Solubilization of poplar biomass using C. thermocellum.
A schematic of the microbial fermentation and analysis procedure used to evaluate the solubilization of poplar biomass by C. thermocellum DSM1313 is shown in Additional File 1: Fig. S1. Fermentations (0.5 L) were performed with 5.68 g chipped and milled poplar biomass at 5 g/L total glucan loading corresponding to approximately 11.4 g/L total solids loading. Changes in total mass yield and cellulosic and non-cellulosic sugar composition in the liquid and solid fractions were measured at times zero, 24, 48 and 120 h (5 d). Microbial fermentation in reactors inoculated with C. thermocellum resulted in total solids solubilization of approximately 5.8%, 10.3%, and 13.7% after 24, 48, and 120 h, respectively, while control reactors without the microbe (henceforth called fermentation controls) resulted in 4.5%, 4.6% and 6.3% total solids solubilization at the same time points (Fig. 1 and Additional File 2: Fig. S2). The total amount of biomass solubilized specifically by C. thermocellum CBP was calculated by subtracting the end-point biomass weight recovered after CBP fermentation from the starting biomass, as described in the material and methods, and substracting from this the amount of biomass solubilized by the fermentation controls which did not contain the microbe. When the amount of biomass solubilized due to the heat and medium conditions in the reactors (i.e., as in the fermentation controls) was subtracted, the results showed that C. thermocellum CBP solubilized 1.3%, 5.7% and 7.4% of the poplar biomass after 24, 48, and 120 h, respectively.
Evaluation of dry mass content and glycosyl residue composition of post fermentation liquor from C. thermocellum CBP fermentations
After 24, 48, and 120 h of microbial and control fermentations, the content of each fermentation reactor was centrifuged to separate the liquor (i.e., the supernatant representing the soluble fraction) from the solid residue. The CBP and fermentation control liquor samples were lyophilized for five days and the dry mass was recorded. There was a significant increase (16% to 23%) in the dry mass recovered from the CBP liquor compared to the controls at 24, 48, and 120 h, respectively (Fig. 2). These results showed that C. thermocellum solubilized biomass over the course of the fermentations. Since the amount of dry mass recovered from the liquor of the fermentation controls and the CBP fermentations was greater than the amount of total solids solubilized from the biomass (Fig. 2 and Additional File 2: Fig. S2), these results suggested that the additional mass recovered in the control liquor may have resulted from dried culture medium components and that the additional mass recovered in the C. thermocellum fermentations may have resulted from the culture medium components plus microbe culture components.
In order to determine which components of the poplar biomass were solubilized by C. thermocellum and recovered in the liquor during the CBP fermentations, the liquor samples from the control and C. thermocellum fermentations were analyzed for glycosyl residue composition by trimethylsilyl (TMS) derivatization and GC-MS. This method specifically quantifies the non-cellulosic sugar composition . Noteworthy, a large amount of non-cellulosic sugars were present in the fermentation controls, indicating that some of the non-cellulosic polysaccharides in the poplar biomass were solubilized by the temperature conditions and medium used during the fermentation. Significantly more sugars; however, were solubilized from the poplar biomass by C. thermocellum fermentations compared to the controls including 41% to 267% more xylose (Xyl), 35% to 80% more mannose (Man), and 19% to 35% more glucose (Glc) over the 120 h fermentation (Fig. 3 and Additional File 3: Fig. S3). The accumulation of these sugars in the liquor suggested that C. thermocellum was able to solubilize hemicelluloses, such as glucuronoxylan and glucomannan in poplar wood but did not completely take up or metabolize these sugars. Only trace amounts of acidic sugars (e.g., galacturonic acid, GalA) were detected in the liquid fractions.
Interestingly, arabinose (Ara), rhamnose (Rha), and galactose (Gal) were also detected in the liquor of both the control and C. thermocellulum fermentations. However, their levels were significantly less in the liquor from the CBP fermentations compared to the liquor from the fermentation controls, being 36 to 57% less for Ara, 24 to 49% less for Rha, and 32 to 46% less for Gal at 48 h and 120 h). The reduction in the levels of these sugars in the liquor of the C. thermocellum fermentations compared to the control fermentations, suggested that C. thermocellum may have assimilated these sugars. However, since the levels of these sugars were greater in the control liquor than in the CPB liquor, it was not possible from these data alone to conclude whether or not C. thermocellum had solubilized these sugars from the biomass, or rather they had all resulted from the process conditions. To determine how much sugar C. thermocellum was able to solubilize from the biomass, the solid residues obtained during the fermentations were also evaluated.
