Functional Characterization of FeoAB in Iron Acquisition and Pathogenicity in Riemerella anatipestifer

ABSTRACT The bacterium Riemerella anatipestifer requires iron for growth, but the mechanism of iron uptake is not fully understood. In this study, we disrupted the Feo system and characterized its function in iron import in R. anatipestifer ATCC 11845. Compared to the parent strain, the growth of the ΔfeoA, ΔfeoB, and ΔfeoAB strains was affected under Fe3+-limited conditions, since the absence of the feo system led to less intracellular iron than in the parent strain. In parallel, the ΔfeoAB strain was shown to be less sensitive to streptonigrin, an antibiotic that requires free iron to function. The sensitivity of the ΔfeoAB strain to hydrogen peroxide was also observed to be diminished compared with that of the parent strain, which could be related to the reduced intracellular iron content in the ΔfeoAB strain. Further research revealed that feoA and feoB were directly regulated by iron through the Fur regulator and that the transcript levels of feoA and feoB were significantly increased in medium supplemented with 1 mM MnCl2, 400 μM ZnSO4, and 200 μM CuCl2. Finally, it was shown that the ΔfeoAB strain of R. anatipestifer ATCC 11845 was significantly impaired in its ability to colonize the blood, liver, and brain of ducklings. Taken together, these results demonstrated that FeoAB supports ferrous iron acquisition in R. anatipestifer and plays an important role in R. anatipestifer colonization. IMPORTANCE In Gram-negative bacteria, the Feo system is an important ferrous iron transport system. R. anatipestifer encodes an Feo system, but its function unknown. As iron uptake may be required for oxidative stress protection and virulence, understanding the contribution of iron transporters to these processes is crucial. This study showed that the ΔfeoAB strain is debilitated in its ability to import iron and that its intracellular iron content was constitutively low, which enhanced the resistance of the deficient strain to H2O2. We were surprised to find that, in addition to responding to iron, the Feo system may play an important role in sensing manganese, zinc, and copper stress. The reduced colonization ability of the ΔfeoAB strain also sheds light on the role of iron transporters in host-pathogen interactions. This study is important for understanding the cross talk between iron and other metal transport pathways, as well as the pathogenic mechanism in R. anatipestifer.

which in turn enables its contribution to various oxidation/reduction processes (4). However, the amount of free iron is very low in vertebrate hosts. In the human body, the majority of iron is bound to heme proteins, such as hemoglobin, iron/sulfur-containing enzymes, iron-binding transport proteins (transferrin), iron-binding glycoproteins (lactoferrin), or the main intracellular iron storage proteins (ferritin) (4)(5)(6)(7)(8). Consequently, pathogenic microorganisms have developed multiple high-affinity iron acquisition systems, including siderophore-mediated iron uptake systems, heme uptake systems, and ferrous iron transport systems (1,(9)(10)(11). In contrast to insoluble ferric iron, many ferrous iron importers have also been characterized in Gram-negative bacteria, such as Feo systems, MntH, ZupT, EfeUOB, IroT, YfeABCDE, and FutABC (12). Among them, the Feo system is mainly committed to the uptake of ferrous iron and widely exists in prokaryotes (2,13). Since iron acquisition represents a critical factor in bacterial virulence, mutations in genes involved in iron acquisition or utilization systems may lead to loss of pathogenicity, and these genes could be used as targets for the development of new antibacterial drugs (14)(15)(16).
