The cyclophilin DIAGEOTROPICA has a conserved role in auxin signaling

The plant hormone auxin has a central role in all aspects of plant growth and development. Auxin acts by regulating transcription through the action of three protein families: the TIR1/AFB auxin receptors (Dharmasiri et al., 2005b; Mockaitis and Estelle, 2008; Greenham et al., 2011), the Aux/IAA transcriptional repressors (Overvoorde et al., 2005) and the ARF transcription factors (Guilfoyle and Hagen, 2007). The TIR1/AFB receptors are F-box proteins and subunits of ubiquitin ligase SCF^TIR1/AFB complexes (Santner and Estelle, 2010). Upon auxin binding, the TIR1/AFB protein interacts with the highly conserved degron domain of the Aux/IAA proteins, resulting in their degradation (Dharmasiri et al., 2005a; Kepinski and Leyser, 2005; Tan et al., 2007; Calderon-Villalobos et al., 2010). The degradation of Aux/IAA proteins releases the repression of ARFs, leading to ARF-dependent transcription.

Our understanding of auxin signaling is based primarily on work in Arabidopsis. Less is known about auxin signaling in non-seed plants. The completion of the genome sequence for the moss Physcomitrella patens (Rensing et al., 2008), together with efficient targeted gene insertion (Schaefer, 2002; Prigge and Bezanilla, 2010), have established P. patens as a useful experimental tool for studying the ancestral role of auxin.

Mosses are one of three lineages referred to as bryophytes, which are early diverging non-vascular land plants (Kenrick and Crane, 1997). The moss life cycle is dominated by a haploid gametophyte. Following germination of a haploid spore, protonemal filaments are produced and proliferate by apical division. Initially, protonemata consist of chloroplast-rich filaments called chloronemata. Later, chloronemal filaments can differentiate to form a distinct type of filament, called caulonemata, which are more elongated and contain fewer chloroplasts. Subapical caulonemal cells, in turn, can form branches that either develop into a secondary chloronemal filament, a caulonemal filament or into a leafy gametophore (Ashton et al., 1979).

Auxin plays a central role throughout the life cycle of P. patens. In the filamentous protonemata stage, exogenous auxin promotes the transition of a chloronemal filament into a caulonemal filament (Ashton et al., 1979). In the adult plant, auxin treatment stimulates stem elongation (Fujita et al., 2008; Eklund et al., 2010), elongation of gametophore phyllids (leaves) (Decker et al., 2006) and the formation of rhizoids from epidermal cells of the gametophore (Ashton et al., 1979).

In a screen carried out three decades ago, Ashton and colleagues (Ashton et al., 1979) isolated P. patens mutants with an altered response to the synthetic auxin 1-naphthalene-acetic acid (NAA). These NAA-resistant (nar) mutants were isolated based on their reduced sensitivity to a high concentration of auxin or by their morphological abnormalities. Although the affected genes remained unknown, the mutants were utilized to characterize the auxin response in moss.

The P. patens genome encodes four auxin receptors, three Aux/IAA and fifteen ARF proteins (Rensing et al., 2008; Paponov et al., 2009; Parry et al., 2009; Prigge et al., 2010). Recently, we demonstrated that the mechanism of auxin action is conserved in P. patens. DNA sequencing revealed eight independent nar mutants harboring gain-of-function mutations in the conserved degron motif of the three PpAux/IAA genes, similar to comparable mutations in Arabidopsis. The moss mutants display an auxin-insensitive phenotype that includes a significant reduction in the number of caulonemal cells and the subsequent absence or reduction in the development of leafy gametophores. Comparable phenotypes were also identified as a result of silencing of the four PpAFB receptor genes. Additionally, PpAFB and PpAux/IAA proteins interact in the presence of auxin, indicating that the mechanism of auxin perception is conserved in non-seed plants (Prigge et al., 2010). This work demonstrates the significant contribution of moss genetic analyses to the study of auxin signaling. In the current work we describe the discovery and analysis of P. patens diageotropica (dgt) auxin-resistant mutants.

