UC Berkeley

Photosynthesis for the generation of isoprene in cyanobacteria was demonstrated with Synechocystis, entailing a process where a single host microorganism acts both as photocatalyst and processor, photosynthesizing and emitting isoprene hydrocarbons. A practical aspect of the commercial exploitation of this process in mass culture is the need to prevent invading microorganisms that might cause a culture to crash, and to provide an alternative to fresh water in scale-up applications. Growth media poised at alkaline pH are desirable in this respect, as high pH might favor the growth of the cyanobacteria, while at the same time discouraging the growth of invading predatory microbes and grazers. In addition, demonstration of salinity tolerance would enable the use of seawater for cyanobacteria cultivations. However, it is not known if Synechocystis growth and the isoprene-producing metabolism can be retained under such theoretically non-physiological conditions. We applied the gaseous/aqueous two-phase photobioreactor system with Synechocystis transformed with the codon-optimized isoprene synthase gene (SkIspS) of Pueraria montana (kudzu). Rates of growth and isoprene production are reported under control, and a combination of alkalinity and salinity conditions. The results showed that alkalinity and salinity do not exert a negative effect on either cell growth or isoprene rate and yield in Synechocystis. The work points to a practical approach in the design of cyanobacterial growth media for applications in commercial scale up and isoprene production.

The concept applied in this work entails application of an isoprene synthase transgene from terrestrial plants, heterologously expressed in cyanobacteria, thereby reprogramming carbon flux in the terpenoid biosynthetic pathway toward formation and spontaneous release of this volatile chemical from the cell and liquid culture. However, flux manipulations and carbon-partitioning reactions between isoprene (the product) and native terpenoid biosynthesis for cellular needs are not yet optimized for isoprene yield. The primary reactant for isoprene biosynthesis is dimethylallyl diphosphate (DMAPP), whereas both DMAPP and its isopentenyl diphosphate (IPP) isomer are needed for cellular terpenoid biosynthesis. This aspect of the work addressed the function of an isopentenyl diphosphate (IPP) isomerase in cyanobacteria and its role in carbon partitioning between IPP and DMAPP, both of which serve, in variable ratios, as reactants for the synthesis of different cellular terpenoids. The project was approached upon the heterologous expression in Synechocystis of the isopentenyl diphosphate isomerase gene (FNI) from Streptococcus pneumoniae, a non-photosynthetic bacterium, using isoprene production as a “reporter process” for substrate partitioning between DMAPP and IPP. It is shown that transgenic expression of the FNI gene in Synechocystis resulted in a 250% increase in the isoprene production rate and yield, suggesting that the FNI isomerase shifted the composition of the endogenous DMAPP-IPP steady-state pool toward DMAPP, thereby enhancing rates and yield of isoprene production.

Efforts to heterologously produce quantities of isoprene hydrocarbons (C5H8) renewably from CO2 and H2O through the photosynthesis of cyanobacteria face barriers, including low levels of recombinant enzyme accumulation compounded by their slow innate catalytic activity. This aspect of the work sought to alleviate the “expression level” barrier upon placing the isoprene synthase (IspS) enzyme in different fusion configurations with the cpcB protein, the highly expressed β-subunit of phycocyanin. Different cpcB*IspS fusion constructs were made, distinguished by the absence or presence of linker amino acids between the two proteins. Length and composition of linker amino acids was variable with lengths of 7, 10, 16 and 65 amino acids designed to confer optimal activity to the IspS through spatial positioning between the cpcB and IspS. Results showed that fusion constructs with the highly expressed cpcB gene, as the leader sequence, improved transgene expression in the range of 61- to 275-fold over what was measured with the unfused IspS control. However, the specific activity of the IspS enzyme was attenuated in all fusion transformants, possibly because of allosteric effects exerted by the leader cpcB fusion protein. This inhibition varied depending on the nature of the linker amino acids between the cpcB and IspS proteins. In terms of isoprene production, the results further showed a tradeoff between specific activity and transgenic enzyme accumulation. For example, the cpcB*L7*IspS strain showed only about 10% the isoprene synthase specific activity of the unfused cpcB-IspS control but it accumulated 254-fold more IspS enzyme. The latter more than countered the slower specific activity and made the cpcB*L7*IspS transformant the best isoprene-producing strain in this work. Isoprene to biomass yield ratios improved from 0.2 mg g-1 (isoprene to biomass, w:w) in the unfused cpcB-IspS control to 5.4 mg g-1 in the cpcB*L7*IspS strain, a 27-fold improvement.

Our final approach unified the lessons previously learned. The combination of an optimized cpcB*IspS fusion construct with the overexpression of the FNI protein resulted in up to a 60-fold increase in isoprene yield, relative to that obtained upon the expression of the IspS gene alone in Synechocystis, reaching a yield of 12.3 mg g-1 (isoprene to biomass, w:w). The latter is the highest verifiable constitutive photosynthetic isoprene production measured with homoplasmic lines, in which the heterologous isoprene synthase gene is encoded by the Synechocystis genomic DNA. Thus, the work constitutes a major step forward in the development of a cyanobacterial engineering platform for isoprene production.