The bacterium DW18 was isolated from the sewage of an animal farm in Wenzhou, China. The collection and isolation methods for DW18 were as follows: a sterile cotton swab was dipped into sewage in the farm sewer line and then immediately placed into a sterile screw-cap specimen collection tube with saline solution. The isolate was streaked on Luria–Bertani agar plates to obtain single colonies, some of which were randomly selected and then preserved in LB broth with 25% glycerol at −80°C. DNA sequencing of DW18 was performed on the Illumina NovaSeq and PacBio RS II platforms (Shanghai Personal Biotechnology Co., Ltd., Shanghai, China). The Illumina short reads were assembled by SKESA v2.4.0 (Souvorov et al., 2018). Unicycler v0.4.8 was used for hybrid assembly of the PacBio long reads and confirming the cyclization of the whole-genome assembly (Wick et al., 2017). Pilon improved the quality of genomic assembly sketches by mapping Illumina short reads onto the assembly to correct possible incorrect assembly (Walker et al., 2014). Species identification was carried out first by 16S rRNA gene homology and then whole-genome average nucleotide identity (ANI) analyses (Konstantinidis and Tiedje, 2005; Richter and Rosselló-Móra, 2009). The calculation of ANI was performed by FastANI (Jain et al., 2018). The bacteria and plasmids used in this study are shown in Table 1.

The agar dilution method was used to determine the minimum inhibitory concentrations (MICs) of the antimicrobials. Susceptibility patterns were interpreted on the basis of the Clinical and Laboratory Standards Institute guidelines M100 (32nd Edition, 2022) for Enterobacteriaceae. The reference strain used for quality control was E. coli ATCC 25922. All 16 antimicrobials used in this work are shown in Table 2.

The promoter region and coding sequence of the novel resistance gene aph(3′)-Id were amplified by polymerase chain reaction (PCR). The primers and annealing temperatures used for PCR are listed in Table 3. The PCR products were inserted into T-Vector pMD19 (Simple) (Takara Bio, Inc., Dalian, China) using a T4 DNA ligase cloning kit (Takara Bio, Inc., Dalian, China). The recombinant plasmid was transformed into competent cells of E. coli DH5α by the calcium chloride method, and then LB broth was added, and the cells were resuscitated by shaking at 37°C. The transformants were screened on LB solid culture medium with 100 μg/mL ampicillin. For the expression of the APH(3′)-Id protein, the ORF of aph(3′)-Id was PCR-amplified using forward and reverse primers with BamHI and HindIII restriction sites, respectively, at the 5′-ends. The PCR product was digested with the restriction endonucleases HindIII and BamHI and then cloned into the pCold I vector treated with the same restriction endonucleases. The recombinant plasmid (pCold I-aph(3′)-Id) was transformed into competent E. coli BL21 cells by the calcium chloride method. The transformants were screened on LB agar plates with 100 μg/mL ampicillin. The cloned insert sequences of the recombinant plasmids were confirmed by PCR and Sanger sequencing.

The recombinant strain (pCold I-aph(3′)-Id/BL21) for the expression of APH(3′)-Id was cultured in 3 mL of LB broth with 100 μg/mL ampicillin at 37°C for 16 h. Then, 1 mL of the bacterial culture was added to 100 mL of LB broth for continuous culture until the OD₆₀₀ value reached 0.6–0.8 (Qing et al., 2004). After refrigeration for half an hour, a final concentration of 1 mM IPTG was added to induce the expression of APH(3′)-Id. Cultivation was continued for 16 h at 18°C, and the culture was centrifuged to collect cells. The cells were resuspended in 4 mL of nondenatured lysate (50 mM NaH₂PO₄, 300 mM NaCl) and disrupted by ultrasonication. The recombinant protein in the supernatant collected after centrifugation at 8,000 × g for 30 min was purified using His-tag Purification Resin and eluted by the nondenatured eluent (300 mM NaCl, 50 mM imidazole, 50 mM NaH₂PO₄) according to the instructions of the His-tag Protein Purification Kit (Reductant & Chelator-resistant) (Beyotime, Shanghai, China). The His tag of the recombinant protein was removed by thrombin with incubation at 20°C for 16 h. The purity of APH(3′)-Id was confirmed by SDS–PAGE, and the concentration was spectrophotometrically examined at 562 nm by the standard curve established using a BCA protein assay kit (Beyotime, Shanghai, China). The protein was stored in dissolution buffer (0.5 M NaCl, 20 mM Tris–HCl; pH 8.0) and preserved at −80°C for subsequent enzyme kinetics analysis.

