Tryptophan-dependent auxin biosynthesis is required for HD-ZIP III-mediated xylem patterning

Abstract

The development and growth of higher plants is highly dependent on the conduction of water and minerals throughout the plant by xylem vessels. In Arabidopsis roots the xylem is organized as an axis of cell files with two distinct cell fates: the central metaxylem and the peripheral protoxylem. During vascular development, high and low expression level of the class III HD-ZIP transcription factors promotes metaxylem and protoxylem identities, respectively. Protoxylem specification is determined by both mobile, ground tissue emanating miRNA165/6 species which downregulate, and auxin concentrated by polar transport which promotes HD-ZIP III expression. However, the factors promoting high HD-ZIP III expression for metaxylem identity have remained elusive. We show here that auxin biosynthesis promotes HD-ZIP III expression and metaxylem specification. Several auxin biosynthesis genes are expressed in the outer layers surrounding the vascular tissue in Arabidopsis root and downregulation of HD-ZIP III expression accompanied by specific defects in metaxylem development is seen in mutants involved in auxin biosynthesis, such as trp2-12, wei8 tar2, or a quintuple yucca mutant, and in plants treated with L-Kynurenine, a pharmacological inhibitor of auxin biosynthesis. Some of the patterning defects can be suppressed by synthetically elevated HD-ZIP III expression. Taken together, our results indicate that polar auxin transport, which was earlier shown to be required for protoxylem formation, is not sufficient to establish a proper xylem axis but that root-based auxin biosynthesis is additionally required.

INTRODUCTION

The plant vascular system, which connects all plant organs, is composed of two conducting tissue types, xylem and phloem. While phloem conducts the photoassimilates produced in the shoot to all other parts of the plant, xylem functions in the conduction of water and minerals taken up by the roots and provides structural support. In the Arabidopsis root, xylem is organized as a central axis consisting of protoxylem (PX) with spiral secondary cell walls at the marginal positions and metaxylem (MX) with pitted walls at the central position (Fig. 1A,C).

Fig. 1.

Xylem phenotype in trp2-12 mutant. (A) Schematic representation of the Arabidopsis root. (B) Seedling phenotype of 5-day-old wild-type (C24), trp2-12 mutant grown on MS medium with and without L-Trp. (Scale bar: 0.5 cm.) (C-D) Basic fuchsin-stained xylem of 5 day-old wild-type and trp2-12. Yellow arrows indicate protoxylem; blue arrows indicate metaxylem. (Scale bars: 25 μm.) (E) Characterization of the xylem phenotype in 5 day-old wild-type (n=28) and trp2-12 (n=28). MX, metaxylem; PX, protoxylem. (F-G) pAHP6::GFPer in wild-type and trp2-12 (Scale bars: 50 μm). White arrows indicate pAHP6::GFPer signal in protoxylem. (H-J) pACL5::GFPer expression in 5-day-old wild-type and trp2-12 roots. (Scale bars: 50 μm for longitudinal and 10 μm for cross-sections).

We have previously shown that the class III HD-ZIP genes, PHABULOSA (PHB), INTERFASCICULAR FIBERLESS1/REVOLUTA (IFL/REV), PHAVOLUTA (PHV), CORONA (CNA/AtHB15) and AtHB8, act in a redundant and dosage-dependent manner to control PX/MX patterning and that their expression along the xylem axis is regulated by the mobile microRNA165/6 species emanating from the ground tissue. High and low levels of HD-ZIP III confer MX and PX specification, respectively, whereas no HD-ZIP III gene expression results in the lack of xylem formation (Carlsbecker et al., 2010; Miyashima et al., 2011).

Besides miRNAs, auxin transport and signaling have been associated with the regulation of HD-ZIP III genes. It is well established that spatio-temporal accumulation of auxin regulates various developmental processes (Sauer et al., 2013; Vanneste and Friml, 2009). Both genetic and pharmacological approaches revealed that auxin concentration maxima (or minima) are established and maintained by polar auxin transport (PAT) employing specific transporters such as PINs, AUXs/LAXs and PGPs (Benkova et al., 2003; Cho et al., 2007; Friml et al., 2002, 2004, 2003; Galweiler et al., 1998; Geisler et al., 2005; Grieneisen et al., 2007; Petrasek et al., 2006; Sorefan et al., 2009; Swarup et al., 2008; Weijers and Jurgens, 2005) and local auxin biosynthesis (Cheng et al., 2006, 2007; Ikeda et al., 2009; Ljung et al., 2005; Petersson et al., 2009; Stepanova et al., 2008; Tao et al., 2008; Zhao, 2010).

Recently, we provided evidence that certain PIN auxin efflux carriers in procambial cells adjacent to the xylem axis transport auxin towards the PX position and that an auxin maximum at this position is required for protoxylem formation (Bishopp et al., 2011; Mahonen et al., 2006).

Previously, it has been demonstrated that HD-ZIP III gene expression patterns overlap with the pattern of auxin distribution (Izhaki and Bowman, 2007). One of the auxin targets at the PX position is the ATHB8 gene (Bishopp et al., 2011). Lowering auxin levels at the PX position by interfering with PAT also lowered AtHB8 expression at this position (Bishopp et al., 2011). In vascular cell specification in leaves AtHB8 is co-expressed with MONOPTEROS/ARF5 (MP), a member of the Arabidopsis AUXIN RESPONSE FACTOR family, and PIN1 and functions as a positive regulator of procambial cell proliferation and xylem differentiation (Baima et al., 1995, 2001; Mattsson et al., 2003; Scarpella et al., 2006; Schuetz et al., 2013; Wenzel et al., 2007). A regulatory element in the promoter sequence of AtHB8 has been found to be directly bound by MP, which thereby regulates auxin-responsive AtHB8 expression (Donner et al., 2009).

