LAZY3 plays a pivotal role in positive root gravitropism in Lotus japonicus

LAZY3, polarly localized to the plasma membrane in root stele cells, is involved in rootward polar auxin transport in roots and required for positive root gravitropism in Lotus japonicus.

Abstract

LAZY1 family genes play important roles in both shoot and root gravitropism in plants. Here we report a Lotus japonicus mutant that displays negative gravitropic response in primary and lateral roots. Map-based cloning identified the mutant gene LAZY3 as a functional ortholog of the LAZY1 gene. Mutation of the LAZY3 gene reduced rootward polar auxin transport (PAT) in the primary root, which was also insensitive to the PAT inhibitor N-1-naphthylphthalamic acid. Moreover, immunolocalization of enhanced green fluorescent protein-tagged LAZY3 in L. japonicus exhibited polar localization of LAZY3 on the plasma membrane in root stele cells. We therefore suggest that the polar localization of LAZY3 in stele cells might be required for PAT in L. japonicus root. LAZY3 transcripts displayed asymmetric distribution at the root tip within hours of gravistimulation, while overexpression of LAZY3 under a constitutive promoter in lazy3 plants rescued the gravitropic response in roots. These data indicate that root gravitropism depends on the presence of LAZY3 but not on its asymmetric expression in root tips. Expression of other LAZY genes in a lazy3 background did not rescue the growth direction of roots, suggesting that the LAZY3 gene plays a distinct role in root gravitropism in L. japonicus.

Introduction

In plants, the direction of gravity is perceived mainly by gravity-sensing cells (statocytes) in both the endodermis of stems and the columella cells of the root cap. In both cell types, the displacement of amyloplasts is thought to trigger signal transduction, which induces the asymmetrical distribution of auxin that generates the curvature response in the elongation zones of each organ (Boonsirichai et al., 2002). Auxin is a crucial plant hormone that has a well-established role in plant gravitropism. In Arabidopsis, polar auxin transport (PAT) is mainly mediated by three families of plasma membrane-associated transporter proteins: AUXIN RESISTANT1/LIKE AUX1 (AUX1/LAX) influx carriers; PIN-FORMED (PIN) efflux carriers; and P-GLYCOPROTEIN ATP-binding cassette (PGP) auxin transporters. Following gravistimulation, PIN3 and PIN7 relocalization establishes an asymmetric auxin flow in root apices. The increased auxin levels on the lower side of the root lead to PIN2 protein retention within the plasma membrane, which facilitates shootward PAT towards the elongation zone at which root tip bending will occur (Blilou et al., 2005; Kleine-Vehn et al., 2010).

A growing body of evidence suggests that the IGT family plays a major role in lateral organ orientation. The IGT genes show a low level of sequence similarity but contain a universally conserved motif (GφL(A/T)IGT) in domain II (Dardick et al., 2013). Phylogenetic analyses have divided IGT genes into two clades, TAC1 and LAZY. The obvious structural difference between TAC1 and LAZY is the conserved C-terminal domain V, which is only present in LAZY (Dardick et al., 2013). Previous research has shown that TAC1 mutations result in narrower branch and tiller angles in both monocots (Yu et al., 2007; Ku et al., 2011) and dicots (Dardick et al., 2013). In contrast, deficiency in LAZY1 results in wider branch angles of lateral organs and reduced shoot gravitropism (Abe et al., 1996; Yoshihara and Iino, 2006; Li et al., 2007; Dong et al., 2013; Yoshihara et al., 2013; Howard et al., 2014). In addition to the positive role in shoot gravitropism, LAZY family genes are also involved in root gravitropism. The LAZY ortholog OsDRO1 controls root gravitropism in both seminal and crown roots, thus affecting general root system architecture (Uga et al., 2013). Mutation of the LAZY ortholog MtNGR in Medicago truncatula caused a negative root gravitropism phenotype (Ge and Chen, 2016). In Arabidopsis, four LAZY genes (LAZY1, 2, 3, and 4) are redundantly involved in shoot and root gravitropism (Taniguchi et al., 2017; Yoshihara and Spalding, 2017).

Although recent research using various lazy mutants has yielded important insights into the molecular mechanisms affecting root gravitropism, the incomplete, and sometimes controversial, evidence still needs to be explained. Uga et al. (2013) have shown that DRO1 does not interfere with auxin transport. However, other researchers have suggested that LAZY1 plays a positive role in gravity-induced lateral transport following observations that the lateral auxin gradient was strongly inhibited in lazy1 mutant coleoptiles (Yoshihara and Iino, 2006; Li et al., 2007; Dong et al., 2013). Furthermore, the asymmetric distribution of PIN3 and DR5rev:GFP expression reversed in Atlazy1,2,3 triple mutants, resulting in negative gravitropism in lateral roots. This suggests that the LAZY family regulate PAT by establishing asymmetric PIN3 expression in the root cap columella (Taniguchi et al., 2017).

