Improving the Efficiency of Microbial Isobutanol Synthesis: Directed Evolution Reveals Improved Enzymes and Optimal Enzyme Ratios

Efficient microbial synthesis of the specialty biofuel isobutanol is crucial for bioindustrial production, but efficiency is limited by significant bottlenecks within isobutanol-producing metabolic pathways.^1,2 Maximal isobutanol titers will require anaerobic production, which while demonstrated feasible,³ has not been fully realized due to these bottlenecks. Harnessing directed evolution, we are optimizing the anaerobic synthesis of isobutanol in Escherichia coli and Zymomonas mobilis, a bacterial strain capable of high-flux sugar catabolism.

For fermentative production of isobutanol in microbes, a valine precursor: α-ketoisovalerate is overproduced from pyruvate using AlsS (Acetolactate synthase), IlvC (Ketol-acid reductoisomerase, NADH dependent) and IlvD (Dihydroxy-acid dehydratase), then converted to isobutanol by Kivd (Ketoisovalerate decarboxylase) and AdhA (Alcohol dehydrogenase, NADH dependent).^3,4 Kinetic predictions and previous reports indicate that α-ketoisovalerate is likely to be limiting due to low catalytic rates of IlvC and IlvD.¹ This issue is compounded in Z. mobilis, where AlsS has a lower affinity for pyruvate than the competing enzyme Pdc (Pyruvate decarboxylase).⁵ We aimed to mitigate these bottlenecks by engineered improved versions of these enzymes and identifying optimal ratios of enzyme expression.

We exploited the growth dependence of fermentation-deficient E. coli on near-quantitative isobutanol conversion from glucose to identify genetic perturbations that improve metabolic flux. By combinatorically varying enzyme expression levels in an OptSSeq⁶ experiment, we identified ratios of enzyme expression that ameriolated bottlenecks in efficient isobutanol synthesis. Additionally, we identified beneficial amino-acid substitutions within enzymes that likely improved catalytic rates, thus enabling even higher efficiency of isobutanol synthesis. We are characterizing the improved enzymatic variants to determine their impacts, and measuring in-vivo levels of intermediate metabolites to confirm that our optimized expression cassettes reduce bottlenecks within the pathway.

Initial experiments confirmed difficulties in producing isobutanol in Z. mobilis. Insights from our studies in E. coli are driving tunings in expression ratios of enzymes with improved catalytic rates. We are also testing various methods of knocking down Pdc in Z. mobilis. Our studies will guide significant improvements in microbial production of isobutanol, including from the industrially relevant Z. mobilis.

¹Generoso, et al (2015) Curr op in biot 33:1-7
²Akita, et al (2015) Appl Microbiol Biotechnol 99:991-999
³Bastian, et al (2011) Metab engg 13:345-352
⁴Atsumi, et al (2008) Nature 451:86-89
⁵Neale et al (1987) J Bacteriol 169:1024-1028
⁶Ghosh, et al (2016) ACS synth biol 5:1519-1534