Author Correction: Complete biosynthesis of cannabinoids and their unnatural analogues in yeast

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

De novo synthesis of the sedative valerenic acid in Saccharomyces cerevisiae

Valeriana officinalis (Valerian) root extracts have been used by European and Asian cultures for millennia for their anxiolytic and sedative properties. However, the efficacy of these extracts suffers from variable yields and composition, making these extracts a prime candidate for microbial production. Recently, valerenic acid, a C15 sesquiterpenoid, was identified as the active compound that modulates the GABA_A channel. Although the first committed step, valerena-4,7(11)-diene synthase, has been identified and described, the complete valerenic acid biosynthetic pathway remains to be elucidated. Sequence homology and tissue-specific expression profiles of V. officinalis putative P450s led to the discovery of a V. officinalis valerena-4,7(11)-diene oxidase, VoCYP71DJ1, which required coexpression with a V. officinalis alcohol dehydrogenase and aldehyde dehydrogenase to complete valerenic acid biosynthesis in yeast. Further, we demonstrated the stable integration of all pathway enzymes in yeast, resulting in the production of 140 mg/L of valerena-4,7(11)-diene and 4 mg/L of valerenic acid in milliliter plates. These findings showcase Saccharomyces cerevisiae's potential as an expression platform for facilitating multiply-oxidized medicinal terpenoid pathway discovery, possibly paving the way for scale up and FDA approval of valerenic acid and other active compounds from plant-derived herbal medicines.

Engineering high-level production of fatty alcohols by Saccharomyces cerevisiae from lignocellulosic feedstocks

Fatty alcohols in the C12-C18 range are used in personal care products, lubricants, and potentially biofuels. These compounds can be produced from the fatty acid pathway by a fatty acid reductase (FAR), yet yields from the preferred industrial host Saccharomyces cerevisiae remain under 2% of the theoretical maximum from glucose. Here we improved titer and yield of fatty alcohols using an approach involving quantitative analysis of protein levels and metabolic flux, engineering enzyme level and localization, pull-push-block engineering of carbon flux, and cofactor balancing. We compared four heterologous FARs, finding highest activity and endoplasmic reticulum localization from a Mus musculus FAR. After screening an additional twenty-one single-gene edits, we identified increasing FAR expression; deleting competing reactions encoded by DGA1, HFD1, and ADH6; overexpressing a mutant acetyl-CoA carboxylase; limiting NADPH and carbon usage by the glutamate dehydrogenase encoded by GDH1; and overexpressing the Δ9-desaturase encoded by OLE1 as successful strategies to improve titer. Our final strain produced 1.2g/L fatty alcohols in shake flasks, and 6.0g/L in fed-batch fermentation, corresponding to ~ 20% of the maximum theoretical yield from glucose, the highest titers and yields reported to date in S. cerevisiae. We further demonstrate high-level production from lignocellulosic feedstocks derived from ionic-liquid treated switchgrass and sorghum, reaching 0.7g/L in shake flasks. Altogether, our work represents progress towards efficient and renewable microbial production of fatty acid-derived products.

Superiority of yeast over bacterial cytosine deaminase for enzyme/prodrug gene therapy in colon cancer xenografts

The enzyme/prodrug strategy using bacterial cytosine deaminase (bCD) and 5-fluorocytosine (5-FC) is currently under investigation for cancer gene therapy. A major limitation for the use of bCD is that it is inefficient in the conversion of 5-FC into 5-fluorouracil. In the present study, we show that the K(m) of yeast cytosine deaminase (yCD) for 5-FC was 22-fold lower when compared with that of bCD. HT29 human colon cancer cells transduced with yCD (HT29/yCD) were significantly more sensitive to 5-FC in vitro than HT29 cells transduced with bCD (HT29/bCD). In tumor-bearing nude mice, complete tumor regression was observed in 6 of 13 HT29/yCD tumors in response to 5-FC treatment (500 mg/kg i.p. daily, 5 days a week for 2 weeks), whereas 0 of 10 HT29/bCD tumors were cured. Our study demonstrates an improved efficacy of the CD/5-FC treatment strategy when yCD was used. This enzyme has, therefore, a high potential to increase the therapeutic outcome of the enzyme/prodrug strategy in cancer patients.