Industrial WWTP Sludge as a Reservoir of Putative Plastic-Degrading Enzymes: Insights Into Microbial Community Structure

Abstract

Plastic pollution has emerged as a major environmental challenge. While various microorganisms and enzymatic mechanisms involved in plastic degradation have been described, the microbial communities present in wastewater treatment plants (WWTPs) - particularly those linked to the oil industry – remain largely underexplored as potential reservoirs of plastic-associated metabolic functions. This study evaluated the potential of microbial communities from an industrial WWTP located in an oil refinery complex to colonise different plastic polymers, as well as their genetic potential for plastic-associated transformations. Industrial sewage sludge was incubated with four polymeric substrates: low-density polyethylene, polyamide, polypropylene, and polystyrene. Metagenomic analysis of the inoculum was performed to assess its potential to harbour genes encoding proteins with sequence similarity to enzymes previously associated with plastic transformation or degradation. Additionally, 16S rRNA amplicon sequencing studied changes in microbial community composition during biofilm formation. Proteobacteria accounted for approximately 50% of the microbial relative abundance in the inoculum. Metagenomic screening identified 773 genes encoding proteins with homology to enzymes previously associated with plastic degradation. Biofilms developed on plastic surfaces were dominated by Proteobacteria and Bacteroidota, although differences in dominant classes and orders were observed depending on the polymer type. Correlation analyses further indicated associations between specific microbial taxa and individual plastic types. Overall, the results highlight industrial WWTP sludge as a potential reservoir of microorganisms associated with plastic colonisation, providing new insights into plastisphere community assembly in engineered environments.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s00248-026-02839-7.

Introduction

Plastic pollution has become a major environmental problem. In 2023, plastic production reached 413 Mt [1], and it is expected to increase in the coming years [2]. The main plastic polymers produced worldwide include polyethylene (PE), polypropylene (PP), polystyrene (PS), poly(vinyl chloride) (PVC), poly(ethylene terephthalate) (PET), polyurethanes (PUs), and polyamides (PAs) [3]. As most plastics are fossil-based polymers with high chemical stability and resistance to biodegradation, together with inadequate waste management strategies, plastic waste accumulates in natural environments. Once in the environment, they provide a durable substratum for microbial colonisation [4, 5]. The resulting biofilms often harbour microbial communities that differ from those in the surrounding environment, leading to the concept of plastisphere, a specific ecological niche associated with plastic surfaces [4, 6]. Hence, plastic-associated microbial communities have been investigated across a wide range of environments, with a particular focus on marine ecosystems [7]. In some cases, biofilm formation may precede biodeterioration or enzymatic modification of plastic surfaces [8].

In addition to marine and terrestrial habitats, wastewater treatment plants (WWTPs) are recognised as an important niche of plastic pollution, mainly in the form of microplastics [8]. Consequently, the microbial communities in these facilities may represent a potential reservoir of microorganisms capable of colonising plastic materials. For instance, Rajcoomar et al. [9] examined factors influencing biofilm formation on PP and high-density polyethylene (HDPE) microplastics in domestic wastewater. Moreover, few other studies have focused on finding plastic-degrading microorganisms in this environment, usually by isolating bacterial strains [8, 10].

To date, several microorganisms able to colonise plastic surfaces and potentially participate in their transformation have been identified, along with enzymes associated with these processes. However, substantial additional evidence is still required to demonstrate actual biodegradation, particularly for polymers such as PE, PP or PS [11].

Nevertheless, the potential involvement of microorganisms in plastic-associated processes is evident. In this context, frequently reported bacterial plastic degraders include members of the Proteobacteria, Actinobacteria, Firmicutes, Bacteroidetes, and Cyanobacteria phyla [11]. Among these, Proteobacteria are particularly relevant, as they are frequently associated with the degradation of a wide range of pollutants, including hydrocarbon-based compounds [3].

