Genomic and Metabolic Insights into Denitrification, Sulfur Oxidation, and Multidrug Efflux Pump Mechanisms in the Bacterium Rhodoferax sediminis sp. nov.

This genus contains both phototrophs and nonphototrophic members. Here, we present a high-quality complete genome of the strain CHu59-6-5T, isolated from a freshwater sediment. The circular chromosome (4.39 Mbp) of the strain CHu59-6-5T has 64.4% G+C content and contains 4240 genes, of which a total of 3918 genes (92.4%) were functionally assigned to the COG (clusters of orthologous groups) database. Functional genes for denitrification (narGHJI, nirK and qnor) were identified on the genomes of the strain CHu59-6-5T, except for N2O reductase (nos) genes for the final step of denitrification. Genes (soxBXAZY) for encoding sulfur oxidation proteins were identified, and the FSD and soxF genes encoding the monomeric flavoproteins which have sulfide dehydrogenase activities were also detected. Lastly, genes for the assembly of two different RND (resistance-nodulation division) type efflux systems and one ABC (ATP-binding cassette) type efflux system were identified in the Rhodoferax sediminis CHu59-6-5T. Phylogenetic analysis based on 16S rRNA sequences and Average Nucleotide Identities (ANI) support the idea that the strain CHu59-6-5T has a close relationship to the genus Rhodoferax. A polyphasic study was done to establish the taxonomic status of the strain CHu59-6-5T. Based on these data, we proposed that the isolate be classified to the genus Rhodoferax as Rhodoferax sediminis sp. nov. with isolate CHu59-6-5T.


Introduction
Rhodoferax species are frequently found in stagnant aquatic environments exposed to light [1], and some have been isolated from fresh water, sewage, sludge and sediments [2][3][4][5][6][7]. In the case of Rhodoferax antarcticus, strains were first isolated from microbial mats collected from saline ponds [2]. Rhodoferax growth occurs in a temperature range from 2 to 35 • C [1][2][3][4][5][6][7]. Rhodoferax fermentans, Rhodoferax ferrireducens, Rhodoferax koreense, Rhodoferax lacus and Rhodoferax bucti are mesophilic species with an optimal growth temperature between 25 and 30 • C [1,3,[5][6][7]. The other two species, R. antarcticus and Rhodoferax saidenbachensis, are psychrotolerant with optimal growth temperatures ranging from 15 to 20 • C, but they are capable of growth at temperatures near 0 • C [2,4]. Rhodoferax is also well known for its diverse metabolic pathways. Members of the Rhodoferax species are able to grow phototrophically, aerobically and anaerobically. Two species in this genus, R. fermentans and R. antarcticus, grow photoheterotrophically using carbon sources such as acetate, pyruvate, lactate and succinate. Genomic features that provide photosynthesis gene clusters of the R. antarcticus have already been reported [8]. In contrast, R. ferrireducens is able to grow anaerobically using organic electron donors to reduce ferric iron (Fe +3 ) to ferrous (Fe +2 ) [1][2][3][4][5][6][7]. As of 2019, seven genomes have been sequenced from the genus Rhodoferax. Some of them contain the genes for the RuBisCo (ribulose 1, 5-bisphosphate carboxylase/oxygenase), several heavy metal resisting genes, and putative arsenite efflux pump genes [9]. Recently, the species R. ferrireducens was applied to the area of sustainable energy microbial fuel cells (MFC), where a bacterial suspension was used as a source of electrons for the bacteria [10]. Additionally, Rhodoferax is one of most abundant groups in the lakes of northeastern Germany [11], understanding the ecology of the genus Rhodoferax and their roles in geochemical cycles, as well as their physiology, would contribute to the understanding of its functions in environments. In the present study, a non-phototrophic member in this genus Rhodoferax obtained from freshwater was investigated at the genome and metabolic level, and a polyphasic approach was applied to establish its taxonomic status.

