Ipecac alkaloid biosynthesis in two evolutionarily distant plants

Colinas, Maite; Morweiser, Clara; Dittberner, Olivia; Chioca, Bianca; Alam, Ryan; Leucke, Helena; Nakamura, Yoko; Serna Guerrero, Delia Ayled; Heinicke, Sarah; Kunert, Maritta; Wurlitzer, Jens; Ploss, Kerstin; Hong, Benke; Grabe, Veit; Lopes, Adriana A.; O’Connor, Sarah E.

doi:10.1038/s41589-025-01926-z

Download PDF

Article
Open access
Published: 03 June 2025

Ipecac alkaloid biosynthesis in two evolutionarily distant plants

Nature Chemical Biology (2025)Cite this article

4845 Accesses
1 Citations
146 Altmetric
Metrics details

Subjects

A preprint version of the article is available at bioRxiv.

Abstract

Ipecac alkaloids are medicinal monoterpenoid-derived tetrahydroisoquinoline alkaloids found in two distantly related plants: Carapichea ipecacuanha (Gentianales) and Alangium salviifolium (Cornales). Here we provide evidence suggesting that both plants initiate ipecac alkaloid biosynthesis through a nonenzymatic Pictet–Spengler reaction and we elucidate the biosynthetic fate of both the 1R and 1S stereoisomers that are produced in this nonstereoselective reaction. Although the biosynthesis of the 1S-derived protoemetine proceeds according to the same chemical logic in both species, each plant uses a distinct monoterpene precursor. Phylogenetic analyses show examples of independent pathway evolution through parallel and convergently evolved enzymes. This work provides insight into how nature can capitalize on highly reactive starting substrates and the manner in which multistep pathways can arise and lays the foundation for metabolic engineering of these important medicinal compounds.

Comparative transcriptomics of two Salvia subg. Perovskia species contribute towards molecular background of abietane-type diterpenoid biosynthesis

Article Open access 06 February 2024

Metabolomics-guided discovery of cytochrome P450s involved in pseudotropine-dependent biosynthesis of modified tropane alkaloids

Article Open access 02 July 2022

Revealing evolution of tropane alkaloid biosynthesis by analyzing two genomes in the Solanaceae family

Article Open access 15 March 2023

Main

Plants produce an enormous diversity of natural products or specialized metabolites. Although natural product pathways are typically lineage specific, in some cases, evolutionarily distant plants independently evolved pathways to synthesize the same molecule^1,2. Examples include glucosinolates³, benzoxazinoids⁴, caffeine^5,6,7, cannabinoids⁸ and cardenolides^9,10,11. Ipecac alkaloids are monoterpenoid-derived tetrahydroisoquinoline alkaloids that occur in Carapichea ipecacuanha (Gentianales) and Alangium salviifolium (Cornales). Both species are known medicinal plants: ipecac syrup made from C. ipecacuanha rhizomes has been used as a vomit-inducing medicine in cases of intoxication, while A. salviifolium, also known as Ankol(a), is used as an emetic and to treat a variety of diseases in traditional ayurvedic medicine^12,13. The active emetic ingredients are the tetrahydroisoquinoline alkaloid pathway products, cephaeline and emetine, both derived from protoemetine (Fig. 1). Additionally, anticancer and antimalarial activities have been described for other protoemetine-derived alkaloids such as tubulosine^14,15. Because Cornales and Gentianales are estimated to have diverged approximately 150 million years ago¹⁶, these plants serve as excellent models to study the evolution of a complex medicinal alkaloid biosynthesis pathway.

**Fig. 1: Ipecac alkaloid biosynthesis in C. *ipecacuanha* and A. *salviifolium* based on the literature.**

In C. ipecacuanha, ipecac alkaloid biosynthesis begins with a Pictet–Spengler reaction that couples the monoterpenoid secologanin 1 with dopamine to generate the initial tetrahydroisoquinoline scaffold. Conflicting studies suggest that either secologanin 1 or secologanic acid 2 can be similarly conjugated with dopamine 3 in A. salviifolium^17,18,19,20, although secologanic acid has been shown to be the monoterpene precursor in indole alkaloid biosynthesis in Camptotheca acuminata, also in the Cornales lineage^21,22. Although previously identified enzymes that catalyze the Pictet–Spengler reaction are stereoselective, unusually, both S and R stereoisomers of the initial tetrahydroisoquinoline Pictet–Spengler product are observed in C. ipecacuanha (deacetylisoipecoside DAII 4a (S-epimer) and deacetylipecoside DAI 4b (R-epimer))²³ (Fig. 1). Similarly, methylated forms of both S and R stereoisomers of the corresponding acids (deacetylisoipecosidic acid DAIIA 5a (S-epimer) and deacetylipecosidic acid DAIA 5b (R-epimer)) are observed in A. salviifolium^17,24. Early work suggested that a Pictet–Spenglerase enzyme is present in A. salviifolium but a gene encoding this enzyme has never been identified^18,25. In both C. ipecacuanha and A. salviifolium, the S-epimers are converted by O-methylation, deglycosylation, reduction and decarboxylation to protoemetine 8, an ipecac alkaloid common to both species^26,27,28. In C. ipecacuanha, protoemetine is converted into cephaeline and emetine²⁶, whereas, in A. salviifolium, proteoemetine is converted to cephaeline, alangimarckine and tubulosine^29,30. In C. ipecacuanha, the R-epimer is N-acetylated to ipecoside 11 (ref. ³¹), whereas, in A. salviifolium, the R-epimer is converted to 6-O-Me-DAIA 6b and 7-O-Me-DAIA 12b (ref. ²⁴) (Fig. 1).

A few ipecac alkaloid biosynthetic genes from C. ipecacuanha encoding glucosidases and O-methyltransferases (OMTs) have been reported^32,33,34 but the biosynthesis of ipecac alkaloids remains largely unknown. Here, we report the complete discovery of protoemetine biosynthesis in both C. ipecacuanha and A. salviifolium, which we show proceeds through an unexpected order of enzymatic reactions. We provide evidence that the Pictet–Spengler reaction initiating the pathway can occur spontaneously in the vacuole, which would explain the presence of both 1R and 1S stereoisomers in these plants. While protoemetine is derived from the 1S-epimer, we also identify biosynthetic genes that derivatize the 1R-epimer in C. ipecacuanha and A. salviifolium. Phylogenetic analyses suggest that the enzymes that convert DAII(A) to protoemetine evolved independently in C. ipecacuanha and A. salviifolium through means of parallel and convergent evolution. This collection of metabolic pathways provides a striking example of evolution of complex, medicinally important compounds in phylogenetically distant plants.

Results

Metabolite profiling of C. ipecacuanha and A. salviifolium

We isolated C. ipecacuanha and A. salviifolium tissues for RNA sequencing and for metabolomic profiling. Protoemetine (8) (1S stereoisomer)-derived alkaloids were found in C. ipecacuanha rhizomes (cephaeline and emetine) and in A. salviifolium root (cephaeline) (Fig. 2a). The 1R-derived products were also found in C. ipecacuanha rhizomes (ipecoside, 11) and A. salviifolium root (6-O-Me-DAIA, 6b) (Fig. 2a,d). Tissue-specific metabolite profiling (Fig. 2e,f and Supplementary Figs. 1 and 2; comparisons to standards in Supplementary Figs. 3–7) revealed that C. ipecacuanha accumulates similar amounts of ipecac alkaloids in both young leaves and rhizomes (Fig. 2f). In contrast, A. salviifolium pathway intermediates up to protoemetine 8 are detected in high levels in leaf buds but cephaeline only accumulates in roots and barks of older stems (Fig. 2e). We used these metabolite profiles to guide the search for protoemetine gene candidates in the corresponding RNA-seq datasets.

Fig. 2: Metabolite profiling and evidence for a nonenzymatic Pictet–Spengler reaction in C. *ipecacuanha* and A. *salviifolium.*

A Pictet–Spengler reaction in ipecac alkaloid biosynthesis

In the metabolite profiling experiments described above, we also noted that secologanin (1) is only observed in C. ipecacuanha while secologanic acid (2) is observed only in A. salviifolium (Fig. 2b,c). This suggests that C. ipecacuanha uses secologanin in ipecac alkaloid biosynthesis while A. salviifolium uses the corresponding acid. Consistent with this, DAII 4a and DAI 4b are detected in C. ipecacuanha but not A. salviifolium (Fig. 2d). Unexpectedly, we noticed that when we infiltrated the starting substrates (secologanin or secologanic acid together with dopamine) into Nicotiana benthamiana leaves, the corresponding Pictet–Spengler products DA(I)I or DAI(I)A (both 1S and 1R stereoisomers) were formed within 24 h (Fig. 2g,h). Moreover, when Catharanthus roseus flower petals, which contain endogenous secologanin 1 but do not produce ipecac alkaloids, were infiltrated with dopamine, accumulation of DAI 4b (1R) and DAII 4a (1S) was also observed after 24 h (Fig. 2i), clearly demonstrating that this nonenzymatic reaction can take place in planta. We were unable to detect enzymatic activity in C. ipecacuanha or A. salviifolium that impacted the rate or stereoselectivity of this nonenzymatic reaction (Supplementary Fig. 8). Dopamine is a highly activated substrate for the Pictet–Spengler reaction, reacting rapidly in a nonstereoselective manner under mild acidic conditions³⁵. In contrast, we showed that Pictet–Spengler reactions with tryptamine, either in N. benthamiana leaves or in C. roseus flower petals, resulted in much lower levels of nonenzymatically formed coupled product, suggesting that dopamine is chemically activated compared to tryptamine (Extended Data Fig. 1). Although enzymes that catalyze the Pictet–Spengler reaction with dopamine are known, these enzymes are stereoselective³⁶. Because C. ipecacuanha and A. salviifolium produce products derived from both stereoisomers and because we demonstrated that this reaction can occur in two different plants lacking a dedicated enzyme, we hypothesize that this Pictet–Spengler coupling may occur nonenzymatically and nonstereoselectively in these plants.

Ipecac alkaloid biosynthetic gene candidates

After the Pictet–Spengler reaction, protoemetine is generated from DAII (4a) or DAIIA (5a) (1S-epimers) by O-methylation, deglycosylation, reduction and, in the case of C. ipecacuanha, deesterification (Fig. 1). The 1R-epimers (DAIIA 5a or DAIA 5b) are either acetylated to form ipecoside (11) in C. ipecacuanha or O-methylated to form 6-O-Me-DAIA 6b and 7-O-Me-DAIA 12b in A. salviifolium. We searched for genes in our generated transcriptomes that encode enzymes that could catalyze these reactions. First, we identified the previously published C. ipecacuanha OMT genes (renamed CiDOMT1, CiDOMT2 and CiDPOMT) and glucosidase (CiDGD)^32,33,34. Expression levels of the three OMT genes in our C. ipecacuanha transcriptome were highest in young leaves and rhizome (Extended Data Fig. 2a), consistent with the metabolite profile of ipecac alkaloid accumulation (Fig. 2f). Using the previously discovered CiDOMT1 gene as bait for coexpression analysis, we identified two reductase and three esterase transcripts. Surprisingly, the previously reported glucosidase gene, CiDGD, was not coexpressed with the CiOMT genes (−0.49 Pearson correlation with CiDOMT1); however, among the highly coexpressed contigs, a new glucosidase, named CiS6DGD (71.4% amino acid identity to CiDGD), was found. A gene annotated as an acetyl transferase was also identified. Additionally, we noticed that orthologs of known precursor-generating enzymes secologanin synthase (SLS; which catalyzes the final step of secologanin biosynthesis) and tyrosine decarboxylase (TyrDC; predicted to be involved in dopamine biosynthesis) were also tightly coexpressed with CiDOMT1.

No pathway genes from A. salviifolium have been reported and the A. salviifolium transcriptome did not contain orthologs of the previously published CiOMT genes. Therefore, we mined the A. salviifolium transcriptome for orthologs of putative precursor genes TyrDC and SLAS (secologanic acid synthase), which catalyzes the last step of secologanic acid biosynthesis³⁷. The expression profile of these two identified orthologs was highest in leaf buds and roots (Extended Data Fig. 2b), consistent with metabolite data (Fig. 2e). Using TyrDC as bait for coexpression analysis, we identified coexpressed genes that had functional annotations consistent with OMT, dehydrogenase and glycosyl hydrolase activity. Because metabolite data indicated that A. salviifolium uses secologanic acid to form the initial Pictet–Spengler product DAIIA, we predicted that an esterase would not be required for protoemetine biosynthesis in this plant.

Comparative discovery of protoemetine biosynthesis

Transient expression of pathway gene candidates along with infiltration of the starting substrate in N. benthamiana enabled us to successfully deconvolute the complete protoemetine (8) biosynthetic pathway from both C. ipecacuanha and A. salviifolium (Fig. 3 and Extended Data Figs. 3–6). Combinatorial expression of the identified gene candidates was followed by assay of individual genes. On the basis of previous proposals and the chemical logic established for other secologanin-derived natural products such as corynantheal (Cinchona pubescens)^32,33,38, we initially hypothesized that, in C. ipecacuanha, DAII (4a) would be deglycosylated, reduced and deesterified, which would in turn lead to spontaneous decarboxylation to form protoemetine. However, because we observed DAIIA (5a) and 6-O-Me-DAIIA (6a) in C. ipecacuanha (Fig. 2f and Supplementary Fig. 2), we speculated that deglycosylation may happen after deesterification. Indeed, the identified esterase (CiDE) deesterifies the glucosylated intermediate DAII (4a) to yield DAIIA (5a) (Fig. 3 and Extended Data Fig. 4). In A. salviifolium, the pathway directly starts with deesterified DAIIA, derived from the Pictet–Spengler condensation of secologanic acid (2) with dopamine. Thus, both C. ipecacuanha and A. salviifolium use DAIIA (5a) as a protoemetine (8) intermediate.

