Direct Conversion of Minimally Pretreated Corncob by Enzyme-Intensified Microbial Consortia

Abstract

The presence of diverse carbohydrate-active enzymes (CAZymes) is crucial for the direct bioconversion of lignocellulose. In this study, various anaerobic microbial consortia were employed for the degradation of 10 g/L of minimally pretreated corncob. The involvement of lactic acid bacteria (LAB) and a CAZyme-rich bacterium (Bacteroides cellulosilyticus or Paenibacillus lautus) significantly enhanced the lactic acid production by Ruminiclostridium cellulolyticum from 0.74 to 2.67 g/L (p < 0.01), with a polysaccharide conversion of 67.6%. The supplement of a commercial cellulase cocktail, CTec 2, into the microbial consortia continuously promoted the lactic acid production to up to 3.35 g/L, with a polysaccharide conversion of 80.6%. Enzymatic assays, scanning electron microscopy, and Fourier transform infrared spectroscopy revealed the substantial functions of these CAZyme-rich consortia in partially increasing enzyme activities, altering the surface structure of biomass, and facilitating substrate decomposition. These results suggested that CAZyme-intensified consortia could significantly improve the levels of bioconversion of lignocellulose. Our work might shed new light on the construction of intensified microbial consortia for direct conversion of lignocellulose.

1. Introduction

Lignocellulose represents one of the most abundant yet challenging renewable carbon resources to harness. Corncob is such an abundant agricultural byproduct resource, with an annual production of 164 Tg across the world and 45.9 Tg in China, and it has drawn much attention for its utilization [1]. These lignocellulosic resources consist of not only three major components, namely cellulose, hemicellulose, and lignin, but also some minor yet more complex ingredients, such as pectin [2]. Generally, the carbohydrate fraction of lignocellulose is mainly built up by a series of glycoside units through various glycosidic bonds as well as ester bonds. Particularly, both the hemicellulose and pectin chains can be decorated with extensive branches [3], and the pectin fraction also consists of at least 20 kinds of glycosidic linkages [4]. These substrates are not liable to hydrolysis simply because of the lack of sufficient carbohydrate-active enzymes [5]. During the hydrolysis process, these complex carbohydrates may generate hydrolysis-resistant oligosaccharides, which could compete with other substrates for enzyme active sites, and in turn inhibit the hydrolysis reaction [6,7]. Besides the lignin recalcitrance, both the diverse glycoside units and glycosidic linkages have already become a major challenge for the utilization of this renewable resource.

Many strategies and methods have been taken toward efficient conversion of lignocellulose, with bioprocessing as one of the mildest and most environment-friendly measures. Several major technologies, such as simultaneous saccharification and fermentation (SSF), separate hydrolysis and fermentation (SHF), and consolidated bioprocessing (CBP), have been established for the efficient conversion of lignocellulosic biomass [8]. CBP stands out as an ideal strategy because of its simple one-pot processing design, which may cut down the enzyme costs and minimize the inhibition by the intermediate products. However, these technologies usually need a process of biomass pretreatment [9], which may require additional costs and result in the generation of some toxic compounds and contamination [10,11]. Hence it would be beneficial to establish a method of CBP without the pretreatment step or only involving minimal pretreatment of biomass, which refers to slight grinding, as well as autoclaving, but no harsh physicochemical pretreatment.

