Temporo-spatial variations in resistance determinants and clonality of Acinetobacter baumannii and Pseudomonas aeruginosa strains from Romanian hospitals and wastewaters

Background Romania is one of the European countries reporting very high antimicrobial resistance (AMR) rates and consumption of antimicrobials. We aimed to characterize the AMR profiles and clonality of 304 multi-drug resistant (MDR) Acinetobacter baumannii (Ab) and Pseudomonas aeruginosa (Pa) strains isolated during two consecutive years (2018 and 2019) from hospital settings, hospital collecting sewage tanks and the receiving wastewater treatment plants (WWTPs) located in the main geographical regions of Romania. Methods The strains were isolated on chromogenic media and identified by MALDI-TOF-MS. Antibiotic susceptibility testing and confirmation of ESBL- and CP- producing phenotypes and genotypes were performed. The genetic characterization also included horizontal gene transfer experiments, whole-genome sequencing (WGS), assembling, annotation and characterization. Results Both clinical and aquatic isolates exhibited high MDR rates, especially the Ab strains isolated from nosocomial infections and hospital effluents. The phenotypic resistance profiles and MDR rates have largely varied by sampling point and geographic location. The highest MDR rates in the aquatic isolates were recorded in Galați WWTP, followed by Bucharest. The Ab strains harbored mostly blaOXA-23, blaOXA-24, blaSHV, blaTEM and blaGES, while Pa strains blaIMP, blaVIM, blaNDM, blaVEB, blaGES and blaTEM, with high variations depending on the geographical zone and the sampling point. The WGS analysis revealed the presence of antibiotic resistance genes (ARGs) to other antibiotic classes, such as aminoglycosides, tetracyclines, sulphonamides, fosfomycin, phenicols, trimethoprim-sulfamethoxazole as well as class 1 integrons. The molecular analyses highlighted: (i) The presence of epidemic clones such as ST2 for Ab and ST233 and ST357 for Pa; (ii) The relatedness between clinical and hospital wastewater strains and (iii) The possible dissemination of clinical Ab belonging to ST2 (also proved in the conjugation assays for blaOXA-23 or blaOXA-72 genes), ST79 and ST492 and of Pa strains belonging to ST357, ST640 and ST621 in the wastewaters. Conclusion Our study reveals the presence of CP-producing Ab and Pa in all sampling points and the clonal dissemination of clinical Ab ST2 strains in the wastewaters. The prevalent clones were correlated with the presence of class 1 integrons, suggesting that these isolates could be a significant reservoir of ARGs, being able to persist in the environment. Supplementary Information The online version contains supplementary material available at 10.1186/s13756-022-01156-1.

The Joint Programme Initiative on Antimicrobial Resistance (JPIAMR) is represented by the elucidation of the role of the environment as a source for the selection and dissemination of AMR. One important goal is mapping the distribution of multidrug-resistant (MDR) pathogens and plasmids of different genomic lineages in different clinical and aquatic compartments. This important insight could be translated into policy measures to monitor AMR and control the emergence and spread of ARB [16,17].
Acinetobacter baumannii (Ab) and Pseudomonas aeruginosa (Pa) are members of the initially designed "ESKAPE", then "ESCAPE" (Enterococcus faecium, Staphylococcus aureus; Clostridioides difficile, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacteriaceae) group [18,19]. Centers for Disease Control and Prevention (CDC) classified them urgent threat due to their often MDR features [20] and therefore requiring concerted research and management efforts [21]. Ab and Pa exhibit all known AMR mechanisms, such as drug inactivation/alteration, modification of drug binding sites/targets, cell permeability modification and biofilm development [22]. One of the most clinically relevant mechanisms of resistance in Ab and Pa strains is represented by the production of antibiotic inactivating hydrolytic enzymes, especially carbapenemases (CPs).
Presently, there is insufficient data on the Ab and Pa dissemination and survival from hospitals in wastewater and finally into the natural recipients. Despite the huge burden of AMR presence in Romania and its significant overall impact on European AMR rates, the genetic relationships between the bacterial strains isolated from different aquatic environmental and clinical compartments were not investigated at a national level.
We aimed to perform a phenotypic and molecular characterization of the acquired resistome of a significant number of MDR Ab and Pa strains isolated in the same temporal sequence from the hospital environments and the receiving wastewater network from different counties in Romania.

