*3.4. Transfer Potential of ARG Types between Fish, Soil and Water Samples*

A comparative analysis was performed between samples from the gut of *C. semilaevis* (A0-A3), water (B0-B3) and soil (C0-C3) to evaluate the possibility of HGT between the different biomes. The correlation of ARGs detected from different samples showed a good cluster in the sample type (Figure 2), and samples from fish (A) were distinguished from water (B) and soil (C) (Adonis: *p*-value = 0.017; Anosim: *p*-value = 0.011). The results (Table S1 in Supplementary Materials) indicated that the average level of ARGs detected in fish samples was lower compared to the soil samples and the numbers of ARGs in soil samples were lower compared to the water samples. Relatively pristine environments (A0 and B0) revealed fewer ARGs than other fish farm (Table S1 in Supplementary Materials). Clustered A1 and A2 separated from A0; B1, 2 and 3 separated from B0 (Figure 2). The Venn diagram showing the distribution of ARGs among fish (A), water (B) and soil (C) samples was depicted (Figure 3). A total of 366 ARGs, all classes combined, was screened from the water samples, while 340 ARGs were screened from the soil samples and only 221 ARGs were obtained from the fish samples. Some individual ARGs followed this trend. For example, 62 ARGs encoding for resistances to aminoglycosides were found in water samples, while only 46 and 30 ARGs were recorded for the soil and fish samples, respectively. In addition, 10 ARGs encoding for resistance to chloramphenicol were obtained from water samples, while 6 and 1 from the soil and fish samples, respectively. The same trend was observed for ARGs encoding for tetracycline, as well as mobile genetic elements (MGEs). CIntI-1 (class1), IS613, Tp614, intI-1 (clinic), tnpA01, tnpA02, tnpA03, tnpA04, tnpA05 and tnpA07 were detected, and cIntI-1 (class1), intI-1 (clinic) tnpA04 and tnpA07 were the most common MGEs among the three sample types. Nevertheless, some disparities were observed. ARGs encoding for β-lactam, multidrug and sulfonamide were more abundant in soil samples than the water and fish samples, in this order. Additionally, the MLSB-associated ARGs were uniformly distributed in the water (32), soil (42) and fish (42) samples.

**Figure 2.** Correlation of ARGs between different samples: detected correlation of ARGs between different samples from the gut of *C. semilaevis* (A0–A3), water samples (B0–B3) and soil samples (C0–C3) to evaluate the possibility of HGT between the different biomes.

MGEs were in direct proportion with ARGs, and the integrase genes and transposases were in the same trend (Figure S2 in Supplementary Materials). The Spearman correlation analysis among ARGs indicated a significant correlation among all the ARGs across the studied samples (Figure S3 in Supplementary Materials). The correlation coefficient R was > 0.7, and the *p*-value was < 0.05. The integrase gene intI-1 (clinic) was significantly and positively correlated with ARGs related to multidrug, MLSB and tetracycline resistance., while the other integrase gene cIntI-1 (class1) was significantly and positively correlated with ARGs related to aminoglycosides and tnpA07. Transposases exhibited more correlations with ARGs. IS613 was significantly and positively correlated with vancomycin ARGs, while Tp614 with MLSB, beta-lactamase, aminoglycoside, multidrug and other ARGs were not. TnpA01 was significantly and positively correlated with MLSB, while tnpA02 mainly with tetracycline and beta-lactamase, tnpA03 with MLSB, multidrug, vancomycin and aminoglycosides and tnpA07 were significantly and positively correlated with aminoglycoside ARGs. TnpA04 and tnpA05 interacted more tightly with ARGs related to tetracycline and beta-lactamase.

**Figure 3.** Venn diagram indicating the abundance of shared and dissimilar ARGs in the three types of samples: Three types of samples were fish samples (**A**), water samples (**B**) and soil samples (**C**). There were 366 ARGs in total, 74 of them shared in the three types of samples, 23 of them shared between fish samples and water samples, 46 of them shared between water samples and soil samples and 13 of them shared between fish samples and soil samples.

