Virus-induced gene silencing identifies Catharanthus roseus 7-deoxyloganic acid-7-hydroxylase, a step in iridoid and monoterpene indole alkaloid biosynthesis

Authors


Summary

Iridoids are a major group of biologically active molecules that are present in thousands of plant species, and one versatile iridoid, secologanin, is a precursor for the assembly of thousands of monoterpenoid indole alkaloids (MIAs) as well as a number of quinoline alkaloids. This study uses bioinformatics to screen large databases of annotated transcripts from various MIA-producing plant species to select candidate genes that may be involved in iridoid biosynthesis. Virus-induced gene silencing of the selected genes combined with metabolite analyses of silenced plants was then used to identify the 7-deoxyloganic acid 7-hydroxylase (CrDL7H) that is involved in the 3rd to last step in secologanin biosynthesis. Silencing of CrDL7H reduced secologanin levels by at least 70%, and increased the levels of 7-deoxyloganic acid to over 4 mg g1 fresh leaf weight compared to control plants in which this iridoid is not detected. Functional expression of this CrDL7H in yeast confirmed its biochemical activity, and substrate specificity studies showed its preference for 7-deoxyloganic acid over other closely related substrates. Together, these results suggest that hydroxylation precedes carboxy-O-methylation in the secologanin pathway in Catharanthus roseus.

Introduction

Catharanthus roseus is the sole source of the important anticancer drugs vinblastine and vincristine, monoterpenoid indole alkaloids (MIAs) that are extensively used for treatment of leukemia and lymphoma (van der Heijden et al., 2004). This medicinal plant accumulates a number of MIAs that originate from coupling of the indole tryptamine (formed by tryptophan decarboxylase, TDC) and the iridoid secologanin by strictosidine synthase to form strictosidine, from which most other MIAs are generated (O'Connor and Maresh, 2006; El-Sayed and Verpoorte, 2007; Oudin et al., 2007).

Iridoids are cyclopenta[c]pyran monoterpenoids that occur widely in numerous medicinal plants, usually as biologically active glucosides with anti-hepatotoxic, antitumor, antiviral, anti-inflammatory and other properties (Dinda et al., 2007, 2011). In plants, iridoids also function as important defense compounds in plant-herbivore interactions through their anti-feeding and anti-microbial activities (Peñuelas et al., 2006). Because of the pharmacological and economic importance of iridoids, there has been intense interest in understanding their biosynthetic mechanisms in order to develop biotechnological approaches for their targeted production. While several genes involved in the biosynthesis of secologanin have been cloned and functionally characterized in the medicinal plant Croseus, several genes remain to be characterized in order to complete this pathway (O'Connor and Maresh, 2006; El-Sayed and Verpoorte, 2007; Oudin et al., 2007).

The first committed step in secologanin biosynthesis (Figure 1) involves oxidation of geraniol at its C-10 position by geraniol 10-hydroxylase (G10H, CYP76B6) (Collu et al., 2001). Further oxidation of 10-hydroxygeraniol to the dialdehyde, 10-oxogeranial, is catalyzed by 10-hydroxygeraniol oxidoreductase (10HGO) (Ikeda et al., 1991). Conversion of this dialdehyde to iridodial occurs by a unique reductive cyclization performed by an NADPH-dependent 10-oxogeranial cyclase named iridoid synthase (IRS) (Geu-Flores et al., 2012). The conversion of iridodial to secologanin involves several enzymes, some of which are dependent on cytochrome P450s, and the precise biochemical reactions have yet to be characterized.

Figure 1.

Iridoid biosynthesis pathway from geraniol to secologanin.

G10H, geraniol 10-hydroxylase; 10HGO, 10-hydroxygeraniol oxidoreductase; IRS, iridoid synthase; GT, glucosyltransferase; DL7H 7-deoxyloganic acid 7-hydroxylase; LAMT, loganic acid-O-methyltransferase; SLS, secologanin synthase. Multiple arrows indicate multiple steps. The alternative pathway that has been proposed involves O-methylation of 7-deoxyloganic acid to 7-deoxyloganin by LAMT, followed by hydroxylation of 7-deoxyloganin to loganin by DL7H.