In contrast to PX specification, impaired polar auxin transport has little impact on MX specification (Bishopp et al., 2011). In order to identify the factors promoting MX specification, we characterized two allelic loss-of-function mutants with dramatically impaired MX development but relatively normal PX formation. Characterization of these mutants, which we isolated from a forward genetic screen, revealed the involvement of auxin biosynthesis in MX specification.

Auxin is primarily produced via tryptophan (Trp)-dependent biosynthesis pathways but a Trp-independent auxin biosynthesis pathway has been also postulated (Chandler, 2009; Cohen et al., 2003; Korasick et al., 2013; Normanly et al., 1993; Ostin et al., 1999; Strader and Bartel, 2008; Woodward and Bartel, 2005; Zhao, 2010). Three gene families are involved in catalyzing specific Trp-dependent auxin biosynthesis steps, the TRYPTOPHAN AMINOTRANSFERASE OF ARABIDOPSIS1/TRYPTOPHAN AMINOTRANSFERASE-RELATED (TAA1/TAR) and YUCCA (YUC), which have recently shown to function in the same pathway (Mashiguchi et al., 2011; Stepanova et al., 2011; Won et al., 2011; Zhao, 2012) and CYTOCHROME P450 79B2/B3 (CYP79B2/B3) which define a distinct pathway (Sugawara et al., 2009; Zhao et al., 2002). The auxin precursor Trp is synthesized from chorismate via several catalytic steps (Radwanski and Last, 1995). The final step, the conversion of indole-3-glycerol phosphate to Trp, is catalyzed by an enzyme complex consisting of TRYPTOPHAN SYNTHASE α- and β-subunits (Barczak et al., 1995; Last et al., 1991). In Arabidopsis two homologous genes encode for α -subunits and four genes for β -subunits (TSB1, TSB2, TSB3 and TSBt2) (Yin et al., 2010; Zhang et al., 2008).

We show that the TSB1 mediated tryptophan biosynthesis and the auxin biosynthesis genes TAR1, TAR2 and YUCCA 3, 5, 7, 8 and 9 are required for high HD-ZIP III expression and MX differentiation in the vascular cylinder of the Arabidopsis primary root. Interestingly, PX differentiation is hardly affected by impaired auxin biosynthesis, whereas inhibition of PAT results in loss of PX but normal MX formation. This suggests that root-based auxin biosynthesis and PAT have to be spatially coordinated during xylem development in Arabidopsis.

RESULTS

Isolation of mutants with reduced HD-ZIP III expression and defects in metaxylem development

In wild-type Arabidopsis plants, GFP expressed under control of the companion cell specific SUCROSE TRANSPORTER 2 promoter (pSUC2) moves into the sieve elements and is unloaded symplastically into the root tip (Stadler et al., 2005). To identify novel genes involved in vascular development, a forward genetic screen was performed and several mutant seedlings that failed to unload GFP into the root tip were isolated (supplementary material Fig. S1B,C). Two of the identified mutants, trp2-12 and trp2-13, exhibited similarly severe overall phenotypes, including small leaves with dark veins and short agravitropic primary roots (Fig. 1B). Additionally the mutants showed altered cellular organization around the quiescent center (supplementary material Fig. S1E,F). An allelism test revealed that the two mutants were allelic (Fig. 3B).

Fig. 3.

Cloning of trp2-12 – mutant with reduced auxin levels in the stele. (A) Schematic view of the intron-exon structure of the TSB1 and sites of the trp2-12 and trp2-13 mutations. The trp2-12 mutation is at position 944 of the coding sequence and causes a C to T change, whereas trp2-13 is at position 929 of the coding sequence and causes a G to A change. (B) Seedling phenotype of 5 day-old wild-type (C24), trp2-12, trp2-13, trp2-8, trp2-12 x trp2-8 F1 progeny and trp2-12 rescued by pTSB1:TSB1:YFP. (Scale bar: 0.5 cm) (C) pIAA2:GFP/GUS in a 5-day old trp2-12 plant germinated on MS with L-TRP. (D) pIAA2:GFP/GUS in a 3-day old trp2-12 plant germinated on MS plates with L-TRP and transferred to MS media without L-TRP. (E,G) pDR5::GFP in 5-day old trp2-12 germinated on MS plates with L-TRP. (F,H) pDR5::GFP in 3-day old trp2-12 germinated on MS plates with L-TRP, transferred to media without L-TRP and grown for a further 3 days (Scale bars: 50 μm for longitudinal and 10 μm for cross-sections). (I) pTSB1::GFP/GUS expression. (Scale bar: 50 μm).

We next analyzed the status of vascular tissue differentiation in trp2-12 roots. Examination of mutant root cross sections and the fuchsin stained xylem axis revealed not only an altered organization of the phloem poles (supplementary material Fig. S1G,H) but also an interesting xylem patterning defect (Fig. 1C-E). Wild-type roots develop a xylem axis consisting of two metaxylem (MX) files flanked by two protoxylem (PX) poles (21/21) (Fig. 1C,E). By contrast, while xylem cells with the morphological characteristics of PX were formed in trp2-12 roots, MX formation was severely impaired (20/28) (Fig. 1D,E). These mutants either lack MX (Fig. 1D), or form discontinuous MX strands or only one strand. We also observed ectopic PX formation in the MX position in some trp2-12 roots (Fig. 1E). Consistent with this observation, the expression of the PX marker pAHP6::GFPer appears to be unaffected in trp2-12 roots (Fig. 1F,G), whereas the expression levels of the MX/procambium marker pACAULIS5 (pACL5)::GFPer (Muniz et al., 2008) is severely reduced or absent (Fig. 1H-J).

