito-Rodrıguez et al. 2007), in addition to Arabidopsis (Zhao et al. 2001; Kim et al. 2007), where these genes were first discovered. The proposed activity for YUCCA, conversion of tryptamine (TRM) to N-hydroxyl-TRM (Fig. 2), is based upon in vitro evidence (Zhao et al. 2001; Kim et al. 2007), and genetic validation of this pathway has not included profiling TRM levels in YUCCA overexpression lines (Zhao et al. 2001) or loss-of-function mutants (Cheng et al. 2006; Chen et al. 2007). N-hydroxyl TRM has been proposed to be a precursor for IAOx (Zhao et al. 2001), but whereas label from [2H5]TRM was incorporated into indole-3-acetaldehyde (IAAld), indole-3-ethanol, and IAA in pea roots, neither N-OH TRM nor IAOX were identified as labeled intermediates (Quittenden et al. 2009). IAOx levels were measured in the Arabidopsis yuc1 yuc2 yuc4 yuc6 quadruple mutant and found to be indistinguishable from wild type (Sugawara et al. 2009); however, given the restricted expression patterns of these four genes (Cheng et al. 2006), it is likely that any localized differences in IAOx levels would have been masked by measuring bulk tissue, even though the sampled tissue exhibited a visible phenotype. The Arabidopsis YUCCA overexpression mutant does not accumulate IAOx, but does accumulate IAN (J. Celenza, J. Cohen, and J. Normanly, unpubl.), which could be the result of ectopically expressing YUCCA. That is, the product of YUCCA may normally be restricted to cells that do not have enzymes capable of converting it to IAOx and subsequently IAN. Additionally, there may be separate IAOx pools in IG-producing plants. IAOx destined for IG synthesis is derived from the CYP79B2 and CYP79B3 pathway, and if IAOx is also derived from YUCCA, it does not contribute to IG synthesis (reviewed in Bender and Celenza 2009). The observation that [15N]TRM did not label IAOx in vivo in wild-type Arabidopsis, but did label IAA (Sugawara et al. 2009) leaves us with more questions about the role of IAOx in the YUCCA pathway.
IAN has been a challenging intermediate to work with in Arabidopsis because of the complexity of IG metabolism (reviewed in Bones and Rossiter 2006; Bender and Celenza 2009). Conversion of IAOx to IAA in Arabidopsis was recently proposed to occur with IAN and IAM as parallel intermediates, based on incorporation of label from IAOx (Sugawara et al. 2009). Label from IAOx was incorporated into both IAN and IAM, but to different degrees, seeming to rule out a precursor product relationship between IAN and IAM. There may be multiple pools of IAN (Bender and Celenza 2009), not all of which are available for synthesis of IAA, and this may complicate the interpretation of the labeling data. For example, IAN is implicated in the synthesis of the defense compound camalexin (Nafisi et al. 2007). The P450 CYP71A13 converts IAOX to IAN in vitro, and exogenous IAN is required to restore camalexin levels in mutants with defects in CYP71A13. IAN levels were not measured in the CYP71A13 mutant, but the fact that exogenous IAN was required implies that if there were indeed other endogenous sources of IAN, they were not available. Tobacco does not accumulate detectable levels of IAOx or IAN (Sugawara et al. 2009), yet ectopic expression of AtCYP79B2 or AtCYP79B3 leads to an accumulation of both and to increases in IAA as well (H. Nonhebel, J. Celenza, J. Cohen, J. Normanly, unpubl.). This implies that enzyme activities exist to convert IAOx to IAN and IAN to IAA in nonglucosinolate plants. Maize, which also does not appear to make IAOx, has at least two nitrilases, Zmnit1 and Zmnit2, that are proposed to have dual functions in kernels and seedlings as a heterodimer in the hydrolysis of b-cyanoalanine (Kriechbaumer et al. 2007), and Zmnit2 is proposed to convert IAN to IAA (purified Zmnit2 does so at a rate of 100 nmol/mg protein/min) (Kriechbaumer et al. 2006; Kriechbaumer et al. 2007). In contrast, IAN is not detectable in maize endosperm, which is a source of highly efficient Trp-dependent IAA synthesis activity; 32 pmol/mg/min (Kriechbaumer et al. 2006) or 140 pmol/mg protein/min (Hendrickson Culler and Cohen 2007). IAN is not a substrate for Trp-dependent IAA synthesis in this system (Hendrickson Culler and Cohen 2007). Interestingly, none of the other proposed intermediates in Trp-dependent IAA synthesis (Fig. 2) were good substrates in this system. Adding radiolabeled Trp to the endosperm assay yielded a small membrane protein to which Trp was covalently attached through a thioester bond (Hendrickson Culler and Cohen 2007). This protein-Trp conjugate was a substrate for IAA synthesis, releasing radiolabeled IAA, and upon hydrolysis of the thioester, racemic Trp.