Tomato dgt mutants exhibit a pleiotropic phenotype that includes a slow gravitropic response, lack of lateral roots, reduced apical dominance, shortened internodes, altered vascular development and reduced fertility and fruit growth (Zobel, 1973; Mito and Bennett, 1995; Rice and Lomax, 2000; Balbi and Lomax, 2003; Ivanchenko et al., 2006). Application of auxin to tomato dgt mutants does not stimulate shoot elongation and the effect on root growth inhibition is reduced compared with the wild type (WT). Ethylene production is also reduced in auxin-treated dgt tomato plants (Bradford and Yang, 1980; Kelly and Bradford, 1986). However, the rate of polar indole-3-acetic acid (IAA) transport in dgt hypocotyls and roots is similar to that of the WT parent (Daniel et al., 1989; Muday et al., 1995) and levels of IAA in dgt shoot apices were shown to be normal (Fujino et al., 1988). These results generally support the hypothesis that tomato dgt is altered in auxin perception or response. The effect of the mutant dgt on the expression of auxin-induced genes is variable. Whereas the induction of several Aux/IAA genes was shown to be normal, the induction of others was significantly reduced (Nebenfuhr et al., 2000; Balbi and Lomax, 2003). The subset of Aux/IAA genes showing reduced expression differs depending on the developmental stage and tissue (Balbi and Lomax, 2003). Similarly, the expression of genes encoding 1-aminocyclopropane-1-carboxylic acid synthases (ACSs), which are key regulatory enzymes of ethylene biosynthesis, is differentially affected in the tomato dgt mutants (Balbi and Lomax, 2003; Coenen et al., 2003; Vidoz et al., 2010). The induction of a SMALL AUXIN UP-REGULATED RNA (SAUR) gene also displays reduced induction compared with WT (Mito and Bennett, 1995; Coenen et al., 2003). The tomato DGT gene was mapped and determined to encode a type A cyclophilin (CyPA), designated CYP1 (Oh et al., 2002; Oh et al., 2006). Each of the three known dgt mutant alleles was shown to contain a single point mutation that generates either an amino acid substitution or an early stop codon (Oh et al., 2006).

Cyclophilins are represented in every organism and belong to a highly conserved large family of peptidyl-prolyl cis-trans isomerase (PPIase) proteins that catalyze the cis-trans isomerization of peptide bonds preceding a proline residue (Galat, 2003). The cellular roles of the cyclophilins have been difficult to determine. In mammals, they have numerous functions, including protein folding, chaperone activities, cell signaling, protein trafficking, ligand-receptor recognition, regulation of transcription and splicing (Wang and Heitman, 2005). The archetypal CyPA has attracted interest due to its involvement in diverse pathological conditions (Nagy et al., 2011; Lee and Kim, 2010). Although extensively studied, its targets and specific roles remain obscure. Cyclophilins are abundantly expressed in Arabidopsis (Chou and Gasser, 1997; Romano et al., 2004), but the functions of only a few are known.

Here we demonstrate that several P. patens nar mutants harbor mutations in the moss ortholog of the tomato DGT (CYP1) gene and describe their reduced sensitivity to auxin. We also characterize the P. patens DGT gene and study its interactions with components of the TIR1/AFB-dependent auxin signaling cascade. We show that DGT has a conserved role in auxin signaling, which affects the induction of auxin-regulated genes downstream of the auxin receptors.

WT ‘Gransden-2004’ P. patens and other strains listed below were grown at 25°C under continuous light at an intensity of 40-70 μmol/m²/second on BCD and BCDAT media supplemented with 1.8 μM para-aminobenzoic acid (PAB) to allow growth of the auxotrophic strains pabA3 and pabB4. Strains NAR 23 (dgt-23), NAR 26 (dgt-26), NAR 171 (dgt-171 pabA3), NAR 176 (dgt-176 pabA3), NAR 180 (dgt-180 pabA3) and pabB4 were described previously (Ashton and Cove, 1977; Ashton et al., 1979). In this work, the pabB4 line was used as an auxotrophic control strain because pabA3 was not available.

Mutations in the PpDGT gene were identified by sequencing PCR products from genomic DNA isolated from nar mutant lines using PpDGT-F/PpDGT-R primers (supplementary material Table S1).

The constructs used for the moss transformations described below are listed in Table 1. Primers are listed in supplementary material Table S1.

