Identification and characteristics of a novel acquired aminoglycoside phosphotransferase, APH(3′)-IVb, from Riemerella anatipestifer

ABSTRACT

A novel acquired aminoglycoside resistance gene, aph(3′)-IVb, was identified via whole-genome sequencing of a multidrug-resistant Riemerella anatipestifer isolate from a duck. The gene encodes a 262-amino-acid phosphotransferase, APH(3′)-IVb, sharing only 39.9% amino acid identity with its closest known homolog, APH(3′)-IVa. Heterologous expression of aph(3′)-IVb in Escherichia coli and a susceptible R. anatipestifer strain conferred resistance to neomycin, paromomycin, and ribostamycin, a phenotype validated by gene deletion and complementation experiments. Kinetic analysis of the purified APH(3′)-IVb enzyme confirmed phosphotransferase activity against these three aminoglycosides, with catalytic efficiencies (k_cat/K_m) ranging from 10⁴ to 10⁵ M⁻¹·s⁻¹. Furthermore, site-directed mutagenesis identified key residues critical for enzymatic function. While the prevalence of aph(3′)-IVb in R. anatipestifer isolates was low (1.6%), analysis of public databases identified 93 aph(3′)-IVb-positive sequences, of which 36.6% originated from human pathogens. Genetic environment analysis revealed that aph(3′)-IVb resides within a genomic resistance island flanked by mobile genetic elements, suggesting its horizontal acquisition. The emergence of this novel enzyme, coupled with its association with mobile elements and distribution among human pathogens, underscores a potential pathway for resistance dissemination across veterinary and clinical environments, posing a significant public health concern.

INTRODUCTION

Aminoglycoside antibiotics remain cornerstones of antimicrobial therapy, widely utilized for treating severe bacterial infections due to their broad-spectrum potency (1, 2). The primary determinant of clinical resistance to these agents is the enzymatic inactivation of the drug, mediated by aminoglycoside-modifying enzymes (AMEs) (3, 4). AMEs are categorized into three major families—N-acetyltransferases, O-nucleotidyltransferases, and aminoglycoside O-phosphotransferases (APHs)—based on the specific chemical modification (acetylation, adenylation, or phosphorylation, respectively) they catalyze (2, 5, 6). Nomenclature within these families is defined by the site of modification (e.g., 3′, 6′, or 2″), followed by a Roman numeral indicating the specific resistance profile and a lowercase letter denoting the unique protein variant (5, 6). Among the nearly 40 identified APH enzymes, the APH(3′) family is the most diverse, comprising seven subclasses (I–VII) with distinct substrate specificities (7, 8). Specifically, the APH(3′)-IV subclass is characterized by its resistance profile against neomycin (NEO), paromomycin (PAR), and ribostamycin (RIB) (5). Historically, aph(3′)-IVa has represented the sole documented gene within this subclass (9). Initially isolated from Niallia circulans, this gene has been extensively characterized regarding its sequence, heterologous expression, and regulatory elements.

Riemerella anatipestifer, a gram-negative bacterium of the family Weeksellaceae, is the etiological agent of infectious serositis (10). This highly contagious pathogen affects a broad range of avian hosts—including ducks, geese, and turkeys—and is responsible for significant outbreaks worldwide. Clinical manifestations range from acute septicemia to chronic polyserositis (characterized by fibrinous pericarditis, perihepatitis, and salpingitis) and meningitis. These infections result in high mortality rates and growth retardation, imposing a substantial economic burden on the global poultry industry (11). While antimicrobials such as aminoglycosides, florfenicol, and macrolide-lincosamide-streptogramin B agents are routinely employed for disease control (12), their extensive use has driven the selection and acquisition of resistance determinants. Consequently, the continuous evolution of the R. anatipestifer resistome has led to the emergence of multidrug-resistant (MDR) isolates (13). High-level resistance to diverse antibiotic classes, including quinolones, aminoglycosides, and tetracyclines, is now prevalent among clinical isolates (14). Furthermore, intrinsic resistance mechanisms, particularly the ubiquitous resistance-nodulation-division (RND) and ATP-binding cassette (ABC) efflux pumps, provide a basal level of protection against aminoglycosides and organic solvents in this species (12, 14).

In this study, we report the identification and functional characterization of a novel APH(3′)-IV variant, designated APH(3′)-IVb. The aph(3′)-IVb gene resides within a novel acquired genomic resistance island in a multidrug-resistant R. anatipestifer clinical isolate.

RESULTS AND DISCUSSION

Antibiotic susceptibility of R. anatipestifer isolate RCAD1101

R. anatipestifer RCAD1101 was isolated from the respiratory tract of a diseased duck on a farm in Yunnan Province, China. The minimum inhibitory concentrations (MICs) of aminoglycosides for RCAD1101, the deletion mutant RCAD1101Δaph(3′)-IVb, the complementation strain RCAD1101Δaph(3′)-IVb(pLMF03-aph(3′)-IVb), and recombinant strains are summarized in Table 1. Notably, the deletion mutant RCAD1101Δaph(3′)-IVb still exhibited high background resistance to neomycin (128 mg/L) and, unexpectedly, an even higher level to ribostamycin (512 mg/L). Although such high-level resistance is often clinically associated with 16S ribosomal RNA (rRNA) methylases, our genomic analysis (Table S3) confirmed the absence of such genes. Furthermore, while the aminoglycoside nucleotidyltransferase gene aadS was identified in the genome, it specifically confers resistance to streptomycin and does not contribute to the resistance against 4,5-disubstituted aminoglycosides (e.g., ribostamycin) observed here (15). Thus, this residual resistance is attributed to the intrinsic multidrug-resistance background of R. anatipestifer, driven by functionally characterized efflux systems, such as the RND-type RaeABCR and RaeE-RaeF-RopN complexes, and the ABC-type RanARanB system (12, 14, 16), which appear to confer high baseline protection in both the mutant and the standard strain ATCC 11845. Consequently, aph(3′)-IVb functions as a “resistance amplifier,” acting on top of this intrinsic efflux activity to boost resistance to clinically high levels (≥1,024 mg/L). The MIC values for an additional 19 antibiotics tested against RCAD1101, with interpretive categories, are presented in Table S2. Based on these MIC data, RCAD1101 was resistant to the majority of antibiotics, confirming its classification as an MDR strain.