**Fig. 3: Discovery of C. *ipecacuanha* and A. *salviifolium* protoemetine biosynthetic genes.**

DAIIA (5a) is then 6-O-methylated by CiDOMT1 (previously identified by Nomura) or by AsDOMT1 and AsDOMT2 (Fig. 3). Although an authentic standard is not available, the resulting product has an m/z value and tandem mass spectroemetry (MS²) consistent with 6-O-Me-DAIIA (6b) and this product was also detected in extracts of both species (Fig. 2d and Supplementary Fig. 5). We observed that CiS6DGD/AsS6DGD deglycosylated 6-O-Me-DAIIA, whereas the pathway could not be reconstituted with the previously reported C. ipecacuanha glucosidase, CiDGD (Fig. 3 and Supplementary Fig. 9). The aglycone generated by CiS6DGD/AsS6DGD is subjected to a two-step reduction, catalyzed in both species by a medium-chain reductase, CiDR1 or AsDR1 (Fig. 3 and Supplementary Fig. 10). Spontaneous decarboxylation is then triggered, yielding 10-O-demethylprotoemetine, which is further methylated by CiDPOMT or AsDPOMT1 and AsDPOMT2. Notably, we could only detect trace amounts of the aldehydes 10-O-demethylprotoemetine and protoemetine in N. benthamiana (Extended Data Figs. 4 and 6); instead, we detected the reduced forms 10-O-demethylprotoemetinol (identified by m/z MS² fragmentation; Supplementary Fig. 11) and protoemetinol (confirmed by comparison with an authentic standard), which is expected because of endogenous N. benthamiana aldehyde reductases^39,40. Paralogs of DR (CiDR2 and AsDR2) were also active but resulted in the formation of lower amounts of protoemetinol (Supplementary Fig. 12). Although CiDOMT2 can methylate the 7-hydoxy group of DAI, it did so with low efficiency (Extended Data Fig. 4).

Discovery of species-specific R-epimer pathways

Our metabolite profiling suggested that the R-epimer DAI (4b) is acetylated to form ipecoside (11) in C. ipecacuanha, while DAIA (5b) is methylated on the 6-hydroxy group (and, to a lesser extent, on the 7-hydroxy group) in A. salviifolium to form 6-O-Me-DAIA or 7-O-Me-DAIA (6b and 12b) (Fig. 2d,e and Supplementary Fig. 2b). A gene annotated as a BAHD-type acetyltransferase in the C. ipecacuanha transcriptome that was highly coexpressed with CiDOMT1 (Extended Data Fig. 2a) led to formation of ipecoside (as evidenced by comparison with an authentic standard) when expressed in N. benthamiana along with DAI (4b) (Fig. 4a–c). This gene was, thus, named ipecoside synthase (CiIpS).

Fig. 4: Reconstitution of R-epimer pathways specific to C. *ipecacuanha* and A. *salviifolium* in N. *benthamiana.*

In the A. salviifolium transcriptome, seven closely related class I OMT genes that showed promising expression profiles were tested (Extended Data Fig. 2b). AsDOMT1 and AsDOMT2 methylate DAIIA (5a) on the 6-hydroxy group on route to protoemetine as described above but also catalyze 6-O-methylation and 7-O-methylation of the R-epimer DAIA (5b) (Fig. 4d–f). An authentic standard of 7-O-Me-DAIA was generated through an in vitro Pictet–Spengler reaction of secologanic acid with 4-O-Me-dopamine; the identity of 6-O-Me-DAIA was established through comparison with this authentic standard and also appears to be identical to a highly accumulating compound in A. salviifolium (Supplementary Fig. 5). AsDOMT3–7 are selective for the R-epimer, DAIA (5b), but produce different O-methylated product profiles (Fig. 4f). A sequence alignment of these seven OMT enzymes revealed a high level of overall sequence identity (83.7%) (Supplementary Fig. 15) but only 55% identity at the N terminus (Fig. 4g), a region known to confer substrate and product specificity⁴¹. Indeed, the similarities in substrate and product profiles correlated with the level of identity at the N terminus (Fig. 4f,g). Surprisingly, when all AsDOMT genes were expressed together in N. benthamiana, their product profile differed from the sum profile expected from the expression of single genes; specifically, much higher levels of 6-O-Me-DAIA compared to 7-O-Me-DAIA were observed, a profile similar to that detected in the native plants (Fig. 4e). It is possible that, upon expression in combination, enzymatic activities influence each other, leading to changed product profiles as previously described⁴².

A vacuolar exporter enhances protoemetine biosynthesis

In reconstitution experiments, the majority of infiltrated DAII (4a) starting substrate was not converted into downstream products (Extended Data Fig. 4). In the related monoterpene indole alkaloid pathway, the Pictet–Spengler product produced from secologanin and tryptamine is produced enzymatically by a vacuolar enzyme and then exported into the cytosol by a transporter (CrNPF2.9)⁴³. We hypothesized that exogenously supplied DAI/I 4b/4a may be imported into the vacuole by N. benthamiana transporters^44,45, rendering these starting materials inaccessible to the downstream cytosolic pathway enzymes. To test whether CrNPF2.9 could export the protoemetine precursor DAII (4a), we expressed CrNPF2.9 with C. ipecacuanha protoemetine biosynthetic genes in N. benthamiana. We observed an almost 12-fold relative increase in protoemetinol levels, suggesting that this vacuolar exporter could transport DAII (4a) out of the vacuole (Extended Data Figs. 4 and 7a). No increase was observed when CrNPF2.9 was expressed in combination with A. salviifolium protoemetine pathway genes (Extended Data Fig. 6), suggesting that this transporter does not recognize DAIIA (5a). Expression of CrNPF2.9 with CiIpS did not lead to an increase of the R-stereoisomer ipecoside, suggesting that this transporter is specific for the S-stereoisomer (Extended Data Fig. 7b). Furthermore, expression of the protoemetine biosynthetic genes with uncoupled secologanin and dopamine (as opposed to DAII 4a) only yielded protoemetinol when CrNPF2.9 was included (Extended Data Fig. 7c). We hypothesize that the glycosylated secologanin would get imported into the vacuole, where the mildly acidic conditions would facilitate the nonenzymatic formation of DAI/I (Extended Data Fig. 7d). Although we tested numerous candidates, the native vacuolar exporters of C. ipecacuanha and A. salviifolium remain to be discovered.

Nuclear-localized glucosidases deglycosylate R-epimers

Glucosidase CiDGD deglycosylates the 1R-derived stereoisomer ipecoside (11)³³ (Fig. 5a,b). In contrast, CiS6DGD, which we showed to be involved in protoemetine biosynthesis, did not turn over ipecoside (Fig. 5b). Leaf disk assays of CiDGD and Ci6SDGD with substrates DAI/I (4b/4a), 7-O-Me-DAI/I (18b/18a), DAI/IA (5b/5a) and 6-O-Me-DAIIA (6a) suggest that CiDGD has broad substrate specificity, whereas the more selective Ci6SDGD most efficiently deglycosylates 6-O-Me-DAIIA 6a, which is on pathway to protoemetine 8 (Extended Data Fig. 8a). Analogously, the A. salviifolium glucosidases AsDGD1 and AsDGD2 or AsS6DGD were similarly assayed with the A. salviifolium substrates DAI/IA 5b/5a, 7-O-Me-DAI/IA 12b/12a, 6-O-Me-DAIIA 6a and 6-O-Me-DAIA 6b. AsDGD1 and AsDGD2 consumed 7-O-Me-DAIA and 6-O-Me-DAIA entirely (Fig. 5d and Extended Data Fig. 8b), while AsS6DGD only consumed the protoemetine pathway intermediate 6-O-Me-DAIIA (Extended Data Fig. 8b). In vitro competition assays using recombinant glucosidases were consistent with the conclusion that CS6DGD and AsS6DGD were more selective for the protoemetine pathway intermediate, whereas CiDGD and AsDGD2 had broader substrate specificity (Supplementary Figs. 16–18). Thus, both C. ipecacuanha and A. salviifolium seem to have a protoemetine pathway-specific glucosidase along with a glucosidase with broader substrate specificity. The substrate specificity of CiS6DGD and AsS6DGD requires that DAIIA is subjected to 6-O-methylation before deglycosylation. In contrast, CiDGD, AsDGD1 and AsDGD2 deglycosylate DAII(A) directly, which prevents formation of protoemetine (Supplementary Fig. 9).

**Fig. 5: Deglycosylation of R-epimer-derived products by nuclear-localized glucosidases.**

CiDGD, AsDGD1 and AsDGD2 each contain a C-terminal bipartite nuclear localization sequence (NLS) (as predicted with DeepLoc2.0)⁴⁶ (Supplementary Table 1). When these genes were fused to eYFP, each colocalized with an mCherry–NLS marker when expressed in N. benthamiana leaves⁴⁷ (Fig. 5e and Supplementary Figs. 19 and 20). C-terminal tagged AsDGD1 and AsDGD2 showed localization across the entire nucleus, whereas the N-terminal tagged fusions were localized to a smaller compartment within the nucleus, which could be aggregates, as previously proposed for the glucosidase involved in monoterpene indole alkaloid biosynthesis⁴⁸ (Supplementary Figs. 19 and 20). Taken together, these results indicate that both species derivatize R-epimers in a species-dependent manner and contain highly active nuclear-localized glucosidases with relatively broad substrate specificity. The specificity of CiS6DGD and AsS6DGD ensures that 6-O-Me-DAIIA rather than DAI/I(A) is deglycosylated by these dedicated protoemetine pathway-specific glucosidases. The protoemetine-specific glucosidase CiS6DGD, which is highly similar to CiDGD, also showed nuclear localization (Supplementary Figs. 19 and 20). The protoemetine-specific glucosidase from A. salviifolium lacked the NLS and was localized to both the nucleus and cytosol (Supplementary Figs. 19 and 20). In comparison, CiIpS, CiDE and CiDR1 each appeared to be localized to the cytosol (Supplementary Fig. 21).

Some biosynthetic genes may have evolved independently

Having elucidated ipecac alkaloid biosynthesis in these two distantly related plants, we performed phylogenetic comparisons of the identified enzymes (Fig. 6, Extended Data Fig. 9 and Supplementary Figs. 22 and 23). These analyses clearly suggest that A. salviifolium (Cornales) and C. ipecacuanha (Gentianales) enzymes evolved independently. AsDOMT enzymes belong to the class I Mg²⁺-dependent caffeoyl CoA 3-OMT family⁴⁹ (Fig. 6a). AsDOMT1–AsDOMT7 form a separate subclade, suggesting that the different substrate and product profiles observed among these enzymes (Fig. 4) likely arose through tandem gene duplication and subfunctionalization. All other identified OMT enzymes in ipecac biosynthesis are class II Mg²⁺ independent. These proteins form two well-separated clades; one clade contains AsDPOMTs and the other contains CiDOMTs and CiDPOMT. Parallel, independent evolution is inferred if phylogenetic analysis reveals that two enzymes, despite the same fold and the same enzymatic activity, are found in separate clades of the tree (that is, each protein clusters more closely together with orthologs from their respective phylogenetic groups rather than with the other enzyme in question)⁵⁰. The positions of the A. salviifolium and C. ipecacuanha proteins on the phylogenetic tree suggest independent (parallel) evolution. All three C. ipecacuanha OMTs are closely related, suggesting that these genes arose through tandem duplications and subsequent neofunctionalization to catalyze O-methylation of either DAIIA or 10-O-demethylprotoemetine.

Fig. 6: Some ipecac alkaloid biosynthetic enzymes appear to have arisen independently in C. *ipecacuanha* and A. *salviifolium.*

The protoemetine-specific CiS6DGD belongs to the GH-1 family of glucosidases (Fig. 6b), whereas AsS6DGD is a member of the GH-3 family (Fig. 6b). Thus, CiS6DGD and AsS6DGD, which have different protein folds, represent a case of convergent, independent evolution^2,50. CiDGD and CiS6DGD are closely related, suggesting that these enzymes evolved through tandem gene duplications and subfunctionalization. The glucosidases that have broad specificity (CiDGD, AsDGD1 and AsDGD2) belong to the same family of glucosidases (GH-1). However, the phylogenetic analysis, which revealed these A. salviifolium and C. ipecacuanha proteins to be in separate clades, also suggests parallel, independent evolution. Lastly, the medium-chain alcohol dehydrogenases DRs share the same protein fold but, for the same reasons as above, phylogenetic analysis also suggests that these enzymes evolved independently through means of parallel evolution (Extended Data Fig. 9).

Discussion

Here, we report the discovery of the protoemetine pathway in two distantly related plants, C. ipecacuanha and A. salviifolium. We show that C. ipecacuanha uses the monoterpenoid precursor secologanin (1) while A. salviifolium uses secologanic acid (2). This is consistent with monoterpenoid indole alkaloid biosynthesis in the Gentianales and Cornales clades, where tryptamine is condensed with secologanin or secologanic acid, respectively, to generate strictosidine or strictosidinic acid^21,51.

Strictosidine (3S) is stereoselectively synthesized from secologanin and tryptamine by a well-characterized vacuolar localized Pictet–Spenglerase enzyme (strictosidine synthase)⁴⁸. The 3R isomer of strictosidine (vincoside) does not appear to be present in these plants. In contrast, all ipecac alkaloid producing plants contain products derived from both 1S and 1R stereoisomers of the initial Pictet–Spengler product (DAI 4b (1R) and DAII 4a (1S) in C. ipecacuanha and DAIA 5b (1R) and DAIIA 5a (1S) in A. salviifolium). Therefore, these plants must generate both 1R and 1S Pictet–Spengler products. Nonenzymatic Pictet–Spengler reactions using dopamine have been well established to occur under physiologically relevant conditions; for example, the presence of phosphate can facilitate efficient coupling of dopamine with aldehydes⁵². We observed that nonstereoselective formation of DAI/I/(A) occurs when (1) dopamine and secologanin or secologanic acid are incubated in aqueous buffer at a pH value consistent with the environment of the vacuole; (2) dopamine and secologanin or secologanic acid infiltrate into N. benthamiana leaves; and (3) dopamine infiltrates into C. roseus flowers. Analogous experiments using tryptamine instead of dopamine resulted in far lower levels of product, suggesting that the dopamine substrate is highly activated. The known plant Pictet–Spenglerases, strictosidine synthase and norcoclaurine synthase, are localized in the vacuole^36,53; however, because these enzymes have optimal catalytic efficiency at neutral pH, the acidic environment of the vacuole is not required for the enzymatic reaction^54,55. However, for a nonenzymatic reaction, the slightly acidic environment of the vacuole could be crucial, even for the highly activated, electron-rich dopamine substrate. We further showed that coexpression of the C. roseus vacuolar strictosidine exporter CrNPF2.9 with C. ipecacuanha protoemetine biosynthetic genes greatly enhanced the levels of the protoemetine product and facilitated formation of the protoemetine-derived product from starting substrates secologanin and dopamine (Extended Data Fig. 7). Lastly, we could not detect the presence of enzymatic Pictet–Spengler activity in crude C. ipecacuanha and A. salviifolium extracts (Supplementary Fig. 8). Collectively these observations support the notion that DAI/I(A) formation occurs nonenzymatically and within the plant vacuole in these pathways. However, the involvement of enzymes in these condensation reactions cannot be definitively ruled out.

