Partial deficiency of isoleucine impairs root development and alters transcript levels of the genes involved in branched-chain amino acid and glucosinolate metabolism in Arabidopsis.

Isoleucine is one of the branched-chain amino acids (BCAAs) that are essential substrates for protein synthesis in all organisms. Although the metabolic pathway for isoleucine has been well characterized in higher plants, it is not known whether it plays a specific role in plant development. In this study, an Arabidopsis mutant, lib (low isoleucine biosynthesis), that has defects in both cell proliferation and cell expansion processes during root development, was characterized. The lib mutant carries a T-DNA insertion in the last exon of the OMR1 gene that encodes a threonine deaminase/dehydratase (TD). TD catalyses the deamination and dehydration of threonine, which is the first and also the committed step in the biosynthesis of isoleucine. This T-DNA insertion results in a partial deficiency of isoleucine in lib root tissues but it does not affect its total protein content. Application of exogenous isoleucine or introduction of a wild-type OMR1 gene into the lib mutant can completely rescue the mutant phenotypes. These results reveal an important role for isoleucine in plant development. In addition, microarray analysis indicated that the partial deficiency of isoleucine in the lib mutant triggers a decrease in transcript levels of the genes encoding the major enzymes involved in the BCAA degradation pathway; the analysis also indicated that many genes involved in the biosynthesis of methionine-derived glucosinolates are up-regulated.

Introduction

Characterized by their small, branched, hydrocarbon residues, leucine, isoleucine, and valine together form a unique group of so-called branched-chain amino acids (BCAAs). Humans and animals cannot synthesize these three amino acids and must obtain them from other sources, such as plants. In plants, some enzymes involved in BCAA biosynthesis are the targets of several herbicides. Because of their importance in human and animal nutrition and agriculture, the biosynthesis pathways of BCAAs in higher plants have been extensively studied and well characterized (Singh and Shaner, 1995; Binder, 2010).

In plants, the conversion of threonine to 2-oxobutanoate is the first and also the committed step towards isoleucine biosynthesis. This reaction is catalysed by threonine deaminase/dehydratase (TD) (Fig. 7A). After four more enzymatic reactions, 2-oxobutanoate is condensed with one molecule of pyruvate and subsequently transformed into isoleucine. A unique feature of BCAA biosynthesis in plants is that the four enzymes that catalyse the reactions for production of isoleucine are also used for biosynthesis of valine starting from the condensation of two molecules of pyruvate. These four enzymes, acetohydroxyacid synthase (AHAS), ketol acid reductoisomerase (KARI), dihydroxyacid dehydratase (DHAD), and branched-chain aminotransferase (BCAT), act in parallel on different substrates and finally lead to the production of isoleucine and valine. The last intermediate of valine biosynthesis, 3-methyl-2-oxobutanoate, is also the initial substrate in the branch leading to leucine biosynthesis. Through four enzymatic reactions, this substrate is converted to leucine, with the first step catalysed by isopropylmalate synthase (IPMS) (Fig. 7A). For all three BCAAs, the enzymes involved in the last step of production are BCATs. BCATs also catalyse the first step of BCAA degradation (Fig. 7B). Interestingly, some enzymes involved in leucine biosynthesis are also used for synthesis of glucosinolates through the methionine chain elongation pathway (Sønderby ).

Fig. 7.

The BCAA metabolic pathway. (A) The BCAA biosynthesis pathway. The substrates and enzymes catalysing each step are shown. Feedback inhibition by the end-product is indicated by a red line. The antagonistic effect of valine on the inhibition of TD by isoleucine is indicated by a green dotted line. The enzymes whose transcript level is increased in lib are shown in red. The numbers along the AGI gene code indicate the ratio of expression of that gene in lib versus the WT. (B) The BCAA degradation pathway. The substrates and enzymes catalysing each step are shown. The enzymes whose transcript level is decreased in lib are shown in blue. The numbers along the AGI gene code indicate the ratio of expression of that gene in lib versus the WT. In (A) and (B), the names of enzymes are boxed.

In plants, BCAA homeostasis is strictly regulated. So far, our understanding of such regulation comes mostly from the studies of allosteric control of three enzymes, namely TD (Halgand ), AHAS (McCourt and Duggleby, 2006), and IPMS (de Kraker ). It is well known that the activity of TD is feedback inhibited by its end-product, isoleucine. An Arabidopsis mutant with a feedback-insensitive TD overaccumulates isoleucine (Mourad and King, 1995). The feedback inhibition of TD by isoleucine is antagonized by valine. Similarly, AHAS is inhibited by each of the BCAAs, and IPMS is feedback regulated by leucine. It remains unclear, however, whether other regulatory mechanisms control BCAA homeostasis. For example, it is not known whether a deficiency of a particular BCAA will affect transcript levels of the genes involved in the BCAA metabolic pathways.

In humans and other animals, some amino acids serve not only as building blocks for protein synthesis but also as signals that control gene transcription, translation of mRNAs, and metabolic activities (Meijer and Dubbelhuis, 2004; Kimball and Jefferson, 2006). For example, leucine plays an important role in the regulation of autophagy, which is associated with many physiological processes, such as development, cell death, and tumour suppression (Levine and Klionsky, 2004). Whether leucine has a similar function in plants is unclear. Previous studies have shown that homeostasis of some amino acids may also have regulatory functions in sustaining plant root growth and development. Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) catalyses the conversion of glyceraldehyde-3-phosphate to 1,3-bisphosphoglycerate. The plastidial GAPDH (GAPCps) has two isoforms in Arabidopsis (GAPCp1 and GAPCp2), which are important for the synthesis of serine in roots (Muñoz-Bertomeu ). The gapcp double mutant, which contains reduced levels of serine in roots and aerial parts, displays a drastic phenotype of arrested root development. All these mutant phenotypes can be rescued by exogenously applied serine. The homeostasis of another amino acid, histidine, also affects root development. The Arabidopsis mutant hpa1, which has a 30% reduction in free histidine content, has a very short root system and exhibits a specific defect in root meristem maintenance (Mo ). Similarly, its mutant phenotype can be fully rescued by exogenously applied histidine or by introduction of a wild-type (WT) functional HPA1 gene. The effect of perturbed histidine homeostasis on plant development was also observed for the Arabidopsis mutant apg10 (Noutoshi ). In addition, exogenously applied glutamate can act as a signal to modulate primary root growth and lateral root formation, probably through the interaction with the auxin signalling pathway (Walch-Liu ). Whether the three BCAAs play any specific roles in regulating plant development is currently unknown.