Because the HD-ZIP III genes play a critical role in xylem specification, we next assessed whether their expression levels are altered in the trp2-12 mutant. We performed a qRT-PCR assay and discovered reduced transcript levels of PHB, AtHB8, AtHB15 and PHV, whereas REV expression was slightly elevated in 5 day old trp2-12 mutant root tips in comparison to wild-type (Fig. 2H). In addition, down-regulation of some of the HD-ZIP III genes, including PHB, AtHB8 and AtHB15, was confirmed by RNA in-situ hybridization (Fig. 2E-G) and by examining the reporter constructs pAtHB8::GFPer (supplementary material Fig. S2D,E) and pPHB::PHB:GFP (Fig. 2A,B). Taken together, our expression analyses all indicate that the expression of most of the HD-ZIPIII genes is downregulated in the trp2-12 mutant.

Fig. 2.

HD-ZIP III gene expression in the trp2-12 mutant. (A-B) pPHB::PHB-GFP in wild-type and trp2-12. (C-D) pPHB::PHB-GFP expression in 6 day-old trp2-12 plants which were first germinated and grown on MS media supplied with L-TRP for three days (C), then transferred to the media without L-Trp and analyzed three days after transfer. (Scale bars: 50 μm). (E-G) In situ hybridization with PHB (E), ATHB8 (F) and ATHB15 (G) mRNA-specific probe on cross-sections of the root meristem in wild-type and trp2-12 mutant. (Scale bars: 10 μM) (H) PHB, ATHB8, PHV, REV and ATHB15 transcript levels in wild-type and trp2-12 root tips. (I) PHB, ATHB8, PHV, REV and ATHB15 transcript levels in 6-day old wild-type and trp2-12 seedlings germinated on MS plates with L-TRP and transferred to MS plates without L-Trp. The root tips were collected 3 days after transfer off L-Trp. The levels of HD-ZIP III transcripts in trp2-12 (red columns) are presented as relative values compared with the mean values of wild type (blue columns). ACTIN2 was used as a reference. Data are shown as mean±s.d.

Tryptophan biosynthesis is required for HD-ZIP III gene expression and vascular patterning

Next we set out to identify the mutant locus via a PCR-based positional cloning approach. Rough mapping revealed that the locus was located between the markers N5-22116697 and N5-22414573 on chromosome 5 within an approx. 300 kb window. Subsequent sequencing of candidate genes identified point mutations for both alleles in the At5g54810 locus, which encodes the TRYPTOPHAN SYNTHASE BETA-SUBUNIT 1 (TSB1/TRP2). The mutant alleles were mis-sense mutations causing the following amino acid changes: A315V (trp2-12) and G310D (trp2-13) (Fig. 3A). In order to verify that the phenotype is indeed caused by the mapped mutations we crossed trp2-12 and trp2-13 (expressing pSUC2::GFP) with the earlier described trp2-8 allele (Barczak et al., 1995). Analysis of the F1 progenies revealed that the mutants are allelic (Fig. 3B). Additionally we were able to complement trp2-12 (and trp2-13) with a 5080 bp genomic DNA encompassing the gene (1930 bp) and the upstream region (3150 bp). We subsequently made a translational and transcriptional fusion to investigate TSB1 expression during root development. The translational pTSB1::TSB1-YFP was able to complement the trp2-12 phenotype (Fig. 3B). The fusion protein was observed in plastids, as predicted in silico (supplementary material Fig. S1O-Q). Introduction of the transcriptional pTSB1::GFP/GUS fusion construct to WT revealed that TSB1 is highly expressed in the root meristem, especially in the early ground tissue and lateral root cap (Fig. 3I).

It has been previously shown that tryptophan synthase mutants can be rescued by externally supplied L-Tryptophan (L-Trp) (Jing et al., 2009; Last et al., 1991; Last and Fink, 1988; Muller and Weiler, 2000). We therefore examined trp2-12 mutants grown on media supplied with 10 μM L-Trp and found that the mutant phenotypes, including phloem unloading and xylem patterning defects, were completely rescued (supplementary material Fig. S1I-K). Consistent with this, expression of pACL5::GFPer (supplementary material Fig. S2A-C) and pSUC2::GFP (supplementary material Fig. S1B-D) was indistinguishable between WT and the L-Trp supplied trp2-12 mutant.

This conditional phenotype allowed us to first grow trp2-12 seedlings on media supplied with L-Trp for 3 days and then transfer the seedlings to media lacking L-Trp to observe whether altered HD-ZIP III gene expression could be detected prior to the onset of morphological alterations in the root meristem. Reduced HD-ZIP III expression levels were observed in trp2-12 seedling three days after transfer based on qRT-PCR (Fig. 2I), in-situ hybridization (supplementary material Fig. S2F-K) and pPHB::PHB-GFP expression (Fig. 2C,D). The downregulation of HD-ZIP III expression occurs prior to obvious developmental changes, as the loss of metaxylem was not seen until 10 days after transfer. These findings show that TSB1-mediated production of tryptophan influences HD-ZIP III gene expression and xylem patterning in the root.