To create pML36, GUS:T_nos was amplified using PML9/PML10 primers and cloned into the XhoI site of pBNRF to create pBNRF-GUS. Approximately 1 kb of genomic sequence upstream (5′ flanking region) and downstream (3′ flanking region) to the DGT coding region were amplified using the primer pairs PML132/PML125 and PML128/PML129, respectively, and each fragment was cloned into pENTR-D/TOPO (Invitrogen, Carlsbad, CA, USA). The 5′ and 3′ flanking regions were recombined into BamHI and SpeI sites in pBNRF-GUS, respectively.

To create pML31, a fragment consisting of 509 bp upstream sequence together with the DGT coding region (P_DGT509:DGT) was amplified using the PML289/PML301A primer pair. pML36 was digested with BamHI and the DGT 5′ flanking region was replaced with the P_DGT509:DGT sequence.

To create pML38, PML31 was subject to site-directed mutagenesis using PML149/PML150.

To create pML32, the YFP-HA:T_OCS fragment was amplified from pEarleyGate102 (Earley et al., 2006) using PML378/PML379 primers followed by site-directed mutagenesis to mutate an XhoI site using PML380/PML381 primers. pML31 was digested with XhoI and the GUS gene was replaced with YFP-HA:T_OCS.

To create pML37, a P_35S:HygR:T_CaMV hygromycin resistance cassette was amplified from pMP1159 (http://web.mac.com/mjprigge/Prigge/Moss.html) using PML371/PML372 primers. The pBNRF-GUS construct harboring a DGT 3′ flanking region inserted into the SpeI site was digested with KpnI to replace the nptII gene with a HygR gene.

To create pML53, pML36 was digested with XhoI to replace a fragment consisting of the DGT 5′ flanking region and the GUS gene with the DGT 5′ flanking region only. The resulting construct was digested with KpnI to replace the nptII gene with a HygR gene.

To create pML59, PpIAA1B cDNA was cloned upstream of GUS:Tnos before amplifying the PpIAA1B-GUS:Tnos fragment using PML230/PML10 primers. The resulting product was cloned into pENTR-D/TOPO and recombined into pMP1180.

To create pML49, PpAFB2 cDNA was amplified using PML203/PML204 primers and cloned upstream of a c-Myc tag before amplifying the PpAFB2-c-Myc fragment using PML203/PML208 primers. The resulting product was cloned into pENTR-D/TOPO and recombined into pMP1155.

Protoplast isolation and PEG-mediated transformation of P. patens was performed as described (Nishiyama et al., 2000). Five days after regeneration, transformants were selected on BCDAT medium containing 20 mg/l G418 or hygromycin.

Transgenic lines were screened by PCR for the presence of both left and right transgene-endogenous sequence junctions to verify the insertion of the transgene at the targeted locus. Genomic DNA extracted from dgtΔ lines was analyzed by PCR to confirm the integration of the transgene into the DGT locus and the absence of the DGT coding region (confirmation of the transgene integration is shown in supplementary material Fig. S3O). cDNA (extracted as described below) was detected by RT-PCR to confirm the absence of a transcribed product.

Protonemal tissue composed mainly of chloronemata was homogenized in a Waring blender and plated in triplicate on BCDAT plates with cellophane overlays for 7 days. For hormone treatment, the cellophane overlays covered with protonemal tissue were transferred into liquid BCD medium containing 10 μM IAA or 10 μM ABA or the equivalent amount of ethanol or methanol solvent, respectively. For heat shock treatment, the cellophane covered with protonemal tissue was transferred into liquid BCD medium and incubated at 37°C (or 25°C control). The tissue was collected after 5 hours and total RNA was isolated using the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA). RNA (500 μg) was reverse transcribed using the Superscript III First-Strand cDNA Synthesis System (Invitrogen). Then 20 μl of the reverse transcription reaction was diluted with water to a final volume of 200 μl. PCR reactions were performed with 4 μl diluted cDNA and the primers listed in supplementary material Table S1 and detected using the CFX384 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA). Normalized expression [ΔΔC(t) method] was calculated using Bio-Rad CFX Manager software with PpACTIN as a reference gene and plotted as values relative to mock-treated WT ± s.e.m.

To express the PpIAA1B-GUS fusion protein, plants were grown on BCDAT plates with cellophane overlays, transferred to liquid BCDAT medium and incubated at 37°C for 1 hour and 30 minutes, followed by a recovery of 30 minutes at 25°C. Tissue was transferred to liquid BCDAT media containing 50 μM IAA or the equivalent amount of ethanol solvent for 5 hours and then stained for GUS.

