Carbapenemase genes in clinical and environmental isolates of Acinetobacter spp. from Quito, Ecuador

Clinical isolates

The National Reference Laboratory for Antimicrobial Resistance provided us a total of 52 Acinetobacter spp. clinical isolates recovered through the surveillance network from January to December 2015. As part of the standard procedure, the Reference Laboratory get back the Acinetobacter isolates from four different hospitals: Hospital Pediátrico Baca Ortiz (H1, n = 42), Hospital Pablo Arturo Suárez (H2, n = 5), Hospital del Sur Enrique Garcés (H3, n = 3), and Hospital San Francisco de Quito (H4, n = 2) in Quito (Fig. 1). The Reference Laboratory isolated Acinetobacter spp. from the following clinical samples: pleural fluid, cerebrospinal fluid, blood, urine, catheters, secretions, wounds, skin, abscesses, sputum, tracheal aspirates, and unidentified samples recovered from eleven units (ICU, general surgery, traumatology, burn unit, external consultation, neonatal unit, infectious diseases, pediatric unit, internal medicine, emergency, cardiology, and unidentified units). The data from each isolate are detailed in Table S1. The Reference Laboratory previously identified all isolates as A. baumannii based on VITEK 2 compact system® (Biomerieux, Marcy-I’Étoile, France).

We inoculated the frozen clinical isolates in sheep blood and MacConkey agars and incubated at 37 °C for 24 h. We used Gram staining, oxidase test, and catalase test to re-confirmed the identity of the isolates. We performed DNA extraction by boiling according to the protocol described by Cuaical et al. (2012); briefly, we boiled an overnight culture diluted in sterile water (500 μl) at 100 °C for 15 min, following centrifugation at 10,000 g for 2 min. We quantified the supernatant by spectrophotometry using the Nanodrop (Thermo Fisher Scientific, Waltham, MA, USA) and stored at −20 °C for subsequent amplification. We confirmed species identification by the amplification of partial RNA polymerase β (rpoB) gene and DNA gyrase subunit B (gyrB) gene according to La Scola et al. (2006) and Higgins et al. (2007, 2010a), respectively. The primers selected are detailed in Table 1.

River water collection and bacterial identification

We collected the water samples from five principal rivers: Monjas (MO, n = 6), Machángara (MA, n = 6), San Pedro (S, n = 6), Pita (P, n = 6), and Guayllabamba (G, n = 3) in March and April 2016 (Fig. 1 and Table S1). We chose those rivers for two reasons: (i) the four rivers are the main ones that run through Quito, all of them flow into the Guayllabamba River, and (ii) sewage is released into the city’s rivers without prior treatment. In addition, we chose the number of water collection points based on the accessibility and proximity to the urban center. We collected 100 ml of water samples at each river point and transported to the laboratory with ice packs according to Zhang et al. (2013) with few modifications. Subsequently, we centrifuged the water samples at 1,000 g for 10 min at room temperature, resuspended the pellets with 5 ml of fluid thioglycollate medium, and incubated at 37 °C overnight. We inoculated thirty microliters of cultured sample onto CHROMagar Acinetobacter plates with antimicrobial selective supplement (CR 202; CHROMagar Company, Paris, France) and incubated at 37 °C overnight. We selected all colonies with red pigmentation and identified through morphological and conventional biochemical tests used for the clinical isolates. We identified the positive potential isolates using VITEK-2 compact system® (Biomerieux, Marcy-I’Étoile, France). We confirmed the species identification by the amplification of DNA gyrase subunit B (gyrB) gene according to Higgins et al. (2007, 2010a).