Trp-dependent auxin biosynthesis via the TAA1/TAR – YUCCA pathway contributes to vascular patterning

During root development in Arabidopsis a distinct auxin maximum is formed at the site of the forming xylem axis (Bishopp et al., 2011). As Trp is a precursor of auxin biosynthesis, we next analyzed whether auxin levels might be altered in trp2-12 seedlings. As expected, the auxin signaling reporter pDR5::GFPer was detected at a reduced level in the root meristem and xylem axis of five day old trp2-12 seedlings (supplementary material Fig. S1L-N). Furthermore, a similar reduction in the expression levels of auxin-inducible reporters pDR5::GFPer and pIAA2::GFP-GUS was observed 2-3 days after the transfer of trp2-12 seedlings from media containing L-Trp to media lacking L-Trp (Fig. 3C-H). Importantly, pDR5::GFPer expression in the xylem axis was severely diminished after transferring to media lacking L-Trp (Fig. 3G,H). We conclude that trp2-12 mutants have reduced auxin levels in the xylem axis.

Screening the Arabidopsis eFP Browser (Winter et al., 2007) revealed a complex expression pattern of genes involved in Trp-dependent auxin biosynthesis with several of these genes preferentially expressed in the quiescent center and its surrounding tissues. We decided to investigate whether TAA1 (WEI8), TAR2 and YUCCA genes might have a function in vascular tissue development, as TAA1/TAR genes act upstream of YUCCA genes in the same biosynthetic pathway (Mashiguchi et al., 2011; Stepanova et al., 2011; Won et al., 2011). pTAA1::GFP-TAA1 and pTAR2::GUS expression has previously been detected in two non-overlapping domains of the root meristem. The strongest pTAA1::GFP-TAA1 expression was detected in the QC and protoxylem as reported before (Fig. 4C). No activity of pTAR2::GUS was observed in the meristem region, though pTAR2::GUS expression was detected in the leaf veins (Fig. 4A,B, supplementary material Fig. S3D and E) (Stepanova et al., 2005, 2008). We confirmed the lack of TAR2 expression in the root meristem region by in situ mRNA hybridization (supplementary material Fig. S3A-C). Both pYUCCA8::GFP and pYUCCA9::GFP are highly expressed in the QC, endodermis/cortex precursor cells and columella (Fig. 4D,E) whereas pYUCCA3::GUS expression can be detected in ground tissue cells in the elongation zone (Fig. 4F).

Fig. 4.

The expression patterns of auxin biosynthesis genes. (A-B) pTAR2::GUS expression. (C) pTAA1::GFP-TAA1 expression. (D) pYUCCA8::GFP expression. (E) pYUCCA9::GFP expression. (F) pYUCCA3::GUS expression (Scale bars: 50 μm).

Next we analyzed the status of xylem formation in the combinatorial mutants for both TAA1/TAR and YUCCA genes. Interestingly, while PX was properly formed at the peripheral positions of the xylem axis in both wei8 tar2 double and quintuple yucca mutant (lacking yuc3, yuc5, yuc7, yuc8 and yuc9), ectopic PX formation was observed in the central MX position (Fig. 5B,E). We observed a reduction of pACL5::GFPer expression in the wei8 tar2 mutant (Fig. 5J-M). These observations all suggest that TRP-dependent auxin biosynthesis via the TAA1/TAR-YUCCA pathway in the non-vascular tissues is required for MX differentiation in the central domain of xylem axis in Arabidopsis root.

Fig. 5.

Auxin biosynthesis via the TAA1/TAR – YUCCA pathway contributes to vascular patterning. (A-I) Basic fuchsin-stained xylem of wild-type (A), wei8 tar2 (B), axr3-3 (C-D), quintuple yucca (E), wild-type plants grown on MS supplemented with L-Kynurenine (Kyn) (F), pTSB1::iaaH plants grown with Kyn (G), pTSB1::iaaH grown with Kyn and IAM (H), pWOX5xve::iaaH grown with Kyn, IAM and β-estradiol (I). (Scale bars: 25 μm). Blue arrows indicate metaxylem and yellow arrows indicate protoxylem. (J-M) pACL5::GFPer expression in wild-type and in the wei8 tar2 mutant (the tip region is shown at higher magnification in L and M). (Scale bars: 50 μm). (N) Characterization of the xylem phenotype in wild-type (n=21), wei8 tar2 (n=21), pTSB1::iaaH (n=21), pTSB1::iaaH+Kyn (n=21), pTSB1::iaaH+Kyn+IAM (n=21).

It was recently shown that L-Kynurenine (Kyn) acts as an alternate substrate for TAA1/TARs and competitively blocks the activity of these enzymes. Kyn treatment mimics the loss of TAA1/TAR function in Arabidopsis and can be used to manipulate TAA1/TAR-dependent auxin biosynthesis (He et al., 2011). To test whether we could pharmacologically block MX differentiation, we treated WT and transgenic lines with Kyn (Fig. 5F-I). Following Kyn treatment, we observed a reduction of pACL5::GFPer expression (Figs 6A-C and 7J). We also observed reduced pIAA2::GFP/GUS expression and elevated levels of auxin-downregulated reporter DII-VENUS (Brunoud et al., 2012) in xylem axis, confirming the inhibitory effect of Kyn on auxin biosynthesis (supplementary material Fig. S4A-J). As expected, Kyn treatment also led to the conversion of MX to PX in the central domain of xylem axis, as observed in wei8 tar2 mutant (Fig. 5B,F). These results suggest that the PX phenotype in wei8 tar2 and Kyn-treated plants is caused by a reduction of auxin levels in the xylem axis.

