Fungal strains were directly isolated by the spread plate method from anti-biodegradable residue which was piled up for 2–5 years at Guizhou Shuangsheng Company (Guiyang, China). Glycine, 2,2′-azino-bis (3-ethylbenzothiazoline-6-sulphonic acid; ABTS), Avicel, Briffon-Robinson 1buffer (pH 4.5), Birchwood, pNP-β-D-glucopyranose (pNP-BGL), pNP-β-D-xylopyranose (pNP-BXL) and other chemicals were purchased from Sigma-Aldrich to determine enzymatic activity; Carboxymethyl cellulose and xylan for solid media were purchased from Hongdaer Biotechnology Company (Guiyang, China) for preliminary screening of fungi.

Fresh anti-biodegradable residue (younger than 2 months) was obtained from Guizhou Shuangsheng Company and ground in a blender (Conair Waring Pulverizer, Fisher Scientific), thereafter, 100 g was sieved through 50 mesh and was placed in a 40°C drying oven for 24 h. After grinding, carbohydrate composition was determined using high-performance liquid chromatography by Shandong Kechuang Quality Testing Company, and the lignin content was determined by a lignin assay kit (Solarbio, China) according to the instructions provided by the manufacturer.

A suitable anti-biodegradable residue tissue block of 10 g was selected and washed with sterile normal saline three times, fungi were isolated by plant tissue separation (Makhuvele et al., 2017; Thi Minh Le et al., 2019), the sample was repeated three times. Genomic DNA of the purified fungi was extracted using Ezup Column Fungal genomic DNA Extraction Kit (Qingke, Chongqing) and used as a template for internal transcribed spacer (ITS) region amplification (Bellemain et al., 2010). Prime pair ITS5 and ITS4 (Mishra et al., 2017) was used to amplify the ITS regions (Thermal Cycler PCR instrument BIO-RAD T100, Japan). The PCR products were sequenced and used for the identification of fungi by nucleotide BLAST (Altschul et al., 1990) against the NCBI databases. The living strains were deposited at Guizhou Medical University. Phylogenetic tree (Supplementary Figure 1) was constructed by using ITS regional sequences of 302 residue isolates and by MEGA10 using the neighbor-joining (NJ) method after multiple alignments of sequences data (Saitou and Nei, 1987). The corrected evolutionary distance was evaluated according to the p-distance model (Nei and Kumar, 2000). In order to estimate the consensus of the branching, the bootstrap resampling analysis of the phylogenetic tree was employed with 1000 replicates of the data set (Felsenstein, 1985). According to the results of phylogenetic analysis, the pie chart was constructed with Adobe Illustrator CS5 software and Office.

The fungi were pre-grown on a PDA agar medium at 25°C for 3–7 days, then mycelium was transferred to three different sole carbon source media. (1) carboxymethyl cellulose Congo red agar medium (CMC-red): 1% sodium CMC, 0.05% KH₂PO₄, 0.05% MgSO₄⋅7H₂O, 1.5% agar, 0.02% Congo red, 0.2% gelatin, pH 6.8–7.2; Gupta et al., 2012; Fen et al., 2014; Wu et al., 2016); (2) xylan Congo red agar medium (Xyl-red): 0.5% xylan, 0.01% Congo red, 0.1% MgSO₄⋅7H₂O, 0.5% yeast extract, 0.1% KH₂PO₄, 1.5% agar, 0.1% (NH₄)₂SO₄, 0.5% NaCl, pH 6.5–7; Liu et al., 2010), and (3) potato dextrose agar containing 0.01% aniline blue (PDA-blue), pH 6–7 (Wang H. et al., 2009). The ratio of the hydrolytic circle (D) to fungal colony diameter (d) was measured after 5 days of constant temperature incubation. Analysis of hydrolytic efficacy was performed by division of the diameter of mycelium growth by that of the clearance zone.

