Bacteria with CMCase activity were abundant (1.14 ± 0.13 × 10⁸ colony forming units (CFU)/gut) in the hindgut of H. parallela. However, bacteria with CMCase activity were seldom found in the midgut (only 20 ± 1.45 CFU/gut). These results are similar to those from another scarab beetle, P. marginata [30]. Cazemier et al. [30] observed that a large number of bacteria with CMCase and xylanase activities were present in the hindgut of P. marginata (2.5 ± 1.1 × 10⁸ CFU (mL_gut) ⁻¹), but that these bacteria were not detected in the midgut. Studies of the gut microbiota of other scarab beetles showed that the hindgut microbiota was dominated by groups of fermentative bacteria such as Clostridiales, Actinobacteria, and Cytophaga-Flavobacterium-Bacteroides, which contain a wide range of species able to ferment cellulose, hemicellulose, pectin and other polysaccharides [34,38–40]. These results indicate that the bacteria in the scarab hindgut play an important role in the degradation of roots and other organic matter consumed by scarab larvae, as suggested by Cazemier et al. [30] and Huang et al. [31]. As bacteria with cellulolytic activity appear to be absent in the midgut, it seems likely that the midgut of scarabs serves a predigestive function for lignocellulose rather than for the microbial degradation of cellulose and hemicellulose [30].

A total of 207 isolates with CMCase activity were obtained from the gut contents of H. parallela either by plating on CMC medium or by enrichment on filter paper. Among these cellulolytic bacteria, 81 isolates were obtained using the filter paper inoculation method, and 126 isolates were obtained from direct plating. These isolates produced variable zones of CMC clearance (Figure 1). Based on the calculation of the ratio of the diameter (mm) of the zone of clearance to the diameter of the colony, it was determined that these bacterial isolates demonstrated large differences in their ability to degrade CMC (Figure 2). This ratio ranged from 1.1 to 9.0 among all the isolates, with 24.1% of the isolates showing high CMC-degrading activity (ratio > 5), demonstrating that multiple bacterial isolates from the scarab gut possess the ability to produce CMCase (Figure 2).

The 207 cellulolytic bacterial isolates obtained in this study were grouped into 21 clusters or genotypic groups (Table 1). Each group displayed a specific ARDRA banding pattern, and the number of isolates belonging to each group was different (Table 1). A total of 21 isolates were chosen to represent each ARDRA group, and these isolates were investigated both by 16S rDNA sequencing and by physical and biochemical characterization (Table 2). Overall, the 16S rDNA sequences from the 21 isolates showed a high degree of similarity (99–100%) to a number of annotated sequences found in the databases (data not shown), and their identification was in agreement with the biochemical and physiological tests. The 21 isolates clustered into four phyla (Proteobacteria, Actinobacteria, Bacteroidetes, and Firmicutes), and represented 17 different genera (Table 1). The cellulolytic bacterial community was represented by members of the phylum Proteobacteria (67.13%), followed by Actinobacteria (23.15%), Firmicutes (4.35%), and Bacteroidetes (1.45%). At genus level, Pseudomonas (31.4%), Cellulosimicrobium (13.53%), Ochrobactrum (12.08%), Rhizobium (11.59%), and Microbacterium (10.63%) were the dominant genera identified, with 65, 28, 25, 24 and 22 isolates, respectively, while Siphonobacter (group 3), Devosia (group 14), Variovorax (group17), Shinella (group 18) each consisted of a single bacteria isolate. Furthermore, the ARDRA grouping results also revealed that bacterial isolates belonging to Bacillus licheniformis, Microbacterium oxydans, Microbacterium binotii, Microbacterium aurum., Cellulosimicrobium funkei, Ochrobactrum cytisi, Rhizobium radiobacter, Labrys neptuniae, Pseudomonas nitroreducens, Stenotrophomonas maltophilia can obtained both by the direct plating method and by the filter papers enrichment method. The fact that these bacterial isolates can be obtained using both methods demonstrates that bacteria with cellulolytic ability are commonly present in the hindgut of H. parallela.

Our results showed that Pseudomonas was the most dominant group in the cellulolytic bacterial community in the gut of soil-dwelling scarab larvae. The dominance of Pseudomonas in the present study is similar to the results of previous studies on cellulolytic bacteria present on native Chaco soil, which showed that the Pseudomonas was the only genus that exists stably in three samples (native forest soil, CMC- and filter paper-enriched samples) [41]. Bacteria of the genus Pseudomonas can be found in many different environments including soil, water, plant and animal tissue, and these bacteria have the ability to metabolize a variety of diverse nutrients [42]. Many Pseudomonas species are opportunistic pathogens that infect humans, animals, and plants [43–45], but other Pseudomonas species also have been reported to degrade cellulose [46–48]. There have been no reports, however, describing the cellulolytic activity of P. nitroreducens, which we observed in this study.

The cellulolytic activity of some of the bacteria found in this study has been reported previously. B. licheniformis is characterized by strong xylanase activity, and also possesses CMCase, mannanase, and pectinase activities [49]. Though Dyadobacter fermentans NS114^T does not hydrolyze cellulose or starch [50], whole genome sequencing of D. fermentans DSM 18053 has revealed several genes encoding for 1,4-β-cellobiosidase, β-glucosidase, and endo-1,4-β-xylanase enzymes [51]. The Microbacterium genus contains many species with cellulolytic or xylanolytic activities. A cellulolytic bacterium that showed 99% 16S rDNA sequence similarity to M. oxydans has been found to produce an array of cellulolytic-xylanolytic enzymes (filter paper cellulase, β-glucosidase, xylanase, and β-xylosidase) [52]. M. binotii have also been reported to produce an enzyme with β-glucosidase activity [53]. Rhizobium species are known to produce cellulolytic and pectinolytic enzymes that can break the glycosidic bonds present in the plant cell wall, and these enzymes are essential for the primary symbiotic infection of legume host roots [54–56]. However, little attention has been paid to their potential ability to degrade organic compounds during their growth as free-living saprophytes [41,57]. An analysis of the genome sequence of R. radiobacter (formerly Agrobacterium tumefaciens) has identified several genes encoding pectinase, ligninase, and xylanase as well as genes encoding regulators of pectinase and cellulase production [58]. Variovorax paradoxus is a microorganism of special interest due to its diverse metabolic capabilities. Whole genome sequencing of V. paradoxus revealed a single gene encoding β-glucosidase, but genes involved in the production of pectinases and other cellulases remained unidentified [59]. The Stenotrophomonas genus contains species ranging from common soil organisms (Stenotrophomonas nitritireducens) to opportunistic human pathogens (S. maltophilia) [41]; one S. maltophilia strain from the mesophilic microbial community BYND-8 has also been reported to be cellulolytic [60].

In addition to those bacterial isolates for which cellulolytic activity has been well described, our results demonstrate cellulolytic activity for several bacterial strains that have not been previously reported to be cellulolytic. To the best of our knowledge, this is the first report describing Siphonobacter aquaeclarae, C. funkei, Paracoccus sulfuroxidans, O. cytisi, Ochrobactrum haematophilum, Kaistia adipata, Devosia riboflavina, L. neptuniae, Ensifer adhaerens, Shinella zoogloeoides, Citrobacter freundii, and P. nitroreducens as being cellulolytic, with some isolates displaying high cellulolytic activity. In the case of S. aquaeclarae, the ratio of the CMC clearance zone to the colony diameter was greater than 7, and for C. funkei, the ratio ranged from 3.3 to 5.3, indicating robust CMC-ase production. These cellulolytic bacterial isolates demonstrate great potential for the study of novel enzymes in cellulose degradation and for improving the bioconversion of lignocellulosic biomass.

Third-instar larvae of H. parallela were collected from a peanut field and were maintained individually in containers with sterile soil. All the larvae were fed with peanuts surface sterilized with 70% ethanol, and the diets were replaced every 3 days. After 3 weeks, 9 healthy larvae were surface sterilized with 70% ethanol to remove contamination, washed twice in sterile distilled water, and allowed to air dry for 1 min. The preparation of the intestinal tract (mid- and hindgut) was performed as described previously by Zhang and Jackson [34].

Medium I and Medium II were prepared as described by Cazemier, et al. [30], with some modifications, as follows:

Medium I: peptone, 5 g/L; yeast extract, 0.1 g/L; K₂HPO₄, 1 g/L; MgSO₄·7H₂O, 0.2 g/L; carboxymethylcellulose (CMC), 10 g/L (sodium salt, low viscosity, Sigma); Na₂CO₃, 10 g/L (sterilized separately); pH 10.3.

Medium II: K₂HPO₄, 1.9 g/L; KH₂PO₄, 0.94 g/L; KCl, 1.6 g/L; NaCl, 1.43 g/L; NH₄Cl, 0.15 g/L; MgSO₄·7H₂O, 0.037 g/L; CaCl₂·2H₂O, 0.017 g/L; yeast extract, 0.1 g/L; CMC, 10 g/L; pH 7.2.

Medium III was prepared as described by Wenzel et al. [25], with the following modifications: yeast extract, 0.04 g/L; malt extract, 0.1 g/L; CaCO₃, 0.5 g/L; filter paper strips, 5 g/L (Whatman Filter Paper No.1); pH 10.3.

The media were sterilized (121 °C, 20 min) and solidified with agar (17 g/L) when necessary.

For viable counts, an individual gut was homogenized and suspended in 10 mL of medium I and serially diluted ten-fold (to 10⁻⁹). From each dilution, 100 μL was spread on plates with solid medium I (midgut) or medium II (hindgut). A triplicate series of dilutions from the midguts and hindguts of three different larvae were incubated at 28 °C. Colonies were counted following 4 weeks of incubation. Only the colonies that were encircled by a clear zone after staining with a solution of congo red (1 mg/mL) were counted.

To isolate cellulolytic bacteria, the midgut or hindgut sections from six individual larvae were pooled and homogenized. Serial dilutions and plating were performed as described above, and the plates were incubated aerobically at 28 °C for up to 4 weeks. In addition to directly plating the gut samples on solid media, 0.5 mL of the homogenized midgut or hindgut suspension was inoculated into 100 mL of medium III and incubated at 28 °C. After 3 weeks of enrichment, 100 μL of the growing cultures were cultivated on solid medium I (midgut) or medium II (hindgut). Bacteria from single colonies were repeatedly grown on solid agar plates until a pure culture was obtained.

CMC degradation by the isolates was tested on solid medium II by covering the Petri dishes with congo red dye, as described by Teather and Wood [61]. Carboxymethylcellulose degradation was indicated by a clear zone around the colonies. Enzyme activity was indexed as the diameter of the colony plus the surrounding clear zone divided by the diameter of the colony [29]. Three measurements were taken from each isolate, and only the isolates that produced a clear zone around the colony were chosen for further study.

Bacterial isolates were grown in LB medium (Tryptone, 10 g/L; yeast extract, 5 g/L; NaCl, 10 g/L; pH 7.0) at 30 °C for 48 h. The cultures were centrifuged at 10,000× g for 1 min, and the supernatant was removed. DNA extraction was performed using a Cell/Tissue Genomic DNA Extraction Kit (BioTeke Corporation, Beijing, China) according to the manufacturer’s instructions, and the genomic DNA was stored at −80 °C until further analysis. Bacterial universal primers 27F (5′-AGAGTTT GATCMTGGCTCAG-3′) and 1492R (5′-TACGGYTACCTTGTTACGACTT-3′) were used to amplify the 16S rDNA from genomic DNA [62]. Polymerase chain reaction (PCR) was performed in a thermocycler (MyCycler, Bio-Rad, USA). Each reaction mixture (50 μL) contained 5 μL of 10× reaction buffer without MgCl₂, 1.5 mM MgCl₂, 0.2 μM of each primer, 0.2 mM of each dNTP, 2.5 U of Taq DNA polymerase (TaKaRa Biotechnology (Dalian) Co., Ltd., China), and 25 ng of template DNA. The amplification was performed as follows: initial denaturation for 5 min at 94 °C, 35 cycles each of denaturation for 30 s at 94 °C, annealing for 30 s at 55 °C, and primer extension for 1.5 min at 72 °C, and a final extension for 10 min at 72 °C. The PCR products were checked by gel electrophoresis in 1.2% (w/v) agarose gels stained with ethidium bromide (10 mg/mL) and cleaned using an EasyPure Quick Gel Extraction Kit (Transgen Biotech, China) according to the manufacturer’s instructions.

Amplified ribosomal DNA restriction analysis (ARDRA) was performed on the PCR-amplified 16S rDNA products from each of the isolates using three specific restriction enzymes: HhaI, AfaI, and MspI (TaKaRa Biotechnology (Dalian) Co., Ltd., China). Five microliters of each PCR product was digested for 2 h at 37 °C with 1.5 U of each restriction endonuclease. Aliquots (5 μL) of each digested product were analyzed by gel electrophoresis in an 8% nondenaturing acrylamide gel (acrylamide: N,N′-Methylenebisacrylamide, 29:1) [63] and by silver nitrate staining, as described previously [64]. Fragment sizes were estimated using a low range, 50 bp DNA ladder (Dongsheng Biotech Co., Ltd., China), and a final grouping of isolates was performed by a visual comparison of the restriction patterns. For each distinct ARDRA group, one bacterial isolate was selected for sequencing and standard physical and biochemical characterization.

Nearly full-length bacterial 16S rDNA fragments were amplified by PCR from each representative isolate using the universal primers 27F and 1492R, as described above. The PCR products were cleaned and cloned using the pEASY-T1 cloning kit (Transgen Biotech, China) with blue-white screening. The clones containing inserts of the correct size were sequenced, and the sequences were aligned against those found in the NCBI database [65], in the RDP II database [66], and on the EzTaxon server [67] using the BLAST (Basic Local Alignment and Search Tool) algorithm [68]. All the sequences have been submitted to the GenBank database under the accession numbers JQ291585-JQ291605.

For each ARDRA group, one representative isolate was identified based on standard physical and biochemical tests [69], including motility, Gram staining, the methyl red (MR) test, the Voges-Proskauer (VP) test, the activities of catalase, oxidase, urease, and arginine dihydrolase, tests for nitrate reduction, the production of indole, the utilization of citrate, and acid and gas production from glucose. Different carbon sources (d-Lactose, d-Glucose, d-Fructose, d-Maltose, Mannose, Xylose, d-Rhamnose, d-Mannitol, and d-Sorbitol) were used to evaluate carbon utilization. Except for the gelatinase activity test (which was performed at 20 °C), all of the tests were performed at 28 °C in the appropriate medium and were conducted according to standard methods [69].