Sampling, Isolation and Cultivation of the Strain
Strain CHu59-6-5 T was recovered from a 67 cm-long sediment core (36 • 22'30" N, 127 • 33'58" E) collected at a water depth of 17 m from the Daechung Reservoir using a modified gravity corer (Wildco, FL, USA) ( Figure 1). The Daechung Reservoir is located in the central region of South Korea, which is a large branch-type lake, and the reservoir has a 72 m-high dam with a gross storage capacity of 1490 Mm 3 , where cyanobacterial blooms occur every year because of eutrophication and global warming [12][13][14][15]. Approximately one gram of sediment sample was applied to the serial dilution method in a 0.85% saline solution. A total of 100 µl aliquot of each serial dilution (10 −6 or 10 −7 ) was spread on onto modified 1/10 R2A medium (Difco, NJ, USA), which is specifically used for isolating aquatic bacteria [16] and incubated at 25 • C under aerobic heterotrophic conditions. A colorless colony CHu59-6-5 T was isolated after 6 days and routinely subcultured on an R2A agar at 30 • C for 48 h. For long term preservation of the culture, glycerol stocks (20% v/v) were prepared and stored at −70 • C. Strains were cultured on R2A agar for most physiological experiments.

Sampling, Isolation and Cultivation of the Strain
Strain CHu59-6-5 T was recovered from a 67 cm-long sediment core (36°22'30" N, 127°33'58" E) collected at a water depth of 17 m from the Daechung Reservoir using a modified gravity corer (Wildco, FL, USA) ( Figure 1). The Daechung Reservoir is located in the central region of South Korea, which is a large branch-type lake, and the reservoir has a 72 m-high dam with a gross storage capacity of 1490 Mm 3 , where cyanobacterial blooms occur every year because of eutrophication and global warming [12][13][14][15]. Approximately one gram of sediment sample was applied to the serial dilution method in a 0.85% saline solution. A total of 100 µ l aliquot of each serial dilution (10 −6 or 10 −7 ) was spread on onto modified 1/10 R2A medium (Difco, NJ, USA), which is specifically used for isolating aquatic bacteria [16] and incubated at 25 °C under aerobic heterotrophic conditions. A colorless colony CHu59-6-5 T was isolated after 6 days and routinely subcultured on an R2A agar at 30 °C for 48 h. For long term preservation of the culture, glycerol stocks (20% v/v) were prepared and stored at −70 °C. Strains were cultured on R2A agar for most physiological experiments. For the comparative whole-cell fatty acid profile, strains CHu59-6-5 T , R. saidenbachensis DSM 22694 T , R. lacus KACC 18983 T , R. bucti KCTC 62564 T and R. koreense KCTC 52288 T were cultured on R2A agar for 3 days at 30 • C. Cell harvesting standardization was done following the method described by Jin et al. [20], and the extracted fatty acids were analyzed by gas chromatography (GC) (Hewlett Packard 6890, Kyoto, Japan) and identified in the TSBA 6 (Trypticase Soy Broth Agar) database using the Sherlock software 6.1. Respiratory isoprenoid quinones were extracted and analyzed by high performance liquid chromatography (HPLC) (Shimadzu, Kyoto, Japan) with the YMC-Pack ODS-A column, following the method described by Komagata and Suzuki [21]. Polar lipids were extracted according to the method of Tindall [22]. The biomass used for lipid extraction was obtained from cultures growing on R2A agar plates at 30 • C for 3 days. Several spraying reagents were applied for visualizing the spots on the two-dimensional thin layer chromatography (TLC) on silica gel 60 F254 plates (Merck): molybdatophosphoric acid for total spots, ninhydrin for aminolipids, molybdenum blue for phospholipids and alphanaphthol solution for glycolipids.

Genomic and Phylogenetic Analyses
Whole genomic DNA was extracted using a FastDNA TM SPIN kit according to the manufacturer's instructions. The 16S rRNA gene of strain Chu59-6-5 T was amplified by PCR (polymerase chain reaction) with the following universal primer sets: 27F (5'-AGA GTT TGA TCM TGG CTC AG-3'; Escherichia coli position 8-27) and 1492R (5'-TAC GGY TAC CTT GTT ACG ACT T-3'; E. coli position 1492-1510) [23], and identified in the server of EzBioCloud [24]. For the phylogenetic analysis, 16S rRNA gene sequences of strain CHu59-6-5 T and closely related species were aligned with clustal x [25] and edited with bioedit [26]. Based on the aligned sequences, the phylogenetic tree was reconstructed using the following algorithms in the mega7 software [27]: neighbor-joining [28], maximum-parsimony [29] and maximum-likelihood methods [30]. In each case, bootstrap values were calculated based on 1000 resamplings of the sequences [31]. The whole genome sequencing was carried out using the SMRT (Single Molecule, Real-Time) platform at Novogene Biotechnology (Beijing, China) together with Illumina next-generation sequencing technology. The low-quality read filtration and assemblage of the draft genome were conducted using SMRT version 5.0.1. The genome annotation of strain CHu59-6-5 T was annotated in the RAST pipeline, and a sequence-based comparison was made using the SEED Viewer [32,33]. The predicted protein coding sequences (CDSs) were submitted to the COG (Clusters of Orthologous Groups) database (http://www.ncbi.nlm.nih.gov/COG/) to generate the functional category and summary statistics [34,35]. The average nucleotide identity (ANI) and digital DNA-DNA hybridization (dDDH) values were calculated using the OrthoANI tool in EZBioCloud web server [36].

