Diversity of bla_POM in carbapenem-resistant opportunistic pathogenic Pseudomonas otitidis in municipal wastewater

Since the first isolation of carbapenemase and metallo-β-lactamase (MBL) from Pseudomonas aeruginosa more than 20 years ago (Watanabe et al. 1991), the detection of carbapenem-resistance bacteria has been increasing at an alarming rate globally (Cornaglia et al. 2011). MBLs are recognized to be clinically important as resistance determinants because of their potent carbapenemase activities, including the broad substrate specificity covering most β-lactam families (Cornaglia et al. 2011). MBLs belong to class B β-lactamases with three major structural subclasses (Bush & Jacoby 2010), encoded either by genes in the chromosomal framework (resident MBL) or by genes acquired through horizontal gene transfer (acquired MBL) (Cornaglia et al. 2011). Intrinsic resistance to carbapenems is less common (Meletis 2016), which has been detected in species of clinical significance including Bacillus spp., Stenotrophomonas maltophilia, Aeromonas spp., Bacteroides fragilis, various flavobacteria, and Pseudomonas otitidis (Walsh et al. 2005; Thaller et al. 2011).

P. otitidis was first found to intrinsically possess an MBL names POM (P. otitidisMBL) isolated from humans infected with acute otitis externa (Clark et al. 2006). POM-1 is a subclass B3 MBL, which demonstrates an overall broad substrate specificity, especially for penicillins and carbapenems (Borgianni et al. 2015). While POM is highly conserved in P. otitidis (Thaller et al. 2011), P. otitidis is genetically closely related to P. aeruginosa (Clark et al. 2006) and phenotypically unable to differentiate the two species (Kim et al. 2016). P. otitidis has been isolated not only from clinical settings (Clark et al. 2006; Lee et al. 2012; Kim et al. 2016; Caixinha et al. 2021) but also from food (Wong et al. 2015), animal carcass (Vieira et al. 2020), and freshwater (Rodríguez-Verdugo et al. 2012; Tacao et al. 2015; Miyazaki et al. 2020).

Wastewater treatment plants (WWTPs) are considered a major source of bacterial antibiotic resistance to the environment (Michael et al. 2013). This is not surprising as sewage contains the microbiomes of human populations (Newton et al. 2015), and the discharge of treated wastewater degrades water quality and cannot be used for potable water and industrial applications directly (Panagopoulos 2022). Most studies (Picão et al. 2013; Galler et al. 2014; Hrenovic et al. 2016; Naquin et al. 2017; Basode et al. 2018; Cooper et al. 2021) on carbapenem-resistant bacteria (CRB) in the WWTP were focusing on carbapenem-resistant Enterobacteriaceae (CRE) as they are considered urgent threats to public health (CDC 2019). On the other hand, carbapenem resistance in non-Enterobacteriaceae bacteria in wastewater has not been extensively studied, even though some of these non-Enterobacteriaceae can also be human pathogens (Thaller et al. 2011; Woodford et al. 2014). While the study of antibiotic-resistant P. otitidis in wastewater is very limited, other researchers studied the use of P. otitidis isolated from wastewater for bioelectricity generation (Thulasinathan et al. 2019), triphenylmethane dyes decolorization (Jing et al. 2009), and biodegradation of sodium dodecyl sulfate (SDS) (Ibrahim & Abd Elsalam 2018).

In this study, we aim to identify opportunistic pathogenic carbapenem-resistant non-Enterobacteriaceae possessing MBL in the wastewater and characterize the MBLs phenotypically (minimum inhibition concentration (MIC) and sensitivity of hydrolyzation of meropenem) and genetically (16S rRNA, pulse-field gel electrophoresis, PFGE; POM-1 amino acid alignments).

Raw municipal wastewater samples (500 mL, n = 3) and activated sludge samples (500 g, n = 3) were collected from the East Honolulu WWTP (Honolulu, Hawaii) in April 2016. The abundance of total heterotrophic bacterial biomass, Escherichia coli, and CRB were enumerated by the direct plating method on Luria-Bertani (LB) and mTEC agar with/without meropenem (2 mg/L; TCI America, Portland, OR, USA). Well-mixed wastewater samples were serially diluted (10⁰ to 10⁻⁷) in sterilized distilled water, and 100 μL of the serial dilutions were plated in triplicate on LB, LB + meropenem, mTEC, and mTEC + meropenem agars. LB and LB + meropenem agar plates were incubated at 30 °C for 24 h, while mTEC and mTEC + meropenem agar plates were incubated at 44.5 °C for 24 h. Colony-forming units (CFU/mL and CFU/g) were counted.