TABLE 1.

Aminoglycosides MICs for R. anatipestifer and recombinant E. coli strains^a

Strains	MIC (mg/L)
	NEO	PAR	RIB	AMK	SPE	MCR	KAN	NET	SIS	TOB	APR	STR	GEN
R. anatipestifer RCAD1101	1,024	1,024	4,096	128	128	32	512	128	256	512	512	64	128
RCAD1101△aph(3′)-IVb	128	256	512	128	128	32	512	128	256	256	512	64	128
RCAD1101△aph(3′)-IVb(pLMF03-aph(3′)-IVb)	1,024	1,024	4,096	128	128	32	512	128	256	256	512	64	128
R. anatipestifer ATCC 11845	128	256	512	64	64	64	512	128	256	512	512	256	32
ATCC 11845/pLMF03	128	256	512	64	128	64	512	128	128	512	512	256	64
ATCC 11845/pLMF03- aph(3′)-IVb	1,024	1,024	4,096	64	128	64	512	128	128	512	512	256	64
E. coli BL21	32	32	64	32	64	32	32	8	8	16	32	32	16
BL21/pET32a	32	32	64	32	128	16	32	8	8	8	32	32	16
BL21/pET32a-aph(3′)-IVb	256	512	＞4,096	32	128	16	32	8	8	8	16	16	16

^a NEO, neomycin; PAR, paromomycin; RIB, ribostamycin; AMK, amikacin; SPE, spectinomycin; MCR, micronomicin; KAN, kanamycin; NET, netilmicin; SIS, sisomicin; TOB, tobramycin; APR, apramycin; STR, streptomycin; GEN, gentamicin.

Genome screening and identification of the novel resistance gene aph(3′)-IVb

To elucidate the genetic basis of the MDR phenotype of RCAD1101, we sequenced its complete genome (GenBank accession no. CM136845.1) and analyzed it using AMRFinderPlus (version 3.10.24) (17) to identify potential resistance determinants. The screening revealed a total of 38 known and putative resistance genes (Table S3). Among these, a putative novel resistance determinant showing limited homology to characterized aminoglycoside phosphotransferases was identified and designated as aph(3′)-IVb. Consequently, subsequent experiments focused on the functional characterization of this novel gene.

Functional characterization of aph(3′)-IVb and its contribution to aminoglycoside resistance

The aph(3′)-IV-like gene consists of 789 bp and encodes a 262-amino acid protein. It shares the highest amino acid sequence identity (39.9%) with the chromosomally encoded aminoglycoside phosphotransferase APH(3′)-IVa from Niallia circulans (9). When expressed in Escherichia coli BL21 and R. anatipestifer ATCC 11845 (Table 1), the APH(3′)-IV-like protein (designated APH(3′)-IVb) conferred resistance to neomycin, paromomycin, and ribostamycin, but not to tobramycin, sisomicin, netilmicin, spectinomycin, apramycin, gentamicin, amikacin, micronomicin, streptomycin, or kanamycin. Additionally, deletion of the gene resulted in increased susceptibility to these three antibiotics, whereas complementation restored resistance to neomycin, paromomycin, and ribostamycin, confirming the predicted function of aph(3′)- IVb.

To definitively classify this enzyme given its relatively low sequence identity (39.9%) to APH(3′)-IVa (9), we integrated phylogenetic, sequence alignment, and structural analyses. The phylogenetic tree indicated that this protein clusters within the APH(3′)-IV-type clade (Fig. S1). Consistent with this clustering, multiple sequence alignment revealed that despite the low overall sequence identity, the critical functional motifs characteristic of the APH family are strictly conserved (Fig. 1). These include the critical residues responsible for ATP binding, the catalytic center, and substrate recognition loops, as well as other conserved functional elements, which have been functionally defined in structurally characterized homologs such as APH(3′)-IIIa and APH(3′)-IIa (18–20).

Fig 1.

Multiple sequence alignment and structural annotation of APH(3′)-IVb. The red regions indicate fully conserved sites. Colored triangles represent specific functional motifs corresponding to characterized residues in APH(3′) homologs: black triangles indicate the catalytic base essential for phosphate transfer; red triangles mark the catalytic core residue stabilizing the transition state; orange triangles denote the conserved Lys-Glu salt bridge required for ATP orientation; blue triangles identify the invariant Asn residue responsible for coordinating magnesium ions (Mg2⁺); lime triangles indicate the residue positioning the beta-phosphate of ATP; dark red triangles mark the Gly-rich P-loop involved in nucleotide binding; purple triangles show hydrophobic residues forming the dimerization interface; green triangles highlight substrate Loop A defining substrate specificity; cyan triangles mark the C-terminal hydrophobic latch essential for pocket integrity; and pink triangles indicate the structural anchor Asp218 maintaining helix integrity. The red dashed line underscores the electrostatic substrate-binding helix (α6) responsible for recruiting polycationic antibiotics. Circles represent the mutation sites in this experiment, with red filled circles marking the sites where the drug resistance phenotype was compromised after mutation, and open circles indicating the sites where the drug resistance phenotype remained unchanged. The numbers correspond to the positions of amino acid residues in APH(3′)-IVb.