The late stage of secologanin biosynthesis from iridodial involves a three-step oxidation to form 7-deoxyloganetic acid, most probably catalyzed by cytochrome P450, followed by glucosylation to produce 7-deoxyloganic acid before progression to two putative pathways for synthesis of loganin, which is further converted to secologanin by secologanin synthase (SLS), a cytochrome P450-dependent enzyme (CYP72A1) (Uesato et al., 1986; Irmler et al., 2000; Yamamoto et al., 2000). In the first pathway, a hydroxylation step precedes a methylation step in which 7-deoxyloganic acid is hydroxylated by 7-deoxyloganic acid 7-hydroxylase (CrDL7H) to produce loganic acid, which is then methylated into loganin by loganic acid methyltransferase (LAMT) (Madyastha et al., 1973), which has been cloned and characterized in Croseus (Murata et al., 2008). This methyltransferase accepts loganic acid and secologanic acid but not 7-deoxyloganic acid as a substrate. In the alternative pathway, the methylation step precedes hydroxylation, with 7-deoxyloganic acid being methylated to 7-deoxyloganin before hydroxylation to form loganin. The 7-deoxyloganin 7-hydroxylases detected in microsomes isolated from Croseus and Lonicera japonica cell cultures are putative cytochrome P450s, as suggested by their requirement for NADPH and oxygen for activity (Irmler et al., 2000; Katano et al., 2001). Although the Lonicera enzyme has been partially purified from cell suspension cultures, no genes encoding the relevant hydroxylase have been cloned and functionally characterized. Therefore, identification of CrDL7H was initiated by searching our EST database of Croseus for cDNAs that encode putative cytochrome P450 genes (CYPs) with high similarities to those found in other MIA-producing plant species and also found in the secologanin-producing species Ljaponica (www.phytometasyn.ca/). The PhytoMetaSyn Project (www.phytometasyn.ca) comprises large annotated transcript databases from 75 non-model plants know to produce biologically active terpenes, alkaloids and polyketides (Facchini et al., 2012; Xiao et al., 2013). This search for putative CYP genes led to the identification of three cytochrome P450 candidates (CYP76A26, CYP72A224 and CYP72A271) that may be involved in iridoid biosynthesis. Virus-induced gene silencing (VIGS) of one candidate CYP (CYP72A224) belonging to the CYP72 sub-family resulted in silenced plants that accumulated 7-deoxyloganic acid rather than secologanin. Recombinant expression and characterization of this CYP in yeast showed it to be a 7-deoxyloganic acid 7-hydroxylase that precedes the O-methylation step catalyzed by LAMT in Croseus.

Results

Bioinformatics-guided screening of the PhytoMetaSyn database provides three CYP candidate genes to be tested for the deoxyloganic acid-7-hydroxylase function

Screening of databases of transcripts from various MIA-producing Apocynaceae plants (Vinca minor, Rauwolfia serpentina, Tabernaemontana elegans and Amsonia hubrichtii), Cinchona ledgeriana (Rubiaceae) (O'Connor and Maresh, 2006) and secologanin-producing Ljaponica (Caprifoliaceae) (Kawai et al., 1988) that have been assembled and annotated for the PhytoMetaSyn Project (http://www.phytometasyn.ca/) (Facchini et al., 2012) identified several putative homologs of known iridoid biosynthetic genes from Croseus. All six species contain candidates for each of the known iridoid biosynthetic genes, including a recently characterized geraniol synthase (GS) (Simkin et al., 2013), G10H, 10HGO, IRS, LAMT and SLS (Table S1). Inspection of G10H, 10HGO, IRS, LAMT and SLS sequence comparisons showed that MIA producing plant species contained 87 candidate genes with 80-90% amino acid sequence identity to Catharanthus genes with known biochemical functions in the secologanin pathway. The 87 candidate hits were then screened against transcriptomes of five randomly selected non-secologanin-producing species (Abies balsamea, Saponaria vaccaria, Cannabis sativa, Lactuca sativa and Papaver bracteatum) in the PhytoMetaSyn Project to eliminate genes that are likely to be involved in other pathways. This screen yielded 29 candidates that were not represented in the non-secologanin-producing species. Further screening of the 29 candidates against secologanin-producing Ljaponica yielded CYP76A26, CYP72A224 and CYP72A271, which were selected for further study by virus induced gene silencing and metabolite profiling. This strategy for candidate gene identification was recently described in a review on MIA biosynthesis in Catharanthus (Salim and De Luca, 2013).

Virus-induced gene silencing of CrCYP72A224 modifies the iridoid metabolite profile of silenced Catharanthus plants

In order to test their possible roles in secologanin biosynthesis, each of the three genes were subjected to modified virus-induced gene silencing (VIGS) (De Luca et al., 2012) based on the method used for successful identification of new genes involved in iridoid biosynthesis (Geu-Flores et al., 2012) and MIA biosynthesis (Liscombe and O'Connor, 2011) in Croseus that utilizes a well-known pTRV vector system (Burch-Smith et al., 2004) that has been successfully applied in several plant species. As described previously (De Luca et al., 2012), additional experiments to silence the phytoene desaturase gene were also performed as the visible photobleached phenotype obtained was useful in determining when it was time to perform transcript and metabolite profiling in plants in which expression of each of the three CYP genes was suppressed. The effects of transcriptional down-regulation of CYP76A26, CYP72A224 and CYP72A271 on the iridoid profiles of silenced Catharanthus tissues were monitored by UPLC-MS. As silencing of CYP72A271 did not produce any decreases in secologanin or monoterpenoid indole alkaloid levels, this candidate gene was eliminated from further study. Silencing resulted in dramatic decreases of secologanin in CrCYP72A224 VIGS plants (Figure 2, DL7H-vigs-1 and DL7H-vigs-2) compared to the levels of secologanin found in plants treated with empty vector (Figure 2, EV). As the loss of secologanin resulted in accumulation of deoxyloganic acid (Figure 2, DL7H-vigs-1 and DL7H-vigs-2), the putative substrate of CrDL7H (Figure 1), CrCYP72A224 appears to be the gene responsible for the DL7H reaction in Catharanthus. Although silencing of CYP76A26 also produced a significant decrease in secologanin levels, silenced plants did not show detectable accumulation of any other intermediates in the secologanin pathway. Studies on the possible function of CYP76A26 in secologanin biosynthesis remain to be performed.