Seventeen individual bacterial colonies that grew on the mTEC agar with 2 mg/L of meropenem were randomly selected and further purified by streaking on mTEC + meropenem (2 mg/L) agar and incubated at 37 °C for 24 h. These colonies did not show the characteristic blue color of E. coli on the mTEC agar. The meropenem-resistant bacteria isolates were collected and stored in Tryptic Soy Broth (TSB) (Fluka Biochemika, Buchs, Switzerland) with 15% (v/v) glycerol at −80 °C for further analysis.

MICs of meropenem of the meropenem-resistant bacteria isolates were determined by using the microbroth dilution assay described by the CLSI (CLSI 2015). Single colonies were inoculated in 10 mL of LB broth + meropenem (2 mg/L) and incubated overnight at 37 °C at 180 rpm. Then, 100 μL of the bacteria-inoculated broth was transferred to a new tube containing 10 mL of LB broth + meropenem (2 mg/L) and incubated at 37 °C until the 0.5 McFarland standards were achieved. Approximately 200 μL of cation-adjusted Mueller Hinton broth (CAMHB) (BD; MD, USA) containing 2–48 mg/L of meropenem were distributed in a 48-well cell culture plates and inoculated with the 0.5 McFarland standards grown isolates. The 48-well plates were incubated at 37 °C for 24 h.

The whole-cell PFGE method on the 17 meropenem-resistant bacterial isolates was conducted using the standard operating procedure from PulseNet PFGE (CDC 2013). Briefly, the CRB isolates were inoculated in LB broth + meropenem (4 mg/L) at 37 °C overnight, and the cells were collected and resuspended in cell suspension buffer (100 mM Tris–HCl/100 mM EDTA, pH 8.0) to achieve a concentration of OD₆₁₀ 0.8–1.0. Two hundred μL of these cell suspensions was embedded in 0.5% Seakem Gold agarose (Lonza, Allendale, NJ, USA). These PFGE plugs were then incubated in cell lysis buffer (50 mM Tris–HCl/50 mM EDTA, pH 8.0, and 1% Sarcosyl) with proteinase K (1 mg/mL) at 55 °C for at least 1.5 h. Restriction digestion of genomic DNA within the agarose plugs was conducted using Xba I (50 U/sample). Electrophoresis was conducted on a CHEF Mapper (Bio-Rad, Richmond, CA, USA) for 18 h at a 10–700 kb molecular weight auto algorithm. Analysis was performed using BioNumerics version 5.10 (Applied Maths, Austin, TX, USA).

Twelve out of the 17 meropenem-resistant bacterial isolates (MR1, MR2, MR4, MR6, MR7, MR8, MR9, MR11, MR14, MR15, MR16, and MR17) that exhibited distinct PFGE patterns were selected for subsequent 16S rRNA gene and POM gene amplification and sequencing. The isolates exhibited the same PFGE patterns and were considered clonal. Briefly, DNA extractions were performed by using GenElute™ Bacterial Genomic DNA Kits (MilliporeSigma, St. Louis, MO, USA) according to the manufacturer's protocol. In short, the CRB strains were inoculated in LB broth + meropenem (4 mg/L) and incubated at 37 °C overnight at 180 rpm. The solutions were centrifuged at 3,000 rpm and the cells were collected and proceeded to DNA extractions. The DNAs extracted were kept at −20 °C for further PCR analysis. 16S rRNA was amplified using the universal oligonucleotide primers 27F (5′-AGA GTT TGA TCC TGG CTC AG-3′) and 518R (5′-ATT ACC GCG GCT GCT GG-3′), and POM-1 genes were amplified using primers designed in this study: lactaF (5′-GCA TTG ACC TGC GCG ACC AGG CAG T-3′) and lactaR (5′- CTT GTC GGC GTA GGC CTT GCA GCT-3′) in a GeneAmp^® PCR System 9700 (Applied Biosystem, Beverly, MA, USA). The PCR mixture with a total reaction volume of 25 μL was comprised of 2.5 μL of 10× PCR buffer, 1.5 mM MgCl₂, 0.04 U/μL of Taq Polymerase, 0.1 mM dNTP, 0.4 mg/mL of BSA, 0.15 μM of each forward/reverse primer pairs, 1 μL of DNA template, and nuclease-free water. The thermocycling program included DNA polymerase activation and initial denaturing at 95 °C for 5 min, thermo cycles (16S rRNA = 45 cycles; POM-1 = 35 cycles) of denaturation at 95 °C for 30 s, annealing (16S rRNA = 56 °C for 30 s; POM-1 = 63 °C for 30 s), and extension (16S rRNA = 72 °C for 45 s; POM-1 = 72 °C for 50 s), following by a final extension at 72 °C for 8 min. The PCR amplicon was separated on 1.5% agarose gel through gel electrophoresis and then visualized by a UVP GelStudio (Analytik Jena, Upland, CA, USA). PCR amplicons were gel purified using the QIAquick^® Gel Extraction Kit (Qiagen, Valencia, CA, USA) and proceeded with sequencing the reaction using the ABI 3730XL sequencer at the Advanced Studies in Genomics, Proteomics and Bioinformatics (ASGPB), University of Hawaii at Manoa. The sequence reads were quality trimmed and checked manually against the chromatograms by using the Sequence Scanner (version 2.0; Thermo Fisher, Waltham, MA, USA), and the species and antibiotic resistance gene identifications were confirmed by using BLASTn. The 16S rRNA sequences (n = 12) and bla_POM sequences (n = 12) were deposited in the GenBank database, accession numbers ON944110–ON944121 and ON959567–ON959578, respectively.