Phenotypic characterization of the Ab and Pa strains Sampling location
Water sampling was performed from September 2018 to August 2019. The collection points were represented by sewage tanks from eight hospital units and their wastewater collecting WWTPs. The eight sampling locations covered several Romanian regions such as the Southern region, with sampling locations in Bucharest (C/H-Glina municipal WWTP 2018/2019), WWTP Târgoviște (E) and WWTP Râmnicu-Vâlcea (J); Central and Western regions: WWTP Cluj (I) and WWTP Timișoara (F); Northern and Eastern regions: WWTP Iași (G) and WWTP Galați (D) (Fig. 1).
The different sampling points from the selected locations were: hospital/WWTP effluent (EF), WWTPs influent (IN), activated sludge (AS) and returned sludge (RS),  where all isolated strains were considered within a single group.

Isolation and characterization of Ab and Pa strains
The water samples were collected in sterile glass containers, transported to the laboratory at 5 ± 3 °C and processed within the first 24 h. The water samples were diluted and filtered through 0.45 μm pore size membrane filters (Millipore, France), as described in SR EN ISO 9308-2/2014 (for coliform bacteria) and then cultivated on ChromID ESBL agar and ChromID CARBA agar (Bio-Mérieux, France). The resistant colonies developed after cultivation at 37 °C for 24 h in aerobic conditions were subsequently inoculated on the corresponding antibioticenriched media for the confirmation of ESBL-(ChromID ESBL) or CP-(ChromID CARBA) producing phenotypes. All resistant strains were identified by MALDI-TOF-MS (Bruker system). In the same time frame with the collection of the water samples (i.e., during a ten-day period prior to the water sampling), Ab and Pa clinical strains were isolated from intra-hospital infections that occurred in the units discharging the wastewater in the sampled WWTPs. The antibiotic susceptibility profiles of the identified strains (220 Ab and 84 Pa), were determined using the standard disc diffusion method according to Clinical and Laboratory Standards Institute (CLSI guidelines) for 2018 and 2019 [25,26] (see Additional file 1: Tables S1  and S2).

Mating experiments
Transferability of bla OXA-23 and bla OXA-24 genes by conjugation was tested using the solid mating method, with rifampicin (RIF) resistant Acinetobacter baylyi ADP1 as recipient. Briefly, equal amounts (100 µL) of overnight cultures of the donors (n = 40 Ab strains from all isolation sampling points) and recipient strains were mixed and incubated in Brain heart infusion agar plates. Cells were resuspended in saline solution and selected in plates containing RIF (300 mg/L) and meropenem (MEM) (0.5 mg/L) [29]. Characterization of the transconjugants was conducted by PCR.

Whole-genome sequencing (WGS), assembling, annotation and characterization
To determine the genetic relationships between the clinical and wastewater isolates, 54 strains (Ab, n = 34 and Pa, n = 20) were selected for WGS based on the isolation source, geographical region, temporal sequence and the presence of MDR phenotype in order to have a complete picture of the antibiotic resistance in different Romanian regions. Total DNA was isolated using DNeasy Ultra-Clean Microbial Kit (Qiagen) and subjected to Illumina (Nextera DNA Flex Library Prep Kit) sequencing on a MiSeq platform (V3, 600 cycles). The sequencing quality was very good (with an average of 88% over Q30 and 95% over Q20 for Ab (78% over Q30 and 89% over Q20 for Pa), and an average of 1.53 million reads per sample for Ab and 1.66 million reads per sample for Pa. Raw reads were checked for quality using FastQC v0.18.8 [30], assembled using Shovill v1.1.0 pipeline [31] and primarily annotated using Prokka v.1.14.6 [32], while the prediction of resistance profiles was performed by using ABRicate v0. 5 [35,40] and Single Nucleotide Polymorphism (SNP) (v3.1) [41]. Comparative gene analyses were performed using Roary v.3.13.0 [42] and the output was used to infer phylogenies using RAxML v8.2.12 (Maximum Likelihood inference using bootstrap value N = 1000) [43] and visualized using iTOL [44]. The assembled sequences have been deposited in GenBank with BioProject ID PRJNA841266.
The 34 selected Ab WGS assemblies were subjected to phylogenetic analysis to further attain their relationship to other Ab strains from the NCBI database. For this, all Ab sequences available (n = 4175) were downloaded from NCBI and from those, 71 were randomly selected. MLST predictions and annotations were performed on this dataset and, together with the 34 selected strains, were subjected to further pangenome analysis using Roary and aligned with Mafft v.7.741. The resulted phylogenetic tree was drawn with