Genomic Analysis: The Taxonomic Status
The 16S rRNA gene sequence (1478 nt) of strain CHu59-6-5 T was compared against the 16S rRNA gene sequences of representative species within the genus Rhodoferax and related genera. We used the EzTaxon-e server (www.ezbiocloud.net) to search for their closest members. The results show that strain CHu59-6-5 T shares a 98.8% pairwise similarity with Rhodoferax saidenbachensis ED16 T , 98.4% with Rhodoferax lacus IMCC26218 T , 98.2% with Rhodoferax bucti GSA243-2 T and less than 98.0% with other species in the genus Rhodoferax. The strain CHu59-6-5 T also shares high similarities with other species than just members of Rhodoferax: 98.1% with Curvibacter delicates NBRC 14919 T Curvibacter fontanus AQ9 T , 97.8% with Curvibacter fontanus AQ9 T , 97.5% with Variovorax boronicumulans BAM-48 T and 97.3% with Variovorax paradoxus NBRC 15149 T . However, strain CHu59-6-5 T clustered clearly with the species of Rhodoferax from the topology of the phylogenetic tree ( Figure 2). To determine genomic relatedness with other Rhodoferax species and its genomic characteristics, whole genome sequence of strain CHu59-6-5 T was obtained by the PacBio platform together with the Illumina MiSeq platform. The draft genome sequence of strain CHu59-6-5 T was deposited at DDBJ/EMBL/GenBank with the accession number CP036282. The genomic DNA G+C content of the strain was 64.4 mol%, which is within the range reported for the genus Rhodoferax (59.52-66.2 mol%). The observed ANI and dDDH values between strain CHu59-6-5 T and all species of Rhodoferax were 74.3%-76.9 % and 21.3%-23.3 % (Table 3), respectively, which were much lower than the species separation threshold of 95%-96% and in fact fall in the intergeneric range [37][38][39]. Data are from the present study. −, not detected. Note: ‡ Summed Feature 3 contains C16:1 ω7c and/or C16:1 ω6c; summed feature 8 contains C18:1 ω7c and/or C18:1 ω6c.

Genomic Analysis: The Taxonomic Status
The 16S rRNA gene sequence (1478 nt) of strain CHu59-6-5 T was compared against the 16S rRNA gene sequences of representative species within the genus Rhodoferax and related genera. We used the EzTaxon-e server (www.ezbiocloud.net) to search for their closest members. The results show that strain CHu59-6-5 T shares a 98.8% pairwise similarity with Rhodoferax saidenbachensis ED16 T , 98.4% with Rhodoferax lacus IMCC26218 T , 98.2% with Rhodoferax bucti GSA243-2 T and less than 98.0% with other species in the genus Rhodoferax. The strain CHu59-6-5 T also shares high similarities with other species than just members of Rhodoferax: 98.1% with Curvibacter delicates NBRC 14919 T Curvibacter fontanus AQ9 T , 97.8% with Curvibacter fontanus AQ9 T , 97.5% with Variovorax boronicumulans BAM-48 T and 97.3% with Variovorax paradoxus NBRC 15149 T . However, strain CHu59-6-5 T clustered clearly with the species of Rhodoferax from the topology of the phylogenetic tree ( Figure 2). To determine genomic relatedness with other Rhodoferax species and its genomic characteristics, whole genome sequence of strain CHu59-6-5 T was obtained by the PacBio platform together with the Illumina MiSeq platform. The draft genome sequence of strain CHu59-6-5 T was deposited at DDBJ/EMBL/GenBank with the accession number CP036282. The genomic DNA G+C content of the strain was 64.4 mol%, which is within the range reported for the genus Rhodoferax (59.52-66.2 mol%). The observed ANI and dDDH values between strain CHu59-6-5 T and all species of Rhodoferax were 74.3%-76.9 % and 21.3%-23.3 % (Table 3), respectively, which were much lower than the species separation threshold of 95%-96% and in fact fall in the intergeneric range [37][38][39].