The Trp-independent IAA synthesis pathway has been established in numerous plant species by way of stable isotope labeling studies (reveiewed in Woodward and Bartel 2005), yet remains undefined genetically. New alleles of iar4, a putative E1a subunit of mitochondrial pyruvate dehydrogenase, provide a possible link to the regulation of this pathway. The phenotypes of iar4-3 and iar 4-4 suggest a dual role for this protein in auxin homeostasis and the normal metabolic function of the acetyl-CoA-generating pyruvate dehydrogenase complex (Quint et al. 2009). The auxin response phenotypes in these mutants can be suppressed by increasing endogenous IAA, either by raising the temperature as in Gray et al. 1998, or ectopic expression of YUCCA, suggesting a deficiency in IAA levels. IAA biosynthesis was not impaired, but dual labeling that was able to distinguish between Trp-dependent and Trp-independent IAA synthesis showed that the Trp-independent pathway had been up-regulated. Levels of IAA-glutamate, a presumed storage form of IAA, were elevated in the mutants, implying a possible defect in conversion to free IAA.
IAA exerts its effects through the formation of local gradients and maxima/minima of IAA; thus, the regulation of IAA concentrations is central to the many roles that it plays in plant growth and development. The formation of IAA gradients in plant organs has been demonstrated directly in several systems by measuring IAA concentrations in segments of tissues, using stable isotope labeled internal standards and gas chromatography-mass spectrometry (GC-MS) (reviewed in Bhalerao and Bennett 2003). Direct measurement of IAA gradients include a 30-fold concentration differential across developing xylem in the wood-forming section of trees (Uggla et al. 1996), along a 15 mm tobacco leaf from petiole (high) to tip (low) (Ljung et al. 2001a), a broad range of concentrations in 3-week old soil grown plants, with siliques having the highest and mature leaves the lowest (Muller et al. 2002) and Arabidopsis roots from the root-shoot junction (high) to the root tip (low) (Bhalerao et al. 2002). Here, the source of the IAA in the gradient is proposed to be shoot-derived. Interestingly, further dissection of the root tip (1–2 mm sections) reveals a much shallower gradient of high to low from the root tip upwards, implying local synthesis, which is supported by measurement of auxin synthesis rates in stable isotope labeling experiments (Ljung et al. 2001a; Ljung et al. 2005; Ikeda et al. 2009).
Concomitant with these sorts of measurements, which are painstaking, IAA gradients have been inferred from the observation of a polar auxin transport (PAT) system in plants and subsequently from correlations between the distribution of auxin transport components (reviewed in Vieten et al. 2007) and the gene expression output from either colorimetric or fluorescent reporter constructs driven by the synthetic auxin-responsive promoter DR5 (Ulmasov et al. 1997), which responds to the TIR1-mediated signal transduction pathway (reviewed in Lau et al. 2008; see also an animation in Laskowski 2006). The readout from the DR5-reporter system has frequently been interpreted to be an indicator of IAA location or amount; however, as pointed out in at least two recent reviews (Ljung et al. 2004; Chandler 2009), this is a risky extrapolation, considering the complexity of auxin signal transduction. In only a very few cases (Nacry et al. 2005; Ruzicka et al. 2007; Swarup et al. 2007; Ikeda et al. 2009), notably all in Arabidopsis roots, the DR5-reporter output has been directly correlated with measurement of IAA in the same tissues. The output from reporter constructs generally correlated with the measured levels, but with much less precision. In the Arabidopsis IAA-overproducing mutant sur2, the only correlation between DR5-GUS staining and IAA levels was in the base of the hypocotyl, even though more dramatic increases in IAA were measured in the apical portions of the plant compared to wt (Ljung et al. 2004). The most definitive example of the inadequacy of DR5-reporters as measures of IAA levels or location comes from IAA measurements on cells that expressed DR5-GFP and were separated from nonexpressing cells by FACS (Petersson et al. 2009). These measurements were compared with those of other specific cell types, isolated in the same manner. DR5-GFP was expressed only in a very few cells in the root tip, whereas IAA levels measured in the other root cell types revealed a much more nuanced and broad range of IAA levels in the root. Another important consideration for the DR5-reporter system is that it depends upon IAA/AUX and ARF interactions, and expression of several ARFs involved in auxin signaling has been shown recently to depend upon a microRNA (Mallory et al. 2005; Wang et al. 2005). In maize, the ARF3-regulating siRNA forms a gradient from the upper to lower sides of leaves, which patterns ARF3 (Chitwood et al. 2009). Interpretation of DR5-reporter readout as an indicator of auxin response, combined with direct measurement of IAA and/or localized manipulation of auxin levels in a variety of mutants, has consistently supported the hypothesis that auxin transport, biosynthesis, and metabolism work in concert to establish, maintain, or alter auxin maxima to affect changes in plant developmental programs. Some recent examples of this follow.