Fungi with high activity were cultured on a PDA plate to produce mycelium and spores. The culture medium with fungus was then placed in a sterile beaker, and spores and mycelia were separated by Tween80. Adding 2 × 10⁶ spores (Li et al., 2016) of each fungus into minimal medium (MM; composed of 4 g/L glucose and 10 g/L peptone), after shaking for 3–7 days at 28°C and a rotational speed of 30× g, mycelia were filtered through a 200-mesh (10 cm diameter) sieve and collected, and fully washed with sterile water. Mycelia weighing 2 g (wet weight) was transferred into a 50 ml culture medium with anti-biodegradable residue (the residue was filtered through a 50-mesh sieve) as the only carbon source (anti-biodegradable residue 3 g, Peptone 0.5 g, Yeast powder 0.25 g, NaCl 0.25 g, KH₂PO₄ 0.05 g, MgSO₄⋅7H₂O 0.025 g, (NH₄)₂SO₄ 0.25 g, CuSO₄ 0.0003 g; Wang J. F. et al., 2009; Cheng et al., 2012; Zhang et al., 2015) at 28°C, 30× g, 2 mL supernatant was taken at the same time on d 3, 5, 7, 9, 11, centrifuged at 4°C, 21,100× g for 10 min, and the clarified supernatant was taken and stored in −80°C for use.

The single filamentous fungi, which can degrade cellulose hemicellulose lignin effectively, were paired by the permutation group (Ming et al., 2019) method, in order to construct fungal combinations that can degrade lignocellulose comprehensively. The specific method is to use a sterile knife to cut the thriving fungal culture medium into 5 mm square pieces, and transfer them to a new PDA culture medium. Each piece of fungus is separated by about 1 cm. Then, the medium is sealed in a 28°C medium for 3–7 days to observe the ability of fungi to grow in co-cultivation conditions.

β-glucosidase and β-xylosidase activities were measured with the substrates para-nitrophenyl β-glucoside (pNP-BGL) and para-nitrophenyl β-xyloside (pNP-BXL), respectively. The reaction contained 40 μl culture supernatant, 10 μl 0.1% pNP-BGL or pNP-BXL, and 50 μl 100 mM sodium acetate, pH 5.0. The reactions were performed overnight (∼16 h) at 25°C followed by the addition of 100 μl 250 mM Na₂CO₃ prior to measurement of the absorbance at 405 nm using a microtiter plate reader (biotek epoch2, US). Para-nitrophenol (0–100 μM) was used as the standard (Patyshakuliyeva et al., 2016). One unit was defined as the amount of enzyme releasing 1 μmol pNP from pNP-BGL or pNP-BXL under the assay condition.

Cellulase and xylanase activities were measured toward avicel cellulose and birchwood xylan, respectively. The reaction contained 20 μl culture supernatant and 180 μl 1% avicel or birch xylan in 50 mM sodium acetate, pH 5.0, and was incubated overnight (∼16 h) at 25°C. Mixing 100 μl of the reaction mixture with 150 μl 3,5-dinitrosalicylic acid solution (1% 3,5-dinitrosalicylic acid, 0.2% phenol, 0.05% Na₂SO₃, and 1% NaOH), incubated at 95°C, 30 min and cooled on ice prior to measurement of the absorbance at 560 nm using a microtiter plate reader (Miller, 1959). 2–20 mM D-glucose or D-xylose was used as standard (Peciulyte et al., 2017).

Peroxidase and laccase activities were measured toward 2,2′-azinobis (3-ethylbenzthiazoline-6-sulfonic acid; ABTS) with or without supplement of hydrogen peroxide, respectively. The peroxidase reaction contained 20 μl culture supernatant, 20 μl 140 μM ABTS, 10 μl 3% H₂O₂, 25 μl 400 mM Britton–Robinson buffer, pH 4.5, and 125 μl ddH₂O. Detection of initial rates of oxidation was performed at 440 nm with a 2 min interval up to 60 min, at 25°C using a microtiter plate reader. The laccase reaction contained 20 μl culture supernatant, 20 μl 140 μM ABTS, 20 μl 500 mM glycine-HCl, pH 3.0, and 140 μl ddH₂O. The initial rates of hydrolysis were determined by measuring absorbance at 440 nm in a 2 min interval for 30 min, at 25°C. Laccase and peroxidase activities were calculated based on Lambert-Beer law, where the extinction coefficient of 3.6× 10⁴ M^–1 cm^–1 was used as mentioned in Childs and Bardsley (1975) and Srinivasan et al. (1995).