Here we report how a Lotus japonicus mutant exhibits negative gravitropism in both primary and lateral roots. The mutated gene cloned through a map-based approach belongs to a LAZY subfamily. There are six LAZY family genes in L. japonicus named according to their location on the chromosomes (see Supplementary Table S1 at JXB online). We therefore named the mutant as Ljlazy3. The LjLAZY3 protein was polarly localized on the plasma membrane in L. japonicus root stele cells. LjLAZY3 mutation both decreased rootward PAT and reversed gravity-induced asymmetric auxin flow in the primary root tip. Mechanisms for how LjLAZY3 could affect root gravitropism are discussed.

Materials and methods

Plant materials and growth conditions

Ljlazy3-1 mutant was obtained from the progeny of L. japonicus ecotype Miyakojima MG-20 (MG-20). Ljlazy3-2 and Ljlazy3-3 mutants were obtained from the LORE1 collection (Urbański et al., 2012). MG-20 is the wild-type (WT) for Ljlazy3-1 mutant and L. japonicus ecotype Gifu B-129 (Gifu) is the WT for LORE1 mutants.

Surface-sterilized seeds were germinated on filter paper. Seedlings were transferred to pots with horticultural soil containing vermiculite (1:3) and grown in a greenhouse at 26 °C (day, 16 h) and 22 °C (night, 8 h). For gravistimulation tests, seedlings were grown on agar plates containing 1/2 B&D medium for 3 d. The seedlings were shifted on the surface of vertically oriented agar plates so that the roots were in a horizontal position. Root curvature in response to gravity was measured at different time points after the start of the experiment.

In the auxin and auxin transport inhibition experiments, seedlings were transferred to plates containing indole-3-acetic acid (IAA), 1-naphthaleneacetic acid (NAA), 2,4-dichlorophenoxyacetic acid (2,4-D) or 1-N-naphthylphthalamic acid (NPA).

Fine mapping of the LjLAZY3 locus

We generated a mapping population by crossing the Ljlazy3-1 line with Gifu. To map the LjLAZY3 locus, 36 pairs of simple sequence repeat (SSR) markers evenly distributed in the L. japonicas genome were selected to screen for polymorphism among the parents and the gene pools. BM2375 and TM1350 revealed polymorphism among the parents and the gene pools, respectively. Then, more than eight SSR markers were designed between BM2375 and TM1350 for fine-mapping. Using BLASTN (http://www.arabidopsis.org), a total of 18 genes were predicted to be present within this region. We then amplified and sequenced all of the genes in this region.

RNA isolation and expression analysis

Total RNA was extracted from L. japonicus tissue using the RNeasy Plant Mini Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. First-strand cDNA was synthesized from 2 µg of RNA using M-MLV reverse transcriptase (Promega, Madison, WI, USA) according to the manufacturer’s instructions. Real-time PCR was performed using a LightCycler 480 instrument (Roche Life Science, Penzberg, Germany). Forty cycles were performed under the following cycling conditions: 95 °C for 30 s; 95 °C for 5 s; 60 °C for 20 s; and 72 °C for 20 s. ATP-synthase (ATPase), Ubiquitin-conjugating enzyme (UBC) and Protein phosphatase 2A (PP2A) served as the reference genes. Relative quantification software (Roche) was used to calculate normalized, efficiency-corrected relative transcript levels.

Plasmid constructs

The β-glucuronidase (GUS) reporter gene was used for the promoter assay. All promoters were amplified by PCR and the resultant fragments were subcloned into pMD19-T (Takara, Kyoto, Japan) for sequence confirmation. All primers used in this study are listed in Supplementary Table S2. The promoter regions of LAZY homologs in L. japonicus and LjPIN2 were amplified with either PstI and EcoRI (pLAZY1, 2054 bp; pLAZY2, 1500 bp; pLAZY4, 2015 bp; pLjPIN2, 1928 bp) sites or HindIII and EcoRI sites (pLAZY3, 2714 bp) using genomic DNA as the template. The confirmed fragments were cloned into either PstI/EcoRI site or HindIII/EcoRI site in the binary vector pCAMBIA1391Z.

To generate a construct that would induce LjLAZY genes overexpression in plants, the full-length cDNA was cloned into a pCAMBIA1302 vector under the control of either the CaM35S promoter (p35S:LjLAZY) or the LjLAZY3 promoter (pLAZY3:LAZY3) using In-Fusion™ PCR Cloning Kits.

