Background

Sulfate-reducing bacteria (SRB) are widely distributed in different environments, e.g., wastewater treatment systems [1], freshwater sediments [2], and marine sediments [3]. Dissimilatory sulfate reduction to sulfide by SRB is a predominant terminal pathway of organic matter mineralization in marine sediments [4]. SRB are considered strictly anaerobic, however, it is now generally accepted that a large number of strains tolerate the exposure to oxygen for periods of some length [5,6,7]. Some members of the genus Desulfovibrio, even have high rates of aerobic respiration [6, 8]. Some of them can couple this respiratory process to ATP synthesis [9,10,11]. Recently, it has been shown that different Desulfovibrio strains were able to couple respiration with oxygen to growth [12, 13]. In addition, filamentous sulfide-oxidizing bacteria, the so-called “cable-bacteria”, are able to reduce oxygen (or nitrate) at one end of the filament and oxidize sulfide at the opposing end, thereby transporting electrons over cm-scale distances [14,15,16]. They represent the only example of a member of the Desulfobulbaceae that successfully couples the reduction of oxygen with growth, most likely by reverting the canonical sulfate reduction pathway for the oxidation of sulfide.

SRBs constitute over 23 genera, which are found both within archaeal and bacterial domains with Deltaproteobacteria being the dominant class [17]. Recently, some newly discovered bacterial phyla recovered from metagenomic data were found to be SRB [18], suggesting a high phylogenetic diversity of SRB in the environment. However, a large-scale comparison of Deltaproteobacteria genomes revealed that the genomes of cultivated Deltaproteobacteria strains are phylogenetically distant with those Deltaproteobacteria metagenome-assembled genomes (MAGs) [19]. It indicates that the genomic contents of the cultured isolates need further investigation to understand their physiological traits, yet many MAGs are recovered from metagenomic data. By now, only three species, Desulfofaba fastidiosa P2 (DSM 15249) [20], Desulfofaba gelida PSv29 (DSM 12344) [21], and Desulfofaba hansenii P1 (DSM 13527) [22] have been isolated and described within Desulfofaba, the genus within Desulfofabaceae family. The genus Desulfofaba is phylogenetically distinct to other groups of SRBs based on a single gene marker (16S rRNA or DrsA/B). All isolates incompletely oxidizing propionate to acetate and CO2 in the presence of sulfate. Interestingly, phylogenetic trees both based on 16S rRNA gene sequences and on dissimilatory sulfite reductase gene amino acid sequences show that the members of the genus Desulfofaba are closer related to complete oxidizers than to incomplete oxidizers [20,21,22].

However, single-gene based study or specific experimental studies may overlook their metabolic potential without exploring their genomic contents. To get a better overview of the metabolic potential of the genus Desulfofaba, here we present results of the analysis of the genomes of the three species. We focused on the genes that relate to the handling of oxygen and its intermediates, as physiological experiments with Desulfofaba hansenii revealed that the strain was able to respire with oxygen. We also identified genes encoding for the synthesis of polyhydroxybutyrate (PHB) in Desulfofaba hansenii and Desulfofaba gelida, which potentially could be used as electron donor when oxygen is expired in Desulfofaba hansenii. We further investigated the distribution of Desulfofaba based on the homologue search of the marker gene, as well as the collection of our MAGs recovered from various marine environments. This is of particular interest as it may provide some clues regarding the habitat adaptation of Desulfofaba genus to understand their distribution in the environment.

