Comparing deconstruction of sugarcane bagasse by a coculture of thermophilic bacteria and a fungal cellulase system

Abstract

Background

Lignocellulosic biomass is a promising feedstock for sustainable bioethanol production due to its abundance, renewability, and high carbohydrate content. Among these feedstocks, sugarcane bagasse (SCB), a fibrous byproduct of sugar and ethanol mills, is particularly abundant in Brazil. However, the complex and recalcitrant structure of lignocellulose hinders its biological conversion into biofuels. There are different enzymatic architectures to overcome this challenge. Fungal enzyme systems rely on non-complexed cellulase components that act independently and are often used in combination with yeast fermentation in a simultaneous saccharification and fermentation (SSF) configuration. By contrast, thermophilic bacteria such as Clostridium thermocellum employ complexed cellulosome systems capable of highly efficient biomass deconstruction, which enables their application in Consolidated Bioprocessing (CBP). Although previous studies have compared these systems on various lignocellulosic feedstocks, no direct comparison exists for unpretreated sugarcane bagasse. This study aimed to compare carbohydrate solubilization between CBP with a coculture of thermophilic bacteria and SSF with a fungal enzyme–yeast system for the conversion of unpretreated SCB, and to assess the effect of particle size.

Results

CBP consistently achieved higher carbohydrate solubilization than SSF across all particle sizes, with improvements ranging from 1.9- to 3.7-fold. CBP was also less affected by particle size variation, showing a drop of 0.52-fold in solubilization for particles sized 0.2–4.00 mm, compared to a 2.2-fold for SSF. A strong linear correlation between C₆ and C₅ sugar solubilization was observed for both strategies. Despite differences in solubilization, residual sugar concentrations were similar between systems. Neither increased enzyme loading nor improved conditions significantly improved SSF performance, indicating intrinsic limitations in the enzymatic hydrolysis of unpretreated SCB.

Conclusions

This study provides the first direct comparison of unpretreated SCB deconstruction mediated by a thermophilic bacterial culture to a fungal cellulase preparation in the presence of yeast. Higher carbohydrate solubilization is seen for the bacterial system regardless of particle size. These findings highlight the potential of CBP with thermophilic bacteria as an alternative to conventional strategies. The results also reinforce the importance of standardized metrics, such as fractional carbohydrate solubilization, for cross-study comparison.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13068-026-02787-x.

Background

Lignocellulosic biomass is widely regarded as a highly relevant renewable feedstock for biofuel production due to its cost-effectiveness, large-scale availability, and potential to enhance environmental sustainability [1]. Among these feedstocks, sugarcane bagasse (SCB), a fibrous byproduct of sugarcane processing for juice extraction in sugar and ethanol mills, is produced in large quantities in many tropical and subtropical countries [2]. In Brazil alone, approximately 165 million tons are generated annually [3, 4]. Although commonly used for electricity cogeneration, increasing process efficiency has resulted in surplus availability at processing sites and SCB is regarded as a strategic feedstock for second-generation ethanol production [5].

Despite its potential, conversion of lignocellulosic biomass into biofuels such as second-generation ethanol remains technically challenging due to biomass recalcitrance, the intrinsic resistance of plant cell walls to biological deconstruction [6]. Overcoming recalcitrance has led to development of multi-step processing strategies involving pretreatment to increase enzyme accessibility, enzymatic hydrolysis to release fermentable sugars, and fermentation to convert these sugars into ethanol [7]. Although substantial progress has been achieved in improving these individual steps and integrating them into viable process configurations, the overall cost and complexity of biofuel production from cellulosic feedstocks via the conventional paradigm remains greater than that of production from sugarcane juice or starch [8, 9]. In light of the dominant impact of biomass recalcitrance on the economic viability of cellulosic biofuel production, it is of considerable interest to compare the relative ability of different biocatalysts to mediate lignocellulose deconstruction.

Biological systems have evolved distinct enzymatic architectures for plant cell wall deconstruction. These architectures can be broadly categorized into non-complexed systems, comprising mono- and multi-functional enzymes, and complexed systems, in which enzymes are organized into highly aggregated cellulosomal structures [10, 11]. In non-complexed systems, individual cellulases, hemicellulases, and accessory enzymes act independently to hydrolyze polysaccharides, and fungal cellulases, particularly those secreted by Trichoderma reesei, represent the most studied examples [12]. Complexed systems are characterized by the assembly of multiple catalytic subunits into a large extracellular structure known as the cellulosome, which enables synergistic and highly efficient hydrolysis of cellulose [6]. The thermophilic anaerobe Clostridium thermocellum exemplifies this architecture, exhibiting exceptional cellulolytic capability primarily due to its elaborate cellulosome complex [12].

Industrial application of non-complexed cellulase systems for second-generation biofuel production, as described in the literature and implemented in practice, involves enzyme production in an aerobic unit operation, followed by their adddition to pretreated lignocellulosic biomass to release soluble sugars that are fermented to ethanol under anaerobic conditions by a non-cellulolytic micro-organism, such as Saccharomyces cerevisiae [1]. This configuration, which integrates enzymatic hydrolysis and fermentation, is commonly implemented as Simultaneous Saccharification and Fermentation (SSF). Depending on process design, hydrolysis and fermentation can also be carried out sequentially (Separate Hydrolysis and Fermentation, SHF) and include the fermentation of pentoses and hexoses simultaneously (Simultaneous Saccharification and Cofermentation, SSCF) [7]. Biofuel production using complexed cellulase systems, though at a considerably lower technology readiness level, is envisioned to occur through Consolidated Bioprocessing (CBP), where enzyme production, biomass deconstruction, and fermentation of released sugars take place in a single anaerobic unit operation [1]. Monocultures and cocultures involving Clostridium thermocellum have received the most attention for CBP, given the highly efficient cellulosome complex of this micro-organism [13].

Prior studies have directly compared deconstruction of various cellulosic feedstocks in the absence of pretreatment other than autoclaving via cultures including C. thermocellum and via fungal cellulase preparations in the presence or absence of yeast [14–21]. All such studies known to us find that thermophilic bacteria achieve higher carbohydrate solubilization. Paye et al. [14] and Holwerda et al. [16] found that C. thermocellum cultures achieved significantly greater carbohydrate solubilization than SSF with commercial fungal enzymes and yeast for switchgrass and Populus. Similarly, Shao et al. [15] reported that C. thermocellum outperformed SSCF with fungal enzymes for unpretreated winter rye, while Yee et al. [17] observed higher total fermentation product yields for C. thermocellum, Caldicellulosiruptor bescii, and Caldicellulosiruptor obsidiansis compared to SSF with fungal cellulase for switchgrass. In agreement with these findings, Kothari et al. [20, 21] demonstrated that C. thermocellum was markedly more effective than commercial fungal enzymes at solubilizing unpretreated switchgrass, achieving nearly fivefold greater glucan conversion. A direct comparison of SCB deconstruction by C. thermocellum and fungal cellulases has not been reported in the literature.