## **4. Discussion**

Studies have shown that approximately 30% to 90% of used antibiotics promote the production and spread of ARGs [23], but the abundance and dissemination of ARGs in the environment and their mechanisms are far from being completely elucidated. The present study aimed to explore the distribution of ARGs in aquatic microecosystems. Eight classes of ARGs, namely, aminoglycoside, multidrug, beta-lactamase, MLSB, chloramphenicol, sulfonamide, tetracycline and vancomycin ARGs, which are all common antibiotics, were those predominantly abundant in the aquatic environment. The abundance of these ARGs differed according to the microsystem and sampling site, mostly abundant in the soils, indicating an accumulation in the sediments. Regardless of the type of samples, the mechanisms adopted by the ARGs were the same and were represented by antibiotic deactivation, cellular protection and efflux pump. These results were the first look into the distribution and spread of ARGs related to the *C. semilaevis* breeding industry.

The results of the high-throughput analysis of the resistome revealed that a high level of ARGs could be found in the soil, water and intestinal tract of *C. semilaevis*. These results signposted that fish farms constitute a great reservoir of ARGs, conforming with previous research demonstrating that fish-rearing ponds, lakes, bays [24–26] and shrimp ponds [16,17] were all enriched in ARGs. In addition, the bacterial community structures also differed, which might affect the composition of ARGs [27–29]. The aminoglycoside-related ARGs were more abundant in the water samples, and this was corroborated by previous findings indicating that aminoglycoside ARGs are abundant and persistent in wastewater plants [9] and drinking waters [30]. The presence of aminoglycoside ARGs in the fish intestinal tract and soil samples was also supported by studies conveying the persistence of aminoglycoside ARGs in the gut of fish and zoo animals [31] and plenty of studies reporting the presence of aminoglycoside ARGs in the soils of diverse types [6,12,13,26,32]. Our findings also indicated that β-lactam ARGs were enriched in soil and water samples, but were found in a lesser extent in the fish intestinal tract. This observation is supported by previous works in the Tibetan environment, sewage sludge field application soils, agricultural soils

and parks [9,30,33–37]. The abundance of chloramphenicol ARGs was relatively low in our samples, but its occurrence was in line with previous works which indicated the presence of chloramphenicol ARGs in feedlots, Baltic sea fish farm soils, water effluents and pristine remote Antarctic soils [6,8,32,38]. MLSB-related ARGs were uniformly distributed in the fish intestinal tract, water and soil samples. Studies reporting MLSB ARGs in fish intestinal tract are limited, but these ARGs have been reported in water effluents and soils treated with sewage sludge [8,26]. A particular enrichment of multidrug ARGs was recorded in the soil samples, followed by water and fish intestinal tract samples. These ARGs have been also reported in soil and water samples by other researchers [6,16,26,30,36,39]. Similarly, sulfonamide, tetracycline and vancomycin-resistant ARGs were all found in the three types of samples, especially tetracycline, which corroborated previous findings [6,8,30,40,41]. Our study revealed a diversity of ARGs with different degrees of abundance in water, the intestinal tract of *C. semilaevis* and soil in the aquatic settings and adjacent natural sea areas.

Disparities were found in ARGs detected from the same type of samples collected from diverse geographical locations. The level of ARGs was lower in the samples from natural sources. In daily production, antibiotics are usually applied through the splashing of the factory culture ponds in the fish farm, mainly florfenicol, quinolones (oxoquine), streptomycin, neomycin (neomycin sulfate and tetracycline), doxycycline hydrochloride, sulfonamide antibiotics (sulfadiazine and sulfamethoxazole) and enrofloxacin, related to the ARGs detected here. A frequent outside release of the culture water may decrease the concentration of antibiotics and ARB in it, which could explain the decreased level of ARGs in the water samples. The discharge of effluent water in the soil may also explain the increased accumulation of ARGs in soil samples. Indeed, previous studies indicated that the effluent water could be released into the natural world, infiltrating into the ground or being dumped into the sea, which leads to the accumulation of antibiotic-resistant bacteria in the soil; thus, causing the enrichment of soil in ARGs, which is equivalent to ARG transfer and enrichment from water to soil [29]. From the geographical map of our sampling, the three fish farms were all near the sea and the water for fish farming came from the sea, a large part of which would flow back in. Therefore, the transfer of ARGs may occur from the coastal fish farm to the sea (including natural seawater, sea mud and marine organisms such as fish species). Accordingly, inflow water was also proved to be a major source of trout farming contamination with salmonella and multidrug-resistant bacteria [27]. Thus, a cycle accumulation pattern seems to exist.

