Harnessing Integrated Omics Approaches for Plant Specialized Metabolism Research: New Insights into Shikonin Biosynthesis

Plants produce an amazing diversity of specialized (secondary) metabolites, which in turn may underpin a plant’s adaptation to its specific ecological niche. Some of these metabolites have also become part of our every-day human lives—as flavorings, fragrances, insecticides, medicines and pigments. The biosynthesis of plant specialized metabolites often requires the co-regulation of underlying genes, proteins and pathway products (Schilmiller et al. 2012). Furthermore, the formation of such specialized metabolites often requires the synergism of multiple enzymes with distinct substrate and product specificities (Pichersky and Lewinsohn 2011, Yazaki et al. 2017). Armed with this knowledge, the elucidation of unknown biochemical steps and their transcriptional regulatory control is increasingly being guided by multi-omics, co-expression and integrated network analyses (see recent examples in Hao et al. 2017, Xu et al. 2017, Vannozzi et al. 2018). Shikonin, a red naphthoquinone pigment, is one such specialized metabolite, whose biosynthesis has been aided by applying these approaches. Shikonin and its derivatives are commonly produced by the roots of select Lithospermum, Alkanna and Echium species of the Boraginaceae. They are thought to fulfill protective roles against biotic agents in the rhizosphere (Yazaki et al. 2017). To humans, shikonin is historically prized for its colorant properties but has since found renewed use as a therapeutic agent (Yazaki 2017). Despite a wealth of information gained from the ‘industrial revolution’ of shikonin production in the 1970 s and 1980 s, the biochemical and genetic basis of only one dedicated shikonin biosynthetic step is known (i.e. p-hydroxybenzoate geranyltransferase, PGT) and many questions relating to its biosynthesis still remain (Fig. 1). One key question concerns the establishment of the naphthalene ring that is crucial for shikonin formation. Overview of the shikonin biosynthesis pathway in Lithospermum erythrorhizon cells. M9-cultured L. erythrorhizon cells grown in darkness accumulate large amounts of red pigmented shikonin derivatives (e.g. acetylshikonin, deoxyshikonin and α-methyl-n-butylshikonin). Geranyl diphosphate and p-hydroxybenzoic acid derived from the mevalonate pathway and shikimate pathway, respectively, are two key precursors of shikonin. The newly proposed mechanism for naphthalene ring formation in shikonin is labeled as i, ii, iii and briefly described in the text. The current status of each enzymatic reaction (black arrows) is indicated with colored circles. Red circles, enzyme activity confirmed (EAC) and its underlying gene(s) identified previously (GI); green circles, proposed enzymatic steps (PE) and candidate genes (CG) from this study; blue circles, EAC but no CG were identified. Molecular mechanisms pivotal for the secretion of shikonin and its derivatives into the apoplastic space (blue arrows) remains a key target for further investigation. In this issue, Takanashi et al. (2019) address this question by employing a hypothesis-driven systems biology approach, combining high-throughput next-generation sequencing and shotgun proteomics techniques, in Lithospermum erythrorhizon. In the first phase, the authors exploited the inducibility of shikonin biosynthesis in this system to capture the gene repertoire relevant to its regulation using de novo RNA sequencing (RNA-seq) analysis. Three combinatorial scenarios for shikonin production were considered using (i) dedifferentiated cells cultured in M9 medium in darkness (on) or in the light (off); (ii) hairy root cultures in darkness (on) or in the light (off); and (iii) intact root bark (on) and peeled root (off). In the second phase, the authors performed an in-depth proteome analysis using liquid chromatography–tandem mass spectrometry techniques (LC-MS/MS) of M9-cultured L. erythrorhizon cells grown in darkness (M9-dark) or in the light (M9-light). These two contrasts were chosen because they showed the greatest expression differences from an initial transcriptome evaluation. As an additional contrast, the authors profiled L. erythrorhizon cells grown in darkness in Murashige and Skoog (MS) medium (MS-dark), a condition suited for cell growth but suboptimal for shikonin production (Yazaki 2017). Metabolite analysis also determined the presence of shikonin and its derivatives in all settings considered, with the greatest abundance observed in the M9-dark condition. In total the authors found close to 3,200 expressed proteins, with a higher presence of metabolic proteins for cells grown in M9 compared with MS medium. Those involved in sugar metabolism, quinone, peroxidase and polyphenol oxidase activities were often enriched in dark conditions, while pyrimidine, purine and cell wall proteins were enriched in MS- compared with M9-grown cells. Hierarchical clustering analysis on the protein abundance profiles revealed eight clusters with distinct patterns. In particular, the authors identified one cluster (cluster B) whose profile shows expression specific to M9-dark conditions and closely mirrors the disproportionally high levels of shikonin derivatives present within them and vice versa in M9-light and MS-dark conditions. Based on the notion that plant specialized metabolism pathway components are often co-regulated (Schilmiller et al. 2012), the authors reasoned that candidates involved in shikonin biosynthesis should be grouped in cluster B. Multiple candidate enzymes of potential downstream biochemical steps were identified from this cluster. This led to a new proposed pathway for the naphthalene ring formation whereby several novel intermediates catalyzed by these newly identified candidate enzymes are predicted (Fig. 1). The key features of this include: (i) cyclization of the geranyl side chain of 3’’-hydroxy-geranylhydroquinone to form a novel phenol-bearing intermediate (intermediate A) by cannabidiolic acid synthase (CBDAS) homologs; (ii) oxidation of intermediate A to a quinone-bearing compound (intermediate B) by short-chain dehydrogenase/reductases (SDR); and (iii) oxidation of intermediates A and B by polyphenol oxidases (PPOs) to yield the penultimate precursor to shikonin. In addition, three plant clade III BAHD acyltransferases (O-acetyltransferase) were predicted to be involved in the O-acetylation of shikonin. Their findings also reconciled their transcriptomic analyses where corresponding transcripts were specifically induced in other shikonin-rich tissues (e.g. root bark and hairy roots cultured in darkness). Beyond the primary goal of this study, candidate enzymes of other specialized metabolites (e.g. the shikimate pathway derivatives, rosmarinic acid and lithospermic acid B) were also prioritized. Nonetheless, functional characterization of pathway candidates is still essential to ascertain their involvement in various proposed steps. Despite being a non-model system, the various breakthroughs of this study were made possible by several key factors: (i) a considerable background knowledge of the putative biosynthetic precursors as well as the chemical and physical factors affecting their biosynthesis; (ii) the compounds of interest are produced in a tissue- and condition-specific manner (Yazaki 2017); and (iii) the strategic application of multi-omics approaches (i.e. comparative transcriptome and proteome analysis in this study) on contrasting active (on) and non-active (off) tissues to aid rapid and robust candidate prioritization. These simple guidelines make the biochemical investigation of other specialized metabolites in non-model plant systems both feasible and rewarding (Wong et al. 2017). While a comprehensive pathway for shikonin biosynthesis has now been proposed for the first time (Fig. 1), the molecular mechanisms relating to the transport of shikonin between cellular compartments remains a key target for further investigation (Tatsumi et al. 2016). A careful experimental set-up involving the use of membrane trafficking inhibitors, single-cell transcriptomics and subcellular proteome analysis may aid our future understanding of these critical processes. The author has no conflicts of interest to declare.