Introduction

Auxin is a critical plant hormone that modulates diverse growth and developmental processes such as tropic responses to light and gravity, root and shoot architecture, organ patterning, vascular development, as well as growth and differentiation in tissue culture (Hagen and Guilfoyle 2002; Woodward and Bartel 2005). Over the past 20 years, auxin has been discovered to exert rapid and specific regulation on the expression of auxin-inducible genes at transcriptional level. Those genes that specifically induced by auxin within minutes of exposure to the hormone have been extensively studied and regarded as primary/early auxin-responsive genes. These genes, including Aux/IAAs, GH3s and SAURs, share auxin response elements (AuxREs) in their promoter regions (Abel et al. 1994; Hagen and Guilfoyle 1985; Walker and Key 1982; Woodward and Bartel 2005). The first member of GH3 gene family was isolated from soybean and accumulation of GH3 transcript is induced by auxin (Hagen and Guilfoyle 1985). There are increasing evidences indicating that GH3 family members function in modulating the level of free auxin, JA as well as SA via amino acid conjugation and light signaling may be involved in this process (Hsieh et al. 2000; Woodward and Bartel 2005). The present review mainly focuses on the distribution, the promoter characteristics, biochemical function and biological functions of the GH3 family in plant.

Distribution of GH3 genes in plant

In the past 20 years, GH3 genes were identified in angiosperms (including both dicotyledons and monocots), gymnospermae, and moss. The first identified GH3 transcript was derived from auxin-treated soybean seedlings (Hagen and Guilfoyle 1985). The soybean GH3 genes was shown to be specifically induced by exogenous auxin treatment within 5 min, and this induction was not affected by treatment with the protein synthesis inhibitor cycloheximide, suggesting that the induction does not need de novo synthesis of protein (Franco et al. 1990; Hagen and Guilfoyle 1985). The Arabidopsis thaliana GH3 gene family consists of 19 members and an additional partial gene encoding only the amino-terminal residues of the protein. Members of Arabidopsis thaliana GH3 gene family located on chromosomes 1, 2, 4 and 5, but not on chromosome 3 (Table 1). Based on the phylogenetic analysis along with the substrate specificities, the GH3 family in Arabidopsis has been classified into three groups and there are two, eight and ten members in Group I–III, respectively (Table 1; Staswick et al. 2002). Group I includes AtGH3-11/FIN219/JAR1 and AtGH3-10/DFL2 (Hsieh et al. 2000; Staswick et al. 2002; Takase et al. 2003). Group II consists of AtGH3-2/YDK1, AtGH3-5/AtGH3a/WES1, AtGH3-9, AtGH3-6/DFL1 and four other members (Nakazawa et al. 2001; Takase et al. 2004; Tanaka et al. 2002). Group III is composed of AtGH3-12/PBS3/GDG1 and nine other members (Jagadeeswaran et al. 2007; Nobuta et al. 2007).

Table 1 Arabidopsis GH3 family members
Full size table

In addition to soybean and Arabidopsis, GH3-like genes were also found in other dicotyledon species. Roux and Perrot-Rechenmann (1997) first isolated a GH3-like gene from Nicotiana tabacum, which was designated as Nt-gh3 sharing 70% identity with the soybean GH3. Since then, Lahey et al. (2004) detected transcript of GH3-like protein in Citrus madurensis. Liu et al. (2005) also identified a GH3-like gene from pungent pepper (Capsicum chinense L.), whose predicted protein shares 95% identity with Nt-gh3.

GH3-like genes were also identified in monocots. Using genomic approaches and gene expression analysis, 13 GH3-like ORFs (including 12 active genes) were identified in rice (Oryza sativa). Unlike AtGH3s, however, OsGH3s are only able to be subdivided into two Groups (Group I, II) based on sequence similarity to Arabidopsis GH3 (Jain et al. 2006; Terol et al. 2006). Group III GH3-like genes have not been found in rice. Furthermore, an extensive survey of the EST database of other monocots, including wheat (Triticum aestivum), corn (Zea mays), Sorghum bicolor, sugarcane (Saccharum officinarum), and barley (Hordeum vulgare), indicates that Group III of GH3 is absence in monocots (Jain et al. 2006).

In the gymnospermae, Pinus pinaster, and the moss, Physcomitrella patens, GH3-like genes were also identified and/or characterized (Bierfreund et al. 2004; Reddy et al. 2006). To our knowledge, Pp-GH3.16 showed the highest homology to Group II GH3 members of Arabidopsis, and was the only GH3 member studied in Pinus pinaste until now. In Physcomitrella patens, there were three GH3-like genes: PpGH3-1, PpGH3-2 (belong to Group I) and a truncated gene. Phylogenetic analyses indicated that GH3 proteins are highly conserved all over the plant kingdom (Hagen and Guilfoyle 2002; Terol et al. 2006).

