XRE-Type Regulator BioX Acts as a Negative Transcriptional Factor of Biotin Metabolism in Riemerella anatipestifer

ABSTRACT

Biotin is essential for the growth and pathogenicity of microorganisms. Damage to biotin biosynthesis results in impaired bacterial growth and decreased virulence in vivo. However, the mechanisms of biotin biosynthesis in Riemerella anatipestifer remain unclear. In this study, two R. anatipestifer genes associated with biotin biosynthesis were identified. AS87_RS05840 encoded a BirA protein lacking the N-terminal winged helix-turn-helix DNA binding domain, identifying it as a group I biotin protein ligase, and AS87_RS09325 encoded a BioX protein, which was in the helix-turn-helix xenobiotic response element family of transcription factors. Electrophoretic mobility shift assays demonstrated that BioX bound to the promoter region of bioF. In addition, the R. anatipestifer genes bioF (encoding 7-keto-8-aminopelargonic acid synthase), bioD (encoding dethiobiotin synthase), and bioA (encoding 7,8-diaminopelargonic acid synthase) were in an operon and were regulated by BioX. Quantitative reverse transcription-PCR showed that transcription of the bioFDA operon increased in the mutant Yb2ΔbioX in the presence of excessive biotin, compared with that in the wild-type strain Yb2, suggesting that BioX acted as a repressor of biotin biosynthesis. Streptavidin blot analysis showed that BirA caused biotinylation of BioX, indicating that biotinylated BioX was involved in metabolic pathways. Moreover, as determined by the median lethal dose, the virulence of Yb2ΔbioX was attenuated 500-fold compared with that of Yb2. To summarize, the genes birA and bioX were identified in R. anatipestifer, and BioX was found to act as a repressor of the bioFDA operon involved in the biotin biosynthesis pathway and identified as a bacterial virulence factor.

IMPORTANCERiemerella anatipestifer is a causative agent of diseases in ducks, geese, turkeys, and various other domestic and wild birds. Our study reveals that biotin synthesis of R. anatipestifer is regulated by the BioX through binding to the promoter region of the bioF gene to inhibit transcription of the bioFDA operon. Moreover, bioX is required for R. anatipestifer pathogenicity, suggesting that BioX is a potential target for treatment of the pathogen. R. anatipestifer BioX has thus been identified as a novel negative regulator involved in biotin metabolism and associated with bacterial virulence in this study.

INTRODUCTION

Riemerella anatipestifer is a Gram-negative, nonmotile, non-spore-forming, rod-shaped bacterium. It belongs to the family Flavobacteriaceae in rRNA superfamily V based on 16S rRNA gene sequence analyses (1). R. anatipestifer causes the anatipestifer syndrome in ducks, characterized by diarrhea, lethargy, and respiratory and nervous symptoms, which can lead to high mortality and consequently great economic losses (2). For ducks under about 8 weeks of age, R. anatipestifer infection is often acute, with the mortality rate usually between 10% and 30%, but mortality of as high as 75% has been recorded for infected duck farms (2–4). Until now, 21 serotypes have been identified in the world, and there is poor cross-protection among these serotypes (5). Various antibiotics are currently used to prevent and control R. anatipestifer infection in ducks, but they accelerate the emergence of drug-resistant strains. Once infection sets within a duck flock, the bacterium can become endemic with repeated infectious episodes possible, making eradication difficult. In order to effectively prevent and control R. anatipestifer infection, researchers have conducted in-depth research on the molecular mechanism of R. anatipestifer infection in recent years. A variety of virulence factors of R. anatipestifer have been found, which are related to outer membrane protein, lipopolysaccharide synthesis, iron metabolism, biofilm formation, biotin synthesis, and so on (6–16). In a previous study, we demonstrated that the R. anatipestifer AS87_RS09170 gene encodes BioF, a protein associated with bacterial biotin synthesis, as well as bacterial growth, morphology, and virulence (16).

Biotin (vitamin B7 or H) is associated with metabolic pathways involving membrane lipid biosynthesis, amino acids, and gluconeogenesis (17–20). Microorganisms obtain biotin in two ways to meet their physiological needs: one is de novo biosynthesis, and the other is a scavenging route (21–25). In human and mammals, de novo biosynthesis is not possible, and the scavenging route is the only way to obtain biotin (20, 21). Biotin is generated from pimelate precursors by sequential catalyzation of four conserved enzymes: BioF (7-keto-8-aminopelargonic acid synthase), BioA (7,8-diaminopelargonic acid synthase), BioD (dethiobiotin synthase), and BioB (biotin synthase) (26). Several studies suggest that actithiazic acid and amiclenomycin isolated from Streptomyces species are potential antibiotics for the treatment of Mycobacterium resistance to isoniazid and rifampin, etc., because they irreversibly inactivate the target enzyme BioA (27–29). These results suggest that biotin biosynthesis is a potential target for the development of new antibiotics. With bioF gene deletion in a previous study, the R. anatipestifer mutant strain Yb2ΔbioF could not grow in tryptic soy broth (TSB) medium in the absence of biotin and its virulence was substantially attenuated (16), suggesting that biotin is important for bacterial growth and virulence.