To further validate this classification, we examined the tertiary structure. Structural superposition analysis provided robust complementary evidence. As shown in Fig. S2, the predicted 3D structure of this novel enzyme aligns well with APH(3′)-IVa, yielding a root mean square deviation of only 1.141 Å (21, 22). The observed structural congruity confirms that despite sequence divergence, the enzyme adopts the canonical catalytic fold of the APH(3′)-IV subfamily. Therefore, strictly adhering to the standard classification criteria for aminoglycoside modifying enzymes, which prioritize regioselectivity and substrate profile (5), we designated this gene as aph(3′)-IVb. This classification is supported by (i) its specific resistance to 4,5-disubstituted aminoglycosides (phenotype), (ii) its highly conserved 3D structure (mechanism), and (iii) its monophyletic clustering (phylogeny). The nomenclature has been submitted to the National Center for Biotechnology Information (NCBI) GenBank database (accession no. PX763553.1).

It is noteworthy that while the prototype APH(3′)-IVa has been reported to confer resistance to a broad spectrum including kanamycin (23), our identified APH(3′)-IVb exhibited a distinct profile limited to neomycin, paromomycin, and ribostamycin. This discrepancy can be reconciled by the evolutionary origins and established classification of this subclass. Foundational studies by Herbert et al. (9, 24) established that the prototype gene from Bacillus circulans functions primarily as a self-defense mechanism against butirosin, a member of the neomycin (4,5-disubstituted) family. Furthermore, nomenclature consensus established by Ramirez and Tolmasky (5) explicitly defines the APH(3′)-IV subclass by its resistance to neomycin and ribostamycin, excluding the kanamycin from its core defining spectrum. The high intrinsic resistance to kanamycin (MIC = 512 mg/L) in R. anatipestifer further obscures any potential contribution of APH(3′)-IVb to this specific drug. Therefore, the “IVb” designation accurately reflects an enzyme that adheres to the ancestral substrate specificity of the APH(3′)-IV lineage while lacking the expanded phenotype reported in some clinical isolates.

Kinetic parameters of APH(3′)-IVb

The phosphotransferase activity and kinetic parameters of purified APH(3′)-IVb were determined using purified enzyme and various aminoglycoside antibiotics substrates. Consistent with MIC data (Table 1), the enzyme efficiently phosphorylated neomycin, paromomycin, and ribostamycin, but no activity was detected against amikacin, gentamicin, streptomycin, kanamycin, or sisomicin (Table 2). Direct kinetic comparison with the homolog APH(3′)-IVa was not feasible, as its activity was previously assessed solely via semi-quantitative phosphocellulose paper binding assays without determining substrate affinity (K_m) or catalytic efficiency (k_cat).

TABLE 2.

Kinetic parameters of APH(3′)-IVb^a with different aminoglycoside substrates^c

Substrate	Km (μM)b	kcat (s−1)b	kcat/Km (M−1/s−1)
NEO	22.13 ± 2.08	4.85 ± 0.01	(2.19 ± 0.20) × 10⁵
PAR	60.10 ± 3.12	10.12 ± 0.15	(1.69 ± 0.07) × 105
RIB	54.20 ± 4.93	4.96 ± 0.12	(9.15 ± 0.85) × 104

^a The protein was initially modified by a His₆ tag, which was removed after purification.

^b k_cat and K_m values represent the mean ± SD of three independent experiments.

^c NEO, neomycin; PAR, paromomycin; RIB, ribostamycin.

To contextualize these findings, we compared the kinetic profile of APH(3′)-IVb with that of APH(3′)-Ie, a homolog possessing a distinct resistance profile (25). A distinct kinetic trend emerged: APH(3′)-IVb consistently displayed higher turnover numbers (k_cat) but lower substrate affinities (higher K_m) than APH(3′)-Ie across shared substrates. Specifically, for neomycin, APH(3′)-IVb exhibited a 2.96-fold higher k_cat. Although its K_m (22.13 μM) was 1.83-fold higher than that of APH(3′)-Ie (12.07 μM), the overall catalytic efficiency (k_cat/K_m) of APH(3′)-IVb remained 2.25-fold superior (2.19 × 10⁵ vs 1.36 × 10⁵ M⁻¹·s⁻¹). Similarly, for paromomycin, APH(3′)-IVb demonstrated an 11.77-fold higher k_cat (10.12 vs 0.86 s⁻¹) and a 1.48-fold higher catalytic efficiency (1.69 × 10⁵ vs 1.14 × 10⁵ M⁻¹·s⁻¹), despite a 7.76-fold increase in K_m. In contrast, for ribostamycin, while APH(3′)-IVb showed a 3.79-fold higher k_cat (4.96 s⁻¹), its affinity was markedly lower (K_m = 54.2 µM) compared to APH(3′)-Ie (K_m = 4.22 µM). Consequently, APH(3′)-Ie exhibited a 3.52-fold higher catalytic efficiency for this substrate (3.23 × 10⁵ vs 9.15 × 10⁴ M⁻¹·s⁻¹).

Site-directed mutagenesis analysis of APH(3′)-IVb

To identify residues critical for APH(3′)-IVb catalysis, we selected 20 positions for alanine scanning mutagenesis. Site selection was guided by multiple sequence alignment with other APH(3′) enzymes, molecular docking models of the APH(3′)-IVb-aminoglycoside complex (Fig. S3), and prior literature (26). Following established protocols, each target residue was individually substituted with alanine. The mutagenesis results (Table 3) revealed three distinct phenotypic classes regarding APH(3′)-IVb-mediated resistance to ribostamycin, paromomycin, and neomycin in E. coli.