Both C. ipecacuanha and A. salviifolium evolved nearly identical chemistry to generate the ipecac alkaloid protoemetine. Secologanic acid is an iridoid, a class of natural product that is ancestral to all Asterids⁵⁶. Iridoid pathway genes from C. acuminata and C. roseus belonging to the phylogenetically distant asterid lineages Cornales and Gentianales, respectively, have been shown to be orthologs with high levels of sequence identities^21,22,57. Thus, the biosynthesis of the secologanic acid precursor is likely also conserved between A. salviifolium (Cornales) and C. ipecacuanha (Gentianales). The addition of a methyl ester to secologanic acid is a chemical innovation that is not present in A. salviifolium, consistent with other members of Cornales^21,51. Analogously, the esterase (CiDE) that eventually removes this methyl ester is present only in C. ipecacuanha.

After the Pictet–Spengler reaction that condenses secologanin or secologanic acid with dopamine, CiDOMT (C. ipecacuanha, class II OMT) and AsDOMT (A. salviifolium, class I OMT) methylate DAII(A). CiDOMT and AsDOMT both methylate DAIIA but are phylogenetically distant and share only partially the same structural fold (Fig. 6a). Thus, these proteins most likely represent an example of parallel evolution, using the terminology established by Weng and Noel⁵⁰. Parallel evolution refers to an event in which ancestor enzymes with a shared structural lineage have independently evolved to have the same biochemical activity. In the next pathway step, CiS6DGD and AsS6DGD, both proposed to deglucosylate 6a in protoemetine biosynthesis, are members of different structural classes of glucosidase (GH-1 versus GH-3) (Fig. 6b). Using the terminology established by Weng and Noel⁵⁰, two enzymes with distinct folds catalyzing the same reactions would be an example of convergent evolution. Then, both C. ipecacuanha and A. salviifolium use a medium-chain alcohol dehydrogenase to catalyze reduction of this deglycosylated product. However, their distant positions on the corresponding phylogenetic tree (Extended Data Fig. 9) suggest that these reductases have evolved independently through parallel evolution. The reduced product is finally methylated by CiDPOMT (C. ipecacuanha) and AsDPOMT (A. salviifolium), which are both class II OMTs. Phylogenetic analysis also suggests that these CiDPOMT and AsDPOMT evolved through parallel evolution because these proteins are found in separate clades in the phylogenetic tree (Fig. 6a). Collectively, phylogenetic analysis suggests that the four pathway steps from DAII(A) to protoemetine appear to be examples of either convergent or parallel evolution. Our phylogenetic analyses that support a model of independent evolution are consistent with analyses from other examples of independent evolution in plant specialized metabolism, such as glucosinolates³, benzoxazinoids⁴, caffeine^5,6,7, cannabinoids⁸ and cardenolides^9,10,11.

Although protoemetine biosynthesis indicates the same chemical outcome, there appear to be striking species-specific evolutionary strategies by which enzymes were recruited to these two protoemetine pathways. In C. ipecacuanha, all involved OMTs (CiDOMT1, CiDOMT2 and CiDPOMT) are closely related and likely arose through gene duplication and subsequent neofunctionalization. This would exemplify the ‘forward hypothesis’ of evolution, in which a pathway evolves from the first biosynthetic enzyme onward through means of tandem gene duplication and neofunctionalization^58,59. Conversely, in A. salviifolium, enzymes from different OMT classes appear to have been independently recruited to ipecac alkaloid biosynthesis (AsDOMT is a class I OMT; AsDPOMT is a class II OMT). This would exemplify the ‘patchwork hypothesis’ of evolution, in which distinct ancestral enzymes are independently recruited to a pathway.

Analogously, C. ipecacuanha and A. salviifolium each evolved glucosidases with broad substrate specificity (DGDs) and glucosidases that appear to be specific to the intermediate on pathway to protoemetine biosynthesis (S6DGDs). While CiDGD and AsDGD are GH-1 glucosidases, CiS6DGD and AsS6DGD are GH-1 and GH-3 glucosidases, respectively. Thus, in C. ipecacuanha, CiDGD and CiS6DGD likely evolved through tandem gene duplication and subfunctionalization (forward hypothesis), whereas, in A. salviifolium, AsS6DGD convergently evolved this function upon recruitment from the GH-3 family, a group of glucosidases with a different fold not commonly associated with specialized metabolism^60,61 but rather with cell wall biosynthesis (patchwork hypothesis)^62,63,64,65.

Within Gentianales, the downstream steps of the ipecac alkaloid pathway appear to have evolved independently from the related monoterpene indole alkaloid pathway enzymes. Although some of the biosynthetic steps of these two pathways have similar chemistry (for example, the reductase DCS from Cinchona spp.³⁸ and the OMT from vinblastine biosynthesis⁶⁶) the enzymes that catalyze the respective steps (DR and OMT) do not cluster together with known monoterpene indole alkaloid biosynthesis enzymes but instead form sister clades (Fig. 6a and Extended Data Fig. 9).

While both C. ipecacuanha and A. salviifolium convert the S-epimer to protoemetine, the R-epimers are derivatized in a chemically different and species-dependent manner through a simple and shorter ‘shunt’ pathway. It is likely that the R-epimers ipecoside (C. ipecacuanha), 6-O-Me-DAIA and 7-O-Me-DAIA (A. salviifolium) accumulate in the vacuole because this is the typical storage ___location in plants for glycosylated specialized metabolites⁶⁷. These products would, therefore, be separated from the nuclear-localized glucosidases (DGDs) that act upon these substrates. Deglycosylation of ipecoside and 6/7-O-Me-DAIA leads to a dialdehyde moiety that is highly reactive^48,68,69. The R-epimer-derived ipecoside, 6-O-Me-DAIA and 7-O-Me-DAIA could be released from the vacuole upon tissue damage, only then coming into contact with the nuclear-localized glucosidases to generate a reactive and, therefore, toxic defense molecule, as previously hypothesized for ipecoside³³. Despite the species-specific chemical derivatization (O-methylation versus acetylation) the R-epimers may serve similar ecological functions as defensive agents. Mechanisms in which the enzyme is spatially separated from its substrates to avoid the constant accumulation of potentially toxic compounds were first described for glucosinolates^70,71. Instances in which a glucosidase is located in the nucleus while the substrate is stored in the vacuole have also been described for monoterpenoid indole alkaloids, saponins and secoiridoids^48,72,73. Specialized metabolites stored in inactive form that are activated upon tissue damage have recently been referred to as phytoavengins⁷⁴. Overall, this comparative pathway elucidation of ipecac alkaloid biosynthesis highlights the diversity in evolutionary strategies to evolve chemically complex molecules and provides a foundation for metabolic engineering of these biologically active molecules.

Methods

Plant material and sampling

C. ipecacuanha plantlets were grown in vitro on propagation medium (1× Murashige and Skoog (MS) medium including vitamins, 3% sucrose, 3 mg L⁻¹ 6-benzylaminopurine, 10 µg L⁻¹ 1-naphthaleneacetic acid and 8 g L⁻¹ agar, pH 5.7). Upon arrival, plantlets were transferred to root induction medium (0.75× MS medium including vitamins, 3% sucrose, 0.5 mg L⁻¹ 1-naphthaleneacetic acid and 8 g L⁻¹ agar, pH 5.7). After approximately 6 weeks, roots formed and the regenerated plants were transferred to soil. For tissue-specific metabolite profiling and RNA-seq, plants were harvested 4 months after transfer to soil. A. salviifolium plants were received as cuttings from the Botanical Garden of Ghent University and rooted on rockwool. After 6 weeks, rooted plants were transferred to soil. For tissue-specific profiling, plants were harvested 14 months after transfer to soil. Both species were grown in a temperature-controlled and light-controlled greenhouse with the following conditions: 12-h light–dark cycle at 28–30 °C and 24–26 °C, respectively, and 70–80% humidity. For tissue-specific analyses, three plants from each species were macrodissected as shown in Supplementary Fig. 1. Tissues were immediately flash-frozen and ground in liquid nitrogen using an IKA A11 basic analytical mill or mortar and pestle. The frozen finely ground powder was kept at −80 °C until further processing.

Metabolite extraction for metabolite profiling

Finely ground material was extracted with 100% methanol supplemented with 10 mg L⁻¹ caffeine as internal standard. The volume of methanol was normalized to fresh tissue weight (1:10 mg:µl). Samples were vortexed for 1 min and sonicated for 5 min at room temperature. After 1 h of incubation at room temperature, samples were centrifuged for 15 min at 18,000g and supernatants were filtered through Fisher PTFE syringe filters (0.22 μm). Filtered extracts were diluted in 80% methanol containing 0.1% formic acid and 1:10 and 1:50 dilutions were analyzed by untargeted ultrahigh-performance liquid chromatography–mass spectrometry (UPLC–MS). C. ipecacuanha samples were analyzed using UPLC–MS method 1 and A. salviifolium samples were analyzed using UPLC–MS method 2. For comparable traces shown in Fig. 2, representative samples of each plant were analyzed in parallel with UPLC–MS method 3.

UPLC–MS/MS methods

Method 1 was used for C. ipecacuanha tissue-specific metabolite profiling. An Elute LC system (Bruker Daltonics) was coupled to an Impact II high-resolution quadrupole time-of-flight MS instrument (Bruker Daltonics). An Acquity UPLC BEH C18 (2.1 × 50 mm, 1.7 μm; 130 Å) column (Waters) was set at 40 °C and 0.6 ml min⁻¹ flow rate and 2 µl of samples were injected. The mobile phase was A:B where A was water with 0.1% formic acid and B acetonitrile with 0.1% formic acid. The gradient was as follows: 5% B at 1 min to 8% B at 3 min, to 13% B at 5 min and to 30% at 8 min. Then the column was flushed at 100% B until 9.8 min and re-equilibrated to 5% B until 12 min. Method 2 was used for A. salviifolium tissue-specific profiling and was identical to method 1 with the exception of the column temperature, which was set to 35 °C. Method 3, used for all other experiments, was identical to method 2 but used an UltiMate 3000 UPLC system (Thermo Fisher Scientific). For all methods, ionization was performed using pneumatic-assisted electrospray ionization (ESI⁺) with 4,500 V (methods 1 and 2) or 3,500 V (method 3) of capillary voltage, a 500-V end-plate offset and a nebulizer pressure of 2.5 bar, with nitrogen at 250 °C and a flow of 11 L min⁻¹ as the drying gas. Acquisition was performed at 12 Hz following a mass range from 80 to 1,000 m/z with data-dependent MS/MS, an active exclusion window of 0.2 min and a reconsideration threshold of 1.8-fold. Fragmentation was triggered on an absolute threshold of 400 and limited to a total cycle time range of 0.5 s. For collision energy, the stepping option model (from 20 to 50 eV) was used. At the start of each run, the m/z values were recalibrated using the expected cluster ion m/z values of a direct source infusion of sodium formate–isopropanol solution. To prevent the contamination of the MS by injection peaks and salt, the first minute of each run was run isocratically at 5% B and redirected to waste.

MS data analysis

Data was converted to mzml or mzxml format and imported to MZmine (versions 3.6.0 and 4.1.0)⁷⁶. Extracted ion chromatogram (EIC) traces and MS² data of compounds of interest were exported from MZmine. Peak areas were calculated using the MZmine processing wizard and exported. Peak areas of compounds of interest were normalized to the internal standard caffeine and converted to intensity per second. Further data analysis and construction of graphs were performed in GraphPad Prism version 10.2.3 and 10.3.0 for Mac OS X. For tentative identification of molecular structures and analysis of MS² fragmentation pattern, SIRIUS version 4 was used^77,78,79. Chemical structures were drawn in ChemDraw Professional 20.1.0.112.

Nonenzymatic Pictet–Spengler reaction

For coupling in N. benthamiana, a solution of 1 mM dopamine hydrochloride 3 or 1 mM tryptamine hydrochloride and 1 mM secologanin 1 or secologanic acid 2, respectively, was infiltrated into leaves of 4-week-old plants (grown as described below). A leaf sample was taken immediately after infiltration for a control and snap-frozen. For coupling in C. roseus, petals of freshly opened flowers of 4-month-old plants (grown in a controlled growth chamber at 16 and 8 h of light and dark, 23 °C, 40–50% humidity) were infiltrated with 1 mM dopamine hydrochloride 3 or 1 mM tryptamine-d₅ hydrochloride solution. Samples were harvested after 24 h. Control samples were taken after 24 h from flower petals infiltrated with water. Metabolite extraction and analysis were performed as described below for N. benthamiana.