Fig. 6.

The different effects of auxin biosynthesis and polar auxin transport on ACL5 and AHP6 expression. (A-C) pACL5::GFPer in wild-type seedlings germinated on MS without Kyn (A) and transferred to a medium with 100 μM Kyn, and analysed after 24 (B) and 48 (C) hours of treatment. (D) pACL5::GFPer expression in 5 day-old roots grown on a medium with 5 μM NPA. (E-G) pAHP6::GFPer in wild-type roots grown on MS medium without Kyn (E), transferred to medium with Kyn and analysed after 24 (F) and 48 (G) hours.<CE1> (H) pAHP6::GFPer expression in 5 day-old roots grown on media with 5 μM NPA (Scale bars: 50 μm for longitudinal and 10 μm for cross-sections). White arrows indicate protoxylem position.

Fig. 7.

Auxin biosynthesis influences xylem patterning by regulating the level of HD-ZIP III expression. (A-D) pPHB::PHB:GFP expression in wild-type plants grown without Kyn and transferred to a medium with Kyn for 24 and 48 h (Scale bars: 50 μm). (E-I) Basic fuchsin-stained xylem of wild-type, phb-7d, phb-7d+Kyn, pCRE1::PHB and pCRE1::PHB+Kyn. Scale bars: 25 μm. Blue arrows indicate metaxylem and yellow arrows indicate protoxylem. (J) PHB, ATHB8, PHV, REV, ATHB15 and ACL5 transcript levels in wild-type root tips grown without Kyn and transferred to a medium with Kyn for 24 and 48 h. The levels of HD-ZIP III transcripts in Kyn-treated wild-type root tips (red and green columns) are presented as relative values compared with the mean values of untreated wild type (blue columns). ACTIN2 was used as a reference. Data are shown as mean±s.e.m. (K) Characterization of the xylem phenotype in wild-type, wild-type grown with Kyn, phb-7d, phb-7d grown with Kyn, four independent pCRE1::PHB lines grown with and without Kyn.

To further confirm that reduced auxin levels cause the conversion of MX to PX in the central domain of xylem axis, we analyzed the effect of Kyn on the two transgenic lines expressing the bacterial auxin biosynthesis gene iaaH (Schroder, 1984; Thomashow et al., 1984) in the root meristem (pTSB1::iaaH) or QC (pWOX5::XVE>>iaaH), respectively. Whereas Kyn treatment caused ectopic PX formation in both the transgenic plants and in wild type, reduced responsiveness to Kyn was observed when seedlings were additionally supplied with IAM (indoleacetamide) or IAM+estradiol, respectively (Fig. 5F-I). This suggests that iaaH-mediated endogenous auxin production in the non-vascular tissue can restore sufficient auxin levels for proper MX formation under conditions inhibiting TAA1/TAR auxin biosynthetic pathway.

Although we confirmed that Trp-dependent auxin biosynthetic enzymes are expressed in the root meristem, it has also been proposed that several types of phytohormone, including IAA and cytokinin, can be transported from shoot to root through the phloem network (Guo et al., 2005; Overvoorde et al., 2010; Reed et al., 1998). In order to verify whether root-based auxin biosynthesis is sufficient for MX formation, we performed a “shoot-off” experiment where we followed the xylem formation in root after cutting the shoot. Four days after removal of the shoot, we still observed proper MX and PX formation, indicating that auxin produced in the QC and its surrounding tissues is sufficient to maintain normal xylem formation (supplementary material Fig. S4K,L).

Examination of the stabilized auxin signaling inhibitor axr3-3 revealed a lack of MX development, delayed MX formation or an all PX phenotype (10/17 develop extra PX) (Fig. 5C,D), providing further evidence that auxin biosynthesis and auxin signaling is required for MX differentiation.

The relationship between auxin biosynthesis and auxin transport during root procambial development

Recently it was shown that PAT generates an auxin response maximum at the PX position of the xylem axis and that this auxin maximum is required for the correct expression of AHP6, an inhibitor of cytokinin signaling promoting PX differentiation. While genetic and pharmacological inhibition of PAT results in the loss of PX, MX formation is not affected, indicating a differential requirement of PAT between the two xylem domains (Bishopp et al., 2011). By contrast, the trp2-12, wei8 tar2 and quintuple yucca mutant with reduced auxin biosynthesis fail to develop MX development but show normal PX development.

To further analyze the differential effects of auxin biosynthesis and polar auxin transport on xylem development, we assayed the expression of pIAA2::GFP/GUS, pAHP6::GFPer and pACL5::GFPer after NPA and Kyn treatment (Fig. 6 and supplementary material Fig. S4A-H). Bishopp et al. (2011) demonstrated that blocking of PAT by NPA results in a complete loss of pIAA2::GFP/GUS expression in the entire stele, loss of pAHP6::GFPer expression in the PX position and expansion of the signal in the pericycle (Bishopp et al., 2011) (Fig. 6H). However, we were able to demonstrate that the expression of the the MX marker pACL5::GFPer was not altered after NPA treatment (Fig. 6D). In contrast, Kyn treatment did not affect the expression pattern of pAHP6::GFPer (Fig. 6F,G) but caused a reduction of pACL5::GFPer expression levels (Fig. 6B,C).

