Environmental Reservoirs of Antibiotic Resistance in Commercial Catfish Ponds

Primary Investigator: Taylor Heckman, Ph.D., Assistant Professor, Mississippi State University

Co-Investigator: Divya Rose, Ph.D. Candidate, Mississippi State University

1. Overview

Antimicrobial resistance (AMR) in aquaculture presents a critical one health challenge, exacerbated by the open nature of fish rearing systems and intensification of modern production practices. In the United States (US), the leading aquaculture industry is catfish farming in warmwater earthen ponds. Antibiotic feeds remain an important tool for reducing animal and economic losses from catfish bacterial diseases, but their use increases risk for selection, amplification and transfer of antimicrobial resistance genes (ARG) between pathogens and environmental bacteria living in the culture ponds. It is thought that pond bacteria maintain ARG and act as reservoirs to perpetuate resistance over multiple production cycles, but this hypothesis has not been fully evaluated.

This project aims to investigate the environmental dimensions of AMR in catfish aquaculture and the influence of antibiotic use on the pond community structure and ARG resistome (Figure 1). We will work with two commercial catfish farms, one that routinely uses antibiotics and one that relies on non-antibiotic methods. Water samples will be collected in late fall, after peak disease season but prior to cessation of feeding for winter. Sample processing, whole-genome sequencing and data analysis will proceed over the next six months. The project will be carried out at the National Warmwater Aquaculture Center within the Mississippi State University Delta Research and Extension Center, which is fully equipped for standard molecular and microbiological procedures and for in-house long-read sequencing. Experimentation and training will be completed in compliance with the MSU Institutional Biosafety Committee and MSU Policies and Procedures in Ethics and Research. Sequencing data will be made publicly available on online repositories and conclusions disseminated to farmers, stakeholders, community members and scientists through regular updates to ResearchHub and by formal research and extension seminars and presentations.

Outcomes from this study will provide valuable insights into the diversity of antibiotic-resistant microbial communities inhabiting catfish aquaculture systems and provide information on whether farm-specific factors or repeated antibiotic use influences their composition. The research will identify resistant bacterial populations and trace mobile genetic elements carrying ARG that may facilitate AMR spread. Together, these findings will improve antibiotic stewardship in aquaculture by enhancing surveillance of resistance, informing and improving management practices, and laying a foundation for further research into the molecular mechanisms of AMR transfer.

Figure 1:Graphical overview of the study design characterizing the microbial diversity and resistance profiles between antibiotic-treated and untreated catfish ponds at two commercial Mississippi farms. Culturable and unculturable microbiomes will be compared by 16S rDNA sequencing and antimicrobial resistance genes (ARG) by whole genome sequencing of selected colonies using Oxford Nanopore Technologies (ONT) platforms

2. Introduction

Antimicrobials were first defined as microbial substances that inhibit or destroy other microbes, with antimicrobial resistance (AMR) understood as the antithetical ability of microbes to resist the effects of such substances^1,2. This duality arose as part of an ancient microbial arms race, long before human intervention. For example, antimicrobial resistance genes (ARG) for β-lactamase, vancomycin and tetracycline drug classes have been found in 30000-year-old permafrost³. Humans entered the race when antibiotics (antimicrobials active against bacteria) were adopted for clinical purposes in the 1940s and rapidly integrated into human and veterinary medicine^4,5. Unfortunately, decades of widespread antibiotic use across aquatic and terrestrial systems exerted significant selective pressure on bacterial populations, leading to the emergence and spread of drug-resistant pathogens threatening human and animal health^6,7.

