Terpene synthesis in the majority of bacterial species, together with plant plastids, takes place via the 1-deoxy-D-xylulose 5-phosphate (DXP) pathway. The first step of this pathway involves the condensation of pyruvate and glyceraldehyde 3-phosphate by DXP synthase (Dxs), with one-sixth of the carbon lost as CO 2 . A hypothetical novel route from a pentose phosphate to DXP (nDXP) could enable a more direct pathway from C 5 sugars to terpenes and also circumvent regulatory mechanisms that control Dxs, but there is no enzyme known that can convert a sugar into its 1-deoxy equivalent. Employing a selection for complementation of a dxs deletion in Escherichia coli grown on xylose as the sole carbon source, we uncovered two candidate nDXP genes. Complementation was achieved either via overexpression of the wild-type E. coli yajO gene, annotated as a putative xylose reductase, or via various mutations in the native ribB gene. In vitro analysis performed with purified YajO and mutant RibB proteins revealed that DXP was synthesized in both cases from ribulose 5-phosphate (Ru5P). We demonstrate the utility of these genes for microbial terpene biosynthesis by engineering the DXP pathway in E. coli for production of the sesquiterpene bisabolene, a candidate biodiesel. To further improve flux into the pathway from Ru5P, nDXP enzymes were expressed as fusions to DXP reductase (Dxr), the second enzyme in the DXP pathway. Expression of a Dxr-RibB(G108S) fusion improved bisabolene titers more than 4-fold and alleviated accumulation of intracellular DXP.T erpenes constitute a very large family of natural products, members of which are produced in virtually all free-living organisms (1). The diverse array of structures within the terpene family is reflected by the variety of applications in society, ranging from nutrition (carotenoids) and medicine (artemisinin, paclitaxel [originally taxol]) to industrial materials (isoprene, linalool) and candidate biofuels (farnesene, bisabolene, pinene) (2). Terpenes can be synthesized via either the mevalonate pathway or the 1-deoxy-D-xylulose 5-phosphate (DXP) pathway, the former predominating in the eukaryotic cytosol and the latter in plastids, while prokaryotes may contain either pathway or, in some cases, both pathways (3).Commercial-scale terpene production has been demonstrated in a variety of organisms, for example, carotenoids (via the DXP pathway) in algae (4, 5), paclitaxel (via DXP) in Taxus sp. (6), and artemisinin (via mevalonate) in Saccharomyces cerevisiae (7). Key metrics in determining the likelihood of commercial viability, particularly when targeting terpenes valued within the range of commodity chemicals or biofuels, are yield, productivity, and titer (8, 9). Of the two metabolic routes, the DXP pathway is considered the better option from the viewpoint of pathway efficiencyfor example, the theoretical maximum yield of isoprene from glucose is around 20% higher when synthesized via DXP instead of mevalonate (9, 10). In considering production of terpenes from hemicellul...