Riemerella anatipestifer is a Gram-negative bacterium belonging to the family Flavobacteriaceae that causes serositis and septicemia in domestic ducks and other birds (17). The mortality rate of ducklings infected with the bacterium can reach 75% or even higher (18). At least 21 different serotypes of R. anatipestifer have been identified, with no cross-protection between them (18). Clinical isolates of R. anatipestifer have been found to be resistant to a variety of antibiotics (19,20). As a result, existing vaccines and antibiotics have limited effectiveness. R. anatipestifer requires iron for growth, and genome sequencing and analysis showed that R. anatipestifer carries several genes related to iron metabolism (21). Our previous study identified the functions of some of these genes. For example, TonB1 and TonB2 are both involved in the utilization of iron/heme, but TonB3 is not (22,23). The TonB-dependent receptors B739_1208 and B739_1343 were found to be involved in iron utilization and the virulence of R. anatipestifer (24,25). RhuR (B739_1416) and RhuA (B739_1417) were characterized as a TonB-dependent heme transporter and an outer membrane exposed heme binding protein, respectively (26). The DNA-binding protein from starved cells, Dps, was found to protect R. anatipestifer from oxidative stress damage by chelating excess iron and binding DNA (27). Iron homeostasis in R. anatipestifer is mainly managed by the regulator Fur (28). Sequence analysis showed that R. anatipestifer possesses homologues of the ferrous iron transporters Feo and MntH. In addition to transporting iron, MntH is primarily an importer of Mn 21 (12). In this study, we investigated the function of the Feo system in R. anatipestifer. It was shown that the Feo system is essential for Fe 21 utilization and is closely linked to metal homeostasis, oxidative stress resistance, and colonization in R. anatipestifer ATCC 11845.

RESULTS
Sequence analysis and deletion of feoA and feoB in R. anatipestifer ATCC 11845. Database searches of the R. anatipestifer ATCC 11845 genome revealed the presence of two predicted ferrous ion transport genes, ferrous iron transport protein A (feoA, RA0C_RS07485) and ferrous iron transport protein B (feoB, RA0C_RS07490). The feoA and feoB genes appear to be in an operonic structure (Fig. 1), and the products of these genes share partial identity at the amino acid level with the Escherichia coli FeoA (43%) and FeoB (54%) proteins, respectively (see Fig. S1 in the supplemental material). To determine the role(s) of the two putative ferrous iron transport proteins, feoA, feoB, and feoAB isogenic mutants were constructed using the natural transformation-based knockout method described in a previous study (29). The deletion of each gene was verified by PCR amplification of each genomic region in the wild-type and deletion strains (Fig. S2A). To confirm that the deletions were nonpolar, total bacterial RNAs were extracted from the wild-type (WT), DfeoA, DfeoB, and DfeoAB strains, and real-time PCR was used to measure the transcript levels of genes downstream of feoA and feoB. The gene downstream of feoA is feoB, while the gene downstream of feoB is RA0C_RS07490, which encodes a GNAT (Gcn5related N-acetyltransferase) family protein. As shown in Fig. S2B, compared to the WT, Functional Characterization of FeoAB in R. anatipestifer Microbiology Spectrum deletion of feoA or feoB did not result in transcriptional changes in downstream genes, suggesting that deletion of the feoA or feoB gene did not cause a polar effect. Deletion of feoA and feoB significantly reduced the growth of R. anatipestifer in the presence of the iron chelator EDDHA. To determine whether feoA and feoB are involved in ferrous iron uptake in R. anatipestifer, we compared the growth ability of R. anatipestifer ATCC 11845 and the DfeoA, DfeoB, and DfeoAB strains in Trypticase soy broth (TSB) and TSB containing the iron chelator ethylenediamine-N,N9-bis( [2-hydroxyphenyl] acetic acid) (EDDHA) at a final concentration of 40 mM. As shown in Fig. 2, there was no significant difference in the growth of the parent strain, three feo mutants, and the complemented strains under iron-replete conditions. In contrast, the growth of the DfeoA, DfeoB, and DfeoAB strains was slower than that of the parent strain in TSB supplemented with 40 mM EDDHA (Fig. 2), and growth was restored in the complemented strains (Fig. 2). Notably, the observed growth defect of the DfeoAB strain under iron-limited conditions was only fully restored by introducing pFY02::feoAB, but pFY02::feoA and pFY02:: feoB did not revert growth to the wild-type level when introduced into the DfeoAB strain (Fig. 2C). The above-described results showed that all the feo mutants exhibited impaired growth in the presence of the iron chelator EDDHA, suggesting that the FeoAB transport system of R. Anatipestifer ATCC 11845 may participate in ferrous iron uptake.