Results and discussion

Phylogeny and distribution of Desulfofaba genus

Desulfofaba genus consists of three species: Desulfofaba fastidiosa P2 (DSM 15249) [20], Desulfofaba gelida PSv29 (DSM 12344) [21], and Desulfofaba hansenii P1 (DSM 13527) [22]. A set of 120 marker genes identified using GTDB-tk and another set of 37 ribosomal protein encoding marker genes identified using Phylosift supported that Desulfofaba is a monophyletic group which is distinct from other families in the order Desulfobacterales (Figs. 1 and S1). A phylogenetic tree based on 16S rRNA gene sequences shows that the genus Desulfofaba is most closely related to the genus Desulfoluna (Fig. S2). The three species of the genus Desulfofaba were isolated from different environments, including the interior of an eelgrass root, polar surface sediments, and the methane-sulfate transition zone 1.5 m below the sediment surface [20]. Related 16S rRNA gene sequences recovered from different geographic locations (Fig. S2) suggests a wide distribution of the genus. However, when we searched 16S rRNA gene sequences against publicly available metagenomes, we could not find samples with high sequence homology (> 95%) to 16S rRNA genes. Additionally, we searched Desulfofaba genus from 194 Desulfobacterales MAGs on phylogenetic trees which were built with protein encoding marker genes extracted using GTDB-tk or Phylosift. Those 194 MAGs were part of our genome collection of over 6000 MAGs. The metagenomes were recovered from various environments, including coastal sediments in the Bohai Sea, cold seep sediments in the South China Sea, and hydrothermal vent sediment in the Southwest Indian Ocean. The search showed that none of them were closely affiliated with members of genus Desulfofaba, suggesting that members of the genus Desulfofaba are either rare in the environment or difficult to amplify by our current approach.

Fig. 1
figure 1

A maximum likelihood phylogenetic tree of 22 genomes including the 3 Desulfofaba genomes. The phylogeny is based on 37 concatenated ribosomal protein encoding genes identified using PhyloSift. Desulfofaba genomes formed a monophyletic group which is distinct from other families in the order Desulfobacterales. Acidobacteria were set as the outgroup

Full size image

Average amino acid identity (AAI) revealed that the genus Desulfofaba is distinct from other described taxa. These three genomes shared at least 60.1% AAI with each other, and at most 57.5% AAI with genomes of other taxa (Fig. S3).

Overview of three Desulfofaba genomes

The overall genomic information of the three draft genomes were summarized in Table 1, and the assignment of genes into COG functional categories demonstrating the general function of three genomes is shown in Table 2. The assembled draft genomes are 99.3, 98.1, and 99.3% complete with less than 2.1% contamination based on CheckM [23] for Desulfofaba fastidiosa, Desulfofaba gelida, and Desulfofaba hansenii, respectively.

Table 1 Genome statistics of three Desulfofaba genomes according to the annotation from IMG
Full size table
Table 2 Number of genes associated with general COG functional categories in three Desulfofaba genomes
Full size table

Complex organic matter degradation

These genomes include genes encoding for diverse carbohydrate-active enzymes (CAZYmes) and peptidases (Fig. S4) with the potential for degradation of complex carbohydrates and detrital proteins into simple sugars and amino acids. Most of the enzymes are assigned with intra-cellular function based on predictions of the cellular localization of the proteins. They also have genes encoding for proteins that are capable of degrading long-chain fatty acids through beta oxidation (Fig. 2), which is consistent with experimental data for Desulfofaba gelida [21]. As an incomplete oxidizer [20,21,22], they are capable of oxidizing propionate to acetate with the methylmalonyl-CoA pathway, which shares several steps with the tricarboxylic acid (TCA) cycle (Fig. 2). The acetate is excreted and may serve as a substrate for adjacent microbes in the environment.

Fig. 2
figure 2

Overview of the metabolic potential of the Desulfofaba genus based on the annotated genomes. Desulfofaba genomes have genes encoding for diverse central metabolic pathways, including glycolysis, the pentose phosphate pathway (PPP), the Wood-Ljungdahl pathway (WLP), the TCA cycle, and the reductive glycine pathway. Desulfofaba genomes have genes involved in nitrogen, sulfur, hydrogen, selenium, and arsenic cycling. Different types of cytochrome oxidases genes were annotated in Desulfofaba genomes

Full size image

Central metabolism and PHB synthesis

Genes encoding central metabolic pathways, including glycolysis, the pentose phosphate pathway (PPP), the Wood-Ljungdahl pathway (WLP), the TCA cycle, and the reductive glycine pathway were identified (Fig. 2, Supplementary Dataset). This suggests a high degree of metabolic versatility of central metabolism in Desulfofaba genus, which enables them to utilize different types of substrates or be active under different environmental conditions.

