Ferric Uptake Regulator Provides a New Strategy for Acidophile Adaptation to Acidic Ecosystems

This study advances our understanding of the acid tolerance mechanism of A. caldus, identifies the key fur gene responsible for acid resistance, and elucidates the correlation between fur and acid resistance, thus contributing to an understanding of the ecological behavior of acidophilic bacteria. These findings provide new insights into the acid resistance process in Acidithiobacillus species, thereby promoting the study of the environmental adaptation of acidophilic bacteria and the design of engineered biological systems.

A cidophiles are the main drivers for the formation of acid mine drainage (AMD) in natural habitats. Studies in the last several decades have shown that microorganisms in AMD ecosystems present considerable diversity and a high richness of acidophilic taxa (1,2). Generally, Proteobacteria, Acidithiobacillia, Nitrospira, Firmicutes, Acidobacteria, Actinobacteria, Aquificae, Euryarchaeota, and Crenarchaeota are ubiquitous in AMD ecosystems (3,4). Knowledge of the microorganisms promoting the material and energy flow of AMD ecosystems provides basic clues for AMD bioleaching and bioremediation, such as the oxidation of iron and sulfur by the acidophilic chemolithoautotrophic microbes, accelerating the biogeochemical cycles of elements and the release of metals (1,4,5). In addition, AMDs, similar to other metal-bearing wastes, are a potential secondary source for recovering metals. By using acidophiles in AMD, the biomining industry has developed new technology for extracting metals from ores through microbial oxidation (6). Moreover, acidophile-dependent bioleaching technology has been applied to the treatment of waste containing heavy metals, such as sewage sludge, spent household batteries, mine tailings, and printed circuit boards (7). In AMDs and bioleaching ecosystems, one of the main challenges to these microorganisms is extremely acidic stress (6)(7)(8). Thus, a comprehensive understanding of the acid resistance mechanisms of acidophiles would be helpful for the development of the AMD bioleaching industry and bioremediation.
Acidity is the major determinant of microbial community composition in AMD ecosystems. Although species diversity significantly decreases as the pH decreases, high richness of acidophilic taxa, such as Gammaproteobacteria and Nitrospira, was detected in AMD (9,10). Members of the proteobacterial class Acidithiobacillia are prevalent in AMD and biomining environments because of the ability of these microbes to utilize sulfur and iron and adapt to extremely acidic environments (4). In these environments, Acidithiobacillia can dramatically accelerate the acid generation process; therefore, strains in this class play important roles in the material and energy flow of the AMD ecosystem (4). Moreover, several species of Acidithiobacillia, such as Acidithiobacillus caldus, Acidithiobacillus ferrooxidans, and Acidithiobacillus thiooxidans, have been found to be the most active and dominant bacteria in bioleaching applications (11). Therefore, the acid resistance of acidophiles is important for these microbes to survive and thrive in extremely acidic environments, thus determining their ability to occupy the ecological niche in these systems.
Microorganisms have evolved diverse acid resistance mechanisms and models to prevent cell damage due to acid stress (12)(13)(14)(15), such as the efflux of protons (16), proton consumption (17), adjustment of cell membrane composition (18), DNA and protein repair systems (19), generation of reversed transmembrane electrical potential (Δ) (20), and alkali production (21). However, most acid resistance mechanisms and models of acidophiles are limited to transcriptome, proteome, and bioinformatic analyses (22)(23)(24)(25). Extremely acidophilic bacteria use a wide variety of strategies similar to those of neutralophiles for acid resistance that facilitate defense against extreme acid challenges, including more cation transporters and a large number of DNA and protein repair systems (12). Some global regulators are also reported to be involved in the regulation of cytoplasmic proton homeostasis, such as OmpR, RpoS, and Fur in most bacteria (14,22).
