The Evolving Complexity of the Auxin Pathway

The plant hormone and signaling molecule auxin is a key player during pattern formation, organogenesis, and various physiological processes. Recent discoveries in auxin biology point toward an auxin pathway with a higher complexity than previously anticipated. This prompted us to review this constantly growing field and to put these novel and exciting findings into a broader developmental and evolutionary context. Auxin signaling can be divided broadly into three layers that contribute to its complexity: the spatio-temporal pattern of its biosynthesis, its directional transport, and cell- or tissue-specific responses. In Arabidopsis thaliana, auxin is synthesized via several pathways in embryos, leaves, and roots (reviewed in Woodward and Bartel, 2005). The active transport of auxin, leading to the establishment of auxin maxima and gradients, is mediated by specific proteins: PIN-FORMED (PIN) efflux carriers, AUXIN1/LIKE AUX1 (AUX1/LAX) influx carriers, and MULTIDRUG RESISTANCE/P-GLYCOPROTEIN transporters (reviewed in Vieten et al., 2007). Upon reaching specific threshold concentrations, auxin induces specific responses that trigger diverse developmental and physiological effects (reviewed in Teale et al., 2006). The abundant interlinkages between these layers and the variety of protein complexes involved in auxin-related processes contribute to the overall complexity of the auxin pathway. In addition, there are numerous inputs, intrinsic and environmental cues, that feed into this pathway, and this is quite frequently mediated by other hormonal pathways. Although the auxin pathway is being unraveled bit by bit, we will illustrate that there is still a long way to go before it will be fully understood. Basics and Evolution of Auxin Signaling and Response. (A) Schematic representation of important domains and motifs of ARFs and AUX/IAAs. LxLxL resembles the repressive motif in ETHYLENE RESPONSE FACTOR–associated amphiphilic repression domains. DBD, B3 DNA binding domain; AD, activation domain; RD, repression domain. (B) Simplified scheme of auxin-responsive gene regulation through an activating ARF in the absence or presence of auxin. ASK1, ARABIDOPSIS SKP1-LIKE; CUL1, CULLIN1; E2, a conjugating enzyme; RBX1, RING-BOX PROTEIN1; RUB1, RELATED TO UBIQUITIN1. (C) Identical phenotypes of gain-of-function AUX/IAA (iaa12/bdl) and loss-of-function ARF (arf5/mp) auxin response mutants. (D) Phylogenetic scheme highlighting the evolution of auxin signaling and response, based on physiological and genomic data. The colored bars above the branches point out positive evidence for the indicated feature. It should be noted that only from mosses onward the available genome sequence has been analyzed in depth, and likewise the observed auxin responses display biological significance. Years taken from Yoon et al. (2004). *, fully sequenced genome; (), unconfirmed finding; NPA, naphthylphthalamic acid (an auxin transport inhibitor). The 29 members of the Arabidopsis AUX/IAA family usually contain four conserved domains (Figure 1A) and display strong genetic redundancy, since even triple loss-of-function mutants have no apparent abnormal phenotype (Overvoorde et al., 2005). By contrast, when AUX/IAA proteins are stabilized through a specific amino acid exchange in the degron sequence within domain II (Figure 1A), auxin responses are disrupted, and this often results in dramatic phenotypes, as exemplified by solitary root (slr)/iaa14 (Fukaki et al., 2002) or bodenlos (bdl)/iaa12 (Hamann et al., 2002) (Figure 1C). Interestingly, one clade within the AUX/IAA protein family contains proteins that lack all or some of domain II and are stable relative to other AUX/IAAs. Overexpression of some of these noncanonical AUX/IAAs causes strong auxin-related aberrant phenotypes, such as malformed vasculature of cotyledons, collapse of the root apical meristem, and defects in gravitropic response (Sato and Yamamoto, 2008). In Arabidopsis, there are 23 ARF genes encoding proteins that contain a B3 DNA binding domain, a repression or activation domain, and two domains that share high similarity with domains III and IV of AUX/IAAs (Figure 1A). ARFs bind to the auxin-responsive element TGTCTC, thereby regulating the expression of auxin-responsive genes. Domains III and IV mediate the dimerization of AUX/IAAs and ARFs, and this specific interaction blocks the activity of at least activating ARFs (Figure 1B; reviewed in Guilfoyle and Hagen, 2007). Therefore, arf loss-of-function mutations, such as arf7 arf19 (Okushima et al., 2005; Wilmoth et al., 2005) or monopteros (mp)/arf5 (Hardtke and Berleth, 1998), often result in phenotypes similar to those caused by the stabilized AUX/IAA interaction partner (Figure 1C). In the past few years, it was shown that TIR1 and some close homologs act as auxin receptors in Arabidopsis. TIR1 is an integral component of the SKP1/CULLIN/F-BOX PROTEIN (SCF)TIR1 complex that eventually mediates the ubiquitination of AUX/IAAs and thereby destines them for 26S proteasome-dependent degradation (reviewed in Abel, 2007). AUX/IAAs are bound by TIR1 in the TIR1 pocket via their domain II, and auxin acts as a molecular glue to enhance this interaction (Figure 1B; Tan et al., 2007). During evolution of land plants, developmental mechanisms arose that resulted in coordinated multicellular growth. It is thought that this may well have been achieved, at least in part, through the evolution of the auxin pathway (Rensing et al., 2008). The current knowledge of this evolutionary