Figures
Abstract
Artemisia annua is the source of antimalarial phytomolecule, artemisinin. It is mainly produced and stored in the glandular secretory trichomes present in the leaves of the plant. Since, the artemisinin biosynthesis steps are yet to be worked out, in this investigation a microarray chip was strategized for the first time to shortlist the differentially expressing genes at a stage of plant producing highest artemisinin compared to the stage with no artemisinin. As the target of this study was to analyze differential gene expression associated with contrasting artemisinin content in planta and a genotype having zero/negligible artemisinin content was unavailable, it was decided to compare different stages of the same genotype with contrasting artemisinin content (seedling - negligible artemisinin, mature leaf - high artemisinin). The SCAR-marked artemisinin-rich (∼1.2%) Indian variety ‘CIM-Arogya’ was used in the present study to determine optimal plant stage and leaf ontogenic level for artemisinin content. A representative EST dataset from leaf trichome at the stage of maximal artemisinin biosynthesis was established. The high utility small scale custom microarray chip of A. annua containing all the significant artemisinin biosynthesis-related genes, the established EST dataset, gene sequences isolated in-house and strategically selected candidates from the A. annua Unigene database (NCBI) was employed to compare the gene expression profiles of two stages. The expression data was validated through semiquantitative and quantitative RT-PCR followed by putative annotations through bioinformatics-based approaches. Many candidates having probable role in artemisinin metabolism were identified and described with scope for further functional characterization.
Citation: Nair P, Misra A, Singh A, Shukla AK, Gupta MM, Gupta AK, et al. (2013) Differentially Expressed Genes during Contrasting Growth Stages of Artemisia annua for Artemisinin Content. PLoS ONE 8(4): e60375. https://doi.org/10.1371/journal.pone.0060375
Editor: Georges Snounou, Université Pierre et Marie Curie, France
Received: December 28, 2012; Accepted: February 26, 2013; Published: April 3, 2013
Copyright: © 2013 Nair et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Department of Biotechnology (DBT), India, and the Council of Scientific and Industrial Research (CSIR), India, financially supported this study through grants GAP138 (BT/PR6544/PBD/17/405/2005) and NWP08 (XI Five Year Plan Network Project), respectively. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors declare that the affiliation of Dr Suman P. S. Khanuja, one of the authors of the company Nutrahelix, does not alter their adherence to all the PLOS ONE policies on sharing data and materials.
Introduction
Artemisia annua (Asteraceae) is the only source for important antimalarial phytomolecule “artemisinin”. This molecule, an endoperoxide sesquiterpene lactone [1] is active against multi-drug-resistant strains of the malarial parasite. In spite of the report on development of resistance against artemisinin [2], the phytomolecule remains a potent weapon in the arsenal against malaria. Since, it is difficult to synthesize artemisinin chemically and its production in cell/tissue culture is very low, the plant remains the only major source for the drug [3]. Artemisinin is detected in aerial parts of the plant, mainly in leaves, stem and inflorescence and little or none in roots or pollen, with the content varying from 0.01–1.2% in different genotypes of A. annua [4], [5], [6]. Low supply of artemisinin against high demand has provided major impetus for research on identification of genes and regulatory factors towards enhancing the in planta and/or heterologous production of the phytomolecule.
In planta, artemisinin is biosynthesized from isopentenyl pyrophosphate (IPP) via farnesyl pyrophosphate (FPP). FPP is converted to amorpha-4,11-diene by the action of amorpha-4,11-diene synthase (ads), the step commiting the metabolic flux towards artemisinin. The next reaction is a three step oxidation of amorpha-4,11-diene to artemisinic acid via formation of artemisinic alcohol and artemisinic aldehyde by a cytochrome P450 monooxygenase (cyp71av1) [7]. Recently two more genes, a double bond reductase (dbr2) [8] and an aldehyde dehydrogenase (aldh1) [9] presumably involved in artemisinin biosynthesis pathway, were also cloned and indicated in the conversion of artemisinic aldehyde to its dihydro form and then to dihydroartemisinic acid (DHAA), respectively. It is believed that the two bifurcated pathways generating either arteannuin B or artemisinin as end products compete with each other for precursors [10]. Wallaart et al. [11], [12] suggested DHAA as a reactive oxygen species (ROS) scavenger and is probably converted to artemisinin by more than one non-enzymatic, spontaneous photo-oxidation reactions [13]. Besides, there may be other unknown diversions from FPP affecting artemisinin biosynthesis. Also, most of the regulatory factors involved in artemisinin biosynthesis are yet to be elucidated. This necessitates searching for novel structural and regulatory genes involved in artemisinin biosynthesis in the plant. Based on this, ‘CIM-Arogya’, a SCAR-marked high yielding (∼1.2% artemisinin content on dry weight basis) variety of A. annua [5], [14], [15], similar in artemisinin content to the European variety ‘Artemis’, was analyzed in the present study. This variety is well accepted by the industry and grown extensively by farmers. Recently, the genetic map of A. annua has also been constructed and used to identify loci affecting yield of artemisinin [16]. EST-based [17] and deep transcriptome sequencing [18] approaches have been attempted in the plant earlier for identifying genes. These efforts will definitely facilitate new gene discovery and shed light on the regulatory mechanism of artemisinin metabolism and trichome function in A. annua. Still, there is a need to have a representative EST dataset from the trichome of elite Indian A. annua varieties like ‘CIM-Arogya’ to enrich the existing information helping to identify candidates for gene bioprospecting and downstream expression analysis.
