Antibiotic-Resistant Desulfovibrio Produces H2S from Supplements for Animal Farming

Sulphate-reducing bacteria, primarily Desulfovibrio, are responsible for the active generation of H2S in swine production waste. The model species for sulphate reduction studies, Desulfovibrio vulgaris strain L2, was previously isolated from swine manure characterized by high rates of dissimilatory sulphate reduction. The source of electron acceptors in low-sulphate swine waste for the high rate of H2S formation remains uncertain. Here, we demonstrate the ability of the L2 strain to use common animal farming supplements including L-lysine-sulphate, gypsum and gypsum plasterboards as electron acceptors for H2S production. Genome sequencing of strain L2 revealed the presence of two megaplasmids and predicted resistance to various antimicrobials and mercury, which was confirmed in physiological experiments. Most of antibiotic resistance genes (ARG) are carried by two class 1 integrons located on the chromosome and on the plasmid pDsulf-L2-2. These ARGs, predicted to confer resistance to beta-lactams, aminoglycosides, lincosamides, sulphonamides, chloramphenicol and tetracycline, were probably laterally acquired from various Gammaproteobacteria and Firmicutes. Resistance to mercury is likely enabled by two mer operons also located on the chromosome and on pDsulf-L2-2 and acquired via horizontal gene transfer. The second megaplasmid, pDsulf-L2-1, encoded nitrogenase, catalase and type III secretion system suggesting close contact of the strain with intestinal cells in the swine gut. The location of ARGs on mobile elements allows us to consider D. vulgaris strain L2 as a possible vector transferring antimicrobials resistance determinants between the gut microbiote and microbial communities in environmental biotopes.


Introduction
Studies of antibiotic resistance genes (ARG) have mainly focused on pathogenic and commensal bacteria and paid less attention to environmentally relevant microorganisms. Sulphate-reducing bacteria (SRB) are a functional guild of prokaryotes of significant environmental importance due to the end product of their metabolism, hydrogen sulfide. Anaerobic SRB are common inhabitants of natural and technogenic ecosystems, such as marine and freshwater sediments, groundwater, soils and wastewater treatment facilities. SRB, primarily members of the genus Desulfovibrio, also occur in the animal and human intestine [1][2][3][4][5][6][7][8]. Possible associations of SRB with pathologies, including ulcerative colitis [3,9,10], inflammatory bowel disease [11], colorectal cancer [12,13], bacteremia [4,[14][15][16][17]], brain abscess [18] and renal infection [19], are discussed in the scientific literature. A recent study revealed a pathogenic effect of Desulfovibrio in the gut on fatty liver in diet-induces obese mice and children with obesity [20]. Little is known about resistance to antibiotics in sulphate-reducing bacteria and the reported cases refer to clinical practice. D. desulfuricans isolated from an immunocompetent patient with bacteremia was susceptible to penicillin

Physiological Tests
The L2 strain was isolated from the settlement lagoon at a large swine finishing facility, as described previously [29]. Briefly, cumulative manure slurry from the settlement lagoon was used as the inoculum for pure culture isolation. Liquid Widdel-Bak (WB) medium with formate as an electron donor supplemented with zero-valent iron was chosen to prevent overgrowth of other heterotrophic microorganisms. Cell morphology was observed by phase contrast microscopy using an Axio Imager A1 microscope and by transmission electron microscopy (TEM) of ultra-thin sections prepared as previously described [37].
Physiological tests were carried out in the modified WB medium that contained (per litre) 4.0 g Na 2 SO 4 , 0.2 g KH 2 PO 4 , 0.25 g NH 4 Cl, 1 g NaCl, 0.4 g MgCl 2 ·6H 2 O, 0.5 g KCl, 0.113 g CaCl 2 , 2 mL of vitamin solution, 1 mL of trace element solution, 1 mL each of Na 2 SeO 3 (final concentration of 23.6 µM) and Na 2 WO 4 (24.2 µM) solutions. Vitamin and trace element solutions were prepared as described by Widdel and Bak [38]. Each cultivation vial received an iron wire (100% Fe) as described previously [39,40]. Lactate (18 mM) was used as an electron donor. Growth was tested at incubation temperatures in the range of 4-50 • C. Growth was determined by microscopic cell counts in triplicate samples. Specific growth rates were calculated from the cell counts during the exponential phase of growth. Further physiological experiments were conducted at 37 • C.
