INTRODUCTION

The genus Argemone (Papaveraceae) comprises more than 25 species. Its center of distribution includes Southwestern USA and Northern Mexico (^{Schwarzbach and Kadereit, 1999}). Several of those species are part of the native Mexican flora, but A. mexicana is the most successful and widely spread one. This plant, commonly known as chicalote or cardo in Mexico and as Mexican prickly poppy in the US, is considered a persistent and difficult to eradicate weed (^{Villaseñor y Espinosa, 1998}). Nevertheless, it is also regarded as a medicinal plant, used since prior the arrival of Europeans to America (^{Rubio-Piña and Vázquez-Flota, 2013}).

Furthermore, the plant has industrial application in the elaboration of environmentally-friendly agrochemicals (^{Hernández-Romero et al., 2005}). The bioactivity of this plant is attributed to the presence of benzylisoquinoline alkaloids (BIA’s), which are distributed through the entire plant (^{Brahmachari et al., 2013}). More than 30 BIA’s have been identified in A. mexicana; the most abundant ones correspond to sanguinarine and berberine, belonging to the benzophenanthridine- and protoberberine-types, respectively (Figure 1; ^{Rubio-Piña and Vázquez-Flota, 2013}). These alkaloids are distributed by following tissue-related patterns in mature plants. Sanguinarine is restricted to the roots and mature seeds, whereas berberine is found in both roots and aerial tissues (leaves and stems) (^{Rubio-Piña and Vázquez-Flota, 2013}). Remarkably, the alkaloid distribution in this species changes according to the developmental stages. Sanguinarine is not found in mature plants but it is detected in leaves of young plantlets instead, while berberine is absent in the underground tissues of developing seedlings (^{Xool-Tamayo et al., 2017a}).

In developing seedlings and young plantlets, the sites of alkaloid accumulation differ from those displaying the transcriptional activity related to its synthesis, suggesting the operation of complex regulatory mechanisms, which could involve posttranscriptional controls, long distance transport or enzyme catalytic flexibility (^{Xool-Tamayo et al., 2017a}, ^b). This aspect has not yet been explored in mature plants which simultaneously could accumulate higher amounts of both berberine and sanguinarine than developing tissues. Since the formation of these alkaloids shares a common set of early biosynthetic intermediaries, the identification of the actual sites involved in the process could reveal the mechanisms sorting them among the final products.

Several members of the Papaveraceae family produce sanguinarine and, in all cases, both synthesis and accumulation take place in roots (^{Hagel and Facchini, 2013}). In contrast, berberine, which rarely occurs in Papaveraceae but commonly occurs in Berberidaceae and Ranunculaceae, can either remain into the same tissue where it is formed, as it happens in Thalictrum flavum (^{Samanani et al., 2005}), or be exported to distantly located accumulation sites, as in Coptis japonica (^{Shitan et al., 2003}). BIA synthesis requires two units of tyrosine, one of them is deaminated by the action of aromatic monoamine oxidase (MAO), while the other one is hydroxylated and decarboxylated, involving the participation of tyrosine decarboxylase (TyDC). The resulting products, 4-hydrophenylacetaldehyde (4-HPA) and L-3, 4-dihydroxyphenylalanine respectively, are condensed forming norcoclaurine by the action of norcoclaurine synthase (NCS), which after two O- and one N-methylations is converted into reticuline (Figure 1).

The formation of an intramolecular carbon-carbon bond between the N-methyl group and the phenolic ring of reticuline, catalyzed by the berberine bridge enzyme (BBE), produces S-scoulerine which is the last common intermediary for the synthesis of both berberine and sanguinarine (Figure 1). This compound is used either by the S-scoulerine O-methyltransferase (SOMT) for berberine synthesis, or by the S-cheilanthifoline synthase (CheSyn) for sanguinarine synthesis (^{Hagel and Facchini, 2013}). The formation of a 2,3 methylenedioxy bridge (O-CH₂-O), catalyzed by the tetrahydroprotoberberine oxidase (STOX) is the last step in berberine biosynthesis (Figure 1), whereas the formation of an additional 9,10 methylenedioxy bridge and further oxidation are required for sanguinarine synthesis (Figure 1).