All three Desulfofaba genomes have a pathway for PHB synthesis from acetyl-CoA [24, 25]. Genes encoding for acetoacetyl-CoA reductase (PhaB), reducing acetoacetyl-CoA to 3-hydroxybutyryl-CoA, were not annotated in any of these three genomes. However, genes encoding for 3-hydroxybutyryl-CoA dehydrogenase, enoyl-CoA hydratase, and 3-hydroxybutyryl-CoA dehydratase, which could reduce acetoacetyl-CoA through 3-hydroxybutanoyl-CoA and crotonoyl-CoA, were annotated in Desulfofaba genomes (Fig. 2).

Sulfur and nitrogen metabolism

The three species within Desulfofaba genus are incomplete oxidizers, i.e., oxidizing propionate incompletely to acetate and CO2 [20,21,22]. The phylogenetic tree of dissimilatory sulfite reductase (DsrAB) genes showed that the three species, together with other identified incomplete oxidizers, formed a monophyletic group (Fig. 3). This phylogenetic position contradicts with the previous study, which showed a closer phylogeny with complete oxidizers based on 16S rRNA gene or DSR sequencing [22]. According to their genetic inventory, they have the potential to reduce polysulfide to sulfide using a sulfhydrogenase and to assimilate thiosulfate into cysteine.

Fig. 3
figure 3

A maximum likelihood phylogenetic tree of genes encoding for alpha and beta subunits of dissimilatory sulfite reductase (DsrAB). DsrAB in Desulfofaba genomes belong to reductive-type and are closely related to sequences in known incomplete oxidizer genomes

Full size image

The three genomes also encode several key pathways for nitrogen cycling (Fig. 2) such as nitrogenase genes involved in the fixation of nitrogen. Also, genes encoding the assimilation of ammonia into glutamine were found. On the dissimilatory side, Desulfofaba gelida encodes genes for nitrate/nitrite assimilation. Hydroxylamine is an intermediate in two important microbial processes of the nitrogen cycle: nitrification [26] and anaerobic ammonium oxidation [27]. The genomes of both Desulfofaba gelida and Desulfofaba hansenii encode enzymes to oxidize hydroxylamine (NH2OH) to nitric oxide (NO) via hydroxylamine dehydrogenase (HAO) [28, 29] or reduce NH2OH to ammonia (NH3) via hydroxylamine reductase (HCP). All genomes encode enzymes to further reduce NO to nitrous oxide (N2O) via anaerobic nitric oxide reductase. N2O is a key byproduct during denitrification and is a potent greenhouse gas and ozone destroying agent. This indicates that Desulfofaba may have important implications for Earth’s climate [30].

Hydrogen metabolism and energy conservation

Hydrogen plays a central role in the energy metabolism of sulfate reducers, which can either use H2 as an energy source or produce H2 during fermentation [31]. The genomes of three species of the genus Desulfofaba encode [NiFe] group 1b hydrogenase, which may transfer H2-liberated electrons through cytochromes to terminal reductase when sulfate, fumarate, nitrate, and metals serve as terminal electron acceptors [32]. The presence of other types of [NiFe] and [FeFe] hydrogenase genes, including cytoplasmic and transmembrane types for hydrogen metabolism, is species-specific (Fig. 2). The phylogenetic position of Desulfofaba hydrogenases is close to hydrogenases of other known sulfate reducers (Figs. S5 and S6). Besides genes encoding for respiratory complex I and II in the electron transport chain, they also encode the membrane-bound Rnf complex that can couple the electron transfer from reduced ferredoxin (Fd2−) to NAD+ with the translocation of proton (H+) for energy conservation [33]. The three species encode F-type ATPase to generate ATP (Fig. 2).

Selenium and arsenate metabolism

The genomes of the three Desulfofaba species have genes encoding for proteins involved in the mobilization of organic selenium. Those proteins catalyze the reaction of selenide with ATP to form selenophosphate via selenide water dikinase (SelD) [34] and incorporate selenium into selenocysteinyl-tRNA (Sec) with L-seryl-tRNA (Ser) selenium transferase (SelA) [35] and selenomethionyl-tRNA (Met) with methionyl-tRNA synthetase (MetG) [36], which are further utilized for protein biosynthesis.