The tested soluble electron acceptors were the following: 28 mM sodium sulfate, 2 mM sodium sulfite, 20 mM sodium sulfite, 20 mM sodium thiosulfate, 10 mM fumarate and 5 mM calcium nitrate. Solid-phase electron acceptors, gypsum and gypsum plasterboard, were sterilized as powder by autoclaving. Each 12 mL culture vial received 100 mg of gypsum or gypsum plaster board. X-ray diffraction did not reveal impurities in the used gypsum specimen ( Figure 1A). Plasterboard contained gypsum as the dominant phase, an alumosilacte, albite ( Figure 1B). All the acceptors were tested in the presence of 18 mM lactate.
Genomic DNA of strain L2 was additionally sequenced on the MinION (Oxford Nanopore, Oxford, UK) using a Ligation Sequencing Kit 1D protocol and an R9.4 flow cell (FLO-MIN106). Sequencing resulted in 74,516 reads with a total length of~552 Mbp. De novo assembly of MinION reads was performed using Flye v.2.7b [42], Illumina reads were mapped to the assembly and used for the correction with three iterations of Pilon v.1.22 software [43]. Three circular contigs of 3,611,640 bp, 175,461 bp and 109,048 bp, representing a chromosome and two plasmids were obtained. Gene search and annotation were performed using the RAST server 2.0 [44].
Genomic DNA of strain L2 was additionally sequenced on the MinION (Oxford Nanopore, Oxford, UK) using a Ligation Sequencing Kit 1D protocol and an R9.4 flow cell (FLO-MIN106). Sequencing resulted in 74,516 reads with a total length of ~552 Mbp. De novo assembly of MinION reads was performed using Flye v.2.7b [42], Illumina reads were mapped to the assembly and used for the correction with three iterations of Pilon v.1.22 software [43]. Three circular contigs of 3,611,640 bp, 175,461 bp and 109,048 bp, representing a chromosome and two plasmids were obtained. Gene search and annotation were performed using the RAST server 2.0 [44].

Desulfovibrio vulgaris L2 Physiology and Resistance to Antibiotics
Cells of strain L2 are slightly curved rods, 1.6-3.0 µm long, and 0.4-0.6 µm wide (Figure 2A,B). The 16S rRNA gene sequence of strain L2 was 99.78% similar to that of Desulfovibrio vulgaris Hildenorough (recently reclassified as Nitratidesulfovibrio vulgaris [45]). The sequence similarity was above of 98.7%, the species boundary cutoff, assuming that L2 is a novel strain of D. vulgaris. The 16S rRNA sequences of strain L2 was 100% similar to that of Desulfovibrio vulgaris DP4, isolated from heavy metal-impacted lake sediment [46]. The phylogenetic position of strain L2 was confirmed by its genome analysis. The average nucleotide identity (ANI) between L2 genome and D. vulgaris Hildenorough was 99.06%, a value above the species boundary cutoff of 95% [47,48]. Strain L2 was capable of growth at temperatures between +15 and +45 • C with an optimum temperature of 37 • C. Strain L2 used lactate, pyruvate, fumarate, succinate, ethanol, glycerol, peptone and tryptone as electron donors for sulphate reduction. Limited growth was observed with malate, fructose and formate. The strain did not grow with citrate, butyrate, propionate, benzoate, palmitate, glucose, sucrose, butanol and propanol.
average nucleotide identity (ANI) between L2 genome and D. vulgaris Hildenorough was 99.06%, a value above the species boundary cutoff of 95% [47,48]. Strain L2 was capable of growth at temperatures between +15 and +45 °C with an optimum temperature of 37 °C. Strain L2 used lactate, pyruvate, fumarate, succinate, ethanol, glycerol, peptone and tryptone as electron donors for sulphate reduction. Limited growth was observed with malate, fructose and formate. The strain did not grow with citrate, butyrate, propionate, benzoate, palmitate, glucose, sucrose, butanol and propanol. vulgaris strain L2 inferred using neighbor-joining inference on the 120 concatenated single copy marker proteins. The optimal tree is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Poisson correction method and are in the units of the number of amino acid substitutions per site. All ambiguous positions were removed for each sequence pair (pairwise deletion option). There was a total of 5040 positions in the final dataset. Evolutionary analyses were conducted in MEGA11.
