Introduction

Campylobacter concisus is a Gram-negative ε-proteobacterium that was first isolated by Tanner and coworkers in 1981 from human gingival crevices of a patient with gingivitis1. C. concisus is commonly found in the oral environment of healthy individuals2 although it is not considered to be a dominant oral species. It is frequently associated with periodontitis, gingivitis and other dental diseases3. A recent study also found elevated levels of C. concisus in the microbiome of potentially malignant oral leukoplakia4. The presence and spectrum of action of C. concisus are not strictly limited to the oral cavity though. Indeed, data collected within the last 20 years indicate that C. concisus can be found throughout the entire gastrointestinal tract (GIT), including (i) the esophagus: high levels of C. concisus were found in 57% of patients with Barrett’s esophagus (BE) syndrome (but none in the control subject), suggesting a link between presence of the bacterium and BE5; (ii) the gastric mucosa: C. concisus is highly active within the gastric fluid (an increase of 444% compared to the total microbiota), irrespective of pH6. Furthermore, C. concisus pathotypes are present at significant levels in patients with gastroenteritis7; (iii) the intestines, including the ileum, jejunum, cecum and rectum3. Besides, higher prevalence of C. concisus were found in children with Crohn’s disease, as well as in adults with inflammatory bowel disease (IBD)8. In addition, C. concisus has been shown to be associated with intestinal pathogenicity in immunocompromised patients9. While they are phenotypically indistinguishable, most C. concisus strains show high degree of genetic variability, irrespective of their preferred niche (oral or enteric) or the diseases they cause. Thus, C. concisus strains can be classified in two main genomospecies, based on various typing methods that include amplified fragment length polymorphisms (AFLP)10,11, 23 S rRNA PCR11 and multi-locus sequence typing (MLST)12,13,14. The latter method has proven to be also useful to discriminate C. concisus against other emerging Campylobacter species15.

The distribution of C. concisus throughout the entire GIT suggests a highly adaptable metabolism, as well as versatile respiratory pathways. Indeed, Tanner et al. first described the organism as “being able to grow under both microaerophilic (5% O2) or anaerobic conditions”, with “formate and H2 used as energy sources”1. Since then, C. concisus has been systematically described as a H2-requiring microorganism16,17, and the use of H2-enriched gas mixture, and more specifically H2-enriched microaerobic gas mixtures have become standard practice to grow C. concisus, highlighting the importance of H2 (and by extension, hydrogenases) in the pathogen’s metabolism. Results from Lee and coworkers indicated that H2 is required for C. concisus to grow under microaerobic conditions, as none of the 63 C. concisus strains tested in the study were able to grow on plates unless H2 was added18. The same study found that most C. concisus strains tested could grow under anaerobic conditions without H2, however presence of H2 in the gas mixture significantly enhanced the pathogen’s growth18. Thus, it appears that H2 is required for optimal growth of C. concisus, especially in presence of microaerobic O2 levels.

Based on analysis of multiple C. concisus genome sequences, the pathogen appears to possess two hydrogenase operons. One, annotated as “hyd”, encodes for a hydrogenase that shares high sequence homology with H2-uptake type hydrogenases found in other ε-proteobacteriae, such as Helicobacter pylori19, Helicobacter hepaticus20, Campylobacter jejuni21 or Wolinella succinogenes22; the second one, annotated as “hyf”, encodes for a putative H2-evolving type hydrogenase similar to Hyd-3 and Hyd-4 complexes found in E. coli23. Together, Hyd-3 and formate dehydrogenase H (FDH-H) form the formate hydrogenlyase (FHL) complex which disproportionates formate to H2 and CO2 under mixed acid fermentative conditions in E. coli24. Although the exact role of Hyd-4 remains elusive, its subunit composition and its homology to Hyd-3 suggest it can also form a FHL-like complex, called FHL-225. Both FHL and FHL-2 are structurally related to the NADH dehydrogenase complex I of the respiratory chain26. In addition, C. concisus possesses a third operon (“hyp”), with genes encoding putative hydrogenase accessory proteins needed for maturation of both hydrogenases. The hyp operon is located on the same locus as the hyd operon.

