Synthetic biosystems for the production of high-value plant metabolites Peter J. Facchini1, Joerg Bohlmann2, Patrick S. Covello3, Vincenzo De Luca4, Radhakrishnan Mahadevan5, Jonathan E. Page3, Dae-Kyun Ro1, Christoph W. Sensen6, Reginald Storms7 and Vincent J.J. Martin7 1 Department of Biological Sciences, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, T2N 1N4, Canada 2 Michael Smith Laboratories, University of British Columbia, 321-2185 East Mall, Vancouver, British Columbia, V6T 1Z4, Canada 3 National Research Council ��� Plant Biotechnology Institute, 110 Gymnasium Place, Saskatoon, Saskatchewan, S7N 0W9, Canada 4 Department of Biological Sciences, Brock University, 500 Glenridge Avenue, St. Catharines, Ontario, L2S 3A1, Canada 5 Department of Chemical Engineering and Applied Chemistry, University of Toronto, 200 College Street, Toronto, Ontario, M5S 3E5, Canada 6 Department of Biochemistry and Molecular Biology, University of Calgary, 3330 Hospital Drive N.W., Calgary, Alberta, T2N 4N1, Canada 7 Department of Biology, Concordia University, 7141 rue Sherbrooke West, Montreal, �� Quebec, �� H4B 1R6, Canada Plants display an immense diversity of specialized meta- bolites, many of which have been important to humanity as medicines, flavors, fragrances, pigments, insecticides and other fine chemicals. Apparently, much of the varia- tion in plant specialized metabolism evolved through events of gene duplications followed by neo- or sub- functionalization. Most of the catalytic diversity of plant enzymes is unexplored since previous biochemical and genomics efforts have focused on a relatively small number of species. Interdisciplinary research in plant genomics, microbial engineering and synthetic biology provides an opportunity to accelerate the discovery of new enzymes. The massive identification, characteriza- tion and cataloguing of plant enzymes coupled with their deployment in metabolically optimized microbes provide a high-throughput functional genomics tool and a novel strain engineering pipeline. Introduction Plants produce a bewildering array of specialized metabo- lites based on a myriad of skeletal structures and function- al group combinations [1,2]. Economically, many of these compounds are among the most valuable bioproducts in their respective markets (Figure 1). Given their enormous structural diversity and the equally staggering numbers of species in which specialized metabolites are produced, the impressive biosynthetic potential of plant metabolism has long been recognized. As with other organisms, the imple- mentation of genomics has expedited our understanding of many basic biological processes in plants [3]. Initial ge- nome-sequencing and genome-assembly efforts focused on a few model species or major crops, such as Arabidopsis, rice, poplar, grapevine, tobacco and Medicago truncatula [4]. Although an increasing number of plants from all major taxonomic lineages are currently targeted for ge- nome sequencing, the selection of species is generally not based on maximizing coverage of diverse specialized me- tabolism. As a result, the diversity of genes encoding enzymes involved in plant specialized metabolism is not well represented, or at least not well annotated in public DNA sequence databases. Tapping into the biochemistry that is often specific to particular plant species via the characterization and cataloguing of potentially novel enzymes requires the establishment of genome or tran- scriptome sequence resources for a large number of plants producing a sufficient diversity of specialized metabolites [5]. The discovery of genes encoding previously unknown biosynthetic enzymes will also require comprehensive me- tabolite profiling of the same plant species, and an effective integration of genomics and metabolomics datasets [6]. The PhytoMetaSyn Project represents a consortium of researchers from across Canada with the following princi- pal objectives: (i) the establishment of a genomics pipeline that integrates massively parallel DNA sequencing, tar- geted metabolomics, advanced bioinformatics and ���plug- and-play��� functional genomics in yeast to efficiently identi- fy, characterize and catalog a continuously expanding collection of biosynthetic genes responsible for the im- mense chemical diversity of plant metabolism (ii) the development of a framework for the commercial production of valuable plant natural products in microbial systems through the optimized production of primary metabolic precursors and (iii) a demonstration of the feasibility of synthetic biology as a platform for the production of six prototype plant natural products in engineered yeast (Figure 2). Prototype natural products (Figure 1) were selected based on their existing or potential commercial markets, the availability of information about their bio- synthetic pathways in plants, and the scientific expertise of PhytoMetaSyn Project investigators. Opinion Corresponding authors: Facchini, P.J. (pfacchin@ucalgary.ca) Martin, V.J.J. (vmartin@alcor.concordia.ca). 0167-7799/$ ��� see front matter �� 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.tibtech.2011.10.001 Trends in Biotechnology, March 2012, Vol. 30, No. 3 127

Gene discovery: accelerated expansion of the parts catalog Genome and transcriptome mining has become the strate- gy of choice for the discovery of genes encoding enzymes responsible for the multistep formation of plant specialized metabolites. Until recently, the most common mining ap- proach involved generating an expressed sequence tag (EST) database based on traditional Sanger sequencing for a plant producing one or more natural products of interest. Generally, candidate genes have been expressed in a microorganism and the recombinant enzymes func- tionally characterized in vitro. Historically, the ability to isolate and identify a novel enzyme is largely dependent on two factors: (i) the representation of gene candidates in often relatively modest EST collections and (ii) the avail- ability of substrates, which are typically not available commercially. Recently, massively parallel sequencing technologies have dramatically extended the opportunity to establish deep transcriptome databases containing a high-percentage of full-length open reading frames for any organism [7,8]. Owing to the accumulation of different, yet often structurally similar specialized metabolites in related plant taxa, high-quality sequence databases for a variety of related species will yield a large repository of biosynthetic gene candidates with different biochemical functions (e.g. substrate- or regio-specificity). The effectiveness of selecting gene candidates that en- code enzymes with predictable metabolic functions is clear- ly proportional to the amount of information available about the catalytic steps in a particular biosynthetic path- way. Genomics resources must be supported by effective bioinformatics tools beyond simple BLASTX annotations to facilitate the rapid identification of novel biosynthetic genes and the establishment of empirical evidence for hypothetical metabolic pathways. Key tools include in- depth functional domain identification through more sen- sitive database comparisons and modeling to identify can- didate genes, as well as the hierarchical clustering of homologous genes from various plants to allow seamless inclusion of new species sequence databases as they be- come available. Convenient platforms for the unambiguous identification of many plant metabolites have been provid- ed through recent advances in mass spectrometry. The integration of deep transcript and targeted metabolite profiles from corresponding plant tissues is a key compo- nent in establishing an integrated platform to select bio- synthetic gene candidates involved in diverse natural product metabolism. Although a fully elucidated pathway is not essential to discover novel biosynthetic genes, more extensive structural characterization will be required in cases where reaction products have not been previously analyzed. H H HO O Abietic acid O Nootkatone HO O HO H H HCH3 Morphine NH H H H O O OGluc N H OCH3 CH3 CH3 H3C Strictosidine HO O O OH OH Xanthohumol HO H H H H OH O Betulinic acid TRENDS in Biotechnology Figure 1. Six natural products and selected source plants. The compounds represent a sesquiterpene (nootkatone from Citrus paradisi), a diterpene (abietic acid from Pinus sylvestris), a triterpene (betulinic acid from Betula alba), a benzylisoquinoline alkaloid (morphine from Papaver somniferum), a monoterpenoid indole alkaloid (strictosidine) and a polyketide (xanthohumol from Humulus lupulus). Opinion Trends in Biotechnology March 2012, Vol. 30, No. 3 128