Enrichment of antibiotic resistant genes and pathogens in face masks from coastal environments

Abstract

Face masks (FMs) are essential to limit the spread of the coronavirus during pandemic, a considerable of which are accumulated on the coast. However, limited is known about the microbial profile in the biofilm of the face masks (so-called plastisphere) and the impacts of face masks on the surrounding environments. We herein performed face mask exposures to coastal sediments and characterized the microbial community and the antibiotic resistome. We detected 64 antibiotic-resistance genes (ARGs) and 12 mobile gene elements (MGEs) in the plastisphere. Significant enrichments were found in the relative abundance of total ARGs in the plastisphere compared to the sediments. In detail, the relative abundance of tetracycline, multidrug, macrolide-lincosamide-streptogramin B (MLSB), and phenicol-resistant genes had increased by 5–10 times. Moreover, the relative abundance of specific hydrocarbonoclastic bacteria (e.g., Polycyclovorans sp.), pathogens (e.g., Pseudomonas oleovorans), and total MGEs significantly increased in the sediments after face mask exposure, which was congruent with the alteration of pH value and metal concentrations in the microcosms. Our study demonstrated the negative impacts of FMs on coastal environments regardless of the profiles of ARGs or pathogens. These findings improved the understanding of the ecological risks of face masks and underlined the importance of beach cleaning.

Graphical Abstract

1. Introduction

The outbreak of coronavirus disease 2019 (COVID-19) made it essential to wear face masks (FMs) in public settings to limit the spread of the disease [18]. Monthly consumption of FMs reached 129 billion during the epidemic, resulting in the contamination of FMs in widespread environments [47]. Coastal regions were found to be important reservoirs of FMs after their disposals [16], [40], [46]. For example, up to 100 FMs were semi-buried in the sediment of a beach [2], [40], and it was estimated that more than 0.39 million tons of FMs were emitted into the ocean within a year [15]. Disposed FMs can adversely affect marine organisms, including wildlife entanglement and accidental ingestion (e.g., eaten by penguin) [45].

The majority of FMs are made up of plastic fibers [42] that can be broken down into small pieces via UV light, waves, or abrasion [64]. A previous study showed that one FM can release millions of microplastics (< 5 mm) [61], [62], causing a blooming threat to the marine environment [19]. Once FMs enter the ocean, they can be rapidly colonized by microorganisms, the so-called “plastisphere” [69]. The microbial community in the plastisphere is significantly distinct from that in the surrounding seawater, including a wealth of unknown species and potential pathogens [22]. Indeed, plastic materials such as FMs could be carriers for a wide range of pollutants, e.g., organic pollutants and heavy metals [68] as well as biological hazards, e.g., pathogens [8] and antibiotic-resistant genes [77].

FMs can absorb and accumulate organic and inorganic pollutants, such as antibiotics [30] and heavy metals [31], conferring the bacteria inside the plastisphere to select or co-select antibiotic-resistant genes (ARGs) [5], [50]. A recent study showed that FMs had even higher adsorption affinity for heavy metals than other existing microplastics [31], suggesting that FM might be a stronger absorbent. The bacterial community and mobile gene elements (MGEs, e.g., transposons) can also drive the ARG dynamics in the plastisphere [20]. Recent studies underlined the selective enrichments for specific ARGs [35], and a higher rate of horizontal gene transfer in the plastisphere compared to the surrounding aquatic environment [4]. Even though studies have investigated a wide spectrum of ARGs in the plastisphere [35], limited is known about the ARG resistome. Since most FMs (> 80%) were considered to be accumulated in the coastal sediments [46,[74], [75], their ARG resistome is needed for further studies [73].

The impact of FMs on their surrounding micro-ecology was another concern. Recent studies showed that the additives of FMs could be released into the ocean (e.g., Pb, Cu, and Zn) [9]. In addition, organic matters such as antioxidants or volatile chemicals in FMs were also found to be released into the ambient environments [33], [34]. The leaching of dissolved organic matter can stimulate bacterial activity in the surrounding environments [51]. Therefore, FMs might impact the microbial communities and ARG profiles in their surroundings, considering their large surface areas.

Based on this, the current study aimed to unveil the microbial community and the ARG profile in the plastisphere of FMs and their surrounding sediments. First, we hypothesized that the microbial community in the plastisphere would be different from that of the sediments, which might result in an enrichment of ARGs in the plastisphere compared to the surroundings. Second, we hypothesized that FM exposure to the sediments could have an impact on its surrounding microbial profile. For this, FMs were exposed to sediments collected from three representative areas, i.e., a beach [2], an aquafarm [27], and a mangrove [1], [38]. Then, the microbial community and the ARG profile were further characterized. This study could provide more insights of FM risks into marine environments, especially human recreational areas (e.g., beaches).