Partial and full amino acid sequences (n = 31) of bla_POM from the GenBank database were downloaded to compare with the bla_POM sequences detected from wastewater isolates in this study (n = 12). The list of reference genes was summarized in Supplementary Table S1. An unrooted neighbor-joining phylogenetic tree of bla_POM amino acid sequences was created using iTOL, Interactive Tree of Life (https://itol.embl.de/). The amino acid sequences of 12 bla_POM isolated from wastewater (this study) and 9 bla_POM reference genes were randomly selected and performed multiple sequence alignments using Clustal Omega by EMBL-EBI (https://www.ebi.ac.uk/Tools/msa/clustalo/).

The sensitivity and speed of hydrolysis of meropenem were measured by using a modified Kirby-Bauer disk diffusion susceptibility test protocol (Hudzicki 2009). Details of the production of the standard curve are summarized in the Supplementary Material. Five P. otitidis isolates (MR1, MR8, MR11, MR16, and MR17) were inoculated in 5 mL of LB broth + meropenem (2 mg/L) and incubated at 37 °C in a shaking incubator overnight. A total of 2 mL of the overnight cultures were transferred to a 10 mL fresh LB broth + meropenem (2 mg/L) and incubated at 37 °C in a shaking incubator overnight. The solutions were centrifuged at 3,000 rpm to collect the bacteria pellets. Approximately 5 mL of the OD₆₀₀ calibrated bacteria suspension (0.2) were transferred to a new tube and added meropenem solution to a final concentration of 64 mg/L. The tubes were incubated at 22 °C in a shaking incubator overnight at 350 rpm. About 100 μL of the samples were extracted every 1 h (up to 12 h), added to a microcentrifuge tube, and centrifuged at 13,000 rpm for 3 min to pellet the bacteria particles. Approximately 5 μL of the supernatant were pipetted and dropped on new sterile antibiotic disks plated on the E. coli K12 agar plates and incubated at 37 °C overnight. Each antibiotic disk was duplicated and each P. otitidis isolate meropenem hydrolysis test was performed at triplication. A negative control was included in the test where the tubes were incubated with meropenem (64 mg/L) without bacteria. The inhibition zones were measured and the concentrations of meropenem at each hour were measured according to the standard curve.

The enumeration of bacteria and CRB in wastewater and sludge samples is shown in Supplementary Table S2. The colony-forming unit (CFU) of general heterotrophic bacteria, E. coli, and non-E. coli other bacteria enumerated in wastewater and sludge samples ranged from 1.4 × 10⁵ ± 3.2 × 10⁴ to 1.9 × 10⁶ ± 3.3 × 10⁵CFU/mL and 3.3 × 10⁴ ± 1.1 × 10⁴ to 1.4 × 10⁶ ± 1.9 × 10⁵CFU/g, respectively. On the other hand, carbapenem-resistant heterotrophic bacteria detected in wastewater and sludge samples were 1.0 × 10⁴ ± 3.8 × 10³ CFU/mL and 3.7 × 10⁴ ± 1.1 × 10⁴CFU/g, respectively. On mTEC agar, carbapenem-resistant other bacteria detected were less than 100 colonies in both wastewater and sludge samples. Since there were no carbapenem-resistant E. coli colonies (Supplementary Table S2), 17 carbapenem-resistant colonies that did not show the characteristic E. coli blue color, but grew on the mTEC agar, were isolated. These bacterial isolates were subsequently identified as P. otitidis (MR1–MR11, MR13–MR18) by 16S rRNA gene sequencing (Table 1). PFGE fingerprinting patterns showed that among the 17 P. otitidis, five isolates were identical to one of the isolates, that is MR3 to MR1, MR5, MR10 to MR4, MR13 to MR11, MR18 to MR16 (Supplementary Figure S1). These identical isolates of P. otitidis were not included in the following bla_POM analysis. MICs of these 17 P. otitidis isolates showed that isolates MR6, MR7, and MR8 showed resistance to meropenem at 8 mg/L. The remaining P. otitidis isolates have MIC of 16 mg/L. PCR and sequencing of bla_POM showed that 12 P. otitidis (MR1, MR2, MR4, MR6, MR7, MR8, MR9, MR11, MR14, MR15, MR16, and MR17) were found to possess a subclass B3 MBL named POM (P. otitidis MBL).