In addition to biocatalyst type, process parameters such as particle size can also influence the efficiency of lignocellulose solubilization by affecting substrate accessibility and surface area [22]. While the influence of particle size has been studied for enzymatic hydrolysis systems [22], it has been less systematically evaluated for consolidated bioprocessing (CBP), particularly using unpretreated sugarcane bagasse across a broad particle size range. Jin et al. [19] reported that CBP with Clostridium phytofermentans on pretreated corn stover achieved similar polysaccharide hydrolysis to SSCF with fungal enzymes, with smaller particles enhancing sugar release. Likewise, Paye et al. [14] observed a slight decrease in solubilization with increasing particle size for both CBP and SSF using switchgrass. Furthermore, under high-solids conditions, particle size affects the rheological properties of biomass slurries, impacting mass and heat transfer as well as mixing and pumping energy requirements [22, 23].

Here we evaluate deconstruction of sugarcane bagasse with no pretreatment other than autoclaving by a coculture of Clostridium thermocellum and Thermoanaerobacterium thermosaccharolyticum and we compare results to an SSF system using a commercial cellulase preparation (Cellic^® Ctec2) and yeast. Particle size was evaluated across a 50-fold range (0.08–4 mm), and enzyme loading and incubation temperature were varied for SSF. In this study, particle size is used as a controlled variable to modulate substrate accessibility, enabling a comparison between biocatalysts under different physical constraints. Low substrate loadings are employed in order to evaluate the intrinsic deconstruction capability of the systems evaluated in the absence of complicating factors encountered under industrial conditions.

Methods

Feedstock preparation and characterization

Sugarcane bagasse (SCB) was kindly provided by Raízen S.A. (Piracicaba, SP, Brazil) in October 2023 (late in the harvest season), from their local first-generation ethanol mill. Upon receipt, the SCB was air-dried for at least 7 days and stored dry and in the dark. Subsequently, it was comminuted to obtain six particle sizes: 0.08, 0.2, 0.5, 1.0, 2.0, and 4 mm (Fig. 1). For this purpose, an ultra-centrifugal mill (Retsch ZM 300, Haan, Germany) equipped with stainless steel sieves of specific cutoff sizes (0.08 mm, 0.2 mm, 0.5 mm, 1.0 mm, 2.0 mm, and 4.0 mm) was used, following the procedures shown in Fig. 1.

Fig. 1.

Sugarcane bagasse comminution. The received bagasse was milled using a centrifugal mill and sieved through specific cutoff sieves. Material retained on the sieve was remilled until it passed through. The process was performed to obtain six distinct particle size sizes: 0.08, 0.2, 0.5, 1.0, 2.0, and 4.0 mm (Centrifugal mill illustration generated with AI/Sieve illustration sourced from BioRender.com)

The total carbohydrate content of the milled bagasse was assessed according to Zambello et al. [5], adapted from Sluiter et al. [24] (Table 1). Particle size distribution was determined by laser diffraction [25] using a MASZ3000/AeroS particle analyzer (Malvern Instruments, Worcestershire, UK). The measurements were conducted at the Biomass Characterization, Analytical Resources, and Calibration Laboratory (LRAC, FEQ/Unicamp), using a dry dispersion unit at room temperature with compressed air as the dispersing medium. The dispersion pressure was set to 2–2.5 bar and feeding rate was adjusted to 80–90%. Samples of 2.0- and 4.0-mm bagasse were not analyzed due to upper limitations of the equipment. Data were interpreted using the Mie theory [26], assuming a refractive index of 1.45 for the samples. Results are expressed as a particle size distribution curve (frequency vs. particle diameter), volume-based percentile (D₁₀, D₅₀, and D90) and Span value:

$$Span=\frac{D_{90}-D_{10}}{D_{50}}$$ 1

Table 1.

Total carbohydrate content of sugarcane bagasse fractions at different particle sizes

Particle size (mm)	Carbohydrate content (g/g solid)
	Glucose	Xylose	Arabinose	Total
0.08	0.432 ± 0.006	0.258 ± 0.006	0.022 ± 0.002	0.712 ± 0.014
0.20	0.433 ± 0.005	0.241 ± 0.003	0.028 ± 0.000	0.702 ± 0.005
0.50	0.433 ± 0.002	0.258 ± 0.002	0.021 ± 0.002	0.712 ± 0.004
1.00	0.432 ± 0.001	0.260 ± 0.002	0.023 ± 0.003	0.715 ± 0.006
2.00	0.436 ± 0.010	0.260 ± 0.006	0.021 ± 0.002	0.717 ± 0.015
4.00	0.433 ± 0.009	0.262 ± 0.002	0.023 ± 0.004	0.718 ± 0.008

Values are expressed as grams of monomeric sugars (glucose, xylose, and arabinose) per gram of dry solids (mean ± standard deviation; n = 3)

Specific surface area was determined by nitrogen adsorption [25] at 100 °C using the BET (Brunauer–Emmett–Teller) method. These measurements were also performed at LRAC, using a Micromeritics ASAP2020 Plus (Norcross, USA) instrument. Prior to analysis, the samples were degassed under vacuum at 100 °C for at least 12 h to remove adsorbed moisture and gases. The BET specific surface area was calculated from the linear region of the adsorption isotherm with six P/P₀ points. The results are expressed in m²/g.

Micro-organisms, fungal enzymes, growth media, and inoculum preparation

Consolidated bioprocessing (CBP) fermentations were carried out using a coculture of Clostridium thermocellum DSM 1313 and Thermoanaerobacterium thermosaccharolyticum HG-8 ATCC 31960 [27]. C. thermocellum was obtained from the Deutsche Sammlung von Mikro-organismen und Zellkulturen GmbH (DSMZ, Leibniz, Germany) and T. thermosaccharolyticum was obtained from the American Type Culture Collection (ATCC, Manasas, VA). For these fermentations, defined medium for thermophilic clostridia (MTC) was used, adapted from Holwerda et al. [28]. Concentrated stock solutions were prepared individually as follows: solution A (50X, MOPS), solution B (25X, tripotassium citrate, citric acid monohydrate, sodium sulfate, potassium phosphate, and sodium bicarbonate), solution C (50X, ammonium chloride), solution D (50X, magnesium chloride hexahydrate, calcium chloride dihydrate, ferrous chloride tetrahydrate, L-cysteine HCl monohydrate) and solution E (50X, pyridoxamine dihydrochloride, PABA, D-biotin, vitamin B12, thiamine hydrochloride and thioctic acid). The final medium composition (g/L) was as follows: MOPS (C₇H₁₄NO₄SNa), 10; tripotassium citrate (K₃C₆H₅O₇), 2.0; citric acid monohydrate (C₆H₈O₇·H₂O), 1.25; sodium sulfate (Na₂SO₄), 1.0; potassium phosphate (KH₂PO₄), 1.0; sodium bicarbonate (NaHCO₃), 2.5; ammonium chloride (NH₄Cl), 1.5; magnesium chloride hexahydrate (MgCl₂·6H₂O), 1.0; calcium chloride dihydrate (CaCl₂∙2H₂O), 0.2; ferrous chloride tetrahydrate (FeCl₂·4H₂O), 0.1; L-cysteine HCl monohydrate (C₃H₇NO₂S·HCl·H₂O), 1.0; pyridoxamine dihydrochloride (C₈H₁₂N₂O₂ · 2HCl), 0.02; p-aminobenzoic acid (PABA) (H₂NC₆H₄CO₂H), 0.004; D-biotin (C₁₀H₁₆N₂O₃S), 0.002; vitamin B12 (C₆₃H₈₈CoN₁₄O₁₄P), 0.002; thiamine hydrochloride (C₁₂H₁₇ClN₄OS · HCl), 0.004; and thioctic (α-lipoic) acid (C₈H₁₄O₂S₂), 0.004 [28]. Each of the solutions was prepared individually, purged by 20 alternate cycles of 45 s of vacuum and 45 s of nitrogen, and then autoclaved for 30 min in a liquid cycle.