In aquatic environments, fish farming sites are the most important source of AMR due to high concentrations of fish, feed, and fecal matter and the regular administration of antibiotics, chemicals and biocides to support intensive culture practices⁸. For this reason, many countries tightly regulate their use. In the United States, there are only three options approved by the Food and Drug Administration (FDA): oxytetracycline (Terramycin®), florfenicol (Aquaflor® and PAQFLOR®) and combined ormetoprim and sulfadimethoxine (Romet®)⁹. Sulfonamides and tetracyclines were among the earliest antibiotics implemented in fish health^10–12 while florfenicol is a comparatively modern addition exclusive to animal medicine¹³. Phenotypic or genetic resistance has been widely documented for all three drugs in aquatic environments, though tetracycline resistance is most commonly represented in global surveillance studies¹⁴. Emergence and transmission of multidrug-resistant (MDR) bacteria in aquaculture also presents a serious threat to the industry and to public health¹⁵.

Since the 2007 approval of florfenicol for treating enteric septicemia of catfish (ESC), there has been a concerning rise in clinical catfish cases involving MDR pathogens. Isolates demonstrate reduced susceptibility to two or all three approved antibiotics, severely limiting treatment options for these cases. Diagnostic trends from the MSU Aquatic Research and Diagnostic Laboratory show MDR pathogens typically appear in mid-to-late summer, subsequent to previous rounds of treatment. Cases diminish in winter, likely due to reduced bacterial activity and lack of selective pressure, then reemerge again the next summer. This cycle of seasonal reappearance suggests environmental reservoirs are present in catfish pond systems that sustain resistance even in the absence of ongoing antibiotic use, acting as persistent sources of transferable ARG. Ictalurid catfish farming is the most significant aquaculture sector in the US, and despite improvements in production practices and vaccination programs, antibiotics remain crucial tools to counter industry-limiting diseases. Catfish are raised in dug-out earthen ponds, and antibiotics delivered as medicated feed, which can cause leaching and accumulation of residues into the static and highly eutrophic pond systems. These conditions are permissive for the development of AMR strains within the larger autochthonous community^16,17.

Bacterial AMR is largely mediated by ARG encoding products that either reduce the accessibility or action of antibiotics on cellular targets. This may involve enzymatic inactivation, reduced drug accumulation through altered permeability or efflux pumps, or both, acting synergistically^18,19. Resistance genes arise primarily through two mechanisms: chromosomal mutations and acquisition of ARG via horizontal gene transfer. Horizontal transfer occurs through transformation, transduction, and, most prevalently, conjugation via plasmids²⁰. Mobilization of plasmid determinants of resistance can be further amplified by embedded genetic elements such as transposons, which can move genetic material by cut and paste recombination, or integrons, versatile gene acquisition systems that can play a role in capturing and expressing resistance gene cassettes^21,22.

Plasmids can therefore serve as a plastic platform for accumulation of genes and often carry more than one ARG to act against different drug classes for multi-drug resistance. This can severely complicate treatment of aquaculture diseases if exposure to one of the approved antibiotics indirectly selects for linked ARG. For example, Aeromonas spp. carrying the floR gene under florfenicol exposure also harbored tetracycline and sulfonamide resistance, limiting efficacy of all three approved aquaculture drugs²³. In addition to ARG, plasmids often carry genes linked to stress tolerance, virulence, and metabolism that may make it beneficial for bacteria to maintain the plasmid even without antibiotic selective pressure²⁴. Such cases amplify the risk for maintenance of ARG plasmids in environmental bacteria leveraging them for a competitive advantage against other cells, or for protection against natural stressors like UV-damage, pH, temperature shifts, or heavy metal toxicity These may also co-select for AMR in the absence of antibiotics²⁵.

To better understand antibiotic resistance dynamics in catfish aquaculture, this study will compare microbial communities in pond water from farms with and without histories of antibiotic use. The total microbiome will be assessed directly from water samples and compared to the culturable microbiome, as determined by plating on non-selective media and media supplemented with oxytetracycline or florfenicol to highlight taxa associated with capable of surviving selective pressure. Taxonomic analysis and genetic characterization of ARG and associated mobile genetic elements will be completed by 16S amplicon sequencing and whole genome sequencing, respectively. Ultimately, this approach will help identify non-pathogenic taxa that may act as reservoirs or vectors of transferable resistance, informing mitigation strategies and antibiotic stewardship in aquaculture.