Antimicrobial susceptibility test

We use the disc diffusion method and the EDTA sinergy test to determinate the antimicrobial susceptibility and the production of metallo β-lactamases of clinical and river Acinetobacter spp. isolates (Clinical Laboratory Standards Institute, 2022). We spread a concentration corresponding to 0.5 of the McFarland scale of each isolate on Muller Hinton agar plates and incubated at 37 °C overnight. We included the following eleven antimicrobial susceptibility test discs (OXOID, Waltham, MA, USA): ampicillin/sulbactam (20 μg), ceftazidime (30 μg), ciprofloxacin (10 μg), imipenem (10 μg), meropenem (10 μg), gentamicin (10 μg), amikacin (30 μg), cefepime (30 μg), trimethoprim/sulfamethoxazole (25 μg), piperacillin/tazobactam (110 μg), and tobramycin (30 μg). We interpreted the resistance based on the susceptibility breakpoints of the Clinical Laboratory Standards Institute (2022). To assess the synergistic inhibition, we included an ethylenediaminetetraacetic (EDTA) disc (5 mg) with two carbapenem discs, imipenem (10 mg) and meropenem (10 mg), located at 2 cm from the center of the EDTA disc.

Molecular identification of carbapenems-resistant genes

We performed two different multiplex PCRs to screen the presence of carbapenems-resistant genes including six genes encoding enzymes type serine β-lactamases (bla_GES, bla_OXA-23, bla_OXA-24, bla_OXA-51, bla_OXA-58, and bla_OXA-143) and three genes encoding metallo-β-lactamases (bla_SPM, bla_SIM, and bla_GIM) using specific primers previously described (Ellington et al., 2007; Woodford et al., 2006) (Table 1). We achieved a PCR amplification in a final volume of 25 μl containing 1X of Master Mix Green GoTaq® (12.5 μl) (Promega, Madison, WI, USA), 10 μM of forward and reverse primers (0.5 μl), and 50 ng DNA (1 μl). To each PCR assay, we put positive, internal, and negative controls. Both multiplex PCR reaction conditions are detailed in Table 2. We identified all PCR products through electrophoresis in a 2% agarose gel stained with GelstarTM Nucleic Acid Gel Stain 10000 X staining (Lonza, Rockland, ME, USA) and visualized using Gel Doc XR system (Bio-Rad, Hercules, CA, USA).

We performed a conventional PCR to amplified the entire sequence of the following genes bla_OXA-23, bla_OXA-24, bla_OXA-51, and bla_OXA-143 genes using primers previously reported (Heritier, Poirel & Nordmann, 2006; Afzal-Shah, Woodford & Livermore, 2001; Jeon et al., 2005; Woodford et al., 2006) (Table 1), following the conditions detailed in Table 2. We cleaned the PCR products using exonuclease I and shrimp alkaline phosphatase (ExoSAP-IT®). We send both DNA strands of all positive bla_OXA genes for sequencing to Macrogen (Seoul, Korea). We aligned manually both forward and reverse sequences for editing and cleaning to obtain a single consensus sequence with the Clustal W program implemented in the MEGA 7 software (Kumar, Stecher & Tamura, 2016). We performed the identification of sequences obtained using a basic local alignment search tool (BLAST) (Altschul et al., 1997). We determined the number of sequences of bla_OXA-51 and bla_OXA-23 genes with DnaSP v.5.10 software (Rozas et al., 2010).

Identification of Acinetobacter spp.

Acinetobacter spp. clinical isolates originated mainly proceeded from pediatric population (77%; 40/52), followed by the adult population (19.2%; 10/52). Isolates derived principally from patients admitted in UCI 32.69% (17/52), followed by unidentified units 15.38% (8/52) and general surgery 13.46% (7/52) (Fig. 2A). According to the clinical specimens, the isolates came from tracheal aspirates 28.85% (15/52), followed by catheters 11.54% (6/52) and abscesses 9.62% (5/52) (Fig. 2B). The majority of clinical isolates were identified as A. baumannii (96.15% 50/52) and A. pittii (3.85%; 2/52) according rpoB gene. Only the identity of A. pittii (3.85%; 2/52) was confirmed with both markers, rpoB and gyrB genes. A. baumannii was found in all type of clinical specimen from all hospital units while A. pittii was found only in tracheal aspirates from UCI. The biochemical and molecular identification of each isolate is detailed in Tables S2 and S3. The nucleotide and amino sequences of partial rpoB gene are detailed in Informations S1 and S2. We submitted all rpoB sequences to GenBank with assigned accession numbers OP796738–OP796781.

