Culture-independent methods reveal high diversity of OXA-48-like genes in water environments

The class D OXA carbapenemases comprises a very diverse group of enzymes that have been identified mostly in outbreaks of carbapenems-resistant Acinetobacter spp. (e.g. OXA-23, OXA-24, OXA-40, and OXA-58), Pseudomonas spp. (OXA-50) and Enterobacteriaceae (OXA-48) (Evans & Amyes 2014). Other class D carbapenemases have been considered as species-specific like the OXA-60 family, naturally present in the genome of Ralstonia pickettii (Girlich et al. 2004), and OXA-62 in Pandoraea pnomenusa (Schneider et al. 2006).

OXA-48 was first identified in a clinical Klebsiella pneumoniae isolate in Turkey (Poirel et al. 2004). Although initially disseminated mostly in Mediterranean countries, nowadays OXA-48 and its variants are an example of widely disseminated carbapenemases that have been detected in all continents (Poirel et al. 2012). These enzymes hydrolyze penicillins and carbapenems, but not third generation cephalosporins (Poirel et al. 2012). However, there are many reports of OXA-48-like producers that carry additional beta-lactamases, thus limiting treatment options (Poirel et al. 2012).

Since its first description (Poirel et al. 2004) bla_OXA-48-like genes detection has been restricted to Shewanella species (Poirel et al. 2012; Tacão et al. 2013), and the large majority to Enterobacteriaceae members worldwide, mostly in clinical settings. However sporadic bla_OXA-48 genes have been reported also on Pseudomonas aeruginosa (Meunier et al. 2016) and Acinetobacter baumannii (Mathlouthi et al. 2015). Furthermore, there are also reports describing OXA-48 producers in Enterobacteriaceae isolated from river water (Potron et al. 2011), wastewater (Galler et al. 2014) and companion animals (Stolle et al. 2013).

Up to now, several OXA-48 variants have been found, differing in 1 to 5 amino acids, as: OXA-162, OXA-163, OXA-181, OXA-199, OXA-204, OXA-232, OXA-244, OXA-245, OXA-247, OXA-370 and the recently described OXA-405 and OXA-416 (Poirel et al. 2012; Gomez et al. 2013; Sampaio et al. 2014; Antonelli et al. 2015; Dortet et al. 2015).

There are strong indications that Shewanella spp. are the putative origin of OXA-48-like genes (Poirel et al. 2012; Zong 2012; Tacão et al. 2013). Members of this genus are mostly identified in water environments, but some have already been implicated in human infections (Janda & Abbot 2014). Although there are reports on OXA-48-like carbapenemases worldwide, it has been pointed out that the spread of this beta-lactamase is silent due to the difficulties in its detection. The fact that OXA-48 producers show low minimal inhibitory concentration values for carbapenems might be masking their detection and thus underestimating their dispersion (Poirel et al. 2012). Therefore, molecular methods have been pointed out as potential alternatives for the recognition of OXA-48-like producers (Poirel et al. 2012).

The study of the diversity of bla_OXA-48-like genes is important for elaborating molecular-based strategies for their rapid detection. Also, revealing the diversity of these molecular determinants may assist in forecasting the likely dissemination of bla_OXA-48-like genes. In this study we aimed to evaluate the diversity of OXA-48-like class D carbapenemase encoding genes in aquatic systems. In order to attain a broader assessment of gene variety in these environments, we applied a culture-independent approach.

Samples were collected from three rivers (three each, halfway river banks) in the Vouga River basin (central Portugal). These rivers are highly polluted due to disposals of domestic, industrial and agricultural origins (for details see Tacão et al. 2012). The estuarine water was collected from Ria de Aveiro, a mesothrophic estuary located in the same basin and highly polluted due to the presence of harbor facilities, aquaculture ponds, industrial plants, diffuse domestic sewage inputs and run-off from agricultural fields (Azevedo et al. 2012). Water was collected into sterile bottles from 50 cm below the water surface.

Environmental DNA was purified by filtering 200–500 ml of water through 0.2-μm-pore-size filters (Poretics Products). Cells were washed from the filter with TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) followed by centrifugation (13,000 rpm, 10 min). The pellet was resuspended in 200 μl TE buffer containing 10 mg/ml of lysozyme, followed by 1 h incubation at 37°C, and then frozen in liquid nitrogen and thawed three times. DNA extraction continued by using the Genomic DNA Extraction Kit (MBI Fermentas, Vilnius, Lithuania) according to the manufacturer's instructions. Purified DNA was stored at −20°C.