In the present study, we aimed at generating mutants in genes belonging to each of the three aforementioned operons, e.g. hyd, hyf and hyp, a technical challenge since, to our knowledge, no C. concisus mutant has been reported yet. While attempts using conventional methods (electroporation or natural transformation) failed to deliver mutants, methylation treatment of the target DNA (using C. concisus cell-free extracts) prior to transformation proved successful. The construction and characterization of H2-evolving hydrogenase hyfB and tetrathionate reductase ttrA mutants show it is possible to inactivate genes in C. concisus by site-directed mutagenesis. Our results highlight the diversity of respiratory pathways in general and the importance of H2 metabolism in particular in this emerging pathogen.

Results

Analysis of C. concisus genome sequence reveals a versatile respiratory system that includes two hydrogenases

Genome sequence analysis of C. concisus ATCC strains 13826 (also known as BAA-1457) and 5156227 revealed the presence of full sets of genes needed for aerobic (microaerophilic) as well as anaerobic respiration (Fig. 1 and data not shown). Indeed, based on its genome sequence, it appears the pathogen can use a variety of electrons donors such as succinate, formate, hydrogen, and NADH, while the list of putative electron acceptors includes oxygen, fumarate, nitrogen-containing compounds (nitrate, nitrite, nitric oxide, nitrous oxide, possibly trimethylamine N-oxide) and sulfur-containing compounds, including tetrathionate, thiosulfate, and possibly dimethyl sulfoxide (Fig. 1).

Figure 1
figure 1

Putative respiratory pathways of C. concisus. Putative genes encoding for structural subunits of each respiratory enzyme complex are shown. Electron donors are shown in green and electron acceptors are shown in blue. Locus tag numbers and gene annotations refer to strain 13826 (BAA-1457), according to the JGI-IMG/M website (img.jgi.doe.gov). FHL: formate-hydrogenlyase complex. FHL-2: formate-hydrogenlyase 2 complex NADH: Nicotinamide Adenine Dinucleotide. BSO: biotin sulfoxide. DMSO: dimethyl sulfoxide. DMS: dimethyl sulfide. TMAO: trimethylamine N-oxide. TMA: trimethylamine.

Full size image

Formate oxidation appears to be driven through two different formate dehydrogenases (FDH) in C. concisus: the first one has homology to FDH-N (or FDH-O), known to couple formate oxidation to nitrate reduction along with nitrate reductases28 and the second one has homology to FDH-H, usually found as part of the FHL complex23; both types of FDH are found in Enterobacteriaceae. Two sets of hydrogenase genes are present in C. concisus (Figs 1 and 2). Hyd genes encode for subunits of an H2-uptake hydrogenase, while hyf-annotated genes encode for a H2-evolving type 3 or 4 hydrogenase, the hydrogenase part of the FHL (or FHL-2) system in E. coli23. Indeed, the C. concisus “hyf” operon contains both hyf (ABCEFGHI) and hyc (HI) genes (Fig. 2) that are found in operons involved in Hyd-3 and Hyd-4 (H2-evolving) biosynthesis, respectively23. A study from Kovach et al. identified HyfI as being among the most immunoreactive proteins in C. concisus29. A 14-gene operon encoding for a full NADH dehydrogenase type I respiratory complex can be found. Two different sets of genes encoding for putative fumarate reductase (Frd)/succinate dehydrogenase (Sdh) enzyme complexes are present. The Cc13826_0424-0426 complex, herein annotated as SdhABC, is highly similar (80% identity) to C. jejuni Cj0408-410, previously shown to be a bifunctional Frd/Sdh enzyme30, while the Cc13826_1281-1283 complex (MfrABE) shares high homology with C. jejuni MfrABE (Cj437-0439), shown to have fumarate reductase activity only30. Regarding O2 respiration, C. concisus appears to have a branched respiratory chain, based on the presence of genes encoding for two terminal cytochrome oxidases, a cbb3-type and a bd-type quinol oxidase, respectively, similar to those found in C. jejuni31. Genes encoding for enzymes involved in respiration of various nitrogen compounds are present in all C. concisus strains: those include a periplasmic nitrate reductase (Nap), a nitrous oxide reductase (Nos), a nitric oxide reductase (Nor) and a putative periplasmic cytochrome c nitrite reductase (Nrf), although the latter is also hypothesized to be a polysulfite reductase27,32. In addition, C. concisus strains possess three sets of genes encoding for membrane bound, molybdenum (Mo)- or tungsten (W)-containing periplasmic enzymes that could be associated with either DMSO, TMAO or BSO respiration, as suggested by the concomitant presence of a Twin Arginine Translocation (TAT) signal peptide and a Mo/W-Bis-PGD binding motifs in their sequence. Finally, C. concisus has the capacity to respire sulfur-containing compounds, based on the presence of genes encoding for tetrathionate reductase (Ttr), thiosulfate reductase (Tsr), dissimilatory sulfide reductase (Dsr) and possibly sulfite reductase (Nrf), as discussed above (Fig. 1). The noticeable presence of high levels of hydrogen sulfide (H2S), one of the biochemical hallmarks of C. concisus1, confirms that the sulfur respiration pathway is operational. To coordinate these various pathways, C. concisus can rely on several putative transcriptional regulators, including putative CRP/FNR (13826_2145), NikR (13826_0355), Fur (13826_1795) and CsrA (13826_0062) regulatory proteins. Genome sequence analysis of the C. concisus GS2 strain 51562 confirmed the presence of all genes described above (data not shown), with the notable exception of the FDH-H gene (fdhf) homolog.