Deletion of feoAB leads to reduced intracellular iron content. To characterize the ability of the Feo system to transport iron, the intracellular iron content was determined. R. anatipestifer ATCC 11845 pFY02 (WT), DfeoAB pFY02 (DfeoAB), and DfeoAB pFY02::feoAB  (DfeoAB C ) were cultured in TSB medium and TSB supplemented with 40 mM EDDHA, and the intracellular iron content was measured by inductively coupled plasma mass spectrometry (ICP-MS) as described in Materials and Methods. As shown in Fig. 3A, the DfeoAB strain had a significantly lower amount of intracellular iron than the parent strain in both TSB medium and TSB medium supplemented with EDDHA, suggesting that iron uptake is markedly attenuated in a strain lacking feoAB. Intracellular levels of other metals were also measured in both TSB medium and TSB medium supplemented with EDDHA. As shown in Fig. 3, the levels of Mn, Zn, and Cu were increased in the DfeoAB strain compared with the parent strain. In addition, the intracellular iron content of all strains under iron-limited conditions decreased, while the levels of other metals (Mn, Zn, and Cu) increased (Fig. 3). The results suggested that the DfeoAB mutation resulted in a decrease in intracellular Fe content and may have contributed to the homeostasis of other metals.
The feoAB mutant strain displayed reduced susceptibility to streptonigrin. Streptonigrin is a metal-dependent antibiotic that requires intracellular free iron for its anatipestifer ATCC 11845 pFY02 (WT), R. anatipestifer ATCC 11845 DfeoAB pFY02 (DfeoAB), and R. anatipestifer ATCC 11845 DfeoAB pFY02::feoAB (DfeoAB C ) were grown in 10 mL of TSB and TSB containing EDDHA at 37°C with shaking until the exponential phase, and the intracellular Fe content (A), the intracellular Mn content (B), the intracellular Zn content (C), and the intracellular Cu content (D) were measured by inductively coupled plasma-mass spectrometry. The total concentrations of ions are expressed as mg of ion per gram of protein content, and the values are the mean (6standard deviation) of at least three independent assays. Asterisks indicate statistically significant differences (*, P , 0.05; **, P , 0.01; ***, P , 0.001; ns, not significant).

Functional Characterization of FeoAB in R. anatipestifer
Microbiology Spectrum bactericidal action through the formation of reactive oxygen radicals (30,31). Since the intracellular Fe content of the DfeoAB strain was lower than that of the parent strain, we next measured the sensitivity of the DfeoAB strain to streptonigrin. As shown in Fig. 4A, the survival rate of the DfeoAB strain was 2-fold higher than that of the WT after treatment with 30 ng/mL streptonigrin, and the feoAB complemented strain showed a recovery of survival rate to the same level as the parental strain. The same change trend was observed after treatment with 50 ng/mL streptonigrin (Fig. 4A). These results suggest that the DfeoAB strain was less susceptible to streptonigrin than the parent strain, which may be due to impaired Feo-dependent ferrous iron transport resulting in reduced intracellular iron levels.
Deletion of feoAB increased the resistance of cells to H 2 O 2 treatment in comparison with the parent strain. A major cause of iron toxicity is the cross-reaction of excess intracellular iron with H 2 O 2 through the Fenton reaction, which results in harmful levels of oxidative stress that affect the growth and viability of bacteria (32). The DfeoAB strain exhibited reduced intracellular iron content, which may lead to an altered response of cells to oxidative stress. Therefore, R. anatipestifer ATCC 11845 pFY02 (WT), DfeoAB pFY02 (DfeoAB), and DfeoAB pFY02::feoAB (DfeoAB C ) were incubated in the presence of 5 mM and 10 mM H 2 O 2 for 30 min, and the survival rates were recorded as described in Materials and Methods. The WT and DfeoAB C exhibited similar survival rates of approximately 60.3% and 61% after exposure to 5 mM H 2 O 2 , respectively (Fig. 4B). However, the survival rate of the DfeoAB strain was higher than that of the parent strain, reaching 88.5% (Fig. 4B). After treatment with 10 mM H 2 O 2 , the survival rate of the DfeoAB strain was still higher than that of the parent strain and the complementation strain (Fig. 4B). These results indicated that knockout of feoAB decreased the sensitivity of R. anatipestifer to H 2 O 2 , which may also be related to the decrease in intracellular iron content in the DfeoAB strain.