In recent years, increasing attention has been given to hydrocarbonoclastic microorganisms in plastic colonisation and transformation due to their frequent detection in plastisphere communities and their ability to use hydrophobic hydrocarbons as carbon and energy sources [12, 13]. Given their metabolic affinity for these substrates, it has been suggested that they may contribute to biofilm formation on petroleum-derived plastics and may participate in their initial surface modification or transformation. However, limited information can be found about specific enzymes involved in this process. Zettler et al. [4] identified hydrocarbon-degrading bacteria among the microbial taxa associated with PP and PE plastics. Additionally, Denaro et al. [12] reported PET surface alterations caused by hydrocarbon-degrading biofilms derived from marine environments after acclimation to representative hydrocarbon fractions. More recently, Zadjelovic et al. [13] isolated an Alcanivorax strain -a well-known hydrocarbonoclastic genus-capable of growing in the presence of weathered low-density polyethylene (LDPE) and provided evidence of primary biodegradation of pristine plastic, supported by increased expression of AlkB and AlmA monoooxygenases. AlkB monooxygenases catalyse short- and medium-chain alkane hydroxylation, while AlmA is proposed to convert sub-terminally hydroxylated alkanes into esters.

These observations suggest that industrial WWTPs associated with petroleum refineries may represent unique microbial reservoirs enriched in hydrocarbon-degrading bacteria, due to their continuous exposure to petroleum-derived compounds. However, information on plastic colonisation and plastisphere communities in these environments remains limited.

This study investigates the microbial communities present in sludge from an industrial WWTP treating petroleum refinery wastewater and their interactions with various plastic polymers. For this purpose, sludge was incubated with various plastic substrates, namely low-density polyethylene, polystyrene, polyamide and polypropylene. Metagenomic analysis of the inoculum was performed to identify genes whose encoding proteins are homologous to those associated with plastic-degradation processes. In addition, changes in microbial community structure and composition during biofilm development were analysed using 16S rRNA amplicon sequencing.

Methods

Inoculum

The inoculum was collected from the aerated tank of the secondary treatment of a WWTP belonging to a petroleum refinery located in northwestern Spain. The facility treats approximately 2,190,000 m³ of wastewater yearly. Its treatment process begins with a primary stage that employs API-type separators and 2 homogenization tanks, followed by coagulation–flocculation basins, a Dissolved Air Flotation (DAF) unit, and a Conventional Activated Sludge (CAS) system. Although most hydrocarbon compounds are removed by the applied treatments, residual oil content was still detected in the aeration tank, with an average concentration of 8.5 mg/L. The inoculum was stored at 4 °C until inoculation. Before the experiments, 2 mL of the sludge were collected and preserved at -20 °C. 1 mL was used for whole metagenome analysis. The remaining 1 mL was used for 16S rRNA amplicon sequencing, which was performed in triplicate.

Plastic Materials

Plastic materials were selected based on their widespread use and environmental occurrence. Plastic fragments of approximately 2 cm² were obtained from common products available on the market: LDPE from a plastic bag, PS from a food tray, PA (specifically polyamide 6) from garments, and PP fabric from a face mask filter. Their composition was analysed using Fourier transform infrared (FTIR) spectroscopy (See Supplementary Materials S1, Figure S1.1–4). All fragments were incubated in 70% ethanol for 24 h before the experiments to minimize potential contamination. They were then rinsed with sterile distilled water, and each side was exposed to UV light for 15 min under sterile conditions (MSC-Advantage™ Class II Biological Safety Cabinet (λ = 254 nm, irradiance = 500 µW/cm²)).

Experimental Set-up

To promote biofilm formation, a biodisk system was used [14]. This device consists of a rotating wheel in continuous contact with a liquid medium containing microorganisms in suspension with a final volume of 700 mL. Fragments of the four plastic polymers were attached to both sides of the wheel (See Supplementary Materials, Figure S1.5). A single biodisk incubation was performed with all polymer types. For each polymer, six fragments were incubated, and three biofilm subsamples were obtained by pooling biofilm from two fragments per sample. A synthetic medium simulating wastewater conditions was used, with the following composition (mg/L): glucose 1000, meat extract 110, peptone 160, K₂HPO₄ 28, urea 30, NaCl 7, CaCl₂·2 H₂O 4, MgSO₄·7 H₂O 4, and sodium dodecyl sulphate 100. The system was kept at room temperature under conditions minimising external contamination for 30 days. Glucose was added initially to support biofilm development. Fresh medium without additional glucose was periodically added to compensate for evaporation losses. Although the replenishment volume and NaCl concentration were low, minor accumulation of non-volatile solutes cannot be ruled out.

After the incubation period, plastic fragments were removed from the disk and rinsed with sterile distilled water to remove planktonic cells. Three biofilm subsamples per plastic polymer were resuspended in phosphate-buffered saline solution (PBS) at pH 7.4 for 16S rRNA amplicon sequencing. Samples were stored at -20 ° C until further analysis.