Figure 2
figure 2

Genome location and organization of the hyp, hyd, hyf and ttr genes in C. concisus 13826 (BAA-1457). Genes annotations are according to the JGI-IMG/M website (img.jgi.doe.gov). Putative gene names are indicated above each gene. Numbers below each gene box indicate locus tag numbers. Genes targeted in this study (hypE, hydB, hyfB, ttrA) are indicated by dashed arrows. An approximate scale is shown bottom left.

Full size image

H2 is required for optimal growth of C. concisus strain 13826 and 51562 both under anaerobic and microaerophilic conditions

Previous results indicated that H2 plays a major role in C. concisus growth and we aimed at confirming these results with the two strains studied herein, using liquid cultures and well-controlled gas atmospheres. To determine the effect of H2 on the anaerobic or microaerophilic growth of strains 13826 and 51562, cells were inoculated in brain-heart infusion supplemented with fetal calf serum (BHI-FCS) liquid cultures in 165-mL bottles, with headspaces filled with four different gas atmospheric conditions (Fig. 3). After 24 h incubation at 37 °C under vigorous shaking, growth yield was determined (CFU/mL). Under anaerobic conditions there was modest growth for both C. concisus WT strains, however addition of H2 significantly enhanced cell yield (Fig. 3). The most dramatic effect of H2 was observed when cells were grown under microaerophilic conditions: neither strain grew under microaerophilic conditions without H2, while addition of H2 led to the highest growth yield observed herein (Fig. 3). Although formate also appears to be a potential electron donor due to the presence of FDH-N or –O genes (see Fig. 1), C. concisus cells did not grow in formate-supplemented medium under micraerobic conditions when H2 was absent (data not shown), suggesting that formate cannot substitute for H2 under these conditions. H2-enriched microaerophilic conditions are the most favorable growth conditions for C. concisus, as electrons generated by H2 oxidation flow along the respiratory chain with O2 as the final electron acceptor. Taken together, these results indicate that H2 is needed under anaerobic conditions to achieve optimal growth, while it is required under microaerophilic conditions, in agreement with results from a previous study18. These results demonstrate that C. concisus has (at least) one functional H2 uptake-type hydrogenase complex.

Figure 3
figure 3

Effect of H2 on the anaerobic and microaerophilic growth of C. concisus WT strains 13826 and 51562. C. concisus WT strains 13826 and 51562 were grown for 24 h at 37 °C with vigorous shaking (200 rpm) in 165-mL sealed bottles containing 10 mL BHI broth supplemented with 10% fetal calf serum. Bottles were flushed with N2 for 15 min then CO2, O2 and/or H2 (5%, 5%, and 20% headspace partial pressure, respectively) were added in each bottle, as indicated on the right. After 24 h, growth yield was determined by measuring bacterial cell concentration, which is based on CFU counts after serial dilution in PBS, and is expressed as CFU/mL. The dashed line indicates the average inoculum for each strain, based on CFU counts. Columns and error bars represent mean and standard deviation, respectively, from three independent growth cultures. Statistically significant differences (Student’s t-test, two-tailed) are indicated above columns.