The ferric uptake regulator (Fur), originally identified as a regulator of iron homeostasis in a variety of bacteria, has been reported to be involved in the regulation of the acid stress response in some neutralophiles, such as Salmonella enterica (26), Helicobacter pylori (27), and Campylobacter jejuni (28). Generally, Fur follows the classic regulation pattern in which holo-Fur binds to the Acidithiobacillus caldus ferric uptake regulator (AcFur) box in the promoter to restrict the binding of RNA polymerase and represses gene transcription (29). Moreover, Fur is involved in regulating a wide variety of crucial physiological metabolic pathways, including DNA synthesis (30), the tricarboxylic acid (TCA) cycle (31), biofilm formation (32), and expression of virulence genes, as well as the expression of some microbial genes responsible for the maintenance of iron homeostasis and defense against oxidative stresses (33,34), in most prokaryotes. In general, the maintenance of cytoplasmic iron levels is regulated by Fur, except in a few Gram-positive and acid-resistant bacteria utilizing the diphtheria toxin repressor (DtxR) to regulate iron homeostasis, such as Corynebacterium diphtheriae (35). The functions of Fur in maintaining iron homeostasis and the response to acid resistance in A. ferrooxidans have been investigated (22,36); however, the role of Fur in acid-adapted regulation in Acidithiobacillus spp. and the universality of the Fur-dependent regulation in acidophiles are still unclear.
Here, we analyzed the distribution of Fur in acidophiles and compared the sequence and structural features of Fur proteins in acidophilic bacteria. By using A. caldus (MTH-04) as a model, we investigated the AcFur and its role in adaptation to low pH. Fur-regulated phenotypes and genes of A. caldus were determined. Finally, the exis-tence of a Fur-dependent acid adaptation strategy in A. caldus, as well as the significance of Fur for acid resistance regulation in acidophiles, was proposed.

RESULTS
Universality and conservation of Fur in AMD ecosystems. To investigate the distribution of Fur in AMD bacteria, we first searched for AMD-related microorganisms and their genomes in the National Center for Biotechnology Information (NCBI) database according to previous reports (4,5). We then determined an e value (Ͻ5.00E Ϫ 05) and searched for genome and metagenome sequences available in the NCBI database. Ultimately, the results based on the microorganism species, the assembly level of the microbial genome, the e value, and annotation were used for analysis (see Table S1 in the supplemental material). The results showed that about 89% of the AMD bacterial genomes contained fur genes. Interestingly, we found that the genus Acidiphilium (strictly aerobic; optimal pH range, 3 to 3.5; chemoorganotrophic bacteria) does not contain fur genes-for example, Acidiphilium cryptum JF-5 (genome accession number NC_009484.1), Acidiphilium multivorum AIU301 (genome accession number NC_015186.1), Acidiphilium angustum ATCC 35903 (genome accession number NZ_JNJH00000000.1), and Acidiphilium rubrum (genome accession number NZ_FTNE00000000.1). This discovery implied that harboring Fur is a rule rather than an exception in the bacteria of AMD ecosystems. Interestingly, Fur homologs are rarely found in most archaea of AMD communities, except for Thermoplasma volcanium GSS1 and Cuniculiplasma divulgatum PM4. However, most archaea possess the DtxR protein, which is similar to Fur in both protein structure and potential function (see Table S1) (37).
The prominent role of AcFur in the extreme acid resistance of A. caldus. To illustrate the role of Fur in AMD bacteria, fur deletion (⌬fur) and fur-complemented [⌬fur(fur)] strains of A. caldus were constructed using markerless knockout/knock-in technology. The ⌬fur strain exhibited an obvious growth disadvantage at the stationary phase compared with the wild-type (WT) and ⌬fur(fur) strains of A. caldus (Fig. 1A). The acid shock assay showed that the cell density of the ⌬fur strain was much lower than that of the WT and ⌬fur(fur) strains when the pH of the cultures was adjusted to 0.5 on the second day (Fig. 1B). Thus, the results identified the essential role of Fur in the response of A. caldus to low pH. Furthermore, a promoter-probe vector using the reporter luciferase gene (luc) was generated to monitor the expression level of fur under low pH conditions. When the WT strain was used as the host, acid shock caused a significant increase in the luciferase activity at 48 and 72 h after stimulation compared with the nonstimulated group (Fig. 1C). A similar phenomenon appeared in the host ⌬fur strain (Fig. 1D). These results suggested that the expression level of fur is closely related to the pH of the environment. Moreover, under the condition of acid shock, the luciferase activity of the fur mutant was twice that of the WT strain at 48 and 72 h after stimulation ( Fig. 1C and D), indicating that Fur also has an effect on its own expression level in A. caldus when responding to low pH.