Biotin biosynthesis de novo is an energy-consuming process in which 20 ATP equivalents are consumed to synthesize one biotin molecule (30). The biosynthesis of biotin is controlled by diverse mechanisms. To date, several different ways (BirA, BirA/BioR, and BirA/BioQ) have been identified that regulate bacterial biotin metabolism (31–33). In Escherichia coli, the biotin ligase protein BirA is a bifunctional protein, working as an enzyme or as a transcriptional repressor, according to the intracellular concentration of biotin and the levels of biotin-dependent enzymes requiring biotin (34). In Agrobacterium tumefaciens, BirA has only catalytic and not regulatory activity because it lacks the N-terminal DNA binding domain (35). Rodionov et al. (41) found that BioR, a new type of GntR family transcriptional factor, can repress the expression of the bioBFDAZ operon. In 2012, a TetR-like transcription factor named BioQ was discovered that was associated with the biotin biosynthesis regulatory pathway of Corynebacterium glutamicum (36). In addition, biotin carboxyl carrier protein (AccB) may regulate biotin synthesis by forming a complex with AccC in Ralstonia (19). Therefore, biotin protein ligase (BPL) is classified into two groups distinguished by the presence of an N-terminal winged helix-turn-helix DNA binding domain (37). In this study, the regulatory mechanisms of R. anatipestifer biotin biosynthesis were investigated. BirA of R. anatipestifer was identified as a group I BPL that catalyzed covalent attachment of biotin to acceptor protein AccB but not bind to the biotin-biosynthetic operon. AccB belongs to acetyl coenzyme A (acetyl-CoA) carboxylase (biotin-dependent enzyme) components. BirA catalyzes biotin covalent linkage to the AccB to transfer carbon dioxide between metabolites in carboxylation, decarboxylation, and transcarboxylation reactions (38, 39). BioX of R. anatipestifer was identified as a member of the helix-turn-helix xenobiotic response element family of transcription factors (HIT-XRE) and regulated the expression of genes associated with biotin biosynthesis. Electrophoretic mobility shift assays (EMSA) confirmed that BioX bound to one promoter sequence of the biotin-biosynthetic operon. Quantitative PCR demonstrated that bioX repressed the expression of the genes associated with biotin biosynthesis. Furthermore, the bioX gene of R. anatipestifer was identified as a virulence factor. Collectively, these findings revealed that BioX in R. anatipestifer was a novel negative regulator of biotin metabolism and therefore was also a novel virulence factor for bacterial infection. These results will contribute to further study of the design of antibiotics as well as of bacterial metabolism and pathogenesis.

RESULTS

Riemerella anatipestifer putative BirA protein is a group I biotin protein ligase.

Multiple-sequence alignment showed that the R. anatipestifer putative BirA protein lacked the N-terminal helix-turn-helix domain of E. coli BirA but shared similarity with Brucella abortus BirA (Fig. 1A). BirA with an N-terminal DNA binding domain (group II) is widely distributed in eubacteria and archaea (35), suggesting that group II BirA regulation of biotin biosynthesis is ancient. Bacteria with a BPL that lacks the N-terminal DNA binding domain (group I) evolved new regulation of biotin biosynthesis (30, 31). To further identify the protein, the His-tagged recombinant protein was generated, purified, and subjected to SDS-PAGE, and a band appeared at ∼33 kDa, which is consistent with the estimated size (Fig. 1B). Furthermore, liquid chromatography-mass spectrometry (MS) analysis demonstrated that the ∼33-kDa protein matched the expected sequence of the R. anatipestifer BirA protein from the National Center for Biotechnology Information (NCBI), with 63.18% coverage (Fig. 1C). The above-described data demonstrated that the R. anatipestifer putative BirA protein was a group I BPL.

FIG 1.

Identification and expression of Riemerella anatipestifer BirA protein. (A) Alignment of the biotin protein ligase. Identical residues are in white letters with a red background, similar residues are in black letters with a yellow background, variable residues are in black letters, and dots represent gaps. (B) SDS-PAGE (12.5%) profile of the purified recombinant BirA protein from R. anatipestifer. (C) MS identification of R. anatipestifer recombinant BirA. The matched amino acid residues are in bold (63.18%).

Riemerella anatipestifer BirA catalyzes biotinylation of acetyl-CoA carboxylase biotin carboxyl carrier protein.

Acetyl-CoA carboxylase biotin carboxyl carrier protein (AccB) is the typical biotin-dependent enzyme that is covalently attached to the biotin. Sequence alignment between the 87-amino-acid biotinylated domain of AccB (AccB87) in E. coli strain MG1655 and R. anatipestifer strain Yb2 indicated that AccB in R. anatipestifer carried a conserved biotinylation site of lysine (K123) (Fig. 2A). To determine the function of the R. anatipestifer putative BirA protein, the open reading frame of protein AccB was expressed. After purification, the protein band appeared at ∼35 kDa in the SDS-PAGE analysis (Fig. 2B). Enzymatic reaction (BirA-AccB) was used to demonstrate the activity of the BPL in vitro (40), and streptavidin-based Western blotting was performed to detect the biotinylation of AccB. As expected, a biotinylated band was observed (Fig. 2C, lane 3). The band in lane 4 should be caused by carryover of biotin during protein purification from E. coli. Given that proteins AccB and BirA were similar in size (∼35 kDa), it was difficult to determine which protein was biotinylated. To solve this problem, the C-terminal 84-amino-acid domain (AccB84) from R. anatipestifer strain Yb2 was expressed in E. coli and purified to demonstrate that AccB was biotinylated by BirA. SDS-PAGE analysis showed an ∼17-kDa band of AccB84 which is distinguishable with ∼35 kDa of BirA (Fig. 2D). Then, the BirA-catalyzed reaction (BirA-AccB84) was determined by streptavidin-based Western blotting. The results showed that the biotinylation of the AccB84 protein continued to increase with the increase of AccB84 concentrations, ranging from 0.5 to 50 μM. AccB84 was not biotinylated without BirA (lane 6), suggesting that AccB was biotinylated by BirA (Fig. 2E). In addition, neutral hydroxylamine was added to the enzymatic reaction to remove biotinoyl-5′-adenylate (bio-5′-AMP). As shown in Fig. 2E, AccB84 was significantly decreased in the biotinylated band with the addition of neutral hydroxylamine in lane 7, compared to that in lane 5. Collectively, the results demonstrated that R. anatipestifer BirA biotinylated the AccB substrate protein via the biotinoyl-5′-AMP intermediate.