CBP coupled with a powerful microbial consortium is a potential solution for the direct conversion of minimally pretreated biomass. Division of labor is one of the advantages of a sophisticated microbial consortium for the conversion of complex substrates [12]. “Bottom-up” and “top-down” strategies have been employed to design microbial consortia [13]. For the past two decades, an array of microbial consortia has been constructed for the conversion of lignocellulose, where a cellulose “degrader” was accompanied by an efficient “metabolizer”, the latter of which will quickly convert the evolved small molecular sugars into various biochemicals [14,15,16,17]. For example, a cellulose degrader, Clostridium thermocellum, has been cocultured with the metabolizer Thermoanaerobacterium thermosaccharolyticum, as well as several hemicellulases, aiming towards the efficient conversion (>90%) of corn fiber, a type of biomass free of lignin [14]. Nevertheless, the direct bioconversion of minimally pretreated lignocellulose is still challenging, with the percentage conversion usually below or close to 50%, although with optimized microbial consortia [18,19]. A recent review suggested that the rumen employs diverse CAZymes to ensure that every component is efficiently degraded [3]. Obviously, a single microbial degrader is insufficient to decompose minimally pretreated lignocellulose, thus additional degraders that contain more diverse CAZymes should be recruited in these microbial consortia, including some enzyme preparations. In addition, since pentoses could usually not be readily metabolized in the presence of hexose, an additional pentose “metabolizer” should also be involved in the consortia [20].

In this study, a major mesophilic cellulolytic bacterium, Ruminiclostridium cellulolyticum [21], was cocultured with two lactic acid bacteria (LAB; Lactococus sp. X1 for hexose fermentation [22] and Lactobacillus pentosus for pentose conversion [23]) to directly decompose minimally pretreated corncob. Moreover, one or two species (Bacteroides cellulosilyticus [24] and Paenibacillus lautus [25]) that are rich in CAZyme genes, and a commercial cellulase cocktail, CTec 2, which is rich in an exoglucanase of glycoside hydrolase family 7 (GH7), were also involved so as to construct more powerful microbial consortia. Most importantly, these bacteria were also chosen because they can grow under similar conditions. Several pivotal factors were determined to check the performances of different microbial consortia, such as chemical productions, enzymatic activities, the composition and structure of the residual biomass, and the population dynamics of the consortia. The findings of this work may shed new light on the construction of powerful microbial consortia for the efficient decomposition of biomass and utilization of agricultural residuals.

2. Materials and Methods

2.1. Strains and Chemicals

R. cellulolyticum DSM 5812, B. cellulosilyticus DSM 14838, and P. lautus DSM 3035 (identical to NBRC 15380) were purchased from DSMZ (Braunschweig, German). Lb. pentosus 1.505 (identical to ATCC 8041) was purchased from GDMCC (Gugangzhou, China). Lactococcus sp. X1 (CGMCC 24341) was previously isolated by our lab and deposited in CGMCC (Beijing, China) [22]. Carboxymethylcellulose (CMC) was purchased from SCRC (Shanghai, China). Xylan and cellulase CTec 2 were purchased from Sigma (Beijing, China). Pectin (from citrus peel) was purchased from Sangon (Shanghai, China). Arabinan was obtained from Megazyme (Bray, Ireland). Corncobs and wheat brans were collected from a local farm, air-dried, ground using a grinder (CS-700, Tianqishengshi, Jinhua, China), and passed through a 60-mesh screen.

2.2. Media Preparation and Bacteria Inoculation

The fermentation medium contained (per liter) the following: tryptone 1.0 g, yeast extract 2.0 g, beef extract 2.0 g, KH₂PO₄ 1.5 g, K₂HPO₄·3H₂O 2.9 g, triammonium citrate 0.5 g, MgCl₂·6H₂O 0.2 g, CaCl₂·2H₂O 75 mg, and MnSO₄ 0.05 g. The medium was boiled for 5 min and then flushed with nitrogen gas before L-cysteine hydrochloride monohydrate crystals were added into the medium to a final concentration of 0.5 g, as described previously [22]. Unless otherwise stated, corncob (10 g/L final), calcium carbonate (2.0 g/L final), and KHCO₃ (6.5 g/L) were weighed and added in anaerobic tubes, where fifteen milliliters of medium was dispensed under a nitrogen gas flow. Finally, the medium was autoclaved at 121 °C for 20 min. R. cellulolyticum, Lactococcus sp. X1, Lb. pentosus, B. cellulosilyticus, and P. lautus were allowed to grow individually at 35 °C to the late logarithmic phase in the aforementioned medium, except for the carbon source, which was replaced with 5 g/L of cellobiose. Subsequently, these cultures were selectively mixed according to

Table 1. Constitution of the microbial consortia.