Enzyme activity tests on native plant protein extracts

Protein extraction and activity assays were performed on the basis of previously described methods^18,25. Young leaves of A. salviifolium or C. ipecacuanha were freshly harvested, snap-frozen, ground in liquid nitrogen with a mortar and pestle and immediately extracted in 2 ml of ice-cold protein extraction buffer (100 mM Tricine-HCl pH 7.5, 10 mM β-mercaptoethanol and 3 mM EDTA) per 1 g fresh weight (FW). Extraction took place by gentle shaking at 4 °C for 30 min. Extracts were then centrifuged at 18,000g and 4 °C for 15 min and the supernatants were desalted into 100 mM Tricine-HCl pH 7.5 using Zeba spin desalting columns with a 7-kDa molecular weight cutoff (Thermo Scientific) according to the manufacturer’s instructions. Desalted extracts were then immediately used in activity assays. A part of each protein extract was heated for 10 min at 98 °C and centrifuged at 18,000g for 25 min; the supernatant served as the boiled enzyme control. Activity assays were performed in triplicates from the same desalted protein extracts. Assays (200 µl of total volume per replicate) contained 1 mM δ-gluconolactone, 1 mM dopamine hydrochloride 3, 1 mM secologanin 1 (C. ipecacuanha assay) or secologanic acid 2 (A. salviifolium assay) and desalted protein extracts or boiled extracts as negative controls. Assays to detect OMT activity in A. salviifolium contained 1 mM δ-gluconolactone, 1 mM mix of DAIIA 5a and DAIA 5b (corresponding to ~500 µM per epimer), 1 mM MgCl₂ and 1 mM S-(5′-adenosyl)-l-methionine (SAM) chloride dihydrochloride. Assays to detect esterase activity in C. ipecacuanha extracts contained 1 mM δ-gluconolactone and 1 mM mix of DAII 4a and DAI 4b mix (corresponding to ~500 µM per epimer). Assays were pipetted on ice and then incubated with gentle shaking at 30 °C. At indicated time points, aliquots (10 µl) were removed and immediately flash-frozen in liquid nitrogen (time point zero was taken immediately upon pipetting and before transfer to the incubator). For metabolite analysis, 90 µl of 60% methanol containing 0.1% formic acid and 1 mg L⁻¹ caffeine as the internal standard was added to each aliquot and precipitants were pelleted by centrifugation at 18,000g at room temperature for 30 min. Supernatants were analyzed with UPLC–MS method 3.

RNA extraction and RNA-seq

Total RNA was extracted using the RNeasy Plant Mini Kit (Qiagen) from replicate 1 of the same tissue material that was used for metabolite profiling according to the manufacturer’s instructions, including on-column DNAse digest. RNA concentrations and purity were determined with a Nanophotometer N60 (Implen). Samples of sufficient concentration and purity were shipped to Novogene where mRNA library preparation and sequencing were performed according to the company’s standard protocol for mRNA-seq. RNA integrity and quantitation were assessed using the RNA Nano 6000 assay kit of the Bioanalyzer 2100 system (Agilent Technologies). All samples were above the required minimum RNA integrity number. Sequencing was performed on an Illumina NovaSeq 6000 PE150 platform with a data output target of 9 GB of raw data. Adaptor cleaved raw data were provided as FASTQ files.

Transcriptome analysis and candidate selection

All data processing was performed in-house. Reads were quality-checked with FastQC and trimmed using Trimmomatic⁸⁰. Transcriptomes were assembled by rnaSPAdes using the trimmed reads of all tissues combined as an input⁸¹ with default settings, except k-mer size was adjusted to 55, 77 and 99 to discriminate isoforms with high sequence identity. The resulting assemblies were assessed with Busco for the Eudicots lineage and were 95.6% complete for C. ipecacuanha and 97.2% complete for A. salviifolium⁸². FastQC (Galaxy version 0.73), Trimmomatic (Galaxy version 0.38.1) and rnaSPAdes (Galaxy version 3.15.4) were all run on an in-house Galaxy server⁸³. Functional annotation of transcripts was performed on OmicsBox (Biobam) using the SwissProt 2021 database (basic local alignment search tool (BLAST) parameters: E value, 1.0 × 10⁻³; number of hits, 10; word size, 6; low-complexity filter, on; number of threads, 40; high-scoring pair length cutoff, 33) and eggNOG-mapper^84,85. Reads were mapped to transcriptomes using CLC Genomics workbench 21.0.4 (Qiagen) with the following parameters: mismatch cost, 2; insertion cost, 3; deletion cost, 3; length fraction, 0.8; similarity fraction 0.85; autodetect paired distances, on; maximum number of hits for a read, 10. Counts per million (CPM) normalized by the trimmed mean of M values were used for downstream analyses.

To identify known C. ipecacuanha pathway genes, the published sequences (National Center for Biotechnology Information (NCBI) AB455576, AB527082, AB527083, AB527084 and AB576187) were BLASTed (BLASTn) against our transcriptome^32,33,34. The level of pairwise sequence identity at the nucleotide level between the published sequences and the amplified sequences was 96.8% for CiDGD (previously called IpeGlu1), 97.2% for CiDOMT1 (previously called IpeOMT1), 98.7% for CiDOMT2 (previously IpeOMT2) and 98.5% for CiDPOMT (previously IpeOMT3). The closest homologous sequence to another published OMT³⁴ was also CiDOMT2 (98.8% pairwise identity). These observed small differences in sequence identity between published and newly cloned sequences can be attributed to single-nucleotide polymorphisms of different plant sources. The expression profile of CiDOMT1 was used as bait for Pearson correlation (Extended Data Fig. 2).

To identify candidates for the missing esterase and reductase, C. pubescens CpDCE (MW456556.1) and CpDCS (MW456554.1)³⁸ were BLASTed (tBLASTn) against the transcriptome, which resulted in 9 contigs with 38–44% identity and 15 contigs with 55.4–62% identity, respectively. The BLAST hits were then filtered for high correlation with the CiDOMT1 tissue-specific expression profile (Pearson correlation > 0.85) and high absolute expression level in young leaf and rhizome (>50 CPM). Further mining the list of coexpressed and highly expressed contigs for relevant functional annotations revealed Ci6SDGD and CiIpS. For A. salviifolium, no pathway genes were previously published and BLASTing CiDOMTs against the transcriptome did not yield any orthologs with high identity. We, therefore, identified a TyrDC ortholog whose expression pattern matched the tissue-specific metabolite profiling and used it as bait for coexpression analysis. Among the highly coexpressed candidates (Pearson correlation > 0.75, CPM in leaf buds and/or roots > 50), we selected those with functional annotations consistent with OMTs, dehydrogenases and glycosyl hydrolases for screening. After positive screening results for AsDOMTs, three additional highly homologous sequences with higher root-specific expression (AsDOMT2, AsDOMT6 and AsDOMT7) were included in pathway reconstitution experiments.

Gene cloning

Complementary DNA (cDNA) was prepared from total RNA of A. salviifolium leaf buds and roots and C. ipecacuanha young leaves and rhizome (extracted as described above) using the RevertAid first-strand cDNA synthesis kit (Thermo Scientific) according to manufacturer’s instructions. AsDOMT5 and AsDPOMT2 could not be amplified and were, therefore, obtained as synthetic sequences from Twist Biosciences. Coding sequences were amplified with the Q5 high-fidelity 2× master mix (New England Biolabs) using cDNA or synthetic genes as templates and gene-specific primers containing overhangs for In-Phusion cloning (Supplementary Table 5). Amplified sequences were gel-purified using the Zymoclean gel DNA recovery kit (Zymo Research) and cloned using the 5× In-Fusion snap assembly master mix (TaKaRa Bio). For expression in N. benthamiana, coding regions were inserted into a modified 3Ω1 vector (containing UBQ10 promoter and terminator from Solanum lycopersicum⁸⁶) previously digested with BsaI-HF v2 (New England Biolabs). For expression in Escherichia coli, coding sequences were cloned into pOPINF⁸⁷ previously digested with KpnI-HF and HindIII-HF (New England Biolabs). Heat-shock-competent E. coli TOP10 cells were transformed and grown overnight in a 37 °C incubator on Luria–Bertani agar plates containing the respective antibiotics. Plasmids were isolated from overnight cultures of single colonies using the Wizard Plus SV Miniprep DNA purification system kit (Promega). Inserted sequences were confirmed by Sanger sequencing.

Sequences for subcellular localization were amplified and purified as described above but using the previously cloned 3Ω1 constructs as PCR templates (described above) and gene-specific primers with overhangs compatible with Golden Gate cloning for N-terminal or C-terminal eYFP or mCerulean3 fusion proteins (Supplementary Table 5). Constructs (level 1) were assembled using BsaI (New England Biolabs), T4 DNA ligase (New England Biolabs), the pDGB3_ɑ1 vector⁸⁸, pUPD_pSIUbq10 and pUPD_TeSlUbq10^86,89 and the gel-purified PCR products. As6SDGD and CiIpS contained a BsaI restriction site that was removed by introducing a silent mutation in an overlapping PCR before assembly. Assembly consisted of 50 cycles of 5 min at 37 °C followed by 5 min at 16 °C and was stopped by incubation at 65 °C for 10 min. Additionally, CrNPF2.9 was cloned through the same Golden Gate cloning procedure described above into pDGB3_ɑ1 (without eYFP tag), whereas the silencing repressor gene p19 was obtained as pUPD_p19 (Addgene, GB0038) and then cloned as above into pDGB3_ɑ1.

Transient expression in N. benthamiana

In all cases, the Agrobacterium tumefaciens GV3101 strain was used and cultured at 28 °C in YEB medium containing rifampicin and gentamycin and the appropriate antibiotic for plasmid selection. N. benthamiana plants used for infiltration were 3–4 weeks old (grown in a greenhouse with 16 and 8 h of light and dark at 23–26 °C and 16–22 °C, respectively, and 40–70% humidity). Cells were transformed through electroporation, recovered in YEB without antibiotics and incubated for 2 days on YEB plates containing antibiotics. Single colonies were confirmed by colony PCR and grown in liquid YEB for 24 h. From these cultures, glycerol stocks were prepared and stored at −80 °C. For agroinfiltration, cells were grown on YEB plates similar to a previously described method with modifications⁹⁰. Cells from glycerol stocks were spread on YEB plates containing antibiotics and 100 µM acetosyringone and grown for 24 h until a visible layer of bacteria appeared. The bacteria were transferred to 1–2 ml of infiltration medium (10 mM MES, 10 mM MgCl₂ and 100 μM acetosyringone, pH 5.7) and gently resuspended. The optical density at 600 nm (OD₆₀₀) was measured in 1:10 dilutions using an Implen OD600 DiluPhotometer.

For all pathway reconstitution experiments, strains were mixed and diluted in infiltration buffer to OD₆₀₀ = 0.1 per strain. A strain harboring a construct with the p19 gene was coinfiltrated in all cases. The culture mixtures were infiltrated into N. benthamiana leaves and plants were kept for 16 h in the dark and subsequently grown under grow light (16 and 8 h of light and dark). Replicates were from three individual plants. Indicated substrates were infiltrated as 500 µM aqueous solutions of epimeric mixtures (described below) 3 days after infiltration. Leaf material was harvested 24 h after substrate infiltration by flash-freezing in tubes containing metal beads. In the case of infiltration of uncoupled secologanin and dopamine substrates, leaf material was harvested 48 h after substrate infiltration to allow time for coupling.

For analysis of N. benthamiana leaves by confocal laser scanning microscopy, A. tumefaciens culturing and infiltration was performed as above but at OD₆₀₀ = 0.1–0.3 per strain. To confirm subcellular localization, each strain harboring an eYFP or mCerulean3 fusion construct was coinfiltrated with a strain harboring a construct with free mCherry or NLS–mCherry as fluorescent markers for cytosol or nuclear localization, respectively⁴⁷, along with a strain for expression of p19. Leaf tissue was analyzed 2–3 days after infiltration.

Leaf disk assays

To test glucosidase activity toward different substrates, a leaf disk assay was used as previously described with modifications^91,92,93. Growth and infiltration of A. tumefaciens were performed as described above and replicates were from three individual plants. Then, 3 days after agroinfiltration, 1-cm leaf disks were cut using a leaf puncher and incubated with 200 µl of substrate solution in 50 mM HEPES buffer pH 7.5 in 48-well plates. Substrate solutions were prepared as master mixes so that concentrations were identical in each well. Substrate concentrations were 400 µM total, corresponding to 200 µM per epimer for chemically produced epimeric substrate mixtures DA(I)I, DAI(I)A, 7-O-Me-DAI(I) and 7-O-Me-DAI(I)A. Commercially available ipecoside was used at 200 µM. The reactions of enzymatically produced substrates (see below) were monitored by UPLC–MS and concentrations were estimated at 250 µM. Plates were sealed with parafilm to avoid evaporation and incubated for 24 h under growth lights (16 and 8 h of light and dark).

Metabolite extraction from N. benthamiana

Leaf material was ground using two 4-mm metal beads and a TissueLyser (Qiagen) with precooled adaptors and extracted with 100% methanol containing 0.1% formic acid and 2 mg L⁻¹ caffeine as the internal standard. For substrate infiltrated leaves, 150 µl per 100 mg leaf material was used. In the case of leaf disk assays, 50 µl was used per single leaf disk. Samples were sonicated for 10 min, incubated on a rotator for 20 min and centrifuged at 18,000g for 15 min. The supernatants were mixed 1:2 with MQ H₂O to improve shape of early eluting peaks and filtered through a 0.45-µm low-binding hydrophilic PTFE filter plate (MultiScreen Solvinert 96, Merck-Millipore) into a 96-well microtiter plate (SureSTART WebSeal, Thermo Scientific) according to the manufacturer’s instructions. Plates were sealed with Rapid Slit Seal (BioChromato) and immediately analyzed with UPLC–MS method 3.

Confocal laser scanning microscopy

For confocal laser scanning microscopy, the leaf disks were put on a glass slide, mounted with water and covered with a coverslip. Fluorescence was observed and imaged with a W Plan-Apochromat ×40 1.0 differential interference contrast M27 water objective on a cLSM 880 (both Zeiss) equipped with two lasers for excitation of the two different fluorophores. mCherry was excited at 543 nm with a helium–neon laser and emission was filtered between 600 and 651 nm. eYFP was excited at 514 nm with an argon laser and emission was filtered between 525 and 561 nm. mCerulean3 was excited at 458 nm with an argon laser and emission was filtered between 466 and 490 nm. The micrographs with eYFP were taken sequentially in two tracks for each image. In the first track, mCherry emissions were captured along with the transmitted light image. In the second track, eYFP emission was captured. ZEN black 2.1 V.14.0.18.201 was used as software (Zeiss). Images were adjusted and processed with ImageJ software⁹⁴.