MGEs involved in the mobility of ARGs are an important factor [33,36,42,43] and were detected in all three types of samples, such as intI-1 (clinic) tnpA04 and tnpA07. Wild fish in a peri-urban river might be recipients and communicators of ARGs in the water environment and ARGs might transfer between fish and water using bacteria as a spreading medium, with nine ARGs and one MGE continuously shared in both sample types [44]. Notably, a differentially ecological niche of the host resulted in the various resistomes of the lower vertebrate, among which the resistomes of seawater fish shared more similarities and were characterized by a high abundance, distinct from other observed habitats, and ARGs in seawater fish were colocalized with mobile elements with a high frequency, suggesting that they were likely spread through HGT [45]. On the whole, these data revealed the significant diversity and heterogeneity of ARGs in lower vertebrates, indicating that these wild species potentially play an essential role in the global spread of ARGs [44,45].

The HGT of ARGs among environmental bacteria in different taxa is proved to be an important pathway that disseminates resistance and, subsequently, acquires resistance through human pathogens and commensals. Since transposition-related genes and integrases were found in the studied samples, HGT might occur between bacteria by transposition or integration in the aquatic environment. This threes in corroboration with previous reports, confirming that ARGs are introduced by DNA transposition into plasmid and subsequent drug selection in the aquaculture sites [46]. Further studies should be carried out to verify these HGT mechanisms.

In this study, mechanisms driven by ARGs were mainly antibiotic deactivation, cellular protection and efflux pump. Efflux pump is the major mechanism of multidrug resistance in bacteria and acts by interfering with the synthesis of nucleic acid or the inhibition of protein synthesis [6,13,17,47]. Antibiotic deactivation has also been reported in previous studies on urban park soils with reclaimed water irrigation [2,13]. Cellular protection has been proved as a mechanism driven by ARGs screened from farmed fish feces [6], which corroborated our results. Antibiotic deactivation was the most abundant ARG mechanism in the fish intestinal extract, water and soil samples, which indicated that the deactivation of antibiotics may be the main mechanism adopted by resistant bacteria in the aquatic environment. The three mechanisms were also reported by Muziasari [6]. Thus, finding ways to counteract this mechanism could be the key to overcome antibiotic resistance, especially for animals left in these environments.

This study investigated the distributions, diversity, abundance and potential transfer mechanism of ARGs in *C. semilaevis*-related samples; however, these results still need to be verified by a larger-scale sample test. This work also outlined the crisis of ARG transmission from fish farms to adjacent sea areas.

#### **5. Conclusions**

Our study indicated that there were abundant ARGs in the aquatic ecosystem and that there was a transfer of these ARGs into the water, which, subsequently, led to the serious enrichment of aquatic sediments in ARGs. This study allowed not only the knowledge of ARGs in the *C. semilaevis*-related niche, but also the understanding of ARG dissemination in aquatic ecosystems. Therefore, the investigation and in-depth study of the distribution and spread of ARGs could allow us to control the spread of ARGs in the environment and minimize the risk to animal and human health.

**Supplementary Materials:** The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/w14060938/s1. Figure S1: distribution and abundance of detected ARGs, Figure S2: correlation between trends of ARGs and MEGs in three types of samples, Figure S3: correlations among the detected ARGs and MEGs, Table S1: Abundance and distribution of ARG types in different samples. Table S2: The information of primers used in the study.

**Author Contributions:** Conceptualization, B.Z.; methodology, H.L.; validation, Q.D. and H.L.; formal analysis, L.J.; investigation, N.Z.; data curation, N.Z.; writing—review and editing, L.J.; project administration, C.Z.; funding acquisition, L.J. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by grants from the Tianjin Natural Science Foundation (17JC-QNJC15000), China Agriculture Research System of MOF and MARA (CARS-47-Z01) and the modern industrial technology system in Tianjin ITTMRS2022001.

**Institutional Review Board Statement:** The animal study protocol was approved by the Institutional Review Board of Tianjin Fisheries Research Institute (protocol code 2019003 and date of approval is 2 April 2019).

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors would like to acknowledge the support of the client project team and their collaborative effort with the wider project team, which included: the Tianjin Fisheries Research Institute; Guangdong Ocean University; Southern Marine Science and Engineering Guangdong Laboratory-Zhanjiang.

**Conflicts of Interest:** The authors declare no conflict of interest.