FIG 2.

Riemerella anatipestifer BirA catalyzes AccB biotinylation in vitro. (A) Alignment of the C-terminal 87-residue-long fragment of biotin carboxyl carrier protein AccB from R. anatipestifer and Escherichia coli. The arrow indicates the biotinylation site of AccB. Identical residues are in white letters with a red background, similar residues are in black letters with a yellow background, and variable residues are in black letters. (B) SDS-PAGE (12.5%) profile of the purified AccB of R. anatipestifer. (C) (Top) Immunoblot (WB) analysis of the AccB biotinylation by the recombinant BirA protein in vitro, using peroxidase-conjugated streptavidin detection of biotin; (bottom) SDS-PAGE profile of the AccB biotinylated by the recombinant BirA in vitro, as a loading control for the WB. (D) SDS-PAGE (12.5%) profile of the purified C-terminal polypeptide of R. anatipestifer AccB84. (E) (Top) immunoblot analysis of the AccB84 biotinylation by the recombinant BirA in vitro, using peroxidase-conjugated streptavidin detection of biotin; (bottom) SDS-PAGE profile of the AccB84 biotinylated by the recombinant BirA in vitro, as a loading control for the WB.

Riemerella anatipestifer biotin biosynthesis gene cluster constitutes an operon.

Figure 3A shows the genetic organization of the putative bioX and biotin biosynthesis-associated genes. To validate that the bioFDA gene cluster is a transcriptional operon, with the promoter located upstream of bioF, the transcriptional start sites of the bioF and bioB genes were identified by 5′ rapid amplification of cDNA ends (RACE) (Fig. 3B and C). The 5′ end of the R. anatipestifer bioF and bioB transcripts were 18 and 24 nucleotides, respectively, upstream from the translation initiation codon (Fig. 3B and C). To determine whether the bioFDA gene was an operon, both PCR and reverse transcription-PCR (RT-PCR) assays were performed using multiple sets of primers (see Table S1 in the supplemental material). First, the intragenic and intergenic regions were validated by PCR. The intragenic amplifications (1 to 6) and the intergenic amplifications (7 to 9) were positive in PCR and RT-PCR assays, verifying that all nine genes had detectable transcription (Fig. 3D). Therefore, bioF, bioD, and bioA were in an operon.

FIG 3.

Transcriptional analyses of the biotin biosynthesis-associated genes. (A) Genetic organization of the bioX gene and biotin biosynthesis-associated genes. The arrows indicate the gene loci. The numbered lines (1, 2, 3, 4, 5, 6, 7, 8, and 9) represent the specific PCR amplicons that were observed in the following PCR and RT-PCR assays. (B) 5′ rapid amplification of cDNA ends for the bioF transcriptional start site. The predicted binding region is underlined. “S” indicates the putative transcriptional start site, and “ATG” is the translation initiation site. (C) 5′ rapid amplification of cDNA ends for the bioB transcriptional start site. (D) PCR and RT-PCR analyses of the putative biotin biosynthesis-related loci. The primer numbering is identical to that in panel A.

BioX but not BirA protein binds to the biotin-biosynthetic promoter.

R. anatipestifer BirA was identified as a group I BPL; thus, we predicted that the BirA could not bind to DNA as a transcription repressor. The lack of binding in the EMSA supports the conclusion that BirA does not bind the bioF or bioB promoter (Fig. 4A). The R. anatipestifer gene AS87_RS09325 encoded a protein that contains the DNA-binding motif and belongs to the HIT-XRE family, according to sequence analysis. To verify the protein, the His-tagged recombinant protein was generated and subjected to MS analysis. As a result, a 58.42% coverage of the protein with the R. anatipestifer BioX sequence from the NCBI was indicated (Fig. 4B and C). A conserved 10-bp sequence (TTATXXAXAX) was the same as in the binding site in A. tumefaciens (41). The bioF promoters in eight strains of R. anatipestifer were compared, and all had the predicted binding site (Fig. 4D). To determine whether BioX was a direct regulator of the biotin-biosynthetic operon, an EMSA was performed using a 200-bp probe to test the ability of BioX to bind to the bioF, bioB, or birA promoter. BioX specifically interacted with the promoter of the bioF probe but not with the promoter of the bioB probe (Fig. 4E) or with the promoter of the birA probe (Fig. S1), suggesting that BioX efficiently bound to the bioF promoter.

FIG 4.