The transcription of feoA and feoB was regulated by iron through Fur and was affected by metal stress. Given that the Feo system is involved in ferrous iron uptake, we next investigated whether the transcription of feoA and feoB is affected by iron availability. First, we measured the transcript levels of feoA and feoB in iron-replete medium and iron-depleted medium. The results showed that the transcription levels of feoA and feoB were ;5-fold higher in cells grown in medium containing 40 mM EDDHA than in cells grown in TSB (Fig. 5A), suggesting that feoA and feoB could be maximally expressed under iron-limited conditions. In most Gram-negative bacteria, including R. anatipestifer, iron acquisition is regulated by the transcription factor Fur (28). Therefore, we tested the transcription levels of feoA and feoB in the Dfur strain and its complementation Since iron homeostasis can also be affected by other metal stresses (33), we next measured the transcript levels of feoA and feoB under different metal stresses. The results showed that in response to 1 mM MnCl 2 , 400 mM ZnSO 4 , and 200 mM CuCl 2 , the transcript levels of feoA and feoB were also significantly increased in R. anatipestifer ATCC 11845 compared to the untreated control ( Fig. 5C to E). Taken together, these results indicated that the iron-dependent regulation of feoA and feoB transcription was mediated by Fur and could be affected by other metals.
The Feo system contributes to colonization by R. anatipestifer ATCC 11845 in ducklings. To investigate whether the Feo system plays a role in the colonization ability of R. anatipestifer ATCC 11845, ducklings were infected with 10 9 CFU of the parent strain and the DfeoAB strain. The results of bacterial colonization in the blood of the heart, liver, and brain are shown in Fig. 6. While the parent strain colonized at an average of approximately 5 Â 10 5 CFU per gram of the blood of the heart, the DfeoAB strain exhibited a severe defect in its ability to colonize. Similarly, the number of colonies recovered from liver tissues and brain tissues was significantly reduced for the DfeoAB strain compared to the parent strain (P , 0.0001). The above-described results indicated that the Feo system contributed to colonization by R. anatipestifer ATCC 11845 in ducklings.

DISCUSSION
Iron is an essential nutrient for most bacterial pathogens, as it serves as a cofactor for numerous enzymes (1). Due to the limited availability of iron in the environment and the host, pathogenic bacteria have evolved numerous mechanisms to acquire iron through iron transport systems, most of which are virulence-associated genes (34). Hence, iron transport systems could potentially promote bacterial survival during infection by facilitating iron transport. In our previous studies, we identified the mechanism of iron transport and the regulation of iron homeostasis, iron efflux, and iron chelation in R. anatipestifer (22,(24)(25)(26)(27)(28). To refine the understanding of the mechanism of iron metabolism in R. anatipestifer, the ferrous iron transport system Feo was studied in detail in this study.
The canonical feo operon was originally identified in Escherichia coli, and it encodes three proteins, the cytosolic proteins FeoA and FeoC and the integral membrane protein FeoB, which constitute the Feo system (35). FeoA and FeoB are strongly conserved, whereas FeoC is found in only ;13% of operons (36)(37)(38). FeoA is a small (;8 kDa) cytoplasmic b-barrel protein, and it has been hypothesized to interact with FeoB to affect functions (2,36,39,40). FeoC is also a small cytoplasmic protein, but it is not widespread among bacteria (36). Although several potential roles for FeoC have been proposed, these roles have not been experimentally tested in more than one species (39). FeoB is a transmembrane protein consisting of a G-protein domain, a GDP dissociation inhibitor (GDI) domain, and a C-terminal transmembrane (TM) domain (41)(42)(43). The G-protein domain of FeoB is not only responsible for binding and hydrolyzing GTP or ATP but is also important for Fe 21 transport (41,44,45). The GDI domain is the link between the G-protein domain and the TM region, which often adopts a novel fold called a "valve" (46). The TM region may serve as a channel region for ferrous iron transport (47,48).