In this work, an Arabidopsis mutant called lib (low isoleucine biosynthesis) that showed severe defects in root development was identified. Genetic and molecular analysis indicated that the mutant phenotype is caused by a T-DNA insertion within the OMR1 gene, which encodes a threonine deaminase/dehydratase (TD). As a consequence, the mutant produces less isoleucine than the WT. Furthermore, it was found that the partial deficiency of isoleucine in lib alters transcript levels of the genes encoding the enzymes involved in BCAA and glucosinolate metabolic pathways.

Materials and methods

Plant materials and growth conditions

Arabidopsis seeds were surface sterilized with 20% (v/v) commercial bleach for 20min, followed by four washes in sterile distilled water. The sterilized seeds were then sown on Petri dishes containing a medium of half-strength Murashige and Skoog (MS) basal salts with 1% sucrose, 0.1% MES (adjusted to pH 5.7 with 1 N NaOH), and 1.2% agar. After 2 d of stratification at 4 °C, the agar plates were placed vertically in the growth room with a photoperiod of 16h light:8h dark at 22–24 °C. The light intensity was 100 µmol m–2 s–1.

Mutant isolation

The seeds of an Arabidopsis T-DNA activation tagging library (Qin ) were directly sown on Petri dishes containing half-strength MS medium as described above. At 8 days after seed germination (DAG), the seedlings with short primary roots were selected and transferred to soil. The progeny from self-pollinated plants were retested for their root growth phenotype. The confirmed root mutant was backcrossed to the WT three times before further characterization. The genetic linkage between the mutant phenotype and T-DNA insertion was tested by growing the segregated F2 progeny from the backcross on half-strength MS medium supplemented with glufosinate ammonium (Basta) (20 µg ml–1).

Histochemical staining for GUS activity

The histochemical analysis of β-glucuronidase (GUS) activity was performed as described (Jefferson ).

Microscopy

For light microscopy, the excised roots or whole seedlings collected at different stages after germination were mounted in HCG clearing solution (30ml of H2O, 80g of chloral hydrate, 10ml of 100% glycerol) and examined with a differential interference contrast (DIC) microscope (Olympus BX51, Japan) or stereomicroscope (Olympus SZ61, Japan).

Eight-day-old WT and lib seedlings carrying the markers SCR–GFP (green fluorescent protein), J0571–GFP, CS9227–GFP, CS9093–GFP, WOX5–GFP, and PLT1–CFP (cyan fluorescent protein) were used for confocal microscopic analysis. The roots were excised from the seedlings and stained with 10 µM propidum iodide (PI) for 5 s, and rinsed with distilled water. The specimens were examined with a confocal laser scanning microscope (Zeiss 710META, Germany). Excitation wavelengths of 546, 488, and 433nm were used to visualize the signals of PI staining, GFP, and CFP, respectively. The emission wavelengths were 573, 507, and 476nm.

Determination of root meristem size and root cell length

To determine the root meristem size, the root tips were excised from seedlings at different developmental stages and examined with a DIC microscope. For the measurement of the lengths of epidermal cells, the roots excised from 8-day-old seedlings were stained with PI for 5 s and observed with a DIC florescence microscope (Olympus BAX51, Japan).

Cloning of T-DNA flanking sequence

The T-DNA flanking sequence in the lib mutant was cloned by thermal asymmetric interlaced PCR (TAIL-PCR) (Liu ). The authenticity of the cloned sequence was confirmed by PCR using two pairs of primers. The first pair comprised a primer within the T-DNA left border (5’-TTGACCATCATACTCATTGCTG-3’) and a primer located at the sixth exon of the OMR1 gene (5’-GGACCAATGAACATAAGCGA-3’); the second pair comprised two primers surrounding the T-DNA insertion site with the sequence 5’-GGACCAATGAACATAAGCGA-3’ and 5’-GAAACC AAAAGGTAACCCAC-3’.

Vector construction and plant transformation

The detailed protocols for vector construction and plant transformation are described in Supplementary Text S1 available at JXB online.

Enzymatic activity assays for plant threonine deaminase/dehydratase

The detailed protocol for the plant TD enzymatic activity assay is described in Supplementary Text S1 at JXB online.

Quantitative real-time PCR analysis

Total RNAs were extracted from root and shoot tissues with the Tiangen RNAeasy kit (Tiangen, Beijing). DNase-treated RNA (1 µg) was reverse transcribed in a 20 µl reaction using M-MLV reverse transcriptase according to the manufacturer’s manual (TaKaRa, Kyoto, Japan). cDNA was amplified using SYBR Premix Ex Taq (TaKaRa, Japan) with the Bio-Rad CFX96 real-time PCR detection system. The amounts of PCR products in each sample were normalized using the ACTIN gene as the internal control. The primers used for OMR1 amplification were: P1 (5’-GAAAGATCACCTGCGTTACTTGA-3’), P2 (5’-GGCATAGAACCTCGTCTCCA-3’), P3 (5’-GTTTGAAGC TGTGGTGGATAATC-3’), and P4 (5’-CTAAAACGCTCATCC GACAG-3’).

