Alternative splicing of the auxin biosynthesis gene YUCCA4 determines its subcellular compartmentation


(fax +44 114 2253064; e-mail


Auxin is a major growth hormone in plants, and recent studies have elucidated many of the molecular mechanisms underlying its action, including transport, perception and signal transduction. However, major gaps remain in our knowledge of auxin biosynthetic control, partly due to the complexity and probable redundancy of multiple pathways that involve the YUCCA family of flavin-dependent mono-oxygenases. This study reveals the differential localization of YUCCA4 alternative splice variants to the endoplasmic reticulum and the cytosol, which depends on tissue-specific splicing. One isoform is restricted to flowers, and is anchored to the cytosolic face of the endoplasmic reticulum membrane via a hydrophobic C-terminal transmembrane domain. The other isoform is present in all tissues and is distributed throughout the cytosol. These findings are consistent with previous observations of yucca4 phenotypes in flowers, and suggest a role for intracellular compartmentation in auxin biosynthesis.


Auxin is an important plant growth hormone, with numerous applications in agriculture and horticulture. Auxin is synthesized in various plant tissues and may be transported long-distance as well as across membranes from cell to cell (Kramer and Bennett, 2006). The most abundant auxin, indole-3-acetic acid (IAA), affects all aspects of plant development and is involved in numerous plant processes, such as apical–basal polarity (Friml et al., 2003), root formation (Benkováet al., 2003), embryo and gynoecium development (Sohlberg et al., 2006), and various tropisms.

The spatial distribution of auxin at tissue level coordinates developmental processes via auxin gradients, which are generated by localized auxin biosynthesis and transport in the vascular system and across cells (Kramer and Bennett, 2006). This results in dose-dependent gradient responses (Bhalerao and Bennett, 2003) or receptor responses via local accumulation (Dharmasiri et al., 2005). Much is known regarding intercellular auxin transport by influx and efflux carriers and auxin perception, as well as factors controlling IAA homeostasis, such as conjugation, hydrolysis of conjugates, transport and biosynthetic pathways. However, the role of each of these processes in auxin signalling remains to be resolved (Woodward and Bartel, 2005; Chandler, 2009).

Biosynthesis of auxin is particularly complex due to the existence of multiple pathways. Depending on the organ, developmental stage or environment (Normanly and Bartel, 1999; Östin et al., 1999), parallel tryptophan-dependent (Figure 1) and -independent pathways (Woodward and Bartel, 2005; Kriechbaumer et al., 2006) may be differentially regulated, creating a metabolic network that changes dynamically to maintain homeostasis or supply IAA for local demands. To understand the mechanism of local auxin action and gradient formation, it is necessary to determine how auxin biosynthesis is controlled and where auxin synthesis occurs.

Figure 1.

 Proposed tryptophan-dependent auxin biosynthetic pathways in Arabidopsis thaliana.
Known enzymes are indicated.

The tryptophan-dependent ‘YUCCA’ route (Figure 1) was characterized using via auxin biosynthetic mutants, and involves a family of 11 YUCCA genes, which encode flavin-dependent mono-oxygenases (FMOs). Their gene expression was shown to be temporally and spatially regulated, with distinct, but also overlapping, expression patterns (Cheng et al., 2006). YUCCA over-expression results in elevated endogenous levels of free IAA (Zhao et al., 2001). Different yucca mutant combinations yield different phenotypes; for instance, the yucca1/4 double mutant shows defects in floral organs, whereas yucca2/6 shows altered stamen development (Cheng et al., 2006). These phenotypes are suggested to be due to changes in auxin levels, as YUCCA over-expression results in elevated endogenous levels of free IAA (Zhao et al., 2001), and yucca mutants display reduced GUS staining under the control of the auxin-induced promoter DR5 (Cheng et al., 2006). Furthermore, the flower phenotype in yucca mutants can be rescued by over-expression of the bacterial auxin biosynthetic gene iaaM under the control of a YUCCA promoter (Chen et al., 2006).

Despite the unique expression patterns of the various YUCCA proteins, functional overlap between YUCCA proteins has been suggested because yucca triple and quadruple mutants show increasingly strong effects, but are not embryo- or seedling-lethal. Single loss-of-function mutants have no altered phenotype, but multiple mutations in four of the 11 YUCCA genes in Arabidopsis thaliana resulted in locally reduced auxin concentrations and defects in development, height, apical dominance and fertility (Cheng et al., 2006). Over-expression of YUCCA1 (Zhao et al., 2001) and YUCCA6 (Kim et al., 2007) induced phenotypes that are characteristic of auxin over-production, and over-expression of YUCCA1 enhances auxin biosynthesis in a tissue-specific manner, suggesting a role in local auxin biosynthesis (Zhao et al., 2001). YUCCA genes have also been identified in other plant genomes such as rice (Oryza sativa) (Yamamoto et al., 2007), maize (Zea mays) (Gallavotti et al., 2008), Medicago truncatula (Tivendale et al., 2010), tomato (Solanum lycopersicum) (Expósito-Rodríguez et al., 2007) and Petunia hybrida (floozy, Tobena-Santamaria et al., 2002). Generally, this enzyme family exhibits low substrate specificity, and it is possible that efficient substrate channelling between co-localized enzymes mitigates this problem (Müller and Weiler, 2000; Kriechbaumer et al., 2006).