Recombinant protein production and purification

Expression and purification of AsDOMT3, CiDOMT1 and CiDE were performed as previously described with modifications⁹⁵. Briefly, E. coli SoluBL2(DE3) cells were transformed by heat shock with pOPINF constructs. Precultures were inoculated from single colonies, grown overnight at 37 °C and used to inoculate 100 ml of 2× YT medium. Cultures were grown at 37 °C until an OD₆₀₀ of 0.5–0.6, cooled to room temperature and induced with 0.2 mM IPTG. After induction, the cultures were grown at 18 °C overnight and harvested by centrifugation the next day. Cell pellets were lysed using B-PER complete bacterial protein extraction reagent (Thermo Scientific) supplemented with EDTA and subsequently centrifuged according to the manufacturer’s instructions. Supernatants were incubated with gentle shaking in Falcon tubes with 250 µl of Ni-NTA agarose (Qiagen) for 1 h at 4 °C to allow binding of His-tagged proteins. Slurry was pelleted gently by centrifugation at 1,000g for 30 s. The supernatant was removed and the slurry was washed three times with ice-cold wash buffer (50 mM Tris-HCl pH 8, 50 mM glycine, 5% glycerol, 500 mM NaCl and 20 mM imidazole) by inversion, centrifugation and removal of supernatant. Proteins were eluted using elution buffer (as wash buffer but containing 500 mM imidazole). Elution fractions were concentrated and buffer was exchanged to storage buffer (20 mM HEPES and 150 mM NaCl, pH 7.5) using Amicon ultracentrifugal filters (Millipore) with the appropriate exclusion size according to the manufacturer’s instructions. Purity was assessed through SDS–PAGE and concentration was determined by using the extinction coefficient and measuring the absorbance at 280 nm. Proteins were flash-frozen in small aliquots in liquid nitrogen and stored at −70 °C for storage. Expression and purification of AsDGD2, AsS6DGD, CiDGD and CiS6DGD were performed as above with some modifications. To increase yield, expression took place in larger cultures (up to 1 L). Cell pellets were lysed on ice for 30 min with approximately 6 ml of lysis buffer per 1 g of cell pellet (50 mM Tris-HCl pH 8, 50 mM glycine, 5% glycerol, 500 mM NaCl, 20 mM imidazole, 0.2 mg ml⁻¹ lysozyme and one tablet (per 50 ml buffer) of complete EDTA free protease inhibitor (Roche)) and sonicated for 2.5 min (2 s on, 3 s off) on ice (Bandelin UW 2070). Subsequent purification and concentration steps were performed as described above.

Phylogenetic analyses

Sequences were obtained through BLAST searches against the publicly available NCBI and 1KP⁹⁶ databases or from C. ipecacuanha and A. salviifolium transcriptomes generated in the frame of this study. Sequences and accession numbers are provided in Supplementary Dataset 1. Full-length amino acid sequences were aligned with webPRANK (https://www.ebi.ac.uk/goldman-srv/webprank/)⁹⁷ and maximum-likelihood phylogenetic trees were built using the IQ-TREE web server (http://iqtree.cibiv.univie.ac.at/)⁹⁸ (automatic substitution model; bootstrap value = 1,000). iTOL was used to visualize and graphically edit trees (https://itol.embl.de/)⁹⁹.

Protein models

Protein models were predicted with AlphaFold3 through the AlphaFold Server (https://alphafoldserver.com/)¹⁰⁰. Models were visualized using ChimeraX version 1.8 for Mac¹⁰¹.

Commercially available chemicals and standards

Secologanin (50741), dopamine hydrochloride (H3502), tryptamine hydrochloride (246557), 4-O-Me-dopamine hydrochloride (H3132), cephaeline dihydrochloride (PHL85887, phyproof reference substance), emetine dihydrochloride (PHL89489, phyproof reference substance) and ipecoside (TA9H93CFC242, TargetMol) were purchased from Sigma.

Chemically produced standards and substrates

Secologanic acid was produced through alkaline hydrolysis of secologanin by incubating secologanin with 0.1 M NaOH (40 µl per 1 mg secologanin) for 5 h, similarly to a previously described protocol¹⁰². The solution was neutralized and completeness of the reaction was confirmed through LC–MS analysis. Secologanic acid was stored in solution at −25 °C.

Isotopically labeled tryptamine-d₅ was obtained as previously described¹⁰³.

Epimeric mixtures of DAI(I), DAI(I)A, 7-O-Me-DA(I)I and 7-O-Me-DAI(I)A were obtained using a previously described protocol with modifications³³. Briefly, 10 mM secologanin or secologanic acid was incubated together with 10 mM dopamine hydrochloride or 4-O-Me-dopamine hydrochloride, respectively, in 0.1 M citrate and 0.2 M phosphate buffer pH 5.5 for 24 h at room temperature. This was sufficient for coupling as confirmed by UPLC–MS analysis. 6-O-Me-DAI(I)A cannot be produced using this protocol and was produced enzymatically (described below). For in vitro competition assays of glucosidases (described below) DAI(I)A and 7-O-Me-DAI(I)A were produced using 20 mM of each substrate and DAI(I) and 7-O-Me-DAI(I) were produced using 50 mM of each substrate.

Protoemetinol was produced from protoemetine (described below) by reduction with NaBH₄. An aliquot of 50 µg of protoemetine was incubated in methanol with 0.5 g L⁻¹ NaBH₄ for 30 min at room temperature. The reaction was quenched by adding an equal volume of acetone followed by a 15-min incubation at room temperature. The reaction was centrifuged at 18,000g for 15 min, diluted with methanol and analyzed with UPLC–MS method 3, which indicated full conversion to a compound consistent with m/z MS² of protoemetinol (Supplementary Fig. 7).

Purification of epimer-pure DAII and DAI

An epimeric mixture of DAII and DAI was produced as described above but with some modifications. A 1.2-fold excess of dopamine hydrochloride was used, while the reaction was bubbled with argon and then incubated for 3 days. Preparative isolation was performed using an Xbridge BEH C18 OBD prep column 130 Å, 5 µm (Waters) on an Agilent 1260 HPLC system (with G7161A prep bin pump, G9328A column organizer and G7165A detector) and an Agilent 1290 G7159B fraction collector. The mobile phase was A:B where A was water with 0.1% formic acid and B was acetonitrile with 0.1% formic acid. The separation method was as follows: 5% B from 1 min to 25% B at 24 min, then the column was flushed at 100% B until 29 min and re-equilibrated to 5% B until 35 min. The flow rate was 8 ml min⁻¹ and detection was performed at λ = 254 nm. Aliquots of fractions of each of the two main peaks were analyzed by UPLC–MS method 3, confirming that m/z and retention times corresponded to either of the two peaks typically observed in epimeric mixtures (Supplementary Fig. 5). Collected fractions of each peak were combined, dried using a freeze-dryer (Labconco) and subjected to nuclear magnetic resonance (NMR).

Deacetylisoipecoside (DAII, 4a) was unstable during 2D NMR measurements in CH₃OH-d₃ and was, therefore, measured in acidic CH₃OH-d₃ (CH₃OH-d₃ + 0.1% formic acid), to prevent decomposition during long-term measurements. Deacetylipecoside (DAI 4b) was lactamized even in acidic CH₃OH-d₃. Thus, the structure of 4b was confirmed as demethylalangiside after complete lactamization. This observation was consistent with previously reported fact that the lactamization was much faster in the 1R-epimer than the 1S-epimer¹⁰⁴.

NMR analysis

NMR measurements were carried out on a 500-MHz Bruker Avance III HD spectrometer (Bruker Biospin) equipped with a TCI cryoprobe using standard pulse sequences as implemented in Bruker Topspin version 3.6.1. (Bruker Biospin) or a 400-MHz Bruker Avance III HD spectrometer (Bruker Biospin). Chemical shifts were referenced to the residual solvent signals of CDCl₃ (δ_H 7.26/δ_C 77.16) or CH₃OH-d₃ (δ_H 3.31/δ_C 49.0) and coupling constants (J, Hz) are expressed in the following format: chemical shift value (multiplicity, coupling constant, integration). ¹H-NMR spectral data are described using the following abbreviations: brs, broad singlet; s, singlet; d, doublet; t, triplet; q, quartet; dd, doublet of doublets; ddd, doublet of doublets of doublets; dt, doublet of triplets; td, triplet of doublets; appd, apparent doublet; m, multiplet.

On the basis of the structure determined from NMR analysis, a molecular model was created in GaussView version 6 (Semichem) and optimized using the semiempirical method PM6 in Gaussian version 16 (Gaussian). The optimized conformation from PM6 was used for rotating frame Overhauser enhancement spectroscopy (ROESY) analysis of protoemetine. For deacetylisoipecoside and demethylalangiside, the resulting structures were used for conformer variation with the GMMX processor of the Gaussian program package. Resulting structures were optimized through density functional theory (DFT) with Gaussian version 16 (B3LYP/6-31G(d), gas phase). The lowest-energy conformer from the DFT calculations was used for the ROESY analysis.

Enzymatically produced standards and substrates

6-O-Me-DAIIA and 6-O-Me-DAIA were produced using recombinant CiDOMT1 and AsDOMT3, respectively. The reaction mix contained 50 mM HEPES pH 7.5, 1 mM total epimeric DAI(I)A (~500 µM per epimer) as substrate, 1 mM of potential metal cofactors MgCl₂ and MnCl₂, 1 mM SAM chloride dihydrochloride and 5 µM CiDOMT1 or 10 µM AsDOMT3. Reactions were incubated for 24 h at 30 °C and stopped by incubation at 98 °C for 10 min to inactivate enzymes. Precipitates were pelleted by centrifugation at 18,000g for 30 min. Supernatants were diluted 1:2 in H₂O and used as substrates in leaf disk assays. Aliquots of the reactions were analyzed by UPLC–MS method 3 to observe expected products (Supplementary Fig. 5).

Confirmation of DAIIA and DAIA peak identity was achieved through deesterification of purified DAII and DAI, respectively, using recombinant CiDE. Reaction mixes contained 50 mM HEPES pH 7.5, 3 µM recombinant CiDE and 200 µM purified DAII or DAI. Reactions were incubated overnight at 30 °C and stopped by the addition of 50 µl of methanol supplemented with 0.1% FA. Eventual precipitates were pelleted by centrifugation at 18,000g for 30 min and supernatants were analyzed by UPLC–MS method 3 (Supplementary Fig. 5).

For in vitro competition assays of glucosidases (described below), 6-O-Me-DAIIA 6a and 6-O-Me-DAIA 6b were produced separately through coupling reactions. For 6-O-Me-DAIIA 6a, the reaction mixture contained 50 mM HEPES pH 7.5, 1 mM purified DAII 4a (described above), 1 mM MgCl₂, 1 mM SAM, 40 µM CiDOMT1 and 50 µM CiDE. For 6-O-Me-DAIA 6b, the reaction mixture contained 50 mM HEPES pH 7.5, 1 mM purified DAI (described above), 1 mM MgCl₂, 1 mM SAM, 10 µM AsDOMT3 and 20 µM CiDE. Reactions were incubated for 24 h at 30 °C and stopped by incubation at 98 °C for 10 min. Precipitates were pelleted by centrifugation at 18,000g for 30 min. Aliquots of the supernatants were analyzed by UPLC–MS method 3, confirming that reactions were complete, and then used as substrates in assays (described below).

In vitro competition activity assays of glucosidases

Compounds were mixed as 2× substrate master mixes (corresponding to calculated 200 µM per compound) and approximate equimolarity was checked by UPLC–MS analysis method 3. The substrate master mix for AsDGD2 and AsS6DGD assays contained all A. salviifolium-specific ipecac alkaloid glucosides: DAIIA 5a and DAIA 5b and 7-O-Me-DAIIA 12a and 7-O-Me-DAIA 12b (chemically synthesized as epimeric mixtures) and 6-O-Me-DAIIA 6a and 6-O-Me-DAIA 6b (enzymatically produced in epimer-pure form). The substrate mix to assay CiDGD and CiS6DGD contained C. ipecacuanha-specific ipecac alkaloid glucosides: ipecoside (commercially available), DAII 4a and DAI 4b, DAIIA 5a and DAIA 5b, 7-O-Me-DAII 18a and 7-O-Me-DAI 18b (chemically synthesized as epimeric mixtures) and 6-O-Me-DAIIA 6a (enzymatically produced in epimer-pure form). Note that 7-O-Me-DAII 18a and 7-O-Me-DAI 18b were not detected in C. ipecacuanha extracts but were included in the assay because 7-O-Me-DAI 18b can be produced by CiDOMT2 (Extended Data Fig. 4). Final assay mixes were prepared in triplicate at a final concentration of 50 mM HEPES pH 7.5, ~100 µM of each substrate and 0.5 µM of enzyme or boiled mixed enzyme as control. Boiled mixed enzymes were obtained by mixing the respective enzymes (AsDGD2 + AsS6DGD and CiDGD + CiS6DGD) at equimolar amounts, heating for 10 min at 98 °C and removing the precipitate through centrifugation at 18 000g for 20 min. Each replicate was pipetted separately on ice and immediately aliquoted into 25-µl samples on ice and then transferred to 30 °C and gentle shaking. At indicated time points, samples were snap-frozen. For the earliest time point (‘assay start’) samples were snap-frozen immediately after aliquoting on ice and before transfer to 30 °C. For UPLC–MS analysis, 75 µl of 70% methanol containing 0.1% formic acid and 2 mg L⁻¹ caffeine as internal standard was added to each 25-µl sample. Samples were centrifuged for 30 min at 18,000g and room temperature to remove precipitates. Supernatants were analyzed using UPLC–MS method 3.