Genome Properties
The genome of strain CHu59-6-5 T consists of a single circular chromosome of 4,387,497 base pairs with a G+C content of 64.4% (Figure 3). Of the 4240 genes identified in the total genome, 4191 were protein-encoding genes, 49 were ribosomal or transfer RNAs, and 133 were putative pseudogenes (Table 3 and Supplementary Table S1). A total of 3918 genes (92.4%) were functionally assigned to the COG database. The distribution of the genes into the COG functional categories is presented in Table 3 and Supplementary Figure S3. Similar to the closely related Rhodoferax strains, abundant genes related to amino acid transport and metabolism (COG category E), transcription (COG category K), and energy production and conversion (COG category C) were observed. Unexpectedly, 27.7% of genes were assigned to the unknown function COG category in the genome of strain CHu59-6-5 T (Supplementary Table S2).

Genome Properties
The genome of strain CHu59-6-5 T consists of a single circular chromosome of 4,387,497 base pairs with a G+C content of 64.4% (Figure 3). Of the 4240 genes identified in the total genome, 4191 were protein-encoding genes, 49 were ribosomal or transfer RNAs, and 133 were putative pseudogenes (Table 3 and Supplementary Table S1). A total of 3918 genes (92.4%) were functionally assigned to the COG database. The distribution of the genes into the COG functional categories is presented in Table 3 and Supplementary Figure S3. Similar to the closely related Rhodoferax strains, abundant genes related to amino acid transport and metabolism (COG category E), transcription (COG category K), and energy production and conversion (COG category C) were observed. Unexpectedly, 27.7% of genes were assigned to the unknown function COG category in the genome of strain CHu59-6-5 T (Supplementary Table S2).

Carbon Metabolism
Two anoxygenic phototrophic species in the genus Rhodoferax, R. fermentans and R. antarcticus have been reported. Both members are confirmed to have gene clusters including genes encoding

Carbon Metabolism
Two anoxygenic phototrophic species in the genus Rhodoferax, R. fermentans and R. antarcticus have been reported. Both members are confirmed to have gene clusters including genes encoding light-harvesting complex and RuBisCo (ribulose 1, 5-bisphosphate carboxylase/oxygenase) genes [1,2,8,40,41]. Genes of complete sets of glycolysis and the citric acid cycle were identified in the genome of strain CHu59-6-5 T . Furthermore, genes for the Entner-Doudoroff pathway and pentose phosphate pathway were also detected ( Figure 4). No genes for encoding RuBisCO (ribulose 1,5-bisphosphate carboxylase/oxygenase) or light-harvesting complex were found in the genome of strain CHu59-6-5 T , which means that strain CHu59-6-5 T is not able to fix CO 2 [41,42]. Strain CHu59-6-5 T contains genes encoding phosphoenolpyruvate (PEP) carboxylase (EC 4.1.1.31), but lacks genes encoding pyruvate-phosphate dikinase. This means that some other enzyme together with PEP carboxylase assists in adding CO 2 to PEP [43]. light-harvesting complex and RuBisCo (ribulose 1, 5-bisphosphate carboxylase/oxygenase) genes [1,2,8,40,41]. Genes of complete sets of glycolysis and the citric acid cycle were identified in the genome of strain CHu59-6-5 T . Furthermore, genes for the Entner-Doudoroff pathway and pentose phosphate pathway were also detected ( Figure 4). No genes for encoding RuBisCO (ribulose 1,5bisphosphate carboxylase/oxygenase) or light-harvesting complex were found in the genome of strain CHu59-6-5 T , which means that strain CHu59-6-5 T is not able to fix CO2 [41,42].  Figure 4. Model predictions for metabolism and some transporters of strain CHu59-6-5 T . Nitrogen metabolic and sulfur oxidation pathways are indicated. Several transporters and electron transport systems are shown. Two electron chains, complex I (NADH dehydrogenase) and complex II (succinate dehydrogenase), were observed. Complex II usually parallel electron transport to complex I. A C-family heme-copper cbb3 oxidase, which presumably serves as the terminal electron acceptor during aerobic respiration, was identified. The genome also encodes genes related to the type IV pilus. PEP, phosphoenolpyruvate; GS/GOGAT: glutamine synthetase/glutamate oxoglutarate aminotransferase.

