JXB Advance Access published online on September 4, 2007
Journal of Experimental Botany, doi:10.1093/jxb/erm179
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
© 2007 The Author(s).
This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/2.0/uk/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.This paper is available online free of all access charges (see http://jxb.oxfordjournals.org/open_access.html for further details)
RESEARCH PAPER |
Integrated metabolomic and transcriptomic analyses of high-tryptophan rice expressing a mutant anthranilate synthase alpha subunit
1CREST, Japan Science and Technology Agency, Tokyo 103-0027, Japan
2Division of Applied Life Sciences, Department of Agriculture, Kyoto University, Kyoto 606-8502, Japan
3Department of Information Science and Technology, National Agricultural Research Center, 3-1-1 Kannondai, Tsukuba, Ibaraki 305-8666, Japan
4Department of Rice Breeding, National Institute of Crop Science, 2-1-18 Kannondai, Tsukuba, Ibaraki 305-8518, Japan
5Tokyo University of Agriculture, Faculty of Agriculture, 1737 Funako, Atsugi, Kanagawa 243-0034, Japan
To whom correspondence should be addressed at Tokyo University of Agriculture, 1737 Funako, Atsugi, Kanagawa 243-0034, Japan. E-mail: k3wakasa{at}nodai.ac.jp
Received 9 March 2007; Revised 18 June 2007 Accepted 26 June 2007
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Transgenic rice plants overexpressing a mutant rice gene for anthranilate synthase alpha subunit (OASA1D) accumulate large amounts of free tryptophan (Trp) with few adverse effects on the phenotype, except for poor germination and weak seedling growth. Metabolic profiling of 8-d-old seedlings of Nipponbare and two high-Trp lines, HW1 and HW5, by high performance liquid chromatography-photo diode array (HPLC-PDA) confirmed that, relative to Nipponbare, only the peak attributed to Trp was significantly changed in the profiles of the OASA1D lines. More detailed and targeted analysis using HPLC coupled with tandem mass spectrometry revealed that the OASA1D lines had higher levels of anthranilate, tryptamine, and serotonin than Nipponbare, but these metabolites were at much lower levels than free Trp. The levels of phenylalanine (Phe) and tyrosine (Tyr) were not affected by the overproduction of Trp. Transcriptomic analysis by microarray validated by quantitative Real-Time PCR (qRT-PCR) revealed that at least 12 out of 21 500 genes showed significant differential expression among genotypes. Except for the OASA1D transgene and a putative IAA ß-glucosyltransferase, these were not related to Trp metabolism. Most importantly, the overexpression of the OASA1D and the consequent accumulation of Trp in these lines had little effect on the overall transcriptome, consistent with the minimal effects on growth and the metabolome. Integrated analysis of the metabolome and transcriptome of these OASA1D transgenic lines indicates that the over-accumulation of free Trp may be partly due to the low activity of Trp decarboxylase or other metabolic genes that directly utilize Trp as a substrate.
Key words: Metabolic profiling, microarray analysis, OASA1D, Oryza sativa, tryptophan
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Metabolic engineering involves the modification of metabolic pathways in living organisms through recombinant DNA technology. It has been used to generate higher levels of desirable plant metabolites such as alkaloids, essential oils, seed oils, carotenoids, etc (Broun and Somerville, 2001). Rice (Oryza sativa L.) is the major staple food of half of the world's population and improving its nutritional qualities could have a major impact on the quality of life of billions of people. Tryptophan (Trp), a key amino acid, is usually found in limited quantities in cereals such as rice and maize. To date, the most dramatic improvement in free Trp levels in plants has been obtained by Tozawa et al. (2001) who introduced a mutant form of the anthranilate synthase (AS) gene of rice (OASA1D) into a rice cultivar, Nipponbare. It has been demonstrated that AS produced by the OASA1D mutant was insensitive to feedback regulation by Trp and transformed plants were obtained that accumulated up to 5354 nmol g–1 fresh weight (FW) of free Trp, 35 times more than the 153 nmol g–1 FW found in the wild-type Nipponbare. HW1 and HW5 are the best performing OASA1D transgenic lines (Wakasa et al., 2006) and they are now undergoing field-testing and feeding tests.
Trp is a key participant in many of the cell's metabolic activities. Accordingly, the overproduction of free Trp caused by the loss of feedback regulation of AS possibly affects various aspects of plant metabolism. Trp is synthesized via the shikimate pathway along with other aromatic amino acids, phenylalanine (Phe) and tyrosine (Tyr). Thus, funnelling of the metabolic flow into the Trp pathway may result in the depression of Phe and Tyr pools because of an insufficient supply of common precursors from the shikimate pathway. In addition, the loss of feedback regulation of AS increases production of anthranilate and downstream intermediates of Trp.
Trp is converted to a diverse range of secondary metabolites such as indole acetic acid (IAA), serotonin and, in many medicinal plants, the pharmacologically important terpene indole alkaloids (Fig. 1). The decarboxylation of Trp to tryptamine is a common metabolic process leading to the formation of secondary metabolites. Indeed, the accumulation of tryptamine has been detected in lesions formed on the leaves of the rice sl mutant (Ueno et al., 2003). The presence of melatonin, a Trp-derived secondary metabolite, has been reported in rice (Hattori et al., 1995). Enhanced AS activity accompanied the production of Trp-derived secondary metabolites in response to exogenous stimuli (Niyogi and Fink, 1992; Bohlmann et al., 1995), indicating the concerted regulation of Trp production and the downstream secondary metabolism. Research into the regulation mechanisms of Trp biosynthesis and metabolism by the analysis of OASA1D transgenic rice should be a basis for rational metabolic engineering of these Trp-derived pathways.
|
The overaccumulation of Trp caused by the introduction of the OASA1D gene may affect unexpected processes in the plant's physiology via a sequence of multiple events that may include altered gene expression. Transgenic Arabidopsis that express OASA1D at high levels showed increased levels of Phe and Tyr but the levels of their downstream products, phenylpropanoids and flavonoids, decreased (Ishihara et al., 2006). Thus, the decrease of Phe-derived secondary metabolites can not be simply explained by the competition between Trp and Phe branches of chorismate metabolism. The effect on global gene expression by altered amino acid metabolism was analysed in the double mutant of pal1 and pal2 genes that encode Phe ammonia-lyase isoforms in Arabidopsis (Rohde et al., 2004). The double mutant accumulated Phe at a high concentration and exhibited the altered expression of a large number of genes involved in sugar and amino acid metabolism.
