1. Project ETB-2012-26 OPTISOLV - Development, optimization and scale-up of biological solvent production 3nd International Meeting PortoMantovano, December 01, 2014 F. Raganati

2.  Biofilm reactors Start Up Glucose/Lactose Lactose Scheduled Activity  The ABE fermentation process by adopting renewable resources. Characterization in terms of kinetics and yields. simple sugars (glucose, fructose and sucrose) typically present in high sugar content beverages; High Sugar Content Beverages Scheduled Activity  Characterization of the ABE fermentation process. Dynamic kinetic modelling Solventogenic Kinetics Scheduled Activity

3. MEDIUMMEDIUM TANKTANK 1 1 2 2 3 3 4 4 V overall = 160 mL Porto Mantovano – December 1, 2014

4. 1 1 2 2 3 3 4 4 Specific lactose production: 4 stages @ D = 0.15 1/h: 0.9 g/Lh 1 stage @ D = 1 1/h: 4.5 g/Lh 1 stage Porto Mantovano – December 1, 2014 4 stages

5. 1 1 2 2 3 3 4 4 1 stage 4 stages

6. 1 1 2 2 3 3 4 4 Porto Mantovano – December 1, 2014 1 stage 4 stages

7. 1 1 2 2 3 3 4 4 Porto Mantovano – December 1, 2014  Series of 4 packed bed reactors Continuous additional in line feeding between bioreactors  addition of glucose to the 2nd or 3rd bioreactor at D= 0.1 h -1.  addition of AA, AB, glucose or combination of those to the 2nd bioreactor at D=0.15-0.2 h -1 in concentrations that simulated conditions in the flow from the 1st to 2nd bioreactor at high dilution rates.  addition of butanol at concentrations of around 0.5 g/to the feed to the 1st bioreactor at D= 0.1-0.15 h -1.  operation of the system at different glucose concentration in the feeding stream.

8.  Biofilm reactors Start Up Glucose/Lactose Lactose Scheduled Activity  The ABE fermentation process by adopting renewable resources. Characterization in terms of kinetics and yields. simple sugars (glucose, fructose and sucrose) typically present in high sugar content beverages; High Sugar Content Beverages Scheduled Activity  Characterization of the ABE fermentation process. Dynamic kinetic modelling Solventogenic Kinetics Scheduled Activity

9. Porto Mantovano – December 1, 2014

10. Unsupplemented Medium (HSCB) Supplemented Medium (HSCB+) Hydrolized Supplemented Medium (HHSCB+) Glucose conversion: complete Fructose conversion: almost complete Glucose conversion: complete Fructose conversion: almost complete about 10 g/L of butanol were produced a significant amount of sucrose was unconverted tests with hydrolysed HSCB+ (HHSCB+) were performed improved solvents production and sugar conversion degree 13 g/L

11. HSCB as complex substrate to feed the 4 PBR in series. 1 1 2 2 3 3 4 4 Porto Mantovano – December 1, 2014

12.  Biofilm reactors Start Up Glucose/Lactose Lactose Scheduled Activity  The ABE fermentation process by adopting renewable resources. Characterization in terms of kinetics and yields. Simple sugars (glucose, fructose and sucrose) typically present in high sugar content beverages; High Sugar Content Beverages Scheduled Activity  Characterization of the ABE fermentation process. Dynamic kinetic modelling Solventogenic Kinetics Scheduled Activity

13. A kinetic dynamic model of acetone–butanol–ethanol (ABE) production by Clostridium acetobutylicum DSM 792 was developped according to the biochemical networks simulator COPASI Substrate effects investigation: glucose, mannose, fructose sucrose, lactose xylose, and arabinose Substrate effects investigation: glucose, mannose, fructose sucrose, lactose xylose, and arabinose The Embden-Meyerhof-Parnas (EMP) pathway for hexoses and disaccharides The pentose phosphate (PP) pathway for pentoses. The Embden-Meyerhof-Parnas (EMP) pathway for hexoses and disaccharides The pentose phosphate (PP) pathway for pentoses. Porto Mantovano – December 1, 2014

14. The proposed model was an update of the model by Shinto et al. (2007, 2008)* Kinetics Shinto et al. Proposed Model Substrate Inhibition + Non Competitive Butanol Inhibition Substrate Inhibition + Complete Butanol Inhibition BA Activation + Non Competitive Butanol Inhibition BA Activation + Complete Butanol Inhibition Non Competitive Butanol Inhibition Multi Product Inhibition Mass Action Specific Butanol Activation Porto Mantovano – December 1, 2014

15. The r 2 increased with respect to that calculated for Shinto’s simulation, whatever the tested sugar The results confirmed that the structure of the present model improved the simulation results. The soundeness of the model has been tested according to two procedures: the assessment of the average squared correlation coefficients (r 2 ) between the simulation results and the experimental data. the comparison of the results of the presenet model with those reported by Shinto et al. (2007-2008). r2r2 r2r2 Porto Mantovano – December 1, 2014

16. R = 0.14, 0.54 and 0.88 D = 0.02 - 0.15 h -1. The system was described by the equation set: BIOMASS BALANCE CELL TRANSFORMATION PATHS REACTION SET Porto Mantovano – December 1, 2014