Full size image

Supplemental H2 induces protein synthesis and nutrient transport in C. concisus

To study the effects of supplemental H2 on protein synthesis in C. concisus, strain 51562 was grown on BA plates under H2-enriched microaerophilic atmosphere, or in liquid broth under the same four different gas mixture conditions as described above. Cells were harvested after 24 h, and the same amount (10 µg total protein) of cell-free extracts was loaded onto SDS-PAGE (Supplementary Fig. S1). Two bands, corresponding to H2-induced proteins with approximate molecular mass of 50 and 45 kDa, respectively, were excised from the gel and subjected to (MALDI-MS) peptide mass fingerprinting. The most abundant protein associated with the 45 kDa-protein band was identified as translation protein EF-Tu (ORF 51562_228, with a predicted mass of 43,628 Da). Interestingly, a previous study identified EF-Tu as one of the 37 most immunoreactive proteins in strain 1382629. Analysis of the 50 kDa-protein band revealed a major outer membrane protein from the OprD family (ORF 51562_1442, with a predicted mass of 46,399 Da) as the predominant protein. Other proteins associated with these two bands include other major outer membrane proteins, as well as hypothetical proteins (see Supplementary Table S3). Thus, these results suggest H2-derived energy can be used by C. concisus to bolster its protein synthesis and import more nutrients, with increased growth as the final outcome.

C. concisus displays one of the highest H2-uptake hydrogenase activity recorded

In order to assess the H2-uptake activity in C. concisus, WT strains 13826 and 51562 cells were grown on plates under H2-enriched microaerophilic conditions and whole cell H2-uptake assays were carried out using an amperometric method, as previously described19. Hydrogenase activity levels ranged from approximately 115 to 200 nmoles of H2 used per min per 109 cells (Table 1). Those H2-uptake activity levels are by far the highest recorded in our lab, between 3-to 60-fold higher than that previously measured (using the same amperometric method) for other pathogenic bacteria studied thus far (Table 1); those include H. pylori20, H. hepaticus20, S. enterica Typhimurium33 and S. flexneri34. Hydrogenase assays carried out in this study were done under aerobic conditions. Therefore, the elevated activity levels measured herein highlight the functionality of an extremely efficient respiratory electron transport chain in C. concisus.

Table 1 C. concisus H2-uptake hydrogenase activities of various pathogenic bacteria.
Full size table

Construction and characterization of hydrogenase accessory hyp mutants

We sought to further understand the role played by H2 and hydrogenases in the pathogen’s metabolism by introducing mutations in putative hydrogenase genes. As stated above, it appears C. concisus possess three hydrogenase-related operons (Fig. 2): one operon contains hyp hydrogenase accessory genes probably needed for maturation of both hydrogenases; a second operon, annotated as hyd, is located on the same locus as the hyp operon and is predicted to encode for subunits of a H2-uptake type complex similar to that found in related ε -proteobacteria; the third operon, hyf, located elsewhere on the chromosome, possesses genes sharing significant homology with those of H2-evolving hydrogenases 3 (or 4). Since Hyp proteins are generally needed for maturation (and therefore activity) of all hydrogenase complexes, we aimed at abolishing both (H2-uptake and H2-evolving) hydrogenase activities at once by disrupting one of the hyp genes, hypE (Fig. 2). Attempts to disrupt the hypE gene using a hypE::cat PCR product (previously methylated with C. concisus cell-free extracts) proved unsuccessful. Additional attempts using an E. coli suicide plasmid containing hypE::cat were partially successful. Indeed, we were able to isolate chloramphenicol resistant clones, however PCR analysis revealed those were merodiploid mutants, with both a WT-like hypE and a (hypE::cat) mutant copy, following single cross-over insertion of plasmid DNA into the chromosome (data not shown). Furthermore, H2 uptake-type activity in the merodiploid mutants was similar to that of the WT (data not shown), suggesting the chromosomal copy of hypE was not disrupted. Similar merodiploid mutants have been described when essential genes, such at nifU or tatC35,36 were targeted in the related species H. pylori. While technical barriers cannot be ruled out, these results (or lack thereof) suggest hypE is an essential gene in C. concisus. One likely explanation is that it is required for the maturation of both hydrogenases, of which one (Hyd) seems to be also required, as suggested below.