In this study, a combination of microarray analysis with non-targeted and targeted metabolic profiling was employed to investigate the effect of OASA1D expression on the transcriptome and metabolome of rice plants. Agilent's rice oligo-array was used to characterize the transcriptome of transgenic rice plants containing the UBI:OASA1D expression cassette grown in vitro, the methods used to analyse microarray data were evaluated, and selection parameters were developed with regard to validation of quantitative Real-Time PCR (qRT-PCR). In addition, metabolic profiles were obtained by high performance liquid chromatography-photo diode array (HPLC-PDA) to detect changes in the levels of major metabolites. The amounts of anthranilate, Trp, tryptamine, and serotonin were measured by HPLC coupled with tandem mass spectrometry (LC-MS/MS). With this integrated, multi-disciplinary approach (Hirai et al., 2004; Kristensen et al., 2005), it was hoped to elucidate the reasons for the overaccumulation of Trp and the low levels of Trp derivatives in OASA1D transgenic rice lines, and to discover how the rice plant deals with the abnormal situation caused by Trp overaccumulation.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Plant material and in vitro culture
HW1 and HW5 lines were isolated from the original Agrobacterium-mediated transformation of Nipponbare callus with OASA1D (Wakasa et al., 2006). Seeds from Nipponbare, HW1, and HW5 were harvested from an isolated field and manually dehulled, surface-sterilized, and then transferred to sterile jars containing a metal mesh for support and 30 ml sterile deionized distilled water. To compensate for slow germination in the OASA1D transgenics, 10, 20, and 15 seeds of Nipponbare, HW1, and HW5, respectively, were sown in each glass jar. The seeds were allowed to germinate in a culture room at 28 °C with a 16/8 h light/dark photoperiod. Eight days after sowing, shoots from each jar were pooled, quick-frozen in liquid nitrogen, and stored at –80 °C until further use. Pooled samples from each jar were considered as independent biological replicates. Samples from at least two replicates (i.e. two jars) were obtained from each genotype.
RNA extraction and quantitation
Total RNA from each pooled replicate were extracted from 100 mg shoots with RNeasy (Qiagen, Valencia, CA, USA) plus DNase treatment according to the manufacturer's instructions. The concentration and quality of total RNA were determined using an Agilent 2100 Bioanalyser and RNA 6000 kit (Agilent Technologies, Palo Alto, CA). Aliquots from the initial extracts were used for both microarray and qRT-PCR assays.
Microarray hybridization
RNA samples:
A reference sample was constructed by pooling the two replicates obtained from Nipponbare at a 1:1 ratio. The other six samples (two each from Nipponbare, HW1, and HW5) were then co-hybridized with this constant reference sample. In addition, hybridization followed a flip-dye design such that the first replicate was labelled with Cy3 and the second with Cy5. This is based on the recommendation of Dobbin et al. (2003) who concluded that flip-dye labelling is required when the identification of genes expressed differently in the reference sample and in the non-reference samples is desired.
Hybridization protocol:
Using Agilent's Low Input RNA Fluorescent Amplification Kit, 400 ng of total RNA from each replicate were reverse transcribed, linearly amplified, and then labelled with either cyanine-3 CTP (Cy3) and cyanine-5 CTP (Cy5). Cy3- and Cy5-labelled cRNAs were mixed with In situ Hybridization Kit Plus (Agilent) and incubated at 60 °C for 30 min to fragment the cRNAs to an average size of 100–200 bp. The hybridization solution and the fragmented cRNAs were injected into a hybridization chamber containing one Agilent's Rice Oligo microarray slide; each slide has 22 575 60-mer probes including 
21 500 probes specific for unique rice full-length cDNA transcripts annotated in the KOME database (http://cdna01.dna.affrc.go.jp/cDNA/). After incubation for 17 h at 60 °C, the slides were washed and the hybridization signals were detected with the G2565BA Microarray Scanner System from Agilent. Scanned microarray images were processed with Feature Extraction 6.1.1 software (Agilent).
Microarray data analysis:
The log of Cy5/Cy3 signal ratios was calculated with Agilent's Feature Extraction software. The probability of significant genetic differences in gene expression was estimated by performing a monofactorial analysis of variance (AOV) of the signal log ratios (Cy5/Cy3). Genes that showed differential expression at P <0.01 were subjected to Hierarchical Cluster analysis by average linkage clustering of Euclidean distances based on the log10 reference adjusted signal ratios. The analysis was performed in TIGR's Multiple experiment Viewer (Mev) v 3.1 (Saeed et al., 2003).
The average mean fold change (MFC) of each gene, based on the Loess-normalized signal intensity of HW1 and HW5 relative to Nipponbare, was also calculated. An MFC 
0.67 or MFC 
1.5 was adopted as an alternate selection parameter.
