The experimental site was a Plateau Radish Production Base in Ganzi Tibetan Autonomous Prefecture, Litang County, China (29°41′22″ N, 100°23′23″ E), which is 3637 m above sea level. This study used PBAT/PLA HA biodegradable silver black or black films and PBAT/PLA biodegradable black films as follows:

All films were purchased from Qingtian Plastic Co., Ltd.

Radishes (radish seed variety: Xiashuai 335; company: Zhengzhou Harvest Seed Co., Ltd.) were sown directly into the soil on July 20, 2023, and covered with the five mulch films (T1–5). The blank control was uncovered soil (CKw). Each treatment was performed three times. Soil samples were collected from the top 0–20 cm of the experimental plots using the five-point sampling method when the radishes were harvested on September 20, 2023. The soil pre-radish sowing was sampled in the same manner (CKq). A total of 105 soil samples (each 100 g) were collected and divided into seven groups according to the sample source after sieving and mixing, and stored at 80°C for microbial community structure and function analyses.

The soil samples were extracted to analyze the bacterial and fungal communities based on single-molecule real-time (SMRT™) DNA sequencing technology. Total DNA was extracted from the seven soil pools using the cetyltrimethylammonium bromide–sodium dodecyl sulfate (CTAB–SDS) method. The bacterial and fungal communities were analyzed by amplifying the 16S rRNA and ITS genes, respectively, of distinct regions with specific primers tagged with barcodes. The PCR was conducted using TransStart^® FastPfu DNA Polymerase (TransGen Biotech). Sequencing libraries were generated using a SMRTbell™ Template Prep Kit (PacBio) following the manufacturer's recommendations. The library quality was assessed on a Qubit 2.0 fluorometer (Thermo Fisher Scientific) and Femto Pulse system. Finally, the library was sequenced on the PacBio Sequel platform. All library preparation and sequencing procedures were performed by Novogene (China).

Total DNA was extracted using the CTAB–SDS method, and the genomic DNA sample was fragmented into short fragments for metagenome sequencing. These DNA fragments were end-polished, A-tailed, and ligated with full-length adapters for Illumina sequencing before size selection. Subsequently, PCR amplification was performed, and purification was conducted through an AMPure XP system (Beverly). The resulting library was assessed on an Agilent Fragment Analyzer System (Agilent) and quantified to 1.5 nM through Qubit (Thermo Fisher Scientific) and quantitative PCR. The qualified libraries were pooled and sequenced on Illumina platforms according to the effective library concentration and data amount required. Novogene performed all library preparation and sequencing procedures.

The 16S rRNA/ITS sequencing reads were trimmed in the PacBio SMART portal to remove low-quality reads, then truncated by cutting off the barcode and primer sequence for further analysis. Clean reads were obtained by removing chimera sequences through comparison with the reference database using the UCHIME algorithm (http://www.drive5.com/usearch/manual/uchime_algo.html). Then, the representative sequence for each operational taxonomic unit (OTU; clean reads with 97% similarity were assigned to the same OTUs) was produced using UPARSE (v7.0.1001) and assigned to the small subunit (SSU) rRNA databases of the SILVA and UNITE databases to complete the species annotation. Species abundance was normalized using a standard of sequence number corresponding to the sample with the most sequences.

Clean data were obtained by pre-processing the Illumina sequencing reads using Fastp (https://github.com/OpenGene/fastp), then de novo assembled using MEGAHIT with the following parameters: -presets meta-large (-end-to-end, -sensitive, -I 200, -X 400). Redundancy was eliminated using CD-HIT software and the unique gene was obtained. Subsequently, unigenes were aligned with those in the KEGG database using DIAMOND software (https://github.com/bbuchfink/diamond/) with the following parameter settings: blastp, -e 1e-5. The best blast hit results from the alignment results of each gene were selected for subsequent analysis. The abundance of each gene was estimated as reads per kilobase per million mapped reads (RPKM) using BBMap. Data with RPKM ≤ 10 were removed to prevent false positives.

The Shannon–Wiene index is an important indicator of alpha-diversity and is used to demonstrate the observed richness or evenness of microbial species within-group. Similarities or differences among the samples were visualized using principal co-ordinate analysis (PCoA) with the Bray-Curtis algorithm. The species or functional gene differences between groups were searched using metagenomeSeq and linear discriminant analysis (LDA) effect size (LEfSe) analyses. metagenomeSeq analysis involves permutation testing between groups on each taxonomy level and obtains a p-value. The LEfSe was conducted using LEfSe software (default LDA score: 4). The other related indices in the samples were calculated with QIIME (version 1.9.1) and displayed with R (version 2.15.3).

An average 375,137 contiguous sequences (contigs) were assembled in each sample and annotated to six sections and 46 domains of the KEGG pathways. Metabolism was the predominant microbiota functional activity, where carbohydrate fumetabolism, amino acid metabolism, energy metabolism, and metabolism of co-factors and vitamins accounted for 65% of all metabolic pathways (Figure 6). The results revealed a similar microbial activity composition in the PBAT/PLA biodegradable film and no-film groups, while the PBAT/PLA HA biodegradable film groups had similar microbial activity compositions, with higher functional activities such as metabolism (amino acid, carbohydrate, energy), cellular processes, and genetic information processing (cellular community-prokaryotes, cell growth and death, transcription, folding, sorting and degradation) (Figure 7). The metagenomeSeq analysis identified 12 differing functional genes, and the PBAT/PLA HA biodegradable film groups had higher abundances of carbohydrate metabolism, amino acid metabolism, energy metabolism, and cellular community_prokaryotes. Furthermore, decreasing the mulch film thickness tended to decrease the abovementioned abundances slightly (Figure 8).