Denitrification
Denitrification can take place in both terrestrial and marine ecosystems and is one of the main branches of the global nitrogen cycle supported by bacteria. Typically, denitrification occurs in anoxic environments, where the concentration of dissolved and available oxygen is depleted and nitrate (NO3 − ) or nitrite (NO2 − ) can be used as a substitute terminal electron acceptor instead oxygen [44][45][46].
Key functional genes (narGHJI, nirK and qnor) for denitrification were identified on the genomes of the strain CHu59-6-5 T except for Nos genes. Currently, denitrification starts with a membrane bound (narGHI) in the mesophilic models of bacteria [47,48]. Genes encoding nitrate reductase (narGHJI) and nitrate/nitrite transporters (naNiT) were identified ( Figure 5A). Nitrite produced in the cytoplasm by NarGHJI will then be secreted to the periplasm through a membrane transporter (NaNiT) and reduced to nitric oxide (NO) by copper-containing nitrite reductase (NirK). Strain . Model predictions for metabolism and some transporters of strain CHu59-6-5 T . Nitrogen metabolic and sulfur oxidation pathways are indicated. Several transporters and electron transport systems are shown. Two electron chains, complex I (NADH dehydrogenase) and complex II (succinate dehydrogenase), were observed. Complex II usually parallel electron transport to complex I. A C-family heme-copper cbb3 oxidase, which presumably serves as the terminal electron acceptor during aerobic respiration, was identified. The genome also encodes genes related to the type IV pilus. PEP, phosphoenolpyruvate; GS/GOGAT: glutamine synthetase/glutamate oxoglutarate aminotransferase.

Denitrification
Denitrification can take place in both terrestrial and marine ecosystems and is one of the main branches of the global nitrogen cycle supported by bacteria. Typically, denitrification occurs in anoxic environments, where the concentration of dissolved and available oxygen is depleted and nitrate (NO 3 − ) or nitrite (NO 2 − ) can be used as a substitute terminal electron acceptor instead oxygen [44][45][46].
Key functional genes (narGHJI, nirK and qnor) for denitrification were identified on the genomes of the strain CHu59-6-5 T except for Nos genes. Currently, denitrification starts with a membrane bound (narGHI) in the mesophilic models of bacteria [47,48]. Genes encoding nitrate reductase (narGHJI) and nitrate/nitrite transporters (naNiT) were identified ( Figure 5A). Nitrite produced in the cytoplasm by NarGHJI will then be secreted to the periplasm through a membrane transporter (NaNiT) and reduced to nitric oxide (NO) by copper-containing nitrite reductase (NirK). Strain CHu59-6-5 T possesses two quinone-dependent nitric oxide reductases (qNor). As the NO is highly toxic, it needs to be reduced to N 2 O immediately by a membrane-bound nitric oxide reductase (Nor) ( Figure 5B). Depending on the type of Nor enzyme used, electrons for this reaction can be transferred by a periplasmic cytochrome c (cNor) or by quinones (qNor) [47,[49][50][51]. Strain CHu59-6-5 T does not contain N2O reductase (Nos) genes for the final step of denitrification. In many denitrifying microorganisms, N 2 O is finally reduced to N 2 by a periplasmic reductase (NosZ), in which a cytochrome c or a type I copper protein has to act as an electron donor [52]. However, the cases in which organisms without nos genes still can produce N 2 gas have been studied, supporting the idea that a different type of N 2 O reductase needs to be discovered [53,54]. CHu59-6-5 T possesses two quinone-dependent nitric oxide reductases (qNor). As the NO is highly toxic, it needs to be reduced to N2O immediately by a membrane-bound nitric oxide reductase (Nor) ( Figure 5B). Depending on the type of Nor enzyme used, electrons for this reaction can be transferred by a periplasmic cytochrome c (cNor) or by quinones (qNor) [47,[49][50][51]. Strain CHu59-6-5 T does not contain N2O reductase (Nos) genes for the final step of denitrification. In many denitrifying microorganisms, N2O is finally reduced to N2 by a periplasmic reductase (NosZ), in which a cytochrome c or a type I copper protein has to act as an electron donor [52]. However, the cases in which organisms without nos genes still can produce N2 gas have been studied, supporting the idea that a different type of N2O reductase needs to be discovered [53,54].