Correlation coefficients between qRT-PCR and the anti-log of the ratios generated by Agilent's Feature extraction software data were performed via Excel. Microarray data were also subjected to arsinh and cube root transformations and normalization by a stably and constitutively expressed ‘housekeeping’ gene, actin (AK072796).
qRT-PCR
Reverse transcription of 5 µg total RNA from each sample was performed with Superscript III (Invitrogen) at 42 °C. PCR primers (19–22 bp) were designed with Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi) to amplify 50–150 bp fragments with a melting point range from 75–85 °C. Following the recommendations of Perkin-Elmer for SYBR Green primers, the last five nucleotides of each primer were limited to no more than two C or G residues to minimize the frequency of primer-dimers and false priming products. QRT-PCR was performed in an ABI Prism 7000 Sequence Detection System (Perkin-Elmer Applied Biosystems, Foster City, CA) using a SYBR Green RT-PCR Reagents kit in 10 µl reaction volumes containing 2.5 ng cRNA template and 2.5 pmol primer. The standard two-step thermal cycling protocol was repeated 50 times and then the melting point of the amplified products was determined with the prescribed dissociation protocol to verify that the signal came from a single amplicon.
Some of the actin homologues included in Agilent's Rice oligo-array showed unstable expression levels between replicate slides or among genotypes. For the qRT-PCR experiment, the actin gene (AK072796) that was used to normalize expression data of other genes was selected precisely because of the relative stability of its signal among replicates and between genotypes.
Metabolic profiling
HPLC-PDA:
After weight measurement, leaf samples were extracted with 10 vols of 80% methanol for 12 h. HPLC-PDA was performed essentially as described by Matsuda et al. (2005a). Chromatograms were generated at UV 254 nm to represent the metabolic profile of each entry.
LC-MS/MS for anthranilate, Trp, and related secondary metabolites:
Rice leaves were extracted in 10 vols of 80% methanol for 12 h. HPLC separation was performed with an Agilent 1100 HPLC system (Agilent Technologies, Palo Alto, CA, USA) equipped with a Mightysil RP-18 GP column (150 mm long, 2.0 mm i.d., 3 µm particle size, Kanto Chemical, Tokyo, Japan) at a flow rate of 200 µl min–1. The solvents used were 0.1% acetic acid (A), methanol (B), 0.01% trifluoroacetic acid (C), and acetonitrile (D). The following 10 min linear gradients were applied for separation of compounds: (i) 30–45% B/(A+B) for anthranilate; (ii) 40–70% D/(C+D) for indole; (iii) 5–25% B/(A+B) for Phe and Tyr; (iv) 20–35% D/(C+D) for melatonin; and (v) 5–50% C/(C+B) for Trp and other compounds. MS detection was performed on an API-3000 triple stage quadrupole mass spectrometer equipped with a TurboIonSpray ionization source (Applied Biosystems, Foster City, CA, USA). Concentrations of metabolites were determined by multiple reaction monitoring (MRM). Monitored mass transitions were from m/z 138 to m/z 120 for anthranilate, from m/z 118 to m/z 91 for indole, from m/z 205 to m/z 188 for Trp, from m/z 166 to m/z 120 for Phe, from m/z 182 to m/z 136 for Tyr, from m/z 161 to m/z 144 for tryptamine, from m/z 177 to m/z 160 for serotonin, from m/z 191 to m/z 174 for 5-methoxytryptamine, from m/z 175 to m/z 130 for gramine, and from m/z 233 to m/z 174 for melatonin. MS conditions were optimized for MRM detection using authentic compounds (Wako Pure Chemical, Osaka, Japan). Optimized MS conditions include nebulizer gas flow (NEB), curtain gas flow (CUR), ionspray voltage (IS), TurboIonSpray temperature (TEM), collision gas pressure (CAD), declustering potential (DP), focusing potential (FP), entrance potential (EP), collision energy (CE), and cell exit potential (CEP).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Phenotype of 8-d-old seedlings
The phenotypes of the seedlings were compared among Nipponbare, HW1, and HW5. Nipponbare showed a more uniform growth rate after germination (% germination: 75.6±13.9, seedling weight: 20.4±3.4 mg) compared with HW1 (% germination: 55.0±8.6, seedling weight: 11.0±2.2 mg) and HW5 (% germination: 93.3±5.8, seedling weight: 21.5±4.0 mg). Both HW1 and HW5 had very variable seedling growth and germination, with HW1 showing the lowest germination and seedling growth rate among the three genotypes. Despite the initially poor growth, however, the three genotypes showed similar vegetative biomass at maturity, indicating that HW1 and HW5 eventually caught up with Nipponbare in terms of vegetative growth (Wakasa et al., 2006). Leaf samples collected in the early seedling phase were pooled carefully to determine the basis for the observed phenotypic differences between Nipponbare and the high Trp lines.
The overaccumulation of Trp in tissues is obviously an unusual status for plants. Nevertheless, no evident negative effects of Trp overproduction has been found except for a low germination rate and retarded seedling growth immediately after germination (Wakasa et al., 2006). It is likely that OASA1D transgenic rice activates various molecular mechanisms to maintain metabolic homeostasis by regulating the expression of genes related to Trp biosynthesis and catabolism. This regulation will increase and decrease the levels of Trp-related metabolites.