Whole Metagenome Sequencing and Data Processing

Metagenomic DNA from the inoculum sample was extracted using the Nucleospin™ microbial DNA extraction kit (Macherey-Nagel™) following the manufacturer’s instructions. Quantification of extracted DNA was performed by spectrophotometry (Thermo Scientific™ Nanodrop 2000 C) and fluorometry (InvitroGen™ Qubit™). Library preparation targeted 350 bp fragments and sequencing was performed using Illumina NovaSeq platform with paired-end 150 bp reads (Novogene Co. Ltd., UK). Whole metagenome sequencing raw data are available through the NCBI Sequence Read Archive (SRA) under accession number SRX28665119.

Raw reads were quality-filtered with fastp [15] and assembled into contigs using MEGAHIT (v1.0.4) [16], configured with the meta-large preset. Unmapped reads were reassembled and merged with individual assemblies using Bowtie2 (v2.2.4) [17]. Mixed assembly was conducted on the unused reads with the same assembly parameters. ORF prediction and dereplication were performed with MetaGeneMark (v2.10) [18] and CD-HIT [19], respectively. The number of reads mapped to each gene was calculated from the Bowtie2 alignment results. Gene abundance was estimated using a length-normalised approach, in which the number of reads aligned to each gene was divided by the corresponding gene length. Taxonomic annotation was performed using DIAMOND (v0.9.9.11) [20]. Unigenes sequences were aligned against the microNR database (v2018.01.0) (parameter setting: blastp, E-value: 1e^{− 5}), followed by LCA assignment using MEGAN. Abundance tables of taxa were exported and further processed with python (v3.12). Bar plots depicting the relative abundance of taxonomic groups were generated with matplotlib/seaborn libraries.

Screening of Genes Encoding Putative Plastic-Degrading Enzymes

Identified translated ORFs were aligned to the sequences present in the Plastic Biodegradation Database (PlasticDB) [21] (Accessed in March 2025). The alignment was performed using DIAMOND (v2.1.10) with the default parameters (BLOSUM62 scoring matrix, sensitive mode, E-value: 0.001) [18]. The files resulting from the alignment were exported and further processed with Python (v3.12). Only results for “PE”, “LDPE”, “PS” and “Nylon” terms were considered. Stacked bar plots showing the genes assigned to PlasticDB enzyme groups were generated with matplotlib/seaborn libraries. Heatmaps in the Supplementary Materials were elaborated using the sns.heatmap function.

16S rRNA Amplicon Sequencing and Data Processing

Genomic DNA extraction and quantification for the inoculum and biofilm samples were carried out following the same procedure as in “Whole metagenome sequencing and data processing” Section. The V3-V4 hypervariable region of the 16S rRNA gene was amplified and sequenced using Illumina MiSeq platform with paired-end 250 bp reads (Novogene Co. Ltd., UK). Raw data are available through the NCBI Sequence Read Archive (SRA) under accession numbers SRX28671707-SRX28671718. The Bioproject accession number is PRJNA1258906.

The raw reads were filtered with Cutadapt v3.3 [22]. Then, flash v1.2.11 [23] and fastp v0.23.1 [15] were used to merge reads and obtain clean tags, respectively. The chimeric sequences in clean tags were detected and removed obtaining effective tags with vsearch 2.16.0 [24]. Further steps were performed with QIIME2 (v2024.10.1). Denoising to obtain Amplicon Sequence Variants (ASVs) was carried out with DADA2 [25]. Taxonomic annotation was conducted using a pre-trained V3-V4 16 S classifier based on the MIDAS 5.3 database [26]. Diversity analysis was carried out with QIIME2 diversity-alpha and core-metrics-phylogenetic plugins, using rarefaction at a sequence depth of 32,834. Principal Coordinate Analysis (PCoA) was performed using Bray-Curtis dissimilarity to visualize compositional differences between samples. A permutational multivariate analysis of variance (PERMANOVA) was conducted to test for differences among groups using the Bray-Curtis distance matrix, using the beta-group-significance command. Differential abundance analysis was conducted at multiple taxonomic levels using ANCOM-BC via QIIME2 composition plugin, using default settings (statistical significance at FDR-adjusted p < 0.05). Kendall’s correlation test (τ) [27] was then applied to assess associations between plastics and biofilm microbial communities, using the Kendall tau function in Python (v3.12) with significance defined as p < 0.05.

