Scaling up and scaling down the production of galactaric acid from pectin using Trichoderma reesei

Background: Bioconversion of d-galacturonic acid to galactaric (mucic) acid has previously been carried out in small scale (50-1000 mL) cultures, which produce tens of grams of galactaric acid. To obtain larger amounts of biologically produced galactaric acid, the process needed to be scaled up using a readily available technical substrate. Food grade pectin was selected as a readily available source of d-galacturonic acid for conversion to galactaric acid. Results: We demonstrated that the process using Trichoderma reesei QM6a Δgar1 udh can be scaled up from 1 L to 10 and 250 L, replacing pure d-galacturonic acid with commercially available pectin. T. reesei produced 18 g L -1 galactaric acid from food-grade pectin (yield 1.00 g [g d-galacturonate consumed] -1) when grown at 1 L scale, 21 g L -1 galactaric acid (yield 1.11 g [g d-galacturonate consumed] -1) when grown at 10 L scale and 14 g L -1 galactaric acid (yield 0.77 g [g d-galacturonate consumed] -1) when grown at 250 L scale. Initial production rates were similar to those observed in 500 mL cultures with pure d-galacturonate as substrate. Approximately 2.8 kg galactaric acid was precipitated from the 250 L culture, representing a recovery of 77% of the galactaric acid in the supernatant. In addition to scaling up, we also demonstrated that the process could be scaled down to 4 mL for screening of production strains in 24-well plate format. Production of galactaric acid from pectin was assessed for three strains expressing uronate dehydrogenase under alternative promoters and up to 11 g L -1 galactaric acid were produced in the batch process. Conclusions: The process of producing galactaric acid by bioconversion with T. reesei was demonstrated to be equally efficient using pectin as it was with d-galacturonic acid. The 24-well plate batch process will be useful screening new constructs, but cannot replace process optimisation in bioreactors. Scaling up to 250 L demonstrated good reproducibility with the smaller scale but there was a loss in yield at 250 L which indicated that total biomass extraction and more efficient DSP would both be needed for a large scale process.