The bla_OXA-48-like gene fragments were amplified from a pool of environmental DNA from rivers and from DNA extracted from estuarine water with the two bla_OXA-48-like-specific primer sets described so far, designed by: (i) Poirel et al. (2011) (fwd: 5′-GCGTGGTTAAGGATGAACAC and rev: 5′-CATCAAGTTCAACCCAACCG) and (ii) Zong (2012) (fwd: 5′ AGCAAGGATTTACCAATAAT and rev: 5′ GGCATATCCATATTCATC). The polymerase chain reaction (PCR) reaction mixtures (25 μl total volume) consisted of 6.25 μl NZYTaq 2x Green Master Mix (2.5 mM MgCl2; 200 μM dNTPs; 0.2 U/μl DNA polymerase) (NZYtech, Portugal), 16.25 μl of ultrapure water, 0.75 μl of each primer (reverse and forward), and 50–100 ng of purified DNA. PCR reactions were performed in a MyCycler Thermal cycler (Bio-Rad, USA) with conditions as described by Poirel et al. (2011) and Zong (2012). Positive and negative controls were included in each PCR reaction. Water was used as negative control and strain Shewanella xiamenensis IR33 carrying a bla_OXA-48 gene was used as positive control (Tacão et al. 2013). PCR products were analyzed by electrophoresis on a 1.5% agarose gel and stained with ethidium bromide.

Clone libraries of bla_OXA-48-like gene fragments were constructed using the TA Cloning Kit, according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA) and Escherichia coli NZYStar competent cells (NZYTech, Portugal). Clones were screened by PCR for the presence of fragments with the expected size, using primers targeting the vector. Amplicons were purified and sequenced in both strands. Similarity searches in the GenBank database were performed using the BLAST tool with the deduced amino acid sequences. A phylogenetic tree was obtained using MEGA 6.0 (Tamura et al. 2013). The Shannon index of diversity (H) was calculated for each library by using the formula H = −Σ(ni/N) log(ni/N), where ni is the abundance of each bla_OXA-48-like type and N is the sum of the analyzed clones in each library.

All the nucleotide sequences obtained in this work have been deposited in the GenBank database under the accession numbers KJ620426-KJ620480.

In this study, we evaluated the diversity of bla_OXA-48-like gene sequences in river and estuarine water by culture-independent methodologies.

From river water DNA, it was possible to amplify bla_OXA-48-like genes using the two primer sets, and both amplicons were used for constructing two libraries. From estuarine water DNA amplification was obtained using the primer set described by Poirel et al. (2011), and only this amplicon was used. Overall, three clone libraries of bla_OXA-48-like genes were constructed.

Although variants here identified may have their origin in Shewanella spp. genomes, it should also be taken into account that there may be other hosts not yet identified or even enterobacteria carrying these genes. Since the aquatic systems explored in this study are highly impacted by anthropogenic pollution, and thus subjected to continuous discharges from agricultural, domestic and industrial origins, a high proportion and diversity of Enterobacteriaceae members is expected in these water environments. Future studies should include pristine environments to test all the hypotheses stated. Most probably, diverse mobilization events have mediated the transfer of bla_OXA-48-like genes from Shewanella spp. to Enterobacteriaceae or other still unidentified hosts which reinforces the importance of these environmental compartments (i.e. aquatic systems where indigenous bacteria such as Shewanella mix with animal-derived bacteria such as Enterobacteriaceae) in the evolution and spread of antibiotic resistance. In fact, there are already reports of Enterobacteriaceae members isolated from river (Potron et al. 2011) and wastewaters (Galler et al. 2014) carrying bla_OXA-48 genes. However, so far, there is only one report on bla_OXA-48-like producers identified in Portuguese clinical settings (Manageiro et al. 2014). This might be related to the national carbapenems prescription policies and awareness campaigns (Henriques et al. 2012). In a worst-case scenario, in the particular case of bla_OXA-48 genes, what might be happening is a silent spread, i.e. one that is undetectable due to the low-level resistance to carbapenems.

By applying environmental DNA-based methodologies, both culturable and unculturable fractions are covered. However, it is important to acknowledge that by using PCR-based methodologies, the diversity found is biased by the primer sets used, which were designed based on previous described sequences. Nevertheless, this molecular approach added relevant information to the current knowledge on the diversity of OXA-48-like carbapenemases. However, the DNA sequences identified may not represent functional beta-lactamase genes or that code for active beta-lactamases.

The diversity of OXA-48-like sequences identified by culture-independent methods suggests that the environment and, in particular, aquatic systems constitute important reservoirs of these genes. Also from this study resulted a list of diverse variants of OXA-48 genes that should be taken into account when designing molecular strategies for detecting this gene.

The observation of such a diverse bla_OXA-48-like gene pool in these aquatic systems indicates the need for further research along at least three lines: identification of the host species, assessment of expression and activity of the gene products, and evaluation of the capability of dissemination among strains of the variants here reported. Even so, these observations may represent a forewarning of bla_OXA-48-like genes dissemination.

This work was supported by Fundação para a Ciência e a Tecnologia (FCT) through project StARE (WaterJPI/0002/2013) and grants I.H. (IF/00492/2013) and M.T. (SFRH/BPD/114855/2016). We wish to thank Catarina Maia and Susana Araújo (Microbiology Lab, University of Aveiro) for technical support during this research.