Isolation of Cellulolytic Gut Bacteria
The morphological characteristics of bacterial isolates 1 to 7 from Leucopholis coneophora and 8 to 17 from Holotrichia serrata are provided in the Supplementary Data (Table S1). In total, seventeen bacterial isolates were obtained from the fermentation chambers (Figure 2) of two white grub species collected from distinct regions. Seven isolates were derived from L. coneophora, while ten were obtained from H. serrata (Fig. 1).
Molecular characterization of cellulolytic gut bacteria
DNA was extracted from all the samples, and 16S rRNA gene amplification using universal 16S rRNA primers produced high-quality amplicons, with sizes ranging from approximately 1450 bp (Figures S1 and S2). These PCR products were subsequently sequenced, and the resulting sequences were aligned using the BioEdit tool. The sequences were subsequently compared to those in the GenBank database, which was accessed through the National Center for Biotechnology Information (NCBI) website.
Comparative analysis revealed that five of the seven bacterial isolates from Leucopholis coneophora exhibited more than 95% similarity to known sequences, while one isolate exhibited 90% similarity (Table S2). For the bacterial isolates from Holotrichia serrata, eight had more than 95% similarity, and the remaining two had more than 90% similarity (Table S3).
Phylogenetic Tree Analysis
Phylogenetic analysis of the 16S rRNA sequences revealed that the bacteria colonizing the gut of Leucopholis coneophora formed two distinct major clades with diverse subgroups. The dominant phyla in the phylogenetic tree were Firmicutes and γ-Proteobacteria, constituting the major clades. The Firmicutes clade was represented by a single genus, Bacillus, with three species and various subbranches, possibly indicating different strains. The second major clade consisted of γ-proteobacteria, represented by Klebsiella sp. and Citrobacter farmeri. The major clades of the phylogenetic tree were analyzed using 1000 bootstrap replications, with Acidobacterium capsulatum serving as the outgroup. The sequence for the outgroup was retrieved from the NCBI database. A phylogenetic tree of the seven identified gut bacteria in L. coneophora is shown in Fig. 3.
A similar phylogenetic tree was constructed for the gut bacteria of Holotrichia serrata, revealing two distinct clades representing different groups: γ-proteobacteria and β-proteobacteria. The major γ-proteobacteria clade comprised the genera Pseudomonas, Citrobacter, Enterobacter, and Acinetobacter. The minor β-proteobacteria clade contained only the genus Achromobacter. This tree was also constructed with an outgroup, and the sequence was retrieved from the NCBI database. Both clades of the phylogenetic tree, along with the outgroup (Aquifex aeolicus), are depicted in Fig. 3, with bootstrap values based on 1000 replicates.
Overall, the taxonomic classification of gut bacteria from L. coneophora and H. serrata provides valuable insights into the diversity of bacterial species present in these insects (Tables S4 & S5).
Cellulolytic indices of the bacterial isolates
The cellulolytic activity of the bacterial isolates was evaluated using the Congo red overlay method, where the presence of a halo zone on CMC agar plates (Figure 4B) indicated the cellulolytic index. Of the seventeen bacterial isolates tested, thirteen exhibited significant cellulolytic activity. The cellulolytic index ranged from a maximum of 1.93 ± 0.037 for the LC3B1 isolate (Bacillus toyonensis) to a minimum of 0.24 ± 0.015 for the LC2B3 isolate (Klebsiella oxytoca) from Leucopholis coneophora. Among the Holotrichia serrata strains, the highest cellulolytic index was 1.49 ± 0.04 for the H4B3 isolate (Enterobacter sp.), while the lowest index was 0.26 ± 0.012 for H4B2 (Enterobacter ludwigii). The cellulolytic indices of the gut bacteria isolates are presented in Fig. 4A (Table S6). The cellulolytic indices varied significantly among the different bacterial strains, as indicated by the ANOVA results: F(2,12)=661.1, p<2×10−. Duncan’s test further confirmed these differences, revealing significant variation in the cellulolytic indices between the strains. Strain LC3B1 exhibited the highest cellulolytic index, followed by strain LC1B1. Based on these findings, four bacterial isolates—LC1B1, LC3B1, H4B3, and H5B2—were selected for further study due to their promising cellulolytic activities (Fig. 5).
Substrate Degradation Ratio of Agricultural Residues
The four selected bacterial isolates were cultured in media containing powdered corncob, paddy straw, and groundnut husk substrates, followed by incubation for eight days. Among the isolates, H5B2 exhibited the highest degradation efficiency, reaching 48.15 ± 1.56% for groundnut husk, whereas LC3B1 exhibited 46 ± 0.608% for paddy straw powder and 43±0.27% for corncob. The substrate degradation ratio was significantly different between the F(2, 3)= 12.236 strain (p <0.05) and the F(2,2)=8.050 strain (p<0.05) but not between substrates. The other isolates also demonstrated significant degradation capabilities, as shown in Fig. 5 (Table S7).
Cellulase (β-1,4-Endoglucanase) Enzyme Assay
Cellulase (β-1,4-endoglucanase) activity assays were conducted using the DNS method at 50°C and pH 4. β-1,4-endoglucanase was statistically significant among the various substrates F(2, 2)= 7.3, p <0.024 but not between strains of bacteria F(2,3)=1.16, p=0.39). Among the tested genes, LC3B1 had the highest beta 1,4-endoglucanase activity on ground nut husk, followed by LC1B1, which had the second highest endoglucanase activity (3.614 ± 0.02 U/mL), particularly on the corn cob powder substrate. LC1B1 also exhibited significant activity on groundnut husk (2.168 ± 0.07 U/mL), highlighting its potential for enzyme production. Isolates H4B3 (2.4 ± 0.03 U/mL) and H5B2 (1.626 ± 0.01 U/mL) also exhibited notable endoglucanase activities on groundnut husk. Additionally, LC1B1 exhibited activity (1.628 ± 0.01 U/mL) on paddy straw, whereas the remaining three bacteria displayed activities less than 0.7 U/mL on all three substrates. Moreover, LC3B1 exhibited considerable endoglucanase activity (2.082 ± 0.02 U/mL) on corncob.
These findings underscore the substrate-specific cellulase activities of the bacterial isolates, with LC3B1 showing significant potential for enzyme production, particularly on groundnut husk and corncob substrates. The results are depicted in Fig. 6 (Table S8).
FTIR and SEM Results for the Breakdown of Cellulosic Bonds in Filter Paper
Structural analysis of the four bacterial isolates was performed using attenuated total reflectance (ATR) mode in the infrared range of 400–4000 cm−1, enabling the detection of bond types and functional groups (Fig. 7). The peak at 1160 cm−1 is attributed to C-O-C stretching in the cellulose/hemicellulose molecules. The peak between 2896 and 2900 cm−1 indicates C-H (β-glycosidic bond) deformation in the polysaccharide molecules. The stretching of carbon and oxygen molecules between 1749 and 1599 cm−1 revealed decomposition products such as ketones, aldehydes, alcohols, and carboxylic acids derived from cellulose and hemicellulose. Additionally, the shifts in the O–H and C–H stretching vibrations at 3333 and 2894 cm−1 can be attributed to changes in the intra- and intermolecular bonds of the cellulosic polymer. The FTIR results revealed broad absorption bands at 897, 1651, and 2900 cm−1, corresponding to COC, CCO, OCH deformation, C-5, C-6 motion stretching, OH bending, and CH stretching, respectively. The intense peak at 3000–3700 cm−1 is attributed to the stretching of OH functional groups and amine groups.
Field emission scanning electron microscopy (FESEM) provided insights into the significant structural changes in the surface morphology of both the control and treated filter paper (FP) samples (Fig. 8). Compared to those of the control substrates, the treated FP surface exhibited a rough appearance, likely due to bacterial adherence and superficial hydrolysis of the biomass. The bacterial cells adhered to the substrate, potentially creating pores and tunnels that allowed access to the internal fibers of the cellulosic biomass. This process suggested that the long cellulose chains in the FPs released their microfibers. The electron micrographs further demonstrated substrate disruption, as evidenced by the accumulation of hydrolyzed biomass on the surface.