Most biosynthetic genes have been isolated in different species and the corresponding orthologous genes have been identified in transcriptomic collections from A. mexicana stems, cell cultures and developing seedlings (^{Farrow et al., 2012}; ^{Gesell et al., 2011}; ^{Hagel et al., 2015}; ^{Xiao et al., 2013}; ^{Xool-Tamayo et al., 2017a}). Moreover, in some cases heterologous expression has confirmed identity and functionality (^{Díaz et al., 2011}). The aim of this study was to set an approach to establish the possible site of alkaloid formation in mature plants. We followed the transcript accumulation of genes participating at different stages of the biosynthetic process and compared them to those of the main alkaloids, sanguinarine and berberine. This analysis included genes involved in the common reactions required for the synthesis of both alkaloids, as well as those participating specifically in one or another branch (Figure 1).

MATERIALS AND METHODS

RESULTS

Based on their chromatographic characteristics and mass fragmentation patterns, main alkaloids in A. mexicana corresponded to berberine (Rt 8.20 min; 336 [M+H]⁺ m/z), sanguinarine (Rt 9.17 min; 332 [M+H]⁺ m/z), dihydrosanguinarine (Rt 6.12 min; 334 [M+H]⁺ m/z), and chelerythrine (Rt 9.00 min; 348 [M+H]⁺ m/z). Combined, these alkaloids accounted for over 70 % of the total content. In addition, allocryptopine, protopine, canadine, columbanine and stylopine were also identified (Table 2). Alkaloids chelyrubine, N-methyl-stylopine and the biosynthetic intermediaries reticuline, methyl-coclaurine and scoulerine were also identified as minor components in the extracts.

A tissue-associated distribution was detected for the aforementioned alkaloids since benzophenanthridines (dihydrosanguinarine, chelerythrine, sanguinarine and stylopine), and protopines (allocryptopine and protopine), mainly occurred in roots, whereas protoberberines (berberine, canadine and columbamine) occurred in both aerial and underground parts (Table 2). Major alkaloids were quantified in leaves, stems and roots (Figure 2). Data shown correspond to the last plant collection, although similar results were obtained in all collected batches. Roots accumulated similar amounts of both berberine and sanguinarine (around 1.2 mg g DW^-1), with lower amounts of chelerythrine. In aerial parts, only berberine was detected, where it was present in higher amounts than in roots (2.0 mg g DW^-1) (Figure 2). Berberine in roots represented around 10 % in relation to the sanguinarine content, whereas chelerythrine, which is formed from dihydrosanguinarine (Figure 1), represented less than 5 % (Figure 2). In this way, roots from mature plants proved capable of accumulating alkaloids from the benzophenanthridine, protopine and protoberberine groups whereas aerial tissues only seem to have the capacity to accumulate protoberberine (Figure 2).

In order to assess the biosynthetic capacity of the A. mexicana tissues, the abundance of selected transcripts was estimated by qRT-PCR (Figure 3). Transcripts corresponding to genes involved in the early, common biosynthetic reactions (TyDC, NCS and BBE) were selected along with those participating specifically in either berberine (SOMT and STOX), or sanguinarine (CheSyn) synthesis (Figure 1). All the selected genes were expressed in the analyzed tissues; however, important differences in transcript relative abundances were noticed (Figure 3). Interestingly, even when roots showed the highest total alkaloid content (per dry matter unit; Figure 2), low transcript accumulation was observed for both genes involved in the early common biosynthetic reactions (TyDC, NCS and BBE) and likewise for those participating only in berberine synthesis (SOMT and STOX) (Figure 3). In contrast, CheSyn transcripts accumulated majorly in roots in comparison to the other organs (Figure 3). Stems and leaves showed similar levels of transcript accumulation in all cases, with the only exception for those corresponding to CheSyn, which were barely detectable in leaves, but accumulated at fair amounts in stems (Figure 3). It is interesting to notice that the two NCS and BBE isoforms were equally expressed in the analyzed tissues (Figure 3). In contrast to aerial tissues, roots displayed low gene expression associated to berberine synthesis (Figure 3).

The accumulation of CheSyn transcripts in stems is noteworthy (Figure 3) since no sanguinarine (nor other of the benzophenanthridine type alkaloids) was detected in aerial tissues (Figure 2). To further analyze the sanguinarine biosynthetic branch, DBO transcript accumulation was also included in this study (Figure 3). The putative AmDBO was previously retrieved from the A. mexicana seedling transcriptome (^{Xool-Tamayo et al., 2017b}). Interestingly, the amounts of AmDBO transcript were similar in stems and roots (Figure 3), regardless the differences of sanguinarine content (Figure 2), supporting the participation of stems in the synthesis of this alkaloid.