Non-targeted profiling by HPLC-PDA
The effect of OASA1D expression on the aromatic components of the metabolic profile of HW1 and HW5 seedlings was analysed by reversed phase HPLC coupled with a photodiode array detector monitoring a wavelength range of 190 to 400 nm. Figure 2 shows chromatograms obtained at 254 nm, which is suitable for the detection of aromatic compounds including anthranilate-related metabolites. The metabolic profiles revealed that the peak attributed to Trp was drastically increased by the expression of OASA1D and no apparent accumulation of other metabolites was observed. This suggests that most of the excess anthranilate supplied to the Trp biosynthetic pathway by the overexpression of OASA1D was used for Trp synthesis and that downstream gene activity was insufficient to deplete the accumulated Trp significantly. Essentially similar results were obtained at wavelengths other than 254 nm (data not shown). Specific accumulation of Trp was also observed upon introduction of the OASA1D transgene into rice calli (Morino et al., 2005), potato (Matsuda et al., 2005b), and Arabidopsis (Ishihara et al., 2006).
|
Targeted metabolic analysis by LC-MS/MS
To evaluate the effects of the introduction of OASA1D on the accumulated amounts of intermediate compounds, targeted analysis was performed of anthranilate, aromatic amino acids (Trp, Phe, Tyr), and indole alkaloids (tryptamine and serotonin) in transgenic lines by LC-MS/MS in MRM mode (Fig. 3). The amounts of Trp in the transgenic lines were 46- and 30-fold larger than that in Nipponbare (49 nmol g–1 FW). The relative amounts of Trp between transgenic and wild-type plants were within the range of previously reported values (Tozawa et al., 2001). The analysis also showed the accumulation of anthranilate at significantly higher levels in HW1 and HW5 (Fig. 3). This is probably due to the increase of metabolic flow in the Trp biosynthetic pathway. The amount of indole was under the detection limit (250 pmol g–1 FW) in all lines. The enhanced Trp synthesis in these OASA1D transgenic lines had little effect on metabolic flow into the branch for Phe and Tyr in the shikimate pathway.
|
Tryptamine, serotonin (5-hydroxytryptamine), 5-methoxytryptamine, melatonin, and gramine were included in targeted metabolic analysis as Trp-derived secondary metabolites. Tryptamine (Ueno et al., 2003) and melatonin (Hattori et al., 1995) have been shown to be present in rice. Serotonin and 5-methoxytryptamine are putative biosynthetic intermediates of melatonin (Murch et al., 2000), while gramine is a defensive secondary metabolite found in barley (Corcuera, 1993). Tryptamine and serotonin were detected in the extracts (Fig. 3), but the amounts of other compounds were under the detection limits (300 pmol g–1 FW for 5-methoxytryptamine; 40 pmol g–1 FW for melatonin, and 110 pmol g–1 FW for gramine).
The analysis indicated that the levels of tryptamine and serotonin were slightly increased by the OASA1D expression. The absolute levels and fold increase of tryptamine (1.3–1.8 nmol g–1 FW, 2.8–3.8-fold) and serotonin (16–29 nmol g–1 FW, 2.1–3.7-fold) in the transformants were far less than that of Trp (more than 1500 nmol g–1 FW, 30–46-fold), suggesting that the conversion from Trp to tryptamine may be strictly regulated in transgenic lines.
On the basis of these findings, the following question arose: why is so much Trp accumulated in OASA1D transgenic plants and, conversely, why is so little channelled into downstream products? Hence microarray analysis was carried out to help explain this conundrum.
Detection of differential gene expression by microarray analysis
Microarray hybridization was executed considering the possibility of differential expression due to ‘dye’ effects in microarrays, one replicate was labelled with Cy5 while the other was labelled with Cy3. The reference design, where each sample is co-hybridized with a common reference, was adopted to facilitate analysis of data. The reference consisted of a pooled sample of RNA independently extracted from each replicate jar containing the Nipponbare seedlings.
Loess-normalized Cy5 or Cy3 signal values were used to calculate MFC, which is defined as the average signal in OASA1D lines relative to the common reference. A selection parameter of MFC 0.67 or MFC 1.5 identified 2211 genes that showed at least 50% change in OASA1D transgenics relative to the wild-type control (Fig. 4). Most (70.6%=1560/2211) of the genes were up-regulated in OASA1D transgenics relative to Nipponbare. The genes were further classified into the following categories: U (unknown or unclassified), CWMT (cell wall, membrane, and transport), CPR (cell processes and reproduction), EF (energy flow), ER (environmental response), and MD (maintenance and development). While basically showing similar proportions in these various categories, it is interesting to note that EF genes accounted for a greater proportion of down-regulated genes as compared to up-regulated genes in OASA1D transgenics.
|
Monofactorial AOV for selecting genes that show relatively low variation between replicates and high variation between genotypes revealed 70 genes that showed differential gene expression at a P <0.01. The relationships among these genes are shown in the hierarchical cluster in Fig. 4. Only two ESTs showed expression patterns that were almost identical to that of OASA1D: one codes for an NBS-LRR-like [nucleotide-binding site (NBS) and leucine-rich repeat (LRR)] protein while the other codes for an unknown protein (indicated by red asterisks in Fig. 4). Most (58.5%) of the differentially expressed ESTs code for unknown proteins while the known proteins are associated with various aspects of plant metabolism. Except for OASA1, none of the differentially expressed genes identified by AOV could be associated with Trp metabolism. Comparing the number of genes identified, it is clear that AOV is more conservative than MFC.
There were only 22 genes (out of 21 500) that satisfied both criteria described above, i.e. MFC 0.67 or MFC 1.50 and significant genotype effects at P <0.01 in AOV, suggesting that the introduction of OASA1D and resulting Trp overaccumulation had minimal impact on global gene expression.
Other methods of data transformation such as arsinh (Huber et al., 2002), cube root (Keay et al., 2003), and normalization (using actin) were evaluated, but these turned out not to be superior to the above methods in identifying significantly different gene expression.