2. Materials and methods

2.1. Experimental setup

A microcosm experiment was set up by exposing surgical FMs to sediments collected from three sampling sites， i.e.， a beach (114°35'E, 22°32'N), an aquafarm (114°31'E, 22°28'N), and a mangrove (114°34'E, 22°29'N) along the coast of Shenzhen, China, in November 2021. These samples were collected from the top surface centimeters together with the surrounding seawater at low tide. The respective sediment was comprised of fine sand (aquafarm), medium sand (beach), and silt (mangrove). After transportation to the laboratory, sediments were homogenized with a stirrer, and impurities (e.g., living organisms and woods) were removed with a net.

For each site, three glass tanks with a capacity of 10 L were used for the FM exposure. In parallel, another three tanks without FM exposure were used as the blanks. Thus, a total of 18 tanks were used for the three sites (Fig. S1). The microcosms were established by depositing a sediment layer (∼ 3 cm) in the tank bottom, followed by an FM in the middle, and a sediment layer (∼ 0.5 cm) again on the surface. Seawater from the three sites was further added to the tanks (∼ 2.5 cm). Pristine FMs were used in this study to avoid the potential influences resulting from the heterogeneous bacterial community or adsorbent materials after wearing. The tanks were maintained at 22 ℃ and illuminated from above under a 12/12 h light/dark rhythm for two months. 0.22 µm filtered milli-Q water was added to each tank per week until reaching its initial level to compensate for the water evaporation.

After two months of incubation, the physiochemical parameters of overlying water were measured. Salinity, dissolved oxygen, pH together with conductivity, total organic carbon (TOC) were measured using a hand-held practical salinity refractometer, a portable oxygen meter (SMAT AR8406, China), a pH meter (Mettler Toledo), and an Apollo 9000 Total Organic Carbon Analyzer (Teledyne Instruments Tekmar), respectively. Heavy metals in the overlying water of the microcosms were determined with the an inductively coupled plasma mass spectrometry (ICP-MS). In addition, the quantity of heavy metals in the FMs (plastic part) and the iron wire in nose clips were determined using ICP-MS after an acid digestion process [66].

During the sampling after two-month incubation, FMs were rinsed using 0.22 µm filtered seawater and cut into small pieces using a sterile scissor. Sediments were homogenized with the stirrer before stored at − 80 ℃ (Fig. S1).

2.2. Scanning electron microscopy

Aliquots of FMs were fixed with 2.5% (v/v) glutaraldehyde overnight and were washed three times with the phosphate-buffered saline. The washed aliquots were dehydrated in a graded series of 75% ethanol, 90% ethanol, 100% ethanol, and 100% acetone [41]. The dehydrated aliquots were dried in a fume hood, carbon coated with a sputter coater (Hitachi, Japan), and then observed under a scanning electron microscope (Hitachi, Japan).

2.3. Fourier transform infrared (FTIR) spectroscopy

Fourier Transform Infrared (FTIR) analyses were performed on FMs before and after their incubations in the microcosms. The biofilm on FMs was removed with a digestion process by FM immersions in hydrogen peroxide (30%) under 55 ℃ overnight. The pristine FMs were used as the controls. After digestion, FMs were washed with milli-Q water and dried at 55 ℃ before measurements via an FTIR spectrometer (PerkinElmer). FTIR spectra were collected using 16 scans from 4000 cm⁻¹ to 400 cm⁻¹, with a resolution of 4 cm⁻¹. The strength of characteristic absorption bands and FM identification were realized via Omnic Specta software (Thermo Fisher Scientific).

2.4. DNA extraction from the FMs and the sediments

The microbial genomic DNA was extracted from the outer layer of FMs (∼ 10 cm²) and sediments (0.25 g) using a Fast DNA Spin Kit for Soil (MP Biomedicals) according to the manufacturer’s instructions. DNA was quantified with the NanoDrop One (Thermo Scientific, Waltham, MA) and stored at − 80 ℃ until analysis. The quality of DNA extract (OD₂₆₀/OD₂₃₀ > 1.8) was checked on 1% agarose gel.

2.5. High-throughput qPCR

To investigate the profile of ARGs and MGEs in the FMs and the sediments, high-throughput qualitative PCR (HT-qPCR) was carried out with a SmartChip Real-Time PCR System (WaferGen Inc., Fremont, CA, USA) for the assessment of a total of 384 primer pairs [76], including 319 ARGs, 7 taxonomy-associated genes (including pathogens), 57 MGEs (i.e., 19 transposase genes, 12 plasmids, 3 integrase genes, and 23 insertional sequences), and a 16S rRNA gene (Table S1). For each sample, three technical replicates were applied for the analyses, and only the gene detected in all three technical replicates was regarded as positive. The relative abundance of ARGs and MGEs was calculated using the marker gene of 16S rRNA as the standard [58].