Identification and antibiotic susceptibility profiles of Ab and Pa strains
A total number of 220 Ab and 84 Pa were isolated from clinical and water samples during the two consecutive years.
Two pilot sampling campaigns were performed in Bucharest in 2018 from which 66 Ab and 50 Pa strains were isolated (Table 1).
In 2019, the sampling campaign was extended, including, in addition to Bucharest, six other cities that are representative of the main country regions, i.e., North-East (Iași, Galați), Central-West (Cluj, Timișoara) and South (Târgoviște, Râmnicu Vâlcea). A total of 154 Ab and 34 Pa resistant isolates were recovered (Table 1).
Within the same time frame, clinical Ab and Pa clinical strains were isolated in hospital units from which wastewater samples were collected. The hospital wastewater was collected and treated by the corresponding WWTPs from the same town.
The analysis of MDR rates from the hospital to the collecting WWTP in the first sampling campaign in 2018 has revealed the following aspects: (1) clinical isolates-all Ab and the majority of Pa strains (93.3-100%) were MDR (Additional file 1: Tables S1 and S2); (2) hospital EF-all Ab isolates were MDR (Additional file 1: Table S1), while the Pa strains exhibited various MDR rates (from 25 to 100%) (Additional file 1: Table S2); (3) WWTP C, collecting the two hospital EFs-all Ab isolates were MDR (Additional file 1: Table S1).

Geographic distribution of CP and ESBL genes in clinical and water Ab and Pa isolates
The 2018 pilot study was limited to the Bucharest region, and then it was extended during 2019 to other regions of the country, allowing us to perform a comparative analysis regarding the geographic distribution of the CP and ESBL encoding genes in the Ab and Pa strains.
The Ab isolates from the Central-Western region revealed the presence of the following CP and ESBL encoding genes: (1) in clinical settings, all Ab strains were bla VEB positive; 50% were bla OXA-23 and bla OXA-24 positive; (2) the aquatic isolates recovered from the two sampled WWTPs were bla OXA-23 and bla VEB positive (23.07% each of them).
The geographical distribution of the CP/ESBLs found in Pa strains isolated from intra-hospital infections, the hospital sewage tank and the sampled WWTP from the corresponding cities was as follows: in the Southern region, 50% of nosocomial Pa strains were VEB producers, while the wastewater Pa strains harbored bla VEB (40%), blaNDM (40%) and bla GES (20%); in the North-Eastern region, 50% of clinical Pa strains were VEB producers; 62.5% of the hospital sewages strains were positive for bla TEM ; 50% respectively 25% of the WWTPs were bla VEB and bla TEM positive. The Pa strains from the Central-Western regions revealed different CP/ESBLs in clinical/hospital sewage (VIM) and in the sampled WWTPs (14.28% were GES positive) (Fig. 3).