Artemisinin content in the plant in various climates and regions has been studied in the past, mainly in China [19]. In the present study, the critical developmental/seasonal stages for the biosynthesis/accumulation of artemisinin were defined and a representative EST dataset from leaf trichome at the stage of maximal artemisinin biosynthesis was established. A high utility small scale (750 target genes) custom microarray chip of A. annua containing all the significant artemisinin biosynthesis-related genes was designed using the in-house established EST dataset and strategically selected candidates from the A. annua Unigene database (NCBI). The custom array was employed to compare the gene expression profiles of two plant developmental stages having contrasting metabolite (artemisinin) levels. The microarray data was validated through quantitative RT-PCR and differentially expressing candidates were identified in both the stages providing leads for further functional analysis in planta. This is the first report on microarray analysis for a limited selected genes in A. annua.
Materials and Methods
Plant Material and Growth Conditions
The seeds of A. annua (cv. CIM-Arogya) were obtained from the National Gene Bank for Medicinal and Aromatic Plants, CSIR-CIMAP, Lucknow and sown in earthen pots (20 cm high and 20 cm internal diameter) containing a mixture of soil and farmyard manure in the ratio of 1∶1. The plants were grown in glass house under standard conditions of light, temperature and humidity. For the field experiments (for chemo-profiling) the nursery (seedlings having 10 cm height) were transplanted in the field with spacing 50 cm between rows and 30 cm between plants and grown using standard agronomic procedures decribed earlier [14]. For the array-based experiment, the two contrasting stages were selected (six-day-old seedling and six-month-old mature plant).
Chemo-profiling for Artemisinin
For in planta artemisinin content analysis, 0.1 g shade-dried leaves/seedlings were ground, boiled in hexane, filtered and evaporated. The residue was used for artemisinin analysis through HPTLC following the protocol described by Misra et al. [20]. All samplings for the artemisinin content analysis were carried out in triplicate for analysis.
Construction of Leaf Glandular Trichome (GT) cDNA Library and EST Sequencing
GTs were isolated from 500 g leaves of a mature six-month old plant (at the stage of maximal artemisinin biosynthesis) using the protocol described by Teoh et al. [7]. For cDNA library construction, total RNA was isolated from the trichome-enriched leaf tissue according to Chomczynski and Sacchi [21] and poly (A)+ mRNA was isolated following the protocol of Shukla et al. [22]. For the microarray experiment, total RNA was isolated from seedling and mature plant leaf using RNeasy Mini Kit (Qiagen) according to manufacturer’s guidelines. RNA concentration was quantified by measuring absorbance at 260 and 280 nm by NanoDrop (NanoDrop, USA) and RNA quality was evaluated in Agilent 2100 Bioanalyzer (Agilent Technologies Inc., Palo Alto, CA). Total RNA meeting the quality standards was released for probe generation. cDNA library was prepared by using the ZAP Express® cDNA Synthesis Kit and packaged by Gigapack® III gold cloning kit (Stratagene, USA). The primary library was amplified to a titer of 1.6 x 107 pfu/ml in E. coli strain XL1-Blue MRF’. This was mass excised using ExAssist™ helper phage and E. coli strain XL1 Blue MRF’ (at a MOI of 1∶10 lambda phage-to-cell ratio and 10∶1 helper phage-to-cells ratio) to obtain recombinant cDNA clones in pBK-CMV phagemid. The excised phagemids were titered using E. coli strain XLOLR and the recombinants were selected by blue/white screening on X-gal/IPTG-coated plates containing kanamycin (50 µg/ml). The clones from the amplified library (over 90% insert frequency) were used for end sequencing to generate a small in-house leaf trichome EST dataset. The ESTs were analyzed as described earlier [23].
A. annua Custom Array Design and Preparation
A small scale microarray chip (with 750 target genes) was designed for A. annua. The sequences for A. annua were taken from three sources - A. annua Unigene (NCBI) unique sequence library (having 9462 entries), 385 CIMAP in-house A. annua (CIM-Arogya) EST sequences (Accession Numbers GT735932-GT736316) and an additional 22 in house cytochrome P450 (cyp) gene sequences (Table S1) [20]. The A. annua Unigene was searched with keywords – decarboxylase, kinase, reductase, reductoisomerase, regulator, synthase, synthetase, transcription, transcription factor, transferase, transporter, etc and a few candidates belonging to categories like cytochromes, transcription factors, transporters, reductases, synthases, dehydrogenases, peroxidases, isomerases and known genes from A. annua were shortlisted based on their putative involvement in artemisinin biosynthesis/metabolism/diversion pathways/regulation. From each Unigene cluster, a single representative EST-clone sequence provided for cluster by NCBI was included. This sequence is the longest and most homologous strand present within the cluster. This set of shortlisted candidates was then analyzed for redundancy with the in house sequences in the subsequent steps.
