Uridine-derived ribose fuels glucose-restricted pancreatic cancer.

Pancreatic ductal adenocarcinoma (PDA) is a lethal disease notoriously resistant to therapy1,2. This is mediated in part by a complex tumour microenvironment3, low vascularity4, and metabolic aberrations5,6. Although altered metabolism drives tumour progression, the spectrum of metabolites used as nutrients by PDA remains largely unknown. Here we identified uridine as a fuel for PDA in glucose-deprived conditions by assessing how more than 175 metabolites impacted metabolic activity in 21 pancreatic cell lines under nutrient restriction. Uridine utilization strongly correlated with the expression of uridine phosphorylase 1 (UPP1), which we demonstrate liberates uridine-derived ribose to fuel central carbon metabolism and thereby support redox balance, survival and proliferation in glucose-restricted PDA cells. In PDA, UPP1 is regulated by KRAS-MAPK signalling and is augmented by nutrient restriction. Consistently, tumours expressed high UPP1 compared with non-tumoural tissues, and UPP1 expression correlated with poor survival in cohorts of patients with PDA. Uridine is available in the tumour microenvironment, and we demonstrated that uridine-derived ribose is actively catabolized in tumours. Finally, UPP1 deletion restricted the ability of PDA cells to use uridine and blunted tumour growth in immunocompetent mouse models. Our data identify uridine utilization as an important compensatory metabolic process in nutrient-deprived PDA cells, suggesting a novel metabolic axis for PDA therapy.

Regulation of THI and PDC genes by Pdc2 in Nakaseomyces glabratus (Candida glabrata) is complex.

Thiamine (vitamin B1) is essential for glucose catabolism. In the yeast species Nakaseomyces glabratus (formerly Candida glabrata) and Saccharomyces cerevisiae, the transcription factor Pdc2 (with Thi3 and Thi2) upregulates pyruvate decarboxylase (PDC) genes and thiamine biosynthetic and acquisition (THI) genes during starvation. There have not been genome-wide analyses of Pdc2 binding. Previously, we identified small regions of Pdc2 regulated genes sufficient to confer thiamine regulation. Here, we performed deletion analyses on these regions. We observed that when the S. cerevisiae PDC5 promoter is introduced into N. glabratus, it is thiamine starvation inducible but does not require the Thi3 coregulator. The ScPDC5 promoter contains a 22 bp duplication with an AT-rich spacer between the two repeats, which are important for regulation. Loss of the first 22 bp element does not eliminate regulation, but the promoter becomes Thi3-dependent, suggesting cis architecture can generate a Thi3-independent, thiamine starvation inducible response. Whereas many THI promoters only have one copy of this element, addition of the first 22 bp element to a Thi3-dependent promoter confers Thi3-independence. Finally, we performed fluorescence anisotropy and ChIP-seq. Pdc2 and Thi3 bind to regions that share similarity to the 22 bp element in the ScPDC5 promoter and previously identified cis elements in N. glabratus promoters. Also, while Pdc2 binds to THI and PDC promoters, neither Pdc2 nor Thi3 appear to bind the evolutionarily new NgPMU3 promoter that is regulated by Pdc2. Further study is warranted because PMU3 is required for cells to acquire thiamine from environments where thiamine is phosphorylated, such as in the human bloodstream.

Pyruvate decarboxylase and thiamine biosynthetic genes are regulated differently by Pdc2 in S. cerevisiae and C. glabrata.

Understanding metabolism in the pathogen Candida glabrata is key to identifying new targets for antifungals. The thiamine biosynthetic (THI) pathway is partially defective in C. glabrata, but the transcription factor CgPdc2 upregulates some thiamine biosynthetic and transport genes. One of these genes encodes a recently evolved thiamine pyrophosphatase (CgPMU3) that is critical for accessing external thiamine. Here, we demonstrate that CgPdc2 primarily regulates THI genes. In Saccharomyces cerevisiae, Pdc2 regulates both THI and pyruvate decarboxylase (PDC) genes, with PDC proteins being a major thiamine sink. Deletion of PDC2 is lethal in S. cerevisiae in standard growth conditions, but not in C. glabrata. We uncover cryptic cis elements in C. glabrata PDC promoters that still allow for regulation by ScPdc2, even when that regulation is not apparent in C. glabrata. C. glabrata lacks Thi2, and it is likely that inclusion of Thi2 into transcriptional regulation in S. cerevisiae allows for a more complex regulation pattern and regulation of THI and PDC genes. We present evidence that Pdc2 functions independent of Thi2 and Thi3 in both species. The C-terminal activation domain of Pdc2 is intrinsically disordered and critical for species differences. Truncation of the disordered domains leads to a gradual loss of activity. Through a series of cross species complementation assays of transcription, we suggest that there are multiple Pdc2-containing complexes, and C. glabrata appears to have the simplest requirement set for THI genes, except for CgPMU3. CgPMU3 has different cis requirements, but still requires Pdc2 and Thi3 to be upregulated by thiamine starvation. We identify the minimal region sufficient for thiamine regulation in CgTHI20, CgPMU3, and ScPDC5 promoters. Defining the cis and trans requirements for THI promoters should lead to an understanding of how to interrupt their upregulation and provide targets in metabolism for antifungals.