Despite the finding that yucca mutants display strong auxin phenotypes, there is still controversy regarding the suggested enzymatic conversion. YUCCA1 proteins were suggested to use tryptamine (TAM) as a substrate, converting it into N-hydroxylated tryptamine (Zhao et al., 2001), but Tivendale et al. (2010) showed that the published MS spectra for product formation are inconsistent with those for synthetic N-hydroxylated tryptamine. It may be that, due to the broad in vitro substrate spectrum of FMOs, tryptamine is not the primary substrate for YUCCA1. Moreover, it has been suggested that YUCCA proteins may act in the same pathway as tryptophan aminotransferase of Arabidopsis (TAA1): the indole-3-pyruvic acid (IPA) pathway (Strader and Bartel, 2008). More recent findings place the YUCCA proteins downstream of TAA1, catalysing the conversion of IPA to IAA (Mashiguchi et al., 2011; Won et al., 2011): over-expression of YUCCA1 partially suppresses shade avoidance and milder sterile phenotypes of taa1 mutants, active TAA1 genes are required for auxin over-production phenotypes in YUCCA1 over-expression lines (Won et al., 2011), multiple yucca mutants accumulate IPA but not TAM, and YUCCA2 is capable of converting IPA to IAA in vitro (Mashiguchi et al., 2011).

One of the YUCCA genes, YUCCA4 (At5g11320), appears to play a specialized role in flowers, because in situ hybridization demonstrated YUCCA4 expression in apical and flower meristems and young floral primordia. GUS staining was observed at the base of floral organs in young flower buds and in late stages of flowers, predominantly in the apical regions of the carpel, stamens and sepals, as well as in young leaves. Mutation of both YUCCA1 and YUCCA4 affects the number and identity of flower organs (Cheng et al., 2006).

YUCCA4 has been shown to be targeted by several transcription factors, resulting in increased auxin biosynthesis and IAA levels. STYLISH1 protein results in increased transcription levels of YUCCA4 as well as increased IAA levels and auxin biosynthetic rates in Arabidopsis seedlings (Sohlberg et al., 2006; Ståldal et al., 2008). The NGA and STY factors act cooperatively to promote style specification, in part by directing YUCCA-mediated auxin synthesis in the apical gynoecium (Trigueros et al., 2009), and YUCCA4 also appears to be a direct transcriptional target of LEC2 (Stone et al., 2008).

Here, we show that alternative splicing of YUCCA4 generates two major protein isoforms, both of which possess a predicted catalytic FMO domain. YUCCA4.2 also features an additional C-terminal transmembrane domain (TMD) that causes localization to the membrane of the endoplasmic reticulum (ER) (Figure 2). Furthermore, tissue-specific splicing raises the interesting possibility that auxin synthesis is regulated at the tissue level by intracellular compartmentalization.

Figure 2.

 Intron–exon structure of YUCCA4 splice variants.
Intron sequences are shown in red; untranslated regions are shown in white. Exons, the predicted FMO domain and the TMD are indicated.


YUCCA4 occurs as two isoforms

Database annotation (TAIR,, and NCBI, indicates that YUCCA4 generates two splice isoforms. YUCCA4.1 consists of four exons encoding an FMO domain, which includes two predicted NADH/FAD binding sites. Alternative splicing causes 72 bp of the third intron to replace the final exon in YUCCA4.2, introducing a predicted (Kriechbaumer et al., 2009) C-terminal hydrophobic transmembrane domain of 24 amino acids. The lack of an N-terminal signal sequence predicts that YUCCA4.2 is a tail-anchor protein, in which the TMD is inserted into the membrane with the protein facing the cytosol.

YUCCA4 transcription is regulated by tissue-specific splicing

To analyse the expression patterns of splice versions of YUCCA4, RT-PCR was performed for various plant tissues, and the resulting PCR products were identified by sequencing. To allow screening for full-length transcripts, cDNAs were created using primers specific for YUCCA4.1 and YUCCA4.2. Primer combinations for full-length YUCCA4.1 resulted in PCR products of the expected size in all tissues tested (Figure 3a). Interestingly, in flower tissue, an additional 2 kb band was detected that was identified by sequencing as unspliced YUCCA4.1 (Figure 3a). YUCCA4.2 mRNA was only detected in flowers, and was present in both spliced and unspliced forms (Figure 3b). Use of primer combinations overlapping the common exons 1 and 2 confirmed the presence of unspliced transcript in flowers, showing that the first intron does not splice in a significant proportion of the RNA population (Figure 3c). The significance of unspliced forms of YUCCA4 is unclear, but the translated product would only possess part of the FMO domain, and is therefore not expected to be catalytically active.