Sulfur Oxidation
Most of the sulfur-oxidizing bacteria are autotrophic and use reduced sulfur as electron donors for carbon dioxide fixation [55]; chemolithotrophic microorganisms obtain energy by oxidizing inorganic compounds for their structural components to survive, grow and reproduce [56,57]. Some inorganic forms of reduced sulfur (H2S/HS − , S 0 ) can be oxidized by chemolithotrophic sulfur-oxidizing bacteria coupled to the reduction of oxygen (O2) or nitrate (NO3 − ) [56,58].
The functional genes for sulfur oxidation (sox) participating in a thiosulfate-oxidizing multienzyme system include SoxXYZABDEFGH, which is able to oxidize various reduced sulfur compounds to sulfates [57]. The strain CHu59-6-5 T possesses genes (soxBXAZY) encoding for sulfur oxidation proteins (Figure 6), soxB, soxXA and soxZY genes are suggested to be essential to thiosulfate oxidation [59,60]. The SoxYZ complex appears as hetero-and homo-dimers with protein disulfide linked subunits, SoxB contains a dinuclear manganese cluster which is proposed to function as sulfate thiohydrolase and interacts with the SoxYZ complex [61,62]. The FSD and soxF gene encoding the monomeric flavoproteins which have sulfide dehydrogenase activities has also been detected. Those Color key: pink, nor genes; light orange, nir genes; blue-grey, cco genes (cytochrome c oxidase subunit); green, nnr genes (NO sensing protein (NnrS); denitrification regulatory protein (NnrU); white, dnr gene (NO responding transcriptional regulator); blue, naNiT genes (nitrate/nitrite transporter); yellow, naRas gene (nitrite reductase [NAD(P)H] large subunit (NaRas1), Nitrite reductase [NAD(P)H] small subunit (NaRas2); purple, reg genes; gold, nar genes; grey, naT gene (nitrate ABC transporter). (B) The substance transformation in the denitrification process was identified from the genome.

Sulfur Oxidation
Most of the sulfur-oxidizing bacteria are autotrophic and use reduced sulfur as electron donors for carbon dioxide fixation [55]; chemolithotrophic microorganisms obtain energy by oxidizing inorganic compounds for their structural components to survive, grow and reproduce [56,57]. Some inorganic forms of reduced sulfur (H 2 S/HS − , S 0 ) can be oxidized by chemolithotrophic sulfur-oxidizing bacteria coupled to the reduction of oxygen (O 2 ) or nitrate (NO 3 − ) [56,58].
The functional genes for sulfur oxidation (sox) participating in a thiosulfate-oxidizing multienzyme system include SoxXYZABDEFGH, which is able to oxidize various reduced sulfur compounds to sulfates [57]. The strain CHu59-6-5 T possesses genes (soxBXAZY) encoding for sulfur oxidation proteins ( Figure 6), soxB, soxXA and soxZY genes are suggested to be essential to thiosulfate oxidation [59,60].
The SoxYZ complex appears as hetero-and homo-dimers with protein disulfide linked subunits, SoxB contains a dinuclear manganese cluster which is proposed to function as sulfate thiohydrolase and interacts with the SoxYZ complex [61,62]. The FSD and soxF gene encoding the monomeric flavoproteins which have sulfide dehydrogenase activities has also been detected. Those sulfur oxidation Sox systems indicate that the strain CHu59-6-5 T appears to be capable of thiosulfate oxidation to sulfate. sulfur oxidation Sox systems indicate that the strain CHu59-6-5 T appears to be capable of thiosulfate oxidation to sulfate.

