Introduction

Diatoms are one of the most ecologically diverse groups of organisms in today’s oceans with estimates ranging from 30,000 to 100,000 species1, all of which play a major role in supporting higher trophic levels and carbon fluxes in marine ecosystems2. Despite the extensive diversity of this group, only a handful of species have been found to produce structurally unique C25 and C30 highly branched isoprenoid (HBI) hydrocarbons3. These HBI’s are of great interest given that they have been extensively used as geochemical markers in marine sediments due to their ubiquitous nature4 and that some of the C25 HBIs have also been found to exhibit cytostatic effects on certain lung cancer cell lines5. The unique branched structure of these HBIs also make them a potential form of biofuel through a hydrocracking process similar to botryococcene isoprenoids from the freshwater alga Botryococcus braunii6,7.

Among the handful of diatoms that produce these unique HBIs, Rhizosolenia setigera has garnered considerable attention since this species was found to produce a variety of isomers of both C25 and C30 HBIs depending on the strain, culture conditions and position in its life cycle5,8,9. Masse and co-workers10 established that the HBIs produced by R. setigera are mainly produced by the mevalonate (MVA) pathway for isoprenoid biosynthesis through a series of experiments using isotopic labeling techniques. Despite this, the study was only limited to using substrates that were utilized in more upstream processes of the MVA pathway (isotopically labeled acetate, glucose and CO2) and the final steps in the actual synthesis of these hydrocarbons still remain to be elucidated. It has been suggested that the C25 and C30 HBIs are formed by the attachment of a C10 or C15 isoprenoid unit respectively at C-6 of another C15 isoprenoid unit to produce the unique T branch point observed in their structures11,12. Farnesyl pyrophosphate (FPP) produced by FPP synthase (FPPS) (EC 2.5.1.10) is the most promising candidate for the C15 unit used for the formation of HBIs (Fig. 1).

Figure 1
figure 1

Simplified hypothetical biosynthetic pathway for the formation of representative C25 and C30 highly branched isoprenoids produced by R. setigera.

Red structures (E,E-FPP) represent the C15 isoprenoid unit to which a C10 (blue structures, GPP or NPP) or another C15 (orange structures, E,E-FPP or Z,E-FPP) isoprenoid unit attaches at C-6 to produce the trans- or cis- forms of C25 and C30 HBIs respectively through 1’-6 coupling. Biosynthetic scheme is patterned after those proposed in previous studies (Masse et al., 2004b; Belt et al., 2006).

Full size image

FPPS is an enzyme that catalyzes the head to tail condensation of two molecules of isopentenyl pyrophosphate (IPP) with dimethylallyl pyrophosphate (DMAPP)13 to form FPP. FPP is subsequently converted into a wide range of other natural products such as squalene (through the coupling of two FPP molecules via squalene synthase)14, 5-epi-aristolochene (through the cyclization of FPP via 5-epi-aristolochene synthase)15 and other volatile sesquiterpenes such as β-farnesene and α-bergamotene16. FPP also plays an important role in the prenylation of proteins wherein protein prenyltransferases catalyze the addition of the carbon moeity of FPP to a cysteine residue in the CaaX motif of a target protein, thus facilitating the functional anchoring of these proteins to cell membranes17. cDNA cloning of FPPS has already been carried out from a wide range of organisms such as human18, rat19, yeast20 and higher plants21,22,23 and several studies have also highlighted the important regulatory function of this enzyme in producing precursors and substrates for more complex essential metabolites24,25,26.

In the context of R. setigera, FPPS presents an interesting research target assuming that a C15 isoprenoid unit is used as one of the precursors for the production of C25 and C30 HBIs. Therefore, the aim of this study was to clone and characterize a potential FPPS cDNA from R. setigera, to seek a possible role of the enzyme encoded by this cDNA as the supplier of precursors for HBI biosynthesis and to determine changes in its gene expression in relation to the position of R. setigera in its life cycle. Furthermore, it is to our knowledge the first instance that an FPP synthase was cloned and characterized from the extensively diverse diatoms.