Tissue was stained in GUS staining solution [50 mM NaH₂PO₄ pH 7.0, 0.5 mM X-Gluc, 0.5 mM K₃Fe(CN)₆, 0.5 mM K₄Fe(CN)₆, 0.05% Triton X-100] at 37°C for 45 minutes or overnight following auxin treatment of P_GH3:GUS lines or the degradation assay, respectively. Plants were cleared in 70% (v/v) ethanol and imaged using a Nikon SMZ1500 dissecting scope.

Proteins from protonemal tissue were extracted in 50 mM Tris pH 7.5, 150 mM NaCl, 10% glycerol, 0.1% NP40 and complete protease inhibitor (Roche, Indianapolis, IN, USA). The resulting protein extract was centrifuged for 10 minutes at 10,000 g and the supernatant was collected. Proteins were resolved by SDS-PAGE and transferred onto nitrocellulose membranes. Membranes were stained with Ponceau S to standardize the input and an HRP-conjugated anti-c-Myc antibody was used for protein detection (Roche). Proteins were visualized using the ECL Plus Western Blot Detection System (GE Healthcare, Piscataway, NJ, USA).

To discover which genes are mutated in the P. patens nar mutants, we took a candidate gene approach and sequenced the moss orthologs of known auxin-related genes. A moss gene encoding the closest homolog to tomato DGT (CYP1) and to human CyPA was identified and designated PpDGT (GenBank accession number XM_001778952; supplementary material Fig. S1). Sequencing of PpDGT in the auxin-resistant moss mutants revealed five different dgt mutant alleles. The mutated alleles contain single-nucleotide point mutations that generate either an amino acid change (dgt-26^S-L, dgt-171^G-E, dgt-176^G-D and dgt-180^C-Y) or a stop codon (dgt-23^Q-stop) affecting highly conserved regions (supplementary material Fig. S1).

The dgt/nar mutants and WT strains (Grandsen-2004 and pabB4 serving as controls for the non-auxotrophic and auxotrophic mutants, respectively) were grown on minimal medium supplemented with ammonium tartrate (BCDAT) or on a medium without ammonium tartrate (BCD) to stimulate the growth of protonemata or gametophores, respectively. When WT plants were grown on auxin-containing BCDAT medium for 1 month, caulonemal filaments developed from chloronemal cells (Fig. 1A, WT and pabB4). By contrast, the dgt/nar mutants developed fewer and much shorter caulonemal filaments than WT. WT plants grown on BCD medium produced more leafy gametophores (Fig. 1B, WT, pabB4; supplementary material Fig. S2). When 0.5 μM NAA was added to the medium, only a few fully developed gametophores grew and the plants produced mainly rhizoid-like cells that emerged from small leafless gametophore-like structures. With a higher concentration of exogenous auxin, no gametophores developed and the plants comprised rhizoid-like cells only. Under the same conditions, dgt/nar mutant plants grown on BCD developed fewer and smaller gametophores. The effect of exogenous auxin on the dgt/nar mutants was less pronounced compared with WT, with fewer rhizoids developing in response to high auxin concentrations. The dgt-176 mutant showed the strongest phenotype, with no gametophores developing on BCD medium (Fig. 1B).

It has been shown that auxin affects the size of the leaves in moss (Decker et al., 2006). To further characterize the differences between WT and dgt/nar mutants in their response to auxin, we quantified the length of the uppermost two or three leaves from plants grown on BCD medium for 1 month (Fig. 1E). Similar to the differences in size between WT and mutant gametophores, the individual leaves of WT gametophores were significantly longer than those of dgt/nar mutants. Whereas growth on auxin resulted in a decrease in WT leaf length, mutant leaves were either the same length or longer when grown on auxin as compared with untreated leaves.