To study the subcellular localization of LjLAZY3, both the LjLAZY3 open reading frame and its truncated sequence were cloned into pSAT6-EYFP-N1. LjLAZY3ΔC53–enhanced yellow fluorescent protein (eYFP) represents the truncated version of LjLAZY3, with a deletion of amino acid residues 211–264.

Plant transformation

The construct LjPIN2-eGFP in pCAMBIA0390 was transformed into Agrobacterium rhizogenes MSU440, after which it was introduced into MG-20 and Ljlazy3 by hairy-root transformation. Other constructs, i.e. pDR5:GUS, pLjLAZY3:GUS, pLjLAZY3:LjLAZY3, and p35S:LjLAZY3, were transformed into Agrobacterium tumefaciens AGL1 and then introduced into MG-20 and Ljlazy3-1 by stable transformation. Plant transformation was performed as described by Lombari et al. (2005). Positive transgenic plants were selected through hygromycin or basta resistance screening, and independent lines of T₂ homozygous progeny were used for further experiments.

Histochemical analysis of GUS activity

Plants were submerged in GUS reaction buffer (1 mM X-Gluc, 0.1 M phosphate buffer pH 7.0, 1 mM EDTA pH 8.0, 0.5 mM K₃[Fe(CN)₆], 0.5 mM K₄[Fe(CN)₆], 20% methanol, 0.1% Triton X-100), infiltrated in a vacuum for 20–30 min, and then maintained at 37 °C for 3 h. After staining, seedlings were fixed in 70% ethanol for observation. Photographs were acquired using a Leica DVM6 digital microscope (Leica, Wetzlar, Germany).

Measurement of auxin transport in seedlings

Auxin transport in roots was measured using a procedure described by Mei et al. (2012), with certain modifications. Root tips 1 cm in length were excised from 5-day-old seedlings grown vertically in 1/2 B&D medium. To test shootward auxin transport, agar blocks containing 200 nM [³H]IAA (26 Ci mmol⁻¹, PerkinElmer, Waltham, MA, USA) were applied at the root tip. After an 8 h incubation in the dark, 30 5-mm root segments from root parts other than the root tip were collected and soaked in 3 ml of scintillation fluid for 18 h. The rate of auxin transport was determined by measuring radioactive decay through a liquid scintillation counter (1450 MicroBeta TriLux; PerkinElmer). To measure rootward auxin transport, agar blocks containing 200 nM [³H]IAA and 200 nM IAA with or without 200 nM NPA were placed 10 mm from the root tip and incubated for 10 h, after which 30 5-mm root sections containing the root tip were harvested to determine the rate of auxin transport as described above. The measurements were repeated five times for each assay.

Confocal microscopy

For subcellular localization of the LjLAZY3–YFP fusion proteins in Arabidopsis protoplasts, fluorescence images were analysed with a Leica TCS SP8 confocal microscope using a ×63, 1.40 NA oil objective. Laser excitation was at 488 nm for YFP/GFP and emission filters were 505–550 nm for YFP/GFP and 650–750 nm for chlorophyll.

To detect PIN2 distribution under gravistimulation, transgenic roots from hairy root transformed plants were collected and placed horizontally in vertically oriented culture plates. Fluorescence signals were observed before and after reorientation for 3 h using either a Leica TCS SP8 or a Zeiss LSM 710 (Zeiss, Oberkochen, Germany) confocal microscope.

Immunolocalization of LjLAZY3

An improved procedure for whole-mount immunolocalization was adopted. In brief, 1 cm root pieces were mounted in 2% formaldehyde in microtubule-stabilizing buffer (MTSB) supplemented with 0.1% Triton X-100. Vacuum infiltration of explants, hydrophilization, cell wall digestion and membrane permeabilization were performed as described previously (Pasternak et al., 2015). After 3 h of blocking in BSA, samples were incubated with a GFP Tag antibody (1:200 dilution) coupled to Alexa Fluor 488 (Thermo Fisher Scientific) for 1 h at room temperature, followed by three washes and then mounted in Slow Fade^TM Diamond anti-fade mountant for confocal microscopy.

Statistical analysis

Student’s t-test was performed to determine any statistically significant differences among values measured from WT, Ljlazy3, and transgenic samples in each experiment.

Results

A L. japonicus mutant displays negative gravitropic response of roots

In general, roots of the L. japonicus plant cannot be seen at the surface of the soil. However, a mutant with roots emerging above the soil surface (Fig. 1A) was identified from the progeny of L. japonicus (Miyakojima MG-20) irradiated with carbon-ion beams. Further investigation using vertically oriented culture plates showed that both the primary and the lateral roots of this mutant always grow upwards (Fig. 1B, C). In WT plants, the primary root always grew in the same direction as the gravity vector, while the primary root of mutant plants grew against the gravity vector 24 h after reorientation (Fig. 1D).