The kinetic parameters of APH(3′)-Id were determined as described in previous reports (Toth et al., 2013; Lu et al., 2021) with slight modifications. The phosphorylation of aminoglycosides was coupled to NADH oxidation by pyruvate kinase (PK) and lactate dehydrogenase (LDH). The reduction of NADH at 340 nm was monitored by a SpectraMax M5 multifunctional microplate reader (Molecular Devices, United States) to reflect the speed of aminoglycoside phosphorylation. The reaction mixture contained 1 mM phosphoenolpyruvate, 1 mM MgCl₂, 100 mM HEPES (pH 7.5), 600 μM NADH, 2 mM KCl, a commercial mixture of PK and LDH (Sigma P0294; 18–26 U/mL PK and 25–35 U/mL LDH in final concentration), 1 mM ATP, 30–40 nM APH(3′)-Id and aminoglycosides (5–400 μM) in a total volume of 250 μL. The steady-state velocities were calculated from the linear period of the reaction process curve and applied to plot the substrate concentration-velocity curve. The data were fit by nonlinear regression with the velocity as a function of substrate using the Michaelis–Menten equation. Curve fitting and calculation of the dynamic parameters K_m and k_cat were completed in GraphPad Prism 8.0.2 (GraphPad Software, Inc.).

Prokka v1.14.6 (Seemann, 2014) was used to predict the potential open reading frames (ORFs) of the assembled genome sequence, and the deduced proteins were annotated using DIAMOND v2.0.11 (Buchfink et al., 2021) according to the Comprehensive Antibiotic Resistance Database (CARD) (McArthur et al., 2013) to identify antimicrobial resistance genes. The annotation parameters were set as follows: E-value <1E-9, identity >95% and coverage >85%. Multiple sequence alignment and visualization were conducted using MAFFT v7.490 (Katoh and Standley, 2013) and GeneDoc v2.7 (Nicholas, 1997). Construction of phylogenetic trees and their visualization were conducted using IQ-TREE v1.6.9 (Nguyen et al., 2015) and iTOL v6.7 (Letunic and Bork, 2021). The crystal structure of the aminoglycoside phosphotransferase APH(3′)-Ia (SMTL ID: 3r78.1), which shared the highest amino acid identity of 77.49% with APH(3′)-Id, was used as a template for homology modeling of APH(3′)-Id in SWISS-MODEL (Waterhouse et al., 2018) to predict the tertiary structure. The tertiary structure alignment of the APH(3′)-Id model and kanamycin-bound APH(3′)-Ia (PDB 4FEU chain B) was performed by PyMOL (Schrödinger, 2023). A map of the genetic environment surrounding aph(3′)-Id was generated by clinker v0.0.24 (Gilchrist and Chooi, 2021). The Expasy ProtParam Tool (Gasteiger et al., 2003) was used for prediction of the molecular weight and pI value of the protein APH(3′)-Id.

The following accession numbers have been assigned for the complete genome sequence of Kluyvera intermedia DW18 in GenBank: CP123488 for chromosome, CP123489 for plasmid pDW18-1 and CP123490 for plasmid pDW18-2. The nucleotide sequence of the novel AME gene aph(3′)-Id has been assigned GenBank accession number OQ819314.