Role of the key domains of Fur in acid resistance of A. caldus. The amino acid sequences of the Fur proteins from acidophiles showed high similarity and conservation in the N-terminal DNA-binding domain, C-terminal dimerization domain, and metal-binding site motifs (29,38). All Fur sequences of AMD bacteria exhibited a C-terminal extension, in contrast to some neutrophilic bacteria, but the sequence of this C-terminal extension is not conserved in these acidophilic bacteria (see Fig. S2). Thus, the evolutionary conservation of the sequence and structure of Fur proteins in acidophilic bacteria indicates that this protein might play similar roles in AMD bacteria adapting to acidic environments.
To compare and identify the key domains of Fur in A. caldus, an alignment of amino acid sequences of AcFur and modeling of its three-dimensional structure of homology were performed ( Fig. 2A and B). Based on the results of multiple sequence alignment, evolutionary analysis, and structural comparison, the key residues (H31, H88, E106, and H123) of AcFur in the binding sites were found to be highly conserved in all orthologs.
We then suggested that H31 and H88 constitute the regulatory site (SI domain) responsible for binding to the DNA sequence, and the other two key residues constitute the structural site (SII domain) responsible for the steady conformation (38,39). The growth analysis showed that the mutation of SI (H31A and H88A) rendered the cells sensitive to low pH, while the substitution of the two key residues in SII (E106A and H123A) did not influence the growth of the ⌬fur(SII) strain (Fig. 2C). This result indicated that the SI domain is crucial for the maintenance of AcFur function in acid resistance. In addition, to test the role of the sequence of the AcFur C-terminal extension, a complemented Δfur-dC strain lacking the C-terminal extension was constructed. The data showed that the deletion of the C-terminal extension of AcFur resulted in a lower cell density of the Δfur-dC strain after 4 days of stimulation (Fig. 2D).

Fur-regulated genes in A. caldus under acid shock.
To further explore the Fur-mediated acid resistance mechanisms, transcriptome sequencing (RNA-seq) was performed to detect the differentially expressed genes (DEGs) in the ⌬fur strain after 48 h of acid shock. Overall, a total of 302 genes were differentially expressed in the ⌬fur mutant compared with the WT strain, including 214 upregulated and 88 downregulated genes (Fig. 3A). The KEGG pathway enrichment of the DEGs indicated that some pathways were influenced by the deletion of fur in A. caldus, and these were mainly related to bacterial chemotaxis, flagellar assembly, sulfur metabolism, nitrogen metabolism, and two-component systems (Fig. 3B). These results suggested that AcFur plays an important role in the response and adaptation of A. caldus to low pH and may affect a variety of physiological processes in A. caldus.