Figure 2.

In planta silencing of CrDL7H leads to a decrease in secologanin levels with accumulation of the 7-deoxyloganic acid intermediate.

Selected UPLC-MS chromatograms of two VIGS-treated plants (DL7H-vigs-1 and DL7H-vigs-2) are compared with plants treated with empty vector (EV).

Transcript analysis of CrDL7H monitored by real-time PCR revealed that the transcript level decreased by approximately 75% (< 0.05) in leaves infiltrated with Agrobactierium tumefaciens harboring pTRV2-CrDL7H compared with those exposed to bacteria containing pTRV2-EV (Figure 3a). More detailed metabolite analyses of CrDL7H -silenced tissues showed that secologanin levels decreased by at least 70% compared with non-inoculated (WT), EV- or mock-inoculated plants (Figure 3b). CrDL7H-silenced tissues accumulated more than 4 mg g−1 fresh weight of deoxyloganic acid (Figure 3b) compared to Catharanthus control, EV- and mock-inoculated tissues. MIA analyses of CrDL7H-silenced tissues showed a >60% decrease in catharanthine accumulation and a >30% decrease in vindoline accumulation compared to Catharanthus control, EV- and mock-inoculated tissues (Figure 3c).

Figure 3.

Effects of silencing CrDL7H in Croseus.

(a) CrDL7H transcript levels in plants infiltrated with Atumefaciens harboring pTRV2-CrDL7H was determined by quantitative real-time PCR to measure the fold change of transcript levels from individual plants expressing the silenced gene relative to mock-infected plants. Values are means ± standard error of at least eight biological replicates performed on cDNA from each individual plant infiltrated with Atumefaciens harboring each construct. (b,c) Effect of suppressing CrDL7H transcript levels on the iridoid contents (b) and the major alkaloids (catharanthine and vindoline) (c) of Croseus leaves from plants infiltrated with Atumefaciens expressing each construct. Values are means ± standard error of at least eight biological replicates. Determination of catharanthine and vindoline levels from wild-type, pTRV2-EV, mock and pTRV2-CrDL7H treated plants showed that the production of MIAs was impaired when secologanin biosynthesis was suppressed by silencing genes involved in the iridoid pathway. Asterisks indicate statistically significant differences (< 0.05, Student's t test) for the transcript analysis and metabolite contents of the silenced line relative to EV-infected plants.

Functional characterization of CrDL7H

To functionally characterize CrDL7H, an in vivo assay was performed using an engineered dual expression yeast system (Ro et al., 2002) comprising pESC-Leu2d containing Croseus NADPH cytochrome c reductase (CrCPR) as a redox partner and the candidate gene on the same vector under the control of a galactose-inducible promoter. Supplementation of the induced CrDL7H-expressing yeast culture with 7-deoxyloganic acid led to production of a more hydrophilic iridoid (Figure S1B) with the same retention time (Rt = 1.9 min) and mass (m/= 375) as authentic loganic acid (Figure S1A), i.e. the 7-hydroxy derivative of 7-deoxyloganic acid (Figure 4b), but no biotransformation of 7-deoxyloganic acid to loganic acid was observed in the yeast cultures transformed with empty pESC-Leu2d expressing only CrCPR (Figure 4c).

Figure 4.

Recombinant CrDL7H yeast cells and isolated microsomes both convert 7-deoxyloganic acid (m/= 359) to loganic acid (m/= 375).

(a) Spectrum for the loganic acid standard.

(b,c) UPLC-MS analysis shows that biotransformation of 7-deoxyloganic acid to loganic acid only occurred in the yeast expressing CrDL7H (b), but not in yeast expressing empty vector containing CrCPR alone (c).

(d,e) UPLC-MS analysis also shows that loganic acid is formed only in the presence of CrDL7H-expressing microsomes, NADPH and 7-deoxyloganic acid (d), but not in the absence of NADPH (e).

The ability of the yeast strain expressing recombinant CrDL7H to convert 7-deoxyloganic acid to loganic acid was further investigated in cell free extracts. Microsomes prepared from CrDL7H-expressing yeast cultures were assayed for the presence of CrDL7H activity. In the presence of NADPH, yeast microsomes were able to convert 7-deoxyloganic acid to a product (Figure S1D) with the same retention time (Rt = 1.9 min), UV spectra and mass (m/z 375) as authentic loganic acid (Figure S1C and Figure 4d). Assays lacking the NADPH cofactor (Figure 4e) or those performed using boiled CrDL7H-enriched microsomes served as negative controls. It was possible to replace the NADPH by NADH, but this co-substrate was only 20% as active. Other iridoid substrates tested (Figure S2; 7-deoxyloganin, 10-deoxygeniposide and loganin) were not accepted by the recombinant enzyme. Kinetic analyses with yeast microsomes expressing CrDL7H (Figure S3) yielded apparent Km and Vmax values of 111.07 ± 14.8 μm and 5.5 ± 0.77 μm min−1 μg protein for 7-deoxyloganic acid and 29.85 ± 3.28 μm and 4.19 ± 0.40 μm min−1 μg−1 protein for NADPH. The Km obtained for the enzyme expressed in yeast microsomes is similar to that for the enzyme expressed in microsome preparations from Ljaponica cell cultures (Km values of 170 and 18 μm for 7-deoxyloganin and NADPH, respectively) (Katano et al., 2001).