IAA transport from the shoot is proposed to be sufficient to generate the gradients that control lateral root initiation and emergence (Grieneisen et al. 2007). This model is being re-examined in light of the evidence for significant IAA biosynthesis activity in the root (Ljung et al. 2001a; Ljung et al. 2005; Petersson et al. 2009). Ethylene response mutants have been particularly instructive here.
The stimulation of ethylene synthesis by high concentrations of auxin is well established (reviewed in Stepanova and Alonso 2005), and recently, a reciprocal interaction has been demonstrated with the discovery that two root-specific Arabidopsis ethylene-insensitive mutants, wei 2 and wei 7, are alleles of genes encoding the anthranilate synthase α (ASA) subunit and the AS β (ASB) subunit, respectively (Stepanova et al. 2005), both of which are required for IAA synthesis (Fig. 2). Direct measurement of IAA levels in the root, combined with either genetic or chemical disruption of auxin response and transport (Ruzicka et al. 2007; Swarup et al. 2007), establish that ethylene stimulates IAA synthesis and transport in root tips, which creates IAA gradients in the root elongation zone, inhibiting cell elongation. The identification of the Arabidopsis TAA1 gene, which encodes the Trp aminotransferase proposed to catalyze the first step in the IPA pathway, adds detail to this model (Stepanova et al. 2008). TAA1 is expressed in the quiescent center of roots. The double mutant combination of wei8 (defective in TAA1) and tar2 (defective in a TAA-related gene) exhibit ethylene insensitivity in the root and in the ethylene-induced apical hook response in etiolated seedlings. The double mutants appear to have wild-type auxin transport and response, but lower levels of IAA, measured directly in roots and hypcotyls. Exogenous IAA restores ethylene sensitivity in roots only. These findings are consistent with a role for the IPA pathway in ethylene-responsive IAA synthesis that is involved in regulating lateral root growth. Interestingly, in triple mutant combinations with wei8tar2, the sur1/rty, and sur2 alleles (Fig. 2), which themselves accumulate IAA, alleviate the root-specific phenotypes of the TAA1 and TAR2 defects, implying convergence of the IAOx and IPA pathways in the root. The TAA1 gene was also identified in two other mutant screens; one for mutants unable to elongate in simulated shade conditions (Tao et al. 2008) and another for hypocotyl length (Yamada et al. 2009). TAA1 is proposed to provide a rapid local increase in IAA in response to changes in R/FR ratios (Tao et al. 2008) and in response to temperature during hypocotyl elongation (Yamada et al. 2009).
Arabidopsis mutants with defects in CTR1, a negative regulator of ethylene signaling, examined in combination with the ASA and ASB mutants, wei2 and wei7, respectively, as well as other ethylene response and auxin transport mutants, demonstrate that IAA synthesized in the root tip contributes to an auxin transport-dependent auxin maxima distal to the root tip that establishes planar polarity of epidermal cells, specifically, the position of root hairs (Ikeda et al. 2009).
Another twist to this story is the finding that ASA1 is responsive to methyl jasmonate (MeJA) in roots. MeJA inhibits primary root growth and induces LR formation. The jdl1/asa1 mutant is not responsive to MeJA induction of LR, but was responsive to MeJA with regard to primary root growth (Sun et al. 2009). MeJA-induced LR formation requires functional COI1, a regulator of MeJA response, whereas ethylene responses in COI1 with regard to LR formation were normal, implying that the response of ASA1 to ethylene and MeJA may be independent.