The deletion of the fur gene in A. caldus MTH-04 resulted in significant downregulation of genes involved in cell wall and membrane synthesis, biofilm formation, and the tetrathionate intermediate thiosulfate oxidation (S 4 I) pathway, such as the indirectly  WT] in Data Set S1). In summary, AcFur affected the expression of a variety of crucial physiological genes, suggesting that for A. caldus, the Fur-dependent regulatory network is an essential adaptive mechanism for successfully coping with an extremely acidic environment. Regulation of iron transporter genes by Fur. Some known genes and operons are involved in iron transport in A. caldus, such as feoABC and feoP. Unlike the feo operon (feoPABC) in A. ferrooxidans, the feo operon in A. caldus is composed of feoABC (Fig. 4), while feoP is located in another cluster (feoP-znuA-htrB) (Fig. 5A). The Fur binding sequences (AcFur box) were predicted upstream of both the feo operon and feoP genes (Fig. 5A). Gel shift assays showed that Fur could bind to the promoter regions of feoA  and feoP (Fig. 5B). Reverse transcription-quantitative PCR (RT-qPCR) indicated that the transcriptional levels of these two genes in the A. caldus WT strain were significantly upregulated after acid shock (Fig. 5C). The mutation of the SI domain resulted in the loss of the ability of Fur to bind to the promoters of feoA and feoP, and the affinity became weaker when the SII domain was mutated ( Fig. 5D and E). The absence of fur markedly increased the transcription of feoA and feoP, and the Fur-dependent regulatory effect was restored when fur was complemented with ⌬fur ( Fig. 5F and G). The mutation of SI and SII also affected the Fur-dependent regulation of feoA and feoP. SI mutation resulted in the obvious upregulation of feoA and feoP, which is similar to the regulatory effect caused by the deletion of fur. The influence of the SII mutation on the regulatory effect was much weaker than that of the SI mutation ( Fig. 5F and G). Therefore, Fur could directly bind to promoters of feoA and feoP, generating an inhibitory effect on the transcription of iron transport operons. These results from mutagenesis, phenotypic characterization, and function studies suggested that iron homeostasis and pH homeostasis of AcFur are not separable. Regulatory effect of Fur on biofilm formation and EPS generation. The differential expression of genes involved in biofilm formation suggested the significance of biofilm in the adaptation of A. caldus to low pH; thus, we tested the influence of fur deletion on biofilm formation by using crystal violet staining. The data showed that the deletion of fur resulted in a decrease in biofilm formation (Fig. 6A). Furthermore, extracellular polymeric substances (EPS) were extracted by using heat treatment, and carbohydrate content in EPS was measured by the anthrone method to assess the effect of the fur gene knockout on EPS. Under acid shock conditions, both the WT and fur deletion strains showed significant increases in EPS production, suggesting the critical role of EPS synthesis in A. caldus adaptation to low pH. The amount of EPS in the Δfur strain was lower than that in the WT both under normal and acidic shock conditions (Fig. 6B), suggesting that the absence of fur could affect EPS synthesis. Therefore, the significant decrease in EPS and biofilm caused by the deletion of fur suggested that AcFur is required for the regulation of biofilm formation and EPS generation under acid shock.

DISCUSSION
In this work, the indispensable role of Fur in A. caldus adapting to low pH was uncovered, thus providing new insights into the adaptive mechanism of acidophiles to extremely acidic environments. Fur and DtxR homologs are ubiquitous in almost all acidophiles of AMD ecosystems (see Table S1), suggesting that both metal response regulators Fur and DtxR are essential for AMD acidophilic bacteria. Fur and DtxR do not exhibit any obvious sequence similarity. However, Fur is similar to DtxR in its structure, such as its DNA-binding domain (37). In addition, both of these proteins regulate iron homeostasis and some basic physiological and metabolic processes (35,41). In some bacteria, such as Campylobacter jejuni (42) and Escherichia coli (43), Fur senses environmental stimuli and stresses and has a key function. The fact that many endosymbionts and those bacteria that live in cultures within nutrient-rich environments do not possess fur genes (44) suggests that the function of Fur is biased toward environmental response or adaptability. Our experimental evidence demonstrated that disruption of the fur gene reduces the acid tolerance of A. caldus and suggested the significance of Fur in acidophilic bacteria.
The sequences and structures of Fur proteins in AMD bacteria are highly conserved and homologous, but there are some differences from those in other bacteria. Unlike the Fur proteins of some neutrophilic bacteria, the nonconserved C-terminal extensions are universally present in the Fur proteins of AMD bacteria ( Fig. 2D; also, see Fig. S2). The deletion of the C-terminal extension resulted in a decrease in the solubility of the recombinant protein (see Fig. S3), which altered the regulatory effect of Fur to some extent (Fig. 2D). In general, Fur proteins contain at least two metal-binding-site motifs, including functional and structural sites (29,39,45). We found that the SI domains (H31, H88) of AcFur in the metal-binding sites were highly conserved in all orthologs and confirmed that the SI domain is important for iron transport and acid resistance in A. caldus ( Fig. 2A and C and Fig. 5D to G; also, see Fig. S2), suggesting that SI might be a functional site. In addition, the SII domain (E106, H123) is not strictly conserved in many bacteria, suggesting that SII probably plays an auxiliary role in A. caldus ( Fig. 2A and C and Fig. 5D to G).