To confirm that the phenotype of the dgt/nar mutants results from the mutated dgt, we replaced the DGT coding region with the GUS gene by homologous recombination (P_DGT:GUS/dgtΔ). Two and five independent lines were obtained using two different constructs carrying either the neomycin resistance (nptII) or hygromycin resistance (HygR) gene, respectively. The insertion of the transgene into the targeted locus and the absence of DGT were confirmed for all lines. All the lines presented similar phenotypes. The two nptII-encoding lines were compared with the dgt/nar mutants (Fig. 1C,D; supplementary material Fig. S2). RT-PCR analyses confirmed the absence of a full-length DGT transcript in the transgenic line, indicating that the dgtΔ strain is a null allele (supplementary material Fig. S3). Loss of the DGT gene resulted in reduced sensitivity to auxin. The dgtΔ plants developed fewer caulonemal filaments than WT when grown on auxin (Fig. 1C). When grown on medium lacking ammonium tartrate (BCD), the size of gametophores was reduced and the length of their leaves was also reduced compared with WT. Application of auxin also resulted in shorter rhizoids (Fig. 1D,E; supplementary material Fig. S2). However, the dgtΔ mutant presented a less severe phenotype than the five dgt/nar mutants.

We next tested whether the dgt/nar mutations can be complemented with the native DGT gene. The mutant locus in dgt-180 was replaced with a DGT-GUS fusion gene and four independent lines were obtained. These lines [P_DGT:DGT-GUS (dgt-180)] were compared with dgt-180 and WT, and the replacement of dgt with the native gene was found to partially restore the native phenotype when grown on auxin, but not when grown without exogenous auxin [two representative P_DGT:DGT-GUS (dgt-180) lines are presented in Fig. 1F]. These data indicate that the auxin-insensitive phenotype of dgt-180 resulted from the mutation in dgt.

The severe phenotypes of the dgt/nar mutants might suggest a dominant effect of the mutant alleles. An alternative explanation is that the more severe phenotypes are related to additional N-methyl-N′-nitro-N-nitrosoguanidine (EMS)-induced mutations. This explanation is supported by the failure of the DGT gene to complement the phenotype of the dgt-180 mutant when grown without exogenous auxin. To address this, we knocked out the mutant dgt gene in the dgt-26 and dgt-180 lines and compared the resulting transgenic lines with the corresponding mutant background lines. In addition, we ectopically overexpressed the dgt-26 and dgt-180 genes in WT plants and compared these lines with DGT ectopic overexpression lines. No differences were observed in either case between the test and control lines (data not shown). These results suggest that the severe phenotypes of the dgt/nar mutants do not result from a dominant effect.

To compare the phenotypes of the dgtΔ null mutant and a dgt point mutant in the same background, we replaced the DGT coding region with a dgt-176-GUS fusion. We found that the dgt-176-GUS and dgtΔ lines were very similar (two representative P_DGT:dgt-176 lines are shown in Fig. 1G). This indicates that the loss of native DGT protein has the same effect as a mutation that produces a non-functional protein and further associates DGT with the phenotype observed in the nar mutants.

Because the dgt/nar lines appear to have defects unrelated to DGT, to further characterize the dgt phenotype we focused on the dgtΔ null lines. One of each of the nptII and HygR lines was grown on BCDAT or BCD media with different concentrations of NAA (supplementary material Fig. S4A,B). Both dgtΔ mutant lines showed reduced sensitivity to applied auxin. When grown on BCDAT medium, shorter protonemal filaments and fewer gametophores developed compared with WT. In the presence of auxin, WT plants developed fewer gametophores and longer filaments, whereas the dgtΔ mutants presented a delayed response (supplementary material Fig. S4A). On BCD medium, WT plants produced gametophores and also caulonemal filaments. The addition of auxin resulted in a decrease in the growth of caulonemata and reduced the number and size of gametophores and leaves (supplementary material Fig. S4B,D,E). Conversely, the growth of rhizoids was increased (supplementary material Fig. S4B). The dgtΔ mutants presented a delayed auxin response. In the absence of exogenous auxin, the mutants developed smaller gametophores than WT but more filaments. In the presence of exogenous auxin, longer filaments developed. In the mutant, the proportion of gametophores to rhizoids was not decreased (in contrast to observations in WT) and gametophores could still be detected in the presence of 12.5 μM NAA.

To exclude the possibility that replacement of the DGT coding region with a GUS insertion affects the dgt phenotype, dgt knockout lines that do not harbor GUS were also generated (these lines were created in a P_GH3:GUS background for an additional purpose described below). Four independent lines were obtained and the absence of DGT transcripts was confirmed (supplementary material Fig. S3B-E). Three independent lines were analyzed further and used to quantify the number of gametophores, the length of leaves, and the proportion of chloronema to caulonema (supplementary material Fig. S4D-G). The phenotype of these lines was indistinguishable from that of the dgtΔ/GUS lines.