Sugars act as signaling molecules (reviewed in Gibson 2005) and intersect with ethylene regulation of lateral root production. An Arabidopsis T-DNA mutant insensitive to the nonmetabolizable sucrose analogue turanose, is defective in the WUSCHEL-related homeobox gene WOX5 and has root phenotypes indicative of disrupted auxin homeostasis (Gonzali et al. 2005). WOX5 is expressed in the quiescent center and is auxin inducible. The WOX5 mutant up-regulates ethylene production, SUR2 expression, and IAA conjugation activity, suggesting that WOX5 normally functions as a negative regulator of IAA inactivation and promotes IAA biosynthesis through the CYP79B2 CYP79B3 IAOx pathway, thereby contributing to the maintenance of the auxin maximum required for lateral root formation.
Arabidopsis mutants with defects in TAA1/TAR2 or a combination of several of the YUCCA genes result in poorly developed gynoecium, defective embryos, and infertility (reviewed in Cheng and Zhao 2007; Lau et al. 2008). The overlap in both the expression of the TAA1 and YUCCA gene family members and in embryo defects in corresponding mutants implies that these two IAA biosynthetic pathways, both incompletely defined at this point, may converge.
A map of auxin response and YUCCA1 and YUCCA2 expression throughout seven defined developmental stages of the Arabidopsis female gametophyte (embryo sac) predicts dynamic and asymmetric distribution of auxin gradients for the duration of embryo sac development (Pagnussat et al. 2009). Down-regulation of auxin response during these seven stages within the embryo sac only, was achieved with a synthetic microRNA that targeted a subset of ARFs. The results included changes in cell identity at the cellularization stages and mis-expression of cell-specific marker genes. Nuclear positioning in the early stages of embryo sac development was not affected. Similar results were obtained by overexpressing YUCCA1 throughout the embryo sac over the seven developmental stages. PIN localization was observed only during early stages, implying the formation of gradients early in the process. These results are consistent with the hypothesis that localized IAA synthesis contributes to the gradients that determine cell patterns in the embryo sac.
Arabidopsis STY1 has a zinc-binding RING-finger-like domain belonging to the SHI gene family (Sohlberg et al. 2006). STY1 activates transcription of YUCCA4. The gynoecia of sty1-1 and sty1-1 sty2-1 mutants exhibit abnormal style morphology and vascular patterning, are hypersensitive to chemical inhibition of PAT, and exacerbate the gynoecia phenotypes of PAT/signaling mutants pin1-5, pid-8, and ett-1. These phenotypes are consistent with a role for STY1 and STY2 in maintaining auxin gradients through regulating localized IAA biosynthesis by way of the YUCCA pathway. Lastly, a maize mutant, defective in a putative YUCCA ortholog is unable to properly initiate axillary meristems and lateral organs during vegetative and inflorescence development (Gallavotti et al. 2008).
Contrary to these observations, localized manipulation of IAA biosynthesis and conjugation in Arabidopsis epidermal cells with the A. tumefaciens IAA biosynthesis gene, iaaM, and the P. syringae IAA-Lys synthase gene, iaaL, respectively, had no effect on embryo pattern formation (Weijers et al. 2005). The conclusion was that localized IAA biosynthesis and metabolism did not contribute to gradients in the embryo. Instead, a functional PAT system was proposed to buffer against changes in IAA levels and to maintain IAA gradients in the embryo. Direct measurement of IAA levels in these lines revealed significant changes in whole leaves and hypocotyls, and the substrate for the ectopic IAA synthesis was not rate limiting in the embryo. However, the IAA gradient in the embryo was extrapolated from DR5-reporter activity, which is not a direct measurement of IAA. This leaves open the possibility that IAA gradients were unchanged in the embryos because of rapid conjugation by endogenous enzymes and thus inactivation of IAA.