Clustering was carried out for the 385 in house ESTs using CAP3 to remove the redundant sequences, which resulted into 174 unigenes (59 contigs +115 singletons). Further the ESTs and their CAP3 cluster contigs, were annotated by using NCBI’s Local-BLAST’s tblastx functionality. This function compared the given sequences with the ‘Arabidopsis’ RefSeq database and a database of RefSeq sequences of other ‘Flowering Plants’ to indicate any putative homology found to annotated plant organism’s mRNA reference sequences. The program output files were analyzed using TblastParse (an internal report analysis program of Ocimum Biosolutions, India) and significant annotation retrieved and reported for each EST. The Unigene selection set was compared to in house sequences using NCBI’s Local-BLAST program to eliminate redundancy occurring between the two datasets. Four singletons (GT735961, GT736012, GT736179, and GT736228) and 4 cyp gene sequences [(CIM-Arog_CYP01 (KC594703), CIM-Arog_CYP04 (JN594505), CIM-Arog_CYP06 (GU318228), CIM-Arog_CYP08 (KC594704)] of in-house sequences were found to be already represented in the NCBI Unigene selections. CIM-Arog_CYP05 (JN594506) could not be spotted on the array. It was also ensured that inclusion of any sequence showing homology to ribosomal genes and targets for which oligos could not be designed were avoided. Finally, a total of 157 in house unigenes (55 contigs +102 singletons) and 17 cyp genes [20] were represented on the array apart from the other selections from the publicly available Unigenes to make a target number of 750 genes. After checking the cross hybridization of the designed 50 mer oligos, the custom A. annua array was prepared by Ocimum Biosolutions, Hyderabad, India and its platform data has been submitted to Gene Expression Omnibus (GEO, NCBI) under accession number GPL15698.
Probe Generation, Hybridization and Analysis
A dual channel procedure with dye-swap arrangement was adopted in the study of the two samples - seedling and mature leaf, to compare the expression levels of target genes in seedling against mature leaf. The expression data was generated on 1,569 probes, with two replicates. Five microgram of total RNA was used for amplification with the help of Amino Allyl MessageAmp™ II aRNA Amplification Kit from Ambion by linear transcription based RNA amplification system to produce cRNA. Briefly, mRNA was reverse transcribed with an oligo (dT) primer bearing T7 promoter at 42°C for 2 h and second strand cDNA synthesis was carried out at 16°C for 2 h. The resulting cDNA was purified and transcribed with T7 RNA polymerase to generate multiple copies of aminoallyl antisense RNA (aRNA) at 37°C for 16 h. aRNA was then labelled with Cy3TM/Cy5TM post-labelling reactive dye pack (GE Healthcare, UK) at room temperature and unincorporated Cy3/Cy5 molecules were removed by purification process using QIAGEN PCR purification kit before hybridization. Ten microgram of the labeled aRNA in 75 µl of Ocimum’s Hyb buffer was used for hybridization with the A. annua custom array chip. Hybridized chips were scanned using Affymetrix 428TM Array Scanner at three different PMT gains (40, 50 and 60) and the data was analyzed using Genowiz software (Ocimum Biosolutions, Hyderabad). Image analysis was carried out using Imagene, version 5.6.1. The signal values obtained at the three PMT settings were averaged, to get the signal mean for further analysis. Replicate genes were also averaged before normalization. Signal values obtained from each channel were log2 transformed and normalized using LOWESS algorithm and median absolute deviation (MAD) scaling. For each sample, Cy3 and Cy5 intensities were averaged, and used to compare the samples. During normalization, paired slide dye-swap method was followed to overcome the dye-bias during the comparison and then MAD was performed to adjust data into same scale [24]. The normalized (adjusted) data was subsequently used in differential expression (DE) analysis, which was performed using fold change technique. The data from this microarray experiment has been submitted to GEO (NCBI) under series accession number GSE39098 [associated sample data GSM956130 (leaf vs seedling) and GSM956131 (seedling vs leaf)]. The ‘OBSca028_’ prefix in the gene/probe IDs in the submitted data has been abbreviated here as ‘Aa’ for convenience.
Semiquantitative and Quantitative RT-PCR Analysis for Validation
The gene expression profiles obtained in the microarray analysis were validated through semiquantitative and quantitative RT-PCR. For semiquantitative RT-PCR, the total RNA was isolated from the six-day-old seedling and six-month-old mature plant leaf of A. annua using TRIzol® reagent (Invitrogen, USA). The quality and quantity of total RNA was assessed through ethidium bromide staining as well as Nanodrop ND1000 spectrophotometer. Equal amount (4 µg) of DNaseI-treated total RNA was used for first strand cDNA synthesis by Thermoscript RT-PCR System (Invitrogen). A. annua actin (EU531837) was used as a control for the gel-based semiquantitative analysis of gene expression. The primer sequences used for the semi-quantitative RT-PCR-based validation of selected target genes were designed using Gene Runner version 3.05 (Hastings Software, Inc.) and are listed in Table S2. The gene expression profiles of known genes of A. annua, mainly belonging to the artemisinin (and other sesquiterpene) biosynthetic pathway were validated for trichomes isolated from leaf through TaqMan chemistry-based Real Time PCR (Table S3). The level of gene expression was analyzed following the protocol described by Misra et al. [20] and finally log10 RQ values were calculated and represented.