Differentially expressed genes according to qRT-PCR
The expressions of differentially expressed genes selected by the microarray analysis were verified since array performance is quite sensitive to variations in each step of the complex assay procedure. Validation of microarray data has been attempted using biological and technical replicates on different slides, microarrays of other platforms (i.e. cDNA versus oligo-based microarrays; Mitchell et al., 2004), qRT-PCR (Allen and Nuss, 2003; Dallas et al., 2005), semi-quantitative RT-PCR (Romualdi et al., 2003), and conventional northern blots (Rabbani et al., 2003). Kimura et al. (2004) used Agilent's rice 22K custom oligo DNA microarray to confirm data on the expression pattern of DNA repair genes obtained via northern and in situ hybridization in the shoot apical meristem (SAM) and mature leaves of rice. In this study, the qRT-PCR method was employed.
Microarray analysis detected the differential expression of 2259 genes based on both MFC and AOV (at P <0.01). The genes were classified into three groups: (1) genes showing significant genotype effects at P <0.01 (48 genes); (2) genes with MFC 0.67 or MFC 1.50 (2189 genes); and (3) genes that satisfied both criteria (22 genes). As shown in Table 1, 4, 20, and 5 genes were selected from groups 1, 2, and 3, respectively, and their expression was analysed by qRT-PCR. The differential expression of 12 genes (2, 5, and 5 genes in groups 1, 2, and 3, respectively) was confirmed by this analysis. Correlation coefficients between raw Cy5/Cy3 data from Agilent's microarray and the corresponding qRT-PCR data are shown in Table 1. In our experiment, 72% (21/29) of the genes showed significant (P <0.05) correlation coefficients, indicating general agreement between qRT-PCR and the rice oligo array as reported by Dallas et al. (2005); however, a large (28%) proportion of the genes gave non-significant correlations between microarray and qRT-PCR signals.
|
Verification with qRT-PCR confirmed the presence of significant differences among genotypes in 50% (2/4), 20% (5/20), and 100% (5/5) of groups 1, 2, and 3, respectively. This indicates that selection using both parameters will have a higher probability of identifying differentially expressed genes. The combined approach using data from MFC and AOV may prove to be the most rigorous method for prioritization of genes for verification by qRT-PCR.
Trp-related genes showing significant differential expression among genotypes
Among the 2211 genes that were identified by MFC selection criteria, four genes, including AK072053 (homologous to the OASA1D transgene), AK106302 (coding for a putative IAA ß-glucosyltransferase protein that functions in the production of IAA-Glc conjugates in plants), AK065830 (coding for a tyrosine/dopa decarboxylase), and AK069031 (annotated as the orthologue of the Trp decarboxlase gene in Catharanthus roseus) were associated with Trp metabolism (Fig. 5). As expected, overexpression of anthranilate synthase increased the levels of OASA1 (and OASA1D) transcripts in the transgenic lines. The mRNA levels of AK106302 and AK069031 increased as well, an indication that their transcriptional mechanisms may be responding to increased levels of the anthranilate synthase precursor. AK065830, however, responded in the opposite direction (i.e. decreased), an indication that it may not directly participate in the tryptophan metabolic pathway in rice.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
The effects on metabolism of the introduction of OASA1D, which is closely related to the Trp pathway, were investigated. Non-targeted profiling by HPLC-PDA revealed that the introduction of OASA1D did not affect the accumulation of major aromatic compounds with the exception of the intended increase in free Trp. Many of the peaks in the UV spectra generated by HPLC-PDA corresponded to flavonoids and phenylpropanoids that are derived from Phe. In addition, targeted analysis by LC-MS/MS indicated the limited effects of the introduction of OASA1D on the pools of Phe and Tyr. Thus, competition between the branch for Trp and the branch for Phe and Tyr was unlikely in OASA1D expressing rice seedlings. Microarray and qRT-PCR analyses did not show significant changes in the transcription of genes encoding the enzymes on Phe and Tyr branches and upstream of the shikimate pathway. The synthesis of aromatic amino acids is allosterically regulated at the enzyme level. Trp has been demonstrated to activate one isoform of chorismate mutase that is the first committed enzyme for the synthesis of Phe and Tyr in various plant species (Singh et al., 1986). This feedback activation may function to sustain metabolic flow into Phe and Tyr.
Transcriptomic analysis by microarray validated by qRT-PCR further demonstrated that only a limited number of genes showed altered expression in OASA1D rice lines. This lack of a global response in the gene expression of OASA1D rice lines is consistent with the unchanged metabolite profile. Among Trp-related genes, only putative TDC (AK069031) showed a slightly enhanced expression, suggesting the absence of an inducible regulatory mechanism that deals with Trp accumulation. In contrast to marginal changes in the gene expression in OASA1D rice plants, the expression of a large number of genes was affected in the pal1 pal2 double mutant of Arabidopsis that accumulated Phe at a high concentration (Rhode et al., 2004). The fundamental reason for this contrast may be the large difference in the metabolic flow between the Trp branch and the branches for Phe and Tyr. Phe serves as a precursor of various flavonoids and phenylpropanoids as well as being a building block of proteins and a substrate for lignin synthesis. Thus, metabolic flow to Phe and Tyr branches from the shikimate pathway is reasonably considered to be much larger than that for the Trp branch.
The introduction of OASA1D significantly increased the anthranilate level; however, the levels in transgenic lines were less than 0.1% of Trp, indicating that anthranilate is effectively processed to Trp in transgenic lines. The lack of response of Trp biosynthesis-related genes such as phosphoribosylanthranilate transferase (Fig. 1) in transgenic lines suggests that the large capacity of the downstream biosynthetic enzymes keeps the anthranilate concentration at a low level by converting it to Trp without increasing transcription levels. Such a large capacity of Trp synthesis suggests that the accumulation of intermediates may be detrimental to wild-type plants. A glucosyltransferase that catalyses the glucosylation of anthranilate has been identified in the Arabidopsis Trp1-100 mutant that is defective in phosphoribosylanthranilate transferase (Quiel and Bender, 2003). This enzyme has been suggested to control the formation of potentially deleterious anthranilate metabolites by rendering excess free anthranilate non-reactive and/or by localizing the conjugate to a subcellular compartment. A similar mechanism does not seem to be operating in transgenic rice because of the absence of peaks that can be assigned to anthranilate glucoside on non-targeted profiling by HPLC-PDA.