Figure 3.

 Tissue distribution of the YUCCA4 splice isoforms.
(a, b) RT-PCR in various tissues using cDNAs created using gene-specific primers for full-length YUCCA4.1 (a) and YUCCA4.2 (b).
(c) Test for the presence of spliced and unspliced variants in flower tissue using a primer combination overlapping exons 1 and 2, which are present in both YUCCA4 isoforms.

These data show that YUCCA4 exists in two major splice variants, with YUCCA4.2 being specifically found in flowers.

Intracellular localization of YUCCA proteins

To determine the intracellular localization of YUCCA4, cell-free targeting assays were performed. YUCCA4.2 (Figure 4, lane 2) is inserted into ER membranes in a carbonate-resistant manner, but YUCCA4.1 is not inserted (Figure 4, lane 6). Treatment of YUCCA4.2 in the ER membrane with the protease thermolysin resulted in protein degradation (Figure 4, lane 3), indicating that the FMO domain is not translocated to the ER lumen, and can therefore be expected to support cytosolic auxin biosynthesis. As the YUCCA4.2 TMD was predicted to cause chloroplast localization (Kriechbaumer et al., 2009), YUCCA4.2 import into chloroplasts was tested, but no such import was observed (Figure 4, lane 4). The functionality and integrity of the ER and chloroplasts were determined using the ER lumenal protein preprolactin and the chloroplast stromal protein small Rubisco subunit. Preprolactin imported into the ER is cleaved to prolactin (Figure 4, lane 7), which is resistant to protease treatment (Figure 4, lane 8), and the stromal protein small Rubisco subunit is imported into chloroplasts (Figure 4, lane 9) and is thermolysin resistant (Figure 4, lane 10). We conclude that YUCCA4.2 is inserted into ER membranes, but YUCCA4.1 is not.

Figure 4.

 YUCCA4.2, but not YUCCA4.1, inserts into ER membranes.
Radiolabelled YUCCA4.2 and YUCCA4.1 protein was incubated co-translationally with ER microsomes, the ER fraction was re-purified, treated with thermolysin (Th) where indicated, and proteins were analysed via SDS–PAGE and phosphor imaging (lanes 2 and 6). Protein inputs are shown in lanes 1 and 5. Radiolabelled YUCCA4.2 protein was also incubated with purified chloroplasts, the chloroplasts were re-pelleted, and the pellet was analysed for the presence of YUCCA4.2 (lane 4). The ER-localized preprolactin/prolactin protein (PPL/PL, lanes 7 and 8) and the plastidial Rubisco subunit (lanes 9 and 10) are included as controls for membrane integrity.

To further investigate the localization under in vivo conditions, YFP-tagged YUCCA4 variants were expressed transiently in leaves of Nicotiana tabacum, and their localization was determined by confocal microscopy. YFP fluorescent tags were fused to the N-termini of the YUCCA4 isoforms to prevent interference with the C-terminal targeting sequence of YUCCA4.2. YUCCA4.1 was found to be cytosolic (Figure 5a, panel 1), and also showed fluorescence in the nucleoplasm (Figure 5a, panel 1, inset), which is common for cytosolic proteins (Brandizzi et al., 2002), but it did not co-localize with calnexin, which is found in the ER and the nuclear envelope (Figure 5a, panel 1, inset). YUCCA4.2 co-localized with calnexin (Figure 5a, panel 2), consistent with its ability to insert into ER membranes in cell-free assays. The punctate structures seen for YUCCA4.2 over-expression (Figure 5a, panels 2 and 3) co-localized with the Golgi marker sialyl transferase signal anchor sequence (ST–CFP, Figure 5a, panel 3), and treatment with brefeldin A redistributed the Golgi labelling of YFP-YUCCA4.2 to the ER (data not shown). Furthermore, ER localization was also observed when only the C-terminus of YUCCA4.2 was fused to a fluorescent YFP tag (Figure 5a, panel 4), showing that the tail anchor is sufficient for targeting. Overall, YUCCA4.2 is mostly resident in the ER, and has the ability to move into the Golgi, although this may only occur at high expression levels.

Figure 5.

 Dual targeting of YUCCA4.1 and YUCCA4.2 during Agrobacterium-mediated transient expression in Nicotiana tabacum.
(a) Confocal imaging of YFP-tagged YUCCA4.1 (row 1) and YUCCA4.2 (row 2) merged with the GFP-tagged ER marker calnexin. Fluorescence in the nucleoplasm is also shown for YUCCA4.1 compared with calnexin staining of the nuclear envelope (inset). Co-localization between YUCCA4.2 and the CFP-tagged Golgi marker sialyl transferase (ST–CFP) is shown in the panels in row 3. Row 4 shows confocal imaging for the C-terminus of YUCCA4.2 fused to a YFP tag and merged with the GFP-tagged ER marker calnexin. Scale bars = 5 μm.
(b) Ratiometric redox assays for roGFP2-tagged YUCCA4.2 (row 1). The ratio of 405 and 488 nm excitation is determined by the redox state of roGFP2: blue represents a cytosolic environment and red represents an ER luminal environment. Cytosolic roGFP2 (row 2) and the ER marker roGFP2–HDEL (3) were used as controls for localization in the cytosol or the ER lumen, respectively. The mean 405 nm/488 nm ratio is shown on the right for each construct. Scale bars = 5 μm.