The selected co-cultivation was incubated at 28°C for 15 days, and the actual hydrolysis capacity of the residue was tested. The residue was taken from the liquid medium and thoroughly dried before weighing to determine the total weight loss rate of the residue as well as the degradation rates of cellulose, hemicellulose, and lignin. Each strain was repeated three times, and the residue liquid medium without inoculation was used as a negative control. The negative control M₀ was used to calculate the hydrolysis rate of residue and cellulose, instead of residue weight before inoculation of fungi M₁, in order to eliminate the natural loss of residue powder placed in the solution during the experiment and ensure the strictest of the experiment. The formula for calculating the total weight loss rate of drug residue hydrolysis is (M₀-M₂)/M₀, where M₀ represents the dry weight of drug residue after 15 days of culture in a negative control group under the same conditions, and M₂ represents the weight of drug residue after hydrolysis (Ming et al., 2017). Hemicellulose and lignin content in the hydrolyzed residue was detected by the content detection Kit (Solarbio BC4445, Solarbio BC4205) and cellulose content detection kit (ZCIBIO ZC-S0876). The specific operation method is described in the instruction, three biological replicates were performed.

The ITS sequences were used to identify the strains from mega blast results in GenBank. In this preliminary study, we only intended to check whether there are fungal taxa that are inhabiting TCM residues. In Supplementary Table 1, we provide results that were generated in the mega blast. It provides the closest hits which are similar to our selected strains.

The determination of glycosylated components in the residue showed that glucose, accounting for 68.32%, glucose is the reducing sugar of cellulose, indicating that the sample contains a lot of cellulose. Plus, there were xylose, arabinose, mannose, galactose, and other five-carbon sugars, accounting for about 30% (Table 1), and the analysis of lignin content in the residue showed that there was 13.66% lignin. The lignin content test suggests the anti-biodegradable residue contains a moderate amount of lignin in liquid samples. In addition, 50 ml of 6% residue is pH 4.9.

A total of 302 fungi were isolated and labeled as ZYJHYZ01-ZYJHYZ302 (Supplementary Table 2). Phylogenetic analysis showed that they belong to 30 genera in 3 phyla. Among the isolates, 72.19% of strains (218 isolates of 16 genera) were identified as Ascomycota, while 19.87% of strains (60 isolates of 12 genera) were identified as Basidiomycota and 7.95% strains (24 isolates of two genera) were identified as Mucoromycota (Figure 1A). At the genus level, 40.07% of isolates (n = 121) belong to Aspergillus, 7.62% to Coniochaeta (n = 23), 6.29% to Filobasidium (n = 19), 5.96% to Penicillium (n = 18), and 5.30% to Mucor (n = 16; Figure 1B).

Three-hundred and two (302) isolates were screened by inoculating onto a solid medium with a single polymer (CMC, xylan, or glucose) as the sole carbon source and indicator added: CMC-red, PDA-blue, Xyl-red, 25°C, the ratio of the hydrolytic circle (D) to fungal colony diameter (d) was measured after 5 days of constant temperature incubation. From this analysis, we were able to find that 35 strains produced transparent circles while 24 strains showed a ratio greater than or equal to 2 (Supplementary Table 3).

Among them, the D/d ratio of ZYJHYZ163, ZYJHYZ254, ZYJHYZ257, ZYJHYZ265, and ZYJHYZ270 were greater than other strains. According to the mega blast results, Aspergillus fumigatus (ZYJHYZ279) and Aspergillus flavus (ZYJHYZ260) were indicated as harmful to human and animal health, thus, further experimental studies were abandoned. Therefore, fungi with a total score of 2 were preliminarily screened through three kinds of solid media for the next experiment, and a total of 24 strains were identified: ZYJHYZ257, ZYJHYZ259, ZYJHYZ261, ZYJHYZ263, ZYJHYZ265, ZYJHYZ267, ZYJHYZ268, ZYJHYZ269, ZYJHYZ270, ZYJHYZ271, ZYJHYZ272, ZYJHYZ273, ZYJHYZ274, ZYJHYZ53, ZYJHYZ163, ZYJHYZ244, ZYJHYZ242, ZYJHYZ28, ZYJHYZ246, ZYJHYZ254, ZYJHYZ255, ZYJHYZ240, ZYJHYZ247, and ZYJHYZ256 (Figure 2).