Results
Ontogeny and Plant Age-related Variation in Artemisinin Content
Experiment was carried out to study the ontogenic variation of artemisinin content in mature (6-month-old) field grown A. annua (Figure 1). The aerial portion of the plant was demarcated into upper (top about 30 cm), middle (about 30 cm), lower (between middle and leafless region, about 30 cm) and leafless (upto 45 cm height from the ground) regions. Leaf samples were collected from different ontogeny levels of primary and secondary branches as indicated in Figure 1. Artemisinin content was always found to be optimum in the young leaves at upper levels of secondary branches. The content of artemisinin in the stem, seed and seed husk was found to be 1/10th, 1/35th and 1/3rd respectively as compared to the leaves, whereas it was undetected in the roots of the plant (data not shown). Another related experiment was carried out to study the developmental variation of leaf artemisinin content in the plants under north Indian conditions (Figure 2). Starting from February (6-day-old seedling stage), monthly sampling of leaves was carried out till September. Leaf artemisinin content was found to increase from a value “undetected” at the six-day seedling stage, reaching the maximum at the pre-flowering stage (6-month-old plant/August) and declining thereafter.
Sampling was carried out in triplicate, and standard deviation ranged between 1.5–12.3% of mean values.
Y axis represents % artemisinin content per g dry weight of young leaves at upper levels of secondary branches. Sampling was carried out in triplicate.
Representative ESTs from Metabolically Active Leaf GTs
After defining the optimal plant stage and ontogenic level for highest leaf artemisinin content through chemo-profiling, a leaf trichome cDNA library was constructed. Around 500 cDNA clones were sequenced generating 458 ESTs and were submitted to dbEST (under accession numbers: GT735901–GT735931, GT735932–GT736316, GT736317–GT736358). More ESTs were not generated as ∼ 85000 ESTs (as well as unigenes) of A. annua became available in the NCBI public database. Hence, strategic selection of candidates [(genes/ESTs (including unigenes)] from the NCBI database and in house generated dataset was opted for limited array-based gene expression profiling in plant developmental stages contrasting for their artemisinin content.
Comparative Gene Expression Analysis
Selected 385 ESTs (GT735932–GT736316) from the in house generated dataset, shortlisted “Unigene” candidates from NCBI and in house generated cyp genes were taken up for custom array-based comparative gene expression analysis in mRNA populations derived from two contrasting plant developmental stages (6-day-old seedling - negligible artemisinin, mature leaf from 6-month-old plants - high artemisinin). For the comparative analysis, genes with log fold change value≥0.5849 (FC≥1.5) were assumed as up-regulated while genes with log fold change ≤−0.5849 (FC≤0.66) were assumed as down-regulated in seedling, compared to mature leaf sample taken as the control. Based on this criterion, of 750 target genes on the chip, 158 genes were found to be down-regulated (Table S4), whereas 73 genes were up-regulated in seedling (Table S5) compared to mature leaf. The heat maps for the differentially expressing genes is provided in Figures S1 (upregulated in seedling) and S2 (downregulated in seedling). However, while selecting the genes for downstream analysis and usage, a more stringent threshold was used, where genes with log fold change value≥1 (FC≥2) were considered to be up-regulated and those with log fold change value≤−1 (FC≤0.50) were considered to be down-regulated (with a few exceptions like the known genes/ESTs of A. annua). On the basis of this stringent criterion, 98 genes qualified for the “down-regulated in seedling” category and 27 for the “up-regulated in seedling” category (Table 1). The results obtained in the microarray experiment were verified through semiquantitative RT-PCR analysis for a representative set of genes (Figure 3) and through TaqMan chemistry-based Real Time PCR for the known pathway genes (Figure 4).
Expression of all the genes in 6-day-old seedling (calibrator) was equilibrated to “0”. White bars represent expression in the leaf of 3-month-old plant and solid bars in 6-month-old plant. Y-axis represents Log10 RQ, where relative quantity (RQ) equilibrating the expression in seedling is 1RQ value. Data represent mean±standard error of at least 3 biological replicates. The genes studied are 1. Farnesyl pyrophosphate synthase (Aa398); 2. Cytochrome P450 reductase (CPR) (Aa15); 3. Amorpha-4,11-diene synthase (Aa408); 4. P450 monooxygenase (CYP71) (Aa561); 5. Linalool synthase (Aa442); 6. Germacrene-A synthase (Aa434); 7. Squalene synthase (Aa407); 8. β–Caryophyllene synthase (Aa417); 9. Sesquiterpene cyclase (Aa560); 10. 1-Deoxy-D-xylulose-5-phosphate reductoisomerase (DXR) (Aa554); 11. HMG-CoA reductase (Aa689).
Discussion
Optimal Leaf Ontogeny and Plant Age for Maximal in Planta Artemisinin Biosynthesis
Artemisinin is biosynthesized and stored in glandular trichomes (GT) of flowers and on both the surfaces of leaves [25], [26], [27]. In the genus Artemisia, the differentiation of foliar cells into GT cells is completed in a very young primordial stage of the leaf [25]. According to Lommen et al. [28], GT densities are highest at the young leaf stage and decreases after attaining maximal size. However, the number remains more or less constant till the maximal size, after which GT number decreases rapidly suggesting the rupture of GT over time in the older leaves. Similar decrease in the GT number in Mentha arvensis from upper expanding young, to the lower level leaves proceeding for senescence is reported [29]. Also, artemisinin concentrations are reported to be higher in upper leaves compared to lower in a branch [30], [31]. In the present study artemisinin content was found to be maximal in the top level leaves of the branches (Figure1). The leaf artemisinin content was found to increase from an undetectable value at the six-day seedling stage, reaching the maximum at the pre-flowering stage (6-month-old plant/August) and declining thereafter. This was also in consonance to the results obtained by Gupta et al. [31], Liersch et al. [32] and Zhang et al. [19]. These results indicated the stages of active GTs in the leaves biosynthesizing high artemisinin and leaf samples were collected for EST and hybridization analysis at this stage.