RND and ABC Efflux Systems
Bacterial efflux pumps are categorized into five families: ABC, the ATP-binding cassette superfamily; MFS, the major facilitator superfamily; MDR, the small multidrug resistance family; RND, the resistance-nodulation division superfamily; and MATE, the multidrug and toxic compound extrusion family [63,64]. RND efflux systems, a category of bacterial efflux pumps, play a prominent role in both intrinsic and acquired multidrug resistance identified in Gram-negative bacteria and catalyze the active efflux of a wide variety of antibacterial substrates including antibiotics and chemotherapeutic agents [65,66]. RND efflux systems are normally composed of three proteins: an inner membrane transporter, an outer membrane protein that functions as a pore, and a periplasmic protein that interacts with both the inner and outer membrane proteins to provide a conduit for the extrusion of small molecules [67].
Two different RND type efflux systems and one ABC type efflux system have been detected in the strain CHu59-6-5 T . The genome of the strain CHu59-6-5 T contains three-gene operon cmeABC ( Figure 7A). CmeABC is also a tripartite efflux system composed of three proteins CmeA, CmeB and CmeC: the inner membrane protein CmeB, the periplasmic fusion protein CmeA, and the outer membrane protein CmeC [68]. A TetR family regulator gene AcrR was present just upstream of the operon. The AcrR regulator is known to act to repress the AcrAB operon in E. coli [69]. Except for the complete CmeABC gene cluster, three more CmeC genes encoding outer membrane lipoproteins were found throughout the genome. The genome also has been found to have complete genes for the AcrAB-TolC efflux pump, which has shown to be resistant to chloramphenicol, fluoroquinolone, tetracycline, novobiocin, rifampin, fusidic acid, nalidixic acid and β-lactam antibiotics [70]. Except for the complete AcrAB-TolC gene cluster, two more AcrB genes encoding the inner membrane proteins were detected through the genome. Interestingly, the genome of the strain CHu59-6-5 T possesses an ABC-type efflux pump MacAB-TolC ( Figure 7B), which has resistance to a variety of macrolides, aminoglycosides and polymyxins [71]. TolC is an outer membrane protein which interacts with several inner membrane efflux pumps to expel antibiotics and export virulence factors from bacteria [72], and works in combination with other RND, ABC, and MFS efflux pumps [73,74]. The genome of strain CHu59-6-5 T possesses only one TolC gene, presumably working cooperatively as outer membrane proteins of AcrAB-TolC and MacAB-TolC. Those efflux systems may function complementarily with them and can work sequentially when one fails. (A)