Binding of BioX to the promoter regions of putative bioFDA operon. (A) BirA failed to bind to the bioFDA operon. A binding band of BirA with the bioF or bioB probes was not observed. (B) SDS-PAGE (12.5%) profile of the purified recombinant BioX from R. anatipestifer. (C) MS identification of R. anatipestifer BioX. The matched amino acid residues are in bold (58.42%). (D) Multiple-sequence alignment of the predicted binding sites from eight collections of R. anatipestifer and the resulting sequence logo. (Top) Identical residues are in white letters with a red background, and varied residues are in black letters; (bottom) the sequence logo was generated using WebLogo (http://weblogo.berkeley.edu/logo.cgi). (E) BioX protein binds to the bioF but not the bioB promoter.

Biotin is a possible BioX regulatory ligand.

As first reported for E. coli (34), BirA binds to biotin followed by ATP to produce bio-5′-AMP. Then bio-5′-AMP itself binds the biotin-biosynthetic operon. Therefore, BioX was predicted to require bio-5′-AMP in order to bind to its operon and inhibit the biosynthesis of biotin. To test this prediction, an EMSA was performed without the addition of biotin or ATP. BioX bound to the bioF probe, indicating that the BioX had possibly acquired its ligand from the E. coli host (Fig. 5A) (32). To remove the ligand from the protein, an incubation was performed with neutral hydroxylamine to cleave bio-5′-AMP. Following this treatment, BioX did not bind to the bioF probe in the presence or absence of biotin (Fig. 5B). To further confirm whether BioX was a biotin receptor protein, the biotinylation of BioX in the BirA-catalyzed reaction (BirA-BioX) was assessed by determination of the peroxidase-conjugated streptavidin binding to biotin in vitro. The results showed a biotinylated BioX band. A weaker band in lane 6 in the SDS-PAGE gel is not the BirA band (molecule weight is a little different from that of BirA) and should be the protein contamination during purification (Fig. 4B). To conclude, BioX could accept biotin as a ligand (Fig. 5C).

FIG 5.

Biotin promotes BioX-DNA binding. (A) Electrophoretic mobility shift assay-based analyses of the effects of biotin and ATP on BioX binding to its operator. Gel shift assays were conducted using 6% native PAGE. (B) BioX failed to bind to the bioFDA operon with hydroxylamine treatment. (C) (Top) Immunoblot analysis of the BioX biotinylated protein by recombinant BirA in vitro, using peroxidase-conjugated streptavidin detection of biotin; (bottom) SDS-PAGE profile of the BioX protein biotinylated by recombinant BirA in vitro, as a loading control for the WB.

BioX acts as a repressor.

To test whether BioX acted as a repressor of the bioFDA operon, a mutant strain, Yb2ΔbioX, and the complementation strain cYb2ΔbioX were constructed (Fig. S2). qPCR confirmed that the transcription of the bioX gene was disrupted, but the transcriptions of the upstream AS87_RS09320 and downstream AS87_RS09330 genes were not affected in the mutant strain Yb2ΔbioX (Fig. S3). Then, qPCR was used to detect the expression of the bioF, bioD, bioA, bioB, and birA genes in the wild-type strain Yb2, the mutant strain Yb2ΔbioX, and the complementation strain cYb2ΔbioX. With the increase in biotin concentration in TSB medium, the levels of expression of the bioFDA operon in the wild-type strain Yb2 and complementation strain cYb2ΔbioX were decreased; however, the opposite was observed in the mutant strain Yb2ΔbioX (Fig. 6A to C). Similarly, the expression levels of biotin biosynthesis-associated genes in the mutant strain Yb2ΔbioX were higher than those in the wild-type strain Yb2 and the complementation strain cYb2ΔbioX in TSB cultures (Fig. 6D). These results showed that the regulation of biotin biosynthesis weakened in the strain lacking BioX, which is consistent with previous findings that biotin biosynthesis has a strict regulatory mechanism (31, 42).

FIG 6.

Riemerella anatipestifer BioX functions as a repressor of expression of biotin biosynthesis-associated genes. (A) qRT-PCR data of R. anatipestifer wild-type strain Yb2 grown in tryptic soy broth (TSB) medium with the addition of 0, 5 nM, or 100 nM biotin. (B) qRT-PCR data of R. anatipestifer mutant strain Yb2ΔbioX grown in TSB medium with the addition of 0, 5 nM, or 100 nM biotin. (C) qRT-PCR data of R. anatipestifer complementation strain cYb2ΔbioX grown in TSB medium with the addition of 0, 5 nM, or 100 nM biotin. (D) qRT-PCR data of R. anatipestifer wild-type strain Yb2, the mutant strain Yb2ΔbioX, and complementation strain cYb2ΔbioX in TSB medium. The expression of the l-lactate dehydrogenase-encoding gene (ldh) served as an endogenous control. All experiments were carried out in triplicate. The changes of mRNA were expressed as fold expression and calculated using the comparative cycle threshold (2^−ΔΔCT) method. Data are presented as means from 3 independent experiments. Error bars correspond to the SD of the means (n = 3). Significance was analyzed by t test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. ns, no significance.

BioX is required for Riemerella anatipestifer virulence in ducks.