Fig. 1.

Phenotypes of Ljlazy3. (A) Representative images of WT and Ljlazy3-1, with the triangles depicting above-ground roots. (B, C) Representative images of WT and Ljlazy3-1 grown on vertically oriented culture plates for 6 d (B) and 3 weeks (C). (D) Three-day-old seedlings were reoriented so that the roots were horizontal at the beginning. Images were taken after 24 h of gravistimulation. The arrow indicates the direction of gravity.

As auxin and amyloplast sedimentation are intimately linked to gravitropism, we investigated starch accumulation in the root tips and the influence of exogenous auxin on root gravity in Ljlazy3-1. We found that Ljlazy3-1 did not show amyloplast sedimentation changes in the primary root tip (see Supplementary Fig. S1), and that Ljlazy3-1 roots still grew upwards, opposite to the gravity vector, when seedlings were treated with exogenous auxin (Supplementary Fig. S2).

Map-based cloning and structural analysis of LjLAZY3

A map-based cloning strategy was used to clone the LjLAZY3 gene. In each F₂ or BC₂ population, approximately one-quarter of the individual plants displayed reverse gravitropism in roots (28:80 in Ljlazy3-1×Gifu F₂ population and 28:86 in Ljlazy3-1×MG-20 BC₂ population), which suggests that the negative gravitropism observed in the roots of Ljlazy3-1 is controlled by a single recessive locus. Subsequently, we screened 4650 plants of the Ljlazy3-1×Gifu F₂ population and obtained 1453 mutant plants for gene mapping. The LjLAZY3 gene was mapped in the region between two SSR markers, TM1775 (1/1453) and TM2358a (1/1453), on chromosome 3 (Fig. 2A), which is a 210-kb stretch with 18 predicted open reading frames (ORFs) (data from http://www.kazusa.or.jp/lotus/index.html). Based on the comparison of MG-20 and Ljlazy3-1 sequences, only one gene, Lj3g3v2576600, was shown to differ, containing a 2-bp deletion of GA in the ORF region of mutant plants. PCR amplification using DNA as a template further confirmed a GA deletion in the fourth exon of Lj3g3v2576600 in the Ljlazy3-1 (Fig. 2A). This deletion results in the substitution of 24 new amino acid residues for 53 amino acid residues at the C-terminus of the encoded protein.

Fig. 2.

Cloning and functional confirmation of LjLAZY3. (A) Fine mapping and a schematic representation of LjLAZY3, with triangle indicating the 2 bp deletion of GA in Ljlazy3-1. (B–D) Phenotypes in transgenic lines of genetic complementation experiments. The germinated seeds were grown on vertically oriented plates for 3 d (B), after which the roots were horizontally shifted, with representative images (C) and statistical data of root angle (D) taken 24 h after gravistimulation. Fifty to ninety-two seedlings were used for each treatment. The arrow indicates the direction of gravity.

For genetic complementation, the predicted gene driven by its own promoter was introduced into Ljlazy3-1, and four independent transgenic lines were generated. Similar to WT plants, the roots from the transgenic plants grew downwards along the gravity vector when the plants were grown on vertically oriented culture plates for 3 d (Fig. 2B). Moreover, the roots in these transgenic plants also grew downwards after horizontal rotation for 1 d (Fig. 2C, D). The results indicate that the negative gravitropism phenotype observed in Ljlazy3-1 is due to the mutation in Lj3g3v2576600.

LjLAZY3 contains five exons that span an approximately 2-kb genomic region. The gene product has 264 predicted amino acids. LjLAZY3 is an ortholog of the LAZY gene, which belongs to the IGT gene family (see Supplementary Fig. S3). The alignment of predicted amino acids with those of its orthologs is shown in Supplementary Fig. S4; a part of the IV domain and the whole of the V domain are missing from the mutated protein. Lotus japonicus LORE1 (Retrotransposon1) insertion mutants Ljlazy3-2 and Ljlazy3-3 also displayed negative gravitropic response in roots (Supplementary Fig. S5), which demonstrated that the negative root gravitropism resulted from the mutation of LjLAZY3 rather than the new protein produced in the Ljlazy3-1.

Expression patterns of the LjLAZY3 gene

Consistent with its role in root gravitropic response, LjLAZY3 was mainly expressed in roots and was especially prevalent in the root tip as assessed by real-time quantitative RT-PCR (qRT-PCR) assay (see Supplementary Fig. S6). This result was confirmed by evaluating LjLAZY3 promoter activity with pLjLAZY3:GUS predominantly observed in root apices and root stele above the elongation zone (Fig. 3A, B).