Results

Plastic Degradative Potential of the Inoculum

The microbial composition of the inoculum and the genetic potential for encoding for plastic-degrading enzymes were analysed. Proteobacteria was the most abundant phylum (approximately 50% of the community), followed by Chloroflexi (approximately 20%), Actinobacteria (around 10%), Planctomycetes and Acidobacteria (Fig. 1a). At lower taxonomic levels, Burkholderiales and Rhizobiales orders were predominant within Proteobacteria, while Ardenticatenales dominated Chloroflexi (See Supplementary Materials S1, Figures S1.6, 7).

Fig. 1.

Whole metagenome sequencing results of the inoculum. A Microbial community composition at phylum level. Only the 10 most abundant phyla are shown, while the remaining taxa are grouped under “Others”. B Bar plots showing the number of genes assigned to each enzyme group in the PlasticDB for polyethylene (PE), polystyrene (PS), and polyamide (PA) [21]. The PA category corresponds to entries annotated as “Nylon” in the database, whereas PE includes both ‘PE’ and ‘LDPE’ entries. Individual genes that matched multiple enzyme groups associated with the same plastic type were counted more than once within that category. The complete DIAMOND blastp results are provided in Supplementary Materials S3

Metagenomic analysis identified 773 genes potentially associated with different plastics from 839,122 predicted coding sequences, with a wide range of sequence identities (See Supplementary Materials S3). As some sequences matched multiple enzymes from the database, unique gene matches are shown in Fig. 1b. For PA, 512 genes were identified, associated with Hydrolases from Pseudomonas sp. and Paenarthrobacter ureafaciens [28, 29]; a Polyamidase from Nocardia farcinica [30]; and a Nylon oligomer-degrading enzyme from Agromyces sp. [31], all described as amidases capable of modifying PA surface [32]. For PE, 255 genes were associated with an Alkane hydroxylase from Pseudomonas sp. [33], an Alkane monooxygenase from Paenibacillus sp. [34], Laccases from Rhodococcus ruber [35] and Psychrobacter sp. [36] and a Copper oxidase from Rhodococcus opacus [37]. Finally, 6 genes matched the Alkane-1-monooxygenase from Acinetobacter johnsonii, linked to PS degradation [38]. Most genes were affiliated with Proteobacteria, Actinobacteria and Chloroflexi phyla (See Supplementary Materials S1, Figures S1.8–10).

Changes in Microbial Community During Biofilm Formation

The experiment lasted 30 days. Biofilm formation was visually detected within the first 15 days for all materials. DNA extractions were performed in triplicate; however, some samples were excluded from the final analysis due to poor-quality sequences (one PA sample and one PP sample). In addition, reads in the remaining PP samples assigned to “chloroplast” were filtered out.

A total number of 4,618 ASVs were identified. Principal Coordinate Analysis (PCoA) based on Bray-Curtis distances showed that PC1 and PC2 explained 68.8% and 15.7% of the variance, respectively, suggesting that the main differences in microbial community structure occurred between the inoculum and biofilms. Within biofilms, PA and PP tended to cluster separately from LDPE and PS (Fig. 2).

Fig. 2.

Principal Coordinate Analysis (PCoA), based on Bray–Curtis dissimilarity, showing differences in microbial community composition between the inoculum and the biofilms formed on different polymers. Each point represents an individual sample and is coloured and labelled according to its origin: Inoculum (green, INO); biofilms grown on polypropylene (yellow, PP), polystyrene (orange, PS), Low Density Polyethylene (blue, LDPE), and polyamide (purple, PA)

PERMANOVA did not detect statistically significant differences in community composition among groups, likely reflecting the limited sample size and variability among replicates (See Supplementary materials S1, Table S1.1). The largest differences (as indicated by pseudo-F values) were observed between inoculum and the plastic-associated communities, particularly for PA and PS (F-values for inoculum vs. PA:97.5; F-value for inoculum vs. PS: 69.1). Although no significant differences were detected in alpha diversity, a slight decrease in diversity was observed in most biofilm samples, except for PP biofilms (See Supplementary Materials, Table S1.2).

However, taxonomic profiles, differential abundance analysis, and correlation test revealed marked shifts between the inoculum and plastic-associated biofilms, as well as differences related to polymer type.