TABLE 3.

Aminoglycoside MICs for E. coli expressing APH(3′)-IVb site-directed mutants^a

Strains	MIC (mg/L)
	NEO	PAR	RIB
E. coli BL21	32	32	64
BL21/pET32a	32	32	64
BL21/pet32a-aph(3′)-IVb	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(E206A)	32	32	64
BL21/pet32a-aph(3′)-IVb(K46A)	32	32	64
BL21/pet32a-aph(3′)-IVb(D188A)	32	32	64
BL21/pet32a-aph(3′)-IVb(N193A)	32	32	256
BL21/pet32a-aph(3′)-IVb(E260A)	32	32	64
BL21/pet32a-aph(3′)-IVb(D259A)	32	32	64
BL21/pet32a-aph(3′)-IVb(G95A)	32	32	512
BL21/pet32a-aph(3′)-IVb(D155A)	64	256	2,048
BL21/pet32a-aph(3′)-IVb(△F262)	32	128	512
BL21/pet32a-aph(3′)-IVb(R224A)	32	32	128
BL21/pet32a-aph(3′)-IVb(G26A)	64	64	512
BL21/pet32a-aph(3′)-IVb(D192A)	32	32	512
BL21/pet32a-aph(3′)-IVb(D218A)	32	32	64
BL21/pet32a-aph(3′)-IVb(D60A)	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(T97A)	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(L93A)	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(T91A)	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(S27A)	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(N29A)	256	512	>4,096
BL21/pet32a-aph(3′)-IVb(N225A)	256	512	>4,096

^a NEO, neomycin; PAR, paromomycin; RIB, ribostamycin.

Group I comprised mutations that completely abolished resistance. Substitutions at positions K46, D188, D192, E206, D218, D259, and E260 yielded MICs indistinguishable from the negative control. These results indicate that these residues are essential for the enzyme’s function, as their mutation resulted in a complete loss of resistance to all three tested aminoglycosides.

Group II comprised mutations that partially impaired resistance. Specifically, mutations N193A, G95A, and R224A abolished resistance to NEO and PAR, while the MIC for RIB decreased to ≤128 mg/L (a 32-fold reduction compared to the wild-type strain). The deletion of F262 (ΔF262) eliminated resistance to NEO, reduced the PAR MIC to 128 mg/L (25% of wild-type levels), and decreased the RIB MIC to ≤512 mg/L. Furthermore, G26A and D155A mutations caused a general reduction in resistance to all three antibiotics, with MICs dropping to 50%–12.5% of wild-type levels.

Group III comprised mutations that did not alter resistance. Substitutions at S27, N29, D60, T91, L93, T97, N225, and F262A produced MICs indistinguishable from the wild type, suggesting that these specific side chains are not strictly required for APH(3′)-IVb-mediated resistance under the tested conditions.

Structural comparisons with APH(3′)-IIIa further elucidated these findings. Previous studies have established that APH(3′)-IIIa shares structural homology with eukaryotic protein kinases (18, 27). Sequence alignment identifies five highly conserved residues in APH(3′)-IIIa—Lys44, Glu60, Asp190, Asn195, and Asp208—which correspond to Lys46, Asp60, Asp188, Asn193, and Asp206 in APH(3′)-IVb (Fig. 1), respectively. In APH(3′)-IIIa, Lys44 is critical for catalysis by forming ionic interactions with ATP phosphates (27); consistent with this, the K46A mutation in APH(3′)-IVb completely abolished resistance. Conversely, Glu60 in APH(3′)-IIIa stabilizes the ATP-binding pocket via a salt bridge but is not catalytic (18); similarly, the D60A mutation in APH(3′)-IVb did not significantly compromise resistance. The catalytic core residues (Asp190, Asn195, and Asp208 in IIIa) are essential for substrate positioning and transition state stabilization (18); our data confirmed this conservation in APH(3′)-IVb, as mutations at the corresponding sites (Asp188, Asn193, and Asp206) resulted in abolished or severely impaired resistance.

Regarding the C-terminal loop region, Kaul et al. (19) demonstrated its importance for drug binding in APH(3′)-IIIa (residues 260–264). In APH(3′)-IVb, the homologous residues are Asp259, Glu260, and Phe262. The D259A and E260A mutations in APH(3′)-IVb completely abolished resistance, mirroring the role of their homologs in APH(3′)-IIIa in engaging amino groups of the substrate. For Phe262, our results showed that while the deletion (ΔF262) significantly impaired resistance, the single substitution F262A had no effect. This aligns with findings for Phe264 in APH(3′)-IIIa, where the residue’s presence is crucial for maintaining pocket integrity through steric effects rather than specific side-chain chemistry (19).

Finally, among the remaining residues, G26, G95, D155, D218, and R224 are highly conserved across APH(3′) enzymes, and our results confirmed their functional importance. Notably, although Asp192 (D192) is not conserved in the primary sequence alignment, molecular docking identified it as a key site for ADP/ATP binding. The D192A mutation abolished resistance to NEO and PAR and reduced RIB resistance by ≥8-fold. Thus, D192 likely represents a unique and critical determinant of APH(3′)-IVb-mediated aminoglycoside resistance.

Prevalence of aph(3′)-IVb in R. anatipestifer field isolates and its distribution across ecological niches

To investigate the prevalence of aph(3′)-IVb in clinical settings, we screened 741 R. anatipestifer field isolates collected from major poultry-farming provinces in China (Sichuan, Shandong, Jiangxi, Yunnan, and Guangdong). PCR analysis identified 14 positive isolates, corresponding to a detection rate of 1.9% (Table S4). The earliest aph(3′)-IVb-positive strain in our collection dates back to 5 July 2010. Despite the relatively low overall prevalence, the identification of positive isolates across diverse geographical regions over a 15-year span indicates that the gene demonstrates a capacity for persistence and transmission within the R. anatipestifer population.

To broaden our understanding of its ecological distribution, we retrieved 93 sequences sharing > 98% protein sequence identity with APH(3′)-IVb from public databases (Table S5). Note that one sequence was curated as two segments due to a premature stop codon. These sequences originated from at least 23 distinct species, spanning a collection period from 1988 to 2025. The geographical distribution of these strains is illustrated in Fig. S4, with the majority (85/93) originating from China, followed by sporadic isolates from the United States, Bangladesh, Denmark, India, Israel, Peru, and Vietnam. The earliest recorded aph(3′)-IVb-positive strain was Parabacteroides goldsteinii BFG-241 (accession no. NZ_CP081906.1), isolated from a human patient with a Bacteroides infection in the United States (28).

Ecological niche analysis classified these 93 strains into 3 categories: environmental sources (n = 5), pathogenic bacteria explicitly associated with disease (n = 43), and host-associated commensals (n = 45). Among the 43 pathogenic isolates, sources included humans (n = 34), ducks (n = 7), Gallus gallus (n = 1), and Sus scrofa domesticus (n = 1). Notably, the 34 human-derived pathogenic strains were primarily isolated from sputum, blood, and urine samples, suggesting that the gene is frequently detected in human clinical isolates.

Further analysis revealed a strong link between aph(3′)-IVb and the tigecycline resistance gene tet(X). Thirty of the 34 human pathogenic strains were derived from a single study by Zhang et al. (29) and belonged to the phylum Bacteroidota (specifically Chryseobacterium bernardetii, Chryseobacterium indologenes, Elizabethkingia meningoseptica, and Sphingobacterium mizutaii); all were confirmed tet(X)-positive. Similarly, 28 additional strains (27 host-associated and 1 environmental) identified in other studies (30–32) were also tet(X)-positive. This high rate of co-occurrence suggests that the dissemination of aph(3′)-IVb may be driven by co-selection pressure exerted by the combined clinical use of tetracyclines and aminoglycosides.

Regarding R. anatipestifer specifically, public databases contained two positive strains (R-4 and R-17) of duck origin. Combining these with our field surveillance data, the overall prevalence of aph(3′)-IVb in the R. anatipestifer population is estimated at 1.6% (16/975).

Genetic environment analysis and natural transformation of aph(3′)-IVb

To elucidate the genetic context of aph(3′)-IVb in R. anatipestifer, we performed a fine-scale structural analysis of the 100 kb sequences flanking the gene in clinical strains (Fig. 2). This region contains a high density of antimicrobial resistance genes interspersed with mobile genetic elements (MGEs). To assess the novelty of this genomic island, we performed a BLASTn search (Megablast algorithm) against the NCBI nucleotide database. The analysis revealed that no existing entry in the database exhibits complete homology to this region, with the highest query coverage being only 54% (identified in R. anatipestifer strain RAf490, GenBank accession no. CP175957.1). Consequently, we identified this region as a novel genomic resistance island (GRI) characterized by a unique mosaic structure.

Fig 2.

Genetic context of the aph(3′)-IVb gene in R. anatipestifer. Schematic representation of the genetic environment of aph(3′)-IVb and comparison of the aph(3′)-IVb-carrying regions in genomes of R. anatipestifer strains. Open reading frames are shown as arrows drawn to scale to indicate the direction of transcription. The aph(3′)-IVb gene is colored in orange, the known resistance genes are colored in green, the mobile elements are colored in sky blue, and the other genes are colored in black. dfrA49, dihydrofolate reductase A49; floR, florfenicol resistance gene R; tet(X5), tetracycline resistance gene X5; catB, chloramphenicol acetyltransferase B; bla_OXA-1327, β-lactamase OXA-1327; aadS aminoglycoside 6′-N-acetyltransferase S; lnu(I), lincosamide nucleotidyltransferase I; bla_OXA-347, β-lactamase OXA-347; tet(X4), tetracycline resistance gene X4; erm(F), erythromycin resistance methylase F.

Specifically, in the 20 kb flanking regions adjacent to aph(3′)-IVb in strain RCAD1101, a multidrug-resistance cluster was identified, comprising dfrA49 (sulfonamide resistance), floR (phenicol resistance), tet(X5) (tetracycline resistance), aadS (streptomycin resistance), catB (chloramphenicol resistance), and bla_OXA-1327 (β-lactam resistance). Furthermore, two MGEs were characterized adjacent to the gene: an IS30-like insertion sequence (designated IS4351) and an IS91-like insertion sequence (designated ISCR2) both located upstream. Comparative analysis across strains RCAD1101, RCAD1082, and the plasmid-borne aph(3′)-IVb in RCAD0416 consistently detected IS4351 in the upstream and downstream regions (Fig. 2).

To investigate the potential for horizontal transfer of aph(3′)-IVb mediated by the natural competence of R. anatipestifer, we conducted natural transformation assays using both genomic DNA extracted from strain RCAD1101 and a PCR-amplified aph(3′)-IVb fragment. Neomycin-resistant transformants were successfully obtained using both donor materials. Sequence analysis revealed that the integration profiles were identical for transformants derived from both genomic DNA and the PCR fragment. Specifically, the 27 bp 5′ untranslated region (UTR) and 140 bp 3′ UTR of aph(3′)-IVb were consistently co-integrated along with the coding sequence (Fig. 3). Analysis of the 100 kb flanking region in ATCC 11845 using VRprofile2 (33)—covering prophages, GRIs, integrative and conjugative elements (ICEs), integrons, and transposons—revealed an absence of potential mobile genetic elements. Notably, this integration event occurred in the absence of flanking sequence homology with the donor fragment (Fig. 3), distinguishing it from canonical RecA-dependent homologous recombination (34, 35). This observation parallels findings reported for the lnu(H) (36) and bla_RATA (37) genes, implying that R. anatipestifer can integrate exogenous resistance determinants through a mechanism that has yet to be fully elucidated.