The sequences of the primers used for amplification of other three genes were as follows: At5g23010, 5’-GAAATTGAACGCTG TCTTCTCA-3’ and 5’-CGCTAGATGTTACTAATGCCTTCA-3’; At5g23020, 5’-CCGTGAACAGTGTTAAGTACGC-3’ and 5’-CAC CCGCTTTTATCGATTCT-3’; and At3g19710, 5’-CGCAACCTCT TCCTGTGAGT-3’ and 5’-TCCGTACGAACAAACTTGAATG-3’.

Measurement of free amino acid content

Root tissue (100mg) of 8-day-old Arabidopsis seedlings was ground into a fine powder in liquid nitrogen. A 300 µl volume of 80% ethanol was added to the powder, mixed well, and incubated for 30min at 45 °C. The samples were then centrifuged at 12,000rpm for 5min to remove cell debris. The supernants were passed through a 0.22 µm filter into new sterile tubes. The chlorophyll was removed from the supernatants by extraction with cyclohexane. The supernatants were concentrated to 80 µl through vacuum evaporation and were analysed for free amino acid contents with a 1100 HPLC (HP, USA) amino acid analyser.

Total protein determination

The contents of total proteins of shoots and roots were determined according to Jing .

Microarray analysis

The microarray experiments were performed at ShanghaiBio Corporation. Total RNA was isolated using the RNase kit (Qiagen) from three replicate samples of the root tissues of 8-day-old WT and lib seedlings grown on half-strength MS medium. Total RNA was amplified, labelled, and purified with the GeneChip 3’IVT Express Kit (Cat#901229, Afftnetrix, Santa Clara, CA, USA) following the manufacturer’s instructions to obtain biotin labelled cRNA. Array hybridization and wash were performed using the GeneChip® Hybridization, Wash and Stain Kit (Cat#900720, Affymetrix) in a Hybridization Oven 645 (Cat#00-0331-220V, Affymetrix) and Fluidics Station 450 (Cat#00-0079, Affymetrix) following the manufacturer’s instructions. Hybridized slides were scanned by GeneChip©Scanner 3000 (Cat00-00212, Affimetrix) and Command Console Software 3.1 (Affymetrix) with default settings. Raw data were normalized by the MAS 5.0 algorithm, Gene Spring Software 11.0 (Agilent Technologies, Santa Clara, CA, USA). Microarray data analysis was performed using the SBC Analysis system, which is available on the website: http://sas.ebioservice.com/index.jsp. The username and password to access this website are available upon request. A general description of the SBC analysis system can be found on the website: http://www.ebioservice.com/eng/index.asp.

The microarray data generated in this study have been deposited in the Gene Expression Omnibus (GEO) database under the accession number GSE42060.

Statistical analysis

The two-sample t-test function of Origin software (OriginLab Corporation, Northampton, USA) was used for statistical analysis in this work.

Results

Identification of the lib mutant

A mutant, designated lib (low isoleucine biosynthesis), was identified from a T-DNA activation tagging library (Qin ). When grown on half-strength MS medium, lib had an abnormally short primary root (Fig. 1A). Time-course studies indicated that the short-root phenotype of lib was due to its low growth rate rather than to growth arrest at a certain developmental stage (Fig. 1B). Lateral roots became visible 5 DAG for lib but 8 DAG for the WT. Lateral root number, however, increased more rapidly for the WT than for the lib mutant; at 13 DAG, the WT had more lateral roots than the lib mutant (Supplementary Fig. S1 at JXB online). When seedlings were grown on half-strength MS medium for 21 d, shoots were significantly smaller for lib than for the WT (Fig. 1A). When grown in soil, however, no morphological differences were observed between lib and the WT (data not shown). In this research, a detailed analysis of the defects in root development of the lib mutant was conducted.

Fig. 1.

Characteristics of growth of the WT and lib mutant. (A) Appearance of the WT and lib seedlings grown on half-strength MS medium at 7 (left upper panel), and 14 (right panel) days after germination (DAG). Left lower panel: appearance of shoots of WT and lib plants grown on half-strength MS medium for 21 d. (B) Changes in primary root length of the WT and lib mutant from 2 to 10 DAG. (C) Number of root meristem cells of the WT and lib mutant, measured over time from 3 to 10 DAG. In B and C, values are means ±SE (n >20). (D) Root meristematic zone of 8-day-old WT and lib mutant seedlings. Black arrows indicate the boundary between the root meristematic and elongation zones. (E) Expression of CycB1::GUS in the meristematic zones of 8-day-old WT and lib mutant seedlings.

The lib mutation impairs both cell division and expansion

Root growth is a coordinated process of cell division and expansion. To investigate which process was affected in the lib mutant, its cell division activity was first examined. Because cell division activity in roots is reflected by the size of the root meristem, meristem sizes in WT and lib roots were compared by counting the number of cortex cells in a file extending from the quiescent centre (QC) to the first elongated cell. In the first 10 DAG, the meristem was 40–60% smaller in lib roots than in WT roots (Fig. 1C, D). To determine the cause of the low cell division activity in lib, the marker gene CycB1::GUS was introduced into lib through a genetic cross. Expression of the CycB1::GUS marker gene specifically occurs at the transition stage from the G2 to M phase (Colón-Carmona ). The number of meristem cells expressing CycB1::GUS was greater in lib than in the WT (Fig. 1E). These results indicated that low cell division activity in lib was due to the slower transition of cells from the G2 to M phase.