Arsenate and arsenite are toxic to organisms by blocking general cell metabolism [37] and are the two dominant forms of inorganic arsenic in marine environments [38]. Desulfofaba has genes of the arsenic detoxification system (Fig. 2). They can reduce arsenate to arsenite via arsenate reductase (ArsC) through thioredoxin [39]. Even though arsenite is more toxic than arsenate, arsenite could be extruded from the cell by an arsenite transporter (ArsAB) or transformed to methyl arsonate, a less toxic form [40], by arsenite methyltransferase (AS3MT).

Oxygen Consumption and Defense against ROS

Most sulfate reducers are obligate anaerobes, yet some species can tolerate oxygen and have developed different strategies to cope with its presence in the environment [41]. We identified genes encoding two types of membrane-bound oxygen reductases: cytochrome bd-I ubiquinol oxidase and cytochrome cbb3-type terminal oxidase (Fig. 2). The first evidence for a membrane-bound oxygen reductase, a canonical bd quinol oxidase, in SRB was reported in Desulfovibrio gigas [11]. The genes of both identified terminal oxidases in Desulfofaba are of a high-affinity-type, and they are usually considered as important terminal oxidases under low oxygen conditions [42]. Our physiological experiments showed that Desulfofaba hansenii was able to reduce oxygen and that oxygen reduction is most likely linked to the oxidation of PHB storage compounds that are present in the cell (Supplementary material). Many sulfate reducers can reduce O2, probably as a protective mechanism, without sustainable aerobic growth [43]. Growth with energy derived from oxidative phosphorylation linked to oxygen reduction was observed in different Desulfovibrio strains [12, 13]. Based on their genomic outfit, it is possible that Desulfofaba sp. produces energy during aerobic respiration. However, we did not observe aerobic respiration linked to growth in Desulfofaba hansenii. Both Desulfofaba hansenii, isolated from Zostera marina roots, and Desulfofaba gelida, isolated from surface sediments, could encounter oxygen in their respective habitats and thus, the ability to cope with oxygen could be an important survival strategy.

The genomes that we studied contain a number of genes that enable them to deal with oxidative stress, as is the case for many other sulfate reducers [44]. These include genes to detoxify reactive oxygen species (ROS) and repair damaged DNA, as well as genes that trigger a behavioral response to the presence of O2 and ROS. ROS, including superoxide, hydrogen peroxide and hydroxyl radical, are formed during oxygen reduction in the oxygen reduction systems and during non-specific reactions of oxygen with reduced substrates, e.g., transition metals and radical species [45]. The toxicity of O2 in cells is mainly due to cellular damages caused by ROS, such as the oxidation of thiols and the release of metallic centers from proteins leading to the increase of free metals in cytosol. The increased free metals, mainly iron, cause DNA damage through fenton-type reactions that produce ROS [44]. The removal of ROS is important for cells to deal with oxidative stress. We found genes encoding nickel-containing superoxide dismutase that eliminates superoxide radicals through disproportionation into H2O2 and O2 [46] and superoxide reductase that converts toxic O2− into less toxic H2O2 [47]. These genes have been found in other SRB, e.g., in Desulfovibrio genus [48, 49]. In addition, genes encoding enzymes that can decompose H2O2 [47], such as catalase, thiol peroxidase, cytochrome c peroxidase, peroxiredoxin, and rubrerythrin, were found. Catalase is a common enzyme in aerobic organisms that catalyzes the detoxification of H2O2 and has been found in many sulfate reducers both the bacterial and archaeal domains e.g., Desulfovibrio gigas [50] and Archaeoglobus fulgidus [51],

The investigated genomes encode several damage repair systems: (1) thioredoxin, thioredoxin reductase, and glutaredoxin for disulphide bonds reductions; (2) methionine sulfoxide reductase (MsrA/MsrB) for oxidized methionines reduction; and (3) NifU like protein for the Fe-S clusters respiration or biosynthesis (Supplementary Dataset) [44]. The enzymes involved in the damage repair system are metal-free. In contrast, the damage in the detoxification system (superoxide reductase and rubrerythrin) contributes to an increase of Fe2+, leading to the formation of ROS, which requires more repair. Therefore, the system for damage repair is more important than the detoxification system in cells when oxidative conditions are more severe. For example, the expression gene that encodes enzymes of damage repair systems were highly upregulated in Desulfovibrio vulgaris after oxidative stress [52]. Apart from mechanisms dealing with oxidative stress, studied genomes encode methyl-accepting chemotaxis proteins (MCP), which are involved in behavioral strategies to avoid and thereby protect cells from contact with oxygen [10, 43, 53, 54].

