Poison to Products: On harnessing the power of microorganisms to convert waste streams into new chemicals

One of the main challenges society currently deals with is the depletion of fossil fuels. To navigate this issue, we must embrace the concept of circularity and turn waste into a resource. Waste streams are omnifarious and their conversion into new chemical building blocks is not always trivial. Luckily, we can take a look at nature’s problem solving skills to help us out. Because nature, in due time, always finds a solution and there is a (micro)organism for everything.

But.. we can also give nature a hand by simplifying the problem. The diversity and complexity of waste streams can be reduced by using gasification, where the waste is combusted at a high temperature with small amounts of oxygen. This yields syngas, a mixture consisting of mainly carbon monoxide, carbon dioxide and hydrogen gas. Syngas can be converted chemically into i.e. ethanol, but the success of this process highly depends on the ratios of CO, CO₂ and H₂ and the absence of impurities in the gas. Microorganisms can deal with much more variability, making them a promising biocatalyst for the conversion of syngas to chemical building blocks. Yet, we have to understand the microorganisms to be able to work together with them in combatting climate change. The work in this thesis is aimed at increasing our understanding of two specific types of microorganisms that can help us to turn waste into new chemicals: syngas fermenting bacteria and chain elongating bacteria. Together, they can form a team that turns a C1 molecule (carbon monoxide) all the way into a C6 molecule (hexanoate). To make the team as effective as possible, we studied both team members in detail. The syngas fermenting bacterium we studied goes by the name Clostridium autoethanogenum, and is already being used at industrial scale by the company LanzaTech. For its chain-elongating counterpart, however, we used a mixed community of microorganisms that was specifically selected to perform chain elongation. We used this mixed community because the single, optimal partner for C. autoethanogenum has yet to be found.

It has been established previously, by other researchers, that producing a lot of hexanoate is easiest when you feed chain elongating organisms a substrate with a high ethanol-to-acetate ratio. C. autoethanogenum naturally produces ethanol and acetate, but usually in a low ethanol-to-acetate ratio. In Chapter 2 we use a theoretical framework based on thermodynamics, as well as data from literature to understand what triggers C. autoethanogenum to make ethanol. We found that acetate conversion into ethanol is a stress response used to deal with a (too) high load of CO, which can be classified as overflow metabolism. We show that this behavior not only takes place when feeding CO alone, but also in the presence of both CO and H₂, underlining its relevance in syngas fermentation processes. The stress response can be induced by tuning the operational parameters of the bioreactor, such as the CO supply rate or the growth rate.

In Chapter 3 we quantify this effect in the laboratory ourselves. We use a steady-state culture of C. autoethanogenum in a chemostat bioreactor and repeatedly disturb it for periods of one hour with increasing amounts of CO in the inlet gas, up to a CO partial pressure of 1.2 atm. We see that ethanol production increases with increasing CO partial pressures, and at a pCO of 0.6 atm or higher external acetate is even consumed to sustain higher ethanol production rates. This proves that the product spectrum of syngas fermentation can be directed by changing the operational conditions. Furthemore, the experimental method that we used allowed for the identification of the CO uptake rate at each CO partial pressure, directly via the off-gas measurements. We observed biomass-specific CO uptake rates of up to –119 ± 1 mmol·g_x⁻¹·h⁻¹, which is much higher than has previously been reported for this organism. The biomass-specific uptake rate is instrumental for obtaining an accurate mathematical description (or: kinetic model) of this microorganism, which in turn allows for more accurate bioprocess design.

Chapter 4 focusses on the chain-elongating counterpart of our syngas fermenter. C. autoethanogenum prefers to grow at a pH of 5 –5.5, and most chain elongators that have been described in literature rather grow at neutral pH (± 7.0). This chapter revolves around this discrepancy. By using enrichment cultures in a sequencing batch bioreactor, we select for chain elongating microorganisms both at pH 7.0 and pH 5.5. In doing so, we establish that chain elongators can live at pH 5.5 and that a very comparable microbial community (on genus-level) develops at both pH. However, the behavior in the bioreactors was not the same. At lower pH, a significantly smaller fraction of the supplied ethanol was converted to hexanoate. Instead, more of the C₄ molecule butyrate was produced, likely because it is less toxic to the microorganisms than hexanoate. This means that pH is an important parameter to control the product spectrum of chain elongation and that establishing an effective microbial team for C₁-to-C₆ conversion likely requires more than finding microbes with the same preferred pH.

In Chapter 5 we delve into the biochemistry of chain elongating microbes. They are known to be very flexible in their metabolism, and they can deal with a wide range of ethanol-to-acetate ratios. Theoretically, this ratio could even be infinite (i.e. feeding only ethanol), which would lead to the production of only hexanoate and no butyrate. We call this ethanol-only chain elongation. This is interesting from a fundamental as well as a process design perspective. Therefore, we test whether it is also possible in practice by using well-monitored batch experiments in bioreactors. We use different initial conditions: only ethanol, ethanol and a small amount of acetate and ethanol and a small amount of butyrate. We observe in the bioreactors that ethanol-only chain elongation is possible, but that it proceeds very slowly. Beside that, the microorganisms prefer the presence of either acetate or butyrate so much that they eventually start producing these compounds from ethanol themselves when they are not available. This behavior has never been observed before, nor was it regarded as possible.

In Chapter 6 we present a dataset of well-controlled bioreactor experiments in 9 different initial conditions, including the experiments described in the previous chapter. This dataset can be used to refine the current mathematical description of chain-elongating microbes. We describe the initial analysis of this dataset and how we assure its quality and usability for kinetic modelling using data reconciliation. With this reconciled dataset we test the accuracy of the currently available kinetic model. From this overall analysis we set out the next steps for the formulation of a more accurate kinetic model of chain elongating microbes in the future.

Chapter 7 recapitulates the significant findings from this thesis, but more importantly provides a list of questions that still remain to be answered. These questions are grouped around three different themes to provide some structure: the inner world of microorganisms, the interactions between (communities of different) microorganisms and the design of efficient (new) bioprocesses for a more sustainable world. To conclude, I reflect on the societal role of a scientist.