To investigate whether the deletion of the bioX gene affected bacterial virulence, 14-day-old Cherry Valley ducks were injected intramuscularly with the wild-type strain Yb2 or the mutant strain Yb2ΔbioX. The 50% lethal dose (LD₅₀) was determined as previously described (7). The LD₅₀ for the mutant strain Yb2ΔbioX was 3 × 10⁹ CFU, which represented a 500-fold attenuation in virulence compared with that of the wild-type strain Yb2, with an LD₅₀ of 6 × 10⁶ CFU (Table 1). To further investigate the role of the bioX gene in systemic infection in vivo, the bacterial burden in the blood of infected ducks was quantified. Twelve 3-week-old Cherry Valley ducks were randomly divided into two groups of six, which were then infected with the wild-type strain Yb2 or the mutant strain Yb2ΔbioX at 1 × 10⁸ CFU per duck. At 6, 12, 24, and 36 h postinfection (hpi), the bacterial burden was evaluated by counting the CFU in blood. The bacterial recoveries in the Yb2ΔbioX mutant strain were 2.1 × 10⁴ CFU/ml at 6 hpi, 4.7 × 10³ CFU/ml at 12 hpi, 4 × 10³ CFU/ml at 24 hpi, and 2 × 10³ CFU/ml at 36 hpi. In contrast, the bacterial recoveries in the wild-type strain Yb2 were 2.5 × 10⁴ CFU/ml at 6 hpi, 7.5 × 10⁴ CFU/ml at 12 hpi, 6.8 × 10⁵ CFU/ml at 24 hpi, and 3.7 × 10⁵ CFU/ml at 36 hpi (Fig. 7). There was no significant difference in the bacterial recovery between the mutant strain Yb2ΔbioX and the wild-type strain Yb2 at 6 hpi. As the time of infection increased, bacterial recovery in the wild-type strain Yb2 increased, and a systemic infection was established. In contrast, with the increase in infection time, bacterial recovery in the mutant strain Yb2ΔbioX remained at levels similar to that at 6 hpi, and the levels much lower than those of the wild-type strain Yb2 at 12, 24, and 36 hpi. Collectively, the results indicated that the bioX gene plays an important role in R. anatipestifer virulence.

TABLE 1.

Bacterial LD₅₀ determination

Strain	LD50 (CFU)
Yb2	6 × 106
Yb2ΔbioX	3 × 109

FIG 7.

Blood bacterial loadings in Riemerella anatipestifer-infected ducks. Six ducks were injected intramuscularly with 1 × 10⁸ CFU of each bacterial strain. Blood samples were collected at 6, 12, 24, and 36 h postinfection, and the bacterial CFU were counted. Data are presented as the means ± SD for six infected ducks. The data were analyzed using t tests.

DISCUSSION

Biotin acts as a cofactor in metabolic processes (38). When biotin biosynthesis is limited, bacterial growth and viability are impaired in vitro, and Mycobacterium tuberculosis, Francisella, and R. anatipestifer infection is inhibited in animals (16, 53, 54). BirA is part of a classic biotin sensing system and is a transcriptional regulator with a conserved binding signal in eubacteria and archaea (35). However, the mechanism regulating biotin biosynthesis in R. anatipestifer remains unknown. In this study, two R. anatipestifer genes associated with biotin biosynthesis were identified. AS87_RS05840 encoded a BirA protein lacking the N-terminal winged helix-turn-helix DNA binding domain, and AS87_RS09325 encoded a BioX protein, which was in the helix-turn-helix xenobiotic response element family of transcription factors. R. anatipestifer putative BirA protein shared similarity with Brucella abortus BirA by BLAST analysis, suggesting that BirA was a group I BPL (Fig. 1A). R. anatipestifer putative BirA catalyzed biotinylation of acetyl-CoA carboxylase AccB. Moreover, a novel regulatory mechanism was defined that used the XRE family factor BioX as a repressor (Fig. 8). Furthermore, the bioX gene was identified as a virulence factor, and its deletion led to a 500-fold decrease in virulence, demonstrating that bioX contributed to establishing infection in duck.

FIG 8.

Schematic diagram of the BioX regulation of biotin biosynthesis in R. anatipestifer. (Top) BirA catalyzes biotin addition to a biotin-dependent carboxylase AccB but does not bind to the bioF promoter to repress transcription initiation at the biotin biosynthetic operon. (Bottom) With a biotin excess, BioX is biotinylated and binds to the bioF promoter to repress biotin biosynthesis. Green dots, biotin; orange triangles, ATP; yellow ovals, BirA; green dots and orange triangles with yellow ovals, BirA–bio-5′-AMP; irregular purple circle, AccB; green dots with irregular purple circles, biotinylated AccB; red ovals, BioX; green dots and orange triangles with red ovals, biotinylated BioX; orange arrow, banding together; green arrow, formation of complex; pink dotted arrow, transport biotin; black cross, no binding to bioF promoter; black arrows, transcripts.