It is possible that the dgt phenotype is related to a defect in auxin uptake or efflux. The P. patens genome encodes orthologs of both the AUX1 uptake carrier and the PIN efflux proteins (Rensing et al., 2008). Because NAA and 2,4-dichlorophenoxyacetic acid (2,4-D) are substrates for the efflux and influx carriers, respectively, differential sensitivity to these auxins is indicative of a defect in auxin transport (Delbarre et al., 1996). We found that 2,4-D had a similar effect to NAA on the growth of dgtΔ, suggesting that the mutant is not deficient in an aspect of auxin transport (supplementary material Fig. S4C). However, a better understanding of auxin transport in P. patens will be needed to address this hypothesis.

To gain a better understanding of the role of DGT in moss development we studied its expression pattern using P_DGT:GUS (the dgtΔ/GUS lines described above). In protonemata, GUS activity was detected in all filaments, with the strongest activity in the apical cells of rapidly growing caulonemata and in side branches differentiating into secondary chloronemata (Fig. 2A,B). In gametophores, GUS activity was detected in rhizoids and in axillary hairs emerging from the adaxial side of the leaves (Fig. 2C,D). To compare the transcription pattern with the expression of the DGT protein, we created transgenic lines harboring a DGT-GUS fusion. The DGT-GUS expression pattern was identical to the GUS expression in the dgtΔ lines (supplementary material Fig. S5) and was unchanged following auxin treatment (data not shown).

To analyze the subcellular localization of DGT, we generated lines in which a DGT-YFP fusion is expressed from the endogenous DGT locus (P_DGT:DGT-YFP). Similar to other CyPA-like proteins (Wang and Heitman, 2005), YFP fluorescence was observed in both the cytoplasm and nucleus (supplementary material Fig. S6). As cyclophilins can function as chaperones and therefore may facilitate the subcellular translocation of their substrate proteins, we examined whether the subcellular localization of DGT is affected by auxin. We treated the P_DGT:DGT-YFP fusion lines with auxin but did not observe any changes (data not shown).

Tomato dgt mutants exhibit reduced auxin induction of several auxin-regulated genes (Mito and Bennett, 1995; Balbi and Lomax, 2003; Coenen et al., 2003; Vidoz et al., 2010). To explore the effect of P. patens DGT on auxin-dependent transcription, we utilized an auxin reporter. A transgenic moss plant expressing a GUS transcriptional fusion driven by a Glycine max GH3 promoter (P_GH3:GUS) was previously used as an auxin reporter and was shown to respond to exogenous auxin within 24 hours (Bierfreund et al., 2003). We knocked out the DGT coding region within the P_GH3:GUS line, as described above, and examined the distribution of GUS protein in the resulting lines (P_GH3:GUS dgtΔ). Following a 5-hour treatment with IAA, GUS expression in all three representative P_GH3:GUS dgtΔ lines was lower in both protonemata (Fig. 3A) and gametophores (Fig. 3B) compared with the control line (Fig. 3, P_GH3:GUS). Moreover, knocking out DGT in the P_GH3:GUS line enabled us to compare the effect of DGT on the distribution of auxin response activity. No differences in the expression pattern were observed, again indicating that DGT does not affect auxin transport.

To further study the effect of DGT on auxin-dependent transcription we examined the expression of several auxin-regulated genes in dgt and WT. Seven-day-old plants composed mainly of protonema were treated with auxin for 5 hours and transcript levels of the target genes were determined by quantitative RT-PCR (qPCR). The selected genes are known to be regulated by auxin in P. patens [the three PpAux/IAA genes (Prigge et al., 2010) and RHD6-like (Jang and Dolan, 2011)] or are orthologs of known Arabidopsis auxin-regulated genes (GCN5, SAUR, ARF19 and AUX1). With the exception of IAA2 and ARF19-likeA, the expression levels of all the sampled genes were generally reduced in dgtΔ compared with WT. Based on a Student’s t-test, the basal levels of IAA1A, IAA1B, RHD6, GCN5, SAUR and AUX1 and also the auxin-induced levels of IAA1A, IAA1B, GCN5, SAUR and AUX1 were significantly different between dgtΔ and WT.