To differentiate between residence in the ER lumen and the cytosol, a redox-sensitive fluorescent tag (roGFP2, Brach et al., 2009) was used. This GFP form enables ratiometric quantification of the redox potential based on the formation of disulfide bonds between surface-exposed cysteine residues, which affects the ratio of excitement by wavelengths of 405 or 488 nm. A high fluorescence ratio at 405/488 nm (Figure 5b, pseudo-coloured in red) indicates that the protein is located in the oxidizing environment of the ER lumen, whereas a low ratio (pseudo-coloured in blue) indicates that it is located in the more reduced environment of the cytosol. roGFP2–YUCCA4.2 has an excitement ratio of 0.044 (Figure 5b, panel 1), which is much closer to that of cytosolic roGFP2 (0.438) (Figure 5b, panel 2) than that of lumenal roGFP2–HDEL (4.31) (Figure 5b, panel 3), indicating that the N-terminus of YUCCA4.2 faces the cytosol, which is consistent with the results of the cell-free targeting assay. The lower excitement ratio at 405/488 nm of roGFP2–YUCCA4.2 compared to cytosolic roGFP may be due to the local environment of the ER membrane surface, or may be caused by enzymatic activity of YUCCA4 itself: links between the cellular redox state and auxin signalling and transport have previously been noted (Bashandy et al., 2011).

Taken together, the results of organellar targeting assays and in vivo localization of fluorescent-tagged YUCCA4 proteins show that YUCCA4.1 is localized in the cytosol, whereas YUCCA4.2 is anchored into ER membranes by its TMD, with the N-terminal domain facing the cytosol.

In silico analysis of targeting features in the YUCCA family

To investigate the potential membrane localization of more YUCCA proteins, a bioinformatics search for N-terminal signal peptides and transmembrane domains using the prediction programs TargetP, SignalP, ChloroP and TMHMM on the ExPASy server was performed (Table 1). This analysis predicts the existence of N-terminal signal peptides for YUCCA6, which has already been shown to be non-cytosolic (Kim et al., 2007), and for YUCCA10 and YUCCA11. YUCCA7 possesses a chloroplast-targeting signal based on TargetP prediction, but this was not confirmed by the ChloroP prediction. Strong hydrophobic domains that could potentially be TMDs were detected in YUCCA4.2, YUCCA3 and YUCCA5 (Table 1). Although these findings require experimental testing, they suggest that several members of the YUCCA family may be membrane-bound, and therefore indicate a wider importance for enzyme localization in auxin biosynthesis.

Table 1.   Predicted signal peptides and TMDs for YUCCA family members using the prediction programs TargetP, SignalP, ChloroP and TMHMM on the ExPASy server, showing the residues involved
YUCCASignal peptide (TargetP)Secretory signal peptide (SignalP)Chloroplast targeting signal (ChloroP)Predicted TMD (TMHMM)
3N-terminal (amino acids 31–53)
4.2C-terminal (amino acids 334–356)
5Internal (amino acids 248–270)
6Signal peptide (amino acids 1–45)
7Chloroplast transit peptide (amino acids 1–8)
10Signal peptide (amino acids 1–17)
11Secretory signal peptide (amino acids 1–20)Signal peptide (amino acids 1–20)

Enzymatic activity of YUCCA4

To test the enzymatic activity of the two YUCCA4 isoforms, they were cloned into the expression vectors pET28b and pHisTrx to produce recombinant protein. Both YUCCA4 isoforms were either highly insoluble or were not expressed at all in Escherichia coli BL21(DE3) cells. Therefore, a large-scale cell-free system was used, which yielded protein suitable for enzymatic activity assays.

Despite the finding that yucca mutants display strong auxin phenotypes, there is continuing discussion concerning which catalytic step is performed by YUCCA proteins. YUCCA4 has greatest sequence similarity to YUCCA1, with the key predicted catalytic element being an FMO domain that features two predicted N-terminal NADH/FAD binding sites. It has been suggested that YUCCA1 converts TAM to N-hydroxylated tryptamine (Zhao et al., 2001), but published MS spectra are inconsistent with those for synthetic N-hydroxylated tryptamine (Tivendale et al., 2010). Recent data show that yucca mutants accumulate IPA, and in vitro data suggest that the YUCCA proteins convert IPA to IAA (Won et al., 2011; Mashiguchi et al., 2011).