Bacteria have evolved several strategies to regulate biotin biosynthesis in addition to BirA, including BioR and BioQ, which are classified as members of the GntR type and TetR-type transcription factor families, respectively. Genome annotation indicates that TetR transcription factors (AKQ39780.1 and AKQ40092.1), but not GntR transcription factors, were found in the R. anatipestifer genome. BLAST analysis showed that AKQ39780.1 and AKQ40092.1 in R. anatipestifer share high homology with their counterparts in Cloacibacterium and Chryseobacterium. Identification of the AKQ39780.1 conserved domain shows that a regulator of FtsZ, which is a DNA-binding protein in the family of TetR/AcrR transcriptional regulators, participates in septum placement and in chromosome capture during the asymmetrical cell division in endospore formation. The five nearly palindromic DNA motifs (RBMs) to which RefZ binds affect chromosomal localization, but do not affect transcription, so RefZ is not considered a transcription factor. Identification of the AKQ40092.1 conserved domain showed a DNA-binding transcriptional regulator in TetR/AcrR family. We also aligned the sequences of AKQ39780.1 and AKQ40092.1 with the BioQ of Mycobacterium, which belongs to TetR family transcription factor and regulates biotin synthesis. It shows a low value of homology. Therefore, we predict that the R. anatipestifer genome encodes TetR family members, but the function remains to be identified. To identify the transcriptional factors that regulate biotin biosynthesis in R. anatipestifer Yb2, a helix-turn-helix DNA-binding transcriptional regulatory gene, AS87_RS09325, was selected for further investigation, according to our comparative transcriptome sequencing data between the wild-type strain Yb2 and its mutant strain Yb2ΔbioF. The AS87_RS09325 gene encodes a putative transcriptional regulator BioX of the XRE family, which is associated with oxidant tolerance and virulence, DNA damage, and antibiotics (43–45). In this study, the recombinant BioX protein was prepared for an EMSA, and the result showed that BioX specifically bound to the bioF promoter (Fig. 4D). Then, an attempt to determine the exact binding site of BioX on the biotin biosynthesis operon by DNase I footprinting assays failed, suggesting that BioX bound to its operator weakly in vitro or interacted with other repressors to regulate the transcription of biotin. To determine whether BioX bound directly to the biotin operator or it formed a biotinylated protein to bind, an EMSA was performed with the addition of neutral hydroxylamine in the reaction buffer. The results showed that BioX was unable to bind bioF promoter when bio-5′-AMP was removed by the addition of neutral hydroxylamine into the reaction system, suggesting that biotin is the ligand of BioX. Notably, the streptavidin-based Western blot analysis also showed that BioX was a receptor of biotin. The metabolic pathways with BioX involvement after biotinylation in vivo will be explored in the future.

To better evaluate the role of the bioX gene in R. anatipestifer, a mutant strain, Yb2ΔbioX, was constructed in which the 205-bp fragment of the bioX gene that corresponds to the C-terminal region that functions in binding to DNA was replaced with an erythromycin resistance cassette. The complementation strain cYb2ΔbioX was constructed to confirm the bioX function. Growth curves in TSB medium and biotin-reduced TSB medium indicated that all three bacterial strains grew more slowly in the biotin-reduced TSB medium than in TSB medium; no significant difference was shown in the bacterial growth among the wild-type strain Yb2, the mutant strain Yb2ΔbioX, and the complementation strain cYb2ΔbioX (see Fig. S4 in the supplemental material), suggesting that biotin plays a role in the bacterial growth. Figure 6A and C show that expression of the biotin synthesis-associated genes in the wild-type strain Yb2 and complementation strain cYb2ΔbioX decreased with addition of extra biotin in the TSB medium, suggesting that biotin synthesis was regulated for saving the energy of oversynthesis. The levels of expression of the bioF, bioA, and bioB in the mutant strain Yb2ΔbioX were all increased with addition of extra biotin in the TSB medium (Fig. 6B), which could be explained as follows: addition of exogenous biotin increased bacterial growth, which accordingly increased the demand for biotin metabolism. The mutant strain Yb2ΔbioX lacks the negative regulator BioX; thus, the expression of biotin synthesis-related genes increased without suppression by BioX. These data further confirmed that BioX regulates the expression of biotin synthesis-associated genes. The expression data of these genes in Yb2, Yb2ΔbioX, and cYb2ΔbioX in TSB culture further confirmed that bioX functions in the regulation of biotin synthesis-associated genes.

R. anatipestifer infection causes septicemia, which is associated with the colonization and development of the bacterium in a host (46). Our previous study report that deleting the R. anatipestifer bioF gene caused a 768,000-fold attenuation of the bacterial virulence, indicating that biotin is very important for bacterial virulence and that abolishing biotin biosynthesis attenuates virulence (16, 55). In this study, deleting the R. anatipestifer bioX gene caused a 500-fold attenuation of the bacterial virulence, with lower levels of virulence attenuation for the bioX mutant than the bioF mutant, which could be explained by the bioX gene functioning in the regulation of biotin synthesis; when the biotin in the growth environment meets the need for the bacterial metabolic activity, loss of bioX regulatory function will not kill the bacteria but only waste the energy for synthesis of biotin exceeding the demand, thus attenuating the virulence. Systemic infection of the ducks revealed that the bacterial CFU in the blood of ducks infected with mutant strain Yb2ΔbioX did not increase significantly, in contrast to the number in wild-type strain Yb2, suggesting that R. anatipestifer bioX regulation of biotin synthesis might be particularly important in establishing systemic infections by bacteria; further study is needed to determine the exact step in which bioX causes widespread inflammation in a host. In conclusion, R. anatipestifer BioX is a novel regulator for biotin synthesis and an important factor contributing to the establishment of systemic infections.

MATERIALS AND METHODS

Bacterial strains, plasmids, and culture conditions.

The bacterial strains used in this study are listed in Table 2. Riemerella anatipestifer strain Yb2 was the wild-type strain used in this study. All R. anatipestifer strains were grown on tryptic soy agar (TSA; Difco, USA) containing 1.5% agar at 37°C with 5% CO₂ or tryptic soy broth (TSB, Difco) with addition of biotin at various concentrations. Escherichia coli strains were grown on Luria-Bertani (LB) plates or in LB broth. Antibiotics were used at the following concentrations: ampicillin, 100 μg/ml; erythromycin, 0.5 μg/ml; kanamycin, 50 μg/ml; and chloramphenicol, 30 μg/ml.