Sulfite, thiosulfate, nitrate and fumarate, in addition to sulphate, could be used as electron acceptors. Strain L2 could grow with gypsum, gypsum plasterboard, and L-lysine sulphate as the sole electron acceptor with the specific growth rates of 0.15-0. vulgaris strain L2 inferred using neighbor-joining inference on the 120 concatenated single copy marker proteins. The optimal tree is shown. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Poisson correction method and are in the units of the number of amino acid substitutions per site. All ambiguous positions were removed for each sequence pair (pairwise deletion option). There was a total of 5040 positions in the final dataset. Evolutionary analyses were conducted in MEGA11.
Sulfite, thiosulfate, nitrate and fumarate, in addition to sulphate, could be used as electron acceptors. Strain L2 could grow with gypsum, gypsum plasterboard, and Llysine sulphate as the sole electron acceptor with the specific growth rates of 0.15-0.18 h −1 ( Figure 3A). Sulphide formation from gypsum and plasterboard reached 3.77 ± 0.24 and 3.56 ± 0.18 mM, respectively ( Figure 3B). The maximum sulphide concentration formed from L-lysine sulphate was 2.12 ± 0.19.

General Genome Properties
The genome of strain L2 consists of a chromosome of 3,611,640 bp and two circular plasmids 175,461 and 109,048 nucleotides long, designated pDsulf-L2-1 and pDsulf-L2-2, respectively. The chromosome contains all the genes required for dissimilatory sulfate reduction in Desulfobacterota, namely, sulfate permease, sulfate adenylyl transferase (sat), adenylsulfatereductase (aprAB) and dissimilatory sulfitereductase (dsrABD and dsrC). Also present are the genes for the adenylsulfatereductase-associated electron transport complex, QmoABC, heterodisulfidereductase, and the sulfite reductase-associated electron transport proteins, DsrMKJOP. The presence of several molybdopterin-dependent oxidoreductases of the Psr family is consistent with the ability of strain L2 to use alternative electron acceptors for anaerobic respiration. Moreover, the genome of strain L2 also contains cytochrome c oxidase, which can be involved not only in oxygen detoxification, but also in respiration under microaerophilic conditions.
The larger plasmid, pDsulf-L2-1, was highly similar (>97% nucleotide sequence identity over >95% length) to plasmids found in D. vulgaris strains RCH1 (plasmid pDEVAL01), Hildenborough (plasmid pDV) and DP4 (plasmid pDVUL01). These strains are closely related to L2 ( Figure 2C) and belong to the same species. pDsulf-L2-1 contains Type IC CRISPR/Cas system with 32 spacer-repeat units and all the essential genes of a type III secretion system. Type III secretion systems are commonly used by pathogenic bacteria to inject effector proteins directly into the host cell cytoplasm to influence host responses [49]. Since the L2 strain presumably originates from the gut of animals, this plasmid may be involved in the interaction of these bacteria with intestinal epithelial cells. Considering metabolically important functions, the plasmid contains the MoFe nitrogenase genes, the katE catalase gene and the chrBA chromate resistance operon.
Plasmids similar to pDsulf-L2-2 have not been found in any of the Desulfovibrio species. Although sequences highly similar to some regions of pDsulf-L2-2 (about 10% of the total length) were found in various gamma-proteobacteria, a GeneBank search did not reveal sequences similar to the entire pDsulf-L2-2. The plasmid encodes several site-specific recombinases, transposases and toxin-antitoxin systems, as well as proteins involved in conjugative transfer and mobilization.