2.6. Illumina sequencing and data processing

Polymerase chain reaction (PCR) amplification was done to target the V4–5 region of the 16S rRNA gene using universal small subunit ribosomal RNA (SSU rRNA) primers (515Y, 5′-GTGYCAGCMGCCGCGGTAA-3′ 926 R, 5′-CCGYCAATTYMTTTRAGTTT-3′) which has been shown to be well-suited for marine samples [43]. Illumina NovaSeq sequencing was performed (Magigene Biotechnology, China) for 27 samples, corresponding to the nine FM samples, nine sediment samples after FM exposure, and nine sediment samples without FM exposure. All the SSU rRNA data are available in the NCBI SRA repository (accession number PRJNA863402).

Processing of SSU rRNA sequences was performed using the DADA2 R package [11]. The 16S and amplicon sequence variants (ASVs) were assigned against the SILVA 128 database [48]. The ASVs corresponding to eukaryotes, archaea, chloroplast, and mitochondria were removed using the phyloseq R package [39]. Alpha-diversity calculation and histogram creation were performed with the online server of the MicrobiomeAnalyst [14]. The potential pathogens were identified using the high-quality sequences to blast with two pathogen databases (fish and human) with the strict criteria of sequence identity > 99%, matched alignment length > 374 bp and E-value < 1 × 10⁻¹⁰, which was detailed in previous studies [21], [77]. The potential bacterial metabolic functions in FMs and sediments were predicted on the basis of their 16S rRNA gene sequences using PICRUSt2 [10].

2.7. Statistical analyses

An unweighted-pair group method with arithmetic (UPGMA) dendrogram based on the Bray-Curtis similarities was used for visualization of beta diversity. A similarity profile test (SIMPROF, PRIMER 6) was performed on the null hypothesis that a specific sub-cluster can be recreated by permuting the entry species and samples. The significant branch was used as a prerequisite for defining a bacterial cluster. The difference in the microbial community composition was tested using the permutational multivariate analysis of variance (PERMANOVA) [3] together with the verification of homogeneity of variances using the vegan R package. The difference of bacterial taxa at the species or the class level was tested using the DESeq2 R package [36]. The null model analysis was performed to assess the assembly mechanisms of the bacteria community, which were elaborated from a previous study [59].

The difference in the profiles of ARGs, MGEs, or their sub-categories was tested using the student’s t-test. The correspondences between ARGs and MGEs were tested using the linear regression analysis, and the graphs were generated using the Graphpad Prism 9. At the gene level for the ARGs and MGEs, student’s t-tests were also adopted and further adjusted with the Benjamini-Hochberg method to reduce the false discovery rate (FDR) at a 5% level of significance using the STAMP software.

Mantel tests based on the Spearman correlation, the Procrustes analysis, redundancy analysis, and variance partitioning analysis (VPA) were performed using the vegan R package. Furthermore, The relationship between bacterial taxa and ARG profile was tested with the Spearman correlation using the psych R package [49]. Significant results (Rho > 0.8 and p < 0.01) were further visualized using the Gephi 0.9.5 [7].

3. Results

3.1. Physicochemical properties

Metals in the FMs were determined for the plastic part (without nose clips) and metal part (iron wires from the nose clips) after an acid digestion process, including silver, zinc, iron, lead, copper and chromium. The copper and zinc had a higher proportion (3.3 and 0.9 μg/kg) in FMs (without nose clips), followed by chromium, silver and lead, ranging from 0.03 to 0.69 μg/kg. Besides, all metals mentioned above were detected in the iron wires of nose clips, iron accounted for 92.5% of the total mass, whilst the rest of the heavy metals was less than 1% (Table S2).

FMs were exposed to sediments collected from the coast, including the beach (designated as “S1″), the aquafarm (designated as “S2″), and the mangrove (designated as “S3″). After the FM exposure, the physiochemical properties of the overlying water were determined for the treatment and the control groups (without FM exposure). The salinity of overlying water for the three sites was 33.3 (beach), 14.7 (aquafarm), and 27.8 PSU (mangrove), respectively (Table S3). The pH of the three sites was 8.1, 8.2, and 8.5, respectively. Interestingly, there was a significant increase in pH (0.1 unit) (t-test, p < 0.01), and a significant increase in the copper concentration (0.3 μg/L) (t-test, p < 0.01) for the beach site after FM exposure when compared to the control group. No significant differences were found for the following parameters after the pairwise comparisons, including dissolved oxygen (DO), oxidation-reduction potential (ORP), total organic carbon (TOC), and zinc concentration (Table S3). Besides, a corrosion behavior was found in the nose clips after FM exposure (Fig. S1).

