Our results suggest that CPE, as well as other MDR pathogens, occur in the coastal waters of a popular public recreational beach offshore Netanya, Israel and in nearby river estuaries. These CPE appear to be abundant and well-acclimated to freshwater and seawater, as they were easy to isolate and cultivate. It is likely that CPE contamination is persistent, as at the time of our sampling (June and July 2018) no considerable fluctuations in seawater quality were reported36. Local residents and tourists visiting popular recreation and leisure sites may be exposed to these MDROs, as was described here.
The fact that isolates with highly similar genomes were cultivated from both the riverine and marine environments, implies a mutual source, however, its origin is still unclear. Rivers are putative reservoirs and sources of MDROs, which were shown to be virulent 37,38. MDROs are less frequently encountered in patients drowning in seawater37. In Alexander River, the possible sources of contamination include effluents from the adjacent WWTPs, a sea-turtles rescue center, an algae plant, an agricultural catchment basin and reservoir waters22,39. Other unknown sources of contamination may also exist. Daily salinity profiles measured by Ruppin Estuarine and Coastal Observatory39 suggest that this process often directs coastal water up to few kilometers inland. Despite the limited mixing of water masses, bacterial cross-contamination may be substantial. Either way, a gradient of salinity along the estuary may select for specific bacterial lineages in distinct ecological niches.
Gut microbes may be adapted to the aquatic environment, in which their survival rates are poorly understood. Yet, cultivation-based studies demonstrate that common gut bacteria, such as Enterobacteriaceae, are frequently detected in freshwater7,8,11−13,20 and seawater5,6,15. These bacteria may be able to cope with different salinities because the human gut environment is characterized by spatial and temporal heterogeneity in osmolarity, based on the kinds of meals consumed40. The fitness of microbes and their growth rates depends not only on osmolarity but also on taxon-specific physiology and additional external factors such as nutrient abundance, pH and oxygen levels41. Thus, the fact that bacteria grew faster in river water than in seawater media, is not surprising, given the oligotrophic properties of the southeastern Mediterranean Sea, as well as the marked difference in salinity of these media. The three sequenced CPE genomes encoded proteins that have been shown to be involved in halotolerance, including the osmosensitive K + channel histidine kinase KdpD, NhaA type Na+/H+ antiporter DNA-binding protein H-NS42. These sequences were found in scaffolds that were > 400 kbp in length and were defined as chromosomal by plasFlow. This indicates that at least some of the salt tolerance-related traits in these strains are not linked to plasmids as in other bacteria, and therefore not to blaIMI genes, which most likely are encoded on plasmids. Salinity has been recently shown to be the most important factor modulating the distribution patterns of antibiotic resistance genes in oceans and river beach soils43. Thus, it is feasible that less-studied traits that mediate salinity tolerance may be linked to antibiotic resistance genes.
Most importantly, these aquatic isolates are only remotely associated with the local clinical epidemiology. CP Enterobacter spp., are uncommon in our hospital’s clinical settings: between August 2013 and February 2019, we identified 129 Enterobacter spp. out of 798 CRE rectal screening isolates (16%). Carbapenemase-producing Enterobacter spp. were detected in 65 isolates. blaIMI genes were found only in three of these CPE isolates, while blaKPC (50 isolates) and blaNDM (10 isolates) were more common. However, the IMI mechanism can often go undetected, because only the five major enzymes (KPC, NDM, VIM, OXA-48 and IMP) are routinely tested. The clinical implication is that Enterobacter spp. carrying an unidentified blaIMI could have been misidentified as non-CP CREs. Since patients carrying non-CP CRE isolates are not cohorted in Israeli hospitals as CPE carriers, such misidentifications increase the potential for hospital cross-infection and outbreaks.
OXA-48 has also not been frequently encountered during this time period in our facility (19 isolates out of 798 CREs, mostly seen in E. coli spp). Nevertheless, IMI carbapenemases appear to be emerging in clinical practice44 and as well as causing nosocomial outbreaks45.