Structure Conservation and Differential Expression of Farnesyl Diphosphate Synthase Genes in Euphorbiaceous Plants

Farnesyl diphosphate synthase (FPS) is a key enzyme of isoprenoids biosynthesis. However, knowledge of the FPSs of euphorbiaceous species is limited. In this study, ten FPSs were identified in four euphorbiaceous plants. These FPSs exhibited similar exon/intron structure. The deduced FPS proteins showed close identities and exhibited the typical structure of plant FPS. The members of the FPS family exhibit tissue expression patterns that vary among several euphorbiaceous plant species under normal growth conditions. The expression profiles reveal spatial and temporal variations in the expression of FPSs of different tissues from Euphorbiaceous plants. Our results revealed wide conservation of FPSs and diverse expression in euphorbiaceous plants during growth and development.


Introduction
Euphorbiaceae is one of the largest plant families and consists of more than 7000 species. Euphorbiaceus species are evolutionally-diversified, carry distinct physiologies, and have complex traits adapting to dynamic environmental conditions [1]. There are many economically-important plants in Euphorbiaceae, such as the rubber tree (Hevea brasiliensis), the cassava (Manihot esculenta), and the castor bean (Ricinus communis). The rubber tree is the most widely cultivated species for OPEN ACCESS commercial production of natural rubber (cis-polyisoprene) for tires and other products [2]. The cassava is a tropical crop that stores important quantities of starch in its roots. The high starch content makes cassava a desirable energy source both for human consumption and industrial biofuel applications [3]. The castor bean is cultivated in the tropical and subtropical areas of the world for oil production and as an ornamental plant [4].
Isoprenoids constitute a versatile class of compounds fulfilling major physiological functions [5]. The isoprenoid pathway constitutes the most diverse and widespread metabolic pathway of all prokaryotes and eukaryotes, resulting in the biosynthesis of a large number of primary as well as secondary metabolites [6]. In plants isoprenoids are formed by the mevalonate (MVA) pathway in the cytosol [7,8] and the 1-deoxy-D-xylulose 5-phosphate (DXP)/2-Cmethyl-D-erythritol 4-phosphate (MEP) pathway in plastids [9,10]. The MVA pathway is primarily responsible for the synthesis of sesquiterpenes, triterpenes including brassinosteroids, larger molecules such as dolichols, and even macromolecular polyisoprene (natural rubber) [6,11,12]. Farnesyl diphosphate synthase (FPS) is a key enzyme in isoprenoids biosynthesis, which catalyzes the consecutive condensations of dimethylallyl diphosphate (DMAPP) or geranyl diphosphate (GDP) with isopentenyl pyrophosphate (IPP) to produce farnesyl diphosphate (FDP) [8]. FDP serves as a precursor for sesquiterpenoids, sterols, brassinosteroids, triterpenoids, polyprenols, side chains of ubiquinone, and polyisoprenoids such as natural rubber [13,14]. However, little is known of the FPS genes in the Euphorbiaceus species. In this study, the gene structure, phylogenetic characteristics, and expression patterns of Euphorbiaceae plants FPSs were identified and described. Our results revealed wide conservation of FPSs and diverse expression profiles in Euphorbiaceous plants during growth and development.

Cloning, Identification and Structure Analysis of the Euphorbiaceous Plants FPSs
To identify the potential members of the FPS family in euphorbiaceous plants, we used Arabidopsis FPSs (AtFPS1 and AtFPS2) as queries and obtained all possible FPSs by searching the genome database of the rubber tree (Hevea brasiliensis), cassava (Manihot esculenta), castor bean (Ricinus communis), and Jatropha (Jatropha curcas). Three members in the rubber tree (designated as HbFPS1, HbFPS2, and HbFPS3), three members in the cassava (designated as MeFPS1, MeFPS2, and MeFPS3), two members in the castor bean (designated as RcFPS1, RcFPS2), and two members in the Jatropha (designated as JcFPS1, JcFPS2) were identified on the basis of the BLASTP search. The full-length cDNAs of the ten FPSs were PCR amplified, cloned and sequenced. The deduced proteins of the FPSs ranged from 342 to 352 amino acids (predicted molecular mass = 39.37 to 40.80 kDa) with isoelectricpoints ranging from 4.85 to 6.06 ( Table 1). The deduced FPS proteins contained the five conserved regions identified by Chen et al. [15] that are characteristic of prenyltransferases that synthesize isoprenoid diphosphates with E-double bonds ( Figure 1). The highly conserved aspartate-rich motif DDXXD was present in domains II and V. Ten FPSs identified from euphorbiaceous plants showed more than 65.2% amino acid identity and the maximum percentage of amino acid sequence identities was found between HbFPS1 and MeFPS1 (95.61%, respectively) ( Table 2).  Identical and conserved amino acid residues are denoted by black and gray backgrounds, respectively. The five conserved domains of prenyltransferases are underlined and numbered. The highly-conserved aspartate-rich motifs (DDXXD) is present in domains II and V.