Proteobacteria and Bacteroidota were the dominant phyla in all samples, consistent with metagenomic results (Fig. 3a). Proteobacteria accounted for nearly 50% of the microbial relative abundance in the inoculum and increased in all biofilms. In contrast, Bacteroidota relative abundance decreased to around 20% in LDPE and PS biofilms and to below 10% in PA and PP samples. Differences between the inoculum and the biofilms were also observed among less abundant phyla. Chloroflexi, Acidobacteria, Actinobacteria, and Myxococcota together accounted for almost 20% of the inoculum, but their relative abundance decreased in biofilms, while Firmicutes and Cyanobacteria became more abundant.

Fig. 3.

Differences in microbial community composition across samples. A Microbial community composition at phylum level. Only the 10 most abundant phyla are shown, while the remaining taxa are grouped as “Others”. B Community composition within the phylum Proteobacteria at order level. Only the 10 most abundant orders are displayed, while the remaining taxa are grouped as ‘Others’. C Community composition within the Bacteroidota phylum at order level. Only the most abundant orders are shown, while the remaining taxa are grouped as ‘Others’. Samples were taken in triplicates, and they are labelled according to their origin: Inoculum (INO), and biofilms developed on Low Density Polyethylene (LDPE), polystyrene (PS), polyamide (PA), and polypropylene (PP). Replicates correspond to 16S rRNA amplicon sequencing samples obtained from the same biodisk incubation and should not be interpreted as independent biodisk reactors. One PP and one PA biofilm sample were removed from the analysis due to poor sequence quality

These changes became more evident at lower taxonomic levels. Within Proteobacteria (Fig. 3b), microorganisms belonging to Alphaproteobacteria class dominated the inoculum community. In contrast, Betaproteobacteria and Gammaproteobacteria increased markedly in biofilm samples.

At order level, Rhizobiales represented more than 30% of the community in the inoculum, whereas in LDPE and PS biofilms, they were replaced by Burkholderiales as the dominant order (over 30%), followed by Enterobacterales. In PA and PP biofilms, similar relative abundances of Burkholderiales and Enterobacterales were observed. A comparable pattern was observed within the Bacteroidota phylum (Fig. 3c). In the inoculum, Chitinophagales accounted for almost 95% of its relative abundance. However, in biofilm samples, the Bacteroidota community was mainly composed of Flavobacteriales, followed by Sphingobacterales.

Differential abundance analysis further confirmed significant shifts in community composition between the inoculum and the biofilms. Only orders with adjusted p-values < 0.05 and large effect sizes (log fold change (LFC) > 4) are displayed in Fig. 4. All significant orders and effect sizes are provided in the Supplementary Materials S2. At this level, Enterobacterales showed a significant increase in LDPE, PA and PP biofilms (LFC > 4 with adjusted-p < 0.001 in all cases), while Flavobacteriales significantly increased in LDPE and PS biofilms (LFC > 4 with adjusted-p < 0.001). Conversely, Chitinophagales significantly decreased in LDPE, PA and PS biofilms (adjusted-p < 0.001 in all cases). The increase in Burkholderiales was also significant in all the biofilms (LFC > 2, adj-p < 0.001), as well as the increase of Flavobacteriales for PP (LFC > 1.5, adjusted-p < 0.001). Changes were also observed in less abundant taxa, such as an increase in Lactobacillales.

Fig. 4.

Bubble plot showing differentially abundant microbial orders in plastic biofilms relative to the inoculum. Only taxa with significant differential abundance (FDR-corrected p-value < 0.05) with values of |log fold change| > 4 are displayed. Bubble size indicates the magnitude of the log fold change (LFC) values with respect to the inoculum, while colour indicates the direction of change (blue = decreased in biofilms; red = increased in biofilms). All statistically significant results are reported in Supplementary Materials S2

Kendall correlation analysis assessed associations between microbial taxa and plastic substrates, with coefficients from − 1 to 1 indicating weak (near 0) to strong negative and positive associations, respectively. The results (Fig. 5) revealed several positive associations between specific microorganisms and individual plastic types. Although no clear overall trends were observed, Proteobacteria, particularly Alphaproteobacteria, were frequently associated with PS and PA biofilms. In contrast, taxa correlated with PP were more taxonomically diverse, including orders from Firmicutes, Desulfobacterota, or Acidobacteriota. Similarly, members of Firmicutes appear to be more strongly associated with LDPE biofilms.