Fig 3.

Natural transformation of aph(3′)-IVb into R. anatipestifer ATCC 11845 genome. The aph(3′)-IVb gene from RCAD1101 was inserted into the genome of ATCC 11845. Linear genomic comparison shows the genetic organization of the insertion site in the wild-type recipient ATCC 11845 (top), the constructed insertion mutant ATCC 11845_aph(3′)-IVb (middle), and the donor strain RCAD1101 (bottom). Arrows indicate coding sequences and their direction of transcription, color-coded by function: red arrows indicate the resistance gene aph(3′)-IVb; yellow arrows indicate the relE gene (encoding a type II toxin-antitoxin system toxin); green arrows indicate the hthR gene (encoding a putative HTH-type transcriptional regulator); and blue arrows indicate the cdf gene (encoding a cation diffusion facilitator family transporter). Gray-shaded regions denote sequence homology. The results confirm the successful site-specific integration of aph(3′)-IVb into the corresponding chromosomal locus of ATCC 11845 flanked by these conserved marker genes.

To further elucidate the potential mechanism underlying this homology-independent integration, we performed a multiple sequence alignment of the 140 bp downstream region across all analyzed aph(3′)-IVb-positive R. anatipestifer isolates, which revealed complete sequence identity. Secondary structure prediction of this segment using mfold (38) demonstrated the formation of stable stem-loop structures (Fig. S6). Notably, the folding topology of these structures exhibits structural features analogous to the canonical attC recombination sites recognized by integrons (39). This structural resemblance, coupled with the absence of flanking sequence homology, suggests that the acquisition of aph(3′)-IVb is likely mediated by a site-specific recombination mechanism relying on these conserved secondary structural motifs.

However, despite this demonstrated mobility, the prevalence of aph(3′)-IVb remains remarkably low (1.6%, 16/975) in the global R. anatipestifer population. This paradox may be explained by functional redundancy. Given that R. anatipestifer inherently possesses robust multidrug efflux systems, such as the RaeABCR and RaeE-RaeF-RopN complexes, and the RanARanB transporter (12, 14, 16), the species exhibits a high basal tolerance to aminoglycosides. Consequently, it may rely less on specific modifying enzymes for survival under standard selection pressures.

Finally, analysis of the broader distribution of aph(3′)-IVb revealed that it is detected across diverse pathogenic species, particularly those affecting humans. To investigate the genetic context in these pathogens, we analyzed the 10 kb sequences upstream and downstream of the gene (Fig. S5). The results showed that a plethora of resistance genes, as well as ICEs, were frequently co-localized with aph(3′)-IVb. This was especially notable in Glaesserella parasuis strain LSR011 (isolated from Sus scrofa domesticus with polyserositis), where the gene was embedded within a complex mobile region. We postulate that aph(3′)-IVb exploits the mobility of these elements via a mechanism of genetic hitchhiking. By integrating into ICEs, the gene leverages their intrinsic conjugative machinery to facilitate efficient interspecies dissemination. The association with ICEs suggests that aph(3′)-IVb possesses high transferability potential. The strong dissemination capacity of these elements likely contributes to the global distribution of the gene across the human-animal interface, underscoring the potential threat it poses to public health.

Conclusion

The identification of aph(3′)-IVb represents a significant expansion of the APH(3′)-IV subclass, providing novel insights into the evolutionary plasticity of aminoglycoside-modifying enzymes. The preservation of the resistance phenotype, despite substantial sequence divergence (~40% identity to APH(3′)-IVa), suggests that the diversity of this enzyme family is far more extensive than previously recognized. Moreover, the demonstrated acquisition of aph(3′)-IVb via natural transformation exemplifies the remarkable genomic plasticity of R. anatipestifer, highlighting its capacity to actively capture and integrate exogenous determinants in response to antimicrobial pressure. Mechanistically, unlike determinants that confer high-level resistance in isolation, APH(3′)-IVb functions as a “resistance amplifier,” synergizing with intrinsic efflux systems to elevate basal tolerance to clinically significant resistance levels. Furthermore, the localization of aph(3′)-IVb within a mobile genomic resistance island, coupled with its detection in diverse human pathogens, reveals a critical pathway for resistance dissemination across the human-animal interface. These findings underscore the urgency of adopting a “One Health” strategy to surveil such cryptic resistance elements and mitigate their potential public health impact.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions

Table S1 summarizes the bacterial strains, plasmids, and primers used in this study. R. anatipestifer strain RCAD1101 was isolated in September 2020 from the respiratory tract of a diseased duck in Yunnan Province, China. Clinical samples were streaked onto Tryptic Soy Agar (TSA; Oxoid Ltd., Basingstoke, UK) supplemented with 5% sheep blood and incubated at 37°C. Individual colonies were purified and identified via Sanger sequencing of the full-length ¹⁶S rRNA gene (40).

E. coli DH5α was utilized as the host for routine cloning, while E. coli BL21(DE3) served as the expression host for the purification of APH(3′)-IVb. The multidrug-susceptible strain R. anatipestifer ATCC 11845 was used as the recipient for natural transformation experiments. Plasmids pET32a and pLMF03 were employed as cloning and expression vectors. The pET28a vector was specifically used for the overexpression and purification of His-tagged APH(3′)-IVb. Unless otherwise specified, bacterial strains were routinely cultured overnight in Luria-Bertani broth at 37°C, supplemented with appropriate antimicrobial agents and, when required, solidified with 1.5% agar.