Production of alcohol insoluble residues from poplar residual biomass and characterization of lignin, cellulose, and non-cellulosic sugars in the residual biomass
In order to determine the amount of lignin, cellulose and non-cellulosic polysaccharides remaining in fermentation residues during C. thermocellum CBP, alcohol insoluble residue (AIR) was produced from the post-fermentation biomass (solid residues) remaining after CBP and in the fermentation controls. The resulting AIR, which represents cell walls, was de-starched for 48 h with α-amylase and the resulting de-starched AIR was analyzed for cellulose, lignin and non-cellulosic sugars. The yield of total AIR per gram of solid residue from the fermentation controls and the C. thermocellum CBP fermentationsafter the first 24 h were comparable (Additional File 4: Fig. S4). However, the yield of AIR per gram of solid residue from the CBP solid residuals was 3% and 5% less than that from the controls after 48 and 120 h incubation, respectively. This slightly lower yield of AIR per gram solid residue in the C. thermocellum CBPfermentations may have been due to components contributed to the solid residues by C. thermocellum over the course of the fermentation that were not recovered by the ethanol and chloroform/methanol extraction process used for AIR preparation.
To evaluate the amount and composition of lignin in the post-fermentation solid residues of poplar wood, the amount of total lignin, guaiacyl (G), p-hydroxyphenyl (H) and syringyl (S) lignin subunits and the lignin S/G ratio were measured by pyrolysis molecular beam mass spectrometry (py-MBMS) . The total lignin content of the CBI reference poplar was significantly less (14%) than the total lignin content of the BESC standard poplar (Fig. 4A and Additional File 5: Fig. S5A). The CBI reference poplar line (GW-9947) was previously identified as a low recalcitrance poplar natural variant with reduced lignin and a high syringyl-to-guaiacyl (S/G) ratio , a difference confirmed by the py-MBMS lignin analysis results reported here.
Analysis of the lignin content of the CBI reference poplar biomass versus the lignin content in the solid residues of the 24 h, 48 h, and 120 h CBP fermentation controls revealed an unexpected reduction in the amount of total, S, and G lignin, suggesting that some lignin-containing wall material is solubilized during the process used for fermentation. The total, S, and G lignin content of the CBP solid residual samples following fermentation of the CBI reference poplar by C. thermocellum increased on average 25%, 24%, and 13% (per mg starting biomass dry weight), respectively, compared to the fermentation controls (Fig. 4A-D and Additional File 5: Fig. S5), whereas no major changes were observed in the H subunit content. The increase in the percentage of lignin in the residual biomass during CBP fermentation by C. thermocellum was at least partially associated with the solubilization of hemicellulose and cellulose (see below), leading to a greater weight percentage of total lignin in the residues.
The lignin S/G ratios were significantly increased (6 to 14%) in CBP solid residue compared to those of controls (Fig. 4E). A higher S/G ratio in the residual solids following C. thermocellum CBP fermentation of poplar wood has been previously described [14, 16-18]. One hypothesis for this result is that some of the hemicellulose or cellulose solubilized by C. thermocellum is associated with lignin that has a higher G/S ratio, which was preferentially solubilized along with the glycans. There were no major changes in the C5:C6 ratio of poplar solid residues in the control or C. thermocellum CBP fermentations (Fig. 4F).