Differential Gene Expression
Although a subtractive hybridization-based approach has been followed in A. annua earlier to compare blooming flowers and flower buds [33], the present study was more elaborate and comprehensively planned. The target of the present study was to analyze differential gene expression associated with contrasting artemisinin content in planta. Since an A. annua genotype having zero/negligible artemisinin content was unavailable, it was decided to compare different stages of the same genotype (CIM-Arogya) having contrasting artemisinin content. So, the seedling stage (negligible artemisinin) and mature plant leaf (high artemisinin) derived mRNA populations were used to analyze the custom A. annua array. Though, some noise in the data due to other variations in the samples (like developmental age related differences) cannot be ruled out, this was the best option available to compare high and low artemisinin content related gene expression profiles in the plant.
Thirty six candidates showing significant fold change in expression in the two contrasting plant stages were identified for further gene prospecting (Table S2). Figure 3 depicts representative genes that were taken up for validation through semi-quantitative RT-PCR. The secondary metabolism related genes were specifically validated for the trichome expression (Figure 4) as these are reported to be expressing in GTs present on both surfaces of the leaf. The increasing trend in expression was similar for all the genes as the plant matured from 6-day-old seedling to 6-month stage. The selection of the genes for microarray was also based on the relevance of their annotated function to the objective of association with artemisinin metabolism. Many genes found to be directly or indirectly influencing terpenoid (specially artemisinin) biosynthesis were found to be differentially expressed in the seedling as compared to the mature leaf and could be mapped on the pathway (Figure 5).
Gene IDs are represented in parentheses. ‘+’ indicates up-regulated in mature leaf and ‘−’ indicates down-regulated in mature leaf as compared to seedling. The mapped genes from this study are highlighted. ADS: amorpha-4,11-diene synthase; ALDH1: aldehyde dehydrogenase 1; AMO/CYP71AV1: amorpha-diene oxidase; AAOH: artemisinic alcohol; AAA: artemisinic aldehyde; AA: artemisinic acid; BAS: β-amyrin synthase; BFS: β-farnesene synthase; CPR: cytochrome P450 reductase; BCS: β-caryophyllene synthase; CYP71AV1: amorphadiene-12-hydroxylase; DBR2: artemisinic aldehyde Δ11(13) reductase; DHAAOH: dihydroartemisinic alcohol; DHAAA: dihydroartemisinic aldehyde; DHAA: dihydroartemisinic acid; ECS: epi-cedrol synthase; DMAPP: dimethylallyl pyrophosphate; FPPS: farnesyl diphosphate synthase; FPP: farnesyl pyrophosphate; GAS: germacrene-A synthase; GPP: geranyl diphosphate; HMGR: 3-hydroxy-3-methyl-glutaryl coenzyme A reductase; HMGS: 3-hydroxy-3-methyl-glutaryl coenzyme A synthase; IPP: isopentenyl pyrophosphate; IDS: IPP/DMAPP synthase; MVA: mevalonic acid; SAH: steroid 23-alpha-hydroxylase; SQC: sesquiterpene cyclase; SQS: squalene synthase; BPS: β-pinene synthase; CDP-MEP: 4-cytidine 5′-diphospho-2-C-methyl-D-erythritol 2-phosphate; cMEPP: 2-C-methyl-D-erythritol 2,4-cyclophosphate; DXP: 1-deoxy-D-xylulose 5-phosphate; DXR: 1-deoxy-D-xylulose-5-phosphate reductoisomerase; DXS: 1-deoxy-D-xylulose-5-phosphate synthase; GPP: geranyl diphosphate; GDS: geranyl diphosphate synthase; HDR: hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate reductase; HDS: hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate synthase; HMBPP: 1-hydroxy-2-methyl-2-(E)-butenyl 4-diphosphate; LS: Linalool synthase; MECS: 2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase; MEP: 2-C-methyl-D-erythritol 4-phosphate.