Preferential expression of CrDL7H in various Catharanthus plant organs

The expression levels of CrDL7H in various Croseus organs were examined by quantitative real-time PCR, and their relative abundance was compared to the levels of secologanin, catharanthine and vindoline (Figure 5). The transcript level of CrDL7H was highest in leaf pair 1, and decreased with leaf age, as seen in leaf pairs 2–5 (Figure 5). The patterns of expression of CrDL7H were very similar to those of LAMT and SLS (Murata et al., 2008), in agreement with the pattern of secologanin and MIA accumulation within these organs (Figure 5). The levels of CrDL7H were consistently higher in other actively growing aerial organs, for example when comparing expression in flower buds with that in more mature open flowers (Figure 5). Expression of CrDL7H was also detected in stems and roots, consistent with the detection of secologanin and/or MIAs in these organs (Figure 5) and with the role of this enzyme in catalyzing the 3rd to last step in secologanin biosynthesis in Catharanthus (Figure 1).

Figure 5.

Expression of CrDL7H transcripts, iridoids and MIAs in various organs of Croseus plants.

Expression of CrDL7H in various organs of Croseus was measured by quantitative real-time PCR in three biological replicates, and each biological replicate comprises three technical replicates. Data show expression relative to that for the 60S RPPOC ribosomal gene.

The carborundum abrasion technique has been successfully used in our laboratory to harvest leaf epidermis-enriched RNA from Croseus and to confirm MIA pathway gene expression in these cells (Murata et al., 2008; De Luca et al., 2012). Analysis of isolated leaf epidermis-enriched transcripts by quantitative real-time PCR showed that the levels of transcripts for loganic acid methyltransferase (LAMT) and secologanin synthase (SLS), which are well known to be preferentially expressed in the leaf epidermis (Murata et al., 2008; Guirimand et al., 2011), were 2.5 and 5 times higher in these cells than in whole leaf extracts (Figure 6). In contrast, the levels of transcripts for G10H and IRS, which have been shown to be expressed in internal phloem parenchyma cells (Burlat et al., 2004; Geu-Flores et al., 2012), and CrDL7H were 4, 7 and 7 times higher, respectively, in whole leaves than in the leaf epidermis (Figure 6). Together, these results suggest that CrDL7H is preferentially expressed within Catharanthus leaves rather than in leaf epidermal cells where the last two steps in secologanin assembly take place.

Figure 6.

Relative abundance of G10H, IRS, CrDL7H, LAMT and SLS transcripts in leaf epidermis-enriched extracts compared to whole leaves.

Expression of CrDL7H in various organs of Croseus was measured by quantitative real-time PCR performed in three biological replicates; each biological replicate comprises three technical replicates. Data show expression relative to that for the 60S RPPOC ribosomal gene.

Phylogenetic analysis

The phylogenetic analysis of CrDL7H (CYP72A224) is most closely associated with a putative secologanin synthase-like gene from Camptotheca acuminata (accession number HQ605982.1; 56% amino acid identity), and shares 54% amino acid identity with Croseus SLS (CYP72A1), which has been functionally characterized (Irmler et al., 2000). All species that accumulate secologanin (Ljaponica) and/or MIAs (Cledgeriana, Vminor, Telegans, Rserpentina and Ahubrichtii) appear to contain CrDL7H-like candidate genes as well as SLS-and G10H-like genes belonging to the CYP72A and CYP76B subfamilies, respectively (Figure 7). The phylogenetic tree shown in Figure 7 and created using the genes listed in Table S2 illustrates the relationship of CYPs involved in iridoid biosynthesis to four functionally characterized CYP72A genes involved in triterpene biosynthesis (Seki et al., 2011; Fukushima et al., 2013). CrDL7H also shows 49% and 48% amino acid identity to CYP72A154 from Glycyrrhiza uralensis and CYP72A63 from Medicago truncatula, which are involved in C-30 oxidation in glycyrrhizin and β-amyrin, respectively (Seki et al., 2011). In addition, CYP72A61v2 and CYP72A68v2, which are 49% identical to CrDL7H from Mtruncatula, performed unusual oxidations of triterpenoids when expressed in transgenic yeast (Fukushima et al., 2013). The phylogenetic analysis (Figure 7) clearly shows that both the CrDL7H and SLS genes from the six species grouped separately from the other members of the CYP72 family involved in triterpene biosynthesis as well as from four other Croseus CYPs with known biochemical functions [flavonoid 3′,5′-hydroxylase (CYP75A8), Kaltenbach et al., 1999; cinnamate 4-hydroxylase, C4H (CYP73A4), Hotze et al., 1995; tabersonine 16-hydroxylase (T16H; CYP71D12), Schröder et al., 1999; tabersonine 19-hydroxylase (T19H; CYP71BJ1), Giddings et al., 2011].

Figure 7.

Phylogenetic relationships among relevant cytochrome P450s to CYP72A proteins.