Construction and characterization of hydrogenase hyd and hyf mutants

Since the construction of hyp mutants proved to be a challenge, we aimed at constructing independent mutants in hyd and hyf operons by targeting hydB and hyfB, respectively (Fig. 2). The hydB gene encodes for the large subunit of the [NiFe] H2-uptake hydrogenase and the hyfB gene encodes for a multi-spanning transmembrane protein with significant homology (36% identity/56% similarity) to the subunit B of the H2-evolving hydrogenase-4 complex found in E. coli25. Independent PCR products containing hydB::cat or hyfB::cat were methylated, using C. concisus cell-free extracts specific for each strain, purified, and used to transform each respective parental strain (13826 or 51562). Chloramphenicol resistant cells were only obtained for hyfB::cat and with 13826 as parental strain. The concomitant insertion of the cat cassette and the partial deletion of the hyfB gene were confirmed by PCR (Fig. 3). Despite several attempts, we were unable to obtain hydB::cat mutants in either WT strain, even when plates were supplemented with 20 mM formate, suggesting that the H2-uptake hydrogenase complex is essential in C. concisus, while the H2-synthesis hydrogenase complex is not. H2 synthesis was determined in WT and hyfB::cat mutant, using (reduced) methyl viologen (MV) as the electron donor. WT strain (13826) displayed approximately 13.4 ± 3.1 µmoles of H2 produced per min per mg of protein, while MV-dependent hydrogenase activity in the hyfB::cat mutant was not detectable (<0.1 µmoles of H2/min/mg), confirming the H2-evolving hydrogenase pathway had been successfully inactivated in the hyfB::cat mutant. The growth of the hyfB::cat mutant was compared to that of the parental strain (13826) by using the same four different gas atmospheric conditions described above (anaerobic or microaerophilic, with or without supplemental H2). There was no significant difference in growth yield (CFU/mL after 24 h) between the WT and the hyfB::cat, as determined by cell counts (data not shown), suggesting the HyfB membrane protein does not play a major role in C. concisus under these conditions.

Construction and characterization of C. concisus tetrathionate reductase mutants

The efficiency of the site-directed mutagenesis method was tested further on another putative respiratory gene, ttrA (Fig. 2). The ttrA gene encodes for the large subunit of the putative TtrAB tetrathionate reductase. Surprisingly, the C. concisus TtrA does not share homology with the bifunctional tetrathionate reductase/thiosulfate dehydrogenase TsdA found in the related species C. jejuni. Rather, its amino acid sequence resembles more that of the Salmonella Typhimurium (mono functional) tetrathionate reductase TtrA subunit (41% identity/56% similarity). In addition, C. concisus also possess genes encoding for a putative thiosulfate reductase (tsrABC, see Fig. 1). Following the same method used to generate hyfB mutants, we generated ttrA::cat mutants in both C. concisus WT genomospecies 13826 and 51562. The concomitant chromosomal insertion of the cat marker and partial deletion of ttrA was confirmed by PCR in both 51562 (Fig. 4) and 13826 (data not shown).

Figure 4
figure 4

Agarose gel with PCR products used to verify cassette insertion. Lanes 1–4 contain PCR products amplified from genomic DNA of WT or mutant strains. Primers CchyfB-1 and CchyfB-4 were used to amplify hyfB from WT strain 13826 (lane 1, expected size: 1,925 bp) or hyfB::cat from 13826 ΔhyfB::cat mutant strain DNA (lane 2, expected size: 2,050 bp). Primers CcttrA-1 and CcttrA-4 were used to amplify ttrA from WT strain 51562 (lane 3, expected size: 2,900 bp) or ttrA::cat from 51562 ΔttrA::cat mutant strain (lane 4, expected size: 2,675 bp). Lane 5 contains a DNA ladder, with sizes indicated on the right. The gel was stained with ethidium bromide. The picture has been digitally processed (black/white inverted).

Full size image

The effect of the ttrA mutation on C. concisus physiology was studied by growing cells from WT strain 51562 and its isogenic ttrA::cat mutant in liquid broth, under H2-enriched anaerobic conditions, with or without tetrathionate (S4O62−) as terminal electron acceptor (Fig. 5). After 24 h incubation at 37 °C under vigorous shaking, growth yield was determined by counting colony forming units (CFU). Supplementation of the growth medium with 10 mM S4O62− resulted in almost 10-fold increase in cell yield for the WT strain (compared to the no S4O62− added condition), suggesting that S4O62− can be used as terminal acceptor under anaerobic conditions. In contrast, supplemental S4O62− had no effect on the growth yield of ttrA mutant cells (Fig. 5), indicating the targeted gene (e.g. ttrA) is indeed involved in S4O62− respiration. To determine whether the ttrA mutation has an effect on thiosulfate (S2O32−) respiration as well, WT and ttrA::cat mutant cells were grown in the presence of 15 mM S2O32− (also in presence of H2). In this case, growth yield of all strains-WT and ttrA::cat alike-was significantly better compared to that of the control (no added terminal electron acceptor) and there was no difference between WT and mutant strains, indicating that both the WT and the ttrA::cat mutant can use S2O32− as terminal electron acceptor, under H2-enriched anaerobic conditions. In summary, the ttrA mutation prevented use of S4O62−, while S2O32− metabolism was not affected.