The 16S rRNA gene of DW18 shared the highest similarity of 100% (100% identity and 100% coverage) with that of K. intermedia MAG 4195 (GCA_029202345.1). The genome sequence of DW18 shared 98.95% ANI with that of K. intermedia N2-1 (GCA_009649915.1), which exceeded the threshold of 95% to define a bacterial species. Currently, the ANI between genomes is considered the best alternative to DDH and 16S rRNA gene homology analyses as the gold standard for prokaryotic classification, and the ANI threshold for species membership was empirically defined as 95% (Konstantinidis and Tiedje, 2005; Richter and Rosselló-Móra, 2009). Therefore, the isolate DW18 in this study was classified into the species K. intermedia and designated K. intermedia DW18.

The complete genome of K. intermedia DW18 was assembled, and it was found to consist of a chromosome and two circular plasmids. The chromosome is approximately 4.62 Mb in length and encodes 4,417 ORFs with an average GC content of 52.79%. The two circular plasmids were designated pDW18-1 and pDW18-2 (Table 4). pDW18-1 is 4,455 bp in length and encodes 9 ORFs, while pDW18-2 is 3,612 bp in length and encodes 6 ORFs. In the NCBI database, a total of 23 K. intermedia genomes were present, and only four (K. intermedia N2-1, GCA_009649915.1, K. intermedia HR2, GCA_009650415.1, K. intermedia NCTC12125, GCA_900635475.1, and K. intermedia MAG 4159, GCA_029202345.1) of them had complete genome sequences. Among the four strains with complete genome sequences, two (K. intermedia N2-1 and K. intermedia HR2) carried a large plasmid each (pN2-1, 245,785 bp in length, and pHR2-1, 245,785 bp in length). The other strains, NCTC12125 and MAG 4159, did not carry any plasmid. DW18, however, carried two small plasmids, pDW18-1 (4,455 bp) and pDW18-2 (3,612 bp). The chromosome size of the five K. intermedia strains with complete genomes (including DW18 in this work) varied slightly, ranging from 4.62 Mb (DW18) to 4.74 Mb (NCTC12125).

Of the 16 antimicrobials from two classes (aminoglycosides and β-lactams) tested, DW18 showed higher MIC levels for eight, including four aminoglycosides (ribostamycin 64 μg/mL, kanamycin 32 μg/mL, paromomycin 32 μg/mL, and spectinomycin 16 μg/mL) and four β-lactams (penicillin 2048 μg/mL, cefoxitin 512 μg/mL, ampicillin 256 μg/mL, and cefazolin 64 μg/mL) (Table 2). On the one hand, for β-lactams, the resistance profile of K. intermedia DW18 was generally consistent with previous reports, but on the other hand, for aminoglycosides, it showed different resistance phenotypes. Previous studies showed that K. intermedia was resistant to a variety of β-lactam antibiotics, such as penicillin, amoxycillin, and cefoxitin, but was susceptible to aminoglycosides (Stock, 2002). Some isolates also showed resistance to carbapenems and other cephalosporins such as imipenem, meropenem, cefepime, and ceftriaxone (Ribeiro et al., 2014; Paskova et al., 2018).

In our present project studying the resistance status of animal bacteria, we sequenced the draft genomes of several strains isolated from animal specimens. When analyzing the resistance genotype of the DW18 genome based on the CARD with thresholds of identity >95% and coverage >85%, we found that no functionally characterized aminoglycoside and β-lactam resistance gene was annotated in the genome even though it showed high MIC levels for some aminoglycosides and β-lactams.