A phylogenetic tree was generated using the neighbor-joining algorithm (Tamura et al., 2011) based on a comparison of amino acid sequences using the ClustalW program (Thompson et al., 1994). The numbers on each branch represent bootstrap values for 10 000 replicates. The scale bar of 0.1 indicates to a 10% change and each number shown next to the branches is the percentage of replicate trees in which the related taxa clustered in the bootstrap test with 10 000 replicates. Cr, Catharanthus roseus; Vm, Vinca minor; Rs, Rauwolfia serpentina; Te, Tabernaemontana elegans; Ah, Amsonia hubrichtii; Cl, Cinchona ledgeriana; Lj, Lonicera japonica; Ca, Camptotheca acuminata; Gu, Glycyrrhiza uralensis; Mt, Medicago truncatula; G10H, geraniol 10-hydroxylase; SLS, secologanin synthase; DL7H, 7-deoxyloganic acid 7-hydroxylase; C4H, cinnamate 4-hydroxylase; F3,5H, flavonoid 3′,5′-hydroxylase; T16H, tabersonine 16-hydroxylase; T19H, tabersonine 19-hydroxylase.

Discussion

This study describes a successful bioinformatics approach, using transcriptomic data from seven plant species that produce secologanin (Ljaponica) and/or MIAs (Croseus, Vminor, Telegans, Rserpentina, Ahubrichtii and Cledgeriana) to identify candidate cytochrome P450/def/mybox{/vrule depth -0.5mm height 4mm width 8mm}genes involved in previously uncharacterized biochemical steps in iridoid biosynthesis. The search identified a secologanin synthase-like CYP72A sub-family member (CYP72A224) in addition to SLS (CYP72A1) that was previously functionally characterized from Croseus cell cultures (Irmler et al., 2000). Although the CYP72A sub-family is widely distributed among plants, including the Arabidopsis and rice model systems (http://drnelson.uthsc.edu/biblioD.html), four CYP72A sub-family genes that have been functionally characterized are involved in triterpene biosynthesis (Seki et al., 2011; Fukushima et al., 2013). In another example, four clustered P450s from the same CYP71C family are involved in 2,4-dihydro-1,4-benzoxazin-3-one (DIBOA) and 2,4-dihydroxy-7-methoxy-1,4-benzoxazin-3-one (DIMBOA) biosynthesis in the Gramineae (Frey et al., 1997). Interestingly, these clustered genes share an overall amino acid identity of 45–60% and show strict substrate specificities, as do 7-deoxyloganic acid hydroxylase (CYP72A224) and secologanin synthase (CYP72A1) that catalyse the 3rd to last and last steps in secologanin biosynthesis.

While a few Croseus P450 genes have been cloned by traditional screening methods combined with in vitro enzyme assays of recombinant proteins to determine substrate specificities (Mizutani and Ohta, 2010), several recent efforts have demonstrated the effectiveness of VIGs for transiently down-regulating pathway steps in a growing number of plant species (De Luca et al., 2012) and for identifying appropriate candidate genes as in the present study. The silencing of CrDL7H (CYP72A224) in Croseus decreased both secologanin and MIA levels in infected plants, and led to the appearance of 7-deoxyloganic acid, which is not usually detected in these plant tissues (Figure 2). In addition, CrDL7H-expressing yeast converts 7-deoxyloganic acid efficiently to loganic acid (Figure 4b,c) and in vitro enzyme assays with CrDL7H-containing microsomes show that NADPH is required for conversion of this iridoid to loganic acid, while control yeast cells or microsomes harboring empty vector were unable to catalyze this reaction (Figure 4d,e). The kinetic analysis of CrDL7H demonstrated its high affinity for 7-deoxyloganic acid, and its inability to oxidize 7-deoxyloganetic acid, 7-deoxyloganin or 10-deoxygeniposide strongly suggests that secologanin biosynthesis preferentially involves hydroxylation followed by methylation of the carboxyl group (Figure 1) rather than the alternative route. In Catharanthus, this preferred route is also supported by the high substrate specificity of recombinant LAMT (Murata et al., 2008) for loganic acid over 7-deoxyloganic acid. The accumulation of 7-deoxyloganic acid but not 7-deoxyloganin (Figure 2) in CrDL7H VIGS plants is also consistent with hydroxylation occurring before methylation in biosynthesis of secologanin in Croseus. In contrast, the substrate preference for 7-deoxyloganin together with the kinetic properties of microsomal 7-deoxyloganin 7-hydroxylase from Ljaponica cell cultures (Katano et al., 2001) suggest that the order of these reactions (Figure 1) may be different in other plant species or even in different organs of a particular plant species. Together, the present study and the Ljaponica and Croseus cell culture studies (Irmler et al., 2000; Katano et al., 2001) indicate that the order of hydroxylation and methylation is flexible (Figure 1). Studies with Galium mollugo (Uesato et al., 1986) suggested that hydroxylation occurs at the stage of 7-deoxyloganic acid, based on feeding various labeled iridoid and secoiridoid glucosides. The ability of Ljaponica cell cultures to convert 7-deoxyloganin to loganin (Yamamoto et al., 1999) deserves further investigation by cloning the relevant hydroxylase(s) and methyltransferase(s) for functional expression and biochemical analysis.