Expectedly, most of the known artemisinin (as well as other sesquiterpene) metabolism-related genes were found to be having higher expression at transcript level in the mature leaf as compared to the seedling stage. These included candidates like cytochrome P450 reductase (Aa15), squalene synthase (Aa407), artemisinin aldehyde delta-11(13) reductase (Aa316), isopentenyl pyrophosphate/dimethylallyl pyrophosphate synthase (Aa409), β-amyrin synthase (Aa401), amorpha-4,11-diene synthase (Aa408), epi-cedrol synthase (Aa413), β-caryophyllene synthase (Aa417), (3R)-linalool synthase (Aa442, Aa443), β-pinene synthase (Aa456), amorpha-4,11-diene C-12 oxidase (Aa557, Aa561), steroid 23-α-hydroxylase (Aa21), taxadiene 5-α-hydroxylase (Aa25), etc. Interestingly, some genes like sesquiterpene cyclase were found to be present in both categories – up-regulated in seedling (Aa555) and down-regulated in seedling (Aa417), which may be due to presence of different isoforms. In addition, the transcript abundance of HMGR (3-hydroxy-3-methyl-glutaryl-CoA reductase) (Aa689) was observed to be lower in the seedlings (fold change 0.5), whereas the transcript abundance of DXR (1-deoxy-D-xylulose 5-phosphate reductoisomerase) (Aa554) was higher in the seedlings compared to the mature leaf (fold change 2.55). This possibly indicates that the plastidial non-mevalonate terpenoid pathway contributing the major metabolic flux for isopentenyl pyrophosphate (IPP) biosynthesis at the seedling stage, whereas the cytosolic mevalonate pathway is the more prominent flux contributor in the leaf at mature stage. This is consistent with the fact that the artemisinin content of the mature leaf in A. annua is much higher than that in the seedling, and artemisinin being a sesquiterpene (C15) is produced in the cytosolic compartment of the plant cell with major metabolic flux for the intermediate IPP coming from the mevalonate pathway. However, this hypothesis needs further validation as interestingly, 1-deoxy-D-xylulose-5-phosphate synthase (DXS, Aa459) and 2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase (MECS; Aa430) of the non-mevalonate pathway showed higher transcript abundance in the mature leaf whereas an HMGR isoform (Aa308) showed higher transcript abundance in seedling. This may be due to differential expression of isoforms in different tissue. But, as reported earlier the non-mevalonate pathway predominates over the mevalonate pathway in the A. annua GTs [34] and is another reason to analyze the function of these isoforms in detail.
Among the selections, a few interesting candidates were observed with the potential for functional characterization. For example, a few genes (like Aa745) showing no significant similarity to database sequences could provide further clues related to artemisinin biosynthesis and/or accumulation in the plant. Among the in-house cyp genes on the array, only Aa590 was found to have higher transcript abundance in the seedling as compared to the mature leaf. Another interesting class of differentially expressing genes belonged to the category – peroxidases. It has been shown earlier in Arabidopsis that from seedling to mature plant, in cotyledons or leaves of different ages, plastidial gene expression is regulated at the transcriptional and post-transcriptional levels, but not by plastome copy number [35]. This emphasises the role of transcriptional regulators in the transition from seedling to mature leaf in a plant and must be true for non-plastidial genes too. This view was reinforced by the identification of several differentially expressing transcription factors in the A. annua mature leaf vis-à-vis seedling. For example, nuclear transcription factor Y subunit C-1 (Aa94), AP2/ERF and B3 domain-containing transcription factor RAV1(Aa127), transcription factor AS1 (Aa185), ethylene-responsive transcription factor ERF025 (Aa216), etc were found to be having higher expression in seedling compared to mature leaf. On the other hand, multi-protein bridging factor 1A (Aa91), homeobox protein knotted-1-like 7 (Aa111), NAC domain containing protein 83 (Aa138, Aa244), DELLA protein RGA (Aa141), BEL1-like homeodomain 1 (Aa143), auxin-responsive protein IAA9 (Aa163), transcription factor ILR3 (Aa167), RNA polymerase sigma factor (Aa171), transcription factor TCP7 (Aa206), basic leucine-zipper 44 (Aa226), protein agamous-like 42 (Aa233), AP2 domain transcriptional regulator (Aa241), nuclear factor Y (Aa247), scarecrow-like protein 30 (Aa266), homeodomain-like transcriptional regulator (Aa271), etc were found to show higher transcript abundance in mature leaf as compared to seedling.
As there is a possible linkage between flowering and artemisinin biosynthesis [36], the genes associated with flowering with higher transcript abundance in mature leaf assume significance. BELL (BEL1) is known to control ovule development through negative regulation of AGAMOUS gene (AG) in Arabidopsis [37] and the DELLA protein is a target for gibberellin signalling [38]. Gibberellins are known to be enhancing flowering in plants. As the artemisinin content in the leaf peaks just before flowering, it is easy to link the flowering regulators with the role to enhance artemisinin content directly or indirectly. But, higher expression of the early flowering gene CONSTANS in mature leaf compared to seedling, did not induce an increase in artemisinin biosynthesis [39]. Taken together, these results suggest that the observed increase of artemisinin content at pre-flowering stage may not be a direct consequence of flowering itself, but may be due to the combined influence of factors preparing the plants to proceed for flowering stage.
Conclusion
Selected genes from the EST database and in-house generated ESTs from trichomes were analyzed for differential expression at contrasting stages. This is the first report on microarray analysis relating the expression to differential stages of artemisinin biosynthesis. More number of genes were found to show higher transcript abundance in the leaf as compared to seedling indicating a higher level of the metabolic complexity of the mature leaf vis-à-vis the seedling. The result was validated by quantitative and semiquantitative RT-PCR for known genes and found to be confirming the expression of metabolites and result of microarray experiment. Several transcriptional regulators were indicated for their higher expression in the mature leaf and their functional characterization will provide further insight into artemisinin biosynthesis. Higher expression of genes associated with flowering in mature leaf indicates the preparation for flowering, which may be also indirectly influencing artemisinin biosynthesis.
Supporting Information
Figure S1.
Heat map for the differentially expressing genes upregulated in seedling.
https://doi.org/10.1371/journal.pone.0060375.s001
(TIF)
Figure S2.