After two months of incubation, scanning electron microscopy showed that the diameter of the outer layer of FM fibers was around 20 µm and also revealed a large diversity of morphologies comprised of the mainly rod-shaped bacteria-like structure patchily distributed on the FM for all sites (S1, S2, and S3) ( Fig. 1). After the removal of biofilm with hydrogen peroxide, the chemical composition of FM was identified using the FTIR, showing that polypropylene was the component of the three layers of FMs. The absence of a characteristic carbonyl band around 1720 cm⁻¹ indicated that FM did not undergo detectable oxidation/degradation with this technique (Fig. S2).

Fig. 1.

Microbial colonization on FMs (outer layer) observed under scanning electron microscopy after two months. Panel a and b: pristine FMs. Panel c and d: FMs exposed to the sediments of the beach (S1). Panel e and f: FMs exposed to the sediments of the aquafarm (S2). Panel g and h: FMs exposed to the sediments of the mangrove (S3). Arrows in the figure represented bacteria-like structures.

3.2. Microbial community structure

Illumina NovaSeq DNA sequencing generated 1 099 525 sequence tags, falling into 5622 ASVs after randomly resampling to the lowest number of sequencing tags of 26624 to provide statistical robustness when comparing diversity measurements among samples. The dendrogram analysis showed that the sampling sites had a strong effect on the microbial community. A strong dissimilarity (> 98%) was observed for samples between the mangrove (S3) and the other two sites characterized by sandy sediments (the S1-beach and the S2-aquafarm) ( Fig. 2). Furthermore, the cluster constituted by the two sandy sediment samples was separated from the cluster of FMs, showing a dissimilarity of 95% between them. The PERMANOVA results confirmed the significant differences among sampling sites (S1, S2, and S3, p < 0.01) or sampling types (plastisphere and sediments, p < 0.01). It was noticeable that the FMs altered the microbial community of the sediments. Indeed, the microbial community of the sediments formed a separate cluster with FM exposure, compared to that without FM exposure in the aquafarm site (SE-S2 versus SEC-S2) (SIMPROF test, p < 0.01) (Fig. 2).

Fig. 2.

Comparison of taxonomic abundances and community structure of bacteria on the biofilm of FMs and sediments by cumulative bar charts comparing relative abundances (left) and by UPGMA dendrogram based on Bray–Curtis dissimilarities between 16S rRNA-based sequencing profiles (right). The sediments were collected from three sites: a beach (S1), an aquafarm (S2), and a mangrove (S3). Each sample type was performed in triplicates. FM represented the bacterial community from the plastisphere of FMs. SE represented the bacterial community with FM exposure, and SEC represented the bacterial community of the sediments without FM exposure.

The differences between the FMs and sediments were also revealed in the alpha diversity indices. The indices of richness (Chao1) and diversity (Shannon) in the FMs and sediments were revealed to be significantly lower than that in the sediments (t-test, p < 0.05) (Fig. S3). For the sandy sediments (the beach and the aquafarm), the value of the diversity was prone to be lower after the FM exposure, while the results were not significant (Fig. S3).

Taxonomic analyses confirmed the specificity of the community structures in terms of sampling sites. The microbial community composition in the mangrove was dominated by the Deltaproteobacteria (64.2 ± 11.3%, 28.5 ± 2.1%, and 26.6 ± 3.4% for the plastisphere of FMs, and for sediment samples with and without FM treatments, respectively). In contrast, the microbial community composition from sandy sediments (the beach and the aquafarm) was dominated by the Gammaproteobacteria, accounting for 36.6 ± 1.2%, 64.7 ± 12.5% for the plastisphere, 28.6 ± 2.7%, 33.9 ± 1.2% for the sediment samples with FM treatments, and 22.4 ± 6.7%, 38.1 ± 3.2% for the sediment samples without FM treatments, respectively (Fig. 2). The difference in class level was indeed significant for specific taxa. For example, Gammaproteobacteria and Alphaproteobacteria had a higher relative abundance than their surrounding sediments for each site (FDR < 0.05). Besides, it showed that the Alphaproteobacteria were more abundant in the sediment samples after FM exposure than that without FM exposure for the beach site (FDR < 0.05), indicating that FMs can alter the microbial community of sediments.

We further curated in particular 36 bacterial ASVs selected from the top 5 ASVs in each sample type (Fig. S4). Among these ASVs, the abundance of Alcanivorax sp. was significantly higher in the plastisphere than that in surrounding sediments (the beach and the aquafarm) (FDR < 0.05). Specific enrichment in the plastisphere included ASVs of Gammaproteobacteria in the beach site, compared to the Youngimonas Vesicularis, Marinobacter mobilis, Sponggibacter marinus, and Ruegeria sp. in the aquafarm site, and Pesudomonas sp., ASVs of Deltaproteobacteria in the mangrove site. In addition, a difference was observed for the sediment sample when comparing the microbial community with and without FM treatments, Polycyclovorans sp. showed significantly higher abundance in the sediment after FM treatment (FDR < 0.05). Interestingly, no difference in species level could be observed for the mangrove sediments after FM treatment (Fig. S4).

Null model analysis was further performed to illustrate the bacterial assembly process by the assessment of the relative importance of two types of bacterial community assembly (i.e., deterministic and stochastic). The results of the beta nearest taxon index (β-NTI) were 0.015 and 0.006 for the samples of plastisphere and sediments ( Fig. 3a), respectively, showing a slight difference between these two sample types (t-test, p < 0.05). Since the β-NTI of both sample results were in the range between − 2 and 2, the stochastic process was the main force influencing bacterial community assembly. The relative importance of different community assembly processes was further clarified, showing that dispersal limitation, followed by the undominated process, was the main process for both the plastisphere and the sediments. In comparison, the result of the plastisphere had a higher value (Fig. 3b).

Fig. 3.

The result of beta nearest taxon index (β-NTI) (a) and the relative importance of different community assembly processes (b). The FMs represent all samples of the plastisphere of FMs, and the Sediments represent all sediment samples from the three sites.

3.3. Occurrence of potential bacterial pathogens

The presence of human and fish pathogens was investigated in this study from the Illumina sequencing data. Twenty-two out of 5622 ASVs were identified as potential pathogens, for which 6 and 8 pathogens targeted humans and fish, respectively, and the rest of 8 pathogens targeted both humans and fish. Ten pathogens were colonized on FMs and undetected in the sediment samples ( Fig. 4). Generally, there was a higher prevalence of pathogens in aquafarm than the other two sites. It is noteworthy that the relative percentage of total pathogens in an FM sample reached 9.81% of total ASVs in aquafarm when those in the corresponding sediment samples were no more than 2.99%.

Fig. 4.

The relative abundance of the potential pathogens in each sample shown by the heatmap. Each cell represented the average value of the triplicates for a sample type (Please refer to Fig. 2 for the sample type labeled in the left part of this figure). The legend of FMs in the right part of the figure indicated pathogens were only detected in FMs and not in sediments. Asterisks corresponding to the FM sample marked in the figure indicated a significant difference for the FM sample compared to their surrounding sediments, and an asterisk corresponding to the SE sample marked in the figure indicated a significant difference for the sediment after FM exposure compared to sediment samples without FM exposure.

The average percentage of pathogens was 2.22% and 0.92%, respectively, for the bacterial communities in all plastisphere and sediment samples, respectively, and it was significantly higher in the mangrove site (t-test, p < 0.05). Specifically, several pathogens were significantly enriched in the plastisphere compared to their surrounding sediments. For example, Acinetobacter baumannii was found to be more abundant in the plastisphere of the beach site, accounting for 0.31% of the relative abundance, whereas Pseudomonas stutzeri (1.06%) was significantly higher in the aquafarm. In terms of mangroves, Priestia megaterium (0.10%) was enriched in the plastisphere compared to that in the surrounding environments (t-test, p < 0.05) (Fig. 4).

When comparing the potential pathogens in the sediments after FM exposure, the Pseudomonas oleovorans (2.13%) significantly increased in the aquafarm (t-test, p < 0.05), compared to the sediment sample without FM exposure (0.99%). Besides, our HT-qPCR targeted several deleterious pathogens and found a significant increase for Acinetobacter baumannii (0.1%) in the sediments of beach and aquafarm sites after FM exposure (t-test, p < 0.0001) (Table S1).

3.4. Dynamics of antibiotic resistome

Samples were selected to examine the presence of ARGs and MGEs in the FMs or the sediments, including the samples from the beach and the aquafarm, corresponding to 6 FM samples (triplicates for each site), 6 sediment samples after FM exposure, and 6 control sediment samples without FM exposure. A total of 76 genes were found in the plastisphere, including 64 ARGs and 12 MGEs, which is lower than that in the sediment (110 genes), including 88 ARGs and 22 MGEs. In addition, 25 ARGs and 2 MGEs were specifically detected in the FMs, indicative of a selective enrichment (Fig. S5). These detected ARGs conferred their resistance in nine major classes of antibiotics, including aminoglycosides, beta-lactamase, fluoroquinolones, macrolide-lincosamide-streptogramin B (MLSB), multidrug, phenicol, vancomycin, tetracycline, and other resistant genes.

Significant enrichments in the relative abundance of total ARGs were observed in the plastisphere compared to the sediment samples (t-test, p < 0.01), with a 5-time difference between them. In detail, the relative abundance of tetracycline-resistant genes had increased by 5 times, whereas multidrug, MLSB, and phenicol-resistant genes had increased by at least 10 times (t-test, p < 0.05). The relative abundance of total ARGs in the plastisphere of the aquafarm was significantly higher than that of the beach (t-test, p < 0.05). In contrast, the relative abundance of total MGEs in the plastisphere was less than half of that in the sediments (t-test, p < 0.05) ( Fig. 5a, b). These results were also confirmed at the gene level. Indeed, czcA (multidrug), tetA(p) (tetracycline)， and VatA (MLSB) were significantly enriched in the plastisphere of the FMs. On the contrary, AAC(3)-via (aminoglycoside), IS6100 (insertional), trb-C (plasmid), and tnpA-2 (transposase) were more abundant in the sediment than that in the FMs (Fig. 5c).