Comparative genomics

The three studied genomes share 1557 and 1537 homologues which is at least 1/3 of their genome content based on two algorithms with BDBH and OMCL options (Fig. 4). These homologues have key functions for cell maintenance processes. There are still large portions of species-specific proteins, over 1/3 of proteins in each genome that do not have homologues in the other two species, that may be attributed to the different isolation sources of these three species. The number of proteins in Desulfofaba fastidiosa shared with the other two species were much lower than number of genes shared between Desulfofaba hansenii and Desulfofaba gelida. This potential syntrophic process of anaerobic oxidation of methane coupled to sulfate reduction in the sulfate-methane transition zone may increase the chance for lateral gene transfer, which further contributes to the smaller genome size in Desulfofaba fastidiosa [55]. Genes related with transcription, signal transduction, and secondary metabolite synthesis tend to lose during genome reduction in symbiotic genomes [56]. Moreover, the numbers of transport genes are positively correlated with Desulfofaba genome sizes (Table 2), which is consistent with the universal relationship with genome size [56]. It further suggests that the sulfate-methane transition zone in marine sediment, where Desulfofaba fastidiosa was isolated from, is more different from the other two environments, e.g., seagrass roots and surface sediment.

Fig. 4
figure 4

Shared protein contents among three Desulfofaba genomes based on two different algorithms: a bidirectional best-hits (BDBH) and b orthoMCL algorithm (OMCL). Over 1500 proteins were shared between the three species, and over 1/3 of protein sequences in each genome were species-specific

Full size image

Conclusions

Strains belonging to the Desulfofaba genus were isolated from different environments. Genomic content of this genus showed a high degree of versatility of central metabolic pathways. The diverse central metabolism indicates that isolates have the potential to utilize a wide range of substrates. The presence of different genes involved in sulfur and nitrogen metabolism suggest that they may play a role in various aspects of sulfur and nitrogen cycling. A 16S rRNA gene homologues-based search in publicly available metagenomes and our collection of MAGs from various marine environments indicates a limited environmental distribution of this genus. This, however, does not exclude members of Desulfofaba genus from having a significant role in biogeochemical cycling in their respective habitats. Their ability to respire oxygen and the presence of genes for ROS damage defense allows them to inhabit environments with regular oxygen intrusion, such as the roots of Zostera marina plants from which Desulfofaba hansenii was isolated from. The incomplete oxidation of propionate to acetate provides easily utilized electron donors to other microbes which benefits the entire microbial community.

Methods

Bacteria cultivation and DNA extraction

Desulfofaba fastidiosa P2 (DSM 15249) [20], Desulfofaba gelida PSv29 (DSM 12344) [21], and Desulfofaba hansenii P1 (DSM 13527) [22] were grown as described in the literature [20,21,22]. The cells were harvested in the late exponential phase. DNA was extracted from the cell pellet using the PowerLyser® PowerSoil® DNA extraction kit (MoBio, Carlsbad, CA, USA) according to the manufacturer’s protocol.