These results indicate a clear deficiency in auxin-regulated transcription in dgtΔ. However, a general effect of DGT on gene expression could not be excluded. To test whether DGT has a specific effect on auxin regulation, we tested the response of dgtΔ to the plant hormone abscisic acid (ABA) and to a heat shock treatment. We selected several genes that display either an ABA (Cuming et al., 2007; Khandelwal et al., 2010) or a heat shock response and tested their expression following 5 hours of treatment (Fig. 4B,C). No differences were observed between dgtΔ and WT with respect to either basal or induced transcript levels, suggesting a specific role for DGT in auxin-regulated transcription.

Reasoning that cyclophilins may function in protein folding, we examined whether the effect of DGT on transcription is due to a role in the interaction between the AFB and Aux/IAA proteins and the subsequent degradation of Aux/IAAs. We created transgenic lines expressing PpIAA1B-GUS fusions driven by a soybean heat shock promoter (P_HSP:IAA1B-GUS). We obtained two independent transgenic lines and first tested their ability to degrade the heat shock-induced protein following auxin treatment. We show that auxin promotes the degradation of IAA1B-GUS (Fig. 5A,D). To test the effect of DGT on IAA1B-GUS, we introduced P_HSP:IAA1B-GUS into dgt-26 and dgt-180 and isolated one and two independent lines, respectively (one of each is shown in Fig. 5B,C). In addition, dgt was knocked out in one of the P_HSP:IAA1B-GUS transgenic lines (Fig. 5D), resulting in three independent lines (two representative lines are shown in Fig. 5E,F; the null mutation is confirmed in supplementary material Fig. S3F-H). In all these lines IAA promoted the degradation of IAA1B-GUS in a similar way, suggesting that DGT does not affect the ubiquitin-dependent degradation of Aux/IAA proteins. However, it is possible that a modest change in the IAA1B-GUS level might not be detectable in our assay.

To further explore the role of DGT in auxin signaling, we studied the genetic interaction between DGT and PpAFB2. We first generated transgenic lines overexpressing PpAFB2 and characterized the resulting phenotype. We targeted an AFB2-c-Myc fusion gene driven by a maize ubiquitin promoter (P_Ubi:AFB2-Myc). Eight lines were selected and the levels of the tagged protein were determined. Two lines with different levels of AFB2-Myc were used for further analysis (Fig. 6A). When these lines were grown on BCDAT medium they developed more and longer caulonemal filaments, which in turn produced more secondary chloronemal filaments, than WT. Furthermore, their phenotype correlated with the levels of AFB2-Myc protein (Fig. 6C): the higher expressing line (P_Ubi:AFB2-Myc #2) developed more caulonemal filaments and more secondary chloronema than the lower expressing line (P_Ubi:AFB2-Myc #1).

The P_Ubi:AFB2-Myc #2 line was characterized further. The auxin hypersensitive phenotype was also displayed during the development of the gametophore; when no exogenous auxin was added, the leafy gametophores appeared smaller than those of the WT and their leaves were longer and narrower, similar to WT plants grown on auxin-containing media (Fig. 6B,F). Whereas gametophores were still observed on WT plants grown on 0.5 μM NAA, the AFB2 overexpression line developed only rhizoid-like cells, consistent with a hypersensitivity to auxin.

To study the genetic interaction between DGT and the AFB2 auxin receptor we knocked out DGT in the P_Ubi:AFB2-Myc #2 line. Two lines were obtained (confirmation of the null mutation is shown in supplementary material Fig. S3I,J) and the AFB2 protein level was analyzed and found to be similar to that of the background line (Fig. 6D). The P_Ubi:AFB2-Myc overexpression line and the two P_Ubi:AFB2-Myc dgtΔ lines were grown together with WT on BCD media with or without varying concentrations of auxin (Fig. 6B). In general, deletion of DGT from the AFB2 overexpression lines dramatically reduced auxin sensitivity. Whereas the number of gametophores and the size of the leaves were dramatically affected by auxin in both WT and the AFB2 overexpression lines, only a mild effect was observed in the P_Ubi:AFB2-Myc dgtΔ lines. Similar to dgtΔ, even on 12.5 μM NAA the P_Ubi:AFB2-Myc dgtΔ lines produced some gametophores whereas the WT and AFB2 overexpression background line grew only rhizoid-like cells (Fig. 6E,F). These results suggest that dgt can suppress AFB2 and might be involved in transcriptional regulation.