Production of protoemetinol standard

Protoemetine was isolated in several batches from a total of 6 g of freshly harvested leaf buds and young leaves. The material was flash-frozen and ground with liquid nitrogen; then, 100% methanol was added at a ratio of 200 µl per 100 mg. Samples were sonicated in an ultrasonic bath for 10 min, incubated on a rotator for 30 min and centrifuged at 18,000g for 15 min. The supernatants were filtered through 0.22-µm syringe filters, diluted 1:10 in 100% methanol and subjected to semipreparative HPLC. Semipreparative isolation was performed using an XBridge BEH C18 column of 100 mm × 4.6 mm, 2.5 μm (Waters) on an Agilent 1260 Infinity HPLC system (with G1311B Quat pumps, G1315C diode array detectors, G1316A oven, G1329B autosampler and G1364F fraction collector). The mobile phase was A:B, where A was 0.1% formic acid in water and B was acetonitrile. The separation method was as follows: 8% B at 1 min to 13% B at 4 min, to 30% at 12 min. Then, the column was flushed at 100% B until 15 min and re-equilibrated to 8% B until 18 min. The flow rate was 2.4 ml min⁻¹ and detection was performed at λ = 250 nm. Aliquots of collected fractions were analyzed using UPLC–MS method 3 and fractions containing the putative protoemetine peak, according to m/z and typical aldehyde peak shape, were combined and freeze-dried using a freeze-dryer (Labconco). Large freeze-dried fractions were solubilized in small volumes of 100% methanol, combined (around 350 µg) and dried under a nitrogen line. NMR was conducted on a Bruker Avance III HD 700-MHz spectrometer (Bruker Biospin), equipped with a TCI cryoprobe using standard pulse sequences implemented in Bruker Topspin version 3.6.1 (Bruker Biospin). Chemical shifts were referenced to CDCl₃ residual solvent signals (J_H 7.26/J_C 77.16). NMR spectra of the sample were as follows: ¹H-NMR (700 MHz,CDCl₃): δ ppm: 0.93 (t, J = 7.5 Hz, 3H), 1.13 (m, 1H), 1.29 (m, 1H), 1.48 (m, 1H), 1.58 (m, 1H), 1.95 (m, 1H), 2.08 (dd, J = 11.3 Hz, 1H), 2.33 (m, 2H), 2.50 (ddd, J = 11.9, 11.4, 4.5 Hz, 1H), 2.63 (m, 1H), 2.72 (dd, J = 16.4, 3.3 Hz, 1H), 2.98 (m, 1H), 3.10 (m, 2H), 3.13 (m, 1H), 3.83 (s, 3H), 3.83 (s, 3H), 6.57 (s, 1H), 6.63 (s, 1H) and 9.88 (bs, 1H); ¹³C-NMR (175 MHz, CDCl₃): δ ppm: 10.8, 23.6, 29.1, 35.7, 38.1, 41.3, 48.1, 52.3, 55.8, 55.8, 61.2, 62.6, 108.0, 111.4, 126.2, 129.4, 147.0, 147.3 and 202.4. This spectrum is in agreement with a previously published NMR spectrum of protoemetine¹⁰⁵ and with that of chemically synthesized protoemetine (Supplementary Methods) in this study (Supplementary Fig. 26 and Supplementary Table 4). Isolated and synthesized protoemetine were, thus, identical.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

There are no restrictions on the availability of the data. Illumina RNA-seq raw reads were uploaded in FASTQ format to the NCBI Sequence Read Archive under BioProject PRJNA1169657. All reported gene sequences were deposited to NCBI Genbank under the following accession numbers (Supplementary Table 6): AsDGD1 (PQ363556), AsDGD2 (PQ363557), AsDOMT1 (PQ363558), AsDOMT2 (PQ363559), AsDOMT3 (PQ363560), AsDOMT4 (PQ363561), AsDOMT5 (PQ363562), AsDOMT6 (PQ363563), AsDOMT7 (PQ363564), AsDPOMT1 (PQ363565), AsDPOMT2 (PQ363566), AsDR1 (PQ363567), AsDR2 (PQ363568), AsS6DGD (PQ363569), CiDE (PQ363570), CiDGD (PQ363571), CiDOMT1 (PQ363572), CiDOMT2 (PQ363573), CiDPOMT (PQ363574), CiDR1 (PQ363575), CiDR2 (PQ363576), CiIpS (PQ363577) and CiS6DGD (PQ363578). Sequences to construct phylogenetic trees were retrieved from public databases (https://blast.ncbi.nlm.nih.gov/Blast.cgi and https://db.cngb.org/onekp/) or assemblies made in this study; accession numbers and amino acid sequence are listed in Supplementary Dataset 1. Source data are provided with this paper.

Code availability

Not applicable.