In the genome of R. anatipestifer ATCC 11845, there were two putative proteins, FeoA and FeoB, which were annotated as ferrous iron transport proteins, but no sequences containing the FeoC domain or a C-rich peptide were found downstream of feoA and feoB. Mutant strains completely lacking feoA, feoB, and feoAB were generated and analyzed to determine if the absence of feo genes affected bacterial growth. The results showed that the loss of feoA, feoB, and feoAB affected growth in defined medium containing restricted iron (in the presence of the iron chelator EDDHA). Intracellular iron content measurement showed a clear distinction between the parent strain and the DfeoAB strain, regardless of the iron-rich condition and the iron-limited condition, indicating that the iron uptake capacity of the DfeoAB strain was weakened. Streptonigrin is an antibiotic that works when the iron concentration inside bacterial cells is high. We found that the DfeoAB strain increased resistance to streptonigrin compared to R. anatipestifer ATCC 11845 (Fig. 4A), which also confirmed the reduction in intracellular iron. Notably, although the DfeoAB strain exhibited lower levels of intracellular iron in TSB medium, FIG 6 Colonization by R. anatipestifer ATCC 11845 and the DfeoAB strain in the blood, brain, and liver of 3-day-old ducklings. Doses of 10 9 CFU of R. anatipestifer ATCC 11845 (WT) and the DfeoAB strain were prepared and injected into the leg muscles of 3-day-old ducklings. At 24 h postinfection, bacteria were isolated from the liver, the brain, and the blood of the heart. The data points represent the CFU/g values of the indicated organs in individual animals; the bars show the mean values (n = 6). Statistical significance was determined by two-way ANOVA (****, P , 0.0001). the growth of the DfeoAB strain in TSB medium was not affected, which could be an indication that the wild-type strain has an excess of iron beyond what is required for growth.
Although iron is an indispensable nutrient for R. anatipestifer, it is also toxic, as both ferrous and ferric iron are known to induce intracellular ROS formation via Fenton reactions and cause oxidative damage to biomolecules (49). As shown above, the Feo system of R. anatipestifer ATCC 11845 may be a major ferrous iron uptake system and may contribute to elevated iron levels in the bacterial cell, which is related to the oxidative stress response. We compared the survival rates of the DfeoAB strain and the parent strain when exposed to 5 mM and 10 mM H 2 O 2 , and the results showed that the DfeoAB strain had a higher survival rate after incubation with H 2 O 2 (Fig. 4B), demonstrating that the presence of the Feo system contributes to the oxidative stress sensitivity of R. anatipestifer. This is consistent with the observations made in previous work with Porphyromonas gingivalis (50), showing that the deleterious role of iron contributes to the generation of oxidative stress. Furthermore, iron or manganese could act as a cofactor of antioxidant enzymes such as superoxide dismutase (SodB and SodA). Although the decreased intracellular iron level in the DfeoAB strain may have led to the inactivation of SodB, the increased intracellular manganese level in the DfeoAB strain may have increased sodA expression, which could be one of the reasons why the feoAB-deficient strain was more resistant to H 2 O 2 .
To determine the regulatory mechanism of the Feo system, we monitored the mRNA levels of feoA and feoB when the parent strain was grown under different iron conditions to determine whether these genes were regulated by iron. The results showed that the transcript levels of feoA and feoB were increased under iron-limited conditions, suggesting that the expression of feoA and feoB was iron-dependent (Fig. 5A). The mechanisms of iron regulation are usually mediated by Fur (28). In this study, we found that feoA and feoB were also regulated by Fur (Fig. 5B), and the Fur box of R. anatipestifer (51) was present in the promoter regions of feoAB (Fig. 1). In the results of intracellular metal content determination, we observed different degrees of increase in intracellular Mn, Zn, and Cu levels ( Fig. 3B to D). Although the DfeoAB strain exhibited a decrease in intracellular iron content, we speculated that the cells may also have increased their uptake of other metal ions in unknown ways to help maintain intracellular metal homeostasis. In our previous study, we found that the transcription level of feo genes was significantly upregulated in the transcriptome data of R. anatipestifer under excessive manganese conditions (33). Here, the determination of the transcript levels of feoA and feoB under different metal stresses, such as Mn, Zn, and Cu, revealed that feoA and feoB were obviously upregulated ( Fig. 5C to E). Whether this was a direct effect or was indirectly caused by changes in iron levels is still unknown, and further investigations are warranted to clarify the detailed molecular mechanisms underlying the role of feoAB in sensing other metals.