Heat map for the differentially expressing genes downregulated in seedling.
https://doi.org/10.1371/journal.pone.0060375.s002
(TIF)
Table S1.
Details of in-house cytochrome P450 sequences of A. annua .
https://doi.org/10.1371/journal.pone.0060375.s003
(DOC)
Table S2.
Sequences of the primers used for semi-quantitative RT-PCR-based expression analysis of selected target genes of A. annua .
https://doi.org/10.1371/journal.pone.0060375.s004
(DOC)
Table S3.
Sequence details of Assay-By-Designs for TaqMan chemistry-based Real Time PCR of A. annua sesquiterpene biosynthetic pathway genes.
https://doi.org/10.1371/journal.pone.0060375.s005
(DOC)
Table S4.
Genes found to be down-regulated in seedling as compared to mature plant leaf of A. annua .
https://doi.org/10.1371/journal.pone.0060375.s006
(DOC)
Table S5.
Genes found to be up-regulated in seedling as compared to mature plant leaf in A. annua .
https://doi.org/10.1371/journal.pone.0060375.s007
(DOC)
Acknowledgments
Help from Director, CSIR-CIMAP for the facility is acknowledged. PN and AM thank CSIR for their Senior Research Fellowships.
Author Contributions
Conceived and designed the experiments: AK Shasany SPSK VG AK Shukla. Performed the experiments: PN AM AS MMG AKG. Analyzed the data: AK Shukla VG AK Shasany. Contributed reagents/materials/analysis tools: AK Shukla VG MMG. Wrote the paper: AK Shasany AK Shukla.
References
- 1. Hsu E (2006) Reflections on the ‘discovery’ of the antimalarial qinghao. Brit J Clin Pharmaco 61: 666–670.
- 2. Dondorp AM, Fairhurst RM, Slutsker L, Macarthur JR, Breman JG, et al. (2011) The threat of artemisinin-resistant malaria. New England J Med 365: 1073–1075.
- 3. Delabays N, Simonnet X, Gaudin M (2001) The genetics of artemisinin content in Artemisia annua L. and the breeding of high yielding cultivars. Curr Med Chem 8: 1795–1801.
- 4. Jain DC, Mathur AK, Gupta MM, Singh AK, Verma RK, et al. (1996) Isolation of high artemisinin-yielding clones of Artemisia annua. Phytochem 43: 993–1001.
- 5. Khanuja SPS, Paul S, Shasany AK, Gupta AK, Darokar MP, et al. (2005) Genetically tagged improved variety ‘CIM-Arogya’ of Artemisia annua for high artemisinin yield. Journal of Medicinal and Aromatic Plant Sciences 27: 520–524.
- 6. Liu C, Zhao Y, Wang Y (2006) Artemisinin: current state and perspectives for biotechnological production of an antimalarial drug. Appl Microbiol Biotechnol 72: 11–20.
- 7. Teoh KH, Polichuk DR, Reed DW, Nowak G, Covello PS (2006) Artemisia annua L. (Asteraceae) trichome-specific cDNAs reveal CYP71AV1, a cytochrome P450 with a key role in the biosynthesis of the antimalarial sesquiterpene lactone artemisinin. FEBS Lett 580: 1411–1416.
- 8. Zhang Y, Teoh KH, Reed DW, Maes L, Goossens A, et al. (2008) The molecular cloning of artemisinic aldehyde D11(13) reductase and its role in glandular trichome-dependent biosynthesis of artemisinin in Artemisia annua. J Biol Chem 283: 21501–21508.
- 9. Teoh KH, Polichuk DR, Reed DW, Covello PS (2009) Molecular cloning of an aldehyde dehydrogenase implicated in artemisinin biosynthesis in Artemisia annua. Botany 87: 635–642.
- 10. Liu B, Wang H, Du Z, Li G, Ye H (2011) Metabolic engineering of artemisinin biosynthesis in Artemisia annua L. Plant Cell Rep. 30: 689–694.
- 11. Wallaart TE, van Uden W, Lubberink HGM, Woerdenbag HJ, Pras N, et al. (1999) Isolation and identification of dihydroartemisinic acid from Artemisia annua and its possible role in the biosynthesis of artemisinin. J Nat Prod 62: 430–433.
- 12. Wallaart TE, Bouwmeester HJ, Hille J, Poppinga L, Maijers NC (2001) Amorpha-4,11-diene synthase: cloning and functional expression of a key enzyme in the biosynthetic pathway of the novel antimalarial drug artemisinin. Planta 212: 460–465.
- 13. Brown GD, Sy LK (2007) In vivo transformations of dihydro-epi-deoxyarteannuin B in Artemisia annua plants. Tetrahedron 63: 9536–9547.
- 14.
Khanuja SPS, Paul S, Shasany AK, Gupta AK, Darokar MP, et al. (2008) High artemisinin yielding Artemisia plant named ‘CIM-Arogya’. US Patent No. 7,375,260 (20 May, 2008).
- 15. Paul S, Khanuja SPS, Shasany AK, Gupta MM, Darokar MP, et al. (2010) Enhancement of artemisinin content through four cycles of recurrent selection with relation to heritability, correlation and molecular marker in Artemisia annua L. Planta Med. 76: 1468–72.
- 16. Graham IA, Besser K, Blumer S, Branigan CA, Czechowski T, et al. (2010) The genetic map of Artemisia annua L. identifies loci affecting yield of the antimalarial drug artemisinin. Science 327: 328–331.