DISCUSSION

In A. mexicana, genes coding for enzymes involved in the common reactions (TyDC, NCS and BBE) were expressed in all the analyzed tissues; however, they showed lower levels, yet higher alkaloid contents in roots than those in the aerial tissues (Figures 2 and 3). Moreover, no preferential expression of any of the NCS or BBE isoforms was noticed in any of the tissues (Figure 3), what suggests that both isoforms might function either as redundant or additive copies, rather than being associated to a differential tissue-specific expression. No differential expression of these isoforms had previously been noticed neither in developing A. mexicana seedlings when exposed to diverse conditions, nor in vitro cultures exposed to a set of chemical elicitors (^{Guízar-González et al., 2016}; ^{Xool-Tamayo et al., 2017a}, ^b), which supports this interpretation.

Expression of the berberine-related SOMT and STOX genes (Figure 1) was found in leaves, stems and roots (Figure 3), coinciding with the presence of the alkaloid in the same tissues (Figure 2). However, as for the previous genes, lower expression was detected in roots, what suggests that some of the berberine accumulated in roots could have been mobilized from the aerial parts (^{Xool-Tamayo et al., 2017a}). Interestingly, even when CheSyn transcripts accumulated at the highest levels in roots, matching with the amounts of sanguinarine, significant transcript levels were also noticed in stems (Figure 3), where no sanguinarine was detected (Figure 2); such findings point towards a possible participation of stems in sanguinarine synthesis. To confirm this, a candidate gene for DBO, the enzyme involved in the last reaction (Figure 1), was included in the analysis (^{Xool-Tamayo et al., 2017b}). The corresponding transcripts accumulated mainly in stems (Figure 3), suggesting its possible participation and pointing towards the operation of a sanguinarine transport mechanism to roots.

Taken together, these data suggest that both the aerial and underground Argemone tissues are involved in the synthesis of its main alkaloids, regardless the site of final accumulation. Roots, displaying the major alkaloid accumulation, presented a lower transcriptional activity compared to stems and leaves (with the only exception of CheSyn). Moreover, stems could be involved in sanguinarine synthesis, where it does not accumulate. Although the operation of stem-to-root transport systems could account for these results, other alternatives must be considered.

Substrate flexibility has been recently described for some enzymes involved in the late reactions of BIA synthesis; for instance, a cytochrome P450 dependent oxidase (CYP719A13), could display both canadine- and stylopine-synthase activities (^{Díaz et al., 2011}). The products of the catalyzed reactions, canadine and stylopine, are required for berberine and sanguinarine synthesis, respectively. Cheilanthifonine, the product of CheSyn, is the substrate for sylopine synthase activity (Figure 1), which is also required for the synthesis of allocryptoine and protopine. These alkaloids have been detected in extracts of aerial parts (Table 2; ^{Díaz et al., 2011}) suggesting that CheSyn in stems and leaves could be participating in their synthesis, rather than taking part in sanguinarine synthesis; nevertheless, it should be kept in mind that only low amounts were detected. Moreover, many FAD depending oxidases from P. somniferum (PsFADOX) accept as substrates different intermediaries involved in late reactions of benzophenanthridine synthesis (^{Hagel and Facchini, 2013}). Possible orthologous genes have been recently reported in A. mexicana (^{Hagel et al., 2015}), suggesting that alternative biosynthetic pathways involving the participation of gene products with similar catalytic properties could be operating in roots and aerial parts.

Finally, the operation of posttranscriptional mechanisms could also account for the discrepancies between transcript and alkaloid accumulation. Posttranscriptional controls have been pointed out as a common trend for plant specialized metabolism; however, it mainly operates as a negative modulator, down-adjusting the effects of an active gene expression (^{Pichersky and Lewinsohn, 2011}). Interestingly, Argemone roots showed high sanguinarine amounts in presence of low biosynthetic related transcripts (Figures 2 and 3).

CONCLUSIONS

Alkaloid synthesis in A. mexicana plants seems to be a complex process, probably involving the participation of different tissues as well as multilevel regulatory mechanisms.