Due to oil shortage and environmental problems, synthetic plastics have to be replaced by different biodegradable materials. A promising alternative could be polyhydroxyalkanoates (PHAs), and the low-cost abundant agricultural starchy by-products could be usefully converted into PHAs by properly selected and/or developed microbes. Among the widely available starchy waste streams, a variety of residues have been explored as substrates, such as broken, discolored, unripe rice and white or purple sweet potato waste. Cupriavidus necator DSM 545, a well-known producer of PHAs, was adopted in a simultaneous saccharification and fermentation (SSF) process through an optimized dosage of the commercial amylases cocktail STARGEN™ 002. Broken rice was found to be the most promising carbon source with PHAs levels of up to 5.18 g/L. This research demonstrates that rice and sweet potato waste are low-cost feedstocks for PHAs production, paving the way for the processing of other starchy materials into bioplastics.

Fermentation is a widely used process in the biotechnology industry, in which sugar-based substrates are transformed into a new product through chemical reactions carried out by microorganisms. Fermentation yields depend heavily on critical process parameter (CPP) values which need to be finely tuned throughout the process; this is usually performed by a biotech production expert relying on empirical rules and personal experience. Although developing a mathematical model to analytically describe how yields depend on CPP values is too challenging because the process involves living organisms, we demonstrate the benefits that can be reaped by using a black-box machine learning (ML) approach based on recurrent neural networks (RNN) and long short-term memory (LSTM) neural networks to predict real time OD_600nm values from fermentation CPP time series. We tested both networks on an E. coli fermentation process (upstream) optimized to obtain inclusion bodies whose purification (downstream) in a later stage will yield a targeted neurotrophin recombinant protein. We achieved root mean squared error (RMSE) and relative error on final yield (REFY) performances which demonstrate that RNN and LSTM are indeed promising approaches for real-time, in-line process yield estimation, paving the way for machine learning-based fermentation process control algorithms.

Thus far, no manufacturing process able to support industrialization has been reported for the recombinant human brain-derived neurotrophic factor (rhBDNF). Here, we described the setup of a new protocol for its production in Escherichia coli (E. coli) and its purification to homogeneity. A synthetic gene, codifying for the neurotrophin precursor, was inserted into an E. coli expression vector and transformed into BL21 (DE3) strain. The recombinant protein accumulates, at high yields, into inclusion bodies. With the developed strategy, more than 50% of the precursor can be refolded. The protein is successively digested by trypsin and the rhBDNF mature form is finally purified by two additional chromatographic steps If the wild-type precursor can be efficiently obtained by the proposed methodology, its pro-peptide remotion, through enzymatic digestion, is however problematic. To circumvent this difficulty, the precursor hinge region, containing the natural furin recognition site, was engineered to be more specifically cleaved by trypsin. Notwithstanding the substitution of three residues in the pro-region carboxyterminal, the precursor correctly refolds and is efficiently cleaved to generate a biologically active mature rhBDNF. This efficient high-yield process fills the current need of a scalable protocol to produce GMP-grade material and unlocks the rhBDNF employment in future clinical investigations.