WGS analysis of clinical and wastewater Ab and Pa isolates
In case of nine Ab strains recovered in 2018 in Bucharest from intra-hospital infections (n = 2), hospital sewage EF (n = 5) and, the corresponding WWTP EF (n = 2) the WGS demonstrated the presence of OXA-72 and OXA-23 encoding genes in the IN and the EF of the collecting sewage tank and of genes encoding aminoglycoside modifying enzymes (AMEs) i.e. aph(3′)-VIa, ant(3″)-IIa, sulphonamides (sul1) and class 1 integrons (qacE∆1 integron-associated gene in 3′ CS region), in all investigated samples ( Table 2).
The WGS analysis of the acquired resistome of the 10 Pa strains recovered in 2018 revealed the dissemination of CP bla IMP-13 and genes encoding AMEs (aph(3′)-IIb, ant(2″)-Ia), fosfomycin (fosA), phenicols (catB7, bcr1) and trimethoprim-sulfamethoxazole (sul1) resistance genes in strains isolated from one Bucharest hospital and its effluent (Table 3). In the case of the second investigated hospital and the corresponding sewage tank, there has been noticed the presence of genes encoding AMEs (aph(3′)-IIb) and phenicols (catB) ( Table 3).
The ARGs distribution of four Pa strains, collected from Northern and Eastern Romania in 2019, revealed the presence of ESBL encoding genes (bla TEM-40 , bla VEB-9 ), AMEs encoding genes (aac(6′)-II, aadA1, aph(3′)-IIb) as well as determinants of resistance to fosfomycin (fosA), phenicols (catB7, bcr1), tetracycline [tet(A)] and class 1 integrons (qacE∆1 integron associated gene) in clinical and wastewater samples. Regarding the six Pa clinical and wastewater strains from Central and Western regions of Romania, the presence of resistance genes encoding for fosfomycin (fosA) and phenicols (catB7, bcr1) has been observed in Pa strains from almost all sources (Table 4).  (Table 5); group III was represented by closely related (26 SNPs) strains isolated from Bucharest hospital and its collecting sewage tank isolates belonging to ST492; group IV included less related (577 SNPs) clinical strains and aquatic strains isolated from Bucharest belonging to ST79; were; group V included less related (> 700 SNPs) wastewater strains isolated from South Romania belonging to ST1; group VI included only environmental strains from the Northern region of the country belonging to ST155 and two related novel STs (Fig. 4).
Pangenome analyses performed on 34 WGS Ab strains and 71 selected genomes from the NCBI database, supported also by conjugation assays revealed clear dissemination of the same circulating clones from the hospital units into different aquatic compartments [i.e., ST2 encountered in Bucharest hospital unit and the corresponding sampled WWTP carrying the same CP encoding gene (bla OXA-23 or bla OXA-72 ) in Ab strains; ST2 carrying bla OXA-72 gene in Ab strains from Târgoviște hospital unit and bla OXA-23 in the EF of the corresponding WWTP E] and similarities with other international ST2 clones (Additional file 3: Fig. S1). The Pa strains (Fig. 5) were also grouped in six phylogenetic groups: group I included wastewater isolates from Timișoara, and one clinical strain from the Bucharest hospital that belonged to three singleton STs (ST252, ST254 and ST395); group II comprises clinical strains from Central Romania (Cluj hospital) and one collecting sewage tank from a hospital unit in Bucharest that belonged to the epidemic clone ST233; group III included strains isolated from Iași hospital sewage and its collecting WWTP G belonging to ST640; group IV was represented by one Bucharest hospital unit collecting sewage tank isolate belonging to ST621; group V included wastewater strains from central Romania belonging to ST620; group VI contained the majority of the strains (clinical and wastewater isolates from South and East Romania) belonging to the widespread ST357 and one unknown ST (Fig. 5).
The spread from the hospital unit into the natural aquatic recipient was observed in some cases, i.e.: for an epidemic clone isolated from Galați hospital unit and the receiving WWTP D (ST357 carrying the bla VEB-9 ESBL encoding gene);two ST640 strains isolated Iași hospital and sludge from the Iași urban WWTP (32 SNPs, thus below the proposed threshold of 37 SNPs). The other Pa strains are more diverse, even within the same clone, the strains being more distantly related (> 100 SNPs); this fact  11:115 was also suggested by the difference between the core genome (4716 genes) and the pan genome (12,395 genes) calculated for all the Pa strains included in this study.

