UCLA

Biofuel and chemical production through microbial catalysts has been heralded as a route for a renewable future; however, several issues must be resolved before microbes become the workhorse of our energy and chemical industries. Chief amongst these problems is the non-renewable use of reduced nitrogen fertilizers and susceptibility to contamination during fermentation. Previously, the use of engineered Escherichia coli has been demonstrated as a means to recycle reduced nitrogen from protein-rich wastes such as algal biomass and spent-fermentation yeast – an a new bioprocess that brought the global nitrogen cycle towards closure by recycling fertilizers. This was accomplished by introducing novel transamination-deamination metabolic cycles, disrupting free ammonia assimilation, and using biofuels as a carbon sink. However, on a large scale, this process required 2 individual pretreatment steps of the protein feedstock: biomass dissolution in alkaline conditions and polypeptide hydrolysis with in vitro protease addition. These pretreatment unit operations complicate industrial deployment of this nitrogen-neutral process. Additionally, it’s use of E. coli as the host microbe does not combat contamination risks.

We set out to consolidate this process by designing new protein-conversion bioprocesses in two other bacterial strains: Bacillus subtilis and Bacillus marmarensis. Bacillus subtilis is a model gram-positive bacterium known for secreting a potent mixture of extracellular proteases. With genetic tools already established for B. subtilis, several strains were quickly developed to test how we can drive the cell’s metabolism to consume protein and produce biofuels. Deletion of the branched-chain amino-acid and GTP-sensing regulator codY and branched-chain alpha ketoacid dihydrolipoamide transferase bkdB generated a B. subtilis strain which grew robustly on protein and still reclaimed high quantities of ammonia. Furthermore, deletion of codY upregulated the valine/isoleucine production pathway, that when combined with plasmid-based overexpressed of alcohol dehydrogenase yqhD of E. coli, leucine dehydrogenase of Thermoactinomyces intermedius, and alpha-ketoacid decarboxylase kivD of Lactococcus lactis led to biofuel and NH3 yields of 19 and 47% from amino acids, respectively. Further deletion of stringent starvation response regulator relA in this production strain increased biofuel and NH3 yields to 32 and 60 % of the theoretical maximum. This strain presented the first consolidated conversion of protein by direct fermentation of polypeptides to 2.0 and 4.8 g/l of biofuels and NH3. However, the system remains highly-susceptible to contamination and retains 1 of 2 pre-treatment steps.

To further consolidate the process, we explored the unstudied extreme alkaliphile Bacillus marmarensis DSM 21297. To begin, we sequenced and annotated the genome of this strain. We then developed the first entirely-alkaline electrotransformation protocol and identified useable genetic tools to metabolically engineer this strain. B. marmarensis was found to grow optimally between pH 9.0 and 10.5, with residual growth observed up to pH 12.5. Through 16s rDNA library studies, we found it to be highly resistant to intentionally-introduced environmental contaminates. An NH3-recycling process was then designed with this strain. In this process, B. marmarensis was culture on protein in alkaline conditions with air-stripping. This removed NH3 from the media as it was secreted from B. marmarensis cells, and it was then collected in a neutral-buffered solution. This setup recovered 27% of the reduced nitrogen fertilizers of proteinaceous material, all while maintaining contamination resistance. Engineering B. marmarensis in this process added about 1 g/l ethanol production on top of NH3 recycling, and opened the door to a contamination-resistant, no pre-treatment protein conversion process.

Additionally, B. marmarensis was found to grow on the complex carbohydrates/cellulose-degradation production cellobiose and xylose. A process was designed in which B. marmarensis expressing pyruvate decarboxylase pdc and alcohol dehyodrgenase adhB of Zymomonas mobilis produced ethanol at 65% the maximum yield and titers of 38 g/l. Similar yields were obtained from cellobiose and xylose mixtures in contaminated media using seawater. This presents a novel approach to the contamination-resistance conversion of cellulosic material, on top of multivariate consolidation of the protein conversion process.