Fig. 3.

LjLAZY3 expression in L. japonicus. (A) Histochemical staining of pLjLAZY3:GUS seedlings. Eighty seedlings were used for GUS staining, with 73 showing the phenotype in (A). (B) Root apices of pLjLAZY3:GUS seedlings. (C) LjLAZY3 expression in response to 50 µM IAA treatment. Error bars represent the SE of three biological replicates; the statistical significance of differences between time points was assessed with a heteroscedastic Student’s t-test (**P<0.01). (D) LjLAZY3 expression in the root tip of pLjLAZY3:GUS plants following 1.5 h of gravistimulation. The roots were shifted to a horizontal orientation and representative images were taken 0 h (n=83) and 1.5 h (n=91) after reorientation. Triangle indicates the accumulation of LjLAZY3 transcripts at the upper side of the root tip. (E) Transverse section of the region indicated by the rectangles in (D). Three-day-old seedlings were used for phenotypic analysis. The arrow indicates the direction of gravity.

The phytohormone auxin plays a crucial role in root gravitropism (Su et al., 2017); for this reason, we examined how auxin affects LjLAZY3 expression. The transcription of LjLAZY3 in the root decreased by over 50% within 2 h of IAA application (Fig. 3C), which suggests that LjLAZY3 is an early auxin-response gene that is negatively regulated by auxin. Due to the well-established relationship between auxin and gravitropism, we subsequently investigated whether gravity affects LjLAZY3 expression by analysing LjLAZY3 promoter activity in the root tip upon gravistimulation. In the horizontal roots of pLjLAZY3:GUS plants (rotated 90° from the original vertical axis), LjLAZY3 transcripts accumulated less on the lower side of the root tip than on the upper side following 1.5 h of gravistimulation (Fig. 3D, E).

LjLAZY3 has a polar localization in the stele cells of L. japonicus root

To determine the subcellular localization of the LjLAZY3 protein, we introduced a full-length LjLAZY3–eYFP fusion protein into Arabidopsis protoplasts. The results show that LjLAZY3–eYFP was mainly localized to the plasma membrane (Fig. 4A). We also produced a truncated version of the protein, LjLAZY3ΔC53–eYFP, to determine whether mutation of the LjLAZY3 gene affects LjLAZY3 localization. This truncated protein showed similar localization to LjLAZY3–eYFP, with a clear signal originating from the plasma membrane (see Supplementary Fig. S7A). Unlike DRO1 from rice, which requires a C-terminal sequence for plasma membrane localization (Uga et al., 2013), the C-terminal sequence of LjLAZY3 is important for its biological function rather than subcellular localization.

Fig. 4.

Subcellular localization of LjLAZY3. (A) LjLAZY3–eYFP images indicate the localization of LjLAZY3–eYFP in the plasma membrane of Arabidopsis protoplasts. (B) Immunolocalization of LjLAZY3-eGFP in pLjLAZY3:LjLAZY3-GFP transgenic L. japonicus roots. LjLAZY3 protein localizes predominantly to the apical or basal side of plasma membrane in the root stele cells. Arrowheads mark polarity of the LjLAZY3 localization. (C) The region indicated by the rectangle in (B).

We further investigated the in situ subcellular location of p35S:LjLAZY3-YFP in L. japonicus root. Ectopically expressed LjLAZY3–YFP in the epidermis localizes to the plasma membrane, whereas LjLAZY3–YFP in the stele cells localizes predominantly to the basal/apical sides of the plasma membrane (Supplementary Fig. S7B, C). Subsequently, LjLAZY3–eGFP driven by its native promoter was expressed in MG-20, and LjLAZY3–eGFP was observed to have polar localization at the plasma membrane in root stele cells of L. japonicus (Fig. 4B, C).

Ljlazy3 shows a deficiency in rootward PAT and is insensitive to the PAT inhibitor NPA

Since the activity of the DR5 reporter reflects not only endogenous auxin abundance, but also local auxin signaling capacities as well as rates of transcription and translation (Band et al., 2012), the pDR5:GUS cassette was used to study auxin distribution and transport in WT and Ljlazy3-1 roots. We treated seedlings expressing pDR5:GUS with exogenous auxin at the root–shoot junction and found that DR5 expression was much lower in Ljlazy3-1 than in WT roots following treatment with either IAA or 2,4-D (Fig. 5A). This finding suggests that PAT had decelerated in Ljlazy3-1 roots.

Fig. 5.

