Cyanobacterial biofuels: new insights and strain design strategies revealed by computational modeling

Abstract

Background: Cyanobacteria are increasingly recognized as promising cell factories for the production of renewable biofuels and chemical feedstocks from sunlight, CO2, and water. However, most biotechnological applications of these organisms are still characterized by low yields. Increasing the production performance of cyanobacteria remains therefore a crucial step.

Results: In this work we use a stoichiometric network model of Synechocystis sp. PCC 6803 in combination with CASOP and minimal cut set analysis to systematically identify and characterize suitable strain design strategies for biofuel synthesis, specifically for ethanol and isobutanol. As a key result, improving upon other works, we demonstrate that higher-order knockout strategies exist in the model that lead to coupling of growth with high-yield biofuel synthesis under phototrophic conditions. Enumerating all potential knockout strategies (cut sets) reveals a unifying principle behind the identified strain designs, namely to reduce the ratio of ATP to NADPH produced by the photosynthetic electron transport chain. Accordingly, suitable knockout strategies seek to block cyclic and other alternate electron flows, such that ATP and NADPH are exclusively synthesized via the linear electron flow whose ATP/NADPH ratio is below that required for biomass synthesis. The products of interest are then utilized by the cell as sinks for reduction equivalents in excess. Importantly, the calculated intervention strategies do not rely on the assumption of optimal growth and they ensure that maintenance metabolism in the absence of light remains feasible. Our analyses furthermore suggest that a moderately increased ATP turnover, realized, for example, by ATP futile cycles or other ATP wasting mechanisms, represents a promising target to achieve increased biofuel yields.

Conclusion: Our study reveals key principles of rational metabolic engineering strategies in cyanobacteria towards biofuel production. The results clearly show that achieving obligatory coupling of growth and product synthesis in photosynthetic bacteria requires fundamentally different intervention strategies compared to heterotrophic organisms.

Figure 1

Selected key reactions of linear and alternate electron flow pathways in Synechocystis sp. PCC 6803 contained in the RN and GN model. Dashed arrows with H ⁺ represent release or pumping of protons into the thylakoid lumen. The resulting proton motive force is then used for ATP synthesis via ATPase. The two boxes at the bottom (right) display ATP and NADPH stoichiometries of electron flow pathways and for biomass and ethanol synthesis. Abbreviations: PSI/PSII: photosystem I and II, Cyt b ₆ f: cytochrome b ₆ f, Cox: cytochrome c oxidase, Cyd: cytochrome bd-type quinol oxidase, FNR: ferredoxin NADP reductase, ATPase: ATP synthase, FQR: ferredoxin quinone reductase, NDH I: NADPH dehydrogenase I, PQ: plastoquinone, PC: Plastocyanin, Fd: Ferredoxin, LEF: linear electron flow, CEF: cyclic electron flow, AEF: alternate electron flow, e ⁻: electrons.

Figure 2

A simple reaction network with its elementary modes, minimal cut sets and constrained minimal cut sets. There are three minimal cut sets blocking synthesis of P1 of which two remain as constrained minimal cut sets if production of P2 is a desired function to be kept.

Figure 3

Elementary modes and constrained minimal cut sets in the RN model. (A) Phenotypic phase plane depicting the specific biomass and ethanol yields of the EMs (day conditions). Each blue circle corresponds to one or several EMs. (B) For cMCSs calculation, EMs from (A) are classified as target and desired EMs by specifying thresholds for minimum desired biomass and product yield. Red circles represent target modes (Y _{ethanol/photon}≤0.03) and green circles desired modes (Y _{ethanol/photon}>0.03 and Y _{biomass/photon}≥0.0001). Blue circles indicate modes that are neither target nor desired modes. (C) Distribution of the cardinalities of cMCSs calculated from the intervention problem posed in (B) (see also Figure 5). (D) Phenotypic phase plane with the remaining EMs of the mutant resulting from a knockout of all reactions contained in cMCS-1 in Figure 5.

Figure 4

Intervention strategies suggested by CASOP and cut set analysis in the RN model. The red crosses represent suggested knockout targets. Cut set analysis reveals that deletion of all these targets blocks all CEF/AEF pathways and leads thus to a fixed ATP/NADPH ratio of 1.28 generated by photosynthesis via the remaining LEF (thick arrows). Ethanol synthesis becomes then mandatory to readjust the ATP/NADPH balance for biomass synthesis. The mechanisms (a)-(d) show overexpression (or flux enhancement) targets suggested by CASOP all of which will enforce an increased turnover (wasting) of ATP. For abbreviations and general explanations see Figure 1.

Figure 5

The complete set of intervention strategies (cMCSs) in the RN model enforcing high-yield growth-coupled ethanol synthesis. For reaction IDs see Table 1.

Figure 6

Influence of the ATP maintenance demand (ATPm) and the requested minimal ethanol yield on the minimum number of required interventions in the resulting cMCSs.