Enzymatic assays were performed using 1 μg YUCCA4 protein and 2 mm of the substrates TAM, IPA or tryptophan (Trp). Conversions were compared to reactions that were immediately stopped by the addition of methanol and snap freezing (Figure 6, lanes 1 and 5), and to an unrelated protein synthesized by the same cell-free system (dihydrofolate reductase) (Figure 6, lanes 2 and 6). TLC analysis shows substantial conversion of IPA by both YUCCA4.1 and YUCCA4.2 compared to dihydrofolate reductase (Figure 6, lanes 2–4), but no activity with TAM (Figure 6, lanes 7 and 8) or Trp (data not shown). Background IAA formation in the absence of YUCCA4 (Figure 4, lanes 1 and 2) is due to spontaneous conversion of IPA to IAA at room temperature (Libbert et al., 1970; Mino, 1970). Indole-3-acetaldehyde (IAAld) and IAA cannot be distinguished by TLC, so either may be the final product. These results show that both isoforms of YUCCA4 have the potential to play a role in auxin biosynthesis by conversion of IPA to IAA or IAAld.

Figure 6.

 Enzymatic assays with YUCCA4.1 and YUCCA4.2 and the substrates IPA and TAM analysed by TLC.
Addition of protein (YUCCA 4.1, 4.2 or the unrelated protein dihydrofolate reductase, DHFR) and the duration of incubation (−, 0 min; +, 4 h) for the various assays is indicated. Percentage conversions normalized to YUCCA4.2 (100%) are indicated. The standards IPA, TAM, IAAld and IAA are shown in lanes 9–12.


YUCCA4: alternative tissue-specific splicing

We have shown that YUCCA4 exists in two major isoforms due to alternative splicing, with YUCCA4.1 present in all tissues and YUCCA4.2 restricted to flowers. Both isoforms have enzymatic activity, but YUCCA4.2 is attached to the ER membrane, which may confer properties relevant to auxin biosynthesis. The tissue-specific presence of YUCCA4.2 suggests that it plays a specialized role in auxin metabolism in flowers, which is supported by the previous observations that the most severe phenotypes of yucca1/4 double mutants are detected in flower organs, including severe sterility due to defects in all floral organs (Cheng et al., 2006). Furthermore, yucca1/2/6 triple mutants produce flowers, whereas yucca1/2/4/6 quadruple mutants do not (Mashiguchi et al., 2011). This severe phenotype could be due to the presence and localization of YUCCA4.2, or to the double gene doses of both YUCCA4 isoforms. YUCCA4.1 is also expressed in other plant tissues, but the observed phenotypes in yucca1/4 double knockouts, such as slightly decreased height and curvature of rosette leaves, are minor (Cheng et al., 2006). Other yucca mutant combinations retaining YUCCA4, such as yucca2/yucca6 knockouts, have similar height and leaf phenotypes, but do not show defects in all flower organs, and therefore have only reduced fertility (Cheng et al., 2006). Taken together, we suggest that YUCCA4 and YUCCA1 are the key YUCCA genes involved in flower development, and that the additional presence of YUCCA4.2 in flowers and the ER localization of YUCCA4.2 are important for its action.

A role for splicing in the diversity and organization of the plant proteome is becoming increasingly evident (Simpson et al., 2008; Filichkin et al., 2010; English et al., 2010). Approximately 20% of plant genes have one or more alternative splice isoform, showing that these events are not rare, although this level is only one-third of the rate of alternative splicing in humans (Barbazuk et al., 2011), and examples of functionality of splicing events are rare (Filichkin et al., 2010). In Arabidopsis and rice, intron retention is the most common form of alternative splicing, in contrast to humans, in whom intron retention is least common and exon skipping is most abundant (Barbazuk et al., 2011).

Tissue-specific alternative splicing was also reported for a family of splicing regulators, serine/arginine-rich proteins, which are part of the spliceosome and are important for constitutive splicing as well as alternative splicing. Fifteen genes produce more than 90 transcripts with varying tissue distribution and abundance, displaying all combinations from fully spliced to retention of all introns, and mostly leading to premature termination codons and loss of functional domains (Palusa et al., 2007). It has been shown that unspliced and truncated isoforms can act as dominant-negative regulators in the case of jasmonate-regulating JAZ proteins, in which alternative splicing generates C-terminally truncated protein versions that lack a domain necessary for protein interaction. Some of these truncated versions function as dominant repressors of jasmonate signalling, protecting the jasmonate response pathway from hyperactivation (Chung et al., 2010). Such a regulatory function is possible for the unspliced version of YUCCA4 in flower tissue, in which all isoforms are present and, due to most severe phenotypes in this tissue, appear to be most physiologically relevant.

Alternative splicing can also produce variable ratios of isoforms with different cellular locations, and may be influenced by environmental factors such as light (Mano et al., 1999): two splice isoforms of hydroxypyruvate reductase with different localizations (HPR1 in peroxisomes and HPR2 in the cytosol) are developmentally regulated and their abundance increases slowly during germination in darkness; under light conditions, the cell produces mainly cytosolic HPR2 (Mano et al., 1999). It is possible that YUCCA4 splicing is regulated by environmental inputs in addition to its tissue specificity, resulting in a different intracellular localization or production of inactive protein, thereby regulating enzymatic activity.