Ljlazy3-1 showed deficient rootward polar auxin transport. (A) pDR5:GUS expression in WT and Ljlazy3 after treatments with 100 µM IAA, 100 µM NAA or 100 µM 2,4-D for 16 h. The control panel represents no treatment. n=60–80 in each treatment and representative images are shown. (B, C) Rootward auxin transport was drastically reduced in 4- (B) and 7-day-old (C) Ljlazy3 roots. [³H]-labeled IAA together with the auxin transport inhibitor NPA were used as the negative control. Error bars represent the SE of five biological replicates; the statistical significance of differences between groups was assessed using Student’s t-test (**P<0.01, ***P<0.001).

The results from the measurement of shootward and rootward PAT in 4- and 7-day-old seedlings showed that the rootward PAT was drastically reduced in Ljlazy3-1 roots relative to WT roots (Fig. 5B, C), confirming that the mutation of LjLAZY3 results in decelerated auxin flow in roots. Furthermore, NPA treatment did not alter the rootward transport of ³H-IAA in Ljlazy3-1, but reduced rootward auxin flow in WT to the levels observed in 4- and 7-day-old Ljlazy3-1 roots (Fig. 5B, C).

LjLAZY3 deficiency results in reversed auxin redistribution under gravistimulation

It is well-established that auxin accumulation on the lower half of a horizontal root triggers the bending of root tips at the elongation zone (Su et al., 2017). The primary root of Ljlazy3-1 displayed significant negative (i.e. upward bending) gravitropism (Fig. 6A). To determine whether the negative gravitropic response observed in Ljlazy3-1 seedlings is accompanied by changes in asymmetric auxin distribution at the root tip, GUS activity under gravistimulation was analysed in WT and Ljlazy3-1 seedlings containing the pDR5:GUS cassette. As shown in Fig. 6B, after 4 h of gravistimulation, about 70% of the WT seedlings exhibited auxin redistribution to the lower side of the root tip, whereas auxin redistribution was reversed in Ljlazy3-1 root tips, i.e. concentrated to the upper side of the root. When the gravistimulation was extended to 6 h, an even higher proportion of the WT seedlings (86%) demonstrated auxin redistribution to the lower side as the root curved along the gravity signal. In contrast, about 56% of the Ljlazy3-1 seedlings exhibited significant pDR5:GUS expression on the upper side of the root, with minimal expression in the lower side of the root columella cells. These results demonstrate that the LjLAZY3 mutation results in the reversal of asymmetric auxin distribution in the primary root tip upon gravitropism. The reversal of asymmetric auxin distribution was further confirmed by the result that the accumulated LjPIN2 protein appeared to be present predominantly at the upper side of the Ljlazy3-1 root apex upon gravistimulation, which is the opposite to that in the WT roots (see Supplementary Fig. S8).

Fig. 6.

Reversal of the auxin gradient leads to negative gravitropism in Ljlazy3-1 primary roots. (A) Negative root gravitropism in Ljlazy3-1 primary roots. Three-day-old seedlings were shifted to vertically oriented culture plates with root tips in a horizontal position. Root tip angles were measured at different time points after gravistimulation. The presented values are the means ±SE from 50–77 seedlings. (B) pDR5:GUS expression pattern in the root tips. Roots of 3-day-old pDR5:GUS seedlings were shifted to a horizontal position, with roots collected for histochemical staining before the reorientation as well as after 4 or 6 h of horizontal orientation. This experiment was repeated four times (n=50–60 per repeat) and representative images are shown. Triangles indicate positions of auxin accumulation. The arrow indicates the direction of gravity. (C) The effects of NPA treatment on root gravitropism. Five-day-old seedlings were shifted to vertically oriented plates containing 0.1 µM of NPA, and the root tips were placed in a horizontal position. Root tip angles were measured at different time points after gravistimulation. The presented values are means ±SE from 42–55 seedlings.

Treatment with 0.1 µM NPA impaired gravitropism in the WT without noticeably affecting growth (see Supplementary Fig. S9). More importantly, treatment of Ljlazy3-1 with 0.1 µM NPA impaired the previously observed reversal of root gravitropism (Fig. 6C). Our data indicate that an auxin gradient formed in the root tips of both WT and Ljlazy3, but that the gradients in these two groups of seedlings formed on opposite sides of the root tip.

LjLAZY family genes have divergent molecular function in root gravitropism

To investigate what occurs in the roots of Ljlazy3-1 seedlings, we expressed the LjLAZY3 gene driven by the CaM35S promoter in Ljlazy3. Ectopic expression of LjLAZY3 was able to rescue the negative gravitropism phenotype of Ljlazy3-1 when seedlings were grown on vertically oriented plates (Fig. 7A). Moreover, the root tips of p35S: LjLAZY3 plants began to bend downward along the gravity vector after reorientation, similar to what was observed in WT (Fig. 7C). These results indicate that root gravitropism is dependent on the presence of LjLAZY3 but not the asymmetric distribution of LjLAZY3 in root tips.

