Functional genomics revealing antibiotic resistance genes in cultivated bacteria of the East Sea

Abstract

Antibiotic resistant bacteria have been characterized in a wide variety of environments away from their original places of use. Natural environments act as reservoirs of antibiotic resistance genes, including soil, glaciers, animals, and the open ocean, all of which are environments far removed from initial use of antibiotics. Resistance mechanisms can be inherited through vertical transfer or through horizontal gene transfer, which occurs when resistance genes are located on mobile elements, such as plasmids, allowing movement amongst diverse bacteria in exposed environments. Other methods of resistance acquisition include natural resistance, genetic mutation, or transposition of a bacteriophage. Recent studies have focused on metagenomic analyses in the ocean, or genomic analyses in land environments. Several studies on antibiotic resistance genes use primers to detect antibiotic resistance genes in bacteria, which is limiting due to primer-specificity. There are many unknown components that can give rise to resistance in bacteria, and using a functional approach can help to detect these novel mechanisms. In this study, a total of eight seawater samples were analyzed from five stations in the East Sea: one at the surface and one at depth of 100 or 200m, except Gijang and TEB04, where only a surface sample was analyzed. One site was adjacent to fish farm waste runoff, while all others were taken from the open sea. A total of 245 isolates were identified with a 16S rRNA gene sequence of greater than 97%. Of these cultivated bacteria from eight samples, 51.4% belong to Gammaproteobacteria, 19.8% to Alphaproteobacteria, 15.2% to Firmicutes, 7.4% to Bacteriodetes, 3.7% to Actinobacteria, 2.1% to Betaproteobacteria, and 0.4% to Lentisphaeria. Genomic DNA was fragmented to a target size range of 1000-3000 base pairs using a nebulization method, ligated into an expression vector and electroporated into electrocompetent Escherichia coli cells. Recovered cells were spread onto selection media containing one of eleven antibiotics to allow for expression. The insertion fragments from selected colonies were amplified, and then subjected for sequencing. Retrieved sequences were combined to form a single sequence. Of the 120 isolates were selected and tested for resistance, 286 insertion sequences were retrieved, leading to 433 protein sequences identified using Prokka. From these sequences, two identified as resistance genes in blastp, confirmed with HHpred, and 24 sequences (5.5%) identified as positive matches in the Antibiotic Resistance Database (ARDB) using the Protein-BLAST function. Of these matches, 58.3% came from surface samples, and over half were identified as drug efflux systems. The highest incidence of resistance was conferred to trimethoprim (49.4%). Aside from the ARDB results, at least 49 (11.3%) of the 433 sequences identified as an efflux protein, membrane protein or transport protein using blastp or HHpred. A total of 54 sequences (12.5%) identified closest with hypothetical proteins, and 23 sequences (5.3%) did not return any significant hits with blastp or HHpred (including two sequences which matched the ARDB). Genes associated with erythromycin, vancomycin, and aminoglycoside resistance were also detected in this study. Several genes retrieved in this study did not match a previously known resistance mechanism. This study helps to unearth some of the unknown mechanisms of antibiotic resistance in marine bacteria, particularly proteins that could have secondary functions and can expand our knowledge and understanding into environmental resistance mechanisms.