17. R=0.88 The concentration of acidogenic cells increases linearly with D while the spore concentration decreases exponentially with D The concentration of solventogenic cells is almost constant with D except for a little increase at low D As D increases, the progressively shift toward a less harsh conditions – low concentration of solvents and acids – promotes the presence of acidogenic cells at spore’s expense Porto Mantovano – December 1, 2014

18. The agreement between the model prediction and experimental data is satisfactory. The production rate of butanol referred to the mass unit of solventogenic cells was calculated for all tests. Lactose, acetic acid and butyric acid were considered as substrate and butanol as the inhibition product (Monod-Boulton model) Porto Mantovano – December 1, 2014

19. Model of a Biofilm PBR Model based on: Porto Mantovano – December 1, 2014 Dynamic Model Acidogenic Kintics Solventogenic Kintics

20. List of contributions  Raganati, F., Olivieri, G., Procentese, A., Russo, M. E., Salatino, P., Marzocchella, A. (2013). Butanol production by bioconversion of cheese whey in a continuous packed bed reactor. Bioresource Technology, 138, 259–265  Raganati, F., Procentese, A., Olivieri, G., Russo, M. E., Marzocchella, A.. MFA of Clostridium acetobutylicum pathway: the role of glucose and xylose on the acid formation/uptake. Chemical Engineering Transactions. 2014 V. 38, p.193-198  A. Procentese, T. Guida, F. Raganati, G. Olivieri, P. Salatino, A. Marzocchella. Process Simulation of Biobutanol Production from Lignocellulosic Feedstocks. Chemical Engineering Transactions. 2014 V. 38, p.343-438  A. Procentese, F. Raganati, G. Olivieri, M.E. Russo, P. Salatino, A. Marzocchella. Butanol production by fermentation of Clostridium acetobutylicum: solventogenic kinetics. Submitted to Bioresource Technology  F. Raganati, A. Procentese, G. Olivieri, P. Gotz, P. Salatino, A. Marzocchella. Kinetic study of butanol production from various sugars by Clostridium acetobutylicum using dynamic model. Submitted to Biochemical Engineering Journal  F. Raganati, A. Procentese, F. Montagnaro, G. Olivieri, A. Marzocchella. Butanol Production from Leftover Beverages and Sport Drinks. BioEnergy Research. 2014 - DOI 10.1007/s12155-014-9531-8 Porto Mantovano – December 1, 2014

21. Teresa Guida Antonio Marzocchella Giuseppe Olivieri Alessandra Procentese Francesca Raganati Maria Elena Russo (IRC – CNR) Piero Salatino

22. Porto Mantovano – December 1, 2014 Italy: Jan 1, 2013 Germany: May 20, 2013 K.O. meeting: May 27, 2013 December 1, 2014

23. Porto Mantovano – December 1, 2014

24. R=0.88 The lactose conversion and the concentration of products (cells and metabolites) decrease with the D The butanol selectivity increased with D and it approached a constant value of about 0.90 g/g Butanol and ABE productivities increased with D. A double slope may be observed in the productivity vs. D data with a discontinuity at D≈0.1 1/h the slope at lower D is higher than that at higher D Butanol and ABE productivities increased with D. A double slope may be observed in the productivity vs. D data with a discontinuity at D≈0.1 1/h the slope at lower D is higher than that at higher D Porto Mantovano – December 1, 2014

25. The µ was typically smaller than D and larger than D OUT the accumulation of acidogenic cells - µ>D OUT - was prevented by the establishment of a cell population controlled by the equilibrium among acidogenic cells, solventogenic cells and spores. The analysis of µ S and of concentration of acids and solvents suggests that acids promote the solventogenic cell formation while solvents inhibit the formation.

26. The study carried out during the present Ph.D. program aimed at investigating the Acetone- Butanol-Ethanol (ABE) production process by fermentation of renewable feedstocks The activities were articulated along three paths The characterization of the ABE fermentation process as regards kinetics and yields using different renewable resources sugars representative of hydrolized lignocelluloe (glucose, mannose, arabinose and xylose) sugars representative of high sugar content beverages (glucose, fructose and sucrose) Characterization of the ABE fermentation process through MFA and dynamic kinetic models the MFA was adopted to investigate the role of the main reaction steps of the C. acetobutylicum metabolic pathway to convert reference sugars A kinetic dynamic model of acetone–butanol–ethanol (ABE) production by Clostridium acetobutylicum DSM 792 was proposed using the biochemical networks simulator COPASI. Development of innovative continuous reactor for the ABE production High sugar content beverages & Cheese Whey Biofilm Packed Bed Reactor

27. Assessment of the model parameters The maximum reaction rate of a reaction step depends on the sugar because it depends on the enzyme concentration The “affinity” constants do not depend on the sugar because they depend on the enzyme responsible of the reaction step but not on its concentration Parameters of the sugar uptake reactions have been assessed for each sugar according to Servinsky et al. (2010): C. acetobutylicum has sugar-specific mechanisms for the transport and metabolism genes. *Servinsky et al., (2010). Microbiol. 156:3478–3491 The soundeness of the model has been tested according to two procedures: the assessment of the average squared correlation coefficients (r 2 ) between the simulation results and the experimental data. the comparison of the results of the presenet model with those reported by Shinto et al. (2007-2008).