All fungi produce different levels of enzymes with different activities. Here, we screened for six different enzymes, viz., β-glucosidase, cellulase, β-xylosidase, xylanase, laccase, and peroxidase. β-glucosidase – In general, most fungi can produce higher β-glucosidase activity on days 7–9 (Figure 3A), among which Fomitopsis sp. (ZYJHYZ247) produces the highest β-glucosidase level on day 7 (151.68 U/L) followed by Nemania sp. (ZYJHYZ263) reached 134.22 U/L on day 9. Cellulase – However, the peak time of production was different among different fungal strains, among which Aspergillus niger (ZYJHYZ242) reached 66.55 U/L on day 7, and Fomitopsis sp. (ZYJHYZ247) reached 64.63 U/L on day 9 (Figure 3B). β-xylosidase – Figure 3C, showed that multiple fungi had the same enzyme production, among which the best enzyme-producing fungus was Coniochaeta sp. (ZYJHYZ246) and Phaeophlebiopsis sp. (ZYJHYZ254). Coniochaeta sp. (ZYJHYZ246) had the highest enzyme production at 80.97 U/L on the eighth day while Phaeophlebiopsis sp. (ZYJHYZ254) had the highest enzyme production at 79.92 U/L on the nineth day. Xylanase – In Figure 3D, Phaeophlebiopsis sp. (ZYJHYZ254) had the highest enzyme production, with the highest enzyme production time reaching 81.59 U/L on the 5th day, Bjerkandera sp. (ZYJHYZ257) and Isaria sp. (ZYJHYZ259) performed higher xylanase activity at day 9 and 7 with 76.14 and 72.98 U/L, respectively, in addition, Talaromyces sp. (ZYJHYZ163) also showed excellent and stable β-xylosidase and xylanase activity. Laccase – From Figure 3E, strain Piloderma sp. (ZYJHYZ268) reached the highest enzyme production (99.39 U/L) on the fifth day, followed by Coniochaeta sp. (ZYJHYZ246) and Cladosporium sp. (ZYJHYZ265) with 99.31 and 99.24 U/L, respectively. Peroxidase – From Figure 3F, including Phaeophlebiopsis sp. (ZYJHYZ254), Piloderma sp. (ZYJHYZ268), Coniochaeta sp. (ZYJHYZ246), Coniochaeta sp. (ZYJHYZ244), Aspergillus sp. (ZYJHYZ242) showed significantly higher enzyme activity than other fungi, with Phaeophlebiopsis sp. (ZYJHYZ254) having the highest enzyme activity, reaching 602.96 U/L on day 7, followed by Coniochaeta sp. (ZYJHYZ246) and Aspergillus sp. (ZYJHYZ242) with 580.42 and 561.06 U/L, respectively.

In conclusion, strains belonging to Coniochaeta, Fomitopsis, Phanerochaete, and Piloderma are capable of producing a higher amount of lignocellulose degrading enzymes compared to other strains.

Through the screening for six fungal extracellular enzymes, the first three fungal strains with the highest activity of each enzyme were selected for further co-cultivation; Aspergillus sp. (ZYJHYZ242), Fomitopsis sp. (ZYJHYZ247), and Nemania sp. (ZYJHYZ263) for cellulose degradation. Talaromyces sp. (ZYJHYZ163), Phaeophlebiopsis sp. (ZYJHYZ254), and Bjerkandera sp. (ZYJHYZ257) for hemicellulose degradation. High lignin-degrading strains numbered as Coniochaeta sp. (ZYJHYZ246), Phaeophlebiopsis sp. (ZYJHYZ254), and Piloderma sp. (ZYJHYZ268), the BLAST results from NCBI as Supplementary Table 3. A total of 27 fungal co-cultivation was constructed by permutation and combination of the above fungi, in order to screen out fungi that can comprehensively degrade lignocellulose. Twelve co-cultivation were growing together through the symbiosis test (Figure 4) and their enzyme activities showed in Figure 5. The other strain combinations were excluded from liquid cultivation.