A total of 27 water samples were collected from five rivers. Twenty-two samples presented suggestive Acinetobacter isolates, four samples (RMO3, RS5, RP4, and RG3) showed no growth, and one sample (RMO4) had isolates not related to Acinetobacter spp. We recovered twenty-six suggestive isolates from San Pedro (2), Machángara (4), Pita (1), and Monjas (3) rivers, of which only 38.5% (10/26) were identified as Acinetobacter spp. based on VITEK 2 compact® (Biomerieux, Marcy-I’Étoile, France) and as A. baumannii based on gyrB gene amplification. The remaining corresponded to other genera (Pseudomonas aeruginosa: 34.6%, Stenotrophomonas maltophila: 19.3%, Burkkolderia cepacia: 3.8%, and Moraxella group: 3.8%). The biochemical and molecular identification of each isolate is detailed in Tables S2 and S3.

Phenotypic antimicrobial susceptibility

Clinical isolates presented the higher resistance (100%) to the three antibiotics: imipenem, meropenem, and piperacillin/tazobactam. Forty-eight percent were susceptible to tobramycin (Fig. 3A). The river isolates displayed an antimicrobial profile that varied between resistant and intermediate to eight antibiotics tested. Nevertheless, the main resistance was to piperacillin/tazobactam (100%). All of them were susceptible to tobramycin and ampicillin/sulbactam (100%) (Fig. 3B). According to the synergistic inhibition, both river and clinical isolates did not present metallo-β-lactamases enzymes. The antimicrobial susceptibility profile of the clinical and river isolates is detailed in Table S2.

Molecular identification of carbapenem-resistance genes

According the screening of carbapenem-resistance genes, clinical isolates presented the following genes: bla_SIM, bla_SPM, bla_GES, bla_OXA-23, bla_OXA-24, bla_OXA-51, and bla_OXA-143; whereas, seven river isolates showed bla_OXA-51 gene (Table S3). The bla_SIM gene, co-harbored with the bla_OXA-72 and bla_OXA-143 genes, was the least frequently detected (1.9%; 1/52; 15-1175) in A. pittii. The bla_SPM gene, co-harbored with the bla_OXA-51 and bla_OXA-23 genes, was detected in 13.46% (7/52; 15-0122, 15-0591, 15-690, 15-780, 15-795, 15-853, 15-1,138) in A. baumannii. The bla_GES gene, co-harbored with the bla_OXA-51, bla_OXA-23, bla_OXA-72, bla_OXA-143, and bla_SIM genes was present in 11.53% (6/52) in A. baumannii (15-0669, 15-690, 15-853, 15-900) and A. pittii (15-691, 15-1,175). The bla_OXA-23 gene, co-harbored with the bla_OXA-51 gene, was found in 98.07% (51/52) present in A. baumannii. The bla_OXA-24 gene, co-harbored with both bla_OXA-23 and bla_OXA-51 genes, was present in 9.6% (5/52) in A. baumannii (15-0659, 15-0985, 15-1064, and 15-1367) and in A. pittii (15-0691). The bla_OXA-143 gene was the least frequently detected (1.9%; 1/52; 15-1,175); this gene was co-harbored with the bla_OXA-24 gene in A. pittii.

Carbapenem-resistance genes sequences analysis

We reported 47 bla_OXA-51 sequences: 42 complete sequences of 813 bp and five partial sequences of 340 bp from A. baumannii clinical isolates. Of bla_OXA-51 sequences, we identified two genes encoding OXA-65 (n = 44), OXA-70 (n = 2) and one unidentified (MF594746). Forty complete sequences (MF594725–MF594728, MF594730–MF594737, MF594735, MF594739, MF594741–MF594745, OP554146–OP554168) and four partial sequences (MF594729, MF594734, MF594736, and MF594737) shared 100% of nucleotide identity with bla_OXA-65 of A. baumannii (GenBank accession number CP033869.1); whereas two entire sequences (MF594740 and MF594747) shared 100% nucleotide identity with bla_OXA-70 of A. baumannii (GenBank accession number NG_049811.1).