Nucleotide and protein sequence analysis

Genomic DNA was extracted from R. anatipestifer RCAD1101 using the TIANamp Bacteria DNA Kit (Tiangen, China). Sequencing libraries were prepared using the MGIEasy DNA Library Prep Kit (MGI Tech Co., Ltd., Shenzhen, China) and the ONT 1D Ligation Sequencing Kit (SQK-LSK109). Sequencing was subsequently performed on the MGISEQ-2000 (BGI, Shenzhen, China) and Oxford Nanopore MinION (Oxford Nanopore Technologies, UK) platforms, respectively, in accordance with the manufacturers’ protocols. The hybrid assembly was generated using Canu (v1.5) with default parameters, followed by two rounds of polishing with Pilon (v1.24) utilizing the short-read data (41, 42).

Putative resistance determinants in RCAD1101 were identified using AMRFinderPlus (version 3.10.24) (17). Multiple sequence alignments were performed using CLUSTAL W (43) and visualized using ESPRipt 3.0 (44), and phylogenetic trees were constructed using MEGA11 (45). The three-dimensional structure of APH(3′)-IVb was predicted using AlphaFold 3 (46). Molecular docking studies with ATP and aminoglycoside substrates were conducted using AutoDock Vina (47). The genetic environment was visualized and analyzed using EasyFig (48).

Cloning, expression, and purification of the APH(3′)-IVb protein

The aph(3′)-IVb gene was cloned into the pET32a and pLMF03 vectors using a seamless cloning strategy (Sangon Biotech, Shanghai, China). The resulting recombinant plasmids, pET32a-aph(3′)-IVb and pLMF03-aph(3′)-IVb, were verified by Sanger sequencing (Sangon Biotech) and subsequently transformed into E. coli BL21(DE3) and R. anatipestifer ATCC 11845, respectively.

For protein characterization, APH(3′)-IVb was overexpressed in E. coli BL21(DE3) using the pET28a expression system. Briefly, the aph(3′)-IVb coding sequence was inserted into pET28a to generate an N-terminal His₆-tagged fusion protein containing a thrombin cleavage site. Cultures were grown at 37°C until the optical density at 600 nm (OD₆₀₀) reached 0.6–0.8. Expression was induced by the addition of isopropyl β-D-1-thiogalactopyranoside to a final concentration of 0.5 mM, followed by incubation for 24 h at 25°C. Cells were harvested by centrifugation (5,000 × g, 10 min, 4°C), resuspended in lysis buffer (20 mM Tris-HCl, 150 mM NaCl, 3 mM β-mercaptoethanol, 0.5% NP-40, pH 8.0), and disrupted via sonication. Cellular debris was removed by centrifugation (10,000 × g, 30 min, 4°C). The supernatant was incubated with pre-equilibrated Ni-NTA agarose (Beyotime, Shanghai, China) for 8 h at 4°C with gentle agitation, followed by purification via affinity chromatography. The His₆ tag was excised by thrombin digestion (25°C, 3 h), and the reaction mixture was passed through a Ni-NTA column to remove free tags and uncleaved protein. Protein purity was assessed via SDS-PAGE, and concentrations were determined spectrophotometrically and using the BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA).

Construction of deletion strain RCAD1101Δaph(3′)-IVb, complementation strain RCAD1101Δaph(3′)-IVb(pLMF03-aph(3′)-IVb), and site-directed mutagenesis

The deletion strain RCAD1101Δaph(3′)-IVb was generated using an unmarked gene deletion strategy (49), with minor modifications. Briefly, the upstream (aph(3′)-IVbL, 800 bp) and downstream (aph(3′)-IVbR, 800 bp) fragments of the aph(3′)-IVb gene were amplified from RCAD1101 genomic DNA using the primers aph(3′)-IVbL-F1/aph(3′)-IVbL-R1 and aph(3′)-IVbR-F1/aph(3′)-IVbR-R1 (Table S1), respectively. Concurrently, the cfxA-SacB cassette was amplified from the plasmid pBAD24::cfxA-SacB using the primers cfxA-SacB-F/cfxA-SacB-R. These three fragments were assembled using a seamless cloning kit to generate the fusion fragment aph(3′)- IVbL-cfxA-SacB-aph(3′)- IVbR fragment which was subsequently amplified using primers aph(3′)- IVbL-F1 and aph(3′)- IVbR-R2.

The purified fusion fragment (2 μg) was incubated with 1 mL of RCAD1101 suspension (OD₆₀₀ = 1) for 12 h to induce natural transformation, as described previously (34). Transformants were selected on TSA plates containing cefoxitin (1 mg/L), and the intermediate mutant strain RCAD1101Δaph(3′)-IVb::cfx-sacB was isolated. Subsequently, the 1,000 bp upstream fragment (up) and 1,000 bp downstream (down) fragments of aph(3′)-IVb were amplified with the primers aph(3′)-IVbL-F2/aph(3′)-IVb-overlap and aph(3′)-IVbR-F/ aph(3′)-IVb-R2. These two PCR fragments were assembled using the seamless cloning kit. The resulting up-down fragment was incubated with the intermediate mutant RCAD1101Δaph(3′)-IVb::cfxA-sacB for 12 h and counter-selected on a TSA plate containing 15% sucrose to obtain the final deletion mutant RCAD1101Δaph(3′)-IVb.

To construct the complementation strain, the recombinant plasmid pLMF03-aph(3′)-IVb was introduced into the mutant strain RCAD1101Δaph(3′)-IVb via natural transformation as described above. Transformants were selected on TSA plates supplemented with cefoxitin (1 mg/L) and confirmed by PCR analysis. The resulting complementation strain was designated RCAD1101Δaph(3′)-IVb(pLMF03-aph(3′)-IVb).