References

Negin, B. & Jander, G. Convergent and divergent evolution of plant chemical defenses. Curr. Opin. Plant Biol. 73, 102368 (2023).
Article CAS PubMed Google Scholar
Weng, J. K. The evolutionary paths towards complexity: a metabolic perspective. New Phytol. 201, 1141–1149 (2014).
Article PubMed Google Scholar
Rodman, J. E., Soltis, P. S., Soltis, D. E., Sytsma, K. J. & Karol, K. G. Parallel evolution of glucosinolate biosynthesis inferred from congruent nuclear and plastid gene phylogenies. Am. J. Bot. 85, 997–1006 (1998).
Article CAS PubMed Google Scholar
Florean, M. et al. Reinventing metabolic pathways: Independent evolution of benzoxazinoids in flowering plants. Proc. Natl Acad. Sci. USA 120, e2307981120 (2023).
Article CAS PubMed PubMed Central Google Scholar
Huang, R., O’Donnell, A. J., Barboline, J. J. & Barkman, T. J. Convergent evolution of caffeine in plants by co-option of exapted ancestral enzymes. Proc. Natl Acad. Sci. USA 113, 10613–10618 (2016).
Article CAS PubMed PubMed Central Google Scholar
O’Donnell, A. J., Huang, R., Barboline, J. J. & Barkman, T. J. Convergent biochemical pathways for xanthine alkaloid production in plants evolved from ancestral enzymes with different catalytic properties. Mol. Biol. Evol. 38, 2704–2714 (2021).
Article PubMed PubMed Central Google Scholar
Vignale, F. A. et al. Yerba mate (Ilex paraguensis) genome provides new insights into convergent evolution of caffeine biosynthesis. eLife 14, e104759 (2025).
Article PubMed PubMed Central Google Scholar
Berman, P. et al. Parallel evolution of cannabinoid biosynthesis. Nat. Plants 9, 817–831 (2023).
Article CAS PubMed Google Scholar
Kunert, M. et al. Promiscuous CYP87A enzyme activity initiates cardenolide biosynthesis in plants. Nat. Plants 9, 1607–1617 (2023).
Article CAS PubMed PubMed Central Google Scholar
Zust, T. et al. Independent evolution of ancestral and novel defenses in a genus of toxic plants (Erysimum, Brassicaceae). eLife 9, e51712 (2020).
Article CAS PubMed PubMed Central Google Scholar
Younkin, G. C. et al. Cardiac glycosides protect wormseed wallflower (Erysimum cheiranthoides) against some, but not all, glucosinolate-adapted herbivores. New Phytol. 242, 2719–2733 (2024).
Article CAS PubMed Google Scholar
Seger, D. & Muelenbelt, J. Position paper: ipecac syrup. J. Toxicol. Clin. Toxicol. 42, 133–143 (2004).
Article Google Scholar
Hu, X. Y. et al. Genus Alangium—a review on its traditional uses, phytochemistry and pharmacological activities. Fitoterapia 147, 104773 (2020).
Article CAS PubMed Google Scholar
Sauvain, M. et al. Antimalarial activity of alkaloids from Pogonopus tubulosus. Phytother. Res. 10, 198–201 (1996).
Article CAS Google Scholar
Kim, B. H. et al. Tubulosine selectively inhibits JAK3 signalling by binding to the ATP-binding site of the kinase of JAK3. J. Cell. Mol. Med. 24, 7427–7438 (2020).
Article CAS PubMed PubMed Central Google Scholar
Zuntini, A. R. et al. Phylogenomics and the rise of the angiosperms. Nature 629, 843–850 (2024).
Article CAS PubMed PubMed Central Google Scholar
Itoh, A. et al. Tetrahydroisoquinoline-monoterpene and iridoid glycosides from Alangium lamarckii. Phytochemistry 56, 623–630 (2001).
Article CAS PubMed Google Scholar
DeEknamkul, W., Ounaroon, A., Tanahashi, T., Kutchan, T. M. & Zenk, M. H. Enzymatic condensation of dopamine and secologanin by cell-free extracts of Alangium lamarckii. Phytochemistry 45, 477–484 (1997).
Article CAS Google Scholar
Jain, S., Sinha, A. & Bhakuni, D. S. The biosynthesis of β-carboline and quinolizidine alkaloids of Alangium lamarckii. Phytochemistry 60, 853–859 (2002).
Article CAS PubMed Google Scholar
Nagakura, N., Höfle, G. & Zenk, M. H.Deacetylisoipecoside: the key intermediate in the biosynthesis of the alkaloids cephaeline and emetine. J. Chem. Soc., Chem. Commun. 1978, 896–898 (1978).
Article Google Scholar
Sadre, R. et al. Metabolite diversity in alkaloid biosynthesis: a multilane (diastereomer) highway for camptothecin synthesis in Camptotheca acuminata. Plant Cell 28, 1926–1944 (2016).
Article CAS PubMed PubMed Central Google Scholar
Kang, M. et al. A chromosome-level Camptotheca acuminata genome assembly provides insights into the evolutionary origin of camptothecin biosynthesis. Nat. Commun. 12, 3531 (2021).
Article CAS PubMed PubMed Central Google Scholar
Battersby, A. R., Lewis, N. G. & Tippett, J. M.The Basic Glucosides Related to the Biosynthesis of Indole and Ipecac Alkaloids. Tetrahedron Lett. 19, 4849–4852 (1978).
Article Google Scholar
Tanahashi, T. et al. A tetrahydroisoquinoline-monoterpene glucoside and an iridoid glucoside from Alangium kurzii. Chem. Pharm. Bull. 48, 415–419 (2000).
Article CAS Google Scholar
De-Eknamkul, W., Suttipanta, N. & Kutchan, T. M. Purification and characterization of deacetylipecoside synthase from Alangium lamarckii Thw. Phytochemistry 55, 177–181 (2000).
Article CAS PubMed Google Scholar
Battersby, A. R., Davidson, G. C. & Harper, J. T.Ipecacuanha alkaloids. Part I. Fractionation studies and isolation of two new alkaloids. J. Chem. Soc. 1959, 1744–1748 (1959).
Article Google Scholar
Battersby, A. R. & Harper, B. J. T.Ipecacuanha alkaloids. Part II. The structure of protoemetine and partial synthesis of (−)emetine. J. Chem. Soc. 1959, 1748–1753 (1959).
Article Google Scholar
Itoh, A., Ikuta, Y., Tanahashi, T. & Nagakura, N. Two alangium alkaloids from Alangium lamarckii. J. Nat. Prod. 63, 723–725 (2000).
Article CAS PubMed Google Scholar
Battersby, A. R., Merchant, J. R., Ruveda, E. A. & Salgar, S. S.Structure, synthesis, and stereochemistry of deoxytubulosine. Chem. Commun. 1965, 315–317 (1965).
Google Scholar
Albright, J. D., Vanmeter, J. C. & Goldman, L. Alkaloid studies 4. Isolation of cephaeline and tubulosine from Alangium lamarckii. Lloydia 28, 212–217 (1965).
CAS Google Scholar
Battersby, A. R. et al. Constitution of Ipecoside—a monoterpenoid isoquinoline. Chem. Commun. 1967, 219–221 (1967).
Google Scholar
Nomura, T. & Kutchan, T. M. Three new O-methyltransferases are sufficient for all O-methylation reactions of ipecac alkaloid biosynthesis in root culture of Psychotria ipecacuanha. J. Biol. Chem. 285, 7722–7738 (2010).
Article CAS PubMed PubMed Central Google Scholar
Nomura, T., Quesada, A. L. & Kutchan, T. M. The new β-d-glucosidase in terpenoid-isoquinoline alkaloid biosynthesis in Psychotria ipecacuanha. J. Biol. Chem. 283, 34650–34659 (2008).
Article CAS PubMed PubMed Central Google Scholar
Cheong, B. E., Takemura, T., Yoshimatsu, K. & Sato, F. Molecular cloning of an O-methyltransferase from adventitious roots of Carapichea ipecacuanha. Biosci. Biotechnol. Biochem. 75, 107–113 (2011).
Article CAS PubMed Google Scholar
Menendez-Perdomo, I. M. & Facchini, P. J. Elucidation of the (R)-enantiospecific benzylisoquinoline alkaloid biosynthetic pathways in sacred lotus (Nelumbo nucifera). Sci. Rep. 13, 2955 (2023).
Article CAS PubMed PubMed Central Google Scholar
Lee, E. J. & Facchini, P. Norcoclaurine synthase is a member of the pathogenesis-related 10/Bet v1 protein family. Plant Cell 22, 3489–3503 (2010).
Article CAS PubMed PubMed Central Google Scholar
Yang, Y. et al. Bifunctional cytochrome P450 enzymes involved in camptothecin biosynthesis. ACS Chem. Biol. 14, 1091–1096 (2019).
Article CAS PubMed Google Scholar
Trenti, F. et al. Early and late steps of quinine biosynthesis. Org. Lett. 23, 1793–1797 (2021).
Article CAS PubMed PubMed Central Google Scholar
Torrens-Spence, M. P., Pluskal, T., Li, F. S., Carballo, V. & Weng, J. K. Complete pathway elucidation and heterologous reconstitution of rhodiola salidroside biosynthesis. Mol. Plant 11, 205–217 (2018).
Article CAS PubMed Google Scholar
Zhang, Y., Nowak, G., Reed, D. W. & Covello, P. S. The production of artemisinin precursors in tobacco. Plant Biotechnol. J. 9, 445–454 (2011).
Article CAS PubMed Google Scholar
Lukacin, R., Matern, U., Specker, S. & Vogt, T. Cations modulate the substrate specificity of bifunctional class I O-methyltransferase from Ammi majus. FEBS Lett. 577, 367–370 (2004).
Article CAS PubMed Google Scholar
Park, M. R., Chen, X., Lang, D. E., Ng, K. K. S. & Facchini, P. J. Heterodimeric O-methyltransferases involved in the biosynthesis of noscapine in opium poppy. Plant J. 95, 252–267 (2018).
Article CAS PubMed Google Scholar
Payne, R. M. et al. An NPF transporter exports a central monoterpene indole alkaloid intermediate from the vacuole. Nat. Plants 3, 16208 (2017).
Article PubMed PubMed Central Google Scholar
de Brito Francisco, R. & Martinoia, E. The vacuolar transportome of plant specialized metabolites. Plant Cell Physiol. 59, 1326–1336 (2018).
PubMed Google Scholar
Martinoia, E., Maeshima, M. & Neuhaus, H. E. Vacuolar transporters and their essential role in plant metabolism. J. Exp. Bot. 58, 83–102 (2007).
Article CAS PubMed Google Scholar
Thumuluri, V., Almagro Armenteros, J. J., Johansen, A. R., Nielsen, H. & Winther, O. DeepLoc 2.0: multi-label subcellular localization prediction using protein language models. Nucleic Acids Res. 50, W228–W234 (2022).
Article CAS PubMed PubMed Central Google Scholar
Ivanov, S. & Harrison, M. J. A set of fluorescent protein-based markers expressed from constitutive and arbuscular mycorrhiza-inducible promoters to label organelles, membranes and cytoskeletal elements in Medicago Truncatula. Plant J. 80, 1151–1163 (2014).
Article CAS PubMed Google Scholar
Guirimand, G. et al. Strictosidine activation in Apocynaceae: towards a ‘nuclear time bomb‘? BMC Plant Biol. 10, 20 (2010).
Article Google Scholar
Liu, Y., Fernie, A. R. & Tohge, T. Diversification of chemical structures of methoxylated flavonoids and genes encoding flavonoid-O-methyltransferases. Plants (Basel) 11, 564 (2022).
CAS PubMed Google Scholar
Weng, J. K. & Noel, J. P. Chemodiversity in Selaginella: a reference system for parallel and convergent metabolic evolution in terrestrial plants. Front. Plant Sci. 4, 119 (2013).
Article PubMed PubMed Central Google Scholar
Miller, J. C. & Schuler, M. A. Single mutations toggle the substrate selectivity of multifunctional Camptotheca secologanic acid synthases. J. Biol. Chem. 298, 102237 (2022).
Article CAS PubMed PubMed Central Google Scholar
Pesnot, T., Gershater, M. C., Ward, J. M. & Hailes, H. C. Phosphate mediated biomimetic synthesis of tetrahydroisoquinoline alkaloids. Chem. Commun. 47, 3242–3244 (2011).
Article CAS Google Scholar
McKnight, T. D., Bergey, D. R., Burnett, R. J. & Nessler, C. L. Expression of enzymatically active and correctly targeted strictosidine synthase in transgenic tobacco plants. Planta 185, 148–152 (1991).
Article CAS PubMed Google Scholar
Samanani, N. & Facchini, P. J. Purification and characterization of norcoclaurine synthase. The first committed enzyme in benzylisoquinoline alkaloid biosynthesis in plants. J. Biol. Chem. 277, 33878–33883 (2002).
Article CAS PubMed Google Scholar
Maresh, J. J. et al. Strictosidine synthase: mechanism of a Pictet−Spengler catalyzing enzyme. J. Am. Chem. Soc. 130, 710–723 (2008).
Article CAS PubMed PubMed Central Google Scholar
Albach, D. C., Soltis, P. S. & Soltis, D. E. Patterns of embryological and biochemical evolution in the asterids. Syst. Bot. 26, 242–262 (2001).
Google Scholar
Awadasseid, A. et al. Characterization of Camptotheca acuminata 10-hydroxygeraniol oxidoreductase and iridoid synthase and their application in biological preparation of nepetalactol in Escherichia coli featuring NADP⁺–NADPH cofactors recycling. Int. J. Biol. Macromol. 162, 1076–1085 (2020).
Article CAS PubMed Google Scholar
Noda-Garcia, L., Liebermeister, W. & Tawfik, D. S. Metabolite–enzyme coevolution: from single enzymes to metabolic pathways and networks. Annu. Rev. Biochem. 87, 187–216 (2018).
Article CAS PubMed Google Scholar
Scossa, F. & Fernie, A. R. The evolution of metabolism: how to test evolutionary hypotheses at the genomic level. Comput Struct. Biotechnol. J. 18, 482–500 (2020).
Article CAS PubMed PubMed Central Google Scholar
Arthan, D., Kittakoop, P., Esen, A. & Svasti, J. Furostanol glycoside 26-O-β-glucosidase from the leaves of Solanum torvum. Phytochemistry 67, 27–33 (2006).
Article CAS PubMed Google Scholar
Suthangkornkul, R. et al. A Solanum torvum GH3 β-glucosidase expressed in Pichia pastoris catalyzes the hydrolysis of furostanol glycoside. Phytochemistry 127, 4–11 (2016).
Article CAS PubMed Google Scholar
Hrmova, M. et al. Barley β-d-glucan exohydrolases with β-d-glucosidase activity. Purification, characterization, and determination of primary structure from a cDNA clone. J. Biol. Chem. 271, 5277–5286 (1996).
Article CAS PubMed Google Scholar
Kim, J.-B., Olek, A. T. & Carpita, N. C. Cell wall and membrane-associated exo-β-d-glucanases from developing maize seedlings. Plant Physiol. 123, 471–486 (2000).
Article CAS PubMed PubMed Central Google Scholar
Takahashi, H. et al. Gtgen3A, a novel plant GH3 β-glucosidase, modulates gentio-oligosaccharide metabolism in Gentiana. Biochem. J. 475, 1309–1322 (2018).
Article CAS PubMed Google Scholar
Sampedro, J. et al. Soluble and membrane-bound β-glucosidases are involved in trimming the xyloglucan backbone. Plant Physiol. 173, 1017–1030 (2017).
Article CAS PubMed Google Scholar
Levac, D., Murata, J., Kim, W. S. & De Luca, V. Application of carborundum abrasion for investigating the leaf epidermis: molecular cloning of Catharanthus roseus 16-hydroxytabersonine-16-O-methyltransferase. Plant J. 53, 225–236 (2008).
Article CAS PubMed Google Scholar
Shitan, N. & Yazaki, K. Dynamism of vacuoles toward survival strategy in plants. Biochim. Biophys. Acta Biomembr. 1862, 183127 (2020).
Article CAS PubMed Google Scholar
Konno, K., Hirayama, C., Yasui, H. & Nakamura, M. Enzymatic activation of oleuropein: a protein crosslinker used as a chemical defense in the privet tree. Proc. Natl Acad. Sci. USA 96, 9159–9164 (1999).
Article CAS PubMed PubMed Central Google Scholar
Kim, D.-H., Kim, B.-R., Kim, J.-Y. & Jeong, Y.-C. Mechanism of covalent adduct formation of aucubin to proteins. Toxicol. Lett. 114, 181–188 (2000).
Article CAS PubMed Google Scholar
Kissen, R., Rossiter, J. T. & Bones, A. M. The ‘mustard oil bomb’: not so easy to assemble?! Localization, expression and distribution of the components of the myrosinase enzyme system. Phytochem. Rev. 8, 69–86 (2008).
Article Google Scholar
Blazevic, I. et al. Glucosinolate structural diversity, identification, chemical synthesis and metabolism in plants. Phytochemistry 169, 112100 (2020).
Article CAS PubMed Google Scholar
Lacchini, E. et al. The saponin bomb: a nucleolar-localized β-glucosidase hydrolyzes triterpene saponins in Medicago truncatula. New Phytol. 239, 705–719 (2023).
Article CAS PubMed Google Scholar
Koudounas, K. et al. The C-___domain of oleuropein β-glucosidase assists in protein folding and sequesters the enzyme in nucleus. Plant Physiol. 174, 1371–1383 (2017).
Article CAS PubMed PubMed Central Google Scholar
Kliebenstein, D. J. & Kvitko, B. H. Better living through phytochemistry: ‘phytoavengins’ and reappraising the production-focused dichotomy for defensive phytochemicals. Physiol. Mol. Plant Pathol. 125, 101978 (2023).
Article CAS Google Scholar
Salim, V., Jones, A. D. & DellaPenna, D. Camptotheca acuminata 10-hydroxycamptothecin O-methyltransferase: an alkaloid biosynthetic enzyme co-opted from flavonoid metabolism. Plant J. 95, 112–125 (2018).
Article CAS PubMed Google Scholar
Schmid, R. et al. Integrative analysis of multimodal mass spectrometry data in MZmine 3. Nat. Biotechnol. 41, 447–449 (2023).
Article CAS PubMed PubMed Central Google Scholar
Bocker, S. & Duhrkop, K. Fragmentation trees reloaded. J. Cheminform. 8, 5 (2016).
Article PubMed PubMed Central Google Scholar
Duhrkop, K. et al. SIRIUS 4: a rapid tool for turning tandem mass spectra into metabolite structure information. Nat. Methods 16, 299–302 (2019).
Article PubMed Google Scholar
Duhrkop, K., Shen, H., Meusel, M., Rousu, J. & Bocker, S. Searching molecular structure databases with tandem mass spectra using CSI:FingerID. Proc. Natl Acad. Sci. USA 112, 12580–12585 (2015).
Article PubMed PubMed Central Google Scholar
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Article CAS PubMed PubMed Central Google Scholar
Bushmanova, E., Antipov, D., Lapidus, A. & Prjibelski, A. D. rnaSPAdes: a de novo transcriptome assembler and its application to RNA-seq data. Gigascience 8, giz100 (2019).
Article PubMed PubMed Central Google Scholar
Manni, M., Berkeley, M. R., Seppey, M., Simao, F. A. & Zdobnov, E. M. BUSCO update: novel and streamlined workflows along with broader and deeper phylogenetic coverage for scoring of eukaryotic, prokaryotic, and viral genomes. Mol. Biol. Evol. 38, 4647–4654 (2021).
Article CAS PubMed PubMed Central Google Scholar
Afgan, E. et al. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2018 update. Nucleic Acids Res. 46, W537–W544 (2018).
Article CAS PubMed PubMed Central Google Scholar
Gotz, S. et al. High-throughput functional annotation and data mining with the Blast2GO suite. Nucleic Acids Res. 36, 3420–3435 (2008).
Article CAS PubMed PubMed Central Google Scholar
Huerta-Cepas, J. et al. eggNOG 4.5: a hierarchical orthology framework with improved functional annotations for eukaryotic, prokaryotic and viral sequences. Nucleic Acids Res. 44, D286–D293 (2016).
Article CAS PubMed Google Scholar
Cardenas, P. D. et al. Pathways to defense metabolites and evading fruit bitterness in genus Solanum evolved through 2-oxoglutarate-dependent dioxygenases. Nat. Commun. 10, 5169 (2019).
Article PubMed PubMed Central Google Scholar
Berrow, N. S. et al. A versatile ligation-independent cloning method suitable for high-throughput expression screening applications. Nucleic Acids Res. 35, e45 (2007).
Article PubMed PubMed Central Google Scholar
Sarrion-Perdigones, A. et al. GoldenBraid: an iterative cloning system for standardized assembly of reusable genetic modules. PLoS ONE 6, e21622 (2011).
Article CAS PubMed PubMed Central Google Scholar
Grzech, D., Hong, B., Caputi, L., Sonawane, P. D. & O’Connor, S. E. Engineering the biosynthesis of late-stage vinblastine precursors precondylocarpine acetate, catharanthine, tabersonine in Nicotiana benthamiana. ACS Synth. Biol. 12, 27–34 (2023).
Article CAS PubMed Google Scholar
Zhang, Y. et al. A highly efficient Agrobacterium-mediated method for transient gene expression and functional studies in multiple plant species. Plant Commun. 1, 100028 (2020).
Article PubMed PubMed Central Google Scholar
Hong, B. et al. Biosynthesis of strychnine. Nature 607, 617–622 (2022).
Article CAS PubMed PubMed Central Google Scholar
Kamileen, M. O. et al. Recycling upstream redox enzymes expands the regioselectivity of cycloaddition in pseudo-Aspidosperma alkaloid biosynthesis. J. Am. Chem. Soc. 144, 19673–19679 (2022).
Article CAS PubMed PubMed Central Google Scholar
Kamileen, M. O. et al. Streamlined screening platforms lead to the discovery of pachysiphine synthase from Tabernanthe iboga. New Phytol. 244, 1437–1449 (2024).
Article CAS PubMed Google Scholar
Schindelin, J. et al. Fiji: an open-source platform for biological-image analysis. Nat. Methods 9, 676–682 (2012).
Article CAS PubMed Google Scholar
Stavrinides, A. et al. Structural investigation of heteroyohimbine alkaloid synthesis reveals active site elements that control stereoselectivity. Nat. Commun. 7, 12116 (2016).
Article CAS PubMed PubMed Central Google Scholar
One Thousand Plant Transcriptomes Initiative. One thousand plant transcriptomes and the phylogenomics of green plants. Nature 574, 679–685 (2019).
Löytynoja, A. & Goldman, N. webPRANK: a phylogeny-aware multiple sequence aligner with interactive alignment browser. BMC Bioinformatics 11, 579 (2010).
Article PubMed PubMed Central Google Scholar
Trifinopoulos, J., Nguyen, L. T., von Haeseler, A. & Minh, B. Q. W-IQ-TREE: a fast online phylogenetic tool for maximum likelihood analysis. Nucleic Acids Res. 44, W232–W235 (2016).
Article CAS PubMed PubMed Central Google Scholar
Letunic, I. & Bork, P. Interactive Tree of Life (iTOL) v6: recent updates to the phylogenetic tree display and annotation tool. Nucleic Acids Res. 52, W78–W82 (2024).
Article PubMed PubMed Central Google Scholar
Abramson, J. et al. Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493–500 (2024).
Article CAS PubMed PubMed Central Google Scholar
Meng, E. C. et al. UCSF ChimeraX: tools for structure building and analysis. Protein Sci. 32, e4792 (2023).
Article CAS PubMed PubMed Central Google Scholar
Chapelle, J. P. Vogeloside et acide secologanique, glucosides secoiridoides d’anthocleista vogelu. Planta Med. 29, 268–274 (1976).
Article CAS PubMed Google Scholar
McDonald, A. et al. Enzymatic epimerization of monoterpene indole alkaloids in Kratom. Preprint at bioRxiv https://doi.org/10.1101/2024.12.13.628308 (2024).
Beke, G., Szabó, L. F. & Podányi, B. Regio- and stereoselectivity in the coupling reaction of secologanin with dopamine derivatives. J. Nat. Prod. 64, 332–340 (2001).
Article CAS PubMed Google Scholar
Lin, S., Deiana, L., Tseggai, A. & Córdova, A. Concise total synthesis of dihydrocorynanthenol, protoemetinol, protoemetine, 3‐epi‐protoemetinol and emetine. Eur. J. Org. Chem. 2012, 398–408 (2011).
Article Google Scholar

Download references

Acknowledgements

We thank the Ghent University Botanical Garden and Chantal Dugardin for providing A. salviifolium living specimens, N. J. Hernández Lozada and M. O. Kamileen for transport of these specimens, the gardeners E. Rothe and E. Goschala for growing and maintaining A. salviifolium and C. ipecacuanha plants, F. Kaltofen for growing N. benthamiana plants and H. Tilger for installing tools on the in-house Galaxy server. We thank D. Grzech for providing pUPD and CrNPF2.9 constructs, P. Sonawane for providing the modified 3Ω1 vector, M. O. Kamileen for providing A. tumefaciens strains harboring mCherry markers and A. McDonald for providing isotopically labeled tryptamine. We thank M. Florean and T. Köllner for advice on phylogenetic analyses. We are grateful to the Max Planck Society for funding and the European Research Council (788301) and the Leibniz Prize, German Research Foundation (505457618) grants awarded to S.E.O.