All 12 bla_POM amino acids isolated from wastewater have one amino acid substitution at position 1, methionine (M) to valine (V), as compared to the bla_POM amino acid isolated from other sources (Figure 2). Interestingly, P. otitidis MR8 and MR11 have additional two amino acid substitutions, whereas P. otitidis MR7 has additional three amino acid substitutions. In P. otitidis MR8, valine (V) was substituted by methionine (M) at positions 18 and 154, respectively. Meanwhile, in P. otitidis MR11, arginine (R) was substituted by histidine (H) at position 44, and threonine (T) was substituted by asparagine (N) at position 189. P. otitidis MR7 has amino acid substitutions at position 48 (asparagine, N to aspartic acid, D), position 152 (valine, V to methionine, M), and position 232 (valine, V to an undetermined amino acid, X). The amino acid substitution at position 18 (V to M) of MR8 was matching with MC11140 isolated from clinical and SLBN-4 isolated from soil. These amino acid substitutions in isolates MR7, MR8, and MR11 supported the phylogenetic cluster analysis where they were closely or grouped in the clinical cluster. The key residues involved in metal binding in subclass B3 enzymes (His/Gln116, His118, His196, Asp120, His121, and His263) (Palzkill 2013) were conserved in all bla_POM isolates (shown in black arrow).

A study by Borgianni et al. (2015) showed that POM is a novel broad-spectrum subclass B3 MBL, which manifests high catalytic activity on carbapenems, including ertapenem and doripenem. This has also been proven by Thaller et al. (2011) where the expression of bla_POM-1 in E. coli and P. aeruginosa increased the MICs of the bacteria hosts to penicillins, cephalosporins, and carbapenems, except aztreonam. Bla_POM-1 was further proven to show notably high catalytic efficiencies, 2- to 14,000-fold higher k_cat/K_m values to imipenem and meropenem than by other subclass B3 enzymes (except bla_AIM-1 for imipenem) (Borgianni et al. 2015). They also found that some of the readings were higher than the globally spread-acquired MBL VIM-2 (Docquier et al. 2003). It was also suggested that bla_POM could be the earliest member within some Pseudomonas species to possess the chromosomal B3 MBLs, considering a similar enzyme bla_PAM-1 (72% similar at the sequence level) has been reported in Pseudomonas alcaligenes (Suzuki et al. 2014).

Thaller et al. (2011) also showed that the POM-1 enzyme has up to 64% amino acid similarity to several known variants of the L1 enzyme of S. maltophilia, which is higher than other subclass B3 enzyme, suggesting a closer ancestry to the L1 enzyme. They also found that bla_POM-1 has an overall high degree of sequence conservation, with an amino acid homology ranging from 97.0 to 100% among the aligned regions. In this study, the bla_POMs isolated from wastewater have an amino acid homology ranging from 98.4 to 99.6% compared with the type strain MC10330^T. Key residues involved in metal binding in subclass B3 enzymes observed by a previous study (Thaller et al. 2011) determined by the BBL numbering scheme (Garau et al. 2004) were found to be conserved in all the bla_POM isolates in this study (Figure 2). Furthermore, the number of amino acid substitutions in the POM enzyme does not determine the MICs to meropenem. According to Thaller et al. (2011) the POM enzymes isolated which have at least 1–3 amino acid substitutions have a wide range of MICs to meropenem (0.25 to >32 mg/L). For example, for an isolate MCC04511 that has one amino acid substitution, the MIC for meropenem was >32 mg/L; in contrary to MCC11140 that has three amino acid substitutions, the MIC for meropenem was 1 mg/L. In addition, MCC51196 has two amino acid substitutions, and three amino acid deletions have the lowest MIC for meropenem (0.25 mg/L). Wong et al. (2015) found that bla_POM isolated with one or two amino acid substitutions was resistant to meropenem with MICs >8 mg/L. Similar MICs were observed in this study where bla_POM from MR7 that has four and bla_POM from MR8 that has three amino acid substitutions have a MIC of 8 mg/L to meropenem, and bla_POM from MR11 that has three amino acid substitutions has a MIC of 16 mg/L to meropenem (Table 1).

