In recent years, to investigate bacterial resistance to antimicrobials, more than 300 isolates have been isolated from marine samples in Wenzhou, Zhejiang Province, China. One isolate, designated Providencia sp. TYF-12, was obtained from a fish intestine. 16S RNA gene homology comparison, whole-genome average nucleotide identity (ANI) and digital DNA–DNA hybridization (dDDH) analyses were used for the identification of the bacteria. The strains and plasmids used in this study are listed in Table 1.

In accordance with the standards of the Clinical and Laboratory Standards Institute (CLSI),¹ the agar dilution method was used to determine the minimum inhibitory concentration (MIC). E. coli ATCC 25922 served as a quality control for antibiotic susceptibility testing. The tested antimicrobial agents included spectinomycin, tobramycin, streptomycin, netilmicin, paromomycin and amikacin (Table 2).

The genomic DNA of Providencia sp. TYF-12 was sequenced on the Illumina HiSeq 2,500 and PacBio RS II platforms by Shanghai Personal Biotechnology Co., Ltd. (Shanghai, China). The long PacBio reads were first assembled via Unicycler (Wick et al., 2017) and then corrected with Illumina sequencing data via pilon (Walker et al., 2014). Prokka v1.14.6 (Seemann, 2014) was used for the prediction of the open reading frames (ORFs). Potential proteins were annotated via the NCBI nonredundant protein database and DIAMOND v2.0.11 (Buchfink et al., 2021). Antimicrobial resistance genes were predicted via Resistance Gene identifier v5.2.0² and the comprehensive antibiotic resistance database (CARD) (McArthur et al., 2013). The ANI was calculated via fastANI v1.33 (Jain et al., 2018). The genetic environment of ant(9)-Id and its homologous genes were analyzed via clinker v0.0.24 (Gilchrist and Chooi, 2021). The multiple sequence alignment diagram and phylogenetic tree were generated via MAFFT v7.490 and MEGAX (Katoh and Standley, 2013; Kumar et al., 2018), respectively. The molecular weight and pI of ANT(9)-Id were predicted via the JavaScript program (Stothard, 2000).

The target gene, along with its promoter region, was amplified via PCR with the primers listed in Table 3. The PCR products were subsequently inserted into the pUCP20 vector via T4 DNA ligase (Takara Bio, Inc., Dalian, China). The recombinant plasmid was introduced into E. coli DH5α cells via the calcium transformation method, and the transformant were subsequently grown on Luria–Bertani (LB) solid media supplemented with ampicillin (AMP) at a final concentration of 100 μg/mL. The inserted sequence was confirmed via first-generation sequencing (Shanghai Sunny Biotechnology Co., Ltd., Shanghai, China).

To obtain the protein ANT(9)-Id, PCR was used for amplification of the ORF of the ant(9)-Id gene. The amplified product and the expression vector pCold I were both digested by the restriction enzymes BamHI and HindIII, and they were ligated. The recombinant plasmid (pCold I-ant(9)-Id) was transformed into E. coli BL21 cells. The transformants were screened on LB agar plates containing 100 μg/mL AMP, and the cloned fragment was verified by first-generation sequencing. The recombinant strain (E. coli BL21/pCold I-ant(9)-Id) was cultured overnight in LB broth supplemented with 100 μg/mL AMP at 37°C and then continuously cultured in LB liquid medium at a ratio of 1:100 until the optical density (OD₆₀₀) reached 0.6 to 0.8. Sterile isopropyl-beta-D-thiogalactopyranoside (IPTG; Sigma Chemicals Co., St. Louis, MO, United States) was added to a final concentration of 0.5 mM, and protein expression was induced in a shaker incubator at 16°C. The bacteria were collected by centrifugation (8,000 × g, 10 min) and then lysed by ultrasonication for 2 min, after which the supernatant was collected. The target protein in the supernatant was purified with a His-tag protein purification kit (Beyotime, Shanghai, China). At 25°C, the samples were treated with recombinant enterokinase (EK enzyme) for 2 h, after which the His-tag was removed. The purity and relative molecular mass of the enzyme were evaluated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS–PAGE). The protein concentration of ANT(9)-Id was determined via a BCA protein assay kit (Beyotime, Shanghai, China).

