Establishing novel synthetic routes for microbial production of chemicals often requires overcoming pathway bottlenecks by tailoring enzymes to enhance bio-catalysis or even achieve non-native catalysis. Diol dehydratases have been extensively studied for their interactions with C2 and C3 diols. However, attempts on utilizing these insights to enable catalysis on non-native substrates with more than two hydroxyl groups have been plagued with low efficiencies. Here, we rationally engineered the Klebsiella oxytoca diol dehydratase to enable and enhance catalytic activity toward a non-native C4 triol, 1,2,4-butanetriol. We analyzed dehydratase's interaction with 1,2-propanediol and glycerol, which led us to develop rationally conceived hypotheses. An in silico approach was then developed to identify and screen candidate mutants with desired activity. This led to an engineered diol dehydratase with nearly 5 fold higher catalytic activity toward 1,2,4-butanetriol than the wild type as determined by in vitro assays. Based on this result, we then expanded the 1,2,4-butanetriol pathway to establish a novel 1,4-butanediol production platform. We engineered Escherichia coli's xylose catabolism to enhance the biosynthesis of 1,2,4-butanetriol from 224mg/L to 1506mg/L. By introducing the complete pathway in the engineered strain we achieve de novo biosynthesis of 1,4-butanediol at 209mg/L from xylose. This work expands the repertoire of substrates catalyzed by diol dehydratases and serves as an elucidation to establish novel biosynthetic pathways involving dehydratase based biocatalysis.