Microbial Consortia	Microbial/Enzyme Constitutions
A	R. cellulolyticum alone
B	R. cellulolyticum + Lactococcus sp. X1
C	Consortium B + Lb. pentosus
D	Consortium C + B. cellulosilyticus
E	Consortium C + P. lautus
F	Consortium D + P. lautus
Cs	Consortium C + CTec2
Fs	Consortium F + CTec2
Gs	Lactococcus sp. X1 + Lb. pentosus + CTec2

to establish different microbial consortia and inoculated into the corncob medium at ratios (v/v) of 1:10, 1: 100, 1:100, 1:1000, and 1:1000 for R. cellulolyticum, Lactococcus sp. X1, Lb. pentosus, B. cellulosilyticus, and P. lautus, respectively. The high inoculation ratio for R. cellulolyticum was decided according to [26], while the other ratios were the optimized results in our preliminary work. In addition, cellulase CTec 2, which was quantified using a BCA kit (Sangon, Shanghai, China) following the manufacturer’s instructions, was supplemented at 20 mg/g of corncob at the inoculation step where indicated. The corncob culture was grown for 3 days at 35 °C with shaking at 180 rpm, and the conversion reached a plateau stage. Afterwards, samples were collected for chemical determination, structural analysis, and enzyme preparation.

2.3. Bioinformatics Analysis

Genomic protein sequences were downloaded from the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/, accessed on 1 June 2024; accession numbers: CP001348.1 for R. cellulolyticum DSM 5810, GCA_028539415.1 for B. cellulosilyticus DSM 14838, and GCA_004000945.1 for P. lautus NBRC 15380), and the CAZymes were identified using dbCAN2 [27]. The distribution of major CAZymes which were associated with polysaccharide decomposition among the three bacterial genomes was subsequently analyzed using an online tool for Venn diagrams [28].

2.4. Chemical Analysis

The production of the fermentation products in the culture was determined through high-performance liquid chromatography (HPLC) with a refractive index detector and a Bio-Rad HPX-87H column (Bio-Rad, Hercules, CA, USA), as described previously [29]. Analytical pure chemicals of lithium lactate, sodium acetate, and ethanol were used as standards. The total mole numbers of carbon in the products and substrate polysaccharide were compared for the calculation of yields (the molecular weights of glucan and xylan units were estimated as 162 and 132 g/mol, respectively).

2.5. Enzymatic Assay

Different types of substrates may induce distinct enzyme categories. At the end of fermentation, using 10 g/L of either corncob or wheat bran as the carbon sources, the cultures were centrifuged at 10,000 rpm for 2 min to collect the crude enzymes in the supernatant. CMC, xylan, pectin, arabinan, and ball-milled corncob (400 rpm and 3 h) were used as substrates for various enzymatic assays, and each supplemented at a final concentration of 5 g/L. A standard reaction system contains 50 μL of 10% (w/v) substrate, 30 μL of 100 mM pH 6.0 phosphate buffer, and 20 μL of crude enzymes. The hydrolysis reactions were carried out at 37 °C with shaking for 1 h, except for the corncob reaction, which was prolonged to 12 h and supplemented with an antibiotic, tetracycline (10 μg/mL final), to prevent microbial growth. The CMCase, xylanase, pectinase, arabinanase, and corncob activities were determined using the 3,5-dinitrosalicylic acid (DNS) method [30]. Each sugar unit of the polysaccharides was used for standard curve preparation as well as the quantification of the production of reducing sugars. One unit of enzyme activity was defined as the production of 1 micromole of reducing sugars within 1 min.