The inoculum for C. thermocellum was prepared by cultivating the micro-organism in 125 mL serum bottles (with a 50 mL working volume) with MTC medium and microcrystalline cellulose PH105 (5 g/L) (Sigma-Aldrich, St. Louis, MO, USA) as substrate for a period of 24 h at 55 °C and 180 rpm. The bottles were then stored at 4 °C for not longer than 30 days, and the same inoculum batch was used to initiate all CBP fermentations, ensuring consistency across experiments.

The inoculum for T. thermosaccharolyticum was prepared using the complex medium CTFüD as described by Olson and Lynd [29] modified with 5 g/L of D-xylose instead of cellobiose. CTFüD medium was purged as described previously. After overnight incubation in a 125 mL bottle (50 mL of working volume) at 55 °C and 180 rpm, the culture was aliquoted in 5 mL bottles and then stored at −80 °C. One day prior to the fermentation experiments, a frozen stock bottle (− 80 °C) was thawed and used to inoculate a 125 mL bottle (50 mL working volume) containing fresh CTFüD medium, which was incubated overnight at 55 °C and 180 rpm. On the following day (the day of SCB bottle inoculation), this overnight culture was used to inoculate a new bottle cultivated under the same conditions until the time of SCB inoculation to ensure an active and viable cell culture. All CBP fermentation runs were inoculated using T. thermosaccharolyticum cultures derived from the same − 80 °C frozen stock, to ensure consistency across experiments.

The simultaneous saccharification and fermentation (SSF) experiments were conducted using the yeast Saccharomyces cerevisiae PE-2 (Fermentec, Piracicaba, SP, Brazil) and the enzyme cocktail Cellic^® CTec2 (> 1000 hydrolysis units/g, density 1.2 g/mL, 92 g protein/L; Novonesis, Franklinton, NC, USA). The commercial enzyme preparation Cellic^® CTec2 was selected for this study based on its availability at the time the experiments were initiated. Although more recent formulations (e.g., CTec3) have been reported to provide improved performance, CTec2 remains used in the literature [30–33] and is a well-characterized enzyme system for lignocellulosic hydrolysis. S. cerevisiae cell stock was prepared by aerobically cultivating lyophilized yeast in 125 mL Erlenmeyer flasks containing 50 mL of YPD medium at 37 °C and 180 rpm for 24 h, according to the manufacturer’s instructions. The culture was aliquoted into 1 mL cryotubes with 5% glycerol and stored at −80 °C. All SSF inocula were prepared from the same − 80 °C cell stock to ensure consistency across fermentation rounds. For each round, the inoculum was prepared by aerobically culturing the content of one cryotube in 125 mL (50 mL of working volume) Erlenmeyer flasks containing YPD medium for 12–14 h [34]. Enzyme loading was set at 5 mg protein/g of solids (> 65.22 U/g of solids) [14], unless otherwise specified and expressed based on protein concentration (mg protein per g of solids), following Resch et al. [35], without direct activity normalization. All experiments were conducted using the same enzyme batch under consistent conditions. The protein concentration of the enzyme stock solution was measured using the Bradford protein assay [36]. Prior to analysis, the enzyme stock was diluted in 50 mM citrate buffer (pH 5.0) to fall within the linear range of the assay. A series of dilutions (200X, 250X, 300X, and 320X) was prepared and analyzed to ensure consistency in the measured protein concentration. No visible precipitation was observed upon dilution, and samples were analyzed without prior clarification. To further assess the accuracy of protein quantification, the enzyme concentration was also measured using the bicinchoninic acid (BCA) assay. The BCA assay yielded higher protein concentrations compared to the Bradford method (approximately 280 mg/mL vs. 92 mg/mL), consistent with known differences in assay response for cellulase preparations. Protein quantification using Bradford and BCA assays resulted in substantially different values, which is consistent with previous reports for cellulase systems. To assess whether this discrepancy affected experimental outcomes, SSF experiments were conducted using enzyme loadings based on both quantification methods (0.5 mm bagasse). No significant differences in carbohydrate solubilization were observed between the two conditions, indicating that enzyme loading was not limiting under the conditions tested (Additional File, Figure S1). YP medium and 50 mM citrate buffer (final pH 5.0) were employed, as described by Dowe and McMilan [34]. Solutions were prepared individually, purged with nitrogen gas as described before, and subsequently autoclaved at 121 °C for 30 min using a liquid cycle.

Consolidated bioprocessing (CBP) using a thermophilic bacterial coculture at different particle sizes

CBP fermentations were conducted in 125-mL serum bottles (50 mL of working volume). Comminuted bagasse was weighed out in the bottles, suspended in MilliQ water, and the bottles were purged and autoclaved as described for the other solutions. After cooling down and the media solutions addition, the bottles were inoculated with C. thermocellum and T. thermosaccharolyticum (2% v/v each) and incubated at 55 °C, 180 rpm for 7 days. To determine this 7-day incubation period, preliminary CBP and SSF experiments were conducted at 5-, 7-, and 10-day incubation time using the smallest (0.08 mm) and largest (4 mm) particle sizes. Carbohydrate solubilization showed no significant differences across timepoints for most conditions, with the exception of SSF with 4 mm particles, where solubilization increased by ~ 8% from days 5 to 7. Based on these results, 7 days was selected to ensure that optimum solubilization was reached for all particle sizes and both strategies without unnecessarily extending the process (Additional File, Figure S2).

After the incubation, the bottles were harvested, and their contents were centrifuged at 10,400 × g for 10 min (Hettich^® Universal 320 centrifuge, Rotor 1556, 9,000 rpm). The supernatant was collected, and the solids were washed with MilliQ water and centrifuged again under the same conditions before being dried at 55 °C for at least 5 days.

Simultaneous saccharification and fermentation (SSF) using fungal enzymes and yeast at different particle sizes

SSF experiments were conducted anaerobically in sealed 125-mL serum bottles with 50 mL of working volume. YP medium (10X concentrated) and citrate buffer (20X concentrated) were prepared, purged, and autoclaved as previously described for CBP. Comminuted bagasse was weighed into the bottles, suspended in MilliQ water, and the bottles were purged and autoclaved. After cooling, medium and buffer addition, the bottles were inoculated with a diluted and filter-sterilized enzyme cocktail aliquot and S. cerevisiae (2% v/v each), followed by incubation at 37 °C and 180 rpm for 7 days. The bottles were harvested as described for the CBP experiments.