Intracellular localization

The membrane localization of YUCCA4.2 may be significant as it it present only in flowers, where the most severe phenotypes in the yucca1/4 mutant background occur. Many enzymes potentially involved in auxin biosynthesis have low substrate specificities and turnover rates, and therefore metabolic channelling in an ‘IAA synthase complex’ has been postulated (Müller and Weiler, 2000). However, purification attempts (Müller and Weiler, 2000; Kriechbaumer et al., 2006) and yeast two-hybrid approaches have not identified the proteins involved in such a complex. A possible explanation is that IAA biosynthesis occurs at membrane surfaces and is catalysed by membrane-anchored enzymes such as YUCCA4.2, which would impede detection by conventional approaches. A previous example of coordination of protein complexes by tail-anchored proteins at membranes is the protein translocases found in the ER, chloroplasts and mitochondria (Jarvis et al., 1998; Werhahn et al., 2001; Van den Berg et al., 2004).

This study revealed the localization of an auxin biosynthetic enzyme to a membrane and a YUCCA protein to a specific intracellular compartment; such a possibility was raised previously by Kim et al. (2007), who showed that a YUCCA6–GFP fusion protein in Arabidopsis protoplasts was localized in an unidentified intracellular compartment. However, these experiments were performed using a different expression system, and the results were inconclusive due to incomplete overlap of YUCCA6–GFP and RFP, and partial overlap with Golgi and ER markers (Kim et al., 2007). In silico analysis predicted an N-terminal signal peptide for YUCCA6 as well as for other YUCCA proteins (Table 1), indicating a broader significance for enzyme localization in auxin biosynthesis.

The ER localization of YUCCA4 may also be linked to other aspects of auxin biology, because the ER harbours factors for auxin metabolism and signalling; these include several deconjugating enzymes (Woodward and Bartel, 2005), the auxin-binding protein ABP1 (Chen et al., 2001) and the cytosol-to-ER auxin transporter proteins PIN5, PIN6 and PIN8 (Mravec et al., 2009). Double mutations in pin5 and yucca1 result in additive phenotypes (Mravec et al., 2009). Furthermore, as manipulation of PIN5 auxin transport results in less free IAA and increased IAA conjugation, subcellular compartmentalization of auxin metabolism is likely (Mravec et al., 2009). The auxin-binding protein ABP1 that is involved in signalling is mainly localized in the ER and the apoplast (Napier et al., 2002). Therefore, auxin action may not only be dependent on the overall auxin content or free and bound IAA, but also on its actual concentration in various subcellular compartments. IAA/amino acid hydrolases are predicted to be ER-localized, and have been shown to contribute to increasing IAA levels (Rampey et al., 2004), suggesting that the ER can also function as a source of IAA in some instances (Friml and Jones, 2010). A current model suggests a mechanism for cell-to-cell transport of auxin regulated by PIN5-dependent auxin transport between the ER and the cytoplasm (Wabnik et al., 2011). All ethylene receptors, as well as EIN2, the regulator of ethylene signalling, are present and interact at the ER membrane (Grefen et al., 2008; Bisson et al., 2009). Therefore, it has been suggested that intracellular localization of both ethylene and auxin signalling may be important for hormonal cross-talk (Friml and Jones, 2010).

The varying phenotypes for various combinations of yucca mutants indicate that, although auxin can be transported long-distance as well as directionally from cell to cell (Kramer and Bennett, 2006), the site of auxin biosynthesis is crucial for plant development. Localization of biosynthesis at the subcellular level adds a further layer of complexity and regulation.

Enzymatic activity

The ability of both YUCCA4 isoforms to use IPA as a substrate places YUCCA4 in the Trp-dependent IPA pathway, downstream of TAA1, which is proposed to convert Trp to IPA (Figure 1) (Stepanova et al., 2008). Due to the similarity in expression patterns and mutant embryo phenotypes of TAA1 (Stepanova et al., 2008) and YUCCA family members (Cheng et al., 2006), it has been suggested that the pathways that these enzymes are involved in may converge (Normanly, 2010). Two recent studies have shown that yucca mutants accumulate IPA, and that the auxin phenotypes in YUCCA over-expression lines are dependent on active TAA genes (Won et al., 2011). Moreover, TAA1 and YUCCA proteins synergistically enhance IAA levels in over-expression lines, and YUCCA2 is capable of converting IPA to IAA in vitro (Mashiguchi et al., 2011). IAAld is discussed as an intermediate, but levels of IAAld other than IAA or IPA are unaltered in auxin-overproducing or auxin-deficient mutant lines (Mashiguchi et al., 2011). Hence it has been suggested that YUCCA proteins directly convert IPA to IAA.

