Switchgrass (Panicum virgatum) Fermentation by Clostridium thermocellum and Clostridium beijerinckii Sequential Culture: Effect of Feedstock Particle Size on Gas Production

Fermentation of cellulosic biomass can be done in a single step with cellulolytic, solventogenic bacteria, such as Clostridium thermocellum. However, the suite of products is limited in consolidated bioprocessing. Fortunately, the thermophilic nature of C. thermocellum can be exploited in sequential culture. Experiments were conducted to determine the effect of feedstock particle size on fermentation by sequential cultures and to demonstrate this effect could be shown by gas production. Dual-temperature sequential cultures were conducted by first culturing with C. thermocellum (63 ̊C, 48 h) before culturing with C. beijerinckii (35 ̊C, 24 h). Switchgrass (2, 5 or 15 mm particle size) was the feedstock in submerged substrate (10% w/v) fermentation. The extent of fermentation was evaluated by gas production and compared by analysis of variance with Tukey’s test post hoc. C. thermocellum alone produced 78 kPa cumulative pressure (approx. 680 mL gas) when the particle size was 2 or 5 mm. The C. thermocellum cultures with 15 mm feedstock particles had a mean cumulative pressure of 15 kPa after 48 h, which was less than the 2 and 5 mm treatments (P < 0.05). When the culture vessels were cooled (to 35 ̊C) and inoculated with C. beijerinckii, and the cumulative pressures were reset to ambient, cumulative pressure values as great as 70 kPa (equivalent to an additional 670 mL gas) were produced in 24 h. Again, the longer (15 mm) particle size produced less gas (P < 0.05). When the substrates were inoculated with C. beijerinckii without previous fermentation by C. thermocellum, the mean cumulative pressures were Corresponding author.


Introduction
The availability and low cost of cellulosic biomass has led to interest in its use as a raw material for bioenergy [1].Cellulosic biomass can be used to produce liquid fuels via fermentation.However, most microorganisms do not produce the enzymes to catabolize lignocellulose.The biomass must be pre-treated (e.g.exogenous cellulases, sodium hydroxide and high temperature) before saccharification.After saccharification, solventogenic bacteria or fungi can ferment the sugars.
Saccharification and fermentation can be done in a single step with cellulolytic, solventogenic bacteria, such as Clostridium thermocellum [2].C. thermocellum conducts a mixed-acid fermentation, in which CO 2 , H 2 , lactate, acetate, formate and ethanol are the primary products [3] [4].Consolidated bioprocessing with C. thermocellum is not sufficient when other products are desired.However, C. thermocellum is a thermophile, and the enzyme system retains ~25% activity at mesophilic temperatures [5].This characteristic enables sequential culture with a non-cellulolytic mesophile.In this way, the complex, highly active cellulosome of C. thermocellum [6] can be used to liberate sugars for use by mesophilic organisms, such as C. acetobutylicum [7].
The following experiments were initiated to determine if biological pretreatment with C. thermocellum would promote the fermentation of switchgrass (Panicum virgatum) by C. beijerinckii.The hypotheses were: 1) C. thermocellum would liberate sugars and increase gas production by C. beijerinckii, and 2) the rate and extent of gas production would be surface area-dependent.

Feedstock Preparation
The switchgrass was grown on the University of Kentucky Research Farm.It was harvested in November 2011.The switchgrass was cut at 15 cm and stored in small square bales that were later ground to pass through a 2, 5 or 15 mm sieve (hereafter called particle sizes 2, 5 and 15 mm) using a hammer mill (C.S. Bell, CO.Tiffin, OH, USA.Model No. 10HMBD, Serial No. 375 Bratt 03/05).The switchgrass was analyzed by DairyOne (Ithaca, NY, USA) using the wet chemistry package.