To determine whether the lib mutation also affects cell expansion, the size of root epidermal cells in lib and the WT was compared. In the root maturation zone of the WT, the epidermal cells were long and rectangular, and their lengths were relatively consistent (Fig. 2A). In the lib mutant, however, regions with short rectangular cells [referred to as long root cells (LRCs)] and regions with nearly square cells [referred to as short root cells (SRCs)] alternated (Fig. 2A, B). In an 8-day-old lib seedling, two regions with LRCs and two regions with SRCs were evident (Fig. 2B). The LRCs and SRCs in the lib mutant were 40% and 80% shorter, respectively, than the epidermal cells in the WT (Fig. 2D). Root hair cells were uniformly distributed along the primary root of the WT, but were unequally distributed in lib (Fig. 2C). In ~80% of 8-day-old lib seedlings, there were two segments on the primary root, which had few or no root hairs (Supplementary Fig. S2 at JXB online). These segments usually appeared at the junction between LRCs and SRCs. The length of the segments that had few or no root hairs varied among the lib roots examined. The length of the root hairs in the maturation zone was also compared. Overall, lib root hairs were ~70% shorter than WT root hairs (Fig. 2E). These results indicated that both cell proliferation and cell expansion were impaired in the lib mutant.

Fig. 2.

Length of root epidermal cells and patterns of root hairs of the WT and lib mutant. (A) Lengths of root epidermal cells in the maturation zone of the WT and the lib mutant. The roots were stained with propidium iodide (PI) and photographed with a confocal microscope. In the mutant, the regions with long (LRCs) and short (SRCs) root cells are indicated by white arrows. (B) The relative positions of the regions with LRCs and SRCs (boxed) along the primary root of an 8-day-old lib seedling; LRCs and SRCs are indicated by black arrows. (C) The patterns of root hairs of 8-day-old WT and lib mutant seedlings. The black arrows indicate regions with continuous formation of root hairs, and the red arrow indicates a region lacking root hairs. (D) Root cell length in maturation zones of 8-day-old WT and lib mutant seedlings. lib LRC, the long root cells of the lib mutant; lib SRC, the short root cells of the lib mutant. Values are the means ±SE of 30 seedlings (10 root epidermal cells were measured per seedling). (E) Length of root hairs of 8-day-old WT and lib mutant seedlings. Values are the means ±SE of 30 seedlings (20 root hairs were measured per seedling). Asterisks indicate a significant difference according to a two-sample t-test (P < 0.05).

The lib mutant has normal cell patterning

To examine whether the cell patterning in the roots of the lib mutant was altered, the mutant was crossed to transgenic lines that carried various cell type-specific markers with GFP as a reporter gene. These marker lines included SCR–GFP (endodermis specific), J0571 (cortex and endodermis specific), CS9093 (epidermis specific), and CS9227 (columella specific) (http://www.plantsci.cam.ac.uk/Haseloff). Expression patterns of the marker genes did not differ significantly between the lib and WT plants, indicating that the lib mutation did not change the cell patterning in roots (Supplementary Fig. S3 at JXB online). To investigate whether the lib mutation affects the activity of the QC, which is responsible for maintaining the functions of root stem cells, the expression patterns of three marker genes were examined. QC46–GUS and WOX5–GFP are specifically expressed in QC cells, and PLT1–CFP is expressed only in the stem cell niche. Again, the expression of these three marker genes did not differ significantly between lib and the WT, suggesting that the lib mutation did not interfere with the functions of the QC and stem cell activity (Supplementary Fig. S3).

The lib phenotype is caused by the insertion of a T-DNA into the OMR1 gene

To find the molecular lesion responsible for the mutant phenotype, the lib mutant was backcrossed to the WT with the same ecotype background. The phenotype of all the F1 plants was similar to that of the WT. Among the F2 progeny, the ratio of plants with the mutant phenotype to the WT phenotype was ~1:3 (84:210), indicating that the mutant phenotype was caused by a monogenic recessive mutation. Further analysis showed that, in F2 progeny, all the lib mutant plants were Basta resistant, whereas WT plants showed both Basta-resistant and Basta-sensitive phenotypes. These results suggested that the lib mutant phenotype is linked to the T-DNA insertion. Using the TAIL-PCR technique, it was found that the T-DNA was inserted into the last exon of the OMR1 gene (At3g10050) (Fig. 3A). The OMR1 gene encodes a TD that catalyses the deamination and dehydration of threonine, which is the first and also the committed step in the biosynthesis of isoleucine (Binder, 2010). The OMR1 gene was first identified in an Arabidopsis mutant (omr1-1) that was resistant to the inhibition of the isoleucine structural analogue l-O-methylthreonine (Mourad and King, 1995). The omr1-1 mutant produces an active but isoleucine feedback-insensitive form of TD because of two point mutations in the OMR1 protein. Plants with this mutation produce 20 times more isoleucine than the WT (Mourad ). Two pairs of primers were used in real-time PCR analysis to examine the expression level of the OMR1 gene in the WT and lib mutant (Fig. 3A). With the primers P1 and P2, which are located within the seventh and eighth exons (before the T-DNA insertion), a similar amount of OMR1 transcripts between these two primers was detected in both leaves and roots of the WT and lib (Fig. 3B). When primers P3 and P4, which amplify a region immediately after the T-DNA insertion, were used for real-time PCR analysis, no OMR1 transcripts were detected in either shoots or roots of the lib mutant. The insertion of the T-DNA sequence also introduced two consecutive stop codons immediately at the insertion site, which caused a truncation of 13 amino acids at the C-terminus of the OMR1 protein. Because the expression level of the part of OMR1 mRNA before the T-DNA insertion site in lib is similar to that of the OMR1 mRNA in the WT (Fig. 3B), the mutant phenotype is probably caused by the truncation of the OMR1 proteins rather than by the blocked expression of the part of OMR1 mRNA after the T-DNA insertion site.

Fig. 3.