The lack of new peaks or changes in peak heights in the chromatograms generated by HPLC-PDA analysis suggest that unexpected metabolism of Trp does not occur in OASA1D rice lines. This is a favourable phenotype from the viewpoint of metabolic engineering because it may support the substantial equivalence of the transgenic lines. A possible explanation for the stability of Trp metabolism in OASA1D rice lines is the isolation of Trp in the compartment where biochemical processes are relatively inactive. Although the function of vacuoles in amino acid storage has not been clarified, putative amino acid transporters have been shown to be localized in the vacuolar membrane (Su et al., 2004).
Decarboxylation of Trp by Trp decarboxylases (Facchini et al., 2000) and N-hydroxylation coupled with decarboxylation by cytochrome P450 enzymes (Hull et al., 2000; Mikkelsen et al., 2000) are two well-characterized reactions leading to Trp-derived secondary metabolites in plants. As the presence of tryptamine (Ueno et al., 2003) and melatonin (Hattori et al., 1995) indicates that the decarboxylation process is indeed active in rice at least under certain physiological conditions, the amounts of tryptamine-derived secondary metabolites were measured. The amounts of tryptamine and serotonin were slightly increased in the transgenic lines; however, the magnitudes of increase were quite small in comparison with that of Trp. The respective increases in the amounts of tryptamine and serotonin in HW1 seedlings were only 0.06% and 0.7% of the Trp levels, suggesting the strict regulation of the secondary metabolism at the first committed step. Similarly, supplemental Trp only slightly increased the accumulation of tryptamine in Cinchona ledgeriana seedlings (Aerts et al., 1990). More importantly, these results suggest that, in addition to the insensitivity to negative feedback regulation of the OASA1D mutant, the apparent low sensitivity of Trp decarboxylase to high substrate (Trp) levels may be a reason for the underutilization of Trp in in vitro grown rice seedlings.
In this context, the down-regulation of the expression of AK065830 (coding for a putative tyrosine/dopa decarboxylase) demonstrated by microarray analysis is of interest because AK065830 is essentially identical to AB162137, a gene that Ueno et al. (2003) demonstrated to have the ability to increase tryptamine levels in vitro (Fig. 5). Since Trp decarboxylase is responsible for the conversion of Trp to tryptamine, the authors subsequently claimed that AB162137 is the Trp decarboxylase gene in rice. On the other hand, the expression levels of AK069031, annotated as the orthologue of the Trp decarboxlase gene in Catharanthus roseus in the Knowledge-based Oryza Molecular Biological Encyclopedia (KOME: cdna01.dna.affrc.go.jp) database, was increased in the transformants. Both microarray (Fig. 4) and qRT-PCR data (Fig. 5) revealed the up-regulation of this transcript in OASA1D transgenics, although neither AOV of the log ratios from the microarray nor qRT-PCR data could detect statistically significant genotypic differences because of the large variation in transcript rates between replications. On the basis of the expression pattern and phylogenetic relationship, AK069031 is a more plausible candidate gene that encodes a functional Trp decarboxylase in rice.
Serotonin, a powerful neurotransmitter in mammals, is naturally present in many food and medicinal plants (Badria, 2002). Tryptamine is converted to serotonin by a constitutively expressed 5-tryptamine hydroxylase, a cytochrome p450 enzyme (Schroder et al., 1999). Berlin et al. (1993) reported increased serotonin after overexpression of TDC in Pergamum harmala. Serotonin increased in OASA1D transgenics in direct proportion to and about 10 times more than the amount of its precursor, tryptamine (Fig. 3). From the relative amounts of Trp, tryptamine, and serotonin, it may be deduced that tryptamine production (i.e. TDC activity) is less efficient than serotonin generation (i.e. 5-tryptamine hydroxylase activity) in these rice lines.
It has been suggested that tryptamine is a candidate intermediate of IAA biosynthetic pathway from Trp. Matsuda et al. (2005a, b) and Morino et al. (2005) revealed that OASA1D transgenic plants have higher levels of IAA and its conjugates, but these compounds were at much lower levels than their free Trp contents. AK106302 may be a key player in the metabolism of IAA in OASA1D transgenic rice. IAA-conjugates form rapidly when IAA homeostasis is perturbed, either from applied IAA or in mutant or transgenic lines in which IAA synthesis is de-regulated (Tam et al., 2000).
|
Conclusions |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
The combination of metabolic profiling with microarray analysis has revealed some insights associated with the overexpression of a metabolic gene in Oryza sativa via Agrobacterium-mediated transformation.