The lack of YUCCA4 activity with TAM is inconsistent with previous reports for YUCCA1 (Zhao et al., 2001). FMOs generally show low substrate specificity, and are able to oxidize a diverse range of substrates (Willetts et al., 1997), so efficient channelling by co-localization at the ER would allow more specific roles in auxin biosynthesis. Furthermore, in contrast to animal FMOs, plant FMOs have been shown to have very specific roles in various biosynthetic pathways. Phenotypes observed for YUCCA in Arabidopsis have also been observed for the YUCCA orthologue FLOOZY in petunia and OsYUCCA in rice. Over-expression of FLOOZY and OsYUCCA1 resulted in increased IAA levels, and OsYUCCA1 antisense plants showed defects similar to those of auxin-insensitive mutants (Tobena-Santamaria et al., 2002; Yamamoto et al., 2007). Other examples of FMO specificity include AtFMO1, which is involved in pathogen defence by influencing the redox state of the cell (Mishina and Zeier, 2006; Koch et al., 2006; Bartsch et al., 2006), and FMOGS-OX, which converts methylthioalkyl glucosinolates to methylsulfinylalkyl glucosinolates (Hansen et al., 2007). Therefore, specific roles for YUCCA enzymes in auxin biosynthesis are possible, especially if present in enzyme clusters.

In conclusion, we show that YUCCA4 expression is regulated by alternative splicing at a tissue level, resulting in a YUCCA4.2 isoform that is localized to the ER membrane. This may be important for regulating auxin signalling by subcellular compartmentation of IAA synthesis. Mutant complementation studies for both YUCCA4 isoforms in the yucca1/4 mutant background are necessary to determine which of the isoforms is responsible for flower phenotypes. Further localization and interaction studies with other YUCCA proteins, auxin biosynthetic enzymes (e.g. TAA1 and Arabidopsis aldehyde oxidase (AAO) in the IPA pathway) and auxin transport genes (e.g. PINOID, PIN5 and ABP1) correlated to YUCCA4 are required to understand the role of auxin compartmentation.

Experimental procedures

Plant material

Arabidopsis Col-0 plants obtained from the Nottingham Arabidopsis Stock Centre were grown on soil with 14 h light (100 μmol m−2 sec−1) at 21°C with 60% humidity. Adult plant tissue was harvested after 8 weeks. Nicotiana tabacum plants were kept in the greenhouse under 16 h light at 24°C and 8 h dark at 16°C.

RT-PCR analysis

RNA was isolated using TRIzol® reagent (Invitrogen, according to the manufacturer’s instructions, and the isolated RNA was treated for 15 min at room temperature with RNase-free DNase I (Invitrogen) at 0.2 U μg−1 RNA (Invitrogen). cDNAs were created using a ProtoScript® AMV first-strand cDNA synthesis kit (New England Biolabs, with gene-specific primers according to the manufacturer’s instructions. PCR was performed for 30 cycles with Phusion® high-fidelity DNA polymerase (New England Biolabs) according to the manufacturer’s instructions, with an annealing temperature of 55°C and primers Exon1for (5′-GCTCCTGGATCCGGTATGGGCACTTGTAGAGAATC-3′), Exon2rev (5′-CGTCACTTTAATTTGTCCTGATCGG-3′), Yuc4.1rev (5′-GGATTTATTGAAATGAAGATG-3′) and Yuc4.2rev (5′-TCACATATACATATACACATTGAC-3′).

To compare the distribution of all YUCCA4 mRNAs, reverse transcription was also performed using oligo(dT) primers, with actin as an internal control using primers Actin1for (5′-TGGAACTGGAATGGTTAAGGCTGG-3′) and Actin1rev (5′-TCTCCAGAGTCGAGCACAATACCG-3′). The primers bind to two different exons and overlap exon–intron borders to eliminate potential detection of genomic DNA in the cDNAs. Actin was equally distributed throughout the tissues, and no genomic DNA amplification was detected (data not shown). Detection of YUCCA4 transcripts was maximized by selecting primers that bind towards the 3′ end of YUCCA4: exon 4 for YUCCA4.1 and exon 3/intron 3 for YUCCA4.2 (Exon4for, 5′-GAGAATATTTTTTCACAAAG-3′; Exon3for, 5′-CAAATTAAAGTGACGCAAGCCGTGAAG-3′; Yuc4.1rev and Yuc4.2rev). Consistent with the data for the full-length gene, YUCCA4.1 is present in all tested tissues, and YUCCA4.2 is only detectable in flowers (data not shown).

Vector constructs

Using cDNA from flower tissues, full-length spliced and unspliced variants were amplified using primers YucBamfor (5′-GCTCCTGGATCCGGTATGGGCACTTGTAGAGAATC-3′), Yuc4.1Sacrev (5′-CCGGCCGAGCTCTTAGGATTTATTGAAATGAAGATG-3′), Yuc4.2Sacrev (5′-CCGGCCGAGCTCTCACATATACATATACACATTGAC-3′), Yuc4.1Bamrev (5′-CCGGCCGGATCCTTAGGATTTATTGAAATGAAGATG-3′) and Yuc4.2Bamrev (5′-CCGGCCGGATCCTCACATATACATATACACATTGAC-3′) for cloning into pVKHEn6 vectors encoding YFP or roGFP2 tags (kindly donated by Andreas Meyer, Institut für Nutzpflanzenwissenschaften und Ressourcenschutz, University of Bonn, Germany) (Brach et al., 2009) via BamHI and SacI restriction sites, as well as for cloning into the expression vectors pET28b (Novagen, and pHisTrx, a pET32a derivative encoding E. coli thioredoxin with an N-terminal six-histidine tag (a gift from Richard Kammerer, Faculty of Life Sciences, University of Manchester, UK) (Kammerer et al., 1998) via BamHI restriction sites.