T-DNA insertion in the OMR1 gene and its effect on OMR1 expression in the lib mutant. (A) Diagram showing the relative position of the T-DNA insertion (black triangle) in the OMR1 gene of the lib mutant. The black box, grey box, and the horizontal line between the boxes represent the exons, untranslated regions, and introns, respectively. The positions of start (ATG) and stop (TGA) codons and two pairs of primers (P1, P2 and P3, P4) for real-time PCR analysis are shown. (B) The relative expression level of the OMR1 gene in leaves and roots as determined by real-time PCR. ND, not detected. (C) Genetic complementation of the lib mutant. The photograph shows 10-day-old seedlings of the WT, lib mutant, and plant lines transformed with constructs of 35S::OMR1 (gDNA), 35S::OMR1 (cDNA), and OMR1::OMR1 (gDNA). The numbers in parentheses indicate the number of independent transgenic lines with rescued phenotypes over the total number of transgenic lines obtained for each construct.

To confirm that the T-DNA insertion in OMR1 was responsible for the lib mutant phenotype, the genomic and cDNA sequences of the WT OMR1 gene were introduced back into lib mutant plants under the Cauliflower mosaic virus (CaMV) 35S promoter or OMR1’s own promoter. The restoration of OMR1 gene expression to the level in the WT or overexpression of the OMR1 gene was confirmed in roots of three independent transgenic lines for each construct (Supplementary Fig. S4 at JXB online). Analysis of the transgenic plants indicated that all three gene constructs, namely 35S::OMR1 [genomic DNA (gDNA)], 35S::OMR1 (cDNA), and OMR1::OMR1 (gDNA), could restore the mutant phenotype of primary root growth (Fig. 3C). The other root growth parameters of an OMR1::OMR1 (gDNA) line were then compared with those of the WT in terms of the size of the root meristem (Fig. 4A, B), length of epidermal cells (Fig. 4C), and length of root hairs (Fig. 4D). The defects in all root growth parameters examined for lib were rescued in the complemented line. These results demonstrated that the T-DNA insertion into the OMR1 gene was indeed the cause of the lib mutant phenotypes.

Fig. 4.

Characteristics of root growth of 8-day-old seedlings of the WT, lib, the transgenic line of OMR1::OMR1 (gDNA) (C), and lib grown on half-strength MS medium supplemented with 5 µM isoleucine [lib+(Ile)]. (A) Photograph of the root meristematic regions of the four genotypes. (B) The number of root meristem cells of the four genotypes. Values are means ±SE (n >20). (C) Length of root epidermal cells of the four genotypes. Values are the means ±SE of 30 seedlings (10 root epidermal cells were measured per seedling). (D) Length of root hairs of the four genotypes. Values are the means ±SE of 30 seedlings (20 root hairs were measured per seedling). In (B), (C), and (D), asterisks indicate a significant difference from the WT according to a two-sample t-test (P < 0.05). (This figure is available in colour at JXB online.)

Isoleucine biosynthesis is reduced in the lib mutant

Because the lib mutant expresses a truncated version of the OMR1 gene, investigations were carried out to determine whether this would affect the total enzyme activity of TD. Total soluble proteins were extracted from 24-day-old WT and lib seedlings, and their TD activities were compared. The TD activity was ~40% lower in the lib mutant than in the WT (Fig. 5A). The profiles of free amino acids in the WT, the lib mutant, and an OMR1::OMR1 (gDNA) complemented line with restored mutant phenotype were then examined. As shown in Table 1, the amounts of free isoleucine in roots were ~50% lower in the lib mutant than in the WT. In the complemented line, the level of isoleucine was restored to that of the WT.

Fig. 5.

Rescue of the lib mutant phenotype by exogenously applied isoleucine. (A) TD enzyme activity of the WT and lib mutant. Values are the means ±SE of four biological replicates. The asterisk indicates a significant difference according to a two-sample t-test (P < 0.05). (B) Morphology of 8-day-old seedlings of the WT and lib mutant grown on half-strength MS medium supplemented with different amounts of isoleucine (concentrations of isoleucine added to the medium are indicated at the bottom). (C) Effect of exogenous isoleucine when applied to WT and lib seedlings from 7 to 11 DAG. The seeds of the WT and lib were directly sown on half-strength MS medium. After 7 d, the seedlings were transferred to half-strength MS medium supplemented with isoleucine and were photographed after another 4 d of growth. The black lines indicate the positions of the root tips when the seedlings were transferred to isoleucine-containing half-strength MS medium. (D) Total protein contents in leaf and root tissues of the WT and lib mutant.

Table 1.

Amount of free amino acids (pmol mg FW–1) in the WT, lib, and the OMR1::OMR1 transgenic line.

Amino acid	Absolute amount (pmol mg FW–1)
	WT	lib	OMR1::OMR1
Aspartate	117±20	125±31	104±4
Glutamate	231±10	252±56	174±7a
Glutamine	201±14	261±64	168±6a
Serine	218±15	265±84	174±7a
Asparagine	2980±131	3160±318	2130±89a
Histidine	8.38±1.09	5.03±0.28	6.72±2.51
Glycine	45.6±7.6	59.4±2.1	114±26b
Threonine	31.6±7.3	36.7±2.3	26.0±4.5
Arginine	26.1±10.7	43.2±17.8	30.1±9.7
Alanine	46.8±7.7	43.5±16.7	47.3±1.5
Tyrosine	1.93±0.30	4.22±2.71	2.33±0.46
Cysteine	2.75±0.17	3.00±0.02	2.74±0.07
Valine	128±8	125±23	90.4±6.6a
Phenylalanine	8.03±1.17	4.93±1.26a	8.47±4.21
Isoleucine	41.0±4.5	20.7±6.0a	34.6±4.0
Leucine	18.5±3.7	24.3±7.9	43.6±12.0b
Lysine	1.15±0.53	4.89±1.80b	6.62±0.97b
Proline	25.6±3.5	33.3±8.2	35.0±8.9

Values are the means ±SD of four replicates.