Fur, as a global regulator, is directly or indirectly involved in a variety of crucial physiological and metabolic pathways in many bacteria (30,32,46). The RNA-seq data showed that 11.3% of genes were influenced by disruption of the fur gene in A. caldus, under the acid shock condition, including those involved in iron transport, chemotaxis and motility, biofilm formation, and energy system (Fig. 3). In addition, we showed that AcFur directly regulates iron transport and found that AcFur is important for A. caldus growth in an extremely acidic environment. Thus, it could be concluded that the Fur-dependent regulatory mode is probably a favored strategy used by acidophilic bacteria in response to acid stress (Fig. 7).
Most prokaryotic organisms can protect themselves from the environmental stress of low pH by forming biofilms and a complex cell envelope that consists of a plasma membrane, a peptidoglycan cell wall, and an outer membrane (15,47). Some genes related to cell envelope synthesis were significantly downregulated, suggesting that cell envelope synthesis and components were affected, and this change might increase proton permeability in the Δfur strain. Furthermore, the regulation of Fur on biofilm formation and EPS synthesis was confirmed by transcriptional and biochemical analyses ( Fig. 6; also, see the supplemental DEG list [fur versus WT] in Data Set S1). Disruption of the fur gene reduced biofilm formation and EPS synthesis (Fig. 6), which explained the acid-sensitive phenotype of the Δfur mutant to a certain degree (Fig. 1A and B). Flagellum-driven bacterial chemotaxis and motility are important for microorganisms responding to unfavorable environments (48,49). Because of the important role of the EPS layer in bacterial elemental sulfur oxidation (50), the decrease in EPS synthesis observed in the Δfur mutant probably reduced the available sulfur substrates and resulted in the adjustment of the two important periplasmic thiosulfate-metabolic pathways (the Sox system and the S 4 I pathway). Because of the close correlation between the Sox system and flagellum in A. caldus (51), the upregulation of the genes involved in these two systems suggested that the mutant strain had to adjust its

Fur-Dependent Acid Adaptation Strategy in Acidophiles
Applied and Environmental Microbiology metabolic flow to increase flagellum biosynthesis to escape unfavorable environments and find suitable habitats. Thus, the deletion of fur caused the readjustment of sulfur metabolism, flagellum biosynthesis, and chemotaxis, indicating the important regulatory role of Fur in A. caldus in response to low pH. The acquisition of iron may be key to the survival of bacteria at low pH; however, high levels of intracellular free iron probably increase oxidative damage to cells (34,52). It has been reported that iron transporters have been found in most acidophilic microorganisms and that these enzymes could be stabilized by "iron rivets" in acidophilic archaea (12). The transcription of feoA and feoP surged to a high level during acid stimulation (Fig. 5C), indicating that metal ion transport systems play a certain role in the acid stimulation process. Moreover, the direct regulation of iron transport genes by AcFur was also confirmed in A. caldus. These results suggested that Fur-regulated iron transport could affect the iron homeostasis of A. caldus under acid stimulation, which is a possible reason for the weak acid-adapted ability of the Δfur and Δfur (SI) strains ( Fig. 1B and Fig. 2C). In addition, some common acid resistance genes related to L-glutamate (Glu), including those for the ␥-aminobutyric acid (GABA) antiporter GadC and the Glu decarboxylase system, and DNA and protein repair system genes (13) were upregulated in the Δfur mutant after the acid shock (see the supplemental DEG list [fur versus WT] in Data Set S1). This result implied that the enhanced activity of these systems could facilitate the survival of the Δfur strain in an extremely acidic environment.