Fig. 5.

Heatmap showing significant positive correlations between microbial orders and individual polymers based on Kendall’s rank correlation test. Only strong positive correlations (τ > 0.5) with statistically significant results (p < 0.05) are shown. Colour intensity indicates the strength of the correlation coefficient (τ)

Taken together, these results indicate that plastic polymers could enrich specific microbial populations from the inoculum, supporting the concept of substrate-driven plastisphere community assembly in engineered wastewater environments.

Discussion

The results of this study provide new insights into the plastic colonisation of microbial communities from industrial WWTP sludge, as well as the structure of plastisphere communities.

Our first aim was to describe the composition and genetic repertoire of the inoculum to better understand its potential for plastic degradation. As shown in Fig. 1, the microbial community was dominated by Proteobacteria, Chloroflexi and Actinobacteria. The clear dominance of Proteobacteria is consistent with the literature, as is the relevant abundance of Chloroflexi [10].

Genes encoding enzymes potentially involved in plastic polymer degradation revealed marked differences, with a clear predominance of genes associated with PA over those linked to PS or PE. This may be partially explained by the enzyme class composition in the database. In the case of PA-related enzymes, these enzymes are often hydrolases or polyamidases, with relatively broad substrate scope, which increases the likelihood of detecting homologues not limited to plastic degradation. Similarly, PE- related hits were dominated by laccases and a copper oxidase. At the same time, a lower number is observed with the alkane hydroxylase and alkane monooxygenases, which tend to be more specialised. Lastly, for PS, only one alkane monooxygenase is included in the database [38]. This bias probably reflects the fact that amide bonds in PA could be more susceptible to enzymatic attack than the stable C-C bonds present in the other polymers [3, 28]. The broad range of sequence identities observed in the blastp results further highlights the limitations of homology-based annotation, since similarity alone does not necessarily indicate the same enzymatic function. In addition, the same genes scoring with different enzymes suggest that certain genes may encode multifunctional enzymes or belong to families with broad substrate specificity. Moreover, variability in nomenclature may introduce ambiguity when interpreting the functional potential of metagenomic datasets [39]. The detection of genes encoding proteins homologous to alkane monooxygenases, alkane hydroxylases and laccases is consistent with the origin of the inoculum, since the literature indicates that microorganisms adapted to such compounds frequently harbour enzymes involved in alkane oxidation [36]. However, as emphasised by Schnoes et al. [39], sequence-based annotation alone can substantially overpredict molecular function, particularly in large public datasets, where homologous proteins may be assigned to activities without sufficient experimental evidence. In this context, the presence of genes encoding proteins with homology to plastic-related enzymes should be interpreted as indicative of potential rather than confirmed biodegradation capacity. As different authors have stated before, there is a lack of strong evidence of actual biodegradation of PS, PE, and PP polymers [11, 40], and reports on plastic polymer degradation sometimes focus on additives, monomers, or oligomers rather than the actual plastic, or even only on surface modification [32]. Moreover, abiotic processes are often considered necessary to enhance biodegradation processes [40, 41]. Overall, our results highlight the refinery WWTP sludge as a reservoir of putative plastic-degradation enzymatic potential. Nevertheless, further research is needed to assess the actual degradative capacity of these microorganisms for these plastics, including the potential role of horizontal gene transfer in shaping these plastic-associated genes.

Focusing on microbial community structure and composition, only a limited number of studies have investigated bacterial biofilms associated with plastics in WWTP, and none of them have specifically focused on the petroleum industry. For instance, Rajcoomar et al. [9] studied the main factors affecting biofilm formation on microplastics in domestic wastewater, specifically PP and PE fragments. Lastly, Huang et al. [42] reviewed biofilm-based WWTPs, but focused on the performance of these systems and the influence of MP. However, the phenomenon of plastic colonisation has been widely studied in marine environments, with particular emphasis on the importance of hydrocarbonoclastic microorganisms [6, 39]. In this context, the observed shifts in the microbial communities should be interpreted with caution, as they may be partially driven by selective pressure by the culture medium and the effect of plastic colonisation. Additionally, the absence of abiotic controls, such as plastics incubated without inoculum and non-plastic surface controls (i.e. glass, cellulose) limits the ability to distinguish biotic contribution from physicochemical processes influencing these patterns.