Fig. 5.

Profiles of ARGs and MGEs in the plastisphere and sediments. The relative abundance of the ARGs (a) and MGEs (b) was illustrated in each sample type, error bars indicated the standard deviation of the triplicates. Comparisons were performed for each gene between the plastisphere and sediments using the student’s t-test, and the p-value was adjusted with the Benjamini-Hochberg method (c). FMs represented the ARGs from the plastisphere of FMs. Sediments represented the ARGs from the sediments. The correlation between ARGs and MGEs was depicted by linear regression (d) and the network analysis (e) for the beach samples.

We further explored the effect of FMs on the ARG and MGE profiles in sediments. The relative abundance of MGEs increased by approximately 2 times after FM exposure (p < 0.01), but not for that of ARGs (Fig. 5b). The correlation of the total abundance between ARGs and MGEs was analyzed for the site of beach and aquafarm separately, and we found that these were significantly correlated for the beach site (S1), resulting in a linear pattern between them (R² =0.71, p < 0.05) (Fig. 5d). A network analysis was further conducted to explore the co-occurrence between the ARGs and the MGEs (Spearman rho > 0.8, p < 0.01). Gene tnpA-3 (transposon) was highly related to gene ceoA (multidrug) (Fig. 5e). Besides, strong correlations were also found for the gene TN5403 (transpose) versus vanTE (glycopeptide), IS630 (transpose) versus czcA (multidrug), IS630 (transpose) versus OXY-1–1 (beta-lactamase), IS1247 (insertional) versus two genes of aminoglycoside (aadA 17 and aadA 21), as well as ISPps1-pseud (insertional) versus the other two genes of aminoglycoside (aadA and ANT(4′)).

3.5. Correlation between ARGs and microbial communities

The correlation between ARGs and microbial communities was performed to understand the microbial community's contribution to the ARGs' variance. The Mantel statistic test showed a significant correlation between the microbial communities and the ARG profiles based on the Bray−Curtis dissimilarity (r = 0.8091, p = 0.001, 999 permutations). Furthermore, the Procrustes analysis revealed that the ARG and bacterial 16S rRNA profiles were well clustered by exhibiting a goodness-of-fit test (sum of squares M² = 0.4738, r = 0.7254, p < 0.0001, and 9999 permutations) based on Bray−Curtis dissimilarity metrics ( Fig. 6a). Redundancy analysis was performed by applying the dominant bacterial phyla and nine categories of ARGs to understand the responses of the ARGs triggered by the microbial community. The first two axes explained 47.35% of the ARG variation in plastisphere and sediment samples. Proteobacteria, Planctomycetes, and Chloroflexi were positively correlated with the ARG composition (Fig. 6b). By setting up the explanatory variables, variance partitioning analysis (VPA) separated the impact of the bacterial community and MGEs. It showed that the bacterial community contributed the highest (60%) in the variance of the ARGs, followed by the joint effect between the bacterial communities and MGEs (8%), and the MGEs (5%) (Fig. 6c).

Fig. 6.

Correlation between the resistome and microbial communities. (a) Procrustes analysis showing the correlation between the ARGs and the bacterial communities based on Bray−Curtis dissimilarity metrics (sum of squares M² = 0.4738, r = 0.7254, p < 0.0001, 9999 permutations). (b) Redundancy analysis (RDA) on the correlation between bacterial community composition and the nine classes of ARGs. (c) Variation partitioning analysis of ARG profiles explained by the bacterial community, MGE profiles, and their joint effects. (d) Network analysis depicting the significant correlation between the resistome and the bacterial community (Spearman rho > 0.8, p < 0.01).

To uncover potential hosts of the ARGs, co-occurrences between the bacterial communities and the ARG profiles were performed for ASVs with a prevalence of over 50%. Fifty bacterial taxa at the genus level and 137 ARGs (together with MGEs) were used for the analyses. It turned out that 20 genera and 37 ARGs showed a strong correlation (Spearman rho > 0.8, p < 0.01) (Table S4-5). Alcanivorax, Youngimonas, and Marinobacter (belonging to the Proteobacteria) were determined as network hubs, which had enriched in the FMs (Fig. S4, Table S4), and correlated with several of the most abundant ARGs in FMs, including AAC(3)-IV (aminoglycoside), vanXA (glycopeptide), mel_1 (multidrug), tetA(P) (tetracycline), etc. (Fig. 6d).