In the present study, HTS technology was used to reveal for the first time the structure of the alpine radish root microbial community and microbial metabolic response to PBAT/PLA biodegradable mulch, providing important microbiological insights for sustainable high-altitude agricultural development. The results revealed that the alpine radish rhizosphere exhibited a distinct microbial community structure dominated by Proteobacteria, Acidobacteriota, Chloroflexi, Actinobacteria, Bacteroidota, and Verrucomicrobiota, as discovered by Huang and Eilers, Proteobacteria and Actinobacteria are specific microbes for cruciferous vegetable and cold-hardy plants, respectively (Huang et al., 2019; Eilers et al., 2010), suggesting strong plant–microbe interactions in high-altitude agroecosystems. The enrichment of Proteobacteria and Acidobacteriota in the rhizosphere aligned with their roles in carbon cycling and root exudate metabolism observed by Dey in low-altitude crops (Dey, 2017). However, the significant decline in Chloroflexi and Actinobacteria supports the “niche exclusion due to competitive root colonization” hypothesis proposed by Jacoby, though the high-altitude environment may intensify this competitive effect (Jacoby and Kopriva, 2019). Additionally, the prevalence of Granulicella and Rhodanobacter (which exhibit acid tolerance and sulfur metabolism capabilities) suggests a unique adaptation to alpine soil conditions, a phenomenon not previously reported in low-altitude radish cultivation systems. The fungal communities were structured by Basidiomycota, Unclassified_Fungi, Ascomycota, and Mortierellomycota and significantly shifted from Basidiomycota (saprotrophs) to Ascomycota (diverse symbionts/pathogens), suggesting that radish roots may selectively recruit beneficial fungi while suppressing wood-degrading taxa. The high proportion of Unclassified_Fungi highlighted the unexplored fungal diversity in alpine soils, which warrants metagenomic binning efforts. Studies have demonstrated that plant species can modulate active microbiota diversity through root exudation and environmental adaptive responses (Achouak et al., 2019; Perez-Mon et al., 2022). Therefore, further investigation into the combined effects of radish metabolome and alpine environmental factors on radish rhizosphere microbiota is warranted.

Given their unique characteristics of particle size, polymer type, additive composition, and content, mulch films have positive, negative, or no significant effects on soil microbial community abundance and diversity (Feng et al., 2022; Ke Li, 2025; Xu et al., 2024). However, MP pollution typically disrupts the microbial community composition in the soil surrounding plant roots. Nevertheless, HA counteracts the harmful effects of MPs on the rhizosphere soil microbial community composition (Virachabadoss et al., 2024). Similarly, our results revealed that the PBAT/PLA HA biodegradable films enhanced bacterial diversity and fostered distinct microbial consortia, likely due to the dual role of HA as a carbon source and microbial activity stimulant. Furthermore, we identified the marker bacteria Chitinophagaceae and Candidatus_Udaeobacter, which degrade complex polymers and may have accelerated mulch degradation by secreting hydrolytic enzymes (chitinase, esterase) and contributed to soil carbon cycling (Bandopadhyay et al., 2020; Eichorst et al., 2013; Hansen et al., 2023) following the application of the PBAT/PLA HA biodegradable films. In contrast, fungal communities were more sensitive to all mulch film types, possibly due to hyphal network disruption by film residues. Chaetomiaceae are a key fungal taxon for lignocellulose degradation, and their laccase gene is highly expressed in composted soils. Furthermore, Chaetomium globosum can degrade polyurethane by secreting esterases and peroxidases, which is a similar mechanism applicable to other polyester-based plastics such as PBAT (Liu et al., 2024; Qian et al., 2024). Therefore, the soil microbial composition was responsive to the physical and chemical properties of the PBAT/PLA HA biodegradable films, which enhanced soil microbial–plant interactions through film degradation.

The functional profiling of microbial communities in the radish rooting soil revealed significant metabolic adaptations in response to the PBAT/PLA biodegradable mulch films. The dominance of metabolic pathways (65% of total KEGG annotations), particularly carbohydrate, amino acid, and energy metabolism, aligned with previous studies demonstrating that polymer-degrading microbiomes prioritize core metabolic functions to utilize mulch-derived substrates (Zhong and Agarwal, 2024). Notably, the PBAT/PLA HA films enhanced functional activity compared to conventional PBAT/PLA films and bare soil, suggesting that HA may stimulate microbial metabolism through two mechanisms: (1) as a supplementary carbon source that co-metabolizes with mulch components, or (2) it improves soil physicochemical properties to facilitate microbial access to polymeric substrates (Li et al., 2022). The distinct functional signature of the HA-modified films was characterized by elevated carbohydrate metabolism and cellular community functions, indicating a more robust microbial network capable of coordinated mulch degradation. This result was supported by the metagenomeSeq data that demonstrated enriched genes for extracellular hydrolases and prokaryotic cell–cell interaction pathways.

The observed thickness-dependent decline in metabolic gene abundance may reflect substrate limitation in thinner films, consistent with a previously proposed carbon-use efficiency theory (Sinsabaugh et al., 2017). These results advance our understanding of how mulch formulation affects soil microbiome functionality. However, the thickness correlation introduced new considerations for agricultural practice: while thicker films may sustain metabolic activity, optimal thickness must balance degradation rates with crop cycle durations. Future studies should quantify the temporal dynamics of these metabolic shifts and their association with mulch mineralization rates.