Values significantly lower than the WT (P < 0.05).

Values significantly higher than the WT (P < 0.05).

Next, the effect of exogenously applied isoleucine on root growth was tested. WT and lib mutant seeds were directly sown on half-strength MS medium supplemented with different amounts of isoleucine and grown for 8 d. The addition of isoleucine at ≥5 µM completely rescued the mutant phenotype (Fig. 5B). The degree of rescue of the mutant phenotype by addition of isoleucine was concentration dependent (Fig. 5B). Addition of phenylalanine, whose level was also decreased in the lib mutant, to the culture medium failed to rescue the mutant phenotype (data not shown). This indicated that the rescue effect was isoleucine specific. To confirm the effect of isoleucine further, mutant seedlings were first grown on half-strength MS medium for 7 d and then transferred to the same medium supplemented with 10 µM isoleucine. In the following 4 d, the growth rate of the primary root (Fig. 5C) and other characteristics of root growth of the lib mutant (Fig. 4), including the size of root meristem, length of epidermal cells, and length of root hairs, became similar to those of the WT.

Because isoleucine is an essential structural component of proteins, whether the partial deficiency in isoleucine in the lib mutant would affect total protein content in the lib mutant was deternined. The analysis of both leaves and roots showed that there was no obvious difference in total protein content between the WT and the lib mutant (Fig. 5D).

OMR1 is expressed in all plant organs and is present in plastids

To determine the expression patterns of the OMR1 gene, a 2.0kb DNA sequence upstream of its transcription start site was fused to a GUS reporter gene and transformed into WT plants. Forty-five independent transgenic lines were obtained, and the OMR1::GUS expression pattern of a representative line is shown in Fig. 6A. In the transgenic plants, the GUS gene was expressed in all plant organs. This is consistent with previous reports that isoleucine is synthesized in all parts of the plant (Singh and Shaner, 1995). In roots, GUS expression was mainly observed in vascular tissues, and no expression was detected in the meristematic region. Similarly, in cotyledons and mature leaves, GUS expression was strong in vascular tissues but weak in interveinal regions. In stems, GUS expression was strongest in young, growing parts. GUS expression, however, was stronger in old leaves than in young leaves. In addition, GUS expression was observed in young pedicels and pollen grains, but not in sepals, petals, or mature siliques.

Fig. 6.

The expression pattern and subcellular localization of OMR1. (A) OMR1::GUS expression patterns in different organs. Top row (left to right): an 8-day-old seedling; magnified views of cotyledons; root maturation zone; root tip. Middle row: aerial part of a 30-day-old plant; a leaf blade; the young part of a stem. Bottom row: an inflorescence; a fully opened flower; immature and mature siliques. (B) Subcellular localization of OMR1–GFP fusion protein. Upper row: localization of OMR1–GFP on a chloroplast in an isolated protoplast. Left to right: bright field micrograph of a protoplast; green fluorescent signal from OMR1–GFP; red autoflorescence from a chloroplast; merged micrograph of the two fluorescence images. Bottom row: co-localization of OMR1–GFP protein and pt-mCherry in the root tissues. Left to right: bright field micrograph of root tissue; green fluorescent signal from OMR1–GFP; red fluorescent signal from the plastid marker pt-mCherry; merged micrograph of the three images.

It has long been known that isoleucine is synthesized in chloroplasts but not in other cellular components (Binder, 2010). The OMR1 protein sequence is predicted to contain a transit peptide of 91 amino acids at its N-terminus for chloroplast targeting (Chlorop 1.1 Prediction Serves at http://www.cbs.dtu.dk/services/ChloroP/). For confirmation of its subcellular location, the coding sequence of OMR1 was fused to the N-terminus of GFP and transformed into lib mutants under the control of the CaMV 35S promoter. The lib mutant phenotypes were complemented by the transformation, indicating that the 35S::OMR1–GFP fusion protein was functional (data not shown). The protoplasts from mature leaves of the transgenic plants were isolated and subjected to confocal image analysis. As shown in Fig. 6B, green fluorescence signals were associated with the chloroplast. This result provided direct evidence that Arabidopsis TD protein is localized in leaf chloroplasts. To examine the subcellular localization of OMR1 in root tissues, the 35::OMR1-GFP transgenic plant was crossed to a plant line carrying a plastid (pt)-specific marker gene, pt-mCherry (Nelson ). As demonstrated by the co-localization of signals of pt-mCherry and OMR1–GFP (Fig. 6B), the OMR1 protein is localized in plastids in root tissues.

Partial isoleucine deficiency affects gene expression involved in BCAA and glucosinolate metabolism