Results

Isolation of RsFPPS cDNA

A tentative FPPS cDNA was isolated from R. setigera CCMP1694 through a series of bioinformatics and PCR techniques. Using amino acid sequences of known FPP synthases as local BLAST queries and limiting the expectation value [E] of the local BLAST function on BioEdit (V 7.2.3) to 1 × 10−6, a single contig was identified from each of the independent R. setigera EST databases from Kochi University and the University of Tokyo. One contig corresponded to 1545 bp while the other corresponded to 1474 bp. Both contigs contained a putative open reading frame (ORF) corresponding to 1299 bp (RsFPPS) with 100% identity to each other and encoding 432 amino acids. Using primers designed from these contig sequences, a 1465 bp fragment was amplified by PCR and cloned as the plasmid pBs-RsFPPS1465. This clone contained an ORF of 1299 bases preceded by 21 bases of 5’-UTR and followed by 145 bases of 3’-UTR (Fig. S1).

Comparison of RsFPPS against FPPS from other organisms

The predicted molecular weight of R. setigera FPPS is 48.94 kDa, which falls within the range for those from mammals (~48 kDa)27,28 and is relatively larger than those from plants and fungi (39-44 kDa)20,21,22,29 and bacteria (~32 kDa)30. A BLAST search against the NCBI online protein database showed that the deduced RsFPPS amino acid sequence shared 58% and 55% identity with annotated FPPSs from eustigmatophytes and brown algae, respectively and 46-51% identity with representative FPP synthase sequences from other organisms (animals, higher plants, green algae and fungi). Phylogenetic analysis of RsFPPS showed that it was more closely related to algal FPP synthase than to those of animals, higher plants, fungi, or bacteria based on sequence conservation (Fig. 2). Sequence alignment of the putative amino acid sequence of RsFPPS against those of mammals, plants, yeast and bacteria (Fig. 3) revealed that it contained all conserved amino acid residues necessary for substrate binding and catalytic activity (domains I-VII) typical of other identified FPPSs31. Two characteristic aspartate-rich motifs were also present in the RsFPPS sequence (Fig. 3).

Figure 2
figure 2

FPPS phylogenetic tree.

R. setigera FPPS was compared against those from animal, plant, yeast, algae and bacteria sources using the Neighbor Joining method based on their amino acid sequences. Alphanumeric codes to the right of the binomial names correspond to accession numbers on the NCBI database. The scale bar on the bottom-left is representative of the degree of difference among sequences wherein a distance of 0.2 infers a 20% difference among sequences.

Full size image
Figure 3
figure 3

Multiple amino acid sequence alignment of R. setigera FPPS.

Comparisons were made against those from mouse, yeast, bacteria and plants. Conserved domains typical of FPP synthases are marked by Roman numerals (I-VII) and a bar above the sequences. The first and second aspartate rich regions in domains II and VI respectively are indicated by boxes.

Full size image

Purification of RsFPPS protein expressed in E.coli

The nucleotide sequence corresponding to the open reading frame of RsFPPS (of which the ATG start codon was removed) was inserted into the pET200/D-TOPO expression vector (pET200-RsFPPS1296) and transformed into BL21(DE3) E. coli cells. Under the conditions described in Supplementary Methods, appreciable amounts of the recombinant protein were obtained as soluble proteins for Ni-NTA purification.

Recombinant RsFPPS, which had an approximate size of 53 kDa due to the inclusion of a 6x-His tag and a linker sequence from the pET200/D-TOPO vector, was isolated to near purity as determined by SDS-PAGE (Fig. S2). The purified protein preparation showed detectable enzyme activity when incubated with IPP and DMAPP or geranyl pyrophosphate (GPP) as seen in the formation of GPP and FPP from liquid chromatography-mass spectrometry (LC/MS) analysis of the reaction mixtures (Fig. 4A,B). Therefore, the R. setigera RsFPPS was found to encode for a functional FPPS enzyme.

Figure 4
figure 4

Representative LC/MS extracted ion chromatograms (EIC).

A) Enzyme reaction with IPP and GPP as substrates, B) enzyme reaction with IPP and DMAPP as substrates, C) GPP and FPP standards at 100 μM concentrations. GPP and FPP eluted at retention times of 4.0 and 4.9 minutes respectively.

Full size image

Product analysis by LC/MS and enzyme kinetics

An LC/MS assay was established to reproducibly identify FPPS reaction products. LC/MS analysis for the detection and quantification of GPP and FPP standards prepared in sample matrices identical to those used for enzyme assays provided sensitive and reproducible results at concentrations ranging from 0.5 to 250 μM. Sufficient separation of GPP and FPP was achieved with sharp peaks of GPP and FPP at retention times of 4.0 and 4.9 min respectively (Fig. 4C). Identification of RsFPPS reaction products (GPP and FPP) was confirmed by similar retention times (Fig. 4A,B) and mass spectra (Fig. S3) to those of the standards used. This allowed for the determination of steady state enzyme kinetics for RsFPPS, wherein enzyme activity was computed as the amount of FPP molecules formed per min per mg protein.