The catalytic activity of ANT(9)-Id was analyzed by coupling the enzyme to UDP-glucose pyrophosphorylase (UGP), phosphoglucomutase (PGM) and glucose-6-phosphate dehydrogenase (G6PD). The increase in NADPH was measured at a wavelength of 340 nm via a SpectraMax multifunctional microplate reader (M5, Molecular Devices, America) to monitor the progress of the reaction. The total reaction system was 0.2 mL and contained a mixture of enzymes and substrates, including 2 U/mL UGP, 20 U/mL PGM, 20 U/mL G6PD, 10 mM MgCl₂, 50 mM HEPES (pH 7.5), 500 μM dithiothreitol (DTT), 500 μM UDP-glucose, 500 μM glucose 1,6-bisphosphate, 500 μM NADP, 2 mM ATP and purified ANT(9)-Id protein. The mixture was incubated at 37°C for 5 min, and a series of concentrations of spectinomycin were finally added to initiate the reaction. The enzyme kinetic parameters (k_cat and K_m) were determined by fitting the Michaelis–Menten equation via Prism (v8.0.2) software (Chen et al., 2019).

The MICs of the 11 antibiotics for the isolates were tested. Approximately 140 isolates with different resistance spectra were randomly selected (Supplementary Table S1), and the genome sequence of each isolate was sequenced on the Illumina HiSeq 2,500 platform. The potential resistance genes in these genomes were predicted (Supplementary Table S2). In this study, our attention was focused on uncovering aminoglycoside resistance mechanisms. According to the annotated antimicrobial resistance gene profiles derived from the genomic sequences of the 140 isolates, 11 prospective aminoglycoside resistance-related genes with <80% aa identity to the functionally characterized resistance genes in the CARD were selected, which included aadA9-, ant(9)-Ia-, aadA25-, aph(3′)-Ia-, aac(3)-Id-, aac(6′)-IIa-, aac(6′)-Ic-, aac(3)-Ia-, aac(3)-IV-, aac(2′)-Ia-, and aac(6′)-If-homologous genes. These genes were subsequently cloned, and their resistance phenotypes were tested. In vitro antimicrobial susceptibility testing revealed that the ant(9)-Ia-like gene (designated ant(9)-Id in this work) was functional.

The novel resistance gene ant(9)-Id consists of 765 bp and encodes a protein of 254 aa (Supplementary Figure S1), which has a molecular weight of 28.67 kDa and an isoelectric point of 5.03. The results of the antimicrobial susceptibility test conducted on the recombinant strain (pUCP20-ant(9)-Id/DH5α) indicated that, of the 11 aminoglycosides tested, the strain demonstrated resistance to spectinomycin only. Compared with that of the control strain (pUCP20/DH5α), the MIC of spectinomycin (512 μg/mL) for the recombinant strain increased 64-fold. As we know that the vector pUCP20 is commonly used to analyze the function of a gene. As it is a high copy number plasmid, the in vivo protein concentration encoded by the cloned resistance gene (ant(9)-Id) in the recombinant (pUCP20-ant(9)-Id/DH5α) would be higher than that of the original host TYF-12 where the resistance gene is located in the chromosome which indicates that one cell has only one copy of the resistance gene. As a result, the MIC level increase of the recombinant carrying the cloned resistance gene (pUCP20-ant(9)-Id/DH5α) to an antimicrobial would be higher than that of the original host. The expression vector pCold I, however, was used to express the protein, so the MIC level of the recombinant (pCold I-ant(9)-Id/BL21) to the antimicrobial was not tested. As expected, the in vitro catalytic activity of ANT(9)-Id aligned with the in vivo MIC data of the ant(9)-Id gene. The enzyme selectively adenylates spectinomycin, with a K_m of 8.94 ± 2.50 μM and k_cat/K_m of 26.15 μM⁻¹·s⁻¹. However, no adenosine transfer effect on spectinomycin or tobramycin was observed (Table 4 and Supplementary Table S3).