Iron availability is an important determinant of virulence, and genes related to iron metabolism have been extensively studied as candidate virulence factors (24,25,27,28,52). Indeed, the involvement of the Feo system in bacterial colonization and virulence has also been reported (53)(54)(55). Since R. anatipestifer ATCC 11845 is an attenuated strain (56), in this study, we evaluated the effect of deletion of feoAB on colonization by R. anatipestifer ATCC 11845 in ducklings. At 1 day after infection, significantly reduced colonization levels were observed in the blood, spleen, and brain of the DfeoAB strain-infected ducklings compared to those of ducklings infected with the wild-type strain (Fig. 6), suggesting that the DfeoAB strain exhibited a severe defect in its ability to colonize.
Given the importance of iron in oxidative stress induction and virulence (55,57), it is crucial to understand the contribution of the major R. anatipestifer ferrous iron transporter to these processes. Here, we have shown that the feoA and feoB genes from R. anatipestifer ATCC 11845 encode a functional component of the ferrous iron acquisition system. The Feo system of R. anatipestifer provides the bacteria with an advantage for growth and colonization under the iron-restricted condition of the host by increasing iron uptake and promoting in vivo persistence. Overall, this study highlights the importance of ferrous iron acquisition in the pathogenesis of R. anatipestifer and will be useful in identifying novel drug and vaccine targets.

MATERIALS AND METHODS
Bacterial strains and plasmids used in this study. The bacterial strains and plasmids used in this study are listed in Table S1. The primers used in this study are listed in Table S2.
Media and growth conditions. E. coli strains were grown in LB at 37°C with shaking at 180 rpm. R. anatipestifer ATCC 11845 strains were grown on LB agar supplemented with 5% sheep blood or in TSB (Solarbio, China) and GC broth (GCB) as described in a previous study (29). EDDHA was purchased from Alfa Chemistry Protheragen, Inc. (USA) and dissolved in distilled water at a concentration of 100 mM. EDDHA was added to the medium to chelate Fe 31 (58). Antibiotics were routinely used at the following concentrations (per milliliter): 1 mg erythromycin (Erm) (screening for deletion strains), 1 mg cefoxitin (Cfx) (screening for complementation strains), and 50 mg kanamycin (Kana) (screening for R. anatipestifer strains) (56).
Construction of the mutant strains. The mutant strains were constructed based on the natural transformation method as described in a previous study (29). To illustrate this gene deletion method, the feoA gene deletion of R. anatipestifer ATCC 11845 was used as an example. The approximately 800bp upstream sequence and 800-bp downstream sequence of the feoA gene were amplified by PCR using the primer pairs feoA up P1/feoA up P2 and feoA down P1/feoA down P2, respectively. The 994-bp sequence containing the ErmR cassette was then amplified from the R. anatipestifer CH-1 genome using the primer pair feoA erm P1/feoA erm P2. The PCR fragments (feoA upstream, ErmR cassette and feoA downstream) were ligated using the overlap PCR method, and 1 mg of the purified fragments was incubated with 300 mL of R. anatipestifer ATCC 11845 (optical density at 600 nm [OD 600 ], 1) for 1 h at 37°C. Mutants were selected on sheep blood plates supplemented with Erm (1 mg/mL) for 2 days at 37°C and identified by PCR. The R. anatipestifer ATCC 11845 DfeoB strain and the R. anatipestifer ATCC 11845 DfeoAB strain were constructed using the same approach.
Construction of the complementation strains. The feoA gene of R. anatipestifer ATCC 11845 was amplified from the genome of R. anatipestifer ATCC 11845 using the primer pair feoA P1/feoA P2, containing the XbaI/XhoI restriction sites. The PCR products were purified, digested, and cloned into the shuttle plasmid pFY02 to generate pFY02::feoA as described in a previous study (59). The plasmid pFY02::feoA was then transformed into Escherichia coli S17-1 cells, and the recombinant plasmid was introduced into the R. anatipestifer ATCC 11845 DfeoA strain by conjugation as described elsewhere (28). Transconjugants were selected on blood agar plates supplemented with Cfx (1 mg/mL) and Kan (50 mg/mL) and identified by PCR amplification.