Strain and Media Composition
The C. thermocellum ATCC 27405 cell line used in the study came from the culture collection of Herbert J. Strobel, University of Kentucky.C. thermocellum cells were grown anaerobically at 63˚C.The basal medium contained (per liter): 1.53 g Na 2 HPO 4 , 1.5 g KH 2 PO 4 , 90 mg NH 4 Cl, 25 mg (NH 4 ) 2 SO 4 , 90 mg MgCl 2 •6-H 2 O, 30 mg CaCl 2 , 2.0 g yeast extract, 10 ml vitamin mixture [8], 5.0 ml modified mineral mixture (Pfennings Metals plus 10 mg Na 2 WO 4 •2H 2 O and 1 mg Na 2 SeO 3 perliter, as described by Strobel [9]) and 1 ml resazurin.The pH was adjusted to 6.7 with NaOH.The medium was autoclaved (121˚C, 104 kPa, 20 min) and cooled under an O 2 -free CO 2 sparge.The buffer, Na 2 CO 3 (4 mg•ml −1 ), was added before the broth was room temperature.Media for batch cultures were anaerobically dispensed into serum bottles and sealed with butyl rubber stoppers, and autoclaved for sterility.
C. beijerinckii ATCC 51743 was obtained from the American Type Culture Collection (Manassas, VA, USA).C.beijerinckii cells were grown anaerobically at 35˚C in Reinforced Clostridial Media (Difco Laboratories, Detroit, MI, USA).Reinforced clostridial medium (RCM) contained (per liter) 10.0 g peptone, 10.0 g beef extract, 3.0 g yeast extract, 5.0 g dextrose, 5.0 g NaCl, 1.0 g soluble starch, 0.5 g cysteine, 3.0 g C 2 H 3 NaO 2 , and 0.5 g agar.The medium was autoclaved and cooled under O 2 -free N 2 .Media for batch cultures were anaerobically dispensed into serum bottles with butyl rubber stoppers, and autoclaved for sterility.

Effect of the Feedstock Particle Size on Gas Production
Gas production was monitored with the Ankom RF Gas Pressure System (Ankom, Macedon, NY, USA).Ground switchgrass (5 g) was added to the fermentation vessels (Pyrex bottle with sidearm port, 1140 mL actual volume).The vessels were sealed and purged of air with O 2 -free CO 2 through the sidearm septum.Basal medium (50 mL) was added through the septa of the sidearm ports and the vessels were warmed in a water bath (63˚C) prior to inoculation.The vessels were inoculated (10% v/v) through the septa with C. thermocellum (48 h cultures, approximately 10 7 viable cells ml −1 ).After 48 h incubation (63˚C), the vessels were removed from the water bath, permitted to cool, and inoculated (10% v/v) with C. beijerinckii (24 h cultures, approximately 10 8 viable cells ml −1 ).Incubation continued at 35˚C for an additional 48 h.Gas pressure was tested in 1 min intervals, and cumulative pressure was recorded every 5 min.The global pressure release was set at 104 kPa.An uninoculated vessel was included as a control in each iteration of the experiment.The pressure values of the uninoculated vessels were subtracted from each treatment to control for the effects of temperature.
To test gas production from C. beijerinckii without C. thermocellum, control bottles were inoculated with C. beijerinckii (10% v/v).Basal medium was added and the vessel was warmed in a water bath (35˚C, 48h).Gas pressure was tested in 1 min intervals, and cumulative pressure was recorded every 5 min for 48 h.

Soluble Product Quantification
Soluble product quantification was performed on cultures in serum bottles containing 10% switchgrass (2, 5 or 15 mm).Basal medium was added and the serum bottles were warmed in a water bath (63˚C, 48 h).The experiment was initiated by inoculation with C. thermocellum (10% v/v).After 48 h, the temperature was decreased to 35˚, and the bottles were inoculated with C. beijerinckii (10% v/v), and sampled daily.Samples (1 ml) were clarified by centrifugation (14800 × g, 2 min), and frozen for later analyses.Acetate, ethanol, butanol, lactate and formate were quantified by HPLC (Dionex, Sunnyvale, CA).The anion exchange column (Aminex 87H; BioRad) was operated at 50˚C, flow rate 0.4 ml min −1 .Eluting compounds were detected by refractive index (Shodex/Showa).

Statistical Analyses
The experiments were performed in triplicate.The data were analyzed in SAS (version 9.3, SAS Inst.Inc) by MANOVA with Tukey's test post hoc.Treatment variables included feedstock particle size, time of measurement and culture type.P values less than or equal to 0.05 were considered significant.