The resistance characteristics of ant(9)-Id were consistent with those of the other ant(9) genes. At present, four ant(9) genes, namely, ant(9)-Ia (formerly named aad9 or spw), ant(9)-Ib, ant(9)-Ic and spd, are available in the CARD database. All of these strains were resistant to streptomycin only (Mahbub Alam et al., 2005; Murphy, 1985; LeBlanc et al., 1991; Sheng et al., 2023; Jamrozy et al., 2014; Wendlandt et al., 2013). Moreover, the proteins ANT(9)-Ib and ANT(9)-Ic also showed specific adenosine transfer effects on spectinomycin (Sheng et al., 2023).

The drug resistance phenotypes of the ANT family are diverse. ANT(9)s mediate resistance to spectinomycin alone (Mahbub Alam et al., 2005). ANT(6)s are resistant to streptomycin (Heo et al., 2022). ANT(3″)s confer resistance to both streptomycin and spectinomycin (Kehrenberg et al., 2005), whereas ANT(2″)s and ANT(4′)s adenylate multiple antimicrobial agents. ANT(2″)s are resistant to gentamicin, tobramycin, kanamycin and other aminoglycoside antibiotics (Wright and Serpersu, 2006); however, the presence of ANT(4′)s enables the host to be resistant to isepamicin, amikacin, tobramycin and other aminoglycoside antibiotics (Jacoby et al., 1990). AACs (Mayer, 1986) and APHs (Ramirez and Tolmasky, 2010) also have a wide substrate spectrum, except APH(9)s, which, like ANT(9)s, confer resistance to spectinomycin only (Fong et al., 2010).

To investigate the molecular characteristics of the ant(9)-Id gene coding sequence, the entire genome of the isolate TYF-12, which carried the novel resistance gene, was sequenced. Free of a plasmid, the whole genome of TYF-12 contains a chromosome that is 4,660,160 bp in length, which encodes 4,613 coding sequences (CDSs) with an average GC content of 46.13% (Table 5).

The 16S rRNA gene sequence of TYF-12 shared the closest relationship, with 99.47% identity and 98.0% coverage, with that of Providencia vermicola OP1 (NR_042415.1), followed by those of P. rettgeri DSM4541 (NR_042413.1, 99.27% identity and 96.0% coverage) and P. rettgeri NCTC11801 (NR_115880.1, 99.19% identity and 96.0% coverage). However, ANI analysis revealed that the genome of TYF-12 shared the greatest identity of 83.99% with that of P. rettgeri NCTC 11801 (GCF_900455085.1, not a type strain), followed by 81.6% with that of the type strain P. vermicola DSM 17385 (GCF_020381325.1). The dDDH analysis results were consistent with the ANI results, with the highest similarities of 53.3% with P. rettgeri NCTC11801 and 49.1% with P. vermicola DSM 17385. None of the genome sequences in the public database reached the threshold of ANI (96.0%) (Ciufo et al., 2018) or dDDH (70%) (Moore et al., 1987) for the classification of TYF-12 into an existing species. This finding indicated that this isolate represents a new species within the genus Providencia and was thus temporarily designated Providencia sp. TYF-12.

In addition to the novel resistance gene ant(9)-Id identified in this work, which encodes a protein with the highest aa similarity of 40.91% (87.6% coverage and 46.7% identity) with ANT(9)-Ic (QWQ57435.1) among the functionally characterized proteins, a total of 9 antimicrobial resistance genes with ≥80% nucleotide similarity to the functionally characterized genes found in the CARD database were identified from the whole genome, including two lincosamide genes (inuF and inuG), five aminoglycoside genes (aadA, aac(6′)-Ib-cr6, aph(3′)-Ia, aph(4)-Ia and aac(3)-IVa), one rifamycin (arr-3) and one macrolide (mphE) resistance gene (Supplementary Table S4). In vitro susceptibility analysis revealed that Providencia sp. TYF-12 was resistant to 5 (spectinomycin, streptomycin, kanamycin, paromomycin and ribostamycin) of the 17 antimicrobial agents tested. The aminoglycoside resistance phenotype was consistent with the genotype of the isolate (Supplementary Table S5).