Varying concentrations of the substrates IPP, DMAPP and GPP were used to determine the steady state kinetic constants for recombinant RsFPPS (Fig. 5A-D). For reactions where the IPP concentration was kept constant (50 μM) and DMAPP or GPP were used as counter substrates, the Km value for DMAPP was 6-fold lower than the derived Km value for GPP (Fig. 5A,C). In parallel reactions where the DMAPP (50 μM) or GPP (100 μM) concentrations were kept constant and IPP was used as a counter substrate, Km values for IPP did not differ significantly (Fig. 5B,D). It was noted though, that at higher concentrations of IPP (200 μM), decreases in enzyme activity were observed. Because of this observation, the enzyme-inhibitor constant (Ki) for substrate inhibition in these set of reactions was determined and it revealed that reactions with DMAPP had around two-fold higher Ki values for IPP than reactions with GPP (Fig. S5). Conversely, in terms of the turnover rate (kcat) of RsFPPS, reactions where GPP was used as the counter substrate (Fig. 5C,D) showed kcat values that were 10-fold higher compared to reactions with DMAPP as the counter substrate (Fig. 5A,B), indicating that FPP is more efficiently produced when GPP is used as its allylic substrate.

Figure 5
figure 5

RsFPPS enzyme kinetics.

Purified recombinant RsFPPS was subjected to enzyme assays to determine the effect of varying concentrations of the allylic substrates A) DMAPP and C) GPP on enzyme activity. Parallel reactions using IPP as the counter substrate against B) DMAPP (50 μM) and D) GPP (100 μM) were also conducted. Inset values (A-D) are the means (±S.D.) of the derived kinetic constants Km (μM) and kcat(min−1). Values for kcat were calculated using the dimeric form of the enzyme.

Full size image

Product analysis by gas chromatography/mass spectroscopy (GC/MS)

Products of large scale enzyme assays using IPP and DMAPP as substrates were further verified by GC/MS after converting the generated prenyl pyrophosphates to their corresponding alcohols using alkaline phosphatase. This was done in order to check for possible formation of reaction product cis-isomers since separation of GPP and FPP cis- and trans-isomers was difficult using the described LC/MS conditions. The formation of geraniol (GOH) and E,E-farnesol (E,E-FOH) was easily detected by GC/MS with peaks appearing at retention times of 12.45 and 18.45 min, respectively (Fig. 6A,B, upper panels; Fig. S4). Interestingly, small amounts of their cis- isomers, nerol (NOH) and Z,E-farnesol (Z,E-FOH), were also detected with retention times of 12.12 and 18.22 min, respectively (Fig. 6A,B, lower panels; Fig. S4). When IPP and GPP were used as substrates, only a small amount of Z,E-FOH was detected and no peak corresponding to NOH was observed. The peaks for the cis-isomers were not detected when control reactions using GPP and E,E-FPP standards were treated with alkaline phosphatase indicating that the expressed RsFPPS produced a small amount of neryl pyrophosphate (NPP) and Z,E-FPP as minor products.

Figure 6
figure 6

GC/MS total ion chromatogram of enzyme reaction products using IPP and DMAPP as substrates.

Upper panels show prominent peaks corresponding to A) geraniol (GOH) and B) E,E-farnesol (E,E-FOH) with minor peaks for nerol (NOH) and Z,E-farnesol (Z,E-FOH) marked with asterisks (*). Lower panels are magnified images of the boxed areas in the upper panels. Other minor peaks in the upper panels are either contaminants or artefacts of treatment with alkaline phosphatase as determined in control alkaline phosphatase reactions using GPP or E,E-FPP standards.