Most of what we know about IAA metabolism in plants is from work done in Arabidopsis and maize (Fig. 3) (reviewed in Normanly 1997; Slovin et al. 1999; Cohen and Gray 2006). The most abundant IAA metabolites are IAA conjugated to a variety of small molecules, peptides, or proteins (Fig. 3) (reviewed in Seidel et al. 2006). IAA-conjugates are presumed to serve as biologically inactive, storage forms of IAA, or in some cases (IAA-Asp in Arabidopsis), as the first step in a degradation pathway (Ostin et al. 1998). IAA conjugates are generally classified by the type of conjugate linkage; ester-linked for IAA-glucose, IAA-myoinositol, and large molecular weight IAA-glycans, and amide-linked for IAA-amino acids, IAA-peptides, and IAA-proteins. The conjugate moieties vary somewhat by species (see Seidel et al. 2006). The predominant IAA conjugates in Arabidopsis are predicted to be amide-linked (Ljung et al. 2002) based upon the amount of free IAA that is released when amide linkages are broken by treatment of plant extracts with 7N NaOH at 100 °C versus treatment with 2N NaOH at room temperature, which releases ester-linkages (see, for example, Chen et al. 1988). However, in IG-producing species such as Arabidopsis, the presence of high levels of IAN can complicate this analysis. IAN is nonenzymatically converted to IAA in the presence of 7N NaOH, thereby overestimating the amount of amide-linked IAA (Ilic et al. 1996). When measured directly, levels of IAA-amino acids in Arabidopsis are quite low, within the same order of magnitude as free IAA (Tam et al. 2000; Kowalczyk and Sandberg 2001; Rampey et al. 2004), and it is presumed that most of the amide conjugates are peptides and small proteins (Ljung et al. 2002), which are not straight forward to quantify (reviewed in Seidel et al. 2006). Other metabolites of IAA include indole-3-butyric acid (IBA), which is typically present in levels lower than free IAA (Ludwig-Muller and Epstein 1994; Jones et al. 2005) and possibly methyl-IAA, which is below the limits of detection in Arabidopsis (Qin et al. 2005), so possibly turned over very rapidly. Thus, inclusion of quantitative data about IAA metabolite levels in studies of IAA homeostasis is fairly infrequent.
The role of IAA-peptides and IAA-proteins is still under investigation. IAA-protein IAP1 was originally isolated from bean (reviewed in Seidel et al. 2006), and antibodies to bean IAP1 have been used to isolate ortholog genes from Arabidopsis. One of the Arabidopis IAP proteins is a seed storage protein and another is present in the embryo. Rather than a storage form of IAA, it is tempting to speculate that IAA may be a “tag” for these proteins, directing them to specific locations in the cell or interactions with other proteins, similarly to other small molecule protein modifiers. Further characterization of the Arabidopsis IAA-proteins should provide new hypotheses regarding the mechanisms of IAA action.
The genes for amidohydrolases specific to IAA-amino acids have been identified (Bartel and Fink 1995; Davies et al. 1999; LeClere et al. 2002) and collectively their expression patterns encompass areas of auxin biosynthesis activity (Rampey et al. 2004), although direct correlations between the expression of these genes and that of CYP79B2, YUCCA, or TAA1, for example, have not been described and should be informative about the interplay between IAA biosynthesis and conjugation in the maintenance of IAA gradients. For IAA-conjugate formation, the GH3 gene family has been identified, some members of which are amidosynthases and can form amide conjugates of IAA with amino acids (Staswick et al. 2002; Staswick et al. 2005). GH3.2 through GH3.8 and GH3.17 are active with several amino acids in vitro. So far, only GH3.6 and GH3.8 have been overexpressed in vivo (Staswick et al. 2005; Ding et al. 2008). IAA-Asp is the predominant conjugate in both cases, but profiling of all possible IAA-amino acid conjugates was not done. The Arabidopsis IAMT1 gene encodes a methyltransferase that methylates IAA in vitro (Qin et al. 2005). A mutant allele that overexpresses this gene exhibited hyponastic leaf phenotypes, suggesting that methylation of IAA plays a role in regulating plant development and auxin homeostasis. Members of an Arabidopsis family of methyl ester transferases are able to hydrolyze MeIAA in vitro (Yang et al. 2008).
Members of the GH3 family conjugate amino acids to IAA, SA, and JA (Staswick et al. 2002), which in the case of IAA and presumably SA, inactivates the signaling molecule (Staswick et al. 2005) and in the case of JA, activates it (Staswick and Tiryaki 2004). Some GH3 family members play dual roles in conjugating signaling molecules, providing a link between signaling pathways. For example, GH3.5 is able to conjugate both SA and IAA, and this impacts resistance or susceptibility to the bacterial pathogen P. syringae (Zhang et al. 2007). Overexpression of GH3.5 resulted in increases in IAA-Asp and enhanced resistance to biotic and abiotic stress, whereas the converse was observed in a GH3.5 T-DNA knockout (Park et al. 2007). Overexpression of GH3.8 in rice gives rise to enhanced disease resistance and growth defects related to decreases in free IAA levels (Ding et al. 2008).