issue is summarized in Figure 1D. An indication of an ancient evolutionary origin of auxin-related mechanisms might be suggested by the as yet unconfirmed finding of at least two AUX/IAAs in Chlamydomonas reinhardtii (Palenik et al., 2007). If confirmed, this finding might suggest that single-celled green algae that diverged from the land plant lineage some 1 billion years ago already possessed specific components of the auxin pathway. However, this conclusion remains in doubt as several researchers have been unable to convincingly reproduce these results (S. Lau and I. De Smet, unpublished data). In addition, the function of such proteins in Chlamydomonas is unknown, and Ostreococcus, a single-celled member of a different class of green algae, appears to have lost the respective genes (Palenik et al., 2007). For morphologically more complex green algae, such as Chara contraria, there is no sequence information available, but their generative cells respond to auxin treatment with the depolymerization of microtubules, which is reminiscent of what happens in vascular plants (Jin et al., 2007). Red and brown algae also appear to contain and/or respond to auxin (reviewed in Cooke et al., 2002). For instance, in Fucus distichus, a model brown alga, the embryo is affected by auxin and by auxin transport inhibitors (Sun et al., 2004). However, it should be noted that these physiological responses and the presumptive presence of some parts of the auxin pathway might not reflect a real endogenous role for auxin in these organisms. While the function and importance of auxin in the case of these diverse and polyphyletic algal lineages is not well understood, there are numerous indications of its emerging role in land plants. For example, several physiological experiments revealed the occurrence of a basic auxin metabolism, an auxin-dependent apical dominance, and polar auxin transport within the bryophytes (liverworts, hornworts, and mosses) and pteridophytes (lycophytes, horsetails, and ferns) (reviewed in Cooke et al., 2002). Moreover, compelling evidence for the existence of at least a basic auxin pathway in mosses can be drawn from the sequenced genome of Physcomitrella patens and the first analysis of its content. All major players of the higher plant auxin pathway, such as biosynthesis proteins (YUCCAs), receptors (TIR1/AFBs), transporters (PINs and AUX1/LAXs), and transcriptional regulators (AUX/IAAs and ARFs), are encoded in the Physcomitrella genome (Rensing et al., 2008). In agreement with these genomic data, it was recently shown that a small auxin-related molecule, which specifically affects the TIR1-mediated response to auxin in Arabidopsis, also antagonizes auxin response in Physcomitrella (Hayashi et al., 2008). Vascular plants are distinguished from Physcomitrella by a higher developmental complexity, and this is accompanied by an expansion of the genetic machinery of the auxin pathway in those species (Goldfarb et al., 2003; Jain et al., 2006; Kalluri et al., 2007; Rensing et al., 2008). Whereas the Physcomitrella genome contains 55 auxin-related genes representing 0.14% of all protein coding loci, the corresponding numbers in, for example, Arabidopsis are 174 and 0.65%, respectively (Rensing et al., 2008). Remarkably, a strikingly low fraction of auxin-related genes in Physcomitrella encodes AUX/IAAs compared with vascular plants (Rensing et al., 2008). This suggests that enrichment of auxin-related gene families has occurred within vascular plant genomes, for example, via preferential preservation of these genes after duplication events. An extended number of genes enables a higher developmental complexity via different nonexclusive means: changes in the expression pattern and evolution of new or different functions of the corresponding gene products. In the case of the auxin pathway, the relative contributions of these mechanisms to the assumed increased complexity have been thoroughly investigated for the AUX/IAAs in Arabidopsis. Several promoter swapping experiments for AUX/IAAs suggest a predominant role for gene-specific expression patterns compared with the contributions of protein specificities. Differences in protein specificity nonetheless confer a certain degree of distinctiveness to the investigated AUX/IAAs, since protein-specific differences in function were noted when their expression was driven by the same promoter (Weijers et al., 2005; Muto et al., 2007). Both altered expression and function appear to have arisen since the last common ancestor of Arabidopsis and Physcomitrella and may have played a significant role in the evolution of greater morphological complexity in vascular plants. The increased number of possible combinations of auxin signaling components in vascular plants, with each combination potentially giving rise to another output, enhances the diversity or at least the fine-tuning of auxin responses. Still, it remains to be elucidated whether, for instance, the enormous number of theoretically possible AUX/IAA–ARF interaction pairs is of any biological relevance. Cell- or tissue-specific auxin responses depend on mechanisms that bring about specific control of transcript and/or protein levels and the subsequent differential interpretation of auxin maxima and gradients through specific combinations of auxin signaling components. Upstream Regulation and Downstream Events of Auxin Signaling and Response. (A) Transcriptional regulation of AUX/IAAs. PAX1 regulation happens via a direct mechanism or through a feedback loop involving AXR3/IAA17. Dotted arrows