We found 48 bla_OXA-23 sequences: 42 complete sequences of 820 bp and six partial sequences of 471 bp from A. baumannii clinical isolates. Of bla_OXA-23 sequences, we identified two genes encoding OXA-23 (n = 38) and OXA-366 (n = 10). Thirty-two entire sequences (MF594756–MF594759, MF594762–MF594767, MF594769, MF594772–MF594777, OP554169, OP554170, OP554172–OP554178, OP554180–OP554182, OP554185–OP554187, OP554189, OP554191–OP554193) and six partial sequences (MF594755, MF594760, MF594761, MF594768, MF594770, and MF594771) shared 100% of nucleotide identity with bla_OXA-23 of A. baumannii (GenBank accession number CP110468.1). Ten complete sequences (MF594757, MF59474–MF59476, OP554171, OP554179, OP554183, OP554184, OP554188, and OP554190) shared 100% nucleotide identity with bla_OXA-366 of A. baumannii (GenBank accession number NG_049659.1).

We reported five bla_OXA-24/40 entire sequences of 817 bp (MF594778, MF594779, MF594781, MF594782, and MF594783) and one partial sequence of 204 bp (MF594780). Of bla_OXA-24/40 sequences, we identified one gene encoding OXA-72 present in three A. baumannii and two A. pitti clinical isolates that shared 100% of nucleotide identity with bla_OXA-72 of A. baumannii (GenBank accession number MN495626) and also A. pittii (GenBank accession number MN481287.1).

Sequencing of the bla_OXA-143 PCR product obtained from A. pitti clinical isolates identified gene encoding OXA-499. Our bla_OXA-143 sequence (MF594724) shared 99% nucleotide and amino acid identity with bla_OXA-499 of A. pittii (GenBank accession numbers NG_049775.1 and ALM96709.1). The nucleotide sequence presented one specific change in nucleotide 280, first position of the codon (GTG) (ATG), resulting in a valine (V) to methionine (M) change in position 94 (Fig. 4).

Sequencing of the bla_OXA-51 PCR products obtained from A. baumannii river isolates identified two genes encoding OXA-65 and OXA-259. Four partial sequences (MF594748–MF594751) from three rivers (Pita, Machángara, and San Pedro) shared 100% of nucleotide identity with bla_OXA-65 of A. baumannii (GenBank accession number OL961415.1) and three partial sequences (MF594752–MF594754) from Machángara River shared 100% of nucleotide identity with bla_OXA-259 of A. baumannii (GenBank accession number CP053098.1).

The nucleotide and amino sequences of bla_OXA genes are detailed in the Informations S3 to S10. We submitted all sequences to GenBank with assigned accession numbers MF594724–MF594783, OP554146–OP554193.

Acinetobacter spp. in clinical and river isolates

Our findings indicated that Acinetobacter baumannii is the most common species found in clinical isolates from tracheal aspirates from ICU. Similarly, studies reported high rates of baumannii species in ICU samples elsewhere (Qin et al., 2021; Bakhshi et al., 2022). Previous Ecuadorian studies reported high percentages of A. baumannii from respiratory tract samples, bronchial secretions, or tracheal aspirates (Jiménez, 2013; Nuñez et al., 2016; Villacís et al., 2019). On the other hand, we found A. baumannii in the most polluted river environments, Machángara and Monjas rivers (Yungán, 2010; Roldós, 2015). Similar, some studies found these bacteria in river water as well as sewage water (Hrenovic et al., 2016; Kisková et al., 2023). Pathogens of clinical origin as A. baumannii can reach aquatic environments as a result of different human activities, such as sewage water discharge directly into the river without pretreatment (Silva et al., 2016). Acinetobacter spp. survive in wastewater contributing to the spread clinical antibiotic resistance genes in aquatic systems (Kisková et al., 2023; Tobin et al., 2024). Particularly, hospital sewage water is considered a major source of multidrug-resistant pathogenic bacteria and antibiotic residues (Zhang et al., 2013). Contrary, we did not isolate any Acinetobacter spp. from the Guayllabamba River probably due to the distance (2 km), the collection points were far away from the wastewater reservoirs of urban areas and the bacterial concentrations decreases and distribution changes (Zhang et al., 2009).