Genetic Determinants of Antibiotic Resistance
A notable feature of the pDsulf-L2-2 plasmid is the presence of two gene clusters that can confer antibiotic resistance. The first is a class 1 integron [50] containing six resistance genes ( Figure 4). A search in the CARD database predicted that they encode the betalactamase of the OXA-2 family conferring resistance to beta-lactams such as carbapenem, cephalosporin and penams, the aminoglycoside nucleotidyltransferase of the ANT(3") family responsible for the inactivation of aminoglycosides, the lincosamidenucleotidyltransferase (LNU), the QacE efflux pump of the MFS superfamily conferring resistance to antiseptics, sulfonamide-resistant dihydropteroate synthase Sul1, and acetyltransferase(act) distantly related to SAT-2 streptothricinacetyltransferases and aminoglycoside acetyltransferase of the AAC(6 ) family. Close homologues of these genes, including 100% identical integronintegrases, have been found in gamma-proteobacteria (Escherichia, Salmonella, Pseudomonas, etc.). The second closely located locus contains genes for the regulatory protein TetR and the tetracycline efflux pump Tet(C).
The chromosome of strain L2 also contains a class 1 integron with an integrase completely identical to that of the pDsulf-L2-2 integron (Figure 4). However, the genes content of this integron is different, and its genes seem to be obtained from different sources. The first gene encodes an aminoglycoside nucleotidyltransferase of the ANT(3") family; it differs from the ANT(3") gene located on the plasmid, but a GeneBank search revealed identical genes in some Enterobacteriaceae. The integron is then interrupted by an insert containing the oppositely oriented genes encoding the tetracycline-resistant ribosomal protective protein Tet(M), a conjugal transfer protein, two transposases, and the 23S rRNA methyltransferase CfrC conferring resistance to linezolid and phenicol antibiotics. Genes identical to tet(M) and cfrC were found in Firmicutes, Streptococcus pyogenes HKU419, and Clostridioides difficile 020696, respectively. The next is the second copy of the first gene of this integron, ANT(3"), it is followed by catB chloramphenicol acetyltransferase gene. This region is identical to the p035_A-VIM-1 plasmid of Klebsiella aerogenes 035. The next three genes are identical to the last three genes of the pDsulf-L2-2 plasmidintegron, qacE, sul1, and act.  The chromosome of strain L2 also contains a class 1 integron with an integrase completely identical to that of the pDsulf-L2-2 integron (Figure 4). However, the genes content of this integron is different, and its genes seem to be obtained from different sources. The first gene encodes an aminoglycoside nucleotidyltransferase of the ANT(3″) family; it differs from the ANT(3″) gene located on the plasmid, but a GeneBank search revealed identical genes in some Enterobacteriaceae. The integron is then interrupted by an insert containing the oppositely oriented genes encoding the tetracycline-resistant ribosomal protective protein Tet(M), a conjugal transfer protein, two transposases, and the 23S rRNA methyltransferase CfrC conferring resistance to linezolid and phenicol antibiotics. Genes identical to tet(M) and cfrC were found in Firmicutes, Streptococcus pyogenes HKU419, and Clostridioides difficile 020696, respectively. The next is the second copy of the first gene of this integron, ANT(3″), it is followed by catB chloramphenicol acetyltransferase gene. This region is identical to the p035_A-VIM-1 plasmid of Klebsiella aerogenes 035. The next three genes are identical to the last three genes of the pDsulf-L2-2 plasmidintegron, qacE, sul1, and act.
Another chromosomal locus contains genes for the regulatory protein TetR and the tetracycline efflux pump Tet(C). The sequences of these genes differed from the corresponding genes of the plasmid pDsulf-L2-2. However, a nearly identical region was found on the pDsulf-L4 plasmid of D. desulfuricans strain L4 isolated from the same manure storage lagoon.

Genetic Determinants of Resistance to Mercury
Plasmid pDsulf-L2-2 contains a mercuric resistance operon merRTPADE containing the genes for the regulatory protein MerR, the transport protein MerT, the periplasmic Hg 2+ -binding protein MerP, the mercuric ion reductase MerA, the transcription regulator MerD, the transport protein MerE and the protein containing the EAL domain. Such mer operons allow bacteria to detoxify Hg 2+ by enzymatic reduction into volatile metallic mercury [51]. In plasmid pDsulf-L2-2 the mer operon is probably part of a transposon, since it is followed by genes for the resolvase and the Tn3 family transposase. Sequenced genomes of other Desulfovibrio species lacked similar mer operons, but gene clusters with identical nucleotide sequences are present in the chromosomes and plasmids of several species of Pseudomonas and Aeromonas.