The gradual decrease of CrDL7H transcripts during leaf development (Figure 5; leaf pairs 1–5) reflects the results obtained for LAMT enzyme activity (Murata et al., 2008), and the pattern is consistent with the levels of secologanin and catharanthine accumulation in various organs. The expression profile of CrDL7H in floral tissues suggests that young flower buds may be more active in iridoid biosynthesis than more mature open flowers (Figure 5), but further detailed studies are required to assess iridoid biosynthesis in this organ. While expression of LAMT and SLS is restricted to leaf epidermal cells (Figure 6) (Irmler et al., 2000; Murata et al., 2008), quantitative real-time PCR analysis of epidermis enriched cDNA showed that G10H, IRS and CrDL7H transcripts are not highly expressed in epidermal cells (Figure 6). The biosynthesis of monoterpenoid precursors in aerial organs of Croseus has been shown to involve at least two cell types: epidermal cells and internal phloem parenchyma cells where G10H (Burlat et al., 2004) and iridoid synthase are expressed (Geu-Flores et al., 2012). Although accumulation of CrDL7H transcripts in non-epidermal cells, such as internal phloem parenchyma cells, requires verification by in situ hybridization studies, these results suggest possible involvement of internal phloem parenchyma cells that would require intercellular translocation of a loganic acid intermediate between internal phloem parenchyma cells and the leaf epidermis.

In conclusion, we describe the functional characterization of CYP72A224 that hydroxylates 7-deoxyloganic acid. This identification of CrDL7H not only provides insight into the detailed reactions in the iridoid biosynthesis pathway, it also reveals the potential of metabolic engineering using plants or microbial hosts for production of valuable monoterpenoid indole alkaloids. Functional analysis of the other P450 candidates involved in oxidation steps from iridodial to 7-deoxyloganetic acid in Croseus is being undertaken to uncover rate-limiting steps in iridoid biosynthesis and its regulation.

Experimental procedures

Plant materials

Catharanthus roseus (L.) Don cv. Little Delicata seeds were germinated in 50-well trays on wet soil in the greenhouse using a 16 h light/8 h dark photoperiod at 28°C. For VIGS treatment, the seedlings were grown for 4 weeks or until they produced at least two true leaf pairs.

Iridoid extraction

Plant tissues harvested for iridoid extraction were frozen and pulverized in liquid nitrogen. Methanol (5 ml) was added and the leaves were homogenized. The samples were placed on a shaker for 1 h before metabolite analysis by UPLC-MS.

Chemicals

The 7-deoxyloganic acid (14 mg) used in this study was purified from an EtOAc extract of olive wood (Olea europaea) as follows: a sample of olive wood shavings (200 g) was extracted using CH2Cl2 under argon in the dark for 2 h at reflux. The CH2Cl2 extract was filtered and concentrated to give a dried residue (1.20 g). Then, the remaining plant material was extracted with EtOAc under argon in the dark for 2 h at reflux. The EtOAc extract was evaporated under vacuum at a temperature not higher than 40°C. The dried EtOAc extract (3.20 g) was submitted to chromatography on a 200 g silica gel 60 (63-200 μm particle size) column and iridoids were eluted using CH2Cl2/EtOH mixtures of increasing polarity. Fractions of 75 ml each were collected, monitored by TLC and HPLC, pooled and evaporated to give ten major fractions (A1–A10). Fraction A8 (290 mg) was submitted to chromatography on a 40 g silica gel 60 (40–63 m particle size) column and iridoids were eluted as in the previous column chromatography. Fractions of 25 ml each were collected, monitored by TLC and HPLC, pooled and evaporated to give ten major fractions (B1–B10). Fraction B5 (42 mg), enriched in 7-deoxyloganic acid, was re-purified by preparative HPLC, flash evaporation and lyophilization to yield 7-deoxyloganic acid (14 mg), after evaporating MeOH using a rotary evaporator and the remaining H2O using a freeze drier. This final HPLC purification was performed using a preparative reversed-phase HPLC column (Spherisorb ODS-2; Waters Corporation, http://www.waters.com, 250 mm length, 10 mm internal diameter, 5 μm particle size) on a Waters (Waters Corporation) 600E instrument equipped with a diode array detector (scan range 190–800 nm), operating at 30°C. The separation was performed by a step gradient using 99.8:0.2 v/v MeOH/AcOH (solvent A) and 99.8:0.2 v/v H2O/AcOH (solvent B) at a flow rate of 5 ml min−1: isocratic conditions of 10% A for 5 min, linear gradient from 10% to 100% A over 35 min, isocratic conditions of 100% A for 5 min, and a further 5 min to return to the initial conditions. The 1H-NMR spectrum of 7-deoxyloganic acid, recorded at 400 MHz on a Bruker (http://www.bruker.com/) Avance 400 spectrometer (using CD3OD as solvent and tetramethylsilane as internal reference), agrees with that previously reported for this compound (Pérez-Bonilla et al., 2006). The solvents used for extraction and chromatographic separations were HPLC grade. The 7-deoxyloganin and 7-deoxyloganetic acid were kindly provided by Hajime Mizukami (Department of Pharmacognosy, Graduate School of Pharmaceutical Sciences, Nagoya City University, Japan).