Genome sequencing, assembly, annotation, and homologues search

DNA was sequenced on a MiSeq platform. The sequencing data were treated as previously described [57]. Briefly, the sequencing library was trimmed with Trimmomatic-0.36 [58] with the following trimming parameters: CROP:290 HEADCROP:25 SLIDINGWINDOW:4:20. Read quality before and after trimming was assessed by FastQC version 0.11.5 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Reads were assembled using SPAdes 3.10.1 [59]. Contigs shorter than 1,000 bp were removed after assembling. The quality of the assembled genomes was estimated using CheckM v1.1.3 [23]. The draft genome was annotated using the standard operation procedure of the DOE-JGI Microbial Genome Annotation Pipeline (MGAP v.4) supported by the JGI (Berkeley, CA; USA) [60]. The predicted protein sequences from MGAP were further annotated using KofamScan v.1.3.0 with the e-value cut-off of 1e-5 and the highest bit-score larger than the pre-set threshold for each gene [61], and further characterized using KAAS (KEGG Automatic Annotation Server) web server [62] using the ‘Complete or Draft Genome’ setting with parameters: GHOSTX, custom genome dataset, and BBH assignment method.

Key metabolic genes were searched using custom databases. Briefly, peptidases were identified using DIAMOND BLASTP v0.9.31.132 [63] to search against MEROPS Peptidase Protein Sequences (Downloaded on 24th, March, 2022) [64] with the settings: -e 1e-10 –subject-cover 80 –id 50 [65]. Carbohydrate active enzymes (CAZYmes) were identified using the dbCAN v2 standalone tool (CAZYDB.09242021, dbCAN-HMMdb-V10, and dbCAN-fam-aln-V9 databases) [66] with default thresholds. The localization of identified peptidases and CAZYmes was determined using the command-line version of Psort v3.0 using the option –negative for the genomes.

Genes encoding for dissimilatory sulfite reductase (DsrAB) and hydrogenase were further identified using DIAMOND BLSATP v0.9.31.132 [63] to search against different custom databases with the thresholds: -e 1e-10 –subject-cover 70 –id 50; -e 1e-10 –subject-cover 50 –id 30; and -e 1e-10 –subject-cover 50 –id 40 for DsrAB and hydrogenase sequences, respectively. The identified sequences were confirmed with the annotation from MGAP and KO assignment. The identified hydrogenase sequences were further compared with the annotation based on the assigned KO number and the web-based hydrogenase classifier (26th March, 2022) [32].

Homologues between different genomes were searched by GET_HOMOLOGUES [67, 68] with BDBH and OMCL options.

Phylogenetic analysis

A set of 120 marker genes was extracted from the three genomes and reference genomes using the Genome Taxonomy Database (GTDB)-Tk v1.7.0 (release 202) [69]. Another set of 37 single-copy, protein-coding housekeeping genes was extracted using Phylosift v1.0.1 [70]. The two sets of marker genes were separately concatenated, and aligned using MAFFT v7.450 [71] with the setting –maxiterate 1000 –localpair, trimmed using trimAl v1.2rev59 [72] with the -gappyout option, and manually checked. The two refined alignments were used to generate two maximum-likelihood trees using RAxML v8.2.4 [73] with the parameters: raxmlHPC-PTHREADS-AVX -m GTRGAMMA -N autoMRE -p 12345—× 12345. Amino acid identity (AAI) of the three genomes and reference genomes was estimated using CompareM (v0.1.2) AAI workflow (‘comparem aai_wf’, https://github.com/dparks1134/CompareM).

16S rRNA sequences were identified using Barrnap v0.9 (https://github.com/tseemann/barrnap) with the default settings, aligned, and manually curated in ARB [74] with the SILVA SSURef NR99 database (release 138). The alignment was exported to generate a maximum-likelihood tree using IQ-TREE v1.6.12 [75] with the settings: -m MFP -bb 1000 -bnni -alrt 1000.

The identified DsrA and DsrB sequences were separately aligned with reference sequences using MAFFT v7.450 with the settings: –maxiterate 1000 –globalpair –anysymbol, trimmed using trimAl v1.2rev59 [72] with the -gappyout option, manually checked, and concatenated. The maximum-likelihood tree was generated using RAxML v8.2.4 [73] with the parameters: raxmlHPC-PTHREADS-AVX -m GTRGAMMA -N autoMRE -p 12345—× 12345.

The final identified hydrogenase sequences with selected references for different types of hydrogenases [76] were aligned using ClustalW v2.1 [77], and the Neighbor-Joining tree was generated using MEGA X [78] under p-distance model with 1,000 bootstrap.