To determine how the partial deficiency of isoleucine in the lib mutant will affect the gene expression at the genomic level, a microarray analysis was performed comparing the gene expression profiles between the WT and lib. The mRNAs used for analysis were isolated from the root tissues of 8-day-old seedlings. All comparisons were made against the gene expression level of WT plants. The corrected P-value of <0.05 was used to select the genes for comparison, and only those genes whose expression was 2-fold higher or lower in lib than in the WT were used for analysis. The results indicated that expression in the lib mutant was up-regulated for 453 genes and down-regulated for 451 genes (Supplementary Tables S1, S2 at JXB online). Using the SBC Analysis program (http://sas.ebioservice.com/index.jsp), it was found that almost all the pathways that were significantly affected in the lib mutant are involved in amino acid and secondary metabolism (Supplementary Table S3). The significant genes over-represented in the metabolic pathways are listed in Supplementary Table S4. The expression levels of the genes that are involved in the BCAA metabolic pathway were then examined. As predicted by the SBC analysis program, the expression of two IPMS-like genes (At5g23010 and At5g23020; Field ; Kroymann ) increased 2.88- and 4.65-fold, respectively, and the expression of the BCAT4 gene (At3g19710; Diebold ) increased 6.77-fold in the lib mutant (Fig. 7A). These three genes have also been demonstrated to participate in the chain elongation pathway in the biosynthesis of methionine-derived glucosinolates (Sønderby ). The increased expression of these three genes in lib was validated by quantitative real-time PCR analysis (Fig. 8). Using quantitative real-time PCR, it was also found that the increased expression of these three genes was reversed in the complemented line and in lib grown on isoleucine-containing MS medium (Fig. 8), supporting the view that the altered expression of these three genes in lib was due to a deficiency in isoleucine biosynthesis. In contrast, the expression of BCAT2 (At1g10070; Diebold ), branched-chain α-keto dehydrogenase (BCKDH; Mooney ) E1α subunit (At1g21400), BCKDH E1β subunit (At3g13450), BCKDH E2 subunit (At3g06850), isovaleryl-CoA dehydrogenase (IVD, At3g45300; Faivre-Nitschke ), and 3-methycrotonyl-CoA carboxylase (MCCase; Song ) α subunit (At1g03090) and β subunit (At4g34030) was lower in the lib mutant than in the WT. The decrease in the transcript levels of these genes was moderate, but was statistically significant, ranging from 27% to 46% of the WT. The fold changes of these transcripts are indicated in Fig. 7B. BCKDH and IVD are the enzymes involved in the early steps of degradation of all three BCAAs, and MCCase catalyses the reaction toward the degradation of leucine (Fig. 7B). The decrease in the transcript levels of the genes encoding these enzymes suggested that the overall degradation activity of BCAA is reduced in lib.

Fig. 8.

Relative expression of the two IPMS-like genes (At5g23010 and At5g23020) and the BCAT4 gene (At3g19710) in the WT, lib, a complemented line (C), and lib grown on half-strength MS medium supplemented with 5 µM isoleucine [lib(+Ile)]. Total RNAs were extracted from roots of 8-day-old seedlings, and the relative expression levels of the three genes were determined by quantitative real-time PCR. The AGI code of the genes is indicated at the bottom. The expression of the gene in the WT was set at 1.

Discussion

In this work, an Arabidopsis mutant that has defects in both root cell proliferation and cell expansion, but not in root cell patterning, was characterized. The phenotypes of the lib mutant are caused by a T-DNA insertion within the OMR1 gene (At3g10050), which encodes a TD. In plants, TD catalyses the deamination and dehydration of threonine to produce ammonia and 2-oxobutanoate, which is the first and also the committed step in the biosynthesis of isoleucine (Binder, 2010). Complete loss of TD activity in a Nicotiana plumbaginifolia mutant or inhibition of TD in Arabidopsis and corn by herbicide results in plant death (Colau ; Szamosi ). Strong silencing of TD in Nicotiana attenuata also caused severely stunted growth (Kang ). Thus, complete loss or strong inhibition of TD enzyme activity might mask the specific roles of isoleucine in plant development. In the lib mutant used here, TD activity was 40% lower than in the WT because of the T-DNA insertion, which results in the expression of a truncated version of the OMR1 mRNA. The amounts of isoleucine in roots of the lib mutant were about half of those in the WT. The lib mutant phenotype could be rescued by application of exogenous isoleucine or by transformation with a WT OMR1 gene, but not by supplementation with phenylalanine, whose level was also decreased in the lib mutant. These results support the notion that the lib mutant phenotype is specifically caused by a deficiency in isoleucine. Thus, the lib mutant provides a useful material to examine the role of isoleucine in plant development.

How does the deficiency in isoleucine impair root development? One possibility is that isoleucine may play a regulatory or signalling role in the control of root development. Previous studies have shown that various amino acid deficiencies have different effects on plant growth and development. For example, the Arabidopsis mutant hpa1 has a reduced amount of histidine (Mo ). Besides having a reduced root meristem, the hpa1 mutant also has altered QC activity and cell patterning, which were not observed in the lib mutant. A gapcp double mutant of Arabidopsis that produces a low level of serine has arrested primary root growth but normal lateral root growth (Muñoz-Bertomeu ). This is also different from the lib mutant. Another Arabidopsis mutant, smo1/trp2-301, which bears a defect in biosynthesis of tryptophan, has small above-ground organs but normal primary root growth (Jing ). Although tryptophan is the precursor for the biosynthesis of auxin, the smo1/trp2-301 mutant phenotypes could only be rescued by exogenously applied tryptophan, but not auxin, indicating that the mutant phenotype specifically reflects a deficiency in tryptophan. Other unique phenotypes that were observed in the roots of lib were the alternative appearance of LRC and SRC regions as well as the regions with or without root hairs (Fig. 2). In Arabidopsis, biosynthesis of isoleucine is significantly slower in the light than in the dark (Coruzzi and Last, 2000). The diurnal change in the isoleucine biosynthesis rate may result in the alternative patterns of LRCs and SRCs as well as in the formation of root hairs. The distinct effects of perturbed homeostasis of different amino acids on root development suggest that these amino acids may have regulatory roles rather than simply serving as building blocks for general protein synthesis. This notion is supported by the fact that general protein synthesis is not impaired in lib and smo1/trp2-301 mutants. In plants, jasmonate (JA) is an important hormone that regulates a variety of biological processes. The most active form of JA is its conjugate with isoleucine (JA-Ile) (Thines ). Tobacco plants with a partial deficiency in isoleucine were highly susceptible to insect attack (Kang ). Further studies indicated that low production of isoleucine in these plants decreased the formation of JA-Ile, thus affecting the JA signalling pathway and finally reducing plant resistance to herbivore attack. The role of JA in regulating root growth has been well recognized (Wasternack, 2007). Recent research demonstrates that JA regulates auxin synthesis and transport during lateral root growth (Sun ). Thus, it is possible that the low level of isoleucine in the lib mutant may reduce formation of JA-Ile, which in turn affects root development through the change of the JA signalling process or the interactions with other hormones.