Site-directed mutants were further constructed using in vitro site-directed mutagenesis (26). The upstream and downstream fragments of the target aph(3′)-IVb gene were amplified using mutagenic primers (Table S1). The amplified fragments were fused by overlap extension PCR to generate mutant alleles, which were subsequently cloned into the pET32a vector for expression and purification as described in “Cloning, expression, and purification of the APH(3′)-IVb protein.”

Antimicrobial susceptibility testing

Minimum inhibitory concentrations (MICs) were determined using the standard broth microdilution method in accordance with Clinical and Laboratory Standards Institute guidelines (50). Notably, butirosin was excluded from the testing panel due to its commercial unavailability. E. coli ATCC 25922 served as the quality-control strain.

Given the absence of standardized antibiotic breakpoints for R. anatipestifer, strain ATCC 11845 was utilized as the reference standard to assess the resistance phenotype of RCAD1101. This selection was based on the fact that ATCC 11845 harbors no known resistance determinants other than the intrinsic aminoglycoside efflux pumps (12, 14). The inoculum density was adjusted to approximately 10⁷ CFU/mL (100 μL per well). Wells containing inoculated broth without antibiotics served as growth controls, while uninoculated wells served as sterility controls. All assays were performed in triplicate.

Enzyme kinetics

The enzyme kinetics of APH(3′)-IVb were assessed using a previously described assay (51, 52) with minor modifications. Briefly, kinetic parameters were determined using a continuous spectrophotometric assay that couples ADP production from aminoglycoside phosphorylation to NADH oxidation via pyruvate kinase (PK) and lactate dehydrogenase (LD). ADP production was quantified by monitoring the decrease in absorbance at 340 nm with a UV-VIS spectrophotometer (U-3900, Hitachi, Japan) at 25°C.

Reactions were initiated by adding APH(3′)- IVb (final concentration 150 nM) to a 250 μL mixture containing 100 mM HEPES (pH 7.0), 10 mM MgCl₂, 20 mM KCl, 2 mM phosphoenolpyruvate, 100 μM NADH, a commercial PK/LD enzyme mixture (Sigma P0294; 18–26 U/mL PK and 25–35 U/mL LD), 2 mM ATP, and aminoglycosides at various concentrations. Steady-state velocities were calculated from the linear phase of reaction progress curves and plotted against substrate concentration. Data were fitted to the Michaelis-Menten equation, v = (V_max[S])/( K_m + [S]), by non-linear least-squares regression, using GraphPad Prism 9 (GraphPad Software, Inc., San Diego, CA, USA) to determine K_m and k_cat values. In this equation, v represents the steady-state velocity, V_max the maximal velocity, [S] the substrate concentration, K_m the Michaelis constant, and k_cat the turnover number, calculated from V_max = k_cat [E], where [E] is the enzyme concentration.

Prevalence of aph(3′)-IVb in R. anatipestifer field isolates and its distribution across ecological niches

To investigate the prevalence of the aph(3′)-IVb gene in field isolates, 741 R. anatipestifer isolates maintained in our laboratory collection were screened via PCR using primers aph(3′)-IVb-F1 and aph(3′)-IVb-R1 (Table S1). PCR amplification was performed using 2× Taq Master Mix (Dye Plus; Vazyme, Nanjing, China) under the following conditions: initial denaturation at 95°C for 5 min, followed by 30 cycles of (95°C/30 s, 67°C/30 s, 72°C/45 s), and a final extension at 72°C for 8 min. All PCR products (approx. 789 bp) were verified by Sanger sequencing.

To characterize the global distribution of aph(3′)-IVb, we performed a BLASTp search of APH(3′)-IVb sequences against the NCBI Non-Redundant database (https://blast.ncbi.nlm.nih.gov/Blast.cgi), retrieving complete protein sequence hits with an identity threshold of >90%. Additionally, we screened publicly available genome sequences (n = 234, data as of 29 August 2025) from the NCBI RefSeq database (https://www.ncbi.nlm.nih.gov/refseq/) in conjunction with the aph(3′)-IVb-positive isolates identified in this study to assess the overall prevalence of aph(3′)-IVb within the R. anatipestifer population.

Natural transformation of aph(3′)-IVb into R. anatipestifer ATCC 11845 genome

To investigate the transferability of the aph(3′)-IVb gene, we utilized the reference strain ATCC 11845 as the recipient host. The genomic DNA of R. anatipestifer RCAD1101 and the aph(3′)-IVb gene, along with its flanking 5′ and 3′ untranslated regions (UTRs), were prepared as donor materials (Table S1). Briefly, different concentrations of genomic DNA (0.5, 1, 5, 10, and 20 μg) or 2 μg of the purified PCR fragment were individually incubated with 1 mL of an ATCC 11845 suspension (OD₆₀₀ = 1) for 12 h to induce natural transformation. Transformants were selected on Tryptic Soy Agar (TSA) plates supplemented with neomycin (256 mg/L), and the resulting strains were subsequently confirmed.

DATA AVAILABILITY

The genome sequence of R. anatipestifer strain RCAD1101 was deposited in GenBank under accession number CM136845.1, while its 16S rRNA gene sequence has been assigned accession number PX365741.1. Additionally, the nucleotide sequence of the novel resistance gene aph(3′)-IVb was deposited under accession number PX763553.1.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aac.01631-25.

Supplemental figures. aac.01631-25-s0001.docx.

Fig. S1 to S6.

Supplemental tables. aac.01631-25-s0002.xlsx.

Tables S1 to S5.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.