Several of the GH3 IAA amidosynthases show activity in vitro with Trp as a substrate, and IAA-Trp is a native compound in Arabidopsis (Staswick 2009). Unlike other IAA conjugates, IAA-Trp is not an inactive form of IAA; rather, it is an IAA antagonist, rendering resistance to the root growth inhibitory effects of IAA, IBA, and 2,4-dichlorophenoxyacetic acid (2,4-D). The mechanism by which IAA-Trp counteracts IAA is not clear, but IAA-Trp is probably not a source of free IAA based on the substrate specificities of known Arabidopsis amidohydrolases (LeClere et al. 2002), and its antagonistic effects require functional TIR1, but not auxin transport. With the identification of a Trp-protein intermediate in Trp-dependent IAA synthesis (Hendrickson Culler and Cohen 2007), it would be interesting to determine the effects, if any of IAA-Trp on this reaction.
IAA is proposed to be enzymatically converted to IBA and vice versa (reviewed in Woodward and Bartel 2005). IBA is classified as an auxin and it remains to be determined if IBA itself is an auxin, or whether IBA effects are through its metabolism to IAA. IBA quantification in Arabidopsis reveals that it is generally present in lower amounts than free IAA (Ludwig Müller et al. 1993), in one case as low as 2% of free IAA levels (Jones et al. 2005). IBA synthetase activity, measured in vitro in a variety of Arabidopsis ecotypes, revealed varying activities (Ludwig Müller 2007). In the original report on the presence of IBA in maize, some cultivars had detectable levels of IBA and others did not (Epstein et al. 1989). A similar situation may exist for Arabidopsis, as inconsistent levels have been observed by groups who have critically examined IBA levels analytically (Ludwig Müller et al. 1993; Jones et al. 2005; K. Ljung, unpubl.; J. Normanly, unpubl.; J. Cohen, unpubl.). We do not as yet understand what developmental, tissue, environmental, or genetic differences might account for this variation in amounts found in the various investigations.
One IBA resistant mutant, ibr5, identified an auxin response pathway that acts independently of the TIR1-mediated proteosome degradation of repressor proteins (Strader et al. 2008a), and a suppressor screen of ibr5 has yielded an ATP-binding cassette transporter implicated in cellular efflux of the synthetic auxin 2,4-dichlorophenoxyacetic acid (Strader et al. 2008b). This screen identified a number of loci that are likely to be new components of auxin signaling, transport, and metabolism.
The catabolism of IAA is proposed to occur either by enzymatic oxidation of the indole nucleus of IAA (to form ox-IAA or ox-IAA-conjugates) (Fig. 3), or through oxidative decarboxylation of the IAA side chain. Most of this work has been done in vitro (see, e.g., Nonhebel et al. 1985; Beffa et al. 1990), and none of the genes for IAA catabolism have been definitively identified. Recently, three new oxidative metabolites of IAA have been identified in Arabidopsis using a sensitive MS screen of HPLC fractions from extracts of 2-week-old plants (Fig. 3) (Kai et al. 2007a). Two new IAA conjugates, IAA-Phe and IAA-Val, were identified indirectly, which are good substrates for the GH3 family of amidosynthases. OxIAA-Glc is proposed to be a primary oxidative metabolite of IAA in Arabidopsis based on quantification with deuterium-labeled internal standard (Kai et al. 2007a; Kai et al. 2007b). This recent observation is consistent with previous work in which OxIAA-hexose was identified as a major metabolite of IAA in Arabidopsis (Ostin et al. 1998). In the case of OxIAA-Glc measurement, the deuterium, label was incorporated into the side chain (2′) (Fig. 1) of OxIAA-Glc, which is subject to exchange with hydrogen, not just in alkaline conditions, but likely in plant extracts as well. It has been known for some time (Caruso and Zeisler 1983) that using side chain deuterium-labeled IAA as an internal standard overestimates endogenous IAA concentrations by as much as 10-fold. Similar results have been observed with side chain deuterium-labeled IAN (threefold overestimation) (J. Normanly, unpubl.).
The identification of new genes involved in IAA biosynthesis and metabolism has continued at a brisk pace, and the use of mutants to characterize their function has been invaluable in revealing the complexity of interactions between IAA and other signaling molecules. The ability to quantify IAA at high resolution has always been a bottleneck for auxin biology, but significant progress has been made here as well, and the next challenge will be to more broadly profile IAA precursors and metabolites along with those of other signaling molecules. The combination of these sort of data with the molecular genetic approaches that have brought us to this point will only aid us in finally answering the long-standing question of how IAA is made in plants.
The author is supported by National Science Foundation (NSF) MCB 0517420 and 0725192.
Editors: Mark Estelle, Dolf Weijers, Karin Ljung, and Ottoline Leyser
Additional Perspectives on Auxin Signaling available at www.cshperspectives.org