To determine the potential resistance mechanism of DW18 against aminoglycosides, we examined the annotation data of the genome sequence and found that among the predicted genes, the gene showing the highest identity with the functionally characterized aminoglycoside resistance gene was an aph(3′)-Ia-like gene (finally designated aph(3′)-Id in this work), which showed amino acid similarity (100% coverage and 77.49% identity) with APH(3′)-Ia (CAE51638.1). To determine whether the predicted resistance gene was indeed functional, we cloned the ORF of the gene with its promoter region into the T-vector pMD19, and the resistance function of this gene was confirmed. Compared with the control strains (DH5α or DH5α carrying pMD19), the recombinant strain (pMD19-aph(3′)-Id/DH5α) showed 1,024-, 512-, 128-, and 16-fold increased MICs for kanamycin, ribostamycin, paromomycin and neomycin, respectively. No significant increases in the MICs for the other 6 aminoglycosides tested were observed (Table 2). There are only three other aph(3′)-I genes [aph(3′)-Ia, aph(3′)-Ib and aphA15] present in CARD, and only aph(3′)-Ia and aphA15 had the resistance profile documented (Siregar et al., 1995; Riccio et al., 2001). aph(3′)-Id, aph(3′)-Ia and aphA15 all showed higher MIC values for kanamycin and neomycin. The aph(3′)-Id gene in this work did not confer resistance to gentamicin or amikacin; however, the aph(3′)-Ia gene showed high MICs for gentamicin (>1,024 μg/mL) and amikacin (8 μg/mL). According to the evolutionary relationship, substrate spectrum and nomenclature proposed for genes encoding AMEs (Shaw et al., 1993), the novel resistance gene in this work was thus designated aph(3′)-Id.

When analyzing the distribution of aph(3′)-Id and aph(3′)-Id-like genes, only 4 genes with amino acid identities >80% were found according to the BLAST search against the NCBI nonredundant protein database. These genes were derived from K. intermedia (WEJ84261.1, WP_153742217.1, WP_062773153.1) and Klebsiella sp. (WP_181482651.1). The deduced amino acid sequence of APH(3′)-Id shared the highest identity of 99.26% with a hypothetical APH(3′)-I family aminoglycoside O-phosphotransferase (WEJ84261.1) from K. intermedia. The other two proteins from K. intermedia also shared high identities of 98.89% (WP_153742217.1) and 97.05% (WP_062773153.1) with APH(3′)-Id. However, the protein from Klebsiella sp. shared the lowest identity of 80.07% (WP_181482651.1) with APH(3′)-Id. This indicated that APH(3′)-Id was relatively conserved within the species K. intermedia.

The novel aminoglycoside 3′-O-phosphotransferase gene aph(3′)-Id was 816 bp in length and encoded a protein of 271 amino acids (30.69 kDa) with a pI value of 4.90. APH(3′)-Id shared the highest amino acid identities of 77.49, 58.67 and 39.85% with the functionally characterized APH(3′) proteins APH(3′)-Ia (CAE51638.1), APH(3′)-Ib (AAA26412.1) and APH(3′)-IIc (ADQ43421.1), respectively. Furthermore, a phylogenetic tree that included all 19 functionally characterized APH(3′) enzymes collected from the CARD database showed that APH(3′)-Id clustered closest to APH(3′)-I enzymes (Figure 1). Currently, the aph(3′)-Id genes were mainly found in the chromosomes of Kluyvera intermedia. The aph(3′)-Ib and aph(3′)-IIc genes were only found in Pseudomonas aeruginosa and Stenotrophomonas maltophilia, respectively. The aph(3′)-Ia gene, however, was initially identified in the translocator element Tn903 of Escherichia coli (Oka et al., 1981), but now is widely distributed in a variety of bacterial genera, such as Acinetobacter, Aeromonas, Avibacterium, Brucella, and Enterobacter. This suggested that aph(3′)-Id may not spread as rapidly and widely across multiple strains as aph(3′)-Ia.