RND and ABC Efflux Systems
Bacterial efflux pumps are categorized into five families: ABC, the ATP-binding cassette superfamily; MFS, the major facilitator superfamily; MDR, the small multidrug resistance family; RND, the resistance-nodulation division superfamily; and MATE, the multidrug and toxic compound extrusion family [63,64]. RND efflux systems, a category of bacterial efflux pumps, play a prominent role in both intrinsic and acquired multidrug resistance identified in Gram-negative bacteria and catalyze the active efflux of a wide variety of antibacterial substrates including antibiotics and chemotherapeutic agents [65,66]. RND efflux systems are normally composed of three proteins: an inner membrane transporter, an outer membrane protein that functions as a pore, and a periplasmic protein that interacts with both the inner and outer membrane proteins to provide a conduit for the extrusion of small molecules [67].
Two different RND type efflux systems and one ABC type efflux system have been detected in the strain CHu59-6-5 T . The genome of the strain CHu59-6-5 T contains three-gene operon cmeABC ( Figure 7A). CmeABC is also a tripartite efflux system composed of three proteins CmeA, CmeB and CmeC: the inner membrane protein CmeB, the periplasmic fusion protein CmeA, and the outer membrane protein CmeC [68]. A TetR family regulator gene AcrR was present just upstream of the operon. The AcrR regulator is known to act to repress the AcrAB operon in E. coli [69]. Except for the complete CmeABC gene cluster, three more CmeC genes encoding outer membrane lipoproteins were found throughout the genome. The genome also has been found to have complete genes for the AcrAB-TolC efflux pump, which has shown to be resistant to chloramphenicol, fluoroquinolone, tetracycline, novobiocin, rifampin, fusidic acid, nalidixic acid and β-lactam antibiotics [70]. Except for the complete AcrAB-TolC gene cluster, two more AcrB genes encoding the inner membrane proteins were detected through the genome. Interestingly, the genome of the strain CHu59-6-5 T possesses an ABC-type efflux pump MacAB-TolC ( Figure 7B), which has resistance to a variety of macrolides, aminoglycosides and polymyxins [71]. TolC is an outer membrane protein which interacts with several inner membrane efflux pumps to expel antibiotics and export virulence factors from bacteria [72], and works in combination with other RND, ABC, and MFS efflux pumps [73,74]. The genome of strain CHu59-6-5 T possesses only one TolC gene, presumably working cooperatively as outer membrane proteins of AcrAB-TolC and MacAB-TolC. Those efflux systems may function complementarily with them and can work sequentially when one fails.
The genome of strain CHu59-6-5 T contains genes for motility via IV pili but not flagella. Type IV pili, important virulence factors in many bacterial pathogens, generate motile forces called twitching or gliding motility in Bacteria and Archaea. The biogenesis of Type IV pili relies on macromolecular assemblies composed of 15 conserved proteins (PilC, PilD, PilE, PilF, PilG, PilH, PilI, PilJ, PilK, PilM, PilN, PilO, PilP, PilQ and PilW) in model Gram-negative bacteria [75,76]. The strain CHu59-6-5 T contains all 15 necessary genes for the proteins, and the type IV pilus genes in the strain CHu59-6-5 T are located in a more typical pattern throughout the chromosome (Supplementary Figure S4). Unfortunately, no pili of any type have been found in the cultured CHu59-6-5 T cells (Supplementary Figure S1), but the gliding motility was observed by a phase-contrast microscope.
complete CmeABC gene cluster, three more CmeC genes encoding outer membrane lipoproteins were found throughout the genome. The genome also has been found to have complete genes for the AcrAB-TolC efflux pump, which has shown to be resistant to chloramphenicol, fluoroquinolone, tetracycline, novobiocin, rifampin, fusidic acid, nalidixic acid and β-lactam antibiotics [70]. Except for the complete AcrAB-TolC gene cluster, two more AcrB genes encoding the inner membrane proteins were detected through the genome. Interestingly, the genome of the strain CHu59-6-5 T possesses an ABC-type efflux pump MacAB-TolC ( Figure 7B), which has resistance to a variety of macrolides, aminoglycosides and polymyxins [71]. TolC is an outer membrane protein which interacts with several inner membrane efflux pumps to expel antibiotics and export virulence factors from bacteria [72], and works in combination with other RND, ABC, and MFS efflux pumps [73,74]. The genome of strain CHu59-6-5 T possesses only one TolC gene, presumably working cooperatively as outer membrane proteins of AcrAB-TolC and MacAB-TolC. Those efflux systems may function complementarily with them and can work sequentially when one fails.
The genome of strain CHu59-6-5 T contains genes for motility via IV pili but not flagella. Type IV pili, important virulence factors in many bacterial pathogens, generate motile forces called twitching or gliding motility in Bacteria and Archaea. The biogenesis of Type IV pili relies on macromolecular assemblies composed of 15 conserved proteins (PilC, PilD, PilE, PilF, PilG, PilH, PilI, PilJ, PilK, PilM, PilN, PilO, PilP, PilQ and PilW) in model Gram-negative bacteria [75,76]. The strain CHu59-6-5 T contains all 15 necessary genes for the proteins, and the type IV pilus genes in the strain CHu59-6-5 T are located in a more typical pattern throughout the chromosome (Supplementary Figure S4). Unfortunately, no pili of any type have been found in the cultured CHu59-6-5 T cells (Supplementary Figure S1), but the gliding motility was observed by a phase-contrast microscope.

Conclusions
To summarize our investigations, the strain CHu59-6-5 T from a freshwater sediment belonging to the genus Rhodoferax was studied through genome analysis, together with a polyphasic approach. The phylogenetic analysis indicated that the strain CHu59-6-5 T was affiliated closely with the species of Rhodoferax, based on the phylogenetic, genomic and physiological differences; thus, we propose the strain CHu59-6-5 T as a novel species, Rhodoferax sediminis sp. nov., in the family Comamonadaceae.

Conclusions
To summarize our investigations, the strain CHu59-6-5 T from a freshwater sediment belonging to the genus Rhodoferax was studied through genome analysis, together with a polyphasic approach. The phylogenetic analysis indicated that the strain CHu59-6-5 T was affiliated closely with the species of Rhodoferax, based on the phylogenetic, genomic and physiological differences; thus, we propose the strain CHu59-6-5 T as a novel species, Rhodoferax sediminis sp. nov., in the family Comamonadaceae.
The genome sequence and comparative genome analyses of Rhodoferax sediminis CHu59-6-5 T provide a genetic blueprint and physiological characteristics which help us to understand the different metabolism and evolutionary life of the genus Rhodoferax.
The GenBank/EMBL/DDBJ accession numbers for the 16S rRNA gene sequence the whole genome sequence of strain CHu59-6-5T are MF770245 and CP035503, respectively.