Fig. 7.

LjLAZY1, LjLAZY3, and LjLAZY4 were unable to rescue the gravitropic phenotype of Ljlazy3 roots. (A) The negative gravitropism phenotype of Ljlazy3 was recovered by overexpression of LjLAZY3, driven by the CaMV 35S promoter. Seedlings were grown on vertically oriented plates for 3 d. n≥80 in each line, and representative images are shown above. (B) Expression patterns of the LjLAZY family genes in roots. GUS staining of roots expressing pLjLAZY1:GUS, pLjLAZY3:GUS, and pLjLAZY4:GUS. (C) Overexpression of LjLAZY family genes in Ljlazy3. One-day-old seedlings were rotated by 90° and representative images were taken 48 h after reorientation. The gravitropic phenotype of Ljlazy3 roots could be completely rescued by the expression of LjLAZY3, but not the expression of other LjLAZY genes. n=20–41 in each group.

By analysing LjLAZY promoter activity, we found the LjLAZY1 promoter was active mainly in the stele, while LjLAZY4 and LjLAZY5 were active in both the stele and the columella cells in the root (Fig. 7B). The promoter activity of LjLAZY2 and LjLAZY6 was not detected in the root. To investigate whether these LjLAZY genes have the same molecular function as LjLAZY3 in root gravitropism, we therefore ectopically expressed LjLAZY1, LjLAZY4, or LjLAZY5 in Ljlazy3-1. The expression of these LjLAZY–GFP proteins in Ljlazy3-1 was confirmed by fluorescence detection (see Supplementary Fig. S10). None of these LjLAZY genes was able to rescue the gravitropic phenotype of Ljlazy3-1 (Fig. 7C). These results suggested that the LjLAZY3 gene play a distinct role in root gravitropism in L. japonicus.

Discussion

In this study, we isolated a monogenic recessive mutant of the legume species L. japonicus that exhibits negative gravitropism in the primary root and lateral roots (Fig. 1). Deficiency of LjLAZY3, an ortholog of LAZY genes, reverses auxin flow in the root tip without influencing amyloplast sedimentation (see supplementary Fig. S1). This result is consistent with previous reports that LAZY proteins act downstream of amyloplast sedimentation but upstream of auxin gradient formation (Godbole et al., 1999; Li et al., 2007; Taniguchi et al., 2017; Yoshihara and Spalding 2017).

Polar localization of LjLAZY3 might be required for PAT in L. japonicus root gravitropism

Two distinct auxin transport streams exist in roots: auxin either moves rootward through the central cylinder or basipetally through the lateral root cap and epidermis to reach the elongation zone (Jones, 1998). In the primary root, LjLAZY3 is mainly expressed in root stele and apices (Fig. 3A, C), a pattern that correlates with PAT. Here, mutation of LjLAZY3 reduces both rootward PAT and sensitivity to PAT inhibitor NPA (Fig. 5), which indicates LjLAZY3 might facilitate rootward PAT through the central cylinder to regulate root gravitropism in L. japonicus. Similar to other LAZY members that regulate root gravitropism, such as AtLAZY2, AtLAZY3, and OsDRO1 (Yoshihara et al., 2013; Taniguchi et al., 2017), LjLAZY3 was observed only on the plasma membrane in Arabidopsis protoplasts. It is noteworthy that most of the LjLAZY3 proteins were observed to be polarly localized in the stele cells in L. japonicus root (Fig. 4B; Supplementary Fig. S7). This finding suggests that the polarization of LjLAZY3 distribution in the root cells might be required for PAT in the root of L. japonicus.

Positive root gravitropism is dependent on the expression of LjLAZY3 in the root tip but not its asymmetric expression

AtLAZY1, AtLAZY2, AtLAZY3, and AtLAZY4 redundantly contribute to positive gravitropism in the inflorescence stem, hypocotyl, and root in Arabidopsis (Guseman et al., 2017; Yoshihara and Spalding, 2017). In the present study, LjLAZY3 is pivotal to gravitropism in the root, which is consistent with its expression patterns in root tips (Figs 2, 3). Additionally, other LjLAZY genes expressed in the stele and/or tips of roots driven by the CaM35S promoter was unable to rescue the gravitropic phenotype of Ljlazy3-1 roots (Fig. 8). Our results suggested that LjLAZY family genes do not share redundant molecular functions in root gravitropism, and the LjLAZY3 gene plays a distinct role in L. japonicus positive root gravitropism.