Hypotheses:

Catfish ponds exposed to antibiotic selective pressure regardless of farm will have increased representation of similar taxa demonstrating resistance to at least one drug class.
Phenotypic antibiotic resistance will be attributable to transferrable antibiotic resistance genes shared between heterogenous genera.

Our specific objectives are to:

Compare the resistant and non-resistant community structures of resident pond bacteria across two catfish farm facilities employing different disease management strategies.
Identify antibiotic resistance gene reservoirs and associated mobile genetic elements in catfish aquaculture systems.

3. Methods

Study Design: Two commercial catfish farms located within a 30-minute radius will participate in the proposed study. Farm A uses antibiotics as needed for bacterial disease outbreaks. Farm B never uses antibiotics and relies on alternative mitigation strategies. Water samples (100 mL each) will be collected from a total of 18 fingerling ponds: six ponds on Farm A that previously received antibiotic treatment, six ponds on Farm A that did not receive antibiotic treatment, and six ponds on Farm B that did not receive any antibiotic treatment. Sampling will take place approximately two weeks following the last feeding of the season (November) prior to the cessation of feeding for the winter months. This timing is selected to minimize the potential reversion of the pond water microbial communities to their baseline state in the absence of feed and antibiotics.

Sample Preparation: Pond water samples will be transported to the research station on ice. Each sample will be plated on three types of media: (i) Trypticase Soy Agar (TSA) supplemented with 5% sheep blood, (ii) TSA with 5% sheep blood and oxytetracycline (15 µg/mL), and (iii) TSA with 5% sheep blood and florfenicol (15 µg/mL). A 50 mL aliquot of each pond water sample will be concentrated by centrifugation at 20,000 × g, and the resulting pellet will be stored at −80°C for downstream molecular analyses. The agar plates will be incubated at 28°C for 72 hours. All bacterial colonies that grow on these plates will be harvested to generate the total culturable microbiome and the culturable antibiotic-resistant microbiomes for each pond. Four types of microbial community samples will be prepared for each pond: (1) unprocessed pond water (representing the total microbial community), (2) culturable bacteria enriched on TSA (representing the total culturable microbiome), and (3 & 4) culturable bacteria enriched on TSA supplemented with oxytetracycline and florfenicol (representing antibiotic-selected culturable microbiomes). Additionally, representative resistant colonies of different morphotypes will be selected and sub-cultured for whole genome sequencing and ARG characterization. DNA extractions from all samples will be performed using the PowerSoil Pro® Kit (Qiagen, Hilden, Germany).

Whole Genome Sequencing and 16S rRNA Amplicon Sequencing: The quality and quantity of DNA will be assessed using fluorometric method (Qubit). DNA libraries will be prepared using the Native Barcoding Kit 24 v14 (SQK-NBD114.24; Oxford Nanopore) and sequenced on R10.4.1 flowcells on a GridIon at NWAC. All samples will be barcoded with adapter ligation and size selection will be performed to obtain fragments of the desired length. The entire 16S rRNA gene in the samples will be amplified using PCR with universal primers (27F_mod/1425R; Older et al. 2024), libraries prepared as above and sequenced on R10.4.1 flowcells. Sequence reads will be analyzed according to Older et al. 2024. Bacterial colonies exhibiting distinct morphotypes will be isolated from both types of medicated agar plates. These isolates will be sub-cultured and expanded, and Whole genome sequencing (WGS) will be performed using the Ligation Sequencing Kit (LSK) in combination with Native Barcoding Expansion (SQK-NBD114.24). Sequencing will be conducted on R10.4.1 flowcells, following the manufacturer specified protocol.