Funding

Open access funding provided by Max Planck Society.

Author information

Authors and Affiliations

Department of Natural Product Biosynthesis, Max Planck Institute for Chemical Ecology, Jena, Germany
Maite Colinas, Clara Morweiser, Olivia Dittberner, Ryan Alam, Helena Leucke, Delia Ayled Serna Guerrero, Sarah Heinicke, Maritta Kunert, Jens Wurlitzer, Kerstin Ploss, Benke Hong & Sarah E. O’Connor
Universidade de Ribeirão Preto, Departamento de Biotecnologia, Ribeirão Preto, Brazil
Bianca Chioca & Adriana A. Lopes
Research Group NMR and Department of Natural Product Biosynthesis, Max Planck Institute for Chemical Ecology, Jena, Germany
Yoko Nakamura
Microscopy Imaging Service, Max Planck Institute for Chemical Ecology, Jena, Germany
Veit Grabe

Authors

Maite Colinas
View author publications
Search author on:PubMed Google Scholar
Clara Morweiser
View author publications
Search author on:PubMed Google Scholar
Olivia Dittberner
View author publications
Search author on:PubMed Google Scholar
Bianca Chioca
View author publications
Search author on:PubMed Google Scholar
Ryan Alam
View author publications
Search author on:PubMed Google Scholar
Helena Leucke
View author publications
Search author on:PubMed Google Scholar
Yoko Nakamura
View author publications
Search author on:PubMed Google Scholar
Delia Ayled Serna Guerrero
View author publications
Search author on:PubMed Google Scholar
Sarah Heinicke
View author publications
Search author on:PubMed Google Scholar
Maritta Kunert
View author publications
Search author on:PubMed Google Scholar
Jens Wurlitzer
View author publications
Search author on:PubMed Google Scholar
Kerstin Ploss
View author publications
Search author on:PubMed Google Scholar
Benke Hong
View author publications
Search author on:PubMed Google Scholar
Veit Grabe
View author publications
Search author on:PubMed Google Scholar
Adriana A. Lopes
View author publications
Search author on:PubMed Google Scholar
Sarah E. O’Connor
View author publications
Search author on:PubMed Google Scholar

Contributions

M.C. designed all experimental work, performed all pathway discovery experiments (generation and analysis of RNA-seq and metabolite data, screening and reconstitution in N. benthamiana and recombinant enzyme activity assays), analyzed all data except NMR and supervised the work of C.M., O.D., H.L. and B.C. C.M. performed subcellular localization experiments assisted by V.G. C.M. and M.C generated phylogenetic trees. O.D. assisted with enzyme screening and experiments on spontaneous coupling. R.A. synthesized protoemetine. B.C. purified protoemetine from A. salviifolium. H.L. purified DAI and DAII. B.H. assisted with secologanic acid synthesis. Y.N. performed NMR and analyzed NMR data. D.A.S.G., S.H. and M.K. developed the chromatography and MS methods. J.W. assisted with cloning. A.A.L. provided C. ipecacuanha living specimens and assisted with growth. K.P. cultured and regenerated C. ipecacuanha in vitro. M.C. and S.E.O. designed the study and wrote the paper.

Corresponding authors

Correspondence to Maite Colinas or Sarah E. O’Connor.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Chemical Biology thanks Peter Facchini and the other, anonymous, reviewer(s) for their contribution to the peer review of this work.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Extended data

Extended Data Fig. 1 Assessing in planta coupling of tryptamine with secologanin or secologanic acid.

Small amounts of reaction products strictosidine (3S)/vincoside (3 R) or strictosidinic acid (3S)/vincosidic acid (3 R) of coupling reaction of tryptamine with secologanin (a) or secologanic acid (b) are observed within 24 hours after infiltration into N. benthamiana leaves. For comparison to the reaction with dopamine, the y axes showing signal intensity or peak areas were scaled identically to the axes of Fig. 2g-h (main text). c, Infiltration of tryptamine-d5 to flower petals of the natural secologanin producer C. roseus leads to large amounts of strictosidine-d4 but only traces of vincoside-d4. The endogenous C. roseus Strictosidine Synthase (STR) exclusively produces strictosidine, therefore the trace amount of vincoside can be attributed to non-enzymatic coupling. Taken together, the non-enzymatic reaction with tryptamine appears to occur much less readily than with dopamine. Stnd, standard. LC-MS peak areas are shown as bars of the mean of three biological replicates, error bars are standard error of the mean.

Extended Data Fig. 2 Selection of pathway gene candidates.

a, selection of C. ipecacuanha candidates by searching for genes similar to previously published active pathway enzymes^12,13. For each of the published sequences the contig with the highest sequencing homology ( > 96%) at cDNA level was considered to be identical to the published enzyme. CiDOMT1 was used as a bait for Pearson correlation. Candidates for deesterification and reduction were found by homology-based search and coexpression with CiDOMT1 (described in in the methods). A highly coexpressed gene predicted to encode for an acetyltransferase was picked as a candidate for ipecoside synthase. b, selection of A. salviifolium candidates was performed by coexpression analysis using orthologous and likely conserved sequences of enzymes supplying the direct precursor genes TyrDC and SLS, as baits, combined with filtering for predicted functional annotations for O-methyltransferases, glucosidases and dehydrogenases. Additionally, we considered candidates that showed either root or leaf bud specific expression (AsDGDs). Additional root specific AsDOMT5-7 were also identified. For both species gene expression patterns are consistent with metabolite abundance (see Fig. 2e-f). Heatmaps depict z-scores of TMM normalized CPM values for each gene.

Extended Data Fig. 3 C. ipecauanha protoemetine biosynthesis network including possible side branch products.

Activities were detected through expression of pathway genes shown in Supplementary Fig. 10. a, S-epimer derived pathways. b, R-epimer derived pathways.

Extended Data Fig. 4 Reconstitution of protoemetine biosynthesis network by expression of C. ipecacuanha pathway genes.

Data is from the same experiment shown in Fig. 3. This figure includes the data for Fig. 3 and additional gene combinations. Additionally, expression was combined with Catharanthus roseus strictosidine exporter CrNPF2.9. LC-MS peak areas are shown as bars of the mean of three biological replicates, error bars are standard error of the mean.

Extended Data Fig. 5 A. salviifolium protoemetine biosynthesis network.

Activities were detected through expression of pathway genes shown in Extended Data Fig. 6.

Extended Data Fig. 6 Reconstitution of protoemetine biosynthesis network by expression of A. salviifolium pathway genes.

Data is from the same experiment shown in Fig. 3. This figure includes the data for Fig. 3 and additional gene combinations. Additionally, expression was combined with Catharanthus roseus strictosidine exporter CrNPF2.9. LC-MS peak areas are shown as bars of the mean of three biological replicates, error bars are standard error of the mean.

Extended Data Fig. 7 Expression of the C. roseus strictosidine exporter gene CrNPF2.9 boosts C. ipecacuanha protoemetine biosynthesis and enables biosynthesis from dopamine and secologanin directly.

a, comparison of protoemetinol accumulation upon expression of indicated C. ipecacuanha biosynthesis genes in N. benthamiana with or without co-expression of transporter CrNPF2.9. Exogenous DAI/DAII are infiltrated as substrate and allowed to react for 24 hours before the leaf was harvested. Data was extracted from experiment shown in Supplementary Fig. 10. Peak areas are shown here relative to areas from samples where CrNPF2.9 was not expressed as bars of the mean of three replicates. b, expression of CiIps with or without co-overexpression of CrNPF2.9 did not lead to enhanced accumulation of the R-epimer derivative ipecoside. c, when C. ipecacuanha protoemetine pathway genes are expressed and uncoupled secologanin and dopamine are fed for 48 hours, only traces of protoemetinol can be detected (indicated with an asterisk). When CrNPF2.9 is co-overexpressed protoemetinol is formed to detectable levels. d, it is hypothesized that spontaneous in planta coupling occurs in the vacuole and the resulting formed DAII is only accessible to the cytosolic pathway enzymes if it is exported by a vacuolar exporter. CrNPF2.9 appears to export DAII. These results suggest the presence of a yet to be identified vacuolar exporter in C. ipecacuanha and A. salviifolium. LC-MS peak areas are shown as bars of the mean of three biological replicates, error bars are standard error of the mean.

Extended Data Fig. 8 Feeding of ipecac alkaloid glucosides to leaf disks expressing DGDs or S6DGDs.

Consumption of glucosides was measured because the unstable aglycone products cannot be reliably detected. a, agroinfiltration of N. benthamiana with strains harbouring constructs for overexpression of CiDGD or Ci6SDGD or empty vector (EV) respectively. Leaf disks were cut and incubated separately with indicated substrates. Note that 7-O-Me-DAIA and 7-O-Me-DAIIA are not detected in native C. ipecacuanha plants. b, agroinfiltration of N. benthamiana with strains harbouring constructs for overexpression of AsDGD1 or 2 or As6S6DGD or EV respectively. Leaf disks were cut out and incubated each with indicated substrates. AsDGDs deglycosylate all derivatives whereas AsS6DGD only shows activity towards 6-O-Me-DAIIA. 6-O-Me-DAIIA and 6-O-Me-DAIA cannot be obtained readily by Pictet-Spengler chemical reaction and were thus produced in vitro by recombinant CiDOMT1 and AsDOMT3, respectively (see methods). Bars are mean peak areas of three biological replicates relative to peak areas measured in EV, error bars are standard error of the mean SEM.

Extended Data Fig. 9 Parallel evolution of A. salviifolium versus C. ipecacuanha DRs.

Maximum likelihood phylogenetic tree of DR amino acid sequences and homologs from other Cornales and Gentianales species. Alphafold3 models indicate same folds for all DRs.

Supplementary information

Supplementary Information

Supplementary Methods, Figs. 1–26, Tables 1–6 and References.

Reporting Summary

Supplementary Data 1

Accession numbers and amino acid sequences of genes used to construct phylogenetic trees.

Source data

Source Data Fig. 2

Normalized peak areas.

Source Data Fig. 3

Normalized peak areas.

Source Data Fig. 4

Normalized peak areas.

Source Data Fig. 5

Normalized peak areas.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Colinas, M., Morweiser, C., Dittberner, O. et al. Ipecac alkaloid biosynthesis in two evolutionarily distant plants. Nat Chem Biol (2025). https://doi.org/10.1038/s41589-025-01926-z

Download citation

Received: 03 October 2024
Accepted: 30 April 2025
Published: 03 June 2025
DOI: https://doi.org/10.1038/s41589-025-01926-z

Subjects

Abstract

Similar content being viewed by others

Main

Results

Metabolite profiling of C. ipecacuanha and A. salviifolium

A Pictet–Spengler reaction in ipecac alkaloid biosynthesis

Ipecac alkaloid biosynthetic gene candidates

Comparative discovery of protoemetine biosynthesis

Discovery of species-specific R-epimer pathways

A vacuolar exporter enhances protoemetine biosynthesis

Nuclear-localized glucosidases deglycosylate R-epimers

Some biosynthetic genes may have evolved independently

Discussion

Methods

Plant material and sampling

Metabolite extraction for metabolite profiling

UPLC–MS/MS methods

MS data analysis

Nonenzymatic Pictet–Spengler reaction

Enzyme activity tests on native plant protein extracts

RNA extraction and RNA-seq

Transcriptome analysis and candidate selection

Gene cloning

Transient expression in N. benthamiana

Leaf disk assays

Metabolite extraction from N. benthamiana

Confocal laser scanning microscopy

Recombinant protein production and purification

Phylogenetic analyses

Protein models

Commercially available chemicals and standards

Chemically produced standards and substrates

Purification of epimer-pure DAII and DAI

NMR analysis

Enzymatically produced standards and substrates

In vitro competition activity assays of glucosidases

Production of protoemetinol standard

Reporting summary

Data availability

Code availability

References

Acknowledgements

Funding

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Competing interests

Peer review

Peer review information

Additional information

Extended data

Supplementary information

Source data

Rights and permissions

About this article

Cite this article

Share this article

Search

Quick links