2.6. Structure and Composition Analysis of Biomass

The composition of the corncob was determined according to the method of the National Renewable Energy Laboratory, as described before [31]. The corncob was composed of 26.7% cellulose, 28.9% hemicellulose, 17.2% lignin, and 10.5% moisture. The corncob powders before and after microbial degradation were washed with pure water, dried at 60 °C and subjected to scanning electron microscopy (SEM, Regulus 8100, Hitachi, Tokyo, Japan) and attenuated total reflectance–Fourier transform infrared spectroscopy (ATR–FTIR, IS50, Thermo, Sunnyvale, CA, USA), as described previously [32]. In addition, the wheat bran used for enzyme preparation was composed of 24.8% glucan (starch and cellulose), 18.0% hemicellulose, 13.4% lignin, and 10.8% moisture, as determined using the method described above.

2.7. Quantitative Polymerase Chain Reaction (qPCR) Analysis

At the end of fermentation, each culture was centrifuged at 10,000 rpm for 2 min to collect the microbial biomass, which was subsequently disrupted using a bacterial genomic DNA isolation kit (Sangon), following the manufacturer’s instructions. In order to achieve quantitative DNA isolation, DNA was isolated from an identical volume of culture of various consortia, and eluted with the same volume of elution buffer at the final elution step [33]. The genomic DNA of each bacterial strain was also individually extracted, quantified, and employed for standard curve plotting against the qPCR Ct values. Thus, the content of genomic DNA for each strain in a consortium at each stage could be quantified according to their qPCR Ct values. The proportion of a specific bacterium in a consortium was calculated according to its percentage content of genomic DNA. The qPCR was carried out using the LightCycler^® 96 System (Roche, Besel, Switzerland) and a SYBR^® green-based qPCR kit (Sangon) with the primers listed in Table S2.

2.8. Statistical Analysis

All the values were averages of three replicates. The data were analyzed using t-tests and one-way ANOVAs (SPSS 13.0, IBM, New York, NY, USA).

3. Results

3.1. Chemical Productions

All of the microbial consortia mainly secreted lactic acid, acetic acid, and ethanol during the anaerobic degradation of the corncob (Figure 1). R. cellulolyticum mainly produced acetic acid from the corncob; by comparison, the involvement of two kinds of lactic acid bacteria in consortia B and C significantly (p < 0.01) enhanced the lactic acid production (from 0.74 to 1.04 and 2.23 g/L, respectively), the latter of which was three times as large as the production of R. cellulolyticum alone. After the addition of Lb. pentosus, the production of lactic acid increased significantly (p < 0.01), suggesting a more thorough conversion of hemicellulose (pentoses). Meanwhile, the additional participation of B. cellulosilyticus or P. lautus (in consortia D and E) continuously increased the production of lactic acid by 19.7% to a very close value (2.66 and 2.67 g/L, respectively). However, the simultaneous addition of B. cellulosilyticus and P. lautus (consortium F) led to a decrease in lactic acid production (2.24 g/L). Moreover, the involvement of more bacteria slightly increased the acetic acid production, but decreased the ethanol production to a minor extent, except for consortium B. Based on the mole numbers of carbon, the maximal polysaccharide conversion could reach 67.6% using consortium D, which was very closely followed by consortium E (67.2%). It is necessary to mention that some gases also accumulated during the fermentation process but were not determined.