Discussion
Hospitals are a concentrated source of MDR bacteria, which besides having clinical consequences (treatment options are limited and expensive), can be released in wastewater and finally into the environment [45]. Previous studies have revealed the presence of β-lactams, tetracyclines, quinolones and sulfonamides resistance genes in both natural and polluted aquatic environments, indicating that these determinants are released from clinical into aquatic environments, and then further disseminated to opportunistic pathogens [46]. Therefore, rapid identification of high-risk clones is essential for isolating infected patients, preventing the spread of resistance and improving the antimicrobial treatment. This requires the knowledge of the genetic environment and the carrying platforms of ARGs, as well as the development of new methods for assessing the spreading potential of ARGs mediated by mobile genetic elements in both aquatic and hospital environments.  11:115 In Romania, in contrast to clinical studies [27,28,[47][48][49][50][51][52][53][54][55][56][57][58][59][60][61], there is little information available on the epidemiology of AMR reservoirs in the environment, especially in polluted water and rivers. We have previously shown that MDR, CP and ESBL-producing K. pneumoniae isolated from clinics, hospital wastewater and urban WWTPs from different regions of the country exhibit multiple antibiotic and antiseptic resistance, as well as virulence genes, with the ST101 clone being the most frequently encountered in all sampling sites [62]. Also, we have demonstrated the spread of K. pneumoniae ST101 from hospital to wastewater influent and its persistence in the wastewater effluent after the chlorine treatment, suggesting its dissemination in the community and in different aquatic compartments [63]. Our previous research showed that enterococci and Enterobacterales strains in four Romanian natural aquatic fishery lowland salted lakes from Natura 2000 Network carried a high diversity of resistance markers correlated with class 1 integrons [64]. Other authors described tetracycline and sulfonamides ARGs in the WWTP and the receiver river from northwestern Romania and demonstrated that some ARGs, such as bla VIM and bla SHV could persist in the chlorinated hospital wastewater, being detected both in the influent and chlorinated effluent [65,66].
The purpose of this study was to characterize the AMR profiles and clonality of two of the most dangerous ESKAPE pathogens, Ab and Pa, isolated for two consecutive years from hospital settings, hospital collecting sewage tanks and the receiving WWTPs from three different geographical regions of Romania. The clinical and environmental Ab isolates recovered from different geographical regions of Romania revealed high AMR and MDR levels. In another study, from a significant number of groundwater, surface water, and soil samples from Hungary, there were isolated different Acinetobacter species (i.e., A. baumannii, A. johnsonii, A. gyllenbergii and A. beijerinckii, with 8.10% of A. beijerinckii) exhibiting an MDR phenotype [67].
In our study, imipenem-resistant Ab and Pa strains were also resistant to other classes of clinically important antibiotics, including quinolones (ciprofloxacin) and aminoglycosides (gentamicin). MDR was defined according to Magiorakos et al. [45], as non-susceptibility to at least one agent in three or more antimicrobial classes. The phenotypic resistance profiles and MDR rates have largely varied by sampling point and by geographic location. The highest MDR rates in aquatic isolates were recorded in Galați WWTP (D) that could be explained by the location of this county on the lower course of Danube River. The Danube River is considered the most important non-oceanic body of water in Europe and the "future central axis for the European Union", with its Danube Delta included in the Biosphere Reserve and Ramsar Sites lists. The Danube River crosses ten countries, so this basin represents an optimal pool for resistant pathogens and anthropogenic pollutants dissemination and accumulation throughout large and distant areas, being assigned as a reservoir of AMR. The following two locations with high MDR rates in the aquatic isolates were  [68]. Previous data has also indicated the presence of different A. baumannii complex species with MDR phenotypes isolated from environmental samples in Hungary [67]. For Pa clinical strains, the following CPs and ESBLs have been detected: IMP, VIM, NDM, VEB, GES and TEM.