Carbapenem-resistant Acinetobacter spp.

In our study, the clinical and environmental Acinetobacter isolates showed different susceptibility profiles. Clinical isolates exhibited a high rate of resistance to imipenem and meropenem; while, river isolates displayed a low resistance only to imipenem. In general, clinical isolates have high resistance to both antibiotics (Cortivo et al., 2015). The percentage of resistance found in this study was higher compared with an Ecuadorian study (Jiménez, 2013). River isolates are usually more sensitive compared to clinical isolates. The difference in the patterns of resistance found can be due to various factors such as, biological conditions, virulence factors, epidemiology, or place of origin (Karmostaj, Najar & Salmanian, 2013). In general, it is expected that the antibiotic concentration is lower in aquatic environments than in clinical environments (Zhang et al., 2009; Zhang et al., 2013). In clinical environments, antibiotics (in particular broad-spectrum antibiotics) are supplied continuously, promoting selective pressure in clinical isolates. Following this, the transmission of genetic determinants among clinical isolates could take place through horizontal gene transfer (Zhang et al., 2009; Castanheira, Mendes & Gales, 2023). Nonetheless, the possibility of clonal dissemination as a mean of increasing antibiotic resistance, should not be discarded. This situation generates a faster antibiotic resistance evolution in clinical isolates than in environmental isolates (Zhang et al., 2013). However, the susceptibility results of environmental samples should be treated with caution, as these profiles were based on clinical standards (CLSI) and a small sample size (n = 10).

Carbapenemases genes in Acinetobacter spp.

The high percentage of carbapenem-resistance found in clinical isolates may be increased due to the presence of bla_OXA-51, bla_OXA-23, bla_OXA-72, and bla_OXA-143 genes. According to some studies, the association of bla_OXA-23 and bla_OXA-51 genes as the main cause for carbapenem resistance (i.e., Martínez & Mattar, 2012; Teixeira et al., 2013; Cortivo et al., 2015). The bla_OXA-72 gene increases its resistance activity in the presence of imipenem and ampicillin/sulbactam (Kuo et al., 2013). The bla_OXA-143 and bla_OXA-72 genes in association with intrinsic mechanisms as the alteration of the of the outer membrane proteins promotes a high resistance to meropenem (Mostachio et al., 2012; Sen & Joshi, 2015). The low percentage of carbapenem-resistance found in environmental isolates may be explained by the absence of bla_OXA genes. It is necessary the presence of insertion elements to promote overexpression of the bla_OXA-51 gene since this gene it generates a weak resistance (Higgins, Zander & Seifert, 2013; Teixeira et al., 2013; Sen & Joshi, 2015). The bla_OXA-51 overexpression would be linked to the resistance to imipenem in A. baumannii river isolates (Martínez & Mattar, 2012; Evans, Hamouda & Amyes, 2013). It would be important to know about possible ISAba1 insertions upstream of the gene or alternative genes conferring resistance in the five isolates with resistance phenotype.

We found a high percentage of bla_OXA-23 and bla_OXA-51 genes in A. baumannii and A. pittii clinical isolates. The high percentage of bla_OXA-23 gene found in A. baumannii, is related to its distribution worldwide since a single sequence was found in the study. The bla_OXA-23 gene distribution has been associated with a clonal spread (Higgins et al., 2010b; Labarca et al., 2016). In the other hand, bla_OXA-51 gene has been reported principally in baumannii species (Lee et al., 2012; Teixeira et al., 2013; Périchon et al., 2014). Since bla_OXA-51 gene can be found in the chromosome (Poirel, Naas & Nordmann, 2010), some authors consider it as an intrinsic marker for A. baumannii identification (Higgins et al., 2010b; Périchon et al., 2014). However, the gene may be in transposons and plasmids (Liu et al., 2015; Wang et al., 2015; Sennati et al., 2016). Its dissemination among other Acinetobacter species from different environments is possible, suggesting that the dissemination may not be by clonal strains (See Labarca et al., 2016).