4. Discussion

The contamination of FMs is an increasing concern, which inevitably intensifies oceanic plastic pollution [19]. In this study, we showed that FMs can enrich specific bacterial taxa and the ARGs in the plastisphere compared to the counterparts from their surrounding sediments. Moreover, MPs significantly increased the presence of MGEs and pathogens in the sediments after FM exposure.

4.1. Dynamics of the microbial community in FMs

The plastisphere of FMs formed a distinct microbial community compared with the sediments (PERMANOVA results, p < 0.05), which is consistent with niche partitioning theories found in seawater or soil environments [22], [77]. FMs were incubated for two months to develop a mature biofilm, according to the temporal profiles established in our previous studies [12], [29]. Our results showed the richness of the plastisphere was lower than that in the sediments, demonstrating specific taxonomic selections from the surrounding microbial organisms.

For the taxonomic composition in the beach and the aquafarm, Gammaproteobacteria were the dominant taxa in the plastisphere. In contrast, Deltaproteobacteria represented the highest relative abundance in the mangrove. It has been suggested that the microbial community in the plastisphere was influenced by the environments [6]. At the genus level, the five top enriched bacterial taxa in the plastisphere (i.e., Alcanivorax sp., Ruegeria sp., Marinobacter sp., Spongiibacter sp., and Pseudomonas sp., except for the Youngimonas sp.) (Fig. S4) were known to be related to the hydrocarbon degradation [24], [25], [26], [53], indicating microbial degradations on FMs. However, FTIR did not show any visible spectral shifts (Fig. S2), we hypothesized that these bacteria potentially participate in the degradation of plastic additives, because FTIR is not an effective technique for the characterization of plastic additive [13]. Recently, several studies had detected the existence of specific plastic additives in FMs, e.g., antioxidants and organophosphate esters [23], [33], which supported our hypotheses. These results suggested that FMs can select the hydrocarbon degraders in the plastisphere after their disposal in the marine environment.

Null model analyses were used to evaluate the bacterial assembly process. The beta nearest taxon index (β-NTI) is a standardized measure of the mean phylogenetic distance to the nearest taxon between samples [57]. FMs had a significantly higher β-NTI than that of sediments (Fig. 3), suggesting that the microbial community in the FMs was more dissimilar in a phylogenetic view. The relative importance of different community assembly processes was further clarified. Dispersal limitation was the primary process in shaping the microbial community structure for the plastisphere, which was consistent with those found in the freshwater and seawater environments [71], [72]. Our result was in agreement with a previous study showing that the dispersion limitation would be the main assembly process when the physiochemical properties among sites were heterogeneous [32]. Moreover, the dispersion limitation acting in concert with the undominated process (i.e., drift) can have a substantial influence on the community composition in sediments [56]. The results were considered reasonable because the sediments had lower regional connectivity, which decreased the probability of active dispersal. The lower migration rate might explain this to some extent [59]. These results highlighted the importance of the surrounding environments in shaping the microbial community in the plastisphere.

4.2. Enrichment of antibiotic resistant genes in FMs

ARG dissemination in coastal regions made human health face threats in these environments. Riverine runoff, wastewater treatment plants, sewage discharge, and aquaculture are responsible for the prevalence of ARGs [73]. In this study, we observed a significant enrichment of ARGs in the FMs that might serve as another dissemination source of ARGs to the marine environments, such as the FMs buried/semi-buried in the sediments of subtidal zones or the seafloor. The ARG abundance in the plastisphere of the aquafarm was significantly higher than that in the beach (p < 0.05), which might be due to the extensive use of antibiotics in aquaculture [37]. In comparison, several previous studies investigated the antibiotic resistome in the plastisphere from the marine environments and reached divergent conclusions compared to ours. For instance, exposures of microplastics and FMs to the coastal seawater did not show enrichment of ARG resistome [67,[74], [75]. In summary, it potentially clarifies that ARG profiles in the plastisphere could be environmentally specific depending on their exposure conditions (seawater versus sediment).

Mantel tests and Procrustes analysis revealed strong correspondences between the microbial community and the ARG profile. Indeed, the variance partitioning analysis showed that the microbial community explained 60% of the ARG variance in the FMs. However, profiles of MGEs only explained 5% of the ARG variance in FMs. The total abundance of the ARGs is not significantly positively correlated with the total abundance of MGEs (R² = 0.48, p = 0.12) for all FM samples analyzed (analysis not presented in the manuscript), indicating that the antibiotic resistance in the FMs could be more of an intrinsic nature for the bacteria in the environment. The resistance was less ascribed to the horizontal transfer of the ARGs but mainly to the physiological properties of the bacterial cells [17,[61], [62], [70]. A recent study also showed that stochastic (i.e., probabilistic dispersal and ecological drift) rather than deterministic (i.e., environmental filtering) is the main driver of the ARG dynamics in the plastisphere [65]. We found that several abundant genera were strongly correlated with the specific ARGs, such as Alcanivorax and Marinobacter (Fig. 6, Table S4). Previous studies proved that Alcanivorax was resistant to multiple classes of antibiotics (e.g., beta-lactam, phenicol, and tetracycline) [55], [63], and also the resistance of Marinobacter to the beta-lactam or glycopeptide [28], [54].