Additional SSF assays were conducted using β-glucosidase from Aspergillus sp. (25 U/mg protein, 40 U/mL; Megazyme, Bray, Ireland), supplemented to Cellic^® CTec2 at 4.5 mg protein/g solids final concentration (> 58.7 U/g of solids) with 0.5 mg protein/g solids of β-glucosidase (12.5 U/g of solids), totaling 5 mg protein/g solids (~ 0.07 mg protein/mL). These assays were performed exclusively with sugarcane bagasse at 0.5 mm, the standard particle size for SSF in the group. SSF with high enzyme loading (20 mg protein/g solids of Cellic^® CTec2, > 260.9 U/g of solids) was also conducted using bagasse at 0.08 mm and 4 mm, representindg the smallest and largest particle sizes analyzed while minimizing the number of experiments. The catalytic functionality of the β-glucosidase preparation was confirmed in additional assays with cellobiose as substrate (Additional File, Table S1). In addition, enzymatic hydrolysis assays were performed under optimal conditions for Cellic^® CTec2 (50 °C, pH 5.0), using 10 g of carbohydrate/L of bagasse, water, 2.5 mM citrate buffer (pH 5.0), and 20 mg protein/g solids of the enzyme cocktail, following [37]. These hydrolysis assays also employed bagasse at 0.08 mm and 4 mm to represent the particle size range while limiting experimental workload.

Control fermentations were conducted with microcrystalline cellulose (20 µm, Sigma-Aldrich, St. Louis, MO, USA) for both CBP and SSF at the same carbohydrate loading (10 g/L) as in the other fermentations. In addition, SSF control fermentations were performed without bagasse to assess the contribution of the medium itself to residual sugar and ethanol concentrations.

Carbohydrate solubilization, residual sugars, and fermentation products

The carbohydrate content of the comminuted SCB, as well as the residual dry solids after solubilization assays was determined by quantitative saccharification (QS), which involves two-step acid hydrolysis; the first with 72% sulfuric acid at 30 °C for 1 h, the second diluted down to 4% at 121 °C and 1 atm for 1 h, as described by Sluiter et al. [24].

For the analysis of residual soluble carbohydrates (RSCs), 1-mL samples of the collected supernatant were subjected to liquid quantitative saccharification (LQS), consisting of acid hydrolysis with 2.5% (w/v) sulfuric acid at 121 °C and 1 atm for 1 h, in order to convert soluble oligomers into monomeric sugars.

For fermentation product analysis, a separate aliquot of the supernatant (1 mL) was mixed with 50 μL of 10% (w/v) sulfuric acid, followed by centrifugation at 10,400 × g for 2 min. This step was used only for sample stabilization and preparation, and did not involve hydrolysis of soluble carbohydrates.

The hydrolyzed sugars from both QS and LQS procedures, as well as fermentation products, were quantified by high-performance liquid chromatography (HPLC) equipped with a refractive index detector (RID). An Aminex HPX-87H column (Bio-Rad, USA) was used at 60 °C with a flow rate of 0.6 mL/min, employing 2.5 mM sulfuric acid as the mobile phase. Pure analytical standards of glucose, xylose, arabinose, and fermentation products acetic acid, ethanol (net ethanol for SSF), formic acid, and lactic acid at known concentrations were used for compound identification and quantification. Prior to analysis, all samples were filtered through 0.22-μm syringe filters.

Fractional total carbohydrate solubilization (FTCS) was calculated according to Eq. 2, where m₀ represents the initial mass (g) of unfermented bagasse, c₀ the total carbohydrate content (g/g solids) of milled unfermented bagasse, m_f the mass of residual solids after fermentation, and c_f its corresponding carbohydrate content. The total carbohydrate content is the sum of glucose, xylose and arabinose contents:

$$FTCS=1-\frac{m_f\times c_f}{m_0\times c_0}$$ 2

Statistical analysis

Statistical analysis was performed using one-way ANOVA followed by Tukey’s test, with a significance level of p ≤ 0.05. A two-way ANOVA was used to evaluate the effects and interaction between biocatalyst type and particle size on solubilization, followed by Tukey’s test for post hoc comparisons, with p ≤ 0.05. All analyses were performed using Statistica 7 (StatSoft Inc., Tulsa, OK, USA). Data presented in the figures represent the averages of n samples (with n indicated in the respective figures), and error bars correspond to one standard deviation.

Results

Deconstruction of sugarcane bagasse with no pretreatment other than autoclaving was evaluated for two systems: a coculture of Clostridium thermocellum and Thermoanaerobacterium thermosaccharolyticum, referred to as CBP, and a commercial cellulase preparation (Cellic^® Ctec2) in the presence of yeast except where otherwise indicated, referred to as SSF.

Particle size distribution and surface area characterization

The sugarcane bagasse provided by Raízen was comminuted into six distinct particle sizes and characterized for carbohydrate composition, particle size distribution (PSD), and specific surface area. PSD and surface area are particularly relevant, as they might influence microbial and enzymatic accessibility in both CBP and SSF systems [10]. The PSD curves are shown in Fig. 2, while the corresponding PSD parameters and BET specific surface area values are summarized in Table 2. Measurements for the 2 mm and 4 mm fractions could not be obtained due to limitations in the analytical range of the equipment. Figure 2 and Table 2 show that the volume-weighted mean particle size for each size closely approximated its respective target range. The D10, D50, and D90 values further reflect a narrowing of the particle size distribution with decreasing particle size. Despite this, the distributions—particularly for the smallest sizes—exhibited notable dispersion, as evidenced by elevated Span values (> 2), indicating broader size variation. As shown in Table 2, there was a general trend of decreasing BET specific surface area with increasing particle size, consistent with the expected increase in surface-to-volume ratio, meaning a larger available surface area per unit mass for smaller particles, which may enhance their exposure to enzymatic or microbial action. The smallest size (0.08 mm) exhibited the highest surface area (1.800 ± 0.012 m²/g), while larger particle size fractions showed progressively lower values, reaching 0.915 ± 0.004 m²/g for the 2-mm sample. Unexpectedly, a slight increase was observed for the 4 mm fraction (0.958 ± 0.015 m²/g), deviating from the expected trend. This variation may reflect the inherent heterogeneity in particle morphology and porosity in larger material, which can introduce variability in surface area measurements despite their overall size.

Fig. 2.

Particle size distribution curves (frequency vs. particle diameter) for the 0.08, 0.2, 0.5, and 1.0 mm bagasse particles. Measurements were performed using a laser diffraction particle size analyzer. Sizes 2.0 and 4.0 mm were not included due to equipment limitations

Table 2.