Acidity is not only a challenge for the application of acidophiles but also a basic clue for understanding the ecological behavior of AMD bacteria. Acidification occurs during bioleaching and bioremediation, which can seriously affect bacterial growth and application efficiency. The unique acid-resistant module of the acidophilic bacteria may be a potential target for application to engineering strains, with far-reaching potential applications (15,53). The findings of our study provide new insights that would augment the current knowledge base pertaining to the acid resistance process in Acidithiobacillus species, and they could also help promote the study of the environmental adaptation of acidophilic bacteria and facilitate the design of engineered biological systems.

MATERIALS AND METHODS
Strains and culture conditions. All bacterial strains, plasmids, and primers are listed in Table 1. A. caldus MTH-04 and its derivatives were grown in liquid Starkey-S 0 inorganic medium (pH 2.5) or on solid Starkey-Na 2 S 2 O 3 medium (pH 4.8). E. coli strains were grown in Luria-Bertani (LB) broth or on LB agar plates. Ampicillin (Amp), streptomycin (Sm), and kanamycin (Km) were used at 100 g/ml in the LB medium, with the concentrations doubled in the Starkey medium. Chloramphenicol (Cm) was used at 60 g/ml in the Starkey medium. The culture conditions were 40°C and 150 rpm for A. caldus MTH-04 and 37°C and 200 rpm for E. coli.
Construction of fur knockout and complementation strains. The fur gene disruption (Δfur) mutant was constructed using markerless knockout technology as described previously (54). To study the properties of the fur gene in vivo, fur complementation strains were constructed. To construct the complementation strains, markerless knock-in technology was applied to the Δfur strains. First, the suicide plasmid pSDUDI::fur (UHAϩfurϩDHA) was constructed. The upstream and downstream homologous arms (UHA and DHA) and the fur gene were linked using fusion PCR. Next, single-crossover mutants were selected on solid Starkey-Na 2 S 2 O 3 medium containing Km and identified by PCR. Finally, the pMSD1-I-SceI plasmid was transferred into the single crossover of A. caldus MTH-04 to induce the second homologous recombination, thereby generating the fur complementation strains. All plasmids, the Δfur strain, and its derivatives of A. caldus MTH-04 were confirmed by PCR and DNA sequencing (Genewiz, Tianjin, China). The construction of the fur mutant complementation strains was performed as described above.
Acid shock assay and transcriptional analysis. The wild type (WT) and the Δfur mutant of A. caldus MTH-04 were cultivated as described previously (54). Briefly, 1 ml of the treated cell culture (optical density at 600 nm [OD 600 ] ϭ 1.0) was inoculated into 150 ml of Starkey-S 0 liquid inorganic medium (pH 2.5). After 3 days of cultivation, the pH of the culture was adjusted to 0.50 using H 2 SO 4 . Additionally, a culture without the addition of H 2 SO 4 was used as the control. Total RNA was extracted 48 h after stimulation, and transcriptional analyses via RNA-seq and RT-qPCR were performed as described previously (51). The original analysis of DEGs is listed in the supplemental DEG list (fur versus WT) in Data Set S1.
Expression analysis of AcFur using a luciferase reporter gene. To investigate the expression levels of the fur gene in A. caldus responding to acid stress, the promoter region of fur (351 bp in front of the fur gene) from the A. caldus MTH-04 chromosome was amplified with the primer pair P fur -F-MluI-P fur - R-NdeI and inserted into the MluI-and NdeI-treated pJRD215-Luc-Cm plasmid containing the firefly luciferase (luc) gene and the chloramphenicol acetyltransferase (cat) gene. The generated pJRD215-P fur -Luc-Cm plasmid was conjugated into WT A. caldus MTH-04 and its Δfur mutant. Luciferase activities were detected at 0, 24, 48, and 72 h after acid shock. Luciferase activity was measured using a TransDetect single-luciferase (firefly) reporter assay kit. Briefly, a culture of each strain was collected and diluted with phosphate-buffered saline (PBS; 10 mM PO 4 3Ϫ , 0.8% NaCl, pH 7.4) to an OD 600 of 1. First, 2 ml of cells was washed three times with PBS buffer and harvested by centrifugation at 10,000 ϫ g for 2 min at 4°C. Next, the cells were lysed by adding 200 l of cell lysis buffer (1ϫ) and incubating for 20 min at 20°C. Finally, after centrifugation at 13,000 ϫ g for 15 min at 4°C, 100 l of the supernatant was collected, and 90 l of luciferase reaction reagent was added. Luciferase activity in the samples was measured using an EnSpire luminometer.