Full size image

Inhibition of HBI production in diatoms treated with an FPPS-specific inhibitor

Upon observing that RsFPPS produces potential precursors for HBI biosynthesis, we conducted further experiments to assess how RsFPPS inhibition would affect overall hydrocarbon production. R. setigera cultures fifteen cycles after an auxosporulation event (which were mainly producing C25 HBIs) were subjected to doses of the FPPS-specific inhibitor risedronate in a range from 0 to 50 μM after a 4-day pre-incubation period (Fig. 7). In terms of overall growth, risedronate did not have a significant effect on the final biomass (cells•ml−1) of treated cells for concentrations up to 25 μM. But, overall biomass was reduced by around 25% in treatments with 50 μM risedronate (Fig. 7). Interestingly, total hydrocarbon content (pg•cell-1) was reduced in a dose dependent manner with total HBI content decreasing by 16% and 30% at risedronate treatments of 12.5 μM and 25 μM, respectively, compared to the control (Fig. 7). The HBI content of cells treated with 50 μM risedronate were not significantly different from those treated with 25 μM (Fig. 7). This amount can be attributed to HBIs produced during the pre-incubation period prior to the addition of risedronate. These results would suggest that FPPS does contribute C15 isoprenoids for HBI production in R. setigera.

Figure 7
figure 7

Effect of risedronate on the growth and hydrocarbon content of R. setigera.

Biomass of R. setigera after two-day incubation with risedronate is presented as cells ml−1. Total HBI content of R. setigera quatified by GC/MS is presented as pg cell−1. Values under different letters denote significant differences (p < 0.05) with capital and small letters corresponding to biomass and hydrocarbon content data respectively.

Full size image

Changes in R. setigera hydrocarbon composition, cell morphology and FPPS expression at different life cycle stages

To monitor changes in the HBI composition of R. setigera, samples were collected for hydrocarbon analysis by GC/MS at the end of each 15 day culture period from the 1st (Cy1), 10th (Cy10) and 20th (Cy20) culture cycles after auxosporulation as representative samples of R. setigera at different stages of its life cycle. Significant changes in HBI composition were notably observed between cycles, wherein a shift from C30 HBIs to C25 HBIs as the major hydrocarbon component in cells was evident as culture cycles progressed after the auxosporulation event (Fig. 8A-C). Decreasing cell size and increasing cell densities were also noted in the succeeding culture cycles after auxosporulation (Table 1).

Table 1 Observed changes in culture cycles of R. setigera in relation to auxosporulation in terms of cell density, cell size (diameter) and relative amounts of RsFPPS mRNA.
Full size table
Figure 8
figure 8

GC/MS total ion chromatograms of hydrocarbon extracts (left panels) and photomicrographs (right panels) of R. setigera.

A) 1st culture cycle upon the onset of auxosporulation, B) 10th culture cycle after auxosporulation and C) 20th culture cycle after auxosporulation. Ion peaks appearing between 23 to 25 minutes were identified as C25 HBIs and peaks appearing between 33 to 36 minutes were identified as C30 HBIs based on comparisons of their mass spectra to R. setigera HBIs identified in previous studies.

Full size image

Measurement of relative RsFPPS mRNA levels for the corresponding culture cycles mentioned above was analyzed using RT-qPCR techniques with actin (RsACT) as a reference housekeeping gene for normalization. An increasing trend in RsFPPS mRNA levels was observed in succeeding culture cycles after auxosporulation with a 6-fold increase at Cy20 wherein cell density was highest and cell size was smallest (Table 1). Control experiments with reverse transcriptase omitted and replaced with nuclease free H2O during cDNA synthesis gave no detectable signals by qPCR indicating the absence of genomic DNA in RNA preparations.

Discussion

The use of bioinformatics has become a widely used and reliable tool for the discovery of functional genes. Through these techniques, we were able to screen two sets of R. setigera EST databases for a candidate FPPS. Using the information obtained from the EST databases to design gene specific primers, we were able to isolate an RsFPPS cDNA sequence spanning 1465 bp with an ORF of 1299 bp encoding a protein with 432 amino acids and an estimated molecular weight of 48.94 kDa.

The deduced amino acid sequence of RsFPPS cDNA had a high degree of similarity with other known FPP synthases20,21,22,27,29,30. In particular, the highly conserved domains I-VII as outlined by Koyama31 were present in the RsFPPS protein. Domains II and VI contained highly conserved aspartate-rich motifs (DDxxD) that were determined to be essential for enzyme substrate binding and catalytic function. The first aspartate-rich motif (FARM) in domain II has been shown to play a role in determination of chain length for the resulting prenyl pyrophosphates through the presence of a conserved Phe residue located five amino acids upstream of the DDxxD motif32,33. Studies by Tarshis and co-workers34 on the crystal structure of recombinant avian FPPS also revealed that the second aspartate-rich motif (SARM) in domain VI is the binding site for the homoallylic substrate IPP. Aside from FPP synthases, both FARM and SARM are also present in other trans-isoprenyl pyrophosphate synthases35. Enzyme assays carried out using the recombinant E. coli expressed and purified RsFPPS protein showed considerable enzyme activity and the formation of both FPP and GPP was observed using both LC/MS and GC/MS analyses. Taken together, these results strengthen our claim that the cDNA isolated in this study represents an FPPS gene from R. setigera and to our knowledge, is the first instance that an FPPS was cloned and characterized from a diatom.