Auxin biosynthesis and auxin signaling promotes HD-ZIP III expression in the root meristem

After having shown that auxin biosynthesis is required for MX development, we next tested whether Trp-dependent auxin biosynthesis is required for expression of the HD-ZIP III genes. A qRT-PCR assay revealed reduced expression of PHB, AtHB8 and AtHB15 24 h after Kyn treatment of WT seedlings, but slightly elevated levels of REV; within 48 h after treatment, expression of all five HD-ZIP III genes was reduced (Fig. 7J). Furthermore, pPHB::PHB-GFP expression levels are reduced after Kyn treatment (Fig. 7A-D). Since HD-ZIP III genes are negatively regulated by miRNA165/166, we analyzed whether Kyn treatment altered the expression of miRNA165/6; however, no change in the promoter activity of either MIRNA165A or MIRNA166B was observed (supplementary material Fig. S5A-F).

To verify that auxin-regulated high levels of HD-ZIP III expression are required for MX development, we assayed transgenic lines expressing PHB under the control of the auxin insensitive pCRE1 promoter (supplementary material Fig. S6). pCRE1::PHB expressing lines developed a normal xylem axis (Fig. 7H) but the seedlings were partially resistant to Kyn; fewer seedlings (<15%) developed ectopic PX in the MX position (Fig. 7I,K). Furthermore, we examined whether the ectopic MX phenotype of the phb7-d mutant, which escapes downregulation by miRNA165/166, could be reverted by Kyn treatment (Fig. 7F,G,K). Although we were not fully able to restore the WT xylem axis pattern (2 marginal PX and 2 central MX files) (Fig. 7E), partial rescue of PX development was observed in phb7d mutants (14/19 develop 2 or more PX files) (Fig. 7G,K), indicating that auxin biosynthesis influences xylem patterning by regulating the level of HD-ZIP III expression.

DISCUSSION

Various developmental processes depend on the establishment and maintenance of local auxin maxima by PAT and auxin biosynthesis (Cheng et al., 2007; Ikeda et al., 2009; Ljung et al., 2005; Stepanova et al., 2008; Tao et al., 2008; Vanneste and Friml, 2009).

Here we provide evidence that Trp-dependent auxin biosynthesis via the TAA1/TAR-YUCCA pathway is required for HD-ZIP III mediated metaxylem (MX) formation in the xylem axis of Arabidopsis roots. Our data further suggest that both processes, PAT and auxin biosynthesis, are required for proper xylem patterning, but that they might have different roles during xylem patterning. PAT specifies PX formation at the xylem axis periphery whereas auxin biosynthesis is required for MX differentiation in the center of the xylem axis (Fig. 8).

Fig. 8.

TRP-dependent auxin biosynthesis is required for HD-ZIP III-mediated metaxylem formation.

Trp is believed to be the main precursor of IAA biosynthesis in plants and is synthesized from chorismate via several catalytic steps (Ljung, 2013; Radwanski and Last, 1995). The final step, the conversion of indole-3-glycerol phosphate to Trp, is catalyzed by an enzyme complex consisting of the TRYPTOPHAN SYNTHASE α- and β-subunits (Barczak et al., 1995; Last et al., 1991). By examining two novel loss of function alleles of the TSB1/TRP2 β subunit, we revealed that TSB1-mediated Trp synthesis participates in the generation of auxin maxima in the root tip, although previous studies detected no altered active auxin levels in trp2 mutant seedlings (Last et al., 1991; Normanly et al., 1993; Radwanski and Last, 1995). This discrepancy might be due to local changes in auxin levels having minimal impact on total auxin levels.

Furthermore, trp2-12 mutants show a misspecification of xylem identity in the central domain of the stele, where metaxylem elements are normally formed. This phenotype could be rescued by externally supplying L-Tryptophan.

Similar xylem patterning defects were observed in seedlings in which Trp-dependent auxin biosynthesis was genetically or pharmacologically impaired. Both the double taa1(wei8)/tar2 and quintuple yuc3,5,7,8,9 mutant, as well as Kyn treated WT seedlings, form ectopic PX instead of MX in the center of the xylem axis, whereas the formation of PX in the peripheral xylem is not altered. Consistent with the xylem phenotypes, expression of the metaxylem marker ACL5 was downregulated under conditions inhibiting Trp-dependent auxin biosynthesis, whereas expression of the protoxylem marker AHP6 was still retained. By expressing the bacterial auxin biosynthesis gene iaaH in the root meristem, we were able to render xylem patterning resistant to Kyn treatment. This indicates that auxin biosynthesis is required for proper MX formation.

Surprisingly, reporter gene analysis revealed that TSB1, TAA1 and the analyzed YUCCA genes are expressed in tissues surrounding the vascular cylinder in the root meristem. Only TAA1 is additionally expressed in the vascular tissue (He et al., 2011; Stepanova et al., 2008). Based on the expression pattern of the tested genes, it appears that Trp synthesis and Trp-dependent auxin biosynthesis occur in the QC and its surrounding tissues – the ground tissue initials and columella. Furthermore, QC-specific expression of iaaH can rescue metaxylem formation in Kyn-treated roots, indicating that auxin produced in non-vascular tissue regulates xylem patterning.