Crystal violet biofilm assay and extracellular polymeric substance extraction. A. caldus MTH-04 strains were incubated in 150 ml of Starkey-S 0 medium for 8 days and then collected and diluted with Starkey liquid inorganic medium to an OD 600 of 1.0. Subsequently, 1 ml of each type of bacterial cells was inoculated into 30 ml of sulfur coupon-containing Starkey medium. After static cultivation for 8 days, the sulfur coupons were washed gently three times with double-distilled water (ddH 2 O), dried for 40 min at 50°C, and stained with 1 ml of 1% (wt/vol) crystal violet (CV) solution for 15 min at room temperature. After staining, the sulfur coupons were washed with ddH 2 O and dried for 40 min at 50°C; 1 ml of 30% acetic acid was added to dissolve the CV, and the absorbance of the suspensions at 590 nm was quantitatively measured using a spectrophotometer (55).
Extracellular polymeric substances (EPS) was extracted using heat treatment (56). The EPS extraction method was adapted from that described by Xiao et al. (57). Briefly, the cells were harvested by centrifugation at 5,000 ϫ g for 10 min at 4°C and washed twice with PBS buffer (10 mM PO 4 3Ϫ , 0.8% NaCl, pH 7.4). The washed cell pellets were resuspended in PBS buffer and heated in a water bath at 60°C for 30 min. After centrifugation at 12,000 ϫ g for 30 min at 4°C, the supernatant was filtered using a 0.22-m filter. Carbohydrate content was measured using the anthrone method (58).
EMSA. AcFur direct binding of the AcFur box was demonstrated by electrophoretic mobility shift assays (EMSA). Using the AcFur box of Pseudomonas aeruginosa as a pattern, the genome of A. caldus MTH-04 was searched with Virtual Footprint (59), and the potential AcFur box sequences were then selected based on the score. A 200-bp feo operator and a 190-bp feoP promoter containing the AcFur box were amplified from the A. caldus MTH-04 genome. PCR products were purified using a gel extraction kit (Omega), and 0.4 nM DNA was incubated with different concentrations of protein (0, 2, 4, and 8 nM) in Tris-HCl (10 mM, pH 8.0), NaCl (100 mM), and MnCl 2 (1 mM) at room temperature for 30 min. The samples were separated in native polyacrylamide gels (5%) containing 0.5ϫ Tris-borate (TBE) buffer and stained with ethidium bromide (EB).
Statistical analysis. All measurements were performed in triplicate, and all assays were repeated at least three times. Statistical analysis was performed using GraphPad Prism7.0 software. Significance was assessed by using an independent-samples t test.
Data availability. The A. caldus MTH-04 genome sequence and the raw RNA-seq data have been submitted to NCBI with the accession numbers PRJNA318365 and PRJNA577835, respectively. The strain A. caldus MTH-04 was isolated from Tengchong area, Yunnan Province, China, and has been deposited in the China General Microbiological Culture Collection Center (CGMCC) with the accession number CGMCC 1.15711. The mutants of A. caldus MTH-04 constructed in this study were stocked in our lab and will be made available upon request.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 1.1 MB. SUPPLEMENTAL FILE 2, XLS file, 0.1 MB.
tance and suggestions in constructing and testing the gene mutants. We also thank the Core Facilities Sharing Platform for Life and Environment Sciences of Shandong University, including Rui Wang for fluorescence detection, Cheng-jia Zhang for growth measurement, Nan-nan Dong for providing the bacteriological incubator, and Zhi-feng Li for RT-qPCR instruction. We thank Editage for English language editing. We declare no conflicts of interest.