Phylogenetic Analysis
Phylogenetic and molecular evolutionary analyses were conducted using MEGA version 6 [16] by comparing ten FPS from euphorbiaceous plants with known FPS sequence from a wide range of different organisms including bacteria, fungi, plants, and animals ( Figure 2). The results indicated that ten FPSs from euphorbiaceous species appeared at the base of the clade of the plant kingdom, and that FPSs evolved from a common ancestor. Moreover, FPSs from euphorbiaceous plants were clustered into two distinct subgroups. One subgroup contained HbFPS1, HbFPS2, MeFPS1, MeFPS2, RcFPS1, EpFPS, and JcFPS1, which was more closely related to the FPS of legume plants. The other subgroup contained HbFPS3, MeFPS3, RcFPS2, and JcFPS2.

Intron and Exon Organization of FPSs
We analyzed the intron and exon structure of ten FPSs from the rubber tree, the cassava, the castor bean, and the Jatropha ( Figure 3). All these FPSs contained twelve exons and eleven introns. Although introns differ in length, these introns were typically flanked by GT and AG boundaries.

Structure Prediction and Homology Modeling of the FPSs
In order to obtain a reasonable theoretical structure of the euphorbiaceous plant FPSs, protein homology modeling was performed using a Swiss model server. To predict the 3D structure of the FPSs, a 3D structure at 2.20 Å of Artemisia Spiciformis FPS1 (PDB id: 4kk2.1) was used as a template, which shares 80.59%, 75.29%, 66.07%, 79.71%, 79.41%, 66.57%, 80.00%, 66.27%, 79.71% and 65.36% sequence identity with HbFPS1-3, MeFPS1-3, RcFPS1-2, and JCFPS1-2, respectively. The predicted 3D model of FPSs was validated with the QMEAN server [17] for model quality estimation. The total QMEAN-score (estimated model reliability between 0 and 1) of the predicted 3D models for the ten  (Figure 4). The Asn residue in motif-I of HbFPS1-2, MeFPS1-2, RcFPS1, and JcFPS1 have the similar predicted 3D structure; also, the similar predicted 3D structure is found in the Tyr residue in motif-I of HbFPS3, MeFPS3, RcFPS2, and JcFPS2. However, the Tyr, instead of ASn, residue forms a different predicted 3D structure, where Asn residue forms an open structure, the Tyr residue forms a cyclic structure.

Expression Analysis of FPSs in Euphorbiaceous Plants Tissues
In order to characterize the expression profile of FPS in euphorbiaceous plants, we analyzed the tissue-specific expression pattern of FPSs in three euphorbiaceous species. In the rubber tree, HbFPS1 was predominant in the latex, revealed more than a 20-fold difference in the expression levels of different organs. HbFPS2 and HbFPS3 had similar expression profiles, HbFPS2 and HbFPS3 were expressed in all the tested tissues at different levels, with the highest transcription occurring in flowers, followed by latex, barks, leaves, and root. HbFPS1 showed more than 30-fold higher levels of transcript abundance than HbFPS2 and HbFPS3 in different organs ( Figure 5A). We also compared the transcripts of FPSs in each tissue in the cassava and found that the expression levels of MeFPS1, MeFPS2, and MeFPS3 had similar expression profiles, but MeFPS3 revealed more than a 100-fold difference in the expression levels than MeFPS1 and MeFPS2 in different organs ( Figure 5B). In the castor bean, RcFPS1 and RcFPS2 were expressed in all the tested tissues at different levels, with the highest transcription occurring in seeds, followed by flowers, stems, leaves, and root ( Figure 5C).

Discussion
Plants contain small farnesyl diphosphate synthase isozyme families. cDNAs encoding FPS have been cloned and characterized from various plant species [18][19][20][21][22][23][24][25][26]. Arabidopsis contains two genes, FPS1 and FPS2, encoding three FPS isozymes: FPS1L, FPS1S and FPS2. The FPS1 encodes FPS1S and FPS1L, which differ only by an N-terminal extension of 41 amino acid residues that targets FPS1L into mitochondria [19,20], whereas the FPS2 encodes FPS2 that shares 90.6% amino acid identity with FPS1 isozymes [21]. Three FPS isoforms have also been discovered in both maize and Artemisia tridentate [22,23]. In humans, only a single FPS encodes for FPS. Due to the alternative splicing in the first exon of human FPS, multiple splice variants are generated which encode two FPS isoforms: a shorter cytoplasmic/peroxisomal form, and a longer isoform which is a mitochondrial targeting peptide [24]. Although one FPS (HbFPS1) from the rubber tree and one FPS (EpFPS) from the Euphorbia had been characterized [25,26], knowledge of the FPS genes of euphorbiaceous plants is limited. In this study, ten FPSs were identified in Euphorbiaceous species, including three members in the rubber tree, three members in the cassava, two members in the castor bean, and two members in the Jatropha. Sequence and phylogenetic analysis results showed wide conservation of FPSs in euphorbiaceous plants.
The members of the FPS family exhibit tissue expression patterns that vary among several plant species. In Arabidopsis, FPSs are expressed in all organs throughout plant development, albeit at greatly different levels. FPS1 is widely expressed in all tissues throughout plant development, whereas expression of FPS2 is mainly concentrated in floral organs, seeds, and the early stages of seedling development [27][28]. In Ginkgo biloba, GbFPS had high transcription in roots and leaves, and low in stems [29], reflecting the fact that the biosynthesis of ginkgolides and bilobalide occurs in roots and leaves [30]. In Euphorbia pekinensis, the highest EpFPS expression level was detected in roots, in which terpenoids are synthesized [26]. In the rubber tree, HbFPS1 is expressed predominantly in the laticifers and is likely to encode the enzyme involved in natural rubber biosynthesis [25]. The expression of HbFPS2 and HbFPS3 is not cell-type specific. HbFPS2 and HbFPS3 are possibly involved in isoprenoid biosynthesis of a housekeeping nature. Our results revealed that all of the eight FPS genes were differentially expressed in all tissues tested either in their transcript abundance or expression patterns under normal growth conditions.
Our results showed that a substantial number of FPSs which were previously identified and characterized in well studied model plants are conserved in important Euphorbiaceous plants. Despite broad conservation across the euphorbiaceous species, these FPSs also exhibited diverse expression patterns.

Plant Materials and Treatments
Rubber tree (Hevea brasiliensis cultivar RRIM 600), castor bean (Ricinus communis cultivar A202), and cassava (Manihot esculenta cultivar SC8) obtained from Institute of Tropical Bioscience and Biotechnology, were planted in the experimental farm of the Chinese Academy of Tropical Agricultural Sciences in Hainan Island in China (20°N, 110°E). Fresh leaves, flowers, roots, fruits, and barks were immediately ground to form powder in liquid nitrogen and stored at −70 °C or immediately used to extract nucleic acid. The latex of rubber tree was allowed to drop directly into liquid nitrogen in an ice kettle. The frozen latex powder was then stored at −70 °C or used immediately to extract RNA.

Cloning and Identification of FPS Genes
Total RNA was extracted from the rubber tree latex [31] and from other tissues [32]. cDNA was synthesized by reversely transcribing 1 μg total RNA using a PrimeScript™ RT-PCR kit (Takara, Dalian, China) according to the manufacturer's instructions. To identify the FPS homologs in H. brasiliensis, we used Arabidopsis FPS genes (AtFPS1 and AtFPF2) as queries and BLAST analysis of genome database of rubber tree (DDBJ/EMBL/GenBank under the accession: GenBank: AJJZ01000000), cassava (http://www.phytozome.net/cassava) [33], castor bean (http://castorbean.jcvi.org) [4], and jatropha (http://www.kazusa.or.jp/jatropha/) [34]. The contigs of putative FPS genes were then assembled. The cDNA of putative FPSs were amplified by primers based on the assembled sequences ( Table 3). The primers were designed using the Primer Generator (http://www.med.jhu.edu/medcenter/primer/primer.cgi). The PCR products were cloned in the pMD19-T cloning vector (TaKaRa, Dalian, China) and sequenced. The sequence was performed using the ABI BigDye ® Terminator Sequencing Kits in ABI3700 DNA sequencer. Afterward, their sequences were analyzed in GenBank by using the BLAST program. The isoelectric point (pI) of FPS was predicted using the compute pI/MW software (http://www.expasy.ch/tools/ pi_tool.html). The percentage of FPS amino acid identity in four euphorbiaceous plants were done with Clustal W2 (http://www.ebi.ac.uk /Tools/msa/clustalw2/). The gene structure schematic of FPSs identified from four euphorbiaceous plants was drawn using the web server GSDS (http://gsds.cbi.pku.edu.cn/). Multiple amino acid sequence alignment and phylogenetic tree analysis were performed using the MEGA 6.0 software.

Homology Modeling and Structure Prediction
Protein sequences of ten FPS were submitted to the Swiss-Model server (http://swissmodel. expasy.org) [35] to perform sequence analysis, and Artemisia Spiciformis farnesyl diphosphate synthase 1 (PDB id: 4kk2.1) was applied as a template. The catalytically-and enzymatically-important residues of FPSs were displayed using the Pymol software (Delino Scientific, San Carlos, CA, USA).