Interestingly, the second mer operon is located on the chromosome of the L2 strain close to the tetR-tetC locus. This operon has a similar gene order and content (mer-RTPCADE), with the exception of an additional merC gene for the mercury transport protein located between merP and merA. The nucleotide sequence identity between the plasmid and chromosomal mer operons is only 81-85%, indicating their independent acquisi- Another chromosomal locus contains genes for the regulatory protein TetR and the tetracycline efflux pump Tet(C). The sequences of these genes differed from the corresponding genes of the plasmid pDsulf-L2-2. However, a nearly identical region was found on the pDsulf-L4 plasmid of D. desulfuricans strain L4 isolated from the same manure storage lagoon.

Genetic Determinants of Resistance to Mercury
Plasmid pDsulf-L2-2 contains a mercuric resistance operon merRTPADE containing the genes for the regulatory protein MerR, the transport protein MerT, the periplasmic Hg 2+binding protein MerP, the mercuric ion reductase MerA, the transcription regulator MerD, the transport protein MerE and the protein containing the EAL domain. Such mer operons allow bacteria to detoxify Hg 2+ by enzymatic reduction into volatile metallic mercury [51]. In plasmid pDsulf-L2-2 the mer operon is probably part of a transposon, since it is followed by genes for the resolvase and the Tn3 family transposase. Sequenced genomes of other Desulfovibrio species lacked similar mer operons, but gene clusters with identical nucleotide sequences are present in the chromosomes and plasmids of several species of Pseudomonas and Aeromonas.
Interestingly, the second mer operon is located on the chromosome of the L2 strain close to the tetR-tetC locus. This operon has a similar gene order and content (merRTPCADE), with the exception of an additional merC gene for the mercury transport protein located between merP and merA. The nucleotide sequence identity between the plasmid and chromosomal mer operons is only 81-85%, indicating their independent acquisition by lateral transfer rather than transfer between pDsulf-L2-2 and the chromosome. Identical regions spanning both mer and tetR-tetC were found in plasmids and chromosomes of various Enterobacteriaceae (Salmonella enterica, Escherichia coli, Klebsiella pneumonia, Klebsiella enterica, etc.), indicating a single acquisition event.

Discussion
The results of our study suggest that D. vulgaris L2 isolated from swine manure originates from the intestinal microbial community. Recent metagenome studies provide evidence that Desulfovibrio is the dominating group of sulphate reducers in the swine intestine. Desulfovibrio were revealed to be the most abundant SRB in piglet cecum by using DsrA-targeted analysis [32]. At the species level, D. intestinalis was the predominant SRB in Meishan and Yorkshire piglets. Desulfovibrio piger was the second abundant SRB in Meishan piglets. The authors hypothesized that the Desulfovibrio role in the gut was intestinal hydrogen removal. A study of the pig intestinal microbiome using fecal samples sequencing and composite genome (MAG) assembly showed that the genes encoding dissimilatory sulfite reductase were only identified in MAGs belonging to the Desulfovibrionaceae [33]. SRB have long been recognized as environmentally relevant prokaryotes and major players in carbon and sulphur cycles [28]. There is a growing amount of evidence that SRB, primarily Desulfovibrio, is an important constituent of microbial community in the animal intestine [1][2][3][4][5][6][7][8].
The involvement of Desulfovibrio in various pathologies has been documented, but the mechanisms remain largely unknown [8,17]. Nie with co-authors [52] showed that D. fairfieldensis damages epithelial barriers and activate inflammation and pyroptosis in macrophages via outer membrane vesicles. The occurrence of type III secretion system (T3SS) on the pDsulf-L2-1 plasmid, revealed in our study, suggests its possible close contact with intestinal epithelial cells. T3SS is used by a number of pathogens, such as Escherichia coli, Salmonella, Yersinia and Pseudomonas [53], as well as commensal bacteria, including Rhizobium [54], to deliver effector proteins into eukaryotic host cells. A ncbi blastp search reveals a number of ATPase SctN closely related to that found in pDsulf-L2-1 in other Desulfovibrio species. Catalase, which plays a role in oxidative stress, has been identified as a putative T3SS effector shared between plant and animal pathogens [55]. The catalase occurrence on pDsulf-L2-1 close to T3SS corresponds to a possible D. vulgaris contact with epithelial cells.
Until now, the resistance determinants to tetracycline and other antimicrobials applied in animal husbandry have not been detected in SRB. Several reports of Desulfovibrio susceptibility to clinically relevant antibiotics are based on physiological tests [4,21,22] and do not decipher the occurrence of antibiotic resistance determinants in species genomes.
The observed resistance of strain L2 to a number of antibiotics is most likely determined by laterally acquired resistance genes identified on the chromosome and plasmid pDsulf-L2-2. Aminoglycoside nucleotidyltransferase of the ANT(3") family and aminoglycoside acetyltransferase of the AAC(6 ) family are probably responsible for resistance to streptomycin, kanamycin and gentamicin. The observed resistance to ampicillin is probably due to beta-lactamases of the OXA-2 family. All of these resistance genes are located on a class 1 integron. Such integrons play a key role in the spread of antibiotic resistance because they can capture and express various resistance genes and are often located on plasmids and transposons facilitating their lateral transfer [50,56]. The observed resistance to tetracycline is likely determined by the efflux pump Tet(C) and the ribosomal protection protein Tet(M).
All of these resistance determinants were acquired by horizontal gene transfer from various Gamma-proteobacteria and Firmicutes, as evidenced by the presence of identical genes in these lineages and the absence of their close homologues in other Desulfovibrio species. The only exception is a region of about 2.4 kb containing tetR-tetC, which is common to the chromosome of strain L2 and the plasmid pDsulf-L4 of D. desulfuricans strain L4 isolated from the same site. This short (10876 bp) plasmid additionally contains strA-strB (aminoglycoside phosphotransferase) streptomycin resistance genes and the sul2 dihydropteroate synthase sulfonamide resistance gene [29], which are absent in strain L2. Since the tetR-tetC locus in the L2 strain chromosome is linked to the mer operon, and the entire cluster was obtained from a member of the Enterobacteriaceae family, its independent acquisition by the L4 strain from the same donor is the most plausible scenario, although transfer from the L2 strain chromosome to the plasmid pDsulf-L4 of strain L4 cannot be ruled out.
Strain L2 belongs to D. vulgaris, a model SRB. The D. vulgaris Hildenborough strain was isolated by J. R. Postgate in 1946 from clay near Hildenborough, Kent (UK) [24]. Two other strains with available genomes originate from arsenic contaminated soil (DP4) [46] and chromium bioremediation site in Hanford, WA [57]. All these strains contain plasmids similar to pDsulf-L2-1, but only the L2 strain harbors an additional resistance plasmid.
H 2 S is considered as one of the most important pollutants associated with livestock production [58]. The high sulphate reduction rates by Desulfovibrio measured with a radioactive tracer in manure slurry at a large swine finishing facility [29] left open the question of the source of the electron acceptor for H 2 S production. Our experiments with D. vulgaris L2 have shown that the gypsum and gypsum plasterboard used for animal bedding constitute a large reservoir of solid-phase sulphate that is used by Desulfovibrio and other SRB for their metabolism. The L2 strain grows at the same rate and produces the same amount of H 2 S from gypsum or gypsum plasterboard as it does from soluble sulphate.
Currently, most of the supplemental lysine used in pig diets is in the form of hydrochloride, which contains 78.8% of lysine; more recently, lysine sulphate containing ≥54.6% of lysine has been introduced as an alternative source of supplemental lysine [31]. Copper sulfate (CuSO 4 ) has been used as a supplement to reduce the incidence of diarrhea and improve growth performance of piglets over the past decades [59]. L-lysine sulphate and other food supplements formulated as sulphate salts can increase the H 2 S production in the animal gut.

Conflicts of Interest:
The authors declare that there are no conflict of interest.