TABLE 2.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic	Origin
Bacterial strains
DH5α	Cloning host	Lab stock
BL21(DE3)	Protein expression host	Lab stock
Yb2	Riemerella anatipestifer serotype 2 strain	Lab stock
Yb2ΔbioX	bioX deletion mutant of R. anatipestifer Yb2	This work
cYb2ΔbioX	bioX complementation strain of Yb2ΔbioX	This work
Plasmids
pET30(a)	T7 promoter Kanr expression vector	Novagen
pET30-birA	pET30a containing R. anatipestifer birA	This work
pET30-accB	pET30a containing R. anatipestifer accB	This work
pET30-accB84	pET30a containing R. anatipestifer accB84	This work
pET30-bioX	pET30a containing R. anatipestifer bioX	This work

Bioinformatics analyses.

The sequences of BirA from E. coli, Brucella abortus, and R. anatipestifer Yb2, as well as those of AccB87 from E. coli and R. anatipestifer Yb2, were subjected to bioinformatics analyses. Sequence alignments were performed using Clustal Omega (http://www.ebi.ac.uk/Tools/msa/clustalo/), and the output was processed by ESPript 3.0 (http://espript.ibcp.fr/ESPript/cgi-bin/ESPript.cgi) to generate the final Fig. 1A, as described previously (47). The DNA binding sites were from the computational predictions of Rodionov and Gelfand (41).

Expression, purification, and identification of the putative BirA, AccB, AccB84, and BioX proteins.

The open reading frames of birA, accB, and bioX, as well as the gene sequence encoding the C-terminal 84 residues of AccB (AccB84), were amplified from the wild-type strain Yb2 by using the primer pairs 30a_birA-F1/30a_birA-R1 (birA), 30a_accB-F1/30a_accB-R1 (accB), 30a_bioX-F1/30a_bioX-R1 (bioX), and 30a_accB84-F1/30a_ accB84-R1 (AccB84). The amplicons were then digested with XhoI and BamHI and cloned into pET30a vectors to generate the plasmids pET30a-birA, pET30a-accB, pET30a-bioX, and pET30a-accB84 (see Table S1 in the supplemental material). E. coli strains DH5α and BL21(DE3) were subsequently used for transformation and expression. The insertion of the plasmids was confirmed by DNA sequencing. The proteins were expressed in LB medium containing 50 μg/ml of kanamycin with induction at an optical density at 600 nm (OD₆₀₀) of 0.8 with the addition of 1.0 mM isopropyl-β-d-thiogalactopyranoside (IPTG) at 16°C for 16 h. The bacterial cells were pelleted by centrifugation (4,000 × g, 8 min), washed twice with ice-cold phosphate-buffered saline (PBS; 101.4 mM Na₂HPO₄, 1.8 mM KH₂PO₄, 137 mM NaCl, 2.7 mM KCl, 8% glycerol [pH 7.4]), and dissolved in binding buffer. Protein purification was performed using a HisTag high-performance column (Beaver, Suzhou, China) according to the manufacturer’s guidelines. The concentrations of the purified proteins were assessed using Pierce bicinchoninic acid (BCA) protein assay reagent (Beyotime, Shanghai, China).

Assays of in vitro BirA biotinylation activity.

The reaction of in vitro BirA-catalyzed biotinylation was performed as previously reported, with minor modifications (32). Briefly, the reaction mixture contained 50 mM Tris-HCl (pH 8.0), 5 μM ATP, 10 mM MgCl₂, 100 nM biotin, 100 mM KCl, 5 mM Tris(2-carboxyethyl)phosphine, 1.5 μM BirA, and various concentrations of the AccB84 protein in a final volume of 50 μl. If necessary, 0.2 M neutral hydroxylamine was added to remove bio-AMP. Each of the reaction mixtures was incubated at 37°C for 16 h. Then, the mixtures were boiled at 100°C for 10 min. The products were separated by SDS-PAGE and then transferred onto nitrocellulose membranes (Sigma-Aldrich, Shanghai, China). Western blot analysis was performed using a streptavidin-peroxidase polymer (1:1,000 dilution; Sigma-Aldrich) (16). Binding was visualized using a chemiluminescent imaging system (Tanon, Shanghai, China).

Electrophoretic mobility shift assays.

The putative DNA binding site upstream of bioF and bioB was PCR amplified from R. anatipestifer strain Yb2 with primer pairs bioF-pro-F1/bioF-pro-R1 and bioB-pro-F1/bioB-pro-R1, respectively. The DNA fragment was 200 bp in length. Negative-control DNA (200 bp) was amplified from R. anatipestifer Yb2 genomic DNA with primers bioF-pro-F2 and bioF-pro-R2. The PCR products were sized on a 2% agarose gel and purified using a Sangon Biotech (Shanghai, China) PCR purification kit. The DNA concentrations were determined at OD₂₆₀ by using a Nano-300 (Allsheng, Hangzhou, China). The DNA binding reaction mixture contained 50 mM Tris-HCl (pH 8.0), 1 mM MgCl₂, 50 mM KCl, 2.5% glycerol, 40 nM DNA, the desired concentrations of BirA or BioX, 1 mM ATP, and 5 nM biotin (48). The binding reaction mixtures were incubated at room temperature for 30 min and then loaded onto a 6% DNA retardation gel. The gel was run in 0.5× Tris-borate-EDTA (TBE) buffer at 100 V for 1 h and 25 min. The gel was visualized using an Odyssey infrared imaging system (Li-COR, NE, USA).