Although LjLAZY3 functions upstream of auxin gradient formation, our results still showed that auxin can repress LjLAZY3 expression (Fig. 3C). This finding is in agreement with the expression pattern of DRO1 in rice roots in response to auxin (Uga et al., 2013). In horizontally oriented root tips, LjLAZY3 transcripts accumulated in the outer cells of the distal elongation zone on the upper side of the root rather than on the lower side (Fig. 3D, E), which is the complete opposite to the auxin distribution. Furthermore, constitutive expression of LjLAZY3 in Ljlazy3-1 caused bending of the roots in the direction of the gravity, as was the case in WT plants (Fig. 7A, B). This suggests that asymmetric distribution of LjLAZY3 transcription is irrelevant to cell elongation and the bending of the root tip. It is likely that auxin acts as a negative feedback regulator of LjLAZY3 expression. According to our data, we also propose that root gravitropism depends on the presence of LjLAZY3, but not on its asymmetric expression.

The negative gravitropic response in Ljlazy3 might stem from gravity signal reversal in the root tip

Upon gravistimulation, auxin accumulates on the lower side of the root tip, which results in decreased cell elongation of the lower side relative to the upper side, thus triggering bending of the root tip at the elongation zone. Both the polar localization of PIN3 within statocytes and formation of the lateral auxin gradient across the root tip upon gravistimulation can be used as cellular markers to study gravitropism defects (Su et al., 2017). For example, a pin3 mutant was shown to exhibit slight gravitropism alterations (Friml et al., 2002). Moreover, the mutation of Arabidopsis ARL2/ARG1 caused the relocalization of PIN3 to statocytes to fail, and subsequently reduced root gravitropism upon gravistimulation (Harrison and Masson, 2008). Taniguchi et al. (2017) suggested that reversed lateral root bending is the result of inverse asymmetric PIN3 expression in the root tip of Arabidopsis Atlzy1,2,3 triple mutant. Unfortunately, the LjPIN3-GFP fluorescence signal was not visible in either the WT or Ljlazy3-1 roots since L. japonicus roots have multiple cell layers.

Nevertheless, the decelerated rootward PAT observed in Ljlazy3-1 seedlings (Fig. 5) supports that auxin transport might have been disrupted. If PAT was blocked, then the asymmetric auxin redistribution at the root apex would not be formed. This would lead to deficient or delayed root gravitropism since the disruption of auxin transport results in loss of the directional gravitropic response. However, the asymmetric auxin flow and polar relocalization of LjPIN2 reversed in Ljlazy3-1 root tips after reorientation (Fig. 6B), suggesting that both auxin flow and early gravity signal transduction are reversed in Ljlazy3-1 roots. Therefore, our data show that LjLAZY3 is required for positive gravitropism in the root, as well as providing evidence for the existence of a factor, or set of factors, that can establish the auxin gradient, albeit in an inverse orientation, during gravistimulation (Yoshihara and Spalding, 2017).

Supplementary data

Supplementary data are available at JXB online.

Fig. S1. Amyloplast development in WT and Ljlazy3-1 mutant root tips.

Fig. S2. Auxin treatment could not rescue the negative gravitropism phenotype of Ljlazy3-1.

Fig. S3. Phylogenetic tree of IGT gene family in different species.

Fig. S4. Alignment of the deduced amino acid sequences of LjLAZY3 and its IGT gene family homologs.

Fig. S5. LORE1 insertions in the LjLAZY3 gene cause the negative root gravitropism in L. japonicus

Fig. S6. qRT-PCR analysis of LjLAZY3 expression in different tissues from the WT.

Fig. S7. Localization of the LjLAZY3 protein in Arabidopsis protoplasts and in L. japonicus roots.

Fig. S8. The localization of LjPIN2–eGFP fusion protein in transformed hairy roots.

Fig. S9. Seedlings were treated with different concentrations of NPA.

Fig. S10. The expression of LjLjLAZY–GFP in Ljlazy3-1 plants.

Table S1. Nomenclature of LAZY1 Family Genes in Lotus japonicus.

Table S2. Primer pairs used in this study.

Author contributions

HJ, GW, and YC designed the experiments and revised the manuscript. YC collected data, performed data analyses, wrote the manuscript, and conducted experiments. SX and GS generated the mapping population and then conducted gene mapping and phenotypic analyses. SS provided sequence information of SSR markers. SX measured auxin transport. LT generated the plasmid constructs and then conducted the root hair and stable transformations. LL conducted histochemical analyses of GUS activity. XX was responsible for microscopic observation. PW conducted gene expression analyses. The authors declare no conflict of interest.