PERMANOVA and alpha diversity analyses did not reveal statistically significant differences; however, the robustness of these analyses is limited by the relatively small sample size, and PCoA patterns should be interpreted as descriptive trends. Nevertheless, these observations are in concordance with previous literature. Specifically, Rajcoomar et al. [9] reported higher diversity in PP compared to PE microplastics in urban wastewater biofilms. Other authors have similarly indicated that composition and richness of microbial communities present in plastic debris are influenced by their physicochemical properties [43, 44].

On the other hand, differential abundance analysis identified significant differences in the relative abundance of specific taxa, particularly in Proteobacteria and Bacteroidota. Several authors have stated the clear dominance of both phyla in plastic colonisation processes [41], with similar increases in Proteobacteria and decreases in Bacteroidota populations described in LDPE marine biofilms [45]. These shifts are consistent with selective enrichment during plastisphere community assembly on plastic substrates, even though they are not directly related to biodegradation. Additionally, metagenomic analysis showed a relevant number of identified genes related to this phylum (See Supplementary Materials S1, Figures S1.8–10). However, metagenomics analysis should be performed to assess the actual enrichment in gene functions in the biofilm communities.

Focusing on the role of Bacteroidota in plastic colonisation, some authors mention this phylum as secondary colonisers [46], although other studies identify them as relevant microorganisms in the initial steps [44]. In any case, all authors agree that they are part of the relevant populations in the biofilms, especially Flavobacteriales order, the main order identified in our study. Flavobacteriales often dominate microbial communities on plastic polymers during the early stages of colonisation, demonstrating strong biofilm-forming capabilities on plastic surfaces [45]. They contribute to the complex biofilm community developed on plastics in aquatic ecosystems, and their ecological prevalence is linked to the secretion of proteins and membrane components that facilitate surface adhesion and interactions within biofilm matrices [43, 47]. Hence, our findings suggest that similar colonisation patterns may occur in plastisphere communities from industrial WWTP environments.

Additionally, Kendall’s correlation results further suggest a degree of substrate-driven specialisation within the plastisphere communities, with strong positive correlations, such as Alphaproteobacteria members linked to PS and PA, or Desulfobacterota and Firmicutes members linked to PP and LDPE.

Notably, many of these associations involved low-abundance taxa, highlighting the potential ecological importance of rare microbial populations within plastisphere communities. Diverse authors have highlighted the relevance of low-abundance taxa, particularly during early colonisation events [48].

Overall, these findings suggest that industrial WWTP sludge represents a unique microbial reservoir in which exposure to hydrocarbon-rich environments may promote the enrichment of community composition capable of colonising plastic substrates.

Conclusions

This study provides insights into the microbial colonisation of plastic surfaces (LDPE, PS, PA and PP) by communities originating from industrial WWTP sludge associated with a petroleum refinery. The most abundant microorganisms in the inoculum belonged to Proteobacteria, Chloroflexi, and Actinobacteria phyla, and metagenomic analysis revealed several genes encoding proteins with homology to enzymes potentially involved in plastic-degradation processes. Focusing on the developed biofilms, although PERMANOVA results were not statistically significant, our analyses showed consistent shifts in microbial composition, with a marked predominance of Proteobacteria and Bacteroidota. Plastic type was associated with the enrichment of less abundant microbial taxa, suggesting a degree of substrate-related specialisation within the community.

Although direct plastic degradation was not demonstrated, the results highlight industrial WWTP sludge as a promising microbial reservoir with functional potential for future studies on plastic biodegradation.

Supplementary Information

Below is the link to the electronic supplementary material.

Authors’ Contributions

C. Vijande: Investigation, Writing - Original Draft C. Fuentes-González : Investigation; M. Lazzari : Writing - Review & Editing, Supervision J. M. Lema : Writing - Review & Editing, Supervision, Conceptualization S. Balboa: Writing - Review & Editing, Supervision, Conceptualization, Funding acquisition.

Funding

This work was financed by the Galician Government through (project ED431F2024/302), the Spanish Investigation Agency (project RYC2021-031220-I) and a PhD Xunta de Galicia Grant (ED481A 2022/391, Carlota Vijande Álvarez de Linera). The authors belong to a Galician Competitive Research Group (ED431C-2021/37) and to Cross-disciplinary Research in Environmental Technologies (CRETUS Research Center, ED431G 2023/12).

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.