Measurement of metal content. ICP-MS analyses were used to measure the metal content in R. anatipestifer strains as described in previous studies (33). Briefly, R. anatipestifer ATCC 11845 pFY02, R. anatipestifer ATCC 11845 DfeoAB pFY02, and R. anatipestifer ATCC 11845 DfeoAB pFY02::feoAB were grown in 10 mL of TSB or TSB containing EDDHA at 37°C for 4 to 6 h until the exponential phase. Cells were washed with phosphate-buffered saline (PBS) containing 0.1 M EDTA. Cell pellets were then resuspended in 400 mL of ultrapure water and lysed using a FastPrep-96 automated homogenizer (MP Biomedicals). The cytosolic content was separated by centrifugation, the total protein concentration was measured using a bicinchoninic acid (BCA) protein assay kit (Thermo Scientific, Waltham, MA), and the supernatants were stored at 280°C for 24 h. Then, 600 mL of 5% HNO 3 with 0.1% Triton X-100 was added to the supernatants. Samples were incubated at 95°C for 30 min and centrifuged at 12,000 rpm for 10 min, and the supernatants were diluted to 2% HNO3 for detection. The metal content was analyzed by ICP-MS (Elan DRC II, Perkin-Elmer) using Ga as an internal standard. The total ion concentration is expressed as micrograms of ion per gram of protein. Triplicate cultures of each strain were analyzed in a single experiment, and the experiment was repeated at least three times.
Streptonigrin sensitivity assay. The streptonigrin resistance assay was performed as previously described (28). Briefly, the tested strains (R. anatipestifer ATCC 11845 pFY02, R. anatipestifer ATCC 11845 DfeoAB pFY02, and R. anatipestifer ATCC 11845 DfeoAB pFY02::feoAB) were cultured in TSB medium Functional Characterization of FeoAB in R. anatipestifer Microbiology Spectrum overnight, and then the bacteria were grown in fresh TSB at 37°C and 180 rpm until the exponential phase (OD 600 , approximately 1 to 1.5). Cells were harvested by centrifugation at 6,000 rpm for 10 min, and pellets were diluted with sterile PBS up to an OD 600 of 0.5 in a volume of 1 mL. Streptonigrin (Sigma-Aldrich, St. Louis, USA) was diluted to 1 mg/mL with sterile PBS and added to each tube of bacterial solution at final concentrations of 0 ng/mL, 30 ng/mL, and 50 ng/mL. After incubation at 37°C for 30 min, the bacterial solution was diluted and plated on GCB plates for colony counting (designated T0, T30, and T50). After incubation at 37°C for 24 h, the grown colonies were counted. The survival rate was calculated as (T30/T0) Â 100% and (T50/T0) Â 100%. Experiments were performed in triplicate using three biological replicates. H 2 O 2 challenge. For H 2 O 2 challenge, bacterial cells were prepared as described in a previous study (27,28). Briefly, the strains (R. anatipestifer ATCC 11845 pFY02, R. anatipestifer ATCC 11845 DfeoAB pFY02, and R. anatipestifer ATCC 11845 DfeoAB pFY02::feoAB) were grown in TSB medium to the exponential phase and washed twice in PBS. The cell suspension was diluted to an OD 600 of 0.5 in a volume of 1 mL, and 5 mM H 2 O 2 and 10 mM H 2 O 2 were added to the samples. The same volume of PBS was added to the control group. The suspension was incubated at 37°C for 30 min. After incubation, the above-described samples were diluted and plated on GCB plates for CFU counting. Colonies from the control and experimental groups were counted, and the number of living bacteria was expressed as T0, T5, and T10. Survival rates were expressed by (T5/T0) Â 100% and (T10/T0) Â 100%. Experiments were performed in triplicate using three biological replicates.