The higher Km values of RsFPPS for GPP compared to DMAPP as the allylic substrate were similar to those reported for FPPS of the trematode Schistosoma mansoni36 and FDS-2 from the sagebrush Artemisia tridentata23 that exhibited preference for DMAPP over GPP. Decreased enzyme activity observed at the highest concentrations of IPP used (200 μM) in the enzyme assays (Fig. 5B,D) point to substrate inhibition typically observed for this enzyme with IPP beyond physiological concentrations37. The computed Ki values for these set of reactions revealed that the Ki for IPP was around two times higher in reactions with DMAPP compared to reactions with GPP (Fig. S5). Following the logic presented in previous studies38,39 that showed substrate inhibition by high concentrations of IPP occurs due to the binding of IPP at the allylic site and that higher concentrations of the allylic substrate are necessary to counter this inhibiting effect, the higher Ki value of IPP for reactions with DMAPP further support the observation that RsFPPS exhibits a preference for DMAPP over GPP as an allylic substrate. In terms of turnover rate (kcat), comparison of reactions using GPP as the allylic substrate showed kcat values that were on a scale 10-fold higher than those reported for the two FPPSs (FDS-1 and FDS-2) from A. tridentata23. Higher kcat values for FPPS have only been reported from the phylogenetically distant bacteria Bacillus stearothermophilus40 and show an interesting characteristic of RsFPPS (Supplementary Table 1). Given the high catalytic rates of RsFPPS observed in this study compared to those in literature, it stands as a potential candidate for heterologous expression in other model diatoms such as Phaeodactylum tricornutum41,42 which can be used as a platform in relation to efforts to engineer photosynthetic organisms for the production of valuable isoprenoids as biofuels, pharmaceuticals, cosmetics, or other compounds of interest.

GC/MS analysis of the corresponding reaction products after alkaline phosphatase treatment further revealed that RsFPPS also produced a small amount of the cis-isomers of GPP and E,E-FPP, namely NPP and Z,E-FPP, respectively. The formation of these cis-isomers tends to occur due to the balancing act that FPPS needs to achieve between proper double bond stereoselectivity of IPP and specific selectivity for IPP against its structural isomer DMAPP at the IPP binding site37. Although this phenomenon is not necessarily unique for FPPS as demonstrated in a study by Thulasiram and Poulter37, the presence of these cis-isomers may provide additional evidence for the hypothetical biosynthetic pathways for the various HBIs produced by R. setigera as proposed by Belt et al.12 and Masse et al.11. Among the various HBIs from R. setigera characterized by Volkman et al.43 and Rowland et al.9, structural cis- and trans- isomers at the T branching point of both C25 and C30 HBIs have been noted. This leads us to suggest the possibility that both the cis- and trans-prenyl pyrophosphates produced by RsFPPS could serve as precursors for these HBIs.

To test our hypothesis that RsFPPS provides the precursors necessary for HBI biosynthesis, we conducted in vivo inhibition experiments to assess whether FPPS inhibition would affect overall HBI production. Risedronate, a drug commonly used for bone resorption therapy, was determined to specifically target FPPS as its primary mode of action44,45,46. Using this drug, we were able to demonstrate that even at lower concentrations algal growth was not significantly affected, but HBI production was considerably impaired. These observations were similar to those made by Masse and co-workers10 using mevinolin, an inhibitor of an upstream enzyme in the MVA pathway, 3-hydroxy-3-methylglutaryl-CoA reductase. Our results therefore point to the possible involvement of RsFPPS in the biosynthesis of HBIs in that inhibition of this enzyme potentially limits the amount of precursors necessary for subsequent HBI formation. Taking this into consideration further supports our argument regarding the involvement of RsFPPS in the proposed hypothetical biosynthetic pathway for HBIs previously put forth.