According to Figure 5, the activities of 5 enzymes (except for lignin peroxidase) have been greatly improved over those strains in single culture conditions. β-glucosidase – contained more than 100 U/L activities in most fungal combinations while only three single-cultured fungi were capable of production at this level. The highest β-glucosidase was produced on day 11 of incubation by fungal co-cultivation no. 1 (313.63 U/L), followed by no. 10 (325.03 U/L; Figure 5A). Cellulase – showed that multiple fungi had the same enzyme production, the cellulase activity performed highest on the seventh after inoculation by fungal combination nos. 8 and 9 with 91.86 and 132.43 U/L, respectively (Figure 5B), the highest enzyme activity of no. 9 reached 132.43 U/L, was about twice that of a single fungus. β-xylosidase – the peak of enzyme production in combinations nos. 5 and 11 was on the third day, and then decreased with time, while the other fungal combinations had an opposite trend, mostly reaching the peak around the ninth day, and β-xylosidase activity performed highest at day 11 and 9 after inoculation by fungal combination nos. 9 and 2 with 213.07 and 197.86 U/L, respectively (Figure 5C). Xylanase – most fungal combinations showed the highest xylanase activity from day 3 to day 5, among which the highest was no. 6, which reached 107.47 U/L on day 5, followed by no. 11 (103.55 U/L at day 3; Figure 5D). Laccase – from Figure 5E, the enzyme activity of no. 4, 9, and 10 were significantly higher than that of other combinations. The enzyme activity of no. 9 reached 215.46 U/L on the seventh day of inoculation, more than twice that of a single fungus, followed by no. 10 (187.69 U/L). Peroxidase – it can be seen from Figure 5F that the peak of enzyme production of almost all fungal combinations was on day 11, and the enzyme activity of fungal combination no. 12 on day 11 was at least twice that of other fungal combinations, reaching 355.65 U/L. In addition, the enzyme activity of fungal combinations no. 6, 9, and 10 are also relatively advantageous, but not as high as that of single fungi (Figure 5F).

Thus, we suggest using co-cultivation viz., no. 1 Fomitopsis sp. (ZYJHYZ247) + Phaeophlebiopsis sp. (ZYJHYZ254); no. 6 Talaromyces sp. (ZYJHYZ163) + Coniochaeta sp. (ZYJHYZ246) + Fomitopsis sp. (ZYJHYZ247); no. 9 Talaromyces sp. (ZYJHYZ163) + Fomitopsis sp. (ZYJHYZ247) + Piloderma sp. (ZYJHYZ268); no. 10 Talaromyces sp. (ZYJHYZ163) + Nemania sp. (ZYJHYZ263) + Piloderma sp. (ZYJHYZ268) to degrade the anti-biodegradable residue rather introducing them as a monoculture.

In the co-cultivation, fermentation broth with anti-biodegradable residue as the sole carbon source was added at a constant temperature of 30× g at 28°C for 15 days (Figure 6). Compared with the negative control, about half of the residue was left in nos. 1 and 6, and about one-third of residue was left in nos. 9 and 10. The overall weight loss rate of TCMR is shown in Table 2. According to Table 2, the weight loss rate of no. 10 hydrolysis for 15 days is the highest, reaching 60.34%. The overall weight loss data is also consistent (Figure 6). Results of lignocellulosic content are shown in Table 3. It can be seen that the weight loss rate of cellulose in no. 1 after 15 days of hydrolysis is the highest, reaching 57.46%. Nos. 9 and 10 had the strongest hydrolysis ability for hemicellulose, and the weight loss rate of no. 10 reached 38.83%. No. 9 had the strongest hydrolysis capacity of lignin, reaching 64.47%.