A phylogenetic tree of ANT(9)-Id with 34 other ANTs (containing 5 ANT(9)s and 29 ANT(3)s proteins with aa identity >30% with ANT(9)-Id) was reconstructed. At present, a total of 46 functionally characterized ANT family proteins are present in the CARD. They consisted of 2 ANT(2)s (4.35%, 2/46), 31 ANT(3)s (67.39%, 31/46), 4 ANT(4)s (8.70%, 4/46), 5 ANT(6)s (10.87%, 5/46) and 5 ANT(9)s (10.87%, 5/46). Among them, ANT(9)-Ic (QWQ57435.1) shared the highest aa similarity of 41.09% with ANT(9)-Id, followed by ANT(9)-Ia (CAA26428.1, 36.85%) and ANT(3)-Ib (QEQ43477.1, 34.08%), while the other sequences had similarities ranging from 29.64–34.03%. The phylogenetic tree shows that ANT(9)-Id has a close phylogenetic relationship with ANT(9)-Ia and ANT(9)-Ic; this further confirmed that the newly discovered resistance gene was indeed a member of the ANT(9) family (Figure 1).

Multiple sequence alignment of ANT(9)-Id with its close relatives revealed that the functionally essential residues of the ANT(9) proteins were conserved in ANT(9)-Id. In AadA (Q8ZPX9), the amino acid residues W173 and D178 are highly important in the adenylation of streptomycin. In the case of spectinomycin, the essential amino acid residues E87, W112, D182 and either H185 or N185 play vital roles (Stern et al., 2018). The last four residues are conserved in ANT(9) family enzymes, including ANT(9)-Id (E82, W107, D174 and N177) (Figure 2).

To analyze the distribution of the ant(9)-Id-like genes, the ANT(9)-Id amino acid sequence was used as a query to search for homologous proteins in the NCBI nonredundant protein database. A total of 14 hits that shared ≥99.21% similarity with ANT(9)-Id were identified, and all of them were derived from the genus Providencia (Supplementary Table S6). The amino acid sequence of the ANT(9)-Id protein was most similar (100.0%) to that of an aminoglycoside adenylyltransferase family protein (WP_129467149.1), whereas for the remaining 13 proteins, 10 and 3 shared 99.61 and 99.21% similarity, respectively (Figure 3). No protein sharing a similarity ranging from 99.21 to 70.0% was present. These 15 ant(9)-Id(−like) genes (including one from this work) were from bacteria of different sources. They were obtained mainly from human clinical samples (60.0%, 9/15). One of these samples was from marine fish (6.7%, 1/15); however, the sources of the other 5 samples were unknown (33.3%, 5/15) (Supplementary Table S6). Further analysis demonstrated that there were about 89 WP_129467149.1 proteins from the different Providencia species present in the database,² which included P. rettgeri, P. huashanensis, P. xianensis, P. alcalifaciens and several species-undefined strains. Different from the one of this work from the marine fish, the others were isolated from human clinical specimens, soil and so on (Supplementary Table S7).

To analyze the genetic background of the ant(9)-Id(−like) genes, the sequence approximately 20 kb in length with the ant(9)-Id gene at the center was intercepted and used as a query to search for similar sequences in the NCBI nucleotide database. Sixteen hits with similarities ranging from 99.37 to 99.72% (> 99.0% identity and 100.0% coverage) were retrieved, each of which encoded an ant(9)-Id(−like) gene. Among the 16 sequences, 62.5% (10/16) were from Providencia rettgeri, and the rest (37.5%, 6/16) were from species-unclassified Providencia strains (Supplementary Table S8). Except for these 16 sequences, no sequences shared ≥80% identity, and ≥ 80% coverage was found. Further structural analysis revealed that these fragments encoded genes related to the following functional categories: transcriptional regulator (lysR family), RidA/YER057c/UK114 superfamily protein, ferrous iron transport permease and so on. No mobile genetic element (MGE) was found within the fragments. These sequences exhibited high consistency in both gene content and gene order, which revealed that the ant(9)-Id(−like) gene-encoding sequences are highly conserved in the bacteria of the genus Providencia (Figure 4).