After demonstrating that RsFPPS produces potential precursors for HBIs, we tried to investigate whether RsFPPS played a regulatory role at the transcriptional level in the biosynthesis of varying types of HBIs produced by R. setigera throughout its life cycle. Based on initial observations of C30 HBI predominance during the auxosporulation phase of R. setigera, we assumed that R. setigera would need more C15 isoprenoid units produced by RsFPPS during this phase (Cy1) than during the phase when C25 HBIs were predominant (Cy20). This assumption was also made based on various hypothetical biosynthetic pathways for HBIs proposed in previous studies11,12. Measurement of relative amounts of RsFPPS mRNA by RT-qPCR revealed that, contrary to our initial expectation that RsFPPS expression would be higher during auxosporulation due to the abundance of C30 HBIs, RsFPPS mRNAs were actually more abundant during cycles when C25 HBIs were predominant. This observation could be more in line with a function related to higher growth rates (i.e. faster cell division) rather than the regulation of the types of HBIs between life stages of R. setigera. The smaller cell size and higher cell densities observed in later culture cycles (when C25 HBIs were predominant) are indicative of faster rates of cell division47. In the diatom Phaeodactylum tricornutum, increased protein production has been correlated with higher cell densities and faster growth rates48. Aside from proteins, sterols, which are derivatives of squalene produced from FPP, are important membrane components in eukaryotes49 and sufficient amounts are needed during times of cell division50. Studies by Fabris and co-workers41 also demonstrated that inhibition of sterol biosynthesis in diatoms was concomitant with reduced growth rates. Given the fact that FPP is located at a branch point in the isoprenoid biosynthetic pathway and is used for various other functions such as protein prenylation and sterol biosynthesis, it could be plausible to assume that the observed increases in RsFPPS mRNA levels are more of a function related to increased growth rates and have little or no regulatory role in terms of the types of hydrocarbons produced by R. setigera at various life stages. Furthermore, the relatively small amounts of HBIs produced by R. setigera (roughly <1% dry weight) and the nearly similar ratios between their cis- and trans- isomers9 may indicate that it only uses a small fraction of the pool of E,E-FPP and GPP to produce the HBIs.

In summary, we have isolated and characterized a cDNA that encodes a functional farnesyl pyrophosphate synthase from the marine diatom R. setigera. In our attempts to better understand how this diatom produces unique HBIs, we have presented several lines of evidence that point to the possible contribution of RsFPPS in overall HBI biosynthesis. In terms of the regulation of the types of HBIs produced by R. setigera throughout its life cycle though, RsFPPS does not seem to play any significant role and this phenomenon is most likely controlled by a different set of enzymes which can probably use both the cis- and trans- forms of GPP and FPP as substrates. We are therefore currently undertaking additional studies to discover tentative enzymes from R. setigera that can catalyze the final formation of these HBIs.

Methods

Algal culture

Rhizosolenia setigera CCMP1694 was obtained from the National Center for Marine Algae and Microbiota (Maine, USA). Cultures of R. setigera were maintained in 300 ml of f/2 medium51 in 500 ml Erlenmeyer flasks under controlled conditions (25 °C, illumination at 250 μmol photons m−2 s−1, 12:12 light:dark cycle). Culture cycles were carried out over a period of 15 days and after each culture cycle, 50 ml of 15 day-old culture was used to inoculate 250 ml of fresh f/2 medium. All culture cycles were carried out in triplicate. To determine the effect of auxosporulation on the hydrocarbon production and FPPS gene expression levels, cultures were constantly monitored by light microscopy for the onset of auxosporulation as determined by morphological changes (i.e. larger cell diameters). In the initial culture cycle when the onset of auxosporulation was notably observed (denoted as Cycle 1), aliquotes of 100 ml were collected on Whatman GF/C glass filters (GE Healthcare, United Kingdom) by vacuum filtration, freeze-dried for 24 hours and stored for hydrocarbon extraction and analysis. The rest of the culture was collected on a 20 μm plankton net and the harvested algal cells were snap frozen with liquid nitrogen and stored at −80 °C for RNA extraction and expression analysis. Similar samplings were carried out in succeding culture cycles.

Hydrocarbon extraction

The hydrocarbon fraction from R. setigera was obtained from the freeze-dried samples collected on GF/C filters by extraction following methods previously reported52. The resulting hydrocarbon fraction was subsequently concentrated with a rotary evaporator and a 2.5 μl portion was subjected to GC/MS analysis under conditions described below.