Metabolic analysis demonstrated that the amounts of Phe and Tyr did not decrease in OASA1D transgenic lines, and the amount of anthranilate remained at a low level although it significantly increased. Trp levels increased dramatically, but no additional major metabolite was detected. Overexpression of the OASA1D gene substantially increased the mRNA levels of a putative IAA ß-glucosyltransferase and decreased the expression of one putative Tyr decarboxylase gene. The rest of the genes associated with Trp metabolism did not respond correspondingly. This indicated that the gene(s) needed to process Trp into other metabolites may not respond proportionately to increased amounts of substrate. Consequently, this may at least partly explain the inordinately high amounts of free Trp accumulated in these high Trp rice lines. The low activity of Trp decarboxylase may be a bottleneck suggesting that molecular engineering to change gene regulation or replace with more efficient enzyme variants may provide the key(s) for the full conversion of excess Trp into other, potentially useful, indole compounds.
|
Acknowledgements |
|---|
The authors would like to express their deep appreciation to Dr Y Nagamura and Ms H Motoyama for providing the facilities, technical advice, and assistance in performing the oligo DNA microarray hybridization and scanning, and to Dr H Hirochika for critical comments that helped improve the manuscript, all from the National Institute of Agrobiological Sciences (NIAS, MAFF, Japan). This research was supported by CREST of Japan Science and Technology Corporation.
|
Footnotes |
|---|
* Present address: Metabolome Research Group, RIKEN Plant Science Center, Yokohama, Kanagawa 230-0045, Japan.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionConclusionsReferences |
|---|
Aerts RJ, Van der Leer T, Van der Heijden R, Verpoorte R. Developmental regulation of alkaloid production in Cinchona seedlings. Journal of Plant Physiology (1990) 13:86–91.[Medline]
Allen TD, Nuss DL. Specific and common alterations in host gene transcript accumulation following infection of the chestnut blight fungus by mild and severe hypoviruses. Journal of Virology (2003) 78:4145–4155.[CrossRef][Web of Science]
Badria FA. Melatonin, serotonin, and tryptamine in some Egyptian food and medicinal plants. Journal of Medicine and Food (2002) 5:153–157.[CrossRef]
Berlin J, Rugenhagen C, Dietze P, Fecker LF, Goddijn OJM, Hoge JHC. Increased production of serotonin by suspension and root cultures of Pergamum harmala transformed with a Trp decarboxylase cDNA clone from Catharanthus roseus. Transgenic Research (1993) 2:336–344.[CrossRef][Web of Science]
Bohlmann J, DeLuca V, Eilert U, Martin W. Purification and cDNA cloning of anthranilate synthase from Ruta graveolens: mode of expression and properties of native and recombinant enzymes. The Plant Journal (1995) 7:491–501.[CrossRef][Web of Science][Medline]
Broun P, Somerville C. Progress in plant metabolic engineering. Proceedings of the National Academy of Sciences, USA (2001) 98:8925–8927.
Corcuera LJ. Biochemical basis of the resistance of barley to aphids. Phytochemistry (1993) 33:741–747.[CrossRef][Web of Science]
Dallas PB, Gottardo NG, Firth MJ, Beesley AH, Hoffmann K, Terry PA, Freitas JR, Boag JM, Cummings AJ, Kees UR. Gene expression levels assessed by oligonucleotide microarray analysis and quantitative real-time RT-PCR: how well do they correlate? BMC Genomics (2005) 6:59.[CrossRef][Medline]
Dobbin K, Shih JH, Simon R. Statistical design of reverse dye microarrays. Bioinformatics (2003) 19:803–810.
Facchini PJ, Huber-Allanach KL, Tari LW. Plant aromatic amino acid decarboxylases: evolution, biochemistry, regulation, and metabolic engineering applications. Phytochemistry (2000) 54:121–138.[CrossRef][Web of Science][Medline]
Hattori A, Migitaka H, Iigo M, Itoh M, Yamamoto K, Ohtanikaneko R, Hara M, Suzuki T, Reiter RJ. Identification of melatonin in plants and its effects on plasma melatonin levels and binding to melatonin receptors in vertebrates. Biochemistry and Molecular Biology International (1995) 35:627–634.[Web of Science][Medline]
Hirai MY, Yano M, Goodenowe BG, Kanaya S, Kimura T, Awazuhara M, Arita M, Fujiwara T, Saito K. Integration of transcriptomics and metabolomics for understanding of global responses to nutritional stress in Arabidopsis thaliana. Proceedings of the National Academy of Sciences, USA (2004) 101:10205–10210.
Huber W, von Heydebreck A, Sultmann H, Poustka A, Vingron M. Variance stabilization applied to microarray data calibration and to the quantification of differential expression. Bioinformatics (2002) 18(Supplement 1):S96–S104.[Abstract]
Hull AK, Vij R, Celenza JL. Arabidopsis cytochrome P450s that catalyze the first step of tryptophan-dependent indole-3-acetic acid biosynthesis. Proceedings of the National Academy of Sciences, USA (2000) 97:2379–2384.
Ishihara A, Asada Y, Takahashi Y, Yabe N, Komeda Y, Nishioka T, Miyagawa H, Wakasa K. Metabolic changes in Arabidopsis thaliana expressing the feedback-resistant anthranilate synthase 
-subunit gene OASA1D. Phytochemistry (2006) 67:2349–2362.[CrossRef][Web of Science][Medline]
Keay S, Seillier-Moiseiwitsch F, Zhang C-O, Chai TC, Zhang J. Changes in human bladder epithelial cell gene expression associated with interstitial cystitis or antiproliferative factor treatment. Physiological Genomics (2003) 14:107–115.
Kimura S, Tahira Y, Ishibashi T, Mori Y, Mori T, Hashimoto J, Sakaguchi K. DNA repair in higher plants; photoreactivation is the major DNA repair pathway in non-proliferating cells while excision repair (nucleotide excision repair and base excision repair) is active in proliferating cells. Nucleic Acids Research (2004) 32:2760–2767.
Kristensen C, Morant MM, Olsen CE, Ekstrøm CT, Galbraith DW, Møller LB, Bak S. Metabolic engineering of dhurrin in transgenic Arabidopsis plants with marginal inadvertent effects on the metabolome and transcriptomes. Proceedings of the National Academy of Sciences, USA (2005) 102:1779–1784.