Real-time PCR. For RNA extraction, the strain R. anatipestifer ATCC 11845 was grown in TSB medium and TSB medium containing 40 mM EDDHA, 1 mM MnCl 2 , 400 mM ZnSO 4 , and 200 mM CuCl 2 . R. anatipestifer ATCC 11845 and R. anatipestifer ATCC 11845 Dfur were grown in TSB medium. All the strains were grown at 37°C and 180 rpm until the exponential phase. Then, 6 Â 10 9 CFU (1 OD 600 = 2 Â 10 9 CFU) (22) of the cells was mixed with 1 mL RNAprotect bacterial reagent (Qiagen, 76506), and the total RNA was isolated using the RNeasy Protect Bacteria minikit (Qiagen, 74524) as described in a previous study (28). cDNA was synthesized from 1,000 ng template RNA with HiScript QRT SuperMix for quantitative PCR (qPCR) (Vazyme, R123-01). The cDNA was quantified by real-time PCR using SYBR green master mix (Vazyme, Q111-01) on a CFX Connect real-time system (Bio-Rad), and the mRNA levels of the feoA and feoB genes were normalized to those of 16S rRNA as described in a previous study (60). Experiments were performed in triplicate using three biological replicates.
Colonization assays. For bacterial colonization assays, R. anatipestifer ATCC 11845 and the R. anatipestifer ATCC 11845 DfeoAB strain were grown in 50 mL of TSB (with the initial OD adjusted to 0.1) at 37°C with shaking until the exponential phase. The bacteria were then collected by centrifugation at 6,000 rpm for 10 min, resuspended in 10 mL of PBS, and centrifuged again. This process was repeated three times to wash the bacteria. The OD 600 values of the bacterial suspensions were measured, and the suspensions were adjusted to 5 Â 10 9 CFU/mL (1 OD 600 = 2 Â 10 9 CFU) (22). Ducklings (3 days old) were infected intramuscularly with R. anatipestifer ATCC 11845 and the R. anatipestifer ATCC 11845 DfeoAB strain (10 9 CFU/duckling, 10 ducklings/group). At 24 h postinfection, six ducklings in each group were randomly selected and euthanized by forced CO 2 inhalation. Liver, brain, and blood from the heart were collected in sterile Whirl-Pak bags (Nasco, B01245WA, USA), weighed, and added to PBS (0.1 g of sample added to 0.9 mL of PBS). All the samples were then transferred to 100 sterile tubes, homogenized using a FastPrep-24 instrument (MP, USA), serially diluted in PBS, and plated on blood agar plates supplemented with 50 mg/mL Kana (for R. anatipestifer ATCC 11845) or on blood agar plates supplemented with 50 mg/mL Kana and 1 mg/mL Erm (for R. anatipestifer ATCC 11845 DfeoAB). The plates were incubated overnight at 37°C to determine the bacterial CFU count.
Ethics approval. The animal study was reviewed and approved by the local animal welfare authorities and the ethics committee of Sichuan Agricultural University. The 1-day-old ducklings were purchased from Grimaud Farms in Chengdu, Sichuan, China, and housed in our animal facilities with free access to food and water.
Statistical analysis. All experimental data are expressed as the mean 6 1 standard deviation (SD). Statistical analysis was performed using Prism 8 (GraphPad Software, California, USA). Independent Student's t test was used to compare two groups, and one-way analysis of variance (ANOVA) or two-way ANOVA was used to compare multiple groups. A P value of ,0.05 was considered significant.
Data availability. The nucleotide sequences of R. anatipestifer ATCC 11845 are deposited in GenBank under accession number CP003388.1. The GenBank accession numbers of the Feo system and its homologues are as follows: WP_004916263.1 for FeoA of R. anatipestifer ATCC 11845, NP_417867.1 for the homologue in Escherichia coli, WP_014411322.1 for FeoB of R. anatipestifer ATCC 11845, and NP _417868.1 for the homologue in Escherichia coli.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, DOCX file, 0.02 MB. SUPPLEMENTAL FILE 2, DOCX file, 0.9 MB. M.L., Mingshu W., and A.C. provided overall guidance and leadership to the research team, overseeing the project from start to finish. M.H., Mengying W., and Y.F. conceived and designed the research, performed experiments, and wrote the first draft of the article. Q.G., R.J., and D.Z. conducted the primary data analysis and developed the statistical models used in the study. S.C. and X.Z. participated in the experiments. Z.W., Q.Y., and Y.W. provided critical insights into the theoretical framework of the research. S.Z., J.H., and B.T. helped to design the study's methodology. Y.H. and X.O. played a key role in collecting and organizing the data used in the study. S.M. and D.S. supervised the studies and corrected the manuscripts. All authors read and approved the final manuscript.

ACKNOWLEDGMENTS
We declare no competing financial interests.