RsFPPS cDNA cloning

A cDNA library was prepared from RNA extracted from 6 day-old R. setigera cells cultured under the same conditions mentioned above. Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA) and mRNA was purified using OligotexTM-dT30 Super mRNA Purification Kit (Takara, Shiga, Japan). ZAP Express cDNA Synthesis Kit and Gigapack III Gold Cloning Kit (Agilent, Santa Clara, CA) were used for cDNA synthesis and library construction according to the manufacturer’s instructions. The original library composed of 200,000 clones was amplified once in XL 1-Blue MRF’ E. coli cells (Agilent).

A putative nucleotide sequence encoding FPPS was mined from two independent R. setigera expressed sequence tag (EST) databases by inputting queries of other known sequences for FPPS using the local BLAST function in BioEdit V 7.2.353. From each database, single contigs with high similarity to other FPP synthases were identified. One contig corresponded to 1545 bp while the other contig was 1474 bp long. Both contigs contained a putative open reading frame (ORF) corresponding to 1299 bp (RsFPPS) encoding 432 amino acid residues with 100% identity with each other. To confirm this sequence, the forward primer Rs-FPPS-est-F and reverse primer Rs-FPPS-est-R (Supplementary Table 2) were designed to amplify a fragment of 1465 bp that included the entire putative coding region and portions of the 5’ and 3’ untranslated regions (UTR) for FPPS. PCR amplification was carried out using KOD DNA polymerase (Toyobo Co. Ltd., Osaka, Japan) and the constructed cDNA library as template. The PCR product was purified by agarose gel electrophoresis and ligated to the EcoRV site of the plasmid pBluescript II KS+ (Agilent). The resulting plasmid was designated as pBs-RsFPPS1465 and transformed into XL 1-Blue MRF’ E. coli cells (Agilent) by the standard CaCl2 protocol. The transformants were plated following protocols for standard blue-white colony selection. Independent positive white colonies were selected to purify plasmids containing the insert for confirmation by sequencing.

Phylogenetic analysis

The deduced amino acid sequence of RsFPPS was compared to the sequences of other known FPPSs using BLAST54. Sequence alignment was carried out using CLUSTAL W255,56 against characterized representative sequences from mouse (Mus musculus)27, yeast (Saccharomyces cerevisiae)20, bacteria (Bacillus stearothermophilus)30 and the higher plants Lupinus albus21, Arabidopsis thaliana22 and Zea mays29. A phylogenetic tree was constructed using the Neighbor Joining function on MEGA software version 4.057,58,59 using the sequences mentioned above along with additional characterized or annotated FPPS sequences from human28, avian34, piscine60, algal61,62,63 and yeast64 sources.

Enzyme assay

Recombinant RsFPPS protein expression and purification are detailed in Supplementary Methods. The purified enzyme’s activity was assayed using methods patterned after those previously described21. Unless otherwise stated, reactions were carried out in 1.5 ml Eppendorf tubes in a total volume of 100 μl containing 35 mM Tris-HCl (pH 7.5), 10 mM MgCl2, 4 mM dithiothreitol (DTT), 50 μM isopentenyl pyrophosphate (IPP; Sigma-Aldrich, St. Louis, MO), 150 μM dimethylallyl pyrophosphate (DMAPP; Sigma) or geranyl pyrophosphate (GPP; Sigma) as the allylic substrates and 100 ng of the purified enzyme. Assay mixtures were prepared in bulk without the substrates and the reactions were initiated by substrate addition after sufficient equilibration of the system temperature. For enzyme kinetic analysis, concentrations of the allylic substrates were varied ranging from 5–200 μM. Parallel assays were done with the allylic substrates at a constant concentration (50 μM DMAPP or 100 μM GPP) and IPP as the counter substrate. Reactions were carried out at 30 °C for 10 minutes and stopped by snap freezing the assay tubes in liquid nitrogen. To denature and precipitate the proteins prior to LC/MS analysis, 100 μl of acetonitrile was added to the mixture, vortexed for 10 seconds and centrifuged at 19,000 g for 5 minutes at 4 °C. Detailed analysis and identification of reaction products by GC/MS were carried following the methods of Thulasiram and Poulter37.