3.2. Enzymatic Activities

In order to determine the hydrolysis functions of these consortia, crude enzymes from the cultures were retrieved for enzymatic activities on an array of substrates, which reflected the total activities on the main chains of cellulose (CMCase), xylan, pectin, and arabinan, as well as the total activity on a real lignocellulose matter, corncob. Meanwhile, the performance of wheat-bran-derived crude enzymes was compared with that of corncob-derived ones. Generally, enzymes from consortium F showed high activities toward most substrates, such as CMC, xylan, pectin, and arabinan (up to 0.24, 0.62, 0.32, and 0.082 U/mL), suggesting the beneficial effects of the involvement of more “degraders” (Figure 2). The xylanase as well as pectinase activities of consortia D and E also increased, as compared to those of consortia A and C, demonstrating the unneglectable role of the involvement of more “CAZyme-rich degraders”. Notably, consortium D showed higher xylanase as well as pectinase activities when growing on wheat bran rather than corncob, probably attributable to the special contribution of B. cellulosilyticus. However, the crude enzymes of consortium A (R. cellulolyticum alone) showed the highest activity toward ball-milled corncob (up to 0.046 U/mL), regardless of the kind of carbon source used for enzyme induction (corncob or wheat bran), suggesting the importance of R. cellulolyticum for corncob saccharification.

3.3. SEM Investigation of the Corncob Residuals

The SEM investigation showed clear variations in the structures and surfaces of corncob powders before and after degradation (Figure 3). As compared to the control (Figure 3A), R. cellulolyticum alone could attack and loosen the pore structure (Figure 3B). Consortia C and D were also able to destroy the corncob structure to some extent (Figure 3C,D, respectively). Interestingly, consortium E resulted in a rough surface of biomass (Figure 3E). Surprisingly, consortium F led to not only a rough surface but also a collapse in the tridimensional structure of the corncob (Figure 3F), suggesting the crucial role of the simultaneous presence of B. cellulosilyticus and P. lautus in the microbial consortium.

3.4. FTIR Analysis of the Corncob Residuals

FTIR analysis showed compositional changes before and after microbial treatments by a series of microbial consortia (Figure 4 and Table S1). The bands between 1000 and 1200 cm⁻¹ are typically related to the C−O−C and C−O−H vibrations in cellulose and hemicellulose. The peak at 1376 cm⁻¹ refers to the bending vibration of C−H₂ and C−H in cellulose and hemicellulose. Clearly, these peaks decreased with the treatments as compared to the control, especially those with consortia D and E, suggesting the substantial utilization of polysaccharides by the microbial consortia. However, the intensity shifts of those peaks with consortium F were quite close to those of consortium C, but less pronounced than those of consortia D and E. A similar decrease pattern was also detected in the peaks ranging from 3300 to 3500 cm⁻¹, which correspond to O−H stretching in cellulose and hemicellulose, as well as lignin. This was also consistent with the change in the treatments of bands 2970 and 2880 cm⁻¹, which were related to the C−H stretching in polysaccharides and lignin. By contrast, the peaks at 1510 cm⁻¹, which correspond to aromatic skeletal vibration in lignin, remained uninfluenced.

3.5. Dynamics of the Microbial Populations

In order to find out the reasons for the inferior performances of consortium F compared to those of consortia D and E in biomass degradation, the dynamics of the microbial consortia were determined using qPCR. During the three days of fermentation, the structure of consortium F changed remarkably as compared to those of consortia C, D, and E (Figure 5). At the end of fermentation, B. cellulosilyticus and P. lautus together accounted for approximately 25% of the total microbes, and the cages of both R. cellulolyticum and Lactococcus sp. X1 were severely compressed, whereas in consortia D and E, B. cellulosilyticus and P. lautus each accounted for less than 3% during the fermentation process. In addition, after 48 h of incubation, the content of the main cellulose degrader, R. cellulolyticum, tended to decrease, and the presence of B. cellulosilyticus or P. lautus would enhance this process (Figure S1). Particularly, B. cellulosilyticus and P. lautus together led to approximately a 60% loss in the content of R. cellulolyticum in consortium F, as compared to that of consortium A. Moreover, B. cellulosilyticus and P. lautus also affected the growth of the two LAB in the manner of firstly enhancing (0 to 24 h) and then weakening it (after 24 h; Figure S1). By contrast, B. cellulosilyticus and P. lautus kept growing during the fermentation process, and they significantly stimulated the growth of each other in consortium F (p < 0.01; Figure S1).