Promoter characteristics of GH3 genes

The GH3 gene of soybean is the first auxin primary/early response gene identified and its expression could be induced by exogenous auxin treatment within 5 min (Hagen and Guilfoyle 1985; Woodward and Bartel 2005). To address how GH3 can be induced by auxin treatment, the promoter of GH3 was analyzed in detail through gel mobility shift assays, methylation interference, deletion analysis, linker scanning, site-directed mutagenesis, and gain-of-function analysis. These analyses identified that the sequence TGTCTC is the core sequence of auxin response element (AuxRE) in GH3 promoter (Liu et al. 1994; Ulmasov et al. 1995). This type of AuxRE and/or its variants were also found to be present in the promoters of other auxin-responsive genes. Furthermore, ARFs are able to specifically bind to the AuxREs to repress or activate expression of these genes (Ulmasov et al. 1995, 1997a, 1997b).

The auxin-responsive ability of the TGTCTC element can be enhanced by the combination with an adjacent or overlapping coupling element (such as CACGCAAT, CCTCGTGtctc). Through combination with different coupling elements, the simple AuxRE, TGTCTC, can make up three kinds of composite AuxREs. Although each shows auxin-inducible activity independently, they can contribute incrementally to the overall level of auxin induction (Liu et al. 1994; Ulmasov et al. 1995). In addition, the simple AuxRE without coupling elements may also function strongly in expression of auxin-inducible genes if the TGTCTC elements occur as tandem direct or palindromic repeats, although this scenario has not yet been found occurring naturally. P3 (4×) consisting of four palindromic repeats spaced by 3 bp is more active than natural AuxREs in response to auxin treatment (Ulmasov et al. 1997a). DR5 (7×) element, which is composed of seven direct repeats of 11 bp fragment including the TGTCTC element, also shows greater auxin inducibility than a natural composite AuxRE and the GH3 promoter (Ulmasov et al. 1997b). Because of their higher auxin inducibility than identified natural promoters, the DR5 (7×) and P3 (4×) constructs have been used as valuable tools to study spatial-temporal expression patterns of auxin-responsive genes in the life cycle of plants (Bierfreund et al. 2003; Hagen et al. 1991; Li et al. 1999; Schwalm et al. 2003). Nevertheless, Goda et al. (2004) found recently that TGTCTC element was not enriched in genes specifically regulated by IAA, but was enriched in genes up-regulated by both brassinolide (BL) and IAA. Therefore, these constructs may be not specific to auxin action, but can also be used as important markers for studying the brassinosteroid (BR)/auxin interaction.

Besides AuxREs, ethylene responsive element (ATTTCAAA) has also been found in promoters of GH3 genes (Liu et al. 2005). Consistent with this, CcGH3 (GH3 in Capsicum chinense L.) was found to be regulated by both auxin and ethylene (Liu et al. 2005).

Biochemical function of the GH3 genes

By sequence analysis and three-dimensional prediction of proteins, GH3s were found to belong to acyl adenylate-forming firefly luciferase superfamily. AtGH3-11 (JAR1, Group I) is the first demonstrated GH3 that is able to specifically catalyze adenylation of JA in vitro. It was suggested that adenylation of JA might initiate conjugation of several amino acids to JA (Staswick et al. 2002). The conjugation of JA to isoleucine (Ile) mediates JA response in Arabidopsis, because exogenous JA-Ile was able to rescue the defect of jar1 in response to JA. Interestingly, the level of JA-ACC conjugates was shown to be up-regulated in jar1-1, suggesting that JAR1 may also play a role in crosstalk between JA and ethylene signaling (Staswick and Tiryaki 2004).

All of the Group II members (except for GH3-1, which needs to be studied) can catalyze adenylation of IAA in vitro (Table 1; Staswick et al. 2002, 2005). This modification converts the free IAA to a conjugated form. Via in vitro studies, IAA-Asp has been found to be the major conjugate when treated with IAA (Staswick et al. 2005). Besides IAA, AtGH3-5 (AtGH3a/WES1, Group II) can also catalyze adenylation SA in vitro (Staswick et al. 2002), indicating that AtGH3-5 is involved in crosstalk between IAA and SA signaling.

Biological functions of the GH3 genes

Using in situ hybridization and analysis of P GH3 :GUS expression, soybean GH3 was found to be expressed in the inner cortex and protoxylem ridges of roots. In addition, it is also transiently expressed during flower and pod development. When treated with 2,4-D (2,4-dichlorophenoxyacetic acid), GH3 transcripts became more abundant in the vascular regions of all organs studied. Furthermore, a high level of GH3 mRNA was also detected in developing palisade mesophyll cells of leaves, cotyledons, and flowers (Gee et al. 1991; Guilfoyle et al. 1993). All of these indicate that GH3 may be involved in auxin-regulated growth and development.

