Production of fungal quinones as electrolytes in redox flow batteries

Renewable energy solutions are important assets if society is to reach the sustainability goals put forth by the United Nations (https://sdgs.un.org/goals/goal7) and fight climate change. However, as opposed to fossil fuel derived energy production, renewable energy from wind and sun is only produced when the wind is blowing and the sun is shining. Thus, in order to make renewable energy competitive, energy storage technologies are required. Conventional batteries are unsuitable for storing large amounts of energy, as they are difficult to scale, require rare-earth metals and are extremely hazardous, e.g. combustable. Quinone redox flow batteries (RFBs) have been suggested as viable alternatives to conventional batteries. Quinones are redox-active small organic molecules that can be sourced from living organisms such as plants, bacteria and filamentous fungi. Filamentous fungi have shown to be particularly good quinone producers, due to their large content of biosynthetic gene cluster encoding this type of polyketides. Filamentous fungi can also be cultivated in industrial settings, for example in bioreactors, and are thus not under influence of seasonal effects, in contrast to plants.

In the work presented in this thesis, we investigated the feasibility of using fungal quinones in RFBs. The thesis is structured as a research and development pipeline, first detailing initial screening efforts to find the most promising fungal strains for quinone production, secondly addressing optimization of production. Lastly, a fungal quinone (phoenicin) was tested in an RFB at laboratory scale as proof of concept, showing that quinones sourced from fungal cultivation can indeed be used in RFBs.

Paper 1 is a mini-review that details the known quinones observed in filamentous fungi of the fungal genera Penicillium, Aspergillus, Talaromyces, Arthrinium, Fusarium and Alternaria. The paper is presented as part of the Introduction chapter and provides examples of the chemical diverse quinones found in nature and especially in filamentous fungi, including their biological purpose and how they can be applied industrially. We categorized quinones into “quinone families” based on their structure, and occasionally their biosynthesis, in order to compare the types of quinones produced by different fungal genera, subgenera and sections.

Paper 2 investigates a high throughput quinone screening method, which utilizes the quinone redox cycle. The assay was originally designed for the detection of ubiquinone in cosmetics products, but we used it to screen for quinones directly in fungal extracts. The purpose of the method was to develop a fast assay for pre-screening a large number of fungal strains, before applying more labor intensive methods such as liquid chromatography hyphenated to mass spectrometry (LC-MS). We found that some quinones showed strong signals after few minutes of reaction, while other showed considerably lower reaction rates. The assay was suited for assessing large differences of phoenicin in some fungal extracts but for other quinones, such as spinulosin and emodin, the assay was not particular successful with the parameters used. Next, a liquid handling robot was used to test seven spinulosin and fumigatin producing strains in an automated fashion, and reaction rates were measured with a spectrophotometer. The same extracts were analyzed by LC-MS, and data between the two detection methods were compared. The two datasets correlated poorly, possibly due to the low reactivity of spinulosin, and potentially fumigatin, in the colorimetric assay. Conclusively, the assay was not suited as a universal screening assay for quinones, as reaction rates were much different for each quinone. However, for some quinones, like phoenicin, the assay seemed to work well.

Paper 3 investigated the production of toluquinone (TQ), terreic acid (TA) and anthraquinones (AQs) by LC-MS analysis. The data was presented as three case studies, highlighting prospects and challenges from working with fungal quinones. TQ is an early shunt product of the patulin biosynthetic pathway, and was not observed in any of the samples, displaying the difficulty in producing shunt products in wild type fungi. Additionally, TQ is poorly ionizing by LC-MS, further increasing the challenge of working with this metabolite. The second example, TA, highlighted the importance of testing different fungal strains and culture conditions in the initial screening process: while both the production strains produced TA on agar medium, only one strain produced the quinone in liquid medium. Thirdly, the AQs were used as examples of the many quinone analogs that some fungi make. The two Talaromyces islandicus production strains are known for their production several AQ monomers and dimers. Feature-based molecular networking (FBMN) was used to identify AQ analogs, and while many quinones were produced, only few were secreted and at a very low level. In all cases, a potential next step for further quinone production was suggested: Production of TQ could potentially be enhanced by discovering or engineering a strain with a truncated patulin biosynthetic pathway. TA production could be increased by optimizing cultivation conditions further. For the AQs, perhaps experimenting with growth conditions could increase the proportion of secreted AQs, or more classical product optimization strategies could be applied to increase production.

Paper 4 presents a thorough investigation of phoenicin, produced by Penicillium atrosanguineum, P. manginii, P. chermesinum and P. phoeniceum. Phoenicin production by P. atrosanguineum on standard liquid media were investigated first. We discovered that the amount and source of carbon was paramount for phoenicin production: With 30 g/L sucrose, no or very little phoenicin was produced while a considerable improvement in production was seen when 90 g/L sucrose was used in the growth medium. We called this mechanism “the phoenicin switch” and discovered that it was conserved across the other species investigated. With a mass spectrometry based metabolomics study, we observed that while several known secondary metabolites with toxic effects were observed (tryptoquialanines and fumiquialanines), most were intracellular, in contrast to phoenicin, which is secreted.

One of the strains investigated (P. phoeniceum) was able to produce 4.94 g/L phoenicin, a very high amount of a natural product in a wild type fungus. For this strain, we tested the effects of the nitrogen sources sodium nitrate and yeast extract in a full factorial design and found that yeast extract had a positive interaction with sucrose, and sodium nitrate had a negative interaction, on the production of phoenicin. A very interesting observation was the production of a tentative phoenicin polymer by P. phoeniceum, which have not previously been described in the scientific literature.

Paper 5 details the culmination of the project, in which phoenicin is tested in an RFB as the anolyte against ferrocyanide. The final battery had an open-circuit voltage of 0.87 V and an initial capacity of 11.75 Ah/L and a capacity fade of 2.85 % per day and 0.35 % per cycle. This paper clearly shows that bio-sourced quinones from filamentous fungi can be used in RFBs.