Figure 5
figure 5

Effect of tetrathionate and thiosulfate on the anaerobic growth of C. concisus WT and ΔttrA mutant strain. C. concisus 51562 WT and 51562 ΔttrA mutant strains were grown in 10 mL BHI broth supplemented with 10% fetal calf serum and either 10 mM sodium tetrathionate (NaS4O62−), or 15 mM sodium thiosulfate (NaS2O32−), or none. Headspace contained 5% CO2, 20% H2 and 75% N2 (partial pressure). After 24 h at 37 °C under vigorous shaking, growth yield was determined in each bottle by determining cell concentration, which is based on CFU counts after serial dilution in PBS and is expressed as CFU/mL. Columns and error bars represent mean and standard deviation, respectively, from four independent growth cultures. The dashed line indicates the average inoculum for each strain, based on CFU counts. Statistically significant differences (Student’s t-test, two-tailed) are indicated above columns. N. S, not significant.

Full size image

Discussion

To our knowledge, the present report is the first to describe the construction and characterization of mutants in the emerging pathogen C. concisus. Both strains chosen for this study, 51562 and 13826 (BAA-1457), are enteric strains that were originally isolated from feces of patients with gastroenteritis1,37. Both strains show great levels of genetic variability, to the extent that they belong to two distinct genomospecies (GS). Despite reports that BAA1457 is too atypical to be a reference strain38, we chose to include it in our study because its genome sequence is available and the strain belongs to the GS1 group39. Strain 51562 was also included in the present study, based on the facts that it is a sequenced strain and belongs to the GS2 group27,39.

Having a long-standing interest in H2 usage by pathogenic bacteria, such as H. pylori19,20,40, H. hepaticus20,41, S. enterica Typhimurium33 or S. flexneri34, we were particularly intrigued by reports on the H2 requirement in C. concisus. Indeed, previous studies suggested that not only H2 enhances C. concisus growth under anaerobic conditions, but also that H2 is actually required for C. concisus to grow in presence of microaerobic O2 concentrations18. Our first goal was to confirm the effect of H2 on C. concisus strains 13826 and 51562. Cells were grown in liquid cultures under N2-CO2 gas atmospheres, supplemented with defined volumes of H2, or/and O2, under vigorous shaking to increase gas diffusion throughout the growth medium. Our results confirmed that H2 is indeed required to achieve optimal growth under anaerobic conditions, and it is required to achieve growth under microaerophilic conditions, e.g. C. concisus cannot grow under microaerophilic conditions without H2. In fact, H2-supplemented microaerobic conditions appear to be the most favorable growth conditions (best yield) for C. concisus. This is in contrast with results from a previous study, which found that H2-supplemented anaerobic conditions are optimal for C. concisus growth18. The discrepancy between results from both studies could be attributable to several factors, including the use of different strains, as well as differences in the growth medium (solid or liquid), the gas-generating system and the quantity of H2 used. In agreement with the H2 requirement, both C. concisus strains 13826 and 51562 had extremely high H2-uptake hydrogenase activities, the highest recorded in our laboratory so far. Based on results obtained with both strains and the fact that genes encoding for the H2-uptake hydrogenase are present in all C. concisus genome sequences analyzed thus far (regardless of the GS they belong to), we hypothesize all members of the C. concisus species will have higher than usual H2-uptake hydrogenase activity. This will have to be experimentally verified though.