To analyze the structure related to the resistance function of APH(3′)-Id, multiple sequence alignment of APH(3′)-Id and other functionally characterized APH(3′) proteins was performed (Figure 2). The results showed that APH(3′)-Id contained three conserved functional domains of the aminoglycoside phosphotransferase family, including Motif 1 (V--HGD----N), Motif 2 (G--D-GR-G) and Motif 3 (D--R/K--F/Y---LDE) (Shaw et al., 1993) (Figure 2). Six of eight active residues responsible for the phosphorylation of kanamycin in APH(3′)-IIa were identified in APH(3′)-Id (D167, D198, R219, E238, E269 and F271) (Nurizzo et al., 2003) (Figure 2). All nine kanamycin binding sites of APH(3′)-Ia were also identified in APH(3′)-Id (D165, F166, D167, D198, N234, E238, D268, E269, and F271) (Stogios et al., 2013) (Figures 2, 3). The tertiary structure alignment of the APH(3′)-Id model and kanamycin-bound APH(3′)-Ia showed that the overall structures of the two were similar, especially the kanamycin binding position. Similar to APH(3′)-Ia, the ATP binding sites of APH(3′)-Id, located in the hinge region between the N- and the C-terminal lobe of the bilobal fold, also had a similar spatial structure to that of eukaryotic protein kinases. The difference was that APH(3′)-Ia had a β-sheet (R6, E7, T8, C9), while APH(3′)-Id had an α-helix (L19, D20, T21, E22, L23), at the N terminus (Figure 3).

The induced expressed protein APH(3′)-Id was soluble (Supplementary Figure S1). After nickel column purification, His-tag removal and ultrafiltration concentration, 2 mL of 1.01 mg/mL (33.06 μM) APH(3′)-Id protein was obtained (Supplementary Figures S2, S3). The kinetic analysis of the phosphotransferase APH(3′)-Id showed that the purified APH(3′)-Id could phosphorylate four substrates, including ribostamycin, kanamycin, neomycin and paromomycin, which was consistent with the results for APH(3′)-I enzymes reported previously (Ramirez and Tolmasky, 2010). In the APH(3′)-I subclass, the enzyme with the closest evolutionary relationship to APH(3′)-Id was APH(3′)-Ia. By comparing the kinetic parameters between APH(3′)-Id and APH(3′)-Ia (Siregar et al., 1995), it was found that the affinity and catalytic efficiencies of APH(3′)-Id for the substrates were slightly weaker than those of APH(3′)-Ia. For example, when kanamycin was the substrate, the K_m of APH(3′)-Ia was lower than that of APH(3′)-Id (1.2 ± 0.2 μM vs. 10.04 ± 2.95 μM), while the k_cat/K_m of APH(3′)-Ia was higher than that of APH(3′)-Id (8.5 × 10⁷ M⁻¹/s⁻¹ vs. 1.10 × 10⁵ M⁻¹/s⁻¹). A similar result was observed when neomycin was the substrate. The K_m of APH(3′)-Ia for neomycin was lower than that of APH(3′)-Id (3.6 ± 0.3 μM vs. 18.70 ± 1.51 μM), while the k_cat/K_m of APH(3′)-Ia for neomycin was higher than that of APH(3′)-Id (5.1 × 10⁷ M⁻¹/s⁻¹ vs. 1.73 × 10⁵ M⁻¹/s⁻¹).

The substrate spectrum of APH(3′)-Id was also consistent with the resistance profile shown by the recombinant strain (pMD19-aph(3′)-Id/DH5α). Consistent with the MIC results, no phosphotransferase activity was found for gentamicin or streptomycin (Table 5). The catalytic efficiencies of APH(3′)-Id varied from the MIC values of the cloned aph(3′)-Id. APH(3′)-Id showed the highest catalytic efficiency for ribostamycin [k_cat/K_m of (4.96 ± 1.63) × 10⁵ M⁻¹/s⁻¹], followed by paromomycin [k_cat/K_m of (2.18 ± 0.21) × 10⁵ M⁻¹/s⁻¹], neomycin [k_cat/K_m of (1.73 ± 0.20) × 10⁵ M⁻¹/s⁻¹], and kanamycin [k_cat/K_m of (1.10 ± 0.18) × 10⁵ M⁻¹/s⁻¹]. However, the cloned aph(3′)-Id showed the highest MIC increase for kanamycin (1024-fold), followed by ribostamycin (512-fold), paromomycin (128-fold) and neomycin (16-fold). Neomycin had a high enzymatic catalytic efficiency but a low MIC value. It is possible that it was poorly transferred into the cell and could not reach the saturation concentration of the enzyme, so the enzyme could not fully perform its catalytic function (Siregar et al., 1995). In addition, APH(3′)-Id showed a high MIC level for kanamycin, while the catalytic efficiency was not very high. Different phosphate donors may result in different catalytic efficiencies of enzymes. ATP was used as the phosphate donor in this study. In a previous kinetic assay, the k_cat of kanamycin phosphorylated by APH(3′)-Ia with GTP as a donor was 10-fold higher than that with ATP as the donor (Stogios et al., 2013).