A number of gh3 mutants have been isolated and characterized in Arabidopsis, and the biological function and significance of GH3s in plant growth and development have been clarified. So far, mutants of seven Arabidopsis GH3 genes were identified via morphological screening and they displayed distinct but interrelated phenotypes (Table 2; Hsieh et al. 2000; Khan and Stone 2007; Nakazawa et al. 2001; Staswick et al. 2002; Takase et al. 2004; Takase et al. 2003; Tanaka et al. 2002; Zhang et al. 2007). AtGH3-11, one Group I gene, is induced by auxin and encodes a protein that specifically adenylates JA but not IAA or other hormones in vitro (Staswick and Tiryaki 2004; Staswick et al. 2002). There are two interesting mutant alleles identified for this locus: jar1-1 and fin219. fin219 exhibited long hypocotyl only under continuous far-red light, suggesting that FIN219 mediated signal transduction of phytochrome A (phyA). Therefore, the FIN219 may be a cross-talk junction between auxin and phyA signaling (Hsieh et al. 2000). jar1, a null mutant of AtGH3-11, exhibited insensitivity to JA. However, fin219 did not show insensitivity to JA and jar1 did not display the specific far-red light long-hypocotyl phenotype (Staswick et al. 2002). The discrepancy of these mutant phenotypes is difficult to explain. Staswick et al. (2002) ascribed this to that fin219 was an epigenetic mutant whose nature needs to be further characterized. Recently, jar1-1 and another jar1 allele were found to exhibit much weaker specific long-hypocotyl phenotype than fin219 under weak continuous far-red light condition (Chen et al. 2007). This result confirms that FIN219/JAR1/AtGH3-11 mediates far-red light response. Unfortunately, the reasons for much weaker far-red light hyposensitivity of jar1 and the JA sensitivity of fin219 are still obscure. To address these questions, it is important to reveal the nature of fin219 mutant.

Table 2 Most notable phenotype (s) of mutants in Arabidopsis GH3 genes
Full size table

DFL2/AtGH3-10, another Group I gene, is a red light-induced gene and is involved in seedling photomorphogenesis. When dark-grown seedlings were exposed to red light, DFL2 expression was up-regulated and maintained for about two hours. Meanwhile, the hypocotyl length was dependent upon the expression level of DFL2 under red-light condition. All of these results suggest that DFL2 is involved in red light signal transduction. Unlike other characterized GH3, DFL2 expression is not induced by exogenous auxin although there are putative AuxREs in the promoter of DFL2 (Takase et al. 2003).

In contrast to DFL2, all other characterized Group II members are auxin up- or down-regulated genes and some of them are also regulated by light. AtGH3a (AtGH3-5, Group II) expression is induced by auxin and by end-of-day far-red light treatment. Furthermore, AtGH3a (WES1, Group II) is involved in the shade-avoidance responses and acts downstream of phytochrome B (Park et al. 2007b; Tanaka et al. 2002). GH3-5 displays adenylation activity not only on IAA but also on SA in vitro, and gh3 activation-tagged mutants (wes1-D and gh3.5-1D) show enhanced auxin resistance and stress adaptation (Park et al. 2007a; Staswick et al. 2002; Zhang et al. 2007). Because some GH3s of Group II act may redundantly, their functions were generally identified through phenotypic characterization of over-expression mutants, such as ydk1-D and dfl1-D. Both YDK1 (AtGH3-2, Group II) and DFL1 (AtGH3-6, Group II) are induced by auxin. ydk1-D displayed a short-hypocotyl phenotype in dark- and light-grown seedlings, but dfl1-D displayed short-hypocotyl only under light conditions. Further analyses on YDK1 and DFL1 expression have shown that YDK1 is inhibited by blue and far-red light, but DFL1 is not influenced by light. Therefore DFL1 protein should function with one or more light-induced partner (s) in regulating hypocotyls elongation (Nakazawa et al. 2001; Takase et al. 2004).

Recently, the function of AtGH3-9 (Group II) is studied through characterization of gh3.9-1 and its RNAi lines. Unlike most other Group II genes, AtGH3-9 expression is down-regulated by low concentrations of exogenous IAA in seedlings. Similar to jar1-1, gh3.9-1 shows moderately JA resistance, indicating that AtGH3.9 is likely to be a juncture between auxin and JA response pathway (Khan and Stone 2007). Since most members of group II proteins including AtGH3-9 have an enzymatic activity for adenylation of IAA in vitro, they may function in auxin homeostasis by reducing the availability of free auxin and they may function redundantly (Khan and Stone 2007; Staswick et al. 2002, 2005).