Specific surface area (BET method) and particle size distribution parameters (D₁₀, D₅₀, D₉₀, Span, and volume-weighted mean diameter) of sugarcane bagasse at different particle sizes

Particle size (mm)	D10 (mm)	D50 (mm)	D90 (mm)	Span	Mean (μm)	BET specific surface area (m2/g)
0.08	0.01	0.08	0.23	2.76	110.59	1.800 ± 0.012
0.20	0.03	0.15	0.41	2.52	206.35	1.512 ± 0.009
0.50	0.05	0.25	0.60	2.17	316.7	0.972 ± 0.017
1.00	0.18	0.48	1.27	2.28	663.91	0.963 ± 0.002
2.00	–	–	–	–	–	0.915 ± 0.004
4.00	–	–	–	–	–	0.958 ± 0.015

Data for the 2.0 and 4.0 mm sizes were not obtained by laser diffraction due to instrumental limitations. Results are expressed as mean ± one standard deviation

Sugarcane bagasse solubilization at different particle sizes

Fractional total carbohydrate solubilization for both CBP with thermophilic bacteria and SSF with fungal enzymes and yeast across the six particle sizes is presented in Fig. 3. In all cases, the thermophilic coculture yielded higher solubilization than the fungal enzyme–yeast system, with values ranging from 1.9-fold to over 3.7-fold higher, depending on particle size. The 0.2 mm size yielded the highest solubilization for both approaches, surpassing even the 0.08 mm particles, and the difference was statistically significant (Tukey test, p = 0.008). For SSF, solubilization dropped by approximately 57% with the increase of particle size, whereas for CBP, the reduction was more moderate, at around 28%. Statistical analysis showed that for CBP and SSF, the solubilization values between 1.0, 2.0, and 4.0 mm for each biocatalyst type were not significantly different (Additional files, Tables S2, S3, and S4). Furthermore, to assess the relative influence of biocatalyst type and particle size on carbohydrate solubilization, a two-way ANOVA was performed. Both factors had significant effects (p < 0.0001), whereas no significant interaction was observed (p = 0.82). The biocatalyst explained approximately 75% of the total variance in solubilization, while particle size accounted for 21%, indicating that the choice of biocatalyst had the dominant impact on the solubilization values observed (Additional files, Table S5).

Fig. 3.

Fractional total carbohydrate solubilization (FTCS) of sugarcane bagasse at six particle sizes via Consolidated Bioprocessing (CBP) with a thermophilic coculture and Simultaneous Saccharification and Fermentation (SSF) with fungal enzymes and yeast and error bars indicate one standard deviation (n = 6)

During fermentation of lignocellulosic feedstocks by C. thermocellum and T. thermosaccharolyticum, not all sugars that are solubilized are utilized, resulting in residual soluble carbohydrates (RSCs) as described for different types of feedstock [5, 27, 38]. Since the strain of yeast used here is unable to take up C₅ sugars, there are also soluble sugars present at the end of SSF experiments. Figure 4 presents the concentration of residual sugars shown as glucose, xylose, and arabinose (quantification by mild acid hydrolysis) in the fermentation broth for each particle size for (A) CBP and (B) SSF. The bars are stacked to allow visualization of both total residual sugar and the relative contribution of each monosaccharide. Despite a significant difference between overall carbohydrate solubilization values, as shown in Fig. 2, the residual sugar concentrations were comparable between CBP and SSF across the particle size range with SSF experiments exhibiting slightly higher values and greater variability, as reflected by larger standard deviations. In both systems, xylose consistently represented the major component of residual sugars, followed by glucose and, to a lesser extent, arabinose. This pattern was observed across all particle sizes.

Fig. 4.

Residual concentrations of glucose, xylose, and arabinose in the fermentation broth for each particle size for A consolidated bioprocessing (CBP) with thermophilic bacteria and B simultaneous saccharification and fermentation (SSF) with yeast and fungal enzymes. Bars are stacked to represent the total residual sugar and to visualize the individual contribution of each sugar. Values represent the average of 6 biological replicates (n = 6) and error bars represent one standard deviation

Fermentation product concentrations for each sugarcane bagasse particle size are shown in Fig. 5. Figure 5A presents the results of CBP with wild-type thermophilic bacteria; while ethanol, acetic acid, lactic acid, and formic acid were quantified, formic and lactic acid concentrations were found to be very low, and are not shown (values can be found in Table S6, Additional files). Figure 5B presents the results of SSF with yeast and fungal enzymes, for which only ethanol was detected as fermentation product, and net ethanol production is shown. In the CBP experiments, ethanol and acetic acid were the main fermentation products across all particle sizes, with ethanol concentrations consistently higher than those observed for SSF experiments. The sum of fermentation products (ethanol + organic acids) in CBP was also greater than the ethanol-only concentrations in SSF.

Fig. 5.

Fermentation product concentrations for A consolidated bioprocessing (CBP) with thermophilic bacteria and B simultaneous saccharification and fermentation (SSF) with fungal enzymes and yeast for each sugarcane bagasse particle size. For CBP, quantified products include ethanol, lactic acid, formic acid, and acetic acid. For SSF, only ethanol was analyzed. Bars represent the average values of 6 biological replicates (n = 6) and error bars represent one standard deviation

To further compare the performance of both approaches across particle sizes, the solubilization of C₆ (glucose) and C₅ (xylose and arabinose) sugars was evaluated. Figure 6A displays the relationship between C₆ and C₅ solubilized mass, combining data from both approaches and all particle sizes. Linear regression showed a strong correlation (R² = 0.964) with a slope of 0.640, which is very similar to the ratio 0.645 for hemicellulose and cellulose content for the SCB used here (Table 1, Methods). It suggests that, despite the different mechanisms, the solubilization of C₆ and C₅ sugars occurs proportionally to the content in unpretreated sugarcane bagasse for both systems. In Fig. 6B, the fractional solubilization for CBP is plotted vs. that for SSF. As all data points are above the y = x line, this clearly indicates that the thermophilic coculture uniformly outperforms the fungal enzyme–yeast system for both C₆ and C₅ solubilization, regardless of particle size. Figure 6C shows the fractional total carbohydrate solubilization (FTCS) as a function of particle size consistent with the expected trend of decreasing solubilization with increasing particle size, and enabling comparison of how each biocatalyst responds to changes in substrate accessibility. To be able compare the impact of particle size for both systems more objectively, each data set was normalized to the lowest solubilization value (FTCS largest particle 4 mm = 1). In Fig. 6D, the increase in normalized FTCS is shown compared to the normalized FTCS of largest particle size for each biocatalyst system, respectively. Note that this results in an increase that equals 0 for normalized FTCS of 4 mm. Both systems were affected by particle size variation; however, SSF seemed to be more affected by particle size than CBP. In SSF, solubilization decreased by approximately 2.2-fold from 0.2 mm (highest value) to 4 mm, whereas in CBP, the decrease was about 0.55-fold. The smallest decrease for SSF is ~ 0.5 at 1.0 mm particle size, which is just shy of the largest difference for CBP of ~ 0.55 at 0.2 mm particle size.

Fig. 6.

Comparative analysis of CBP with thermophilic bacteria and SSF with fungal enzyme and yeast. A Solubilized mass of C₅ (xylose + arabinose) sugars vs. C₆ (glucose). B Fractional total carbohydrate solubilization (FTCS) in CBP vs. SSF for both C₆ and C₅ sugars. C Fractional total carbohydrate solubilization vs. particle size. D Difference in normalized solubilization relative to the 4 mm particle size using 4 mm particle size solubilization as normalization reference, for both CBP and SSF, respectively

Control experiments with microcrystalline cellulose (MCC) were conducted and the results for solubilization, residual soluble sugars, and fermentation products are shown in Fig. 7. Solubilization was higher with MCC than with SCB for both approaches, MCC is a model substrate that consists almost entirely of cellulose and lacks the lignin–hemicellulose matrix that contributes to lignocellulosic recalcitrance. While cellulose crystallinity itself also hinders hydrolysis [39], the small particle size of MCC (0.02 mm; 20 µm) theoretically provides a larger available surface area than the smallest particle size SCB tested here. Under these conditions the CBP with wild-type thermophilic bacteria yielded higher solubilization and more total products, including ethanol, lactic acid, formic acid, and acetic acid, whereas SSF with fungal enzyme and yeast produced solely ethanol. However, CBP resulted in more residual sugars, consisting exclusively of glucose, compared to SSF. In pH-controlled bioreactors, monocultures of C. thermocellum have been shown to solubilize > 100 g/L MCC to near completion in a 7 day time frame [20, 40].