Two methods were used to determine the cellulose content in the post-fermentation solid residues. The first method involved measuring the amount of glucose present in non-crystalline polysaccharides (i.e., non-cellulosic cell wall glycans) by GC–MS of TMS‑derivatized methyl glycosides and substracting this value from the total cellulosic plus non-cellulosic glucose as obtained by Saeman (1945) hydrolysis of crystalline cellulose [19, 20] followed by glycosyl residue composition analysis (Additional File 6: Fig. S6A-C). This measurement of cellulose (Fig. 5A and Additional File 6: Fig. S6C) was in good agreement with published values of cellulose content for poplar wood [21, 22]. The second method used to measure cellulose content was the recently developed fully methylated alditol (MA) procedure which enables analysis of insoluble polysaccharides through their permethylation in DMSO to produce methyl alditol derivatives that can be analyzed by GC-MS  (Fig. 5B and Additional File 6: Fig. S6D-F). The MA method thus also provides a measure of total Glc which, upon substracting the amount of non-cellulosic Glc (see subsection below), provides the cellulose content. CBP fermentation of CBI reference poplar biomass for 5 d by C. thermocellum resulted in a significant 25% total reduction in the cellulose content (mg Glc/starting biomass) based on the Saeman hydrolysis method, which corresponded to a 21% reduction after subtracting the fermentation controls (Fig. 5A and Additional File 6: Fig. S6A-C). The MA method indicated a 29% total reduction in cellulose (mg Glc/starting biomass) and 25% reduction after subtracting the fermentation controls (Fig. 5B, Additional File 6: Fig. S6 D-F, Table 1). These values are comparable to those of Dumitrache and coworkers  who showed a 28% reduction in total glucose content of P. deltoides biomass after incubationof C. thermocellum ATCC27405 for 92 h at 60°C and confirmed that the microbe can solubilize a portion of the cellulose in non-pretreated poplar woody biomass.
Glycosyl residue composition of non-cellulosic sugars in solid residues
To study the solubilization of non-cellulosic cell wall glycans by C. thermocellum CBP, the glycosyl residue composition of CBI reference poplar solid residues after 24, 48, and 120 h of microbial hydrolysis was measured and compared to those of the solid residues in the fermentation controls (Fig. 6). Glycosyl residue composition analysis of the solid residues before and after C. thermocellulm CBP revealed that the mg per mass starting biomass of six sugars was significantly decreased in the presence of C. thermocellum (Fig.6) with7% to 14% decreased Xyl, 13% to 36% decreased Man, 13% to 23% decreased Glc, and 56% to 71% decreased GalA from 24 h to 120 h. There was also 9% to 13% decreased Gal at 48 h to 120 h and 15% decreased Ara at 120 h indicating hydrolysis by C. thermocellum. Similar relative changes in post-fermentation glucose content (18% decreased Glc/gram solid) has been reported for other high S/G woody phenotypes . The particularly large reduction in the amount of GalA in the solid residue but little GalA in the liquor suggests the possible utilization of pectic sugars by C. thermocellum. Conversely, there was a significant trend for increased Rha (14% to 27%) in the CBP solid residues compared to the fermentation controls, suggesting that Rha containing polymers in the wall (e.g., RG-I) were not being solubilized and thus were found to increase in the final residue recovered. Although there was a small reduction in the total non-cellulosic sugars in the solid residue from the fermentation control, the 10% to 18% greater reduction in total non-cellulosic sugars in the solid residues following CBP fermentation by C. thermocellum indicates that the microbe solubilizes at least some of the non-cellulosic wall polymers including xylans and homogalacturonan (Fig. 6 and Additional File 7: Fig. S7).
A comparison of the amount of lignin, cellulose, and non-cellulosic sugars in the solid residues recovered over the 120 h fermentation (Fig. 7) shows that 24% percent of the cellulose was solubilized over 5 d of CBP fermentation by C. thermocellum (Fig. 7A, C). Simultaneouly, 17% of the non-cellulosic sugars were solubilized suggesting that C. thermocellum hydrolyzes these sugars during fermentation (Fig. 7B). The negative value of mass and percentage of lignin solubilization is due to the increase in the lignin content in the residual biomass as the cellulose and non-cellulosic sugars are solubilized (Fig. 7C and Table 1). An analysis of the percentage of the specific sugars solubilized (Fig. 7D and Table 1) reveals that Rha was the only sugar that was not solubilized by C. thermocellum during the fermentation (-21% compared to starting biomass, Fig. 7D; -27% compared to fermentation control, Table 1). This result suggests that Rha-containing cell wall polymers (i.e. RG-I) may be a key determinant limiting solubilization of poplar biomass during CBP fermentation by C. thermocellum.