Considering the results obtained from this work, however, the possibility that the impairment in root development in lib is due to a reduction of general protein biosynthesis cannot be completely excluded. Although the total protein content in lib is not reduced per gram of fresh weight or per cell, it is reduced per plant. The hypothesis that the lib phenotype results from reduced protein synthesis is supported by the fact that, like lib root growth, lib shoot growth is also reduced on medium lacking isoleucine. The suppression of the mutant phenotypes in soil-grown lib plants might result from the presence of a small amount of isoleucine (derived from decomposed plants or microorganisms) in soil.

To maintain normal growth and development, plants must tightly regulate both primary and secondary metabolism, including the pathway that controls BCAA homeostasis. Arabidopsis has seven annotated genes encoding BCAT, and six of them are actively transcribed (Diebold ). Although it has been biochemically demonstrated that BCATs are required for the last step of biosynthesis and the first step of degradation for all three BCAAs, it is still unclear how each of these six BCATs contributes to the biosynthesis and degradation process. The present microarray analysis indicated that among these six active BCAT genes, only BCAT2 (At1g10060) and BCAT4 (At3g19710) showed significant changes of expression in the lib mutant. Because the partial deficiency of isoleucine in lib down-regulates the transcript level of BCAT2 but up-regulates BCAT4, it seems unlikely that these two genes both act in either biosynthesis or degradation. Considering that the three major enzymes involved in BCAA degradation are all down-regulated in lib (as discussed below), it was believed that BCAT2 is involved in the first step of the BCAA degradation pathway (Fig. 7B).

Interestingly, although there are five enzymes involved in biosynthesis of isoleucine, isoleucine deficiency in lib did not alter transcript levels of any of the genes encoding these enzymes. Instead, the transcript levels of the genes encoding the enzymes involved in the three early steps of degradation is decreased (Fig. 7B). These enzymes include BCKDH, IVD, and MCCase. BCKDH catalyses the second step in the BCAA degradation pathway. This enzyme is a protein complex consisting of multiple subunits of E1α, E1β, E2, and E3 (Mooney ). In lib, the expression of the genes encoding E1α, E1β, and E2 subunits was coordinately down-regulated. Similarly, MCCase, which is involved in leucine degradation, consists of two subunits, α and β (Alban ; Song ). In the lib mutant, the expression of both of these subunits is decreased. The strong co-expression of E1α, E1β, and E2 subunits of BCKDH, α and β subunits of MCCase, IVD, and BCAT2 was also observed when plant cells were exposed to carbon starvation, dark, or abscisic acid (ABA) treatment (Fujiki ; Mentzen ; Urano ; Binder, 2010). Based on these results, it is proposed that in the biosynthesis pathway, isoleucine homeostasis may mainly be achieved through the regulation of enzyme activity, such as feedback control of TD activity by isoleucine, rather than through transcriptional effects. In the degradation pathway, however, the regulation of BCAA homeostasis may be achieved mainly through the alteration of the transcription of degradative enzymes.

Because BCAT4 acts on leucine but not on isoleucine or valine (Schuster ), it was speculated that BCAT4 is involved in the last step of leucine biosynthesis (Fig. 7A). Also, the increased expression of two IPMS-like genes (At5g23010 and At5G23020) suggested that more carbon might flow into the leucine biosynthesis pathway through the branch point from the valine biosynthesis pathway (Fig. 7A). Our analysis, however, did not detect a significant increase of leucine content in lib (Table 1). Instead, BCAT4 and these two IPMS-like genes are well known to be involved in the chain elongation pathway in the biosynthesis of methionine-derived glucosinolates in Arabidopsis. This pathway starts with a deamination reaction by a BCAT that is followed by a condensation reaction catalysed by methylthioalkylmalate synthase (MAM) (Sønderby ). BCAT4 has been implicated in catalysis of the initial deamination step of methionine (Schuster ), and the products of two IPMS-like genes have been demonstrated to have MAM activity (Kroymann ; Field ). In addition, the expression of several other genes involved in the glucosinolate biosynthesis pathway (Hirai ; Sawada ) was up-regulated in lib (Supplementary Tables S3, S4 at JXB online). Taken together, the partial deficiency of isoleucine in lib may enhance biosynthesis of methionine-derived glucosinolates rather than biosynthesis of leucine. The increased accumulation of glucosinolate in lib, however, requires experimental confirmation. Furthermore, because a change in glucosinolate content can affect plant growth and development (Kim ; Reintanz ; Tantikanjana ), the retarded growth of lib could also be due to the indirect effect resulting from the increased accumulation of glucosinolates.

Supplementary data

Supplementary data are available at JXB online.

Figure S1. Changes in the number of the first-order lateral roots of WT and lib mutants from 5 to 13 DAG.

Figure S2. The patterns of root hair development in 8-day-old WT and lib seedlings.

Figure S3. The expression patterns of seven cell-specific markers in the roots of 8-day-old WT and lib mutant seedlings.

Figure S4. The relative expression of the OMR1 gene in roots of the WT and transgenic plants.

Text S1. Methods for vector construction and plant transformation.

Table S1. The genes whose expression is up-regulated in the lib mutant.

Table S2. The genes whose expression is down-regulated in the lib mutant.

Table S3. The metabolic pathways that are perturbed in the lib mutant.

Table S4. The significant genes over-represented in the metabolic pathways.