WGS bioinformatic analysis of Ab strains highlighted that the international clone ST2 is broadly spread in our country (Additional file 3: Fig. S1), with 56% of the analyzed Ab strains belonging to this clone. Two ST2 strains were included in ST2 branch since the ST492 is a single locus variant of ST2 [61]. The other STs (in the set highlighted with grey in Additional file 3: Fig. S1) have phylogenetic relationships in accordance with the reference sequences, meaning that the whole cluster highlighted with grey is not homogenous, due to random selection of the reference sequence for the phylogenetic analysis. Therefore, the following STs belong to the cluster (in the same order as in the phylogenetic tree): ST499, ST78, unknown, unknown, ST155, ST622, ST46, ST16, ST40, ST403, unknown, ST429, unknown, ST71, ST113, ST25, unknown, ST10, ST10, ST10, ST108, ST514, unknown (Additional file 3: Fig. S1). Worldwide CP producers are mostly associated with international clone II and OXA-23 [61,69,70]. Other clinical and wastewater isolates belonged to ST636, ST1, ST79, ST492 and ST2 and were correlated with OXA-72. In Zagreb (Croatia) carbapenem-resistant Ab positive for bla OXA-23 recovered from different sampling points of a WWTP and the sewage of a nursing home belonged to the international clonal lineage IC2, the OXA-72 producers belonged to IC1, while the susceptible ones were unclustered [71,72]. The OXA-23, OXA58, and OXA-72 CPs linked to ST2 in hospital environments have also been reported in other countries such as Croatia, Serbia, Bosnia and Herzegovina [73][74][75][76].
The WGS analysis of Pa strains has shown that ST357 was correlated with IMP-13 in one clinical strain; in the sewage effluent, ST357 was correlated with VEB-9 and with both IMP-13 and VEB-9; the epidemic clone ST233 was correlated with VIM-2 in clinical and wastewater Pa strains; ST640 with TEM-40 in hospital sewage and WWTP; ST621 with IMP-13 in hospital sewage; ST620 with GES-1 in WWTP, while an unknown ST was  correlated with VIM-2 in a sewage strain. The singletons ST252, 254, 244, 395 were not CP or ESBL producers. A class 1 integron (qacE∆1 integron-associated gene) was present in most of the identified Ab clones. This association could be the result of co-selection processes due to the spread of successful clones (such as ST2, ST636 and ST1 Ab) that were selected by antibiotic treatment in the hospital settings and were able to accumulate various CPs and ESBLs (i.e., OXA-23, OXA-72, TEM-12, ADC-30, ADC-74, ADC-73, ADC-81). Class 1 integrons were revealed in 80% of Pa strains and 50% of Pa belonging to the epidemic clones ST233, ST357 and the ST244.
Since the prevalent clones have a great potential for transmission among patients, the observation that those prevalent clones are correlated with the presence of class 1 integrons suggests that those isolates could be a significant reservoir of ARGs and can persist in the environment.
One of the limitations of this study arises from the fact that we have selected, using antibiotic supplemented culture media, only the resistant Ab and Pa strains, while the total population structure of these pathogens (including the antibiotic-sensitive strains) could not be assessed by this approach. However, we have isolated non-MDR strains in few cases (e.g., Ab strains from Cluj WWTP I and Pa strains from Târgoviște, Vâlcea, Iași hospitals and from Iași WWTP G and Timisoara WWTP F), but these isolated were not investigated at the genetic level (Additional file 1: Tables 1 and 2).

Conclusion
Our study emphasized the presence of carbapenemresistant MDR Ab and Pa belonging to international high-risk clones in all investigated sampling points (hospital units, their collecting sewage tanks and the sampled WWTPs) and the clonal dissemination of clinical Ab ST2 strains in the wastewaters. The reported data highlight the importance of the screening for acquired AMR in the environment and could provide important knowledge for monitoring the ARB and ARGs transmission from hospital into water bodies.