In vitro transcription and translation

Both YUCCA4 splice variants were fused to the pSPUTK SP6 promoter by overlapping extension PCR (Urban et al., 1997) using primers Yuc4Fusfor (ctttggcagatctaccatgggcacttgtagagaatc), Yuc4.1rev and Yuc4.2rev. Transcription was performed using 10 μg PCR fusion product and SP6 RNA polymerase (New England Biolabs) according to the manufacturer’s instructions. Protein translations were performed in wheatgerm extract (Promega, according to the manufacturer’s instructions using Easy Tag Express 35S (Perkin Elmer,

Targeting assays

ER targeting assays for YUCCA4.1 and 4.2 were performed co-translationally for 15 min using 50 A280 U ml−1 canine pancreatic microsomes. Microsomes were pelleted at 180 000 g for 45 min, and washed for 10 min with 0.1 m Na2CO3. Pelleted microsomes as well as microsomal import samples treated with 40 U ml−1 thermolysin for 5 min at 30°C were analysed by SDS–PAGE and Cyclon phosphor screening (Packard, Chloroplast imports were performed as described by Kriechbaumer et al. (2009). Briefly, chloroplasts from pea leaves (Pisum sativum) were incubated with pre-spun translated protein and incubated at 30°C for 20 min. Organelles were re-purified by centrifugation and Na2CO3 washes.

Transient expression and ratiometric imaging

Nicotiana tabacum leaves were infiltrated as described by Sparkes et al. (2006). All YUCCA constructs were infiltrated at OD600 = 0.1, and the marker constructs were infiltrated at OD600 = 0.05. Three days after infiltration, images were taken using a Zeiss LSM510 Meta laser scanning confocal microscope ( with 63× oil immersion objectives. For imaging of GFP/YFP combinations, samples were excited using 458 and 514 nm laser lines in multi-track mode with line switching. Images were edited using the LSM510 image browser and Adobe Photoshop ( Ratiometric imaging of roGFP2 was performed as described byWang et al. (2011) with one additional step: samples were treated with 10 μg ml−1 brefeldin A (to block ER to Golgi transport) for 1 h prior to imaging, allowing Golgi-localized YUCCA4 to be fully re-distributed to the ER.

In silico protein analysis

Predicted transmembrane domains and N-terminal target peptides were investigated using TMHMM, SignalP, TargetP and ChloroP on the ExPASy server, respectively, using parameters described by Kriechbaumer et al. (2009). ChloroP ( and TMHMM ( were used with standard settings. SignalP is a program predicting the probability of protein secretion in vivo based on properties of the primary sequence (, with high sensitivity (98.2%) and a low rate of false positives (0.4%) in eukaryotes. The parameters used here were: eukaryotes, neural networks and hidden Markov model; truncated to first 70 residues. TargetP ( was used to detect secretory signal peptides and N-terminal plastidial or mitochondrial targeting sequences, and shows 80–100% specificity depending on the reliability class.

Cell-free protein expression and activity tests

For protein expression, a PURExpress™in vitro protein synthesis kit (New England Biolabs) was used according to the manufacturer’s instructions. This kit contains ribosomes, glycerol, dithiothreitol, creatine phosphate, aminoacyl tRNA synthetases, methionyl tRNA formyltransferase, T7 RNA polymerase and plasmid DNA encoding a control protein (dihydrofolate reductase), YUCCA4.1 or YUCCA4.2. To determine enzymatic activities, 1 μg YUCCA4 protein in the cell-free extract (amount estimated from SDS–PAGE in comparison with the protein marker), 10 mm NADPH/NADP and 2 mm of the substrates IPA, IAAld, TAM or Trp were incubated for 4 h at 30°C, and analysed on silica gel-coated TLC plates with chloroform/methanol/acetic acid (75:20:5) as the mobile phase. The TLC plates were stained using Ehrlich’s reagent (Sigma,, and bands were quantified using an ODYSSEY infrared imaging system (LI-COR Biosciences,


The authors thank Andreas Meyer (Institut für Nutzpflanzenwissenschaften und Ressourcenschutz, University of Bonn) for kindly supplying the roGFP2 vector, and Richard Kammerer (Faculty of Life Sciences, University of Manchester) for the pHisTrx vector. We thank Adrian Hall for critical discussion of the splicing experiments, and Malcolm Lock and Patrick Harrison for advice on expression of mono-oxygenases. This work was supported by a Biotechnology and Biological Sciences Research Council research grant awarded to B.M.A. (grant number BB/E01559X/1).