The significant effect of auxin on the development and growth of the moss P. patens has been known for many years. The sequencing of the P. patens genome revealed that the core components of the auxin signaling machinery and other auxin-related genes are conserved in this early land plant. Recently, a few studies of P. patens genes have demonstrated a link between conserved auxin-related genes and auxin-dependent phenotypes. In a recent study (Prigge et al., 2010), we sequenced a set of severe auxin-insensitive mutants and found them to be affected in the three PpAux/IAA proteins. We also found that knockdown of the four TIR1/AFB auxin receptor genes resulted in defects in the chloronema-to-caulonema transition. The Arabidopsis SHORT INTERNODS/STYLISH (SHI/STY) family genes serve as positive regulators of auxin biosynthesis genes including YUCCA4. Knockout and overexpression of the P. patens homologs of SHI/STY resulted in lower and higher levels of IAA, respectively (Eklund et al., 2010). Menand and colleagues (Menand et al., 2007) present a double-knockout mutant of PpRSL1 and PpRSL2 (homologs of Arabidopsis ROOT HAIR DEFECTIVE 6, which encodes a basic helix-loop-helix transcription factor), which fails to develop caulonemal filaments and rhizoids. A recent study reports that the RSL genes are positively regulated by auxin (Jang and Dolan, 2011). In the current study, we show another intriguing example of a conserved role of auxin action that is shared between early land plants and angiosperms. We have sequenced a number of moss mutants displaying an auxin-insensitive phenotype and found five independent alleles of a moss ortholog of the tomato DGT gene.

The tomato dgt mutated lesion was shown to affect several complex developmental processes in tomato, including root growth, fruit development, ethylene production and the development of adventitious roots (Zobel, 1973; Kelly and Bradford, 1986; Balbi and Lomax, 2003; Ivanchenko et al., 2006; Vidoz et al., 2010). By contrast, the developmental stages and body plan of moss are simple. This suggests that DGT has a basic and central role in the auxin response that is expressed as different, lineage-specific outputs. Indeed, the P. patens dgt mutants display defects in every auxin-dependent process, including the chloronema-to-caulonema transition and the development of rhizoids.

Prior to this study, dgt mutants were known only in tomato. The moss dgt mutants highlight the central role of this gene in the auxin signaling cascade. The Arabidopsis genome encodes four close homologs of DGT, and analysis of single-mutant lines has not revealed noticeable phenotypes. Therefore, it is possible that the clear phenotype of tomato and P. patens dgt mutants is due to decreased numbers of DGT genes in these two plant species. Higher-order mutant analysis in Arabidopsis, a plant for which the auxin signaling cascade is better characterized, might contribute to our understanding of the conserved role of DGT in plants.

The role of DGT in auxin signaling is unknown. Phenotypic studies in tomato together with direct measurements of auxin concentration and transport suggested a role for DGT in auxin signaling rather than in auxin biosynthesis or transport (Fujino et al., 1988; Daniel et al., 1989; Muday et al., 1995). The tomato dgt mutants were established as a useful tool to study auxin signaling in tomato and in angiosperms in general. Several studies took advantage of the pleiotropic phenotype of the dgt mutants to characterize the auxin signaling pathway as well as the crosstalk between auxin and other hormones. These studies advanced our understanding of the effect of auxin on plant development and growth. However, they did not reveal a specific biological role for DGT. The position of DGT within an auxin signaling cascade also remains unclear. One mechanistic aspect that has been investigated is the effect of DGT on auxin-regulated transcription (Nebenfuhr et al., 2000; Balbi and Lomax, 2003). These studies demonstrate that DGT affects only a subset of auxin-regulated genes, suggesting that it is a component of a specific auxin signal transduction pathway. In this study, we provide evidence that P. patens dgt has a broad effect on the induction of auxin-regulated genes. Further studies focusing on the role of DGT in gene expression will elucidate the direct targets of DGT.

We thank Neil Ashton for supplying the nar mutant lines; Ralph Quatrano, Mattias Thelander and Eva Sundberg for sharing materials; and Assaf Mosquna for helpful advice on moss transformation.