Recently, it has been shown that PAT is required to establish an auxin response maximum at the periphery of the xylem axis which is needed for PX (Bishopp et al., 2011) but not for MX differentiation. As shown by Bishopp et al., after polar auxin transport inhibition by NPA the auxin-inducible reporter pIAA2::GFP/GUS signal disappeared from the xylem axis and became radially symmetric in the pericycle and outer tissues (Bishopp et al., 2011), indicating that auxin signaling is attenuated in the stele under NPA treatment. On the other hand, NPA treatment also results in the disappearance of auxin-downregulated reporter DII-VENUS in the same region (Santuari et al., 2011), suggesting that NPA treatment increases the auxin level in the whole vascular tissue. As reported before (Brunoud et al., 2012), the expression pattern of DII-VENUS is complementary with those of auxin signaling markers, such as DR5 and IAA2. However, under NPA treatment both IAA2 and DII-Venus expression result in the similar pattern. The similar behavior of IAA2, an output of the auxin response pathway, and DII-Venus, a target for auxin interaction that triggers its degradation, after NPA treatment could raise a possibility of a feedback loop which currently remains elusive and might involve some additional levels of regulation in MX formation in the central xylem. Alternatively, it might be possible that other unknown auxin-independent pathway, which is not visible in wild type condition, enhance the HD-ZIP III expression in MX under NPA treatment specifically, and this retains MX formation under NPA treatment. Surprisingly, in this work we discovered that MX differentiation in the central xylem domain depends on auxin biosynthesis whereas genetic and pharmacological interference with auxin production hardly affects peripheral PX specification. Although certain developmental processes such as leaf formation and root hair positioning require the combined action of both local auxin biosynthesis and PAT (Cheng et al., 2007; Ikeda et al., 2009) it was surprising to find that Trp-dependent auxin biosynthesis and PAT have different roles in xylem patterning. How the root-based auxin biosynthesis contributes to the creation of an auxin maximum and specifically regulates MX differentiation remains elusive. Since the inhibition of PIN-mediated auxin transport affects xylem formation only in the peripheral PX position, local auxin accumulation in the central xylem position might be regulated primarily by a PIN-independent mechanism which would be easily affected by local auxin levels.

Xylem cell fate has been shown to be determined by the HD-ZIP III transcription factors in a dose-dependent manner; high expression levels specify metaxylem fate while low levels specify protoxylem (Carlsbecker et al., 2010). Several reports indicate that AtHB8 is induced by auxin in the vasculature and might play a role in PX differentiation (Bishopp et al., 2011). Here, we show that Trp-dependent auxin biosynthesis promotes the expression of most HD-ZIP III genes in the root, revealing a novel mechanism regulating these transcription factors during vascular development. Evidence that auxin biosynthesis influences xylem patterning by regulating HD-ZIP III gene expression is provided by the phb-7d gain-of function mutant which expresses high levels of miR165/6-resistant PHB transcript form ectopic MX (Carlsbecker et al., 2010); blocking auxin biosynthesis in this mutant causes the conversion of MX to PX. However, in seedlings expressing PHB under the control of the auxin insensitive pCRE1 promoter, interfering with auxin synthesis has little effect on xylem patterning. Despite residual low HD-ZIP III expression levels, trp2-12 mutants often lack xylem in the center of the stele, while the wei8 tar2 and quintuple yucca mutants develop an all PX xylem axis. This phenotypic difference suggests that TSB1-mediated Trp synthesis also promotes xylem differentiation by an auxin-independent mechanism downstream of the HD-ZIP IIIs or by post-transcriptionally regulating them.

It has been shown that REV is able to stimulate auxin biosynthesis by regulating TAA1 and YUC5 in the shoot (Brandt et al., 2012). Therefore, it is likely that specific feedback mechanisms regulate auxin biosynthesis and HD-ZIP III expression in the root to control vascular patterning.

MATERIALS AND METHODS

Plant materials and growth conditions

The Arabidopsis thaliana accessions C24 and Col-0 were used in all experiments. The mutants and transgenic marker lines were obtained as follows: axr3-3 (Leyser et al., 1996), phb-7d (Carlsbecker et al., 2010), pDR5rev::GFP (Friml et al., 2003), pAHP6::GFPer (Mahonen et al., 2006), pAtHB8::GFP, pPHB::PHB:GFP, pMIR165a::GFP (Carlsbecker et al., 2010), pAtSUC2::GFP (Imlau et al., 1999), pIAA2::GFP/GUS (Bishopp et al., 2011), DII-Venus (Brunoud et al., 2012). wei8-1+/− tar2-1−/− (Stepanova et al., 2008) and trp2-8 (Barczak et al., 1995) were described previously and obtained from the Arabidopsis Stock Centre. The genotypes of the mutants were verified using PCR (primers used for genotyping are listed in supplementary material Table S4).

Seeds were sterilized in 70% ethanol and 0.05% Triton X-100, rinsed several times with sterile water, then sown on germination medium containing ½ MS (Murashige and Skoog, 1962), 1% sucrose, 0.05% MES (2-(N-morpholino) ethanesulfonic acid) and 1% phytagel. Seeds were then transferred to a growth room at 22°C under a 16-h light/8-h dark photoperiod. For treatment and transfer assays, the media was supplemented with 5 μM NPA (N-(1-naphthyl) phthalamic acid) (TCI), 20 μM L-Tryptophan (Sigma), 100 nM IAM (indoleacetamide) (Sigma), 100 μM Kyn (L-Kynurenine) (Sigma) or 5 μM 17β-estradiol (Sigma) as indicated in the text.

For anatomical, histological and reporter gene analyses primary roots of 4-6 days old vertically grown seedlings were used.