Cloning and construction of VIGS vectors

A fragment of 363 bp from the CrDL7H gene was amplified using Ex-Taq polymerase (Takara Bio Inc, http://www.takara-bio.com/) and inserted to a pGEM®-T Easy vector (Promega corporation, http://www.promega.com) using primers 5′-CCAAGGCTCTTTTTTTCCAT-3′ (forward) and 5′-AGAATCCAGTATCCCCAAAT-3′ (reverse). The CrDL7H sequence was harvested from the pGEM-T Easy vector using EcoR1 restriction enzyme and inserted into the EcoRI site between the CaMV 35S promoter (2 × 35S) and the NOS terminator of the pTRV2 vector. The orientation of the insert was verified by sequencing. Assembly of the pTRV1 and pTRV2 vectors (provided by Peter Facchini, Department of Biology, University of Calgary, Canada) was performed as previously described (Dinesh-Kumar et al., 2003).

Virus-induced gene silencing

Virus-induced gene silencing (VIGS) was performed as described previously (Liscombe and O'Connor, 2011; De Luca et al., 2012). Agrobacterium tumefaciens strain GV3101 harboring pTRV1, the empty pTRV2 vector (pTRV2-EV), or the pTRV2 vector containing the Catharanthus DL7H transcript (pTRV2-CrDL7H) were cultured overnight at 28°C in 300 ml Luria–Bertani medium containing 10 mm MES, 20 μm acetosyringone and 50 μg ml−1 kanamycin. These cultures were centrifuged at 5000 g for 10 min, and the bacterial pellets were resuspended in 5 ml infiltration buffer (10 mm MES, 200 μm acetosyringone and 10 mm MgCl2), and further incubated at 28°C for 3 h with shaking.

Croseus cv. Little Delicata seeds were germinated and grown in a greenhouse under a 16 h light/8 h dark photoperiod at 28°C for 3–6 weeks to produce at least two true leaf pairs. Young plants were wounded using toothpicks through the stem just below the apical meristem, and infiltrated with sterile water (mock inoculation), with a 1:1 v/v A. tumefasciens cultures harboring pTRV1 and either pTRV2-EV or pTRV2-crDL7H constructs. Other young plants were also treated with Agrobacterium cultures harboring pTRV1 and pTRV2 constructs containing a Catharanthus phytoene desaturase (PDS). Suppression of PDS, that is involved in chlorophyll biosynthesis, produces visible photobleaching and is a useful marker for determining the best time to perform transcript and metabolite analyses of DL7H suppressed plants. Typically, the PDS phenotype was observed 3 weeks after inoculation of seedlings, and leaves from control uninoculated plants and mock-, EV- and CrDL7H-inoculated plants were harvested at this stage. After recording the fresh weights of harvested materials, one member of a leaf pair was used for RNA extraction, while the other was used for metabolite analysis. Leaf tissues were frozen in liquid nitrogen and subjected to extraction using a tissue lyser (TissueLyser II, Qiagen, http://www.qiagen.com) for quick pulverization. Frozen 2 ml microfuge tubes containing leaf materials and 100 μl of 1 and 2 mm glass beads in a 4: atio were transferred to a TissueLyser adapter set that had been frozen at −80°C for at least 2 h and accommodates 24 samples/plate. Tissue lysis was performed at 30 Hz for 1 min, after which samples were cooled in liquid nitrogen for 1–2 min, and tissue lysis was repeated for another minute. Tissue lysed powders were extracted with Trizol reagent (Invitrogen, http://www.lifetechnologies.com/invitrogen) in order to harvest RNA.

Cloning of CrDL7H into yeast expression vector

Total RNA from the first pair of Croseus leaves was isolated using TRIzol® reagent (Invitrogen). cDNA synthesis was performed using AMV reverse transcriptase (Promega) with an oligo d(T) primer, according the manufacturer's instructions. This cDNA pool was used to clone CrDL7H (Genbank accession number KF415115) and a NADPH-cytochrome P450 reductase (CPR) from Croseus (Genbank accession number X69791.1) into the pESC-Leu2d yeast dual expression vector (Agilent Technologies, http://www.agilent.com), as described by Ro et al. (2008). This yeast expression vector was modified by deleting some regions in the promoter allele of the LEU2 gene, which causes Saccharomyces cerevisiae to produce a higher plasmid copy number to compensate for the weakened LEU2 expression. The coding region of the CPR gene was amplified by PCR using high-fidelity Phusion DNA polymerase (New England Biolabs, http://www.neb.ca) using the primers 5′-GCGACGGAGTTGGGATTTTAT-3′ (forward) and 5′-TTGAAAACATCTGGAGGGGTG-3′ (reverse) The PCR fragment of the CPR gene was inserted into the pGEM®-T Easy vector (Promega), digested with NotI and inserted into the corresponding site behind the GAL10 promoter in the correct orientation. The full-length open reading frame (ORF) for CrDL7H was first cloned into the pGEM®-T Easy vector (Promega) using the primers 5′-TACAGCGGGCCCAGGATGGAATTGAACTTCAAATC-3′ (forward) containing an ApaI site (underlined) and 5′-GCGCGTCGACGAGTTTGTGCAGAATCAAATGA-3′ (reverse) containing a SalI site (underlined), excised from the pGEM®-T Easy vector, and ligated to the correpsonding Apa1, Sal1 sites behind the GAL1 promoter. Its sequence was submitted to the P450 nomenclature committee (D.R. Nelson, Department of Biochemistry, University of Tennessee at Memphis, TN) for assignment of a P450 family name.