indicate an unclear regulatory mechanism. (B) Regulation of the levels of ARF transcripts and proteins. Dotted arrow indicates an unclear regulatory mechanism. (C) Scheme of putative interactions between AUX/IAAs (IAA12/BDL), ARFs (ARF5/MP), a corepressor (TPL), a coactivator (MYB77), and proteins involved in histone acetylation (HAG1) and deacetylation (HDA19) in the presence and absence of auxin. Orange/yellow balloons indicate acetyl residues. (D) Model for hypophysis specification during embryogenesis, highlighting the role of IAA12/BDL and ARF5/MP in the embryo proper and an additional auxin response in combination with an MP-dependent putative mobile signal in the hypophysis. Hatched lines indicate MP-mediated auxin responses; red disc with a question mark indicates putative mobile signal. (E) Model for the asymmetric cell division of pericycle cells during lateral root initiation, focusing on SLR/IAA14, ARF7, ARF19, and their downstream targets LBD16 and LBD29. In addition, a means of modulating or impeding auxin responses is via the negative posttranscriptional regulation of the expression of ARFs through small RNAs (Figure 2B; reviewed in Teale et al., 2006) For example, transcript levels of ARF8 are regulated by miR167, and this regulation was shown to play a role in auxin-mediated lateral root development (Gifford et al., 2008). Another mechanism of fine-tuning auxin responses could be the (developmentally) controlled degradation of ARFs. This topic has not received much attention, but a recent report describes the degradation properties of ARF1, which appears to be degraded in a pathway distinct from the way AUX/IAAs are degraded (Salmon et al., 2008). These researchers showed that the degradation of ARF1 at moderate rates is proteasome dependent but does not require a CULLIN1-based SCF complex and is not affected by auxin (Figure 2B; Salmon et al., 2008). Provided that the degradation of ARF proteins is regulated in a developmentally relevant manner, this opens up intriguing new possibilities for an even more subtle regulation of auxin responses. A possible role of targeted ARF degradation could be the prevention of (specific) auxin responses in certain tissues, which could be especially important during the restriction of the expression domain of ARFs. With respect to events occurring during auxin responses, most attention has been paid to the interaction of AUX/IAAs and ARFs. Undoubtedly, AUX/IAA–ARF interaction pairs are principal regulators of auxin-responsive gene expression, but recent studies have revealed that other proteins are also recruited to this core unit. One protein that is able to act as a coactivator with ARFs is MYB DOMAIN PROTEIN77 (MYB77). It was recently shown that MYB77 interacts in vitro with several supposedly activating and repressing ARFs via their C termini and with IAA19 to promote auxin-responsive gene expression (Shin et al., 2007). In the case of ARF7, this was confirmed in planta, and based on the lateral root phenotype of the arf7 myb77 double mutant, it was demonstrated that ARF7 and MYB77 act in a synergistic manner (Shin et al., 2007). While MYB transcription factors had not been implicated previously in auxin signaling, there were indications that the expression of some MYB family members is itself regulated by auxin (Kranz et al., 1998). Because both auxin response elements and putative MYB binding motifs are present in close proximity in promoters that are potentially regulated by ARFs and MYB77, it is feasible that ARFs and MYB77 could interact at those promoters and gene expression (Figure However, since MYB77 to the ARF dimerization domain, it is possible that these MYB binding are not and that MYB77 is recruited to promoters in a similar manner as AUX/IAAs (Shin et al., 2007). In the mutant, which acts as a negative for the RELATED the is into a root et al., 2006). was to be a transcriptional corepressor based on the properties of its domains as well as on genetic interactions with (HDA19) and (HAG1) et al., 2006). Recent demonstrated that can all caused by the in the AUX/IAA gene A of the mechanisms this from the finding that acts in a complex with and to domain of specifically to a motif the ETHYLENE RESPONSE FACTOR–associated amphiphilic repression motif in this domain, which is for repression and et al., 2008). These results are with a report that a putative thought to function in with a histone is for the of lateral root In addition, of a histone was to the lateral root defects in the (Fukaki et al., 2006). The of histone in auxin response could be via auxin-dependent changes in the histone acetylation and for instance, in the mutant, this leading to aberrant developmental This mechanism of developmental might be as a of the developmental control by complexes in plants and (reviewed in and 2007). For auxin maxima or auxin gradients to any trigger or developmentally relevant processes. processes in plants are to be auxin the establishment of the of in the the specification of the the of and the of lateral roots (reviewed in De and and various physiological responses, such as (reviewed in Vieten et al., and et al., 2008). In addition, a number of genes changes of expression and/or contains auxin-responsive elements in the promoter et al., et al., them for ARF In of of specific auxin-dependent events that depend on the activity of ARFs have but the first important are being The developmental pathway during is the one that to the establishment of the root via the proper specification of the hypophysis. The of and in this has been demonstrated (reviewed in De and 2007). both and are not in the but above that in the embryo it has been that a MP-dependent signal be to the hypophysis from the above be it auxin or Because auxin does not the root of the mutant, it is that an signal is (Weijers et al., 2006). It is to that the of which is to the phenotype and which could a auxin response, does not rise to developmental defects in the embryo but to (Weijers et al., 2006). The for this is but a is that specific as for the MYB77, are in combination with to the auxin responses (Figure has been in the of the downstream events of the auxin pathway during lateral root and In the auxin responses can be observed during of pericycle cells for lateral root in the The of which does not with the with the of lateral The of the auxin response may be caused least in by in auxin within the root et al., 2007). from these of the predominant expression patterns the root those of several genes of the auxin pathway, showed expression changes that developmental et al., 2007). ARF7 and ARF19, which appear to be able to interact with and et al., et al., 2005; et al., 2005; et al., are to the expression of DOMAIN and and thereby promote the of lateral ARF7 in vitro to promoter auxin-responsive elements (Figure et al., 2007). These from Arabidopsis the finding of a similar in et al., 2005). Another possible for ARF transcription factors during lateral root development is which contains relevant auxin-responsive elements in its promoter encodes a putative PROTEIN transcription and is involved in the regulation of lateral root via control of the pattern of cell during the of most in combination with other factors et al., 2007). events within the root can be as a for the of the different components of the auxin pathway. to a auxin et al., being on a of which is about by the proteins et al., 2007). The of proteins can be as of the auxin since the expression of genes on auxin and response and at least on and protein levels are eventually into distinct responses. The different levels appear to processes as diverse as the of cell the activity of cell and cell et al., 2007). This a of auxin in with the However, the relevant mechanisms this of in Arabidopsis. The scheme the putative pathways some additional branches important and and the that are in the The pathway is most specific and from via to there are several that are not The and have long been implicated in the biosynthesis of auxin, into (Figure reviewed in Woodward and Bartel, 2005). Recent have been with respect to the of into by the family of et al., 2007; et al., and of into by and its four homologs to et al., et al., (Figure The of the the same combinations of the to more specific the of the differential expression of the various different combinations to a auxin response only in those these are et al., 2006). The developmental that was eventually giving rise to an is by mutants et al., 2007). These four are all in the and and expression in the apical of the embryo et al., the that has long been to be the major of auxin during the and its homologs that have been investigated are also in cell For example, in the expression of is in the is in the out both genes to the of the root cells and eventually to root In embryogenesis, is and from the in the apical and root The importance of this expression of and most also its homologs and for the of auxin during can be from the auxin-related phenotype in the triple et al., 2008). the and triple mutants the of mutants in the auxin pathway that an development and to a the et al., et al., (Hamann et al., and mutants (Hardtke and Berleth, 1998). The of all these mutants the importance of each of the auxin pathway for a proper Therefore, both and auxin biosynthesis contribute to the of auxin during there is no genetic between the two it is unclear mutants in the and the auxin biosynthesis pathways display the same phenotype as mutants in auxin transport, and It might be that both pathways are thereby leading to especially low auxin levels when one pathway is combinations between both pathways and/or analysis of activity when one pathway is out could an to this intriguing Although recent have indicated that auxin transport is to and auxin maxima and gradients et al., these novel suggest the of auxin biosynthesis for and/or auxin In to the of important auxin biosynthesis these findings the role of the spatio-temporal regulation of auxin biosynthesis gene expression in Arabidopsis. the molecule auxin to be the of plant development and of physiological processes. The of auxin and the occurrence of auxin maxima and gradients within the plant appear to be the result of the of auxin biosynthesis and The subsequent different auxin responses depend on a regulation of gene expression, transcript and of proteins of the key components of the auxin signaling all these regulatory processes are controlled by the plant to specific is being revealed by The manner in which cells an auxin-mediated developmental might be revealed and several recently key such as those involved in histone are the for this the was recently suggested to auxin response in an novel way et al., thereby up a new of auxin It appears that some of to auxin in the lineage leading to land plants and several in other However, the biological of these and whether, for example, red and brown algae have auxin and signaling mechanisms as those present in members of the land plant lineage are knowledge of plant and the of new that specifically with auxin responses will most more on these in the three and for and This was by the and the