Since the first publications of the presence of intrinsic β-lactamases in non-fermenting Gram-negative species (Richmond & Curtis 1974; Suginaka et al. 1975), the studies of intrinsic β-lactamases in non-Enterobacteriaceae have been increasing, including Aeromonas spp. (Bakken et al. 1988), Burkholderia cepacia complex (BCC) (Poirel et al. 2009), P. aeruginosa (Fajardo et al. 2014), Shewanella oneidensis (Poirel et al. 2004), and S. maltophilia (Saino et al. 1982). However, the study of this mechanism in P. otitidis was very limited. Carbapenem resistance could also be caused by other mechanisms including the overproduction of AmpC cephalosporinase, changes in the permeability of the outer membrane, and overexpression of efflux pumps (Dantas et al. 2017). Wong et al. (2015) found that while carbapenem resistance in P. otitidis was mainly caused by bla_POM, the overexpression of MexAB-OprM efflux pump and the absence of OprD porin were the reason for carbapenem resistance in P. aeruginosa; while the overexpression of the TtgABC efflux system in P. putida was responsible for carbapenem resistance. The permeability of β-lactams on the outer membrane barrier was also an important role in the hydrolysis of β-lactams by Gram-negative bacteria (Matsumura et al. 1999). A study done by Minami (1993) showed that in the addition of β-lactams (ceftazidime and cefpirome), P. aeruginosa changed from intermediate to highly resistant to imipenem. Although it is not scientifically proven, these carbapenem mechanisms found in P. aeruginosa could have evolved in P. otitidis in the future because these two species are genetically closely related as they are in the same genus Pseudomonas sensu strict (Peix et al. 2009). Other studies also found that originally chromosomally encoded β-lactamases were later found in plasmids, such as IMP-1 in P. aeruginosa (Watanabe et al. 1991) and IMI-2 in Enterobacter asburiae (Aubron et al. 2005) and Enterobacter cloacae (Yu et al. 2006).

This study focused on characterizing the bla_POM amplified from P. otitidis isolated from wastewater, which was found to be phylogenetically different from bla_POM isolated from other sources (e.g. clinics and freshwater) (Figure 1). While most of the studies of P. otitidis possessing bla_POM were focused on the clinical isolates (Clark et al. 2006; Thaller et al. 2011; Lee et al. 2012; Borgianni et al. 2015; Kim et al. 2016), the study of bla_POM in wastewater is very limited. The bla_POM isolated in wastewater was found to have more mutations (i.e. amino acid substitutions) than bla_POM isolated from patients in a previous study (Thaller et al. 2011) (Figure 2). Some P. otitidis isolates MR1, MR8, MR11, MR16, and MR17, which have more than two amino acid substitutions compared to the wild-type strain, were able to hydrolyze meropenem in a shorter time. These were not tested in previous clinical studies.

The prevalence of opportunistic pathogen P. otitidis encoding MBL POM isolated in the wastewater emerged as a substantial threat to public health because wastewater is known as the source of antibiotic-resistant bacteria and genes in the environment. This study isolated bla_POM encoding P. otitidis in wastewater, which was phylogenetically different from other sources (including freshwater, clinics, soil, and animal). Several isolates with multiple amino acid substitutions in the POM region compared to the wild-type strain showed a higher meropenem hydrolysis rate and were closely clustered to the clinical reference isolates in the phylogenetic tree. This phenomenon not only increases the risk for the development of acquired resistance, but it might also limit and complicate drug selections for treatments because P. otitidis is a clinically recognized pathogen.

Due to the lack of the study of bla_POM encoding P. otitidis in wastewater, our findings have shown the difference in its amino acid sequences compared to other isolation sources and its potential to hydrolyze carbapenems in a shorter time. This information is important for health authorities to formulate plans to prevent the spread in the healthcare system and the environment. More in-depth studies of P. otitidis and bla_POM should be done to fill in the knowledge gap of this opportunistic pathogen and MBL in wastewater and the environment.

D.Y.W.D. analyzed the data and wrote the manuscript. G.X.C. and C.Z. designed and performed the experiments. D.Y.W.D. and G.X.C. corrected and revised the manuscript. All authors read and approved the final manuscript. D.Y.W.D. and G.X.C. contributed equally to the manuscript. T.Y. conceived the idea, supervised the research, and reviewed the manuscript.

This material is based upon work supported by the National Science Foundation under grant no. CBET-2027059.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.