Our knowledge for biological roles of Group III genes still remains rudimentary (Woodward and Bartel 2005). Two latest studies suggest that AtGH3-12/PBS3/GDG1 (Group III) plays an important role in the metabolism and signal transduction of SA, which may increase stress adaptation of plants. Synthesis of SA in plant can be induced by both abiotic stress and biologic stress. The damages of the stress on plants can be alleviated by SA (Fujita et al. 2006). Salicylic acid-2-O-β-glucoside (SAG) is the conjugated and primary storage form of SA (Dean et al. 2005). Comparing with the wild type, pbs3-1 and gdg1-1, two loss-of-function mutants of AtGH3-12, exhibited lower-level SAG in the process of pathogen infection. Meanwhile, expression of the SA-dependent pathogenesis related marker, PR1 (pathogenesis-related protein 1, a key component in SA signaling), was down-regulated in this process, and exogenous SA application was able to restore PR1 expression and resistance to pathogens in these mutants (Jagadeeswaran et al. 2007; Nobuta et al. 2007). Surprisingly, Jagadeeswaran et al. (2007) reported that free SA level was decreased in gdg-1 while Nobuta et al. (2007) found that free SA level was elevated in pbs3-1 and pbs3-2. The reason for this contrary result may be the different nature of mutant alleles and/or the different conditions for SA analysis. Further work, such as measuring the SA contents of pbs3 and gdg1 under the same condition, will be useful to resolve this discrepancy.

So far, only the function of OsGH3.8 has been revealed (Ding et al. 2008), although there are at least 12 members of GH3 family in rice (Jain et al. 2006). Similar to group II GH3 proteins in Arabidopsis, OsGH3.8 (group II) is an IAA-amino synthetase which prevents free IAA accumulation. The overexpression line of GH3-8 displayed enhanced resistance to the rice pathogen Xanthomonas oryzae pv oryzae and abnormal plant morphology and retarded growth and development. The mechanism underlining both abnormal development and enhanced resistance may be the inhibition of the expression of expansins, proteins that control cell wall loosening and expansion, by preventing the accumulation of free IAA (Ding et al. 2008). This discovery is helpful to understand the interaction of plant defense systems and auxin signaling.

Relationship between GH3s and ARFs

Auxin response factors (ARFs) can bind specifically to the AuxREs within promoters of early/primary auxin response genes and regulate their expression (Ulmasov et al. 1997a; Woodward and Bartel 2005). There are 23 ARF genes in Arabidopsis, and some of which may bind to GH3’s promoters to regulate gene expression. ARF8 was the first ARF that was demonstrated to regulate expression of three AtGH3 genes, AtGH3a, DFL1, and YDK1 (Tian et al. 2004). These genes were down-regulated in arf8-1 mutant and up-regulated in ARF8 overexpression lines. Although free auxin level was not remarkably elevated in arf8-1, it was indeed decreased in ARF8 overexpression lines. These results suggest that ARF8 might positively regulate expression of GH3, which resulted in adenylating IAA to form IAA-AA. This might be one of the pathways in maintaining auxins homeostasis in vivo (Tian et al. 2004). Recently, Yang et al. (2006) found the microRNA167-ARF8-GH3-IAA pathway in rice. Through analysis of ARF17 overexpression lines by expressing a microRNA160-resistent ARF17 mRNA, it was found that ARF17 could negatively regulate expression of GH3-5 and DFL1, but positively regulate gene expression of GH3-2 and YDK1 (Mallory et al. 2005). In contrast to ARF17, a mutation in ARF7 causes reduced expression of some GH3 genes including YDK1, suggesting that ARF7 also positively regulates YDK1 expression (Stowe-Evans et al. 1998; Takase et al. 2004).

Conclusions

Our understanding of biological functions of GH3 genes has advanced rapidly in recent years. Together with the progresses in functional studies on ARF, Aux/IAA and SAUR genes, an outline of primary/early auxin response pathways was revealed. Phenotype characterizations of the mutant lines indicated that GH3s are involved in different growth and developmental processes. These analyses also provided evidences to identify the biochemical function of GH3. Although GH3 transcripts were first identified from auxin-treated seedlings of soybean, the finding that GH3s belong to acyl adenylate-forming firefly luciferase superfamily and can catalyze adenylation of IAA, JA and SA suggests that GH3s play a role not only in auxin signaling but also in other signal transduction pathways. Some GH3 genes are also regulated by light. Hence GH3 proteins may be key linkers among different signal transduction pathways, although their physiological functions need to be further studied. However our knowledge about GH3s mainly focus on group I and II, the little is known about the group III, especially in terms of their biochemical activities. Further understanding on the function of the GH3s will help to elucidate the complex signal transduction network in plants.