Yeast strain growth for in vivo activity

Scerevisiae MKP-0 was used as the host strain for CrDL7H expression from pESC–Leu2d. The corresponding ‘empty’ vector expressing only the CPR gene was also introduced into this yeast strain via the standard polyethylene glycol/lithium acetate procedure (Gietz et al., 1992). Scerevisiae harboring the plasmid was grown in selective synthetic media without Leu and with His, Uracil, Adenine, Tryp, Lys, 2% glucose as a carbon source, and protein expression was induced using SG–Leu dropout medium with 2% galactose as the sole carbon source as previously described (Ro et al., 2002). For the in vivo assay in yeast, 50 ml late log phase cells (5 × 107 and 2 × 108 cells/ml) were transferred to the same medium to be induced for 12 hr with 2% Galactose rather than glucose. Induced cells were harvested by centrifugation at 1000 g, and resuspended in 25 ml Tris/EDTA buffer (pH 7.5) with 0.2 mm 7-deoxyloganic acid. The yeast culture was incubated with deoxyloganic acid at 28°C for 24 h with shaking at 150 rpm, the cells were removed by centrifugation, for 10 min at 1000 g at 4°C and the medium was freeze-dried overnight and resuspended in 1 ml water. The sample was passed through a Supelclean™ ENVI-18™ column (Sigma Aldrich, http://www.sigmaaldrich.com) equilibrated with 2 ml acetonitrile followed by 2 ml water. The sample was eluted using 200 μl methanol, and analyzed by UPLC-MS.

Enzyme assays and substrate specificity

Yeast microsomal preparations (Pompon et al., 1996) were analyzed for protein content (Bradford, 1976) and were used for in vitro enzyme assays (0.1 ml) containing 460 μg microsomal protein, 1 mm NADPH, 4 mm dithiothreitol, initiated by addition of 138 μm deoxyloganic acid. After 120 min at 30°C, assays were stopped by chilling on ice and by adding water to total volume of 1 ml. Samples were treated on a Supelclean™ ENVI-18™ column as described above. Substrate specificity assays contained 0.1 mm of each iridoid tested.

Enzyme assays (0.2 ml reaction volume) containing 2.3 mg of CrDL7H-enriched microsomes and 4 mm dithiothrietol with serial dilutions of 7-deoxyloganic acid (27.7, 55.5, 83.3, 138, 277, 555, 694, 972 and 1510 μm) at 1 mm NADPH and of NADPH (10, 15, 20, 25, 30, 40, 80 100 and 120 μm) at 0.2 mm 7-deoxyloganic acid were prepared for kinetic analysis. The apparent Michaelis–Menten constant (Km) and maximal reaction velocity (Vmax) for 7-deoxyloganic acid and NADPH were determined by Lineweaver–Burk plots using GraphPad (http://www.graphpad.com/).

Metabolite analysis by LC-MS

Iridoids and MIAs were isolated and analyzed as described previously (Roepke et al., 2010).

Quantitative real-time PCR expression analysis

Expression of G10H, IRS, CrDL7H, LAMT and SLS as well as RPPOC (encoding the 60S acidic ribosomal protein P0-C) was monitored by quantitative real-time PCR using appropriate transcripts and primers (Table S3). The parameters used for quantitative real-time PCR were 95°C for 3 min, 40 cycles of 95°C for 10 sec, 50°C for 20 sec and 72°C for 30 sec, using the CFX96 real-time SYBR system (Bio-Rad, http://www.graphpad.com/) and SYBR Green detection according to the manufacturer's instructions.

Phylogenetic analysis

The amino acid alignments were performed using ClustalW (Thompson et al., 1994). The neighbor-joining phylogeny was generated using MEGA 5.1 (http://www.megasoftware.net/mega.php) with bootstrap analysis of 10 000 replicates (Tamura et al., 2011). The GenBank accession numbers for the genes used for construction of the phylogenetic tree are listed in Table S2.

Acknowledgements

This work was funded by the National Sciences and Engineering Research Council of Canada (to V.D.L), a Canada Research Chair in Plant Biotechnology (to V.D.L.) and by the Centro de Instrumentación Científico-Técnica of the University of Jaén (to J.A.). We also acknowledge funding by Genome Canada, Genome Alberta, Genome Prairie, Genome Quebec, Genome British Columbia, the Canada Foundation for Innovation, the Ontario Ministry of Research and Innovation, the National Research Council of Canada and other government and private sector partners. We thank Peter Facchini (Biology Department, University of Calgary, Canada) for providing the VIGS vectors, and Dae Kyun Ro (Biology Department, University of Calgary, Canada) for providing the yeast expression system. We recognize the skilled technical work of next-generation sequencing personnel at the McGill University/Genome Québec Innovation Centre. We are grateful to Christoph Sensen, Mei Xiao and Ye Zhang (all from Biology Department, University of Calgary, Canada) for their dedicated bioinformatic support and large-scale gene annotation efforts that produced searchable databases (hosted on PhytoMetaSyn web site) and led to the identification of various genes described in this report. We thank David Nelson (University of Tennessee at Memphis, TN) for his assistance in classifying CYP72A224.

Ancillary