Matsuda F, Miyazawa H, Wakasa K, Miyagawa H. Quantification of indole-3-acetic acid and amino acid conjugates in rice by liquid chromatography-electrospray ionization-tandem mass spectrometry. Bioscience, Biotechnology and Biochemistry (2005a) 69:778–783.[CrossRef][Medline]
Matsuda F, Yamada T, Miyazawa H, Miyagawa H, Wakasa K. Characterization of Trp-overproducing potato transgenic for a mutant rice anthranilate synthase -subunit gene (OASA1D). Planta (2005b) 222:535–545.[CrossRef][Web of Science][Medline]
Mikkelsen MD, Hansen CH, Wittstock U, Halkier BA. Cytochrome P450 CYP79B2 from Arabidopsis catalyses the conversion of tryptophan to indole-3-acetaldoxime, a precursor of indole glucosinolate and indole-3-acetic acid. Journal of Biological Chemistry (2000) 275:33712–33717.
Mitchell SA, Brown KM, Henry MM, Mintz M, Catchpoole D, LaFleur B, Stephan A. Inter-platform comparability of microarrays in acute lymphoblastic leukemia. BMC Genomics (2004) 5:71.[CrossRef][Medline]
Morino K, Matsuda F, Miyazawa H, Sukegawa A, Miyagawa H, Wakasa K. Metabolic profiling of Trp-overproducing rice calli that express a feedback-insensitive alpha subunit of anthranilate synthase. Plant Cell Physiology (2005) 46:514–521.
Murch SJ, KrishnaRaj S, Saxena PK. Tryptophan is a precursor for melatonin and serotonin biosynthesis in in vitro regenerated St John's wort (Hypericum perforatum L. cv. Anthos) plants. Plant Cell Reporter (2000) 19:698–704.[CrossRef]
Niyogi KK, Fink GR. Two anthranilate synthase genes in Arabidopsis: defense-related regulation of the tryptophan pathway. The Plant Cell (1992) 4:721–733.
Quiel JA, Bender J. Glucose conjugation of anthranilate by the Arabidopsis UGT74F2 glucosyltransferase is required for tryptophan mutant blue fluorescence. Journal of Biological Chemistry (2003) 278:6275–6281.
Rabbani MA, Maruyama K, Abe H, Khan MA, Katsura K, Ito Y, Yoshiwara K, Seki M, Shinozaki K, Yamaguchi-Shinozaki K. Monitoring expression profiles of rice genes under cold, drought, and high-salinity stresses and abscisic acid application using cDNA microarray and RNA gel-blot analyses. Plant Physiology (2003) 133:1755–1767.
Rohde A, Morreel K, Ralph J, et al. Molecular phenotyping of the pal1 and pal2 mutants of Arabidopsis thaliana reveals far-reaching consequences on phenylpropanoid, amino acid, and carbohydrate metabolism. The Plant Cell (2004) 16:2749–2771.
Romualdi C, Trevisan S, Celegato B, Costa G, Lanfranchi G. Improved detection of differentially expressed genes in microarray experiments through multiple scanning and image integration. Nucleic Acids Research (2003) 31:e149.
Saeed AI, Sharov V, White J, et al. TM4: a free, open-source system for microarray data management and analysis. Biotechniques (2003) 34:374–378.[Web of Science][Medline]
Singh BK, Lonergan SG, Conn EE. Chorismate mutase isoenzymes from selected plants and their immunological comparison with the isoenzymes from Sorghum bicolor. Plant Physiology (1986) 81:717–722.
Schroder P, Abele C, Gohr P, Stuhlfauth-Roisch U, Grosse W. Latest on enzymology of serotonin biosynthesis in walnut seeds. Advances in Experimental Medicine and Biology (1999) 467:637–644.[Web of Science][Medline]
Su Y-H, Frommer WB, Ludewig U. Molecular and functional characterization of a family of amino acid transporters from Arabidopsis. Plant Physiology (2004) 136:3104–3113.
Tam YY, Epstein E, Normanly J. Characterization of auxin conjugates in Arabidopsis. Low steady-state levels of indole-3-acetyl-aspartate, indole-3-acetyl-glutamate, and indole-3-acetyl-glucose. Plant Physiology (2000) 123:589–596.
Tozawa Y, Hasegawa H, Terakawa T, Wakasa K. Characterization of rice anthranilate synthase -subunit genes OASA1 and OASA2. Trp accumulation in transgenic rice expressing a feedback-insensitive mutant of OASA1. Plant Physiology (2001) 126:1493–1506.
Ueno M, Shibata H, Kihara J, Honda Y, Arase S. Increased tryptophan decarboxylase and monoamine oxidase activities induce Sekiguchi lesion formation in rice infected with Magnaporthe grisea. The Plant Journal (2003) 36:215–228.[CrossRef][Web of Science][Medline]
Wakasa K, Hasegawa H, Nemoto H, et al. High level tryptophan accumulation in seeds of transgenic rice and its limited effects on agronomic traits and seed metabolite profile. Journal of Experimental Botany (2006) 57:3069–3078.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  R. Angelovici, A. Fait, X. Zhu, J. Szymanski, E. Feldmesser, A. R. Fernie, and G. Galili Deciphering Transcriptional and Metabolic Networks Associated with Lysine Metabolism during Arabidopsis Seed Development Plant Physiology, December 1, 2009; 151(4): 2058 - 2072. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. Kang, Y.-S. Kim, S. Park, and K. Back Senescence-Induced Serotonin Biosynthesis and Its Role in Delaying Senescence in Rice Leaves Plant Physiology, July 1, 2009; 150(3): 1380 - 1393. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||