Fig. 7.

Fractional total carbohydrate solubilization (A), residual glucose (B), and fermentation products (C) in control fermentations with microcrystalline cellulose (MCC). Glucose was the only residual soluble sugar detected, and ethanol was the only fermentation product observed in SSF. Values are the average of 3 biological replicates (n = 3) with error bars representing one standard deviation

To investigate whether solubilization by SSF with fungal enzyme and yeast could be improved, a number of tests were done. SSF carbohydrate solubilization results for different enzymatic strategies are shown in Fig. 8. The comparison included Simultaneous Saccharification and Fermentation (SSF) at standard enzyme loading (5 mg protein/g solids), SSF with increased enzyme loading (20 mg protein/g solids), and enzymatic hydrolysis conducted under optimal conditions for fungal cellulases (50 °C, pH 5.0, without yeast). These strategies were applied to sugarcane bagasse at two particle sizes (0.08 mm and 4.00 mm) and to microcrystalline cellulose (MCC). For the smallest particle size (0.08 mm), increasing the enzyme loading fourfold resulted in only a modest increase in carbohydrate solubilization (around 5%). For the biggest particle size (4.00 mm) and MCC, higher enzyme loading produced no measurable improvement in solubilization. Enzymatic hydrolysis under optimal temperature and pH conditions slightly improved MCC solubilization (~ 8%) but had no effect on either particle size of sugarcane bagasse. Across all strategies, solubilization remained substantially lower for unpretreated SCB compared to MCC.

Fig. 8.

Fractional total carbohydrate solubilization under different enzymatic strategies: simultaneous saccharification and fermentation (SSF) with 5 mg protein/g solids, SSF with increased enzyme loading (20 mg protein/g solids), and enzymatic hydrolysis under optimal conditions for fungal cellulases in the absence of yeast (50 °C, pH 5.0). Experiments were performed using the smallest and largest sugarcane bagasse particle sizes (0.08 mm and 4.00 mm) and microcrystalline cellulose (MCC) and the incubation time was 7 days. Values are the average of 3 biological replicates (n = 3) with error bars representing one standard deviation

It is also noteworthy that one potential source of uncertainty in the estimation of residual carbohydrate in SSF experiments is the contribution of glucan from Saccharomyces cerevisiae cell walls (~ 55–65%) [41], which may be hydrolyzed during the two-step acid-hydrolysis procedure. This effect may be more pronounced under the low-solid conditions used in this study, where microbial biomass can represent a non-negligible fraction of the total glucan measured in residual solids. In contrast, this contribution is expected to be less significant in CBP systems due to differences in microbial size and cell wall composition. However, as shown in Fig. 8, enzymatic hydrolysis experiments conducted in the presence and absence of yeast—across different substrate conditions—did not result in significant differences in carbohydrate solubilization. This suggests that the contribution of yeast-derived glucan is limited under the conditions tested. Therefore, while this factor may introduce some uncertainty in absolute values, it does not affect the main comparative trends observed in this study.

Because most fungal enzyme cocktails have limited β-glucosidase activity, the enzyme responsible for hydrolyzing cellobiose into glucose, supplementation with additional β-glucosidase has been proposed as a strategy to overcome this bottleneck in lignocellulose saccharification [42]. Accordingly, SSF experiments were performed with β-glucosidase from Aspergillus niger to assess whether carbohydrate solubilization could be improved. Carbohydrate solubilization results for sugarcane bagasse (0.5 mm) and microcrystalline cellulose (MCC) under supplemented and non-supplemented enzymatic conditions are shown in Fig. 9. SSF was performed using Cellic^® CTec2 at a standard loading of 5 mg protein/g solids, and also under supplemented conditions with the addition of fungal β-glucosidase (0.5 mg protein/g solids), labeled as SSF_Sup and MCC_Sup. No significant increase in solubilization was observed in the supplemented conditions for either substrate, indicating that the addition of β-glucosidase did not enhance overall carbohydrate release.

Fig. 9.

Fractional total carbohydrate solubilization for 0.5 mm sugarcane bagasse and microcrystalline cellulose (MCC) under different conditions. Experiments included SSF with Cellic^® CTec2 at 5 mg protein/g solids (labeled as SSF and MCC), and supplemented SSF with Cellic^® CTec2 (4.5 mg protein/g solids) plus a fungal β-glucosidase (0.5 mg protein/g solids), labeled as SSF_Sup and MCC_Sup. Bars represent mean values of biological replicates (n = 3), with error bars indicating one standard deviation

Discussion

The particle size distribution profiles obtained in this study align with those reported in the literature for milled sugarcane bagasse [25]. The high Span values observed in smallest sizes suggest substantial variability within size classes. These broad distributions can be attributed in part to the sequential milling steps needed for repeated milling steps to achieve finer particles, as successive mechanical stresses—especially shear forces, in this case—fragment biomass along multiple structural planes, leading to the formation of more diverse particle populations. In addition, the irregular morphology of lignocellulosic particles may allow some larger fragments to pass through sieves oriented toward nominal sizes, further contributing to the dispersion. Nevertheless, the overall size differentiation was sufficient to evaluate particle size effects on carbohydrate solubilization. Specific surface area increased with decreasing particle size, which can be associated with improved accessibility for enzymes and micro-organisms. Furthermore, comminution may also disrupt polysaccharide chains and enhance their exposure at the particle surface, further contributing to biomass accessibility. However, while surface area contributes to differences in carbohydrate solubilization, it is unlikely to be the predominant factor for untreated lignocellulosic biomass, where structural recalcitrance—primarily due to lignin–carbohydrate interactions—plays a major limiting role. Thus, particle size should be interpreted as a modulating factor, rather than the primary determinant of solubilization [43].

The higher solubilization observed in CBP with thermophilic bacteria across all particle sizes is consistent with previous findings using other unpretreated lignocellulosic feedstocks [14, 16, 19]. The smaller decline in solubilization with increasing particle size in CBP compared to SSF suggests a greater robustness of the thermophilic coculture to variations in feedstock particle size. Although such trends were studied for enzymatic systems [22], their quantitative evaluation for CBP—particularly on unpretreated substrates—remains limited, and the present results help to clarify how these systems respond differently to physical variations in the substrate. This behavior should be interpreted in the context of the inherent recalcitrance of untreated lignocellulosic biomass, in which lignin and hemicellulose form structural barriers that limit enzymatic accessibility. In such systems, particle size reduction alone is insufficient to overcome these constraints. This trend may be attributed to the ability of cellulolytic bacteria such as Clostridium thermocellum to deconstruct lignocellulose through cellulosome-mediated mechanisms, enhancing substrate attack even in larger particles [13]. In addition, the broader metabolic capacity of the CBP coculture, which can utilize both hexose and pentose sugars, contrasts with the yeast used in the SSF experiments reported here. This yeast primarily ferments glucose and has not been reported to naturally utilize xylose or arabinose [44], potentially limiting sugar conversion. The unexpectedly higher solubilization at 0.2 mm compared to 0.08 mm for both biocatalyst systems suggests that extremely fine particles may introduce additional limitations, such as the formation of aggregates [45], counteracting the benefits of increased surface area. These findings point out the need for further investigation into the physical and biological interactions governing substrate accessibility at microstructural levels. The proportional relationship observed between the solubilization of C₆ and C₅ sugars (Fig. 6A), despite the differing microbial and enzymatic systems in CBP and SSF, highlights a shared behavior of the C₆ and C₅ solubilization of unpretreated lignocellulose and aligns with previous studies using other feedstocks [13]. Together, these observations reinforce that differences between CBP and SSF on untreated biomass are governed not only by substrate accessibility, but also by the distinct mechanisms through which each system interacts with the lignocellulosic matrix.