Identification and quantification of assay products by LC/MS

Identification of assay products and analysis of enzyme activity were patterned after modified methods of Nagel et al.65 and Zhang and Poulter66. Analysis was carried out on a Shimadzu LC-20AD HPLC system (Shimadzu, Kyoto, Japan) coupled to an AmaZon SL ion trap mass spectrometer (Bruker Daltonics, USA). Separation of reaction products (GPP and FPP) was achieved on an Inertsil ODS-3 column (particle size 3 μm, 4.6 mm ID x 50 mm, GL Sciences, Tokyo, Japan) using a mobile phase consisting of 5 mM NH4HCO3 as solvent A and 100% acetonitrile as solvent B. A gradient system starting at 0% solvent B, 100% solvent A and increasing to 100% solvent B, 0% solvent A in 8 minutes and a hold at 100% solvent B for an additional 2 minutes was used. A change to 0% solvent B held for 2.5 minutes was done prior to the next injection. Flow rate was at 0.8 ml min−1, column temperature set at 40 °C and injection volume set at 10 μl for all samples and standards. The mass spectrometer was set in negative electrospray ionization mode with ion spray voltage set at −4500 V. For more precise quantification, multiple reaction monitoring was done with transitions of m/z 313.4/159.0 for GPP and m/z 381.5/159.0 for FPP. To establish calibration curves, FPP and GPP standards (Sigma) were prepared and diluted in the same sample matrix used for enzyme assays. Enzyme activity was computed as the amount of FPP produced per minute per mg of protein. Data analysis was done on DataAnalysis V1.5 and QuantAnalysis V1.3.2 software (Bruker Daltonics). Graphs and statistical analysis for enzyme kinetics were done using GraphPad PRISM software V 6.0 (GraphPad Software Inc., La Jolla, CA, USA).

GC/MS analysis

Algal hydrocarbon extracts and the corresponding prenyl alcohols obtained by enzymatic dephosphorylation of the enzyme reaction products were analyzed by GC/MS based on methods previously reported52. Identification of algal hydrocarbons was based on relative retention times and mass spectra of previously characterized HBIs from R. setigera while identification of enzyme reaction products was based on their relative retention times and mass spectra of prenyl alcohols in the mass spectra library on the Shimadzu GCMS LabSolution software V 2.71.

In vivo inhibition of RsFPPS with risedronate

For inhibition studies with risedronate (Tokyo Chemical Industry, Tokyo, Japan), algal cultures fifteen cycles after auxosporulation were used. Using the same conditions mentioned above, a 1L cell culture was allowed to grow for 4 days and subsequently subdivided into 60 ml aliquotes in 100 ml erlenmeyer flasks upon which risedronate was added at concentrations of 0, 12.5, 25 and 50 μM. Each treatment was done in triplicate. The cultures were then allowed to incubate for an additional two days before 50 ml was collected on GF/C glass filters for hydrocarbon analysis.

For quantitative analysis of hydrocarbons in risedronate treated cultures, 10 ng of hexadecane was added to each filter paper as an internal standard prior to extraction and a 2 μl portion was used for GC/MS analysis. Quantification was done by creating a calibration curve using an authentic C25 HBI isomer isolated from R. setigera as a standard.

RsFPPS gene expression analysis by RT-qPCR

The relative expression levels of RsFPPS were monitored in cultures of R. setigera at various stages in its life cycle. Total RNA was extracted from diatom samples stored at −80 °C as described above. cDNA synthesis for qPCR was done using iScript cDNA Synthesis Kit (Bio-Rad) following the manufacturer’s protocols. qPCR reaction mixtures contained 10 μl of SsoFast EvaGreen Supermix (Bio-Rad), 7 μl of PCR-grade water, 1 μl each of forward and reverse primers (10 μM) and 1 μl of the synthesized cDNA for a total volume of 20 μl. Monitoring of real-time PCR reactions was done on a CFX96 Real-Time System coupled to a C1000 Thermal Cycler (Bio-Rad). Relative expression levels of RsFPPS were normalized against the actin gene (RsACT) as a reference housekeeping gene67. Primers for RsFPPS were designed from the confirmed sequence of plasmid pBs-RsFPPS1465 while primers for RsACT were designed based on similar contigs mined from the two R. setigera EST databases mentioned earlier (Supplementary Table 2). To test for contamination of genomic DNA, extracted RNA was subjected to similar methods mentioned above with the exception that reverse transcriptase was replaced with nuclease free H2O in the cDNA synthesis step. The nucleotide sequences for RsFPPS and RsACT have been submitted to the NCBI GenBank™ database with accession numbers KM360174 and KM360175 respectively.

Additional Information

How to cite this article: Ferriols, V. M. E. N. et al. Cloning and characterization of farnesyl pyrophosphate synthase from the highly branched isoprenoid producing diatom Rhizosolenia setigera. Sci. Rep. 5, 10246; doi: 10.1038/srep10246 (2015).