Bioinformatic and Statistical Analysis: Samples sizes are calculated based on the maximum number of ponds matching our study criteria on the two participating farms (channel catfish fingerlings, vaccinated with a live attenuated E. ictaluri vaccine). Alpha and beta diversity indices for the relative abundance of the microbiome community will be generated utilizing the statistical computing environment R. Dimensionality reduction will be performed via principal component analysis (PCA), and plots and visualization will be generated using the ggplot2 package in R (https://github.com/tidyverse/ggplot2). The significance of alpha diversity will be assessed through the utilization of the non-parametric Kruskal-Wallis statistical test, while the assessment of significance in beta diversity will be tested using Analysis of Similarity (ANOSIM) function on R under the vegan package (https://github.com/vegandevs/vegan). Diversity within each treatment groups but between the farms will be statistically analyzed by conducting a pairwise ANOVA test using the permanova function in R. Individual bacterial colonies will undergo genotypic characterization using a suite of bioinformatics tools. Briefly, genome annotation will be performed using the Rapid Annotation using Subsystem Technology (RAST) platform. Mobile genetic elements, conjugative elements, and antimicrobial resistance (AMR) genes will be identified through localized BLAST searches conducted within Geneious Prime®. Antimicrobial resistance profiles will be further characterized using the Resistance Gene Identifier (RGI) tool in conjunction with the Comprehensive Antibiotic Resistance Database (CARD).

Pitfalls, Problems, and Limitations: Both participating farms have already confirmed the availability of their operations for sampling and the same ponds used to generate preliminary data are currently in active production. We do not anticipate any issues collecting samples for this study. Participating farms are located near Moorhead, MS, which is approximately 30 miles from the Delta Research and Extension Center in Stoneville, MS. Working with commercial production facilities comes with the drawback of limited control over external factors like rainfall patterns, differences in farm personnel attending to each sampling unit (pond), complete data on terrestrial wildlife population, and access to proprietary information. Within the limits, the experiment is structured to minimize the influence of external variables.

Experimental Timeline: The experimental process will begin in November with a one-time pond sampling from both antibiotic-treated and non-treated ponds. Samples will be cultured on non-selective and antibiotic-supplemented TSA plates, followed by bacterial isolation and identification. DNA isolation will occur in December, with sequencing starting in January. Bioinformatics analysis of sequencing data, including identification of resistance genes and mobile elements, will take place in January and February. By March and April, analysis will focus on microbiome composition and resistance gene prevalence. Data analysis will continue into May, with a draft manuscript or report prepared by the end of the month. The final manuscript or presentation will be completed by June 30th.

Compliance with Ethical Principles: No animals will be used in this research as it is limited to environmental sampling on commercial farms. All researchers and staff will receive the appropriate training in compliance with the Mississippi State University Institutional Biosafety Committee and the MSU Policies and Procedures in Ethics and Research.

4. Pilot Data

Initial studies conducted during the summer of 2024 laid the foundation for this proposal. Our preliminary data supports the existence of transferable ARG present in environmental catfish pond bacteria but was limited to a single farm and only sequenced a subset of culturable bacteria. This follow-up study will expand upon the characterization and consequences of environmental reservoirs in commercial catfish aquaculture systems by contextualizing resistant taxa against the larger unculturable bacterial community and by including two regional commercial operations to assess farm-specific effects.

In our preliminary study, water samples were collected from production ponds at a commercial catfish farm with a known history of antibiotic use. At the time of sampling, some ponds had recently been treated with florfenicol or oxytetracycline, while others had not. To isolate resistant bacterial populations, pond water samples were plated on non-selective TSA and TSA supplemented with florfenicol or oxytetracycline. Morphotypes that grew on these antibiotic-containing media were sub-cultured, expanded in broth, and sequenced using ONT platforms. Lawns from the antibiotic plates were also harvested and subjected to whole-genome shotgun metagenomic sequencing using ONT’s adaptive sampling protocol, which enriched for genes associated with antibiotic resistance, plasmid mobilization, and conjugal transfer. Our analyses consistently identified resistance genes including tetA, tetD, tetR, floR, and sul2 (Figure 2). Notably, these genes were detected in ponds regardless of recent antibiotic administration, suggesting resistance determinants persist in aquaculture environments even without direct selection pressure. Many of these genes were located on mobile plasmids harbored by non-pathogenic environmental bacteria, highlighting their potential role as reservoirs capable of facilitating horizontal gene transfer.