Our findings detected two different bla_OXA-51 variants OXA-65 and OXA-70 were circulating in the four hospitals; while, in rivers probably are OXA-65 and OXA-259, suggesting that the genes were under selective pressure. Similarly, previous studies OXA-65 was reported in A. baumannii clone ST79 from Guayaquil (Rodríguez et al., 2016). In Latin America, is one of the variants found in A. baumannii in Brazil, Argentina, Paraguay, Mexico, and Honduras (Rodríguez et al., 2016; Nodari et al., 2020; Graña-Miraglia et al., 2020). There is not report about the presence of OXA-70 neither OXA-259 variant in South America. The information about OXA-70 variant is scare worldwide. OXA-70 has been reported in Hong Kong, Malaysia, Ghana, and Honduras (Alattraqchi et al., 2020; Ayibieke et al., 2020; Brown & Amyes, 2005; Galac et al., 2020). OXA-70 belong to the most diverse collection of class D carbapenemases, subgroup 3 that is the major mechanism of carbapenem resistance in A. baumannii (Brown & Amyes, 2005). OXA-259 variant has been reported in A. baumannii clinical isolates from China, Tanzania, and Thailand (Jia et al., 2019; Moyo et al., 2021; Chukamnerd et al., 2022) as well in bulk tank milk in Germani (Sykes et al., 2023). The variant OXA-259 is located on the chromosome of A. baumannii and constitute an intrinsic oxacillinase genes with low-level carbapenemase activity that naturally occur (Chukamnerd et al., 2022), due there is a not ISA_balike elements (Fedrigo et al., 2022). To the best of our knowledge, this is the first report of the OXA-70 and OXA-259 in A. baumannii in Ecuador.

Our findings showed the presence of bla_OXA-366 gene in both species A. baumannii and A. pittii clinical isolates from Ecuador. The information about this variant is scare. GenBank previously reported a sequence of bla_OXA-366 in A. baumannii from USA. Whereas the OXA-366 variant has been reported also in A. baumannii clinical isolates in Iran (Bahador et al., 2015). To the best of our knowledge, this is the first report of the OXA-366 in A. baumannii and A. pittii in South America.

The bla_OXA-72 gene co-harbored with bla_OXA-23 and bla_OXA-51 genes was identified in low levels in A. baumannii and A. pittii. This variant was documented previously in A. baumannii clinical isolates from Guayaquil (Nuñez et al., 2016; Rodríguez et al., 2016) and from neighboring countries Colombia, Peru, and Brazil (Saavedra et al., 2014; Saavedra et al., 2017; Werneck et al., 2011; de Sá Cavalcanti et al., 2013; Levy-Blitchtein et al., 2018; Nodari et al., 2020) and in A. pittii from Colombia (Montealegre et al., 2012), Brazil (Brasiliense et al., 2019), European and Asian countries (Al Atrouni et al., 2016b; Bonnin et al., 2014). The presence of this gene in two species may be due to the bla_OXA-72 gene spreads through horizontal gene transfer mediated by plasmids (Montealegre et al., 2012; Kuo et al., 2013; Nuñez et al., 2016; Rodríguez et al., 2016). To the best of our knowledge, this is the first report of the bla_OXA-72 gene in A. pittii in Ecuador.

This is also the first report of the bla_OXA-143 gene in A. pittii clinical isolates from Ecuador. In South America, this gene and its variants (OXA-231, OXA-253, and OXA-255) have been reported in A. baumannii clinical isolates in Brazil (Higgins et al., 2009; Gionco et al., 2012; Girlich et al., 2014), Perú (Levy-Blitchtein et al., 2018), and Colombia (Saavedra et al., 2017). The variant found in this study is very different from OXA-143 variants reported in A. baumannii from Brazil, Honduras, and Korea (Higgins et al., 2009; Zander et al., 2014; Kim et al., 2010). Interestingly, the sequence alignment showed a high similarity between our variant with two OXA-143 variants (OXA-499 and OXA-255) described to A. pittii in Korea (D’Souza et al., 2017) and United States (Zander et al., 2014). In particular, the difference between the variant reported in this study differs by a change in amino acid sequence with OXA-499 reported by D’Souza et al. (2017). To the best of our knowledge, this is the first report of a variant similar to OXA-499 in A. pittii in South America.