One of the major challenges in comparing the efficiency of microbial systems for lignocellulose deconstruction across studies is the lack of standardized metrics. While the present study adopts fractional total carbohydrate solubilization—which accounts for glucose, xylose, and arabinose contents before and after fermentation, as well as overall mass loss—most published SSF studies report alternative metrics, such as total reducing sugars (TRS), individual glucose and xylose conversions or ethanol titers and yields often excluding arabinose from the analysis. This metric difference complicates direct comparisons and indicates the need for unified reporting standards particularly for comparison purposes. To address this, we compiled and estimated fractional solubilization values from selected studies (Table 3), based on available results and reported biomass composition. Importantly, comparisons were restricted to control experiments using untreated sugarcane bagasse, when available, to ensure consistency with the substrate conditions used in this study. As the majority of studies in this field focus on pretreated substrates, these control conditions represent the most appropriate available basis for contextual comparison. The purpose of this analysis is not to enable direct or definitive comparison between studies, but rather to position the results obtained here within the range of values reported in the literature under broadly comparable conditions. These estimates should, therefore, be interpreted cautiously due to differences in enzyme systems, substrate preparation, analytical methods, and data limitations. This study was designed to compare CBP and SSF under controlled and non-optimized conditions using untreated sugarcane bagasse, with the aim of isolating intrinsic differences between the two systems. Therefore, the results presented here should not be interpreted as directly representative of industrial performance, where factors such as pretreatment, higher solids loading, and process optimization play a critical role. When comparing the fractional total carbohydrate solubilization as well as glucan and xylan solubilization values obtained in this study with those estimated from selected literature reports (Table 3), the SSF results in this study are generally within or above the range reported. Some variability is observed, likely due to differences in enzyme quality, loading, substrate type, and analytical methods. Within the limitations described above, the CBP results in this study fall within the upper range of values reported for SSF/SHF systems using untreated bagasse. However, these comparisons should be interpreted with caution and are intended primarily to provide context rather than to support direct performance ranking across studies:

$$\text{Fractional total carbohydrate solubilization}=\frac{\text{TRS}\left(\text{g}\left./\phantom{\text{gL}}\right)\text{L}\right)}{\left[\text{SL}\times \left(\text{C}\times 1.10+\text{H}\times 1.14\right)\right]}$$

Table 3.

Summary of enzymatic hydrolysis reports to date using unpretreated sugarcane bagasse (SCB) as the substrate

References	Process type	Particle size	Conditionsa	Reported results	Fractional total carbohydrate solubilizationb
This study	Enzymatic hydrolysis (no fermentation)	0.08 mm 4 mm	14 g/L; 50 °C; 168 h; Cellic®CTec2 (> 260.9 U/g substrate); citrate buffer (50 mM, pH 5)	–	0.257 (0.08 mm) 0.093 (4 mm)
[46]	Enzymatic hydrolysis (no fermentation)	Crushed (particle size not specified)	100 g/L; 50 °C; 72 h; Cellic®CTec3 (8.5 FPU/g cellulose); citrate buffer (50 mM, pH 4.8)	Total reducing sugars released: 3.06 g/L	~ 0.048c
[32]	Enzymatic hydrolysis (no fermentation)	100 mesh (~ 0.15 mm)	20 g/L; 50 °C; 72 h; Cellic®CTec2 (20 FPU/g substrate); sodium acetate buffer (0.1 M, pH 4.8) + antibiotics	Total reducing sugar content (TRS): 1.31 g/L (glucose and xylose)	~ 0.094d
[47]	Enzymatic hydrolysis (no fermentation)	0.35–0.85 mm	25 g/L, 50 °C, 72 h, Cellic®HTec2/CTec2 (15 FPU cellulase + 1206.7 U xylanase/g biomass), citrate buffer 50 mM pH 4.8 + 0.02% NaN₃	Glucose conversion: 29.7% Xylose conversion of 23.6%	~ 0.278e
[48]	Enzymatic hydrolysis (no fermentation)	≤ 9.8 mm (as received; no washing/milling)	25 g/L, 50 °C, 72 h, Accellerase 1500 (25 FPU/g) + Novozyme 188 (50 BGU/g), citrate buffer (50 mM, pH 5)	~ 22% hydrolysis yield (from cellulose only)	0.22 (glucose conversion)
[49]	Enzymatic hydrolysis (no fermentation)	1–3 mm	5 g/L, 55 °C, 24 h, Myceliophthora thermophila enzymes (47, 1, and 220 U/g) or Celluclast (47/14.6/7.3 or 500/156.6/78.3 U/g), acetate buffer (0.1 M, pH 5)	Total reducing sugars released: ~ 24 mg/g of bagasse	~ 0.034f

^aConditions include solids loading, temperature, time, enzyme loading, and medium.

^bWhile the metric of carbohydrate solubilization (i.e., the proportion of total structural carbohydrates released as soluble sugars) is commonly applied in CBP studies to assess biomass conversion, it is less frequently reported in SSF literature. Therefore, the values estimated here should be interpreted with caution. They rely on compositional assumptions, exclude arabinose unless otherwise stated, and may differ from more conventional metrics, such as glucose yield or cellulose conversion. Carbohydrate solubilization was calculated when possible, based on reported reducing sugars and biomass composition. “TRS” refers to total reducing sugars. Arabinose content was not considered in these estimates.

^cSolubilization calculated based on the concentration of total reducing sugars (TRS) released after enzymatic hydrolysis (3.06 g/L), relative to the total carbohydrate content in raw sugarcane bagasse (39.84% cellulose and 17.19% hemicellulose), according to the equation below. The theoretical maximum release of carbohydrates was corrected by hydrolysis stoichiometry using conversion factors of 1.10 for cellulose-to-glucose and 1.14 for hemicellulose-to-pentosan, resulting in 63.41 g/L of potentially releasable sugars. Note the calculation below (arabinose content was not considered in the calculation due to lack of compositional data). The reported solubilization percentage may, therefore, slightly underestimate the total carbohydrate release. Where TRS = total reducing sugars released after hydrolysis (g/L), SL = solid loading (g/L), C = cellulose content in the biomass (fraction of dry weight), H = hemicellulose content in the biomass (fraction of dry weight), 1.10 = conversion factor from anhydroglucose to glucose, 1.14 = conversion factor from anhydroxylose to xylose.