Our choice of sequencing method is also supported by preliminary research performed by our larger collaborative group demonstrating that using ONT to sequence the entire length of the 16S rRNA gene provides better taxonomic resolution than sequencing individual variable regions²⁶. While ONT was previously criticized for lower-quality reads and the potential for misclassification, recent advancements in ONT basecalling and chemistry have improved sequencing accuracy. These advances have opened new avenues of research, allowing high-throughput community assessments without the need for culture. This is particularly valuable for identifying and characterizing fastidious organisms, and it eliminates the need for logistically complicated sampling methods that rely on traditional culture techniques.

Figure 2: Adaptive sampling results (count tables) reveal the presence of antibiotic resistance genes (ARGs), mobile genetic elements (MGEs), and conjugative elements across the 12 ponds studied. The red rectangle highlights the two ponds treated last season, indicating that genes involved in antibiotic resistance transfer and maintenance persist despite the absence of antibiotic selection pressure. These findings suggest the need for further detailed investigation to understand the community dynamics and host bacteria associated with these elements.

5. Budget

The budget for this project includes materials specific to the study aims to maximize the number of samples processed. Standard microbiological supplies (media, petri dishes, sterile tubes etc.) will be covered by the PI’s discretionary funds.

For molecular analysis, the PowerSoil Pro® DNA Extraction Kit will be used to isolate DNA from bacterial colonies, with a fluorometer (e.g., Qubit) to assess the quality and quantity of the extracted DNA. The Native Barcoding Kit 96 v14 from Oxford Nanopore will be used to prepare DNA libraries for sequencing, and R10.4.1 flowcells will facilitate high-throughput sequencing. Bioinformatics and statistical analysis will be performed using statistical software such as R, including the ggplot package for visualizing data, and computational resources (either cloud or local servers) will support data processing and analysis.

Item	Type	Cost	# needed	Total
Antibiotics	Florfenicol (5g)	$350	1	$350
Antibiotics	Oxytetracycline (5g)	$40	1	$40
DNA Extraction Kit	Qiagen Powersoil Pro Kit (50)	$510	1	$510
Sequencing Kits	ONT Native Barcoding Kit 96 v14	$800	2	$1,600
Flowcells	R10.4.1	$500	5	$2,500