3.6. Enhancement of Degradation by Fungal Cellulases

In order to further enhance the corncob conversion, cellulase CTec2, which is rich in fungal GH6 and GH7 cellobiohydrolases, was supplemented into the microbial consortia as an additional source of CAZymes. As shown in Figure 6, without the major CAZyme suppliers, the consortium of the two LAB only produced 0.64 g/L of lactic acid and 0.43 g/L of acetic acid in the presence of 20 mg CTec2 per gram corncob (consortium Gs), as opposed to 2.23 g/L of lactic acid and 1.62 g/L of acetic acid with consortium C and slightly higher productions with consortium F. With the aid of CTec2, the lactic acid production of consortia Cs and Fs significantly (p < 0.01) increased from approximately 2.23 g/L to 3.18 and 3.36 g/L, respectively, while the acetic acid production of consortia Cs and Fs slightly decreased from approximately 1.63 g/L to 1.30 and 1.44 g/L, respectively. In addition, the ethanol production remained at relatively low levels across all the consortia (less than 0.26 g/L). Finally, the maximal polysaccharide conversion by consortia Cs and Fs reached 75.8% and 80.6%, respectively.

4. Discussion

It is well-known that the construction of a microbial consortium is an effective measure of efficient biodegradation and bioconversion through the strategy of “division of labor”. Plenty of studies have shown that a specific “metabolizer” was able to enhance the performance of a “degrader” [14,34]. However, when confronting an extremely complex substrate such as lignocellulose, knowing how to continuously enhance the degradation capability of a microbial consortium remains a question. In addition, it is interesting to explore the potential of microbial consortia to effect direct bioconversion of minimally pretreated lignocellulose. Our findings validated the importance of involving a CAZyme-rich “degrader” in the conversion of minimally pretreated corncob, revealed the potential risk of microbial interaction, and might provide some clues for the construction of functional intensified microbial consortia to effect efficient biomass conversion.

Diverse CAZymes are crucial for the efficient cleavage of various chemical bonds within lignocellulose. According the CAZyme database (http://www.cazy.org/, accessed on 1 June 2024), there are at least 189 glycoside hydrolase families, 43 polysaccharide lyase families, 20 carbohydrate esterase families, and 17 families for auxiliary enzymes, all of which are closely related to lignocellulose decomposition [35]. Each family of these CAZymes showed a unique pattern of activities toward special kinds of substrates [36]. Synergistic effects have been found among different CAZyme families [35,37]. However, the numbers of CAZyme families and genes within a single microbial genome are usually quite limited; the establishment of a CAZyme-enriched microbial consortium can be a useful solution. As expected, the xylanase and pectinase activities increased substantially in the crude enzymes of consortia D and E, as opposed to consortia A and C (Figure 2). This is consistent with the Venn diagram analysis, that as many as sixty-five and sixty-two additional CAZyme families were introduced in consortia D and E, respectively, as compared to consortium C (Figure S2). In particular, the crude enzymes of consortium F, which contained the most diverse CAZyme genes among the five consortia, showed the highest number of activities toward up to four out of the five kinds of substrates, suggesting the power of a diverse CAZyme repertoire. Our results are also consistent with the findings that the enzyme activities on rice straw were positively correlated with the diversity of heterogeneously expressed glycoside hydrolases from Hungateiclostridium thermocellum [38].

The involvement of CAZyme-rich microbes can thus be beneficial for the improvement of degradation and conversion of lignocellulose, on the condition that the microbial population structure is not significantly shifted. The significantly higher chemical productions of consortia D and E than those of consortium C could be partially attributed to the involvement of new microbes, which brought in additional CAZymes (Figure 1 and Figure S2). However, consortium F, which had more diverse CAZymes than those of consortia D and E, did not produce more chemicals than the latter, probably because of the overgrowth of B. cellulosilyticus and P. lautus, which competed with R. cellulolyticum for the carbon sources, sped up the lysis or sporulation of this major cellulose degrader after 48 h of incubation (only accounting for 27% at the end of the fermentation; Figure 5 and Figure S1), and in turn decreased the lignocellulose utilization. The fact that the lysis and sporulation of R. cellulolyticum occur on condition of starvation has been reported previously [39]. A decrease in biomass conversion was also found to be correlated with a drop in the content of a cellulose degrader, R. thermocellum, in another consortium [19]. From this point of view, maintenance of the dominant position of the major cellulose degrader in a consortium is crucial for the efficient decomposition of lignocellulose.

The direct conversion of minimally pretreated lignocellulosic biomass into biochemicals is challenging but can be achieved by intensified microbial consortia. Froese et al. [19] employed one cellulolytic and two hemicellulolytic bacteria to anaerobically degrade 3.6 g/L of raw wheat straw, and the percent conversion polysaccharide achieved 39.1%. Puentes-Téllez and Salles [18] reported up to 50.6% degradation of 1% sugarcane straw by a series of minimally active microbial consortia under aerobic conditions. Targeted solely on the hemicellulose conversion, Lu et al. [40] achieved 12.51 g/L succinic acid as well as some other acids from 80 g/L minimally pretreated corncob using a bacterial consortium. Our yields with consortium D and consortium Fs (67.6% and 80.6% polysaccharide conversion, respectively (Figure 1 and Figure 6), were much higher than the above-mentioned findings, which also describe the conversion of minimally pretreated biomass using various bacterial consortia, suggesting the power of diverse CAZymes over the efficiency of microbial consortia in lignocellulose decomposition. The yields of consortium D and consortium Fs were also much higher than that of consortium Gs, which only used the CTec2 cellulase to decompose the corncob, suggesting the essential roles of these CAZyme-rich “degraders”, and a synergistic relationship between bacterial and fungal cellulases. Moreover, the degradation pattern of fungal-derived CTec2 is distinct from those of bacterial cellulases, and it has been well applied in the SSF of pretreated biomass into various biochemicals [41], while our results suggest the potential application of this enzyme in enhancing the performance of microbial consortia during the conversion of minimally pretreated lignocellulosic biomass.

In addition, microbial interactions are crucial for the construction of a microbial consortium. Generally, in consortium C, a commensalism relationship existed between R. cellulolyticum and the two LAB, because the former decomposed the substrate and supported the growth of the latter, and the latter did not remarkably affect the growth of the former (Figure S1). The subsequent involvement of a new member (consortium D or E) provided fresh CAZymes, which in turn successfully enhanced the substrate utilization, but it also competed with R. cellulolyticum as well as the two LAB for carbon sources. Interestingly, the simultaneous addition of B. cellulosilyticus and P. lautus (consortium F) greatly stimulated their own growth, suggesting a strong mutualism relationship between them (Figure S1). Both of these two strains were isolated from the human gut, and they might have evolved a special mutualistic mechanism that could be studied in the future. For example, the cross-feeding phenomenon has been discovered in gut bacteria, such as between Bacteroides and Bifidobacterium [42], and between Bacteroides and Methanobrevibacter [43]. Future work might focus on the screening of alternative CAZyme providers, which do not significantly affect the composition of the lignocellulose-degrading microbial consortium.

5. Conclusions

Despite the strong recalcitrance of minimally pretreated lignocellulose, the strategy of involving additional CAZyme families in a lignocellulose-degrading microbial consortium could substantially enhance the substrate conversion and partially increase several kinds of hydrolysis activities. The further addition of a fungal cellulase preparation could continuously improve the substrate conversion. Nevertheless, the continuous addition of some hydrolyzers increases the risk of microbial interactions, which might disrupt the structure of the microbial consortium and hamper its performance. Hence, strict microbial interaction studies are suggested for the construction of complex microbial consortia. Our work might give a guideline for the construction of CAZyme-rich microbial consortia for better lignocellulose utilization.