As mentioned above, the substrate spectrum of APH(3′)-Id appeared to be similar to that of the other APH(3′)-I enzymes, which showed modification activity against several aminoglycosides, including kanamycin, ribostamycin, neomycin, paromomycin, lividomycin, gentamicin, butirosin and amikacin (Shaw et al., 1993; Siregar et al., 1995; Ramirez and Tolmasky, 2010). The substrate spectrum of the APH(3′)-II enzymes was the same with APH(3′)-I subclass (Shaw et al., 1993; Lu et al., 2021). APH(3′)-III enzymes had a broader substrate spectrum than APH(3′)-II, showing modification activity for isepamicin (Shaw et al., 1993; Ramirez and Tolmasky, 2010). In general, the substrate profile of APH(3′)-Id was consistent with that of the APH(3′)-I enzymes, and it further supported that this novel enzyme can be concluded into the APH(3′)-I subclass.

The aph(3′)-Id gene is located in the chromosome of K. intermedia DW18. To explore its genetic context, the structure of an approximately 20 kb sequence with aph(3′)-Id at the center was analyzed (Figure 4). When using the nucleotide sequence of aph(3′)-Id as a query to search the NCBI nonredundant nucleotide database, a total of six aph(3′)-Id-like genes with identities of more than 80.0% were found. Of these six aph(3′)-Id-like genes, four were from the same species, K. intermedia, and each shared particularly high similarity (100% coverage and > 98% identity) with aph(3′)-Id, while the other two aph(3′)-Id-like genes were from Klebsiella sp. (CP056483.1 and CP055481.1) and shared a lower similarity (both with 100% coverage and 81.13% identity) with aph(3′)-Id. The genetic context of aph(3′)-Id with the four aph(3′)-Id-like genes that shared higher identities with aph(3′)-Id was then analyzed.

Multisequence alignment of the five 20 kb sequences from K. intermedia (including one from this work) demonstrated that the downstream regions of the aph(3′)-Id(−like) genes were similar among them. This region of DW18 is mainly composed of genes related to transcription (osmE and decR_1) and transport (eriC, rhtB and kdgM). Nitrilotriacetate is an industrial chelating agent extensively used in textile and wastewater treatment processes. The product of the ntaA gene, nitrilotriacetate monooxygenase component A, is the core component of the nitrilotriacetate biodegradation process and may help prevent the harm caused by chelating agents in sewage (Kim K. J. et al., 2006). However, the upstream fragments were completely different between the gene from this work and any of the other four, except the orfL gene next to the aph(3′)-Id(−like) genes. The product of the orfL gene belongs to the multifunctional glutathionine S-transferase family, members of which can detoxify several chemical pollutants in the environment, such as 2,4,6-trinitrotoluen (TNT), toxic chlorinated organic compounds and anthracene (Mcguinness et al., 2007; Dixit et al., 2011; Gunning et al., 2014). No MGE was found in the flanking regions of the aph(3′)-Id(−like) genes (Figure 4). These results suggest that aph(3′)-Id may be an intrinsic gene of Kluyvera, and modification of aminoglycosides may be one of its normal physiological functions to protect the organism from environmental drugs. Although no MGE was detected near aph(3′)-Id, the possibility of horizontal transfer of this resistance gene between bacteria of different species or genera in the wild in the long run cannot be ruled out.