Results
The composition of the switchgrass was (dry matter basis): 54.9% acid detergent fiber, 88.9% neutral detergent fiber, 8.8% lignin, 10.0% calculated non-fibrous carbohydrate, 3.8% crude protein, 1.2% crude fat, 1.2% ash.Gas production caused the pressure in the fermentation vessels to increase when C. thermocellum was inoculated into basal media with 5 g switchgrass as the substrate (Figure 1).The lag phase was approximately 30 h, after which rapid gas production was observed.When the particle size was 2 or 5 mm, the cumulative pressure after 48 h fermentation was 78 kPa, which is equivalent to the production of approximately 680 mL of gas.The C. thermocellum cultures with 15 mm feedstock particles had a mean cumulative pressure of 15 kPa after 48 h, which was significantly less than the 2 and 5 mm treatments (P < 0.05).
The culture vessels were cooled (to 35˚C) and inoculated with C. beijerinckii, and the cumulative pressures were reset to ambient (Figure 2).The cultures continued to produce gas through a 10 h lag phase.An increase in the rate of gas production was observed between 10 and 15 h in the 2 and 5 mm treatments and then production decreased into stationary phase.Cumulative pressure values as great as 70 kPa (equivalent to an additional 670 mL gas) were observed.Again, the longer (15 mm) particle size produced less gas (P < 0.05).When the substrates were inoculated with C. beijerinckii without previous fermentation by C. thermocellum, the mean cumulative pressures were approximately 10 kPa.In sequential cultures on switchgrass, acids (acetate, formate, lactate) and ethanol were produced by C. thermocellum (Table 1).After inoculation with C. beijerinckii, butyric acid and butanol were also produced.Acetone was not detected.Sugars (xylose, glucose) were detected close to or below the limits of quantification early in the fermentations.There was no significant effect of particle size (P > 0.05).However, the acetate concentration was numerically higher in the treatment with the smallest particle size.

Discussion
Physical pretreatment of lignocellulosic biomass, such as milling, increases the surface area and generally makes the feedstock more amenable to chemical or biological conversion [10].Decreasing particle size has been shown to increase microbial digestion of the feedstock bypure cultures [10] and by natural, poly-microbial fermentations, like the bovine rumen [11].
Co-cultures, either sequential or simultaneous, have been previously used to maximize conversion [12].Yu and coworkers used a sequential culture of C. thermocellum followed by C. acetobutylicum to produce ethanol and butanol from an artificial substrate [7].These results were repeated using corn (Zea mays) residuals that were also cultured with a fungus that produces a lignin peroxidase [13].In the current study, we employed a sequential culture of C. thermocellum followed by C. beijerinckii to ferment switchgrass.Please note that C. beijerinckii and C. acetobutylicum are very similar bacteria that are formerly categorized as the same species [14].
In the current experiment, switchgrass was first fermented with the thermophile, C. thermocellum prior to fermentation with C. beijerinckii.The cellulosome system retained a portion of its fibrolytic activity at mesophilic temperatures, which was originally observed by Ng and co-workers [5] and is consistent with previous results [7] [13].In this way, C. thermocellum could be considered a biological pretreatment for the mesophile C. beijerinckii.This type of sequential culture could be considered a biological saccharification to replace enzymatic treatment of lignocellulose.However, it is important to note that much of the cost of acetone-butanolethanol production comes from downstream processing of the fermentation broth [1].
Gas production is a universal measure of fermentation because carbon dioxide is the most common metabolic product of fermentative organisms [15].Furthermore, gas production has long been used to evaluate the ability of rumen microorganisms to digest cellulosic feeds [16].More gas was produced from smaller rather than from larger switchgrass particles, which was consistent with idea that milling increased surface area for catalysis.The effect of particle size on gas production was consistent during both thermophilic, C. thermocellum fermentation and mesophilic C. beijerinckii fermentation.C. beijerinckii alone produced very little gas pressure from the substrate.This latter observation supports the hypothesis that the C. thermocellum cellulosome system was deconstructing lignocellulose or other non-soluble carbohydrate into substrates that could then be catabolized by C. beijerinckii.

Table 1 .
Product formation by sequential culture of C. thermocellum and C. beijerinckii.t, trace; -, not detected.