5′ rapid amplification of cDNA ends.

To determine the transcriptional start site, 5′ RACE was performed using a SMARTer RACE 5′ kit (Vazyme, Nanjing, China). Total RNA was extracted from the wild-type strain Yb2 using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. A total of 3 μg of RNA was treated with gene-specific primers and deoxynucleoside triphosphate (dNTP) mix. The treated RNA samples were used for cDNA biosynthesis using random hexameric oligonucleotide primers. Nested PCRs were applied using the 5′ RACE outer primers plus gene-specific outer primers and the 5′ RACE inner primers plus gene-specific inner primers (Table S1). The pMD19-T Simple vector (TaKaRa Biotechnology, Tokyo, Japan) was used for further sequencing of the PCR products. The first nucleotide adjacent to the 5′ RACE adaptor was assumed to be the transcriptional start site of a gene.

Construction of bioX deletion mutant strain Yb2ΔbioX and complementation strain cYb2ΔbioX.

The mutant strain was created by allelic exchange with a nonpolar erythromycin (Erm) gene cassette, which was amplified from plasmid pcp29 using primer pairs ErmF1/ErmR1 (Table S1) as described previously (49). Briefly, a two-step PCR procedure was used to produce a PCR product in which the Erm gene cassette was flanked by arms of approximately 500 to 1,000 bp, corresponding to sequences upstream from the start codon and downstream from the stop codon of the bioX gene. The three fragments were joined by overlap PCR using the primers bioX-up-F1 and bioX-down-R1. Then, 1.0 μg of purified DNA product was added to the R. anatipestifer wild-type Yb2 culture in logarithmic growth in TSB for further growth at 37°C and 220 rpm for 1 h. The bacterial cultures were collected and plated onto TSB agar plates with Erm to screen the bioX deletion mutant strain Yb2ΔbioX, which was further identified using PCR analysis. Complementation strain cYb2ΔbioX was constructed by using plasmid pCP29 (50). Briefly, the bioX gene was amplified from wild-type strain Yb2 using primers 9325 comp-F/9325comp-R (Table S1). The PCR product was inserted into pCP29 at XhoI and SphI restriction sites, resulting in plasmid pCP29-bioX. Plasmids were first introduced into E. coli S17-1 by transformation. Next, they were transferred into mutant strain Yb2ΔbioX by conjugation. Transformants were selected on TSA containing 5 mg/ml of cefoxitin and 50 mg/ml of kanamycin and identified by PCR amplification using primers 9325 comp-F/9325comp-R. The complementation strain was named cYb2Δbio (Fig. S1B).

RNA isolation and RT-PCR.

Cells were grown overnight in TSB or TSB with the addition of biotin. Total RNA was isolated from the mid-logarithmic phase using TRIzol reagent (Invitrogen, Carlsbad, CA) as described previously (7). Quantitative PCR was performed using gene-specific primers to confirm transcriptional levels of the bioFDA operon of R. anatipestifer in response to varied biotin concentrations in the growth environment (Table S1). Data analysis was conducted according to the cycle threshold (2^−ΔΔCT) method via normalization to the expression of the reference gene encoding l-lactate dehydrogenase (ldh) (33, 51).

Determination of bacterial virulence.

To determine whether deletion of the bioX gene affected the virulence of R. anatipestifer, the bacterial LD₅₀ and loadings in the blood of infected ducks were determined (46). Briefly, to determine the LD₅₀, a total of 60 14-day-old Cherry Valley ducks were randomly divided into 10 groups of 6 ducks. Ducks in groups 1 to 5 were inoculated intramuscularly with 10³, 10⁴, 10⁵, 10⁶, or 10⁷ CFU of Yb2 bacteria, and ducks in groups 6 to 10 were injected with 10⁶, 10⁷, 10⁸, 10⁹, or 10¹⁰ CFU of Yb2ΔbioX bacteria. The number of deaths was recorded for 7 days postinfection. Bacterial LD₅₀ values were calculated using the modified Karber method (52).

The bacterial loadings in the blood of infected ducks were measured as described previously, with modifications (7). Briefly, 12 3-week-old Cherry Valley ducks were divided into two groups of 6 ducks. The ducks were infected with the wild-type strain Yb2 or mutant strain Yb2ΔbioX at the dose of 10⁸ CFU/duck. Blood samples were collected at 12, 24, and 36 hpi, diluted appropriately, and plated on TSA to count bacteria.

Statistical analyses.

Statistical analyses were conducted using GraphPad (La Jolla, CA) Prism software v 6.0. A two-tailed independent Student t test was used to analyze bacterial loads in blood. Statistical significance was established at a P value of <0.05.

Ethics statement.

One-day-old Cherry Valley ducks were obtained from the Zhuang Hang Duck Farm (Shanghai, China) and raised at a controlled temperature (28°C to 30°C). The ducks were housed in cages with free access to food and water to maintain conditions of biological safety. Animal experiments were conducted according to the Institutional Animal Care and Use Committee (IACUC) guidelines set by the Shanghai Veterinary Research Institute, Chinese Academy of Agricultural Sciences (CAAS). The animal study protocol was approved by the IACUC of Shanghai Veterinary Research Institute, CAAS, China (permit no. SHVRI-SZ-20200619-01). All surgery was performed under sodium pentobarbital anesthesia, and all efforts were made to minimize suffering.