^dCarbohydrate solubilization estimated based on TRS, using the same approach described in^c

^eBased on glucose and xylose conversion yields

^fCarbohydrate solubilization was estimated by dividing the amount of released reducing sugars (24 mg/g bagasse) by the theoretical maximum sugar release, calculated from the glucan and xylan content (46.9% and 16.3%, respectively) using conversion factors of 1.11 (glucan → glucose) and 1.13 (xylan → xylose). The estimated maximum releasable sugar was 0.7048 g/g bagasse, resulting in a solubilization of ~ 3.4%. Arabinose content was not considered in this estimation

The predominance of xylose in the residual sugar profile for both systems is consistent with certain limitations in microbial utilization of pentoses in these systems. In the case of SSF, the high xylose concentration and total residual soluble carbohydrates (RSCs) were expected, given that the strain of Saccharomyces cerevisiae—the fermenting organism used here—lacks the native metabolic capacity to consume xylose. The CBP system employs a coculture designed to broaden sugar utilization, including a hemicellulose-fermenting partner (T. thermosaccharolyticum), with acid-hydrolysis monomerized xylose and glucose remaining as residual sugars in similar amounts as reported by Zambello et al. [5] for 10 g/L of solids loading. As reported in the literature for switchgrass and corn stover, the absence of T. thermosaccharolyticum in the bacterial culture yields much higher values of RSCs [27]. The total concentration of fermentation products in CBP was higher than the ethanol concentrations observed in SSF across all particle sizes. Notably, ethanol levels in the wild-type CBP also surpassed those in SSF, likely due to the greater carbohydrate solubilization achieved by the thermophilic bacterial coculture. The ethanol concentrations as well as total products concentrations obtained under CBP conditions are consistent with the values reported by Zambello et al. [5] for similar solid loading (10 g/L) for non-engineered bacteria. It is noteworthy that for the SSF experiments, ethanol concentrations obtained with particle sizes of 2 and 4 mm were slightly higher than the expected theoretical values, based on the amount of solubilized glucose and the yeast ethanol yield. This may reflect the inherent challenges of accurately measuring fermentation products in unpretreated bagasse, particularly when using YPD medium. Such deviations likely arise from as-yet-unknown factors and highlight the complexity of fermenting unpretreated lignocellulosic substrates. It should be noted that SSF experiments were conducted in sealed serum bottles, which may have led to CO₂ accumulation and partial inhibition of yeast fermentation. While this may have affected ethanol production, CO₂ is not expected to significantly impact enzymatic hydrolysis. Therefore, its influence on carbohydrate solubilization—the primary metric used in this study—is considered to be limited. The MCC results suggest that CBP achieved higher solubilization and product formation also in the absence of recalcitrance associated with lignocellulose, indicating greater efficiency in solubilizing and utilizing highly crystalline cellulose [21].

The limited effect of increased enzyme loading on carbohydrate solubilization is in accordance with what was reported by Paye et al. [14] using switchgrass. Even with a fourfold increase in protein concentration, solubilization improved only slightly for the finest SCB particles and remained unchanged for larger particles and MCC. Similarly, performing enzymatic hydrolysis under optimal conditions for fungal cellulases (50 °C, pH 5.0) or supplementing with β-glucosidase did not enhance solubilization for SCB. These findings suggest that enzyme accessibility, rather than enzyme quantity, is the primary constraint under these conditions. Taken together, these results underscore the limited efficacy of conventional enzymatic strategies for unpretreated SCB, even under favorable loading and conditions.

These limitations are consistent with the known performance of CTec2 on untreated lignocellulosic substrates, for which the enzyme cocktail was not specifically designed. In the absence of pretreatment, structural barriers such as lignin–carbohydrate complexes and limited pore accessibility further restrict enzyme effectiveness, contributing to the low solubilization observed under SSF conditions. In addition, CTec2 alone has limited hemicellulolytic activity, and the absence of accessory enzymes such as those present in HTec2 may further constrain xylan hydrolysis. It is noteworthy, however, that more advanced commercial enzyme cocktails are available than the one tested here (Cellic^® CTec2); CTec3 and subsequent generations have shown improved activity profiles [50]. Alternatively, newer experimental enzyme preparations as reported by Santos et al. [51] would also be of interest to be tested side-by-side with CBP biocatalysts.

Conclusions

This study directly compared CBP with a thermophilic bacterial coculture and SSF with a fungal enzyme–yeast system for the conversion of unpretreated sugarcane bagasse (SCB), focusing on carbohydrate solubilization across a wide particle size range (0.08–4.00 mm). CBP consistently resulted in higher solubilization values than SSF regardless of particle size and for microcrystalline cellulose, a model substrate. In addition, CBP seemed to be less affected by particle size variation, with a drop of around 0.5-fold from the 0.2-mm to the 4-mm particle size, whereas for SSF, this drop was around 2.2-fold, suggesting a greater ability to overcome structural recalcitrance, even in greater particle sizes. The results also showed a strong linear relationship between glucan and xylan and arabinan solubilization in both systems, supporting the idea that these polymers are deconstructed proportionally under the tested conditions for both systems. Finally, this work reinforces the potential of CBP with thermophilic bacteria as a promising strategy for second-generation ethanol production from SCB without the need for thermochemical pretreatment. The adoption of standardized metrics like fractional total carbohydrate solubilization is also encouraged to improve comparability across studies and advance the field of lignocellulosic biofuels.

Abbreviations

CBP

Consolidated bioprocessing

SSF

Simultaneous saccharificantion and fermentation

SHF

Separate hydrolysis and fermentation

SSCF

Simultaneous saccharification and cofermentation

BET

Brunauer–Emmett–Teller

SCB

Sugarcane bagasse

MCC

Microcrystalline cellulose

MTC

Medium for thermophilic clostridia

YPD medium

Culture medium with yeast extract, peptone, and dextrose

MOPS

4-Morpholinopropanesulfonic acid

QS

Quantitative saccharification

LQS

Liquid quantitative saccharification

LRAC

Biomass Characterization, Analytical Resources, and Calibration Laboratory (FEQ/Unicamp)

FTCS

Fractional total carbohydrate solubilization

RSCs

Residual soluble carbohydrates

PSD

Particle size distribution

HPLC

High-performance liquid chromatography

Author contributions

MNPS: acquisition, analysis and interpretation of data; drafting the work; and revision of the work. LRL.: conception and design of the work; resources; revision of the work; funding acquisition; supervision; and project administration. EKH.: conception and design of the work; analysis and interpretation of data; resources; drafting the work; revision of the work; funding acquisition; supervision; and project administration.

Funding

This work was supported by the São Paulo Research Foundation (FAPESP), grant numbers 18/25682-0 and 22/05220-8. MNPS was financially supported by FAPESP, grant number 21/05397-2. Funding for contributions of EKH and LRL to this work was provided by the Center for Bioenergy Innovation (CBI), U.S. Department of Energy, Office of Science, Biological and Environmental Research Program under Award Number ERKP886.

Data availability

All data generated or analyzed during this study are included in the published article and its additional information files.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

LRL is a shareholder and CSO in a start-up company focusing on biofuel. All other authors declare to have no competing interests.