- 17. Zeng Q, Zhao C, Yin L, Yang R, Zeng X, et al. (2008) Cloning of artemisinin biosynthetic cDNAs and novel ESTs and quantification of low temperature-induced gene overexpression. Science in China Series C: Life Sciences 51: 232–244.
- 18. Wang W, Wang Y, Zhang Q, Qi Y, Guo D (2009) Global characterisation of Artemisia annua glandular trichome transcriptome using 454 pyrosequencing. BMC Genomics 10: 465.
- 19. Zhang L, Ye HC, Li GF (2006) Effect of development stage on the artemisinin content and the sequence characterized amplified region (SCAR) marker of high-artemisinin yielding strains of Artemisia annua L. J Integr Plant Biol. 48: 1054–1062.
- 20. Misra A, Chanotiya CS, Gupta MM, Dwivedi UN, Shasany AK (2012) Characterization of cytochrome P450 monooxygenases isolated from trichome enriched fraction of Artemisia annua L. leaf. Gene 510: 193–201.
- 21. Chomczynski P, Sacchi N (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal Biochem 162: 156–159.
- 22. Shukla AK, Shasany AK, Khanuja SPS (2005) Isolation of poly (A)+ mRNA for downstream reactions from some medicinal and aromatic plants. Indian J Exp Biol 43: 197–201.
- 23. Shukla AK, Shasany AK, Gupta MM, Khanuja SPS (2006) Transcriptome analysis in Catharanthus roseus leaves and roots for comparative terpenoid indole alkaloid profiles. J Exp Bot 57: 3921–3932.
- 24. Yang YH, Dudoit S, Luu P, Lin DM, Peng V, et al. (2002) Normalization for cDNA microarray data: a robust composite method addressing single and multiple slide systematic variation. Nucleic Acids Res 30(4): e15.
- 25. Duke SO, Paul RN (1993) Development and fine structure of glandular trichomes of Artemisia annua L. Int J Plant Sci. 154: 107–118.
- 26. Duke MV, Paul RN, Elsohly HN, Sturtz G, Duke SO (1994) Localization of artemisinin and artemisitene in foliar tissues of glanded and glandless biotypes of Artemisia annua L. Int J Plant Sci. 155: 365–372.
- 27. Olsson ME, Olofsson LM, Lindahl AL, Lundgren A, Brodelius M, et al. (2009) Localization of enzymes of artemisinin biosynthesis to the apical cells of glandular secretory trichomes of Artemisia annua L. Phytochem. 70: 1123–1128.
- 28. Lommen WJM, Schenk E, Bouwmeester HJ, Verstappen FWA (2006) Trichome dynamics and artemisinin accumulation during development and senescence of Artemisia annua leaves. Planta Med 72: 336–345.
- 29. Shanker S, Ajayakumar PV, Sangwan NS, Kumar S, Sangwan RS (1999) Essential oil gland number and ultrastructure during Mentha arvensis leaf ontogeny. Biol Plantarum 42: 379–387.
- 30. Laughlin JC (1993) Effect of agronomic practices on plant yield and antimalarial constituents of Artemisia annua L. Acta Hortic. 331: 53–62.
- 31. Gupta SK, Singh P, Bajpai P, Ram G, Singh D, et al. (2002) Morphogenetic variation for artemisinin and volatile oil in Artemisia annua. Ind Crop Prod 16: 217–224.
- 32. Liersch R, Soicke H, Stehr C, Tullner HU (1986) Formation of artemisinin in Artemisia annua during one vegetation period. Planta Med 52: 387–390.
- 33. Liu S, Tian N, Li J, Huang J, Liu Z (2009) Isolation and identification of novel genes involved in artemisinin production from flowers of Artemisia annua using suppression subtractive hybridization and metabolite analysis. Planta Med 75: 1542–1547.
- 34. Wang H, Ma C, Ma L, Du Z, Wang H, et al. (2009) Secondary metabolic profiling and artemisinin biosynthesis of two genotypes of Artemisia annua. Planta Med 75: 1625–1633.
- 35. Zoschke R, Liere K, Börner T (2007) From seedling to mature plant: arabidopsis plastidial genome copy number, RNA accumulation and transcription are differentially regulated during leaf development. Plant J 50: 710–22.
- 36. Weathers PJ, Elkholy S, Wobbe KK (2006) Artemisinin: The biosynthetic pathway and its regulation in Artemisia annua, a terpenoid-rich species. In Vitro Cell Dev Biol–Plant 42: 309–317.
- 37. Ray A, Robinson-Beers K, Ray S, Baker SC, Lang JD, et al. (1994) Arabidopsis floral homeotic gene BELL (BEL1) controls ovule development through negative regulation of AGAMOUS gene (AG). Proc Natl Acad Sci USA 91: 5761–5765.
- 38. Djakovic-Petrovic T, de Wit M, Voesenek LA, Pierik R (2007) DELLA protein function in growth responses to canopy signals. Plant J 51: 117–126.
- 39. Wang H, Liu Y, Chong K, Liu BY, Ye HC, et al. (2007) Earlier flowering induced by over-expression of CO gene does not accompany increase of artemisinin biosynthesis in Artemisia annua. Plant Biol 9: 442–446.