Screening and map-based cloning of trp2-12 and trp2-13 mutants

The trp2-12 and trp2-13 mutants were isolated in a genetic screen based on EMS mutagenesis of the mobile phloem marker pSUC2::GFP (Imlau et al., 1999) generated previously (Carlsbecker et al., 2010). To map the TSB1 locus, the trp2-12 mutant in C24 background was crossed to Columbia-0 (Col-0) and homozygous mutants were selected in the F2 population for mapping. Genomic DNA was extracted from F2 plants showing the trp2-12 phenotype and analyzed for co-segregation with respect to cleaved amplified polymorphic sequence (CAPS) and simple sequence length polymorphism (SSLP) markers (Bell and Ecker, 1994; Konieczny and Ausubel, 1993). These markers were derived from the Arabidopsis Information Resource (TAIR; http://www.arabidopsis.org). For oligonucleotide sequences, see supplementary material Tables S1 and S2. The rough mapping delimited trp2-12 to a 300 kb region between the CAPS markers N5-22116697 and N5-22414573 in the chromosome 5. A point mutation in TSB1 gene (At5g54810) was identified when the candidate genes were sequenced (primers used for sequencing are listed in supplementary material Table S3.

Gene expression analyses

To examine the transcript levels of HD-ZIP III genes, root tips from 5-day-old seedlings were collected. Differences in gene expression were measured by real time PCR as described previously (Carlsbecker et al., 2010).

In situ hybridization experiments

In situ hybridization and probe preparation were performed as in Carlsbecker et al. (2010).

Histology and confocal imaging

Transverse sectioning, fuchsin staining and confocal imaging were performed on 5-DAG seedlings as described previously (Mahonen et al., 2000), the confocal analysis was done on a Leica SP5 confocal (Leica TCS SP5 Laser Scanning Confocal Microscope, with given excitation (Ex) and emission (Em) wavelengths (Ex/Em): 488 nm/505-530 nm for GFP; 514 nm/525-555 nm for YFP and 488-514 nm/613-679 nm for propidium iodide). GUS staining was performed as described by Miyawaki et al. (2004). The plastic cross-sections for GUS samples were performed as described previously (Bishopp et al., 2011). The analysis of xylem tissue was carried out with 5 days old seedlings embedded to chloral hydrate solution (chloral hydrate – glycerol – water 8:1:2).

Generation of transgenic plants

pAHP6::GFPer, pPHB::PHB-GFP, pDR5::GFP, pIAA2::GFP/GUS, pSUC2::GFP and pATHB8::GFPer markers were introduced into trp2-12 mutant background by crossing and analyses were done on segregating F2 generations. The transgenic constructs were generated using the Gateway or Multisite Gateway system (Invitrogen) and all were introduced into plants via floral dip method (Clough and Bent, 1998).

To generate pTSB1::GFP-GUS, a 3.15 kb sequence corresponding to the pTSB1 promoter was amplified from Col-0 genomic DNA with primers 5′-AAAAAGCAGGCTACCATTGAAACAGAAGTGTGGTGATG -3′ (forward) and 5′-AGAAAGCTGGGTTACTGAATGAATCTCTCTCTGAGTC-3′ (reverse) primers, cloned into pDONR(Zeo) and recombined into the destination vector pBGWFS7 (Karimi et al., 2002).

To generate pTSB1::TSB1-YFP construct, a 1.93 kb sequence corresponding to the TSB1 gene was amplified from Col-0 genomic DNA using primers 5′-AAAAAGCAGGCTATGGCAGCCTCAGGCACCTCTGC-3′ (forward) and 5′-AGAAAGCTGGGTTAACATCAAGATATTTAGCCACTG-3′ (reverse) primers and cloned into pDONR(Zeo). pTSB1 promoter was amplified from genomic Col-0 DNA using primers ATAGAAAAGTTGCCATTGAAACAGAAGTGTGGTGATG (forward) and TTGTACAAACTTGTACTGAATGAATCTCTCTCTGAGTC (reverse) and cloned into pDONRP4-P1R (Invitrogen). The resulting entry clones were recombined together with YFP into the multi-site Gateway vector pBm43GW (Karimi et al., 2002).

To generate pCRE1::PHB, the entry clones containing CRE1 promoter (pCRE1) (Mahonen et al., 2000) and PHABULOSA (PHB) gene (Carlsbecker et al., 2010) were recombined into the multisite destination vector pBm43GW.

pTSB1 promoter and iaaH gene (Blilou et al., 2005) were recombined into the multisite destination vector pBm43GW to produce pTSB1::iaaH construct.

The inducible pWOX5 promoter (Sarkar et al., 2007) and iaaH gene were recombined into the multisite destination vector pBm43GW to produce pWOX5-xve::iaaH.

For the construction of pACL5-GFPer, 2.6-kb upstream region of ACL5 (At5g19530) was amplified from Col-0 genomic DNA with the primers.

5′-AAGTCGACGGCAAGCTATTCATTAGGC -3′ and 5′-TCGGATCCATCCAAGTTGAGGAGAAGATATAGAGA-3′ digested with SalI and BamHI, and placed upstream of the GAL4:VP16 (GV)-coding region of pBIN-UAS-GFPer-hssb-GV (Waki et al., 2013). The resulting construct, pBIN-UAS-GFPer-pACL5-GV, was introduced into the Arabidopsis Col-0 plant to give the pACL5-GFPer reporter line.