To get a better understanding of H2 metabolism in C. concisus, we aimed at inactivating hydrogenase maturation or synthesis genes using a classical site-directed mutagenesis approach. This, however, was anticipated to be a challenge since no mutant had been reported prior to the current study. A chloramphenicol resistance marker (cat gene) was chosen, based on the following: first, C. concisus has been shown to be Cm sensitive, with MIC of only 4 µg/mL reported in two independent studies1,42; second, the cat cassette was originally isolated from a related species, Campylobacter coli43; third, the cassette has been successfully used to generate numerous mutants (including hydrogenase mutants) in the related ε-proteobacteriaceae species H. pylori and H. hepaticus40,41; and fourth, the cassette has its own promoter and has been shown not to cause polar effects36. Therefore, the cat cassette appeared to be a suitable marker to disrupt genes in C. concisus. We aimed at inactivating both the H2-uptake and H2-synthesis hydrogenase pathways at the same time by targeting one of the hyp genes involved in hydrogenase maturation. Our first attempt to construct a hypE::cat mutant by natural transformation or electroporation was unsuccessful, suggesting that either the hypE gene was essential in C. concisus, or our transformation methods were not suitable for this microorganism, or both. Given that C. concisus belongs to the ε-proteobacterium group, a group whose members (Helicobacters or other Campylobacters for instance) are known to be naturally transformable, we hypothesized that transformation (DNA uptake) was not the reason our strategy was unsuccessful. Instead, the failure to introduce foreign DNA within C. concisus has probably more to do with its restriction/ modification system. Thus, we used a DNA methylation method originally developed to overcome the restriction barrier in H. pylori44. The method was successfully applied to C. concisus: first we were able to recombine hypE::cat along with a suicide plasmid into the chromosome. While this single cross-over recombination did not yield a hypE mutant per se, nevertheless it proved that both transformation and recombination into the C. concisus chromosome are possible, especially after proper methylation treatment. Using the same method and PCR products, we successfully inactivated two independent genes, hyfB and ttrA, encoding for a membrane component of the H2-evolving hydrogenase complex and the large subunit of tetrathionate reductase, respectively27.

The use of molecular hydrogen as source of energy by bacteria, including human pathogenic bacteria, has been well documented (for a review, see45). However, in all pathogenic bacteria studied so far (H. pylori, H. hepaticus, S. enterica Typhimurium or S. flexneri), H2 is needed but it is not required, e.g. mutants devoid of H2-uptake hydrogenase activity are viable under laboratory conditions34,40,41,46. It seems this is not the case for C. concisus. The remarkable importance of H2 in the pathogen’s metabolism is highlighted by the fact that the H2-uptake hydrogenase appears to be essential for C. concisus, since we could neither inactivate hypE, a gene needed for maturation of hydrogenases in bacteria, including in H. pylori47, nor hydB, the gene encoding for the large subunit of the H2-uptake Hyd complex. The observation that C. concisus only requires exogenous (supplemented) H2 under microaerophilic conditions, but not under anaerobic conditions is puzzling, however it can be tentatively explained by the redox-dependent expression of the H2-evolving hydrogenase. Indeed, in E. coli hyc (Hyd-3) genes are only expressed under fermentative growth conditions i.e. in absence of all exogenous terminal electron acceptors, including O248; likewise, hyf (Hyd-4) genes are also expressed under anaerobic conditions49. Applied to C. concisus, this means that hyf genes are likely to be expressed under anaerobic conditions, leading to endogenous production of H2 by the FHL or FHL-2 complex, as depicted in our proposed model (Fig. 6). Formate oxidation could be coupled to hydrogen production in C. concisus, as it is the case in E. coli with FHL23. Alternatively, other compounds (NADH or organic acids) could be oxidized instead of formate; indeed both the oxidized compound and the oxidizing enzyme (FDH-H counterpart) are still unknown with respect to the E. coli FHL-2 complex23,26. Regardless of whether formate or another electron donor plays a role in the C. concisus FHL-2 system, it appears C. concisus can produce H2, as shown in this study; this endogenous H2 could in turn be used by the Hyd hydrogenase, after diffusing through membranes. Based on this model one would expect the hyf mutant to have a lower growth yield compared to WT when cells are grown under anaerobic conditions, however there was no significant difference between strains, as both the WT and hyf mutant strains grew poorly, even in presence of formate. Thus, additional electron donors (e.g. NADH or organic acids, as discussed above) might be required to augment H2 synthesis and support anaerobic cell growth. In contrast, under aerobic or microaerophilic conditions, hyc or hyf genes are expected to be turned off, preventing cells from synthesizing H223. Under these conditions, the only source of H2 C. concisus can rely on is exogenous H2 (Fig. 6). H2-enriched microaerophilic conditions are presumably the most favorable conditions for C. concisus, as electrons generated by H2 oxidation flow along the respiratory chain with O2 as the final electron acceptor (C. concisus possess both terminal cytochrome oxidases cbb3 and bd). This was confirmed in the current study.