4.3. Impact of FMs on the surrounding environments

In this study, we further examine the impact of FMs on the surrounding environments. A significant shift in the microbial community in sediments was observed after the FM treatment compared to the control groups without FM exposure. The dendrogram showed that samples formed one cluster after the FM exposure (Fig. 2). The taxonomic composition also proved this result (e.g., the higher abundance of Alphaproteobacteria). At the species level, pathogens (e.g., Pseudomonas oleovorans and Acinetobacter baumannii) were found to be more abundant in the sediments. Studies up-to-date underlined the pathogens in the plastisphere [8]. We argue here that FMs could promote pathogen proliferation in their surrounding sediments, which could not be ignored for these FMs discarded in the coastal environments.

Specific hydrocarbon degradation bacteria became more abundant in sediments after FM exposure, e.g., the taxon of Polycyclovorans sp. at the beach site (Fig. S4). It was worth noting that Polycyclovorans sp. was a group of obligate hydrocarbonoclastic bacteria [60], suggesting that the hydrocarbons were released and degraded by the microbes in the surrounding environments after the exposure of FMs. These results were also supported by the PICRUSt2 results by predicting functional abundances based only on 16S rRNA marker gene sequences [10], showing the significant higher gene abundance related to the xenobiotic degradation in the sediment with FM exposure, compared to that without the FM exposure (Fig. S6). The shift of physiochemical property of the overlying water was consistent with this result. A significant increase in pH value was observed in the beach site (Table S3), indicative of microbial metabolisms of the hydrocarbons deriving from the plastics, as the previous study showed the classical perturbation of pH value during microbial incubation [52].

The abundance of MGEs was significantly increased in sediment samples after FM exposure. Linear regression analysis showed that the ARGs significantly correlated with the MGE profiles at the beach site (Fig. 5), indicating that the increase in the relative abundance of the ARGs could be due to the horizontal gene transfer carried by the MGEs. A previous study revealed the crucial roles of MGEs in disseminating the ARGs [44]. To decipher the potential mechanism of the increased ARG profile, heavy metals were measured from the overlying water of the microcosms, the significantly elevated copper concentration after FM exposure was consistent with the abundance profiles of the ARGs in the beach site. Previous studies underlined the co-selection between the ARG and metal resistance [5]. Heavy metals released in overlying water can be ascribed to FMs, as determined from this study (Table S1-2).

5. Conclusion

This study characterized the microbial community and the antibiotic resistome of FMs and their surrounding sediments. The microbial community in the plastisphere is significantly different from that of the surrounding sediment, indicating that FMs served as a distinct ecological niche compared to the sediments. Enrichments were found for the ARGs and pathogens in FMs and their surrounding environments. The acquisition of antibiotic resistance and pathogens in FMs might impose health risks to marine organisms and humans in coastal regions. Our results might reflect the circumstances of coastal plastic/microplastic contamination. These findings provide important insights into the understanding of threats of FMs in coastal environments, highlighting the necessity of raising awareness of environmental protection and beach cleaning in the pandemic scenario.

Environmental Implication

Face masks (FMs) are made from polypropylene and have a very low biodegradability in the marine environment. Therefore, it is regarded as a persistent pollutant. The outbreak of Covid-19 drastically increased the use of FMs, resulting in worldwide plastic contamination in coastal regions that are strongly associated with human activities. We evaluated the environmental risks of FMs from the perspective of microbial ecotoxicology. The significant enrichment of ARGs and pathogens in the plastisphere of FMs and their surrounding sediments, indicates that FMs were indeed hotspots of these biological hazards. This study suggests that coastal regions are vulnerable to plastic pollution.

CRediT authorship contribution statement

Jingguang Cheng: Conceptualization, Investigation, Formal analysis – original draft, Funding acquisition, Writing - original draft. Daochao Xing: Conceptualization, Investigation, Formal analysis, Writing – review & editing. Pu Wang: Investigation. Si Tang: Investigation. Zhonghua Cai: Conceptualization, Formal analysis, Validation, Writing – review & editing. Jin Zhou: Conceptualization, Validation, Supervision, Funding acquisition, Writing - review & editing. Xiaoshan Zhu: Conceptualization, Validation, Supervision, Funding acquisition, Writing - review & editing.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Editor: Jianhua Guo

Appendix A. Supplementary material

Supplementary material

.

Data availability

Data will be made available on request.