			Total	$5000

6. References

Tang, K. W. K., Millar, B. C. & Moore, J. E. Antimicrobial Resistance (AMR). Br. J. Biomed. Sci. 80, 11387 (2023).
Waksman, S. A. What is an Antibiotic or an Antibiotic Substance? Mycologia 39, 565–569 (1947).
D’Costa, V. M. et al. Antibiotic resistance is ancient. Nature 477, 457–461 (2011).
Aminov, R. I. A Brief History of the Antibiotic Era: Lessons Learned and Challenges for the Future. Front Microbiol 1, 134 (2010).
Cherian, T., Ragavendran, C., Vijayan, S., Kurien, S. & Peijnenburg, W. J. G. M. A review on the fate, human health and environmental impacts, as well as regulation of antibiotics used in aquaculture. Environmental Advances 13, 100411 (2023).
Selvarajan, R., Obize, C., Sibanda, T., Abia, A. L. K. & Long, H. Evolution and Emergence of Antibiotic Resistance in Given Ecosystems: Possible Strategies for Addressing the Challenge of Antibiotic Resistance. Antibiotics (Basel) 12, 28 (2022).
Despotovic, M., de Nies, L., Busi, S. B. & Wilmes, P. Reservoirs of antimicrobial resistance in the context of One Health. Current Opinion in Microbiology 73, 102291 (2023).
Hossain, A. et al. Antibiotics, antibiotic-resistant bacteria, and resistance genes in aquaculture: risks, current concern, and future thinking. Environ Sci Pollut Res Int 29, 11054–11075 (2022).
Medicine, C. for V. Approved Aquaculture Drugs. FDA (2025).
Nelson, M. L. & Levy, S. B. The history of the tetracyclines. Annals of the New York Academy of Sciences 1241, 17–32 (2011).
Gutsell, J. S. Sulfa drugs and the treatment of furunculosis in trout. Science 104, 85 (1946).
McCarthy, D. H., Stevenson, J. P. & Salsbury, A. W. Combined in-vitro activity of trimethoprim and sulphonamides on fish-pathogenic bacteria. Aquaculture 3, 87–91 (1974).
Trif, E. et al. Old Antibiotics Can Learn New Ways: A Systematic Review of Florfenicol Use in Veterinary Medicine and Future Perspectives Using Nanotechnology. Animals (Basel) 13, 1695 (2023).
Heckman, T. I. et al. Florfenicol- and erythromycin-medicated feeds are similarly efficacious in reducing mortality from Lactococcus petauri infections in Rainbow Trout. Journal of Aquatic Animal Health 37, 110–121 (2025).
An, T. et al. Prevalence and transmission risk of colistin and multidrug resistance in long-distance coastal aquaculture. ISME COMMUN. 3, 115 (2023).
Burridge, L., Weis, J. S., Cabello, F., Pizarro, J. & Bostick, K. Chemical use in salmon aquaculture: A review of current practices and possible environmental effects. Aquaculture 306, 7–23 (2010).
Liu, X., Steele, J. C. & Meng, X.-Z. Usage, residue, and human health risk of antibiotics in Chinese aquaculture: A review. Environmental Pollution 223, 161–169 (2017).
Munita, J. M. & Arias, C. A. Mechanisms of Antibiotic Resistance. Microbiol Spectr 4, 10.1128/microbiolspec.VMBF-0016–2015 (2016).
Schwarz, S., Kehrenberg, C., Doublet, B. & Cloeckaert, A. Molecular basis of bacterial resistance to chloramphenicol and florfenicol. FEMS Microbiol Rev 28, 519–542 (2004).
von Wintersdorff, C. J. H. et al. Dissemination of Antimicrobial Resistance in Microbial Ecosystems through Horizontal Gene Transfer. Front. Microbiol. 7, (2016).
Khedkar, S. et al. Landscape of mobile genetic elements and their antibiotic resistance cargo in prokaryotic genomes. Nucleic Acids Res 50, 3155–3168 (2022).
Deng, Y. et al. Resistance integrons: class 1, 2 and 3 integrons. Annals of Clinical Microbiology and Antimicrobials 14, 45 (2015).
Gordon, L. et al. Complete sequence of the floR-carrying multiresistance plasmid pAB5S9 from freshwater Aeromonas bestiarum. J Antimicrob Chemother 62, 65–71 (2008).
Srivastava, S. Plasmids: Their Biology and Functions. in Genetics of Bacteria (ed. Srivastava, S.) 125–151 (Springer, India, 2013). doi:10.1007/978-81-322-1090-0_6.
Martak, D., Henriot, C. P. & Hocquet, D. Environment, animals, and food as reservoirs of antibiotic-resistant bacteria for humans: One health or more? Infectious Diseases Now 54, 104895 (2024).
Older, C. E. et al. Comparison of high-throughput sequencing methods for bacterial microbiota profiling in catfish aquaculture. North American Journal of Aquaculture 86, 39–54 (2024).v

Environmental Reservoirs of Antibiotic Resistance in Commercial Catfish Ponds

Environmental Reservoirs of Antibiotic Resistance in Commercial Catfish Ponds

Funding Progress

Peer Reviews

Related Funding

Request for Proposals - Antimicrobial Resistance (AMR) in Aquaculture

Nonprofit Recipient

Topics

DOI