Figure 6
figure 6

Model showing the redox-dependent expression of Hyd and Hyf hydrogenases complexes in C. concisus. Based on their protein subunit composition, both hydrogenase enzyme complexes are expected to be membrane-bound, however the presence of a TAT motif in the (HydA) sequence indicates the Hyd hydrogenase complex faces the periplasmic space, whereas the Hyf complex (as part of FHL) is supposed to be cytoplasmic. (A) Under anaerobic conditions both hydrogenase complexes are present, allowing the cells to use Hyf-produced H2 and grow without exogenous H2. (B) Under microaerophilic conditions, expression of FHL genes is inhibited and C. concisus only synthesizes Hyd. No growth of C. concisus under microaerophilic conditions is observed unless (host produced) exogenous H2 is available.

Full size image

The fact that C. concisus can use H2 both under anaerobic and microaerophilic conditions likely explains why it can be found in various niches of the human body, including the oral cavity2. Despite the fact that this habitat is considered mostly anaerobic, C. concisus can probably rely on FHL-produced H2; in addition, exogenous H2 is also available, as suggested by several studies. For instance, colonic bacteria continuously produce H2 as part of their metabolism50; the gas is able to move into other tissues (including the lungs) through a combination of cross-epithelial diffusion51 and vascular-based transport52. As a consequence, approximately 14% of intestinal-produced H2 is predicted to be eventually excreted through the breath53. Furthermore, Kanazuru et al. found that the concentration of H2 in the oral cavity of non-expirating healthy volunteers was around 20–30 ppm, spiking to 120 ppm following glucose intake54. Most of this oral H2 was attributed to fermentation by Klebsiella pneumoniae. Such ranges correspond to millimolar H2 concentrations, which are high enough to be detected in the breath through a breath analyzer. While the affinity constant of C. concisus H2-uptake hydrogenase for the substrate (H2) is not yet known, most hydrogenases have a Km in the micromolar range, therefore H2 levels in the oral cavity are likely not a limiting factor for C. concisus in the oral cavity; rather H2 is more likely to be found in excess.

Likewise, in the human gut, another natural niche for C. concisus, there is also abundant H2. Indeed, the colonic flora (predominantly composed of anaerobic bacteria) breakdown host-undigested carbohydrates, producing a variety of catabolites such as short chain fatty acids, lactate, CO2, formate and H250,55. The latter can in turn be used by H2-uptake hydrogenase-containing bacteria, including pathogenic bacteria such as C. concisus and S. enterica Typhimurium. The role of hydrogenases in S. Typhimurium’s host colonization have been studied: Maier et al. showed that S.T. hydrogenase mutants are unable to colonize the colon of mice56. In addition, S. Typhimurium can respire S4O62−, produced from host-driven S2O32− during inflammation57. The use of S4O62− as terminal electron acceptor confers S. Typhimurium a selective advantage over the competing microbiota that cannot respire S4O62– 57. Thus, S. Typhimurium thrives under inflammatory conditions. Interestingly, C. concisus also possess a S4O62− reductase, that appears to be structurally closer to the S. Typhimurium enzyme than to the bifunctional (S4O62− reductase/ S2O32− oxidase) enzyme found in C. jejuni. In the present study, we were able to disrupt the ttrA gene encoding for the large subunit of S4O62− reductase in C. concisus. The phenotype associated with the mutation was as expected: addition of S4O62− enhanced growth in the WT, but not in the ttrA mutant. This suggests that the S4O62− reduction pathway is operational. Since C. concisus’s association with gut inflammatory diseases (such as ulcerative colitis and Crohn’s disease) is well documented8,58, it is very likely the pathogen can use host-produced S4O62− at its own advantage, similar to what has been described for S. Typhimurium. It is worth noting however that not all sequenced C. concisus strains possess tetrathionate reductase genes27.

Taken together, our results shed some light on C. concisus’s versatile respiratory system. Its H2-uptake and H2-synthesizing abilities, coupled to its capacity to respire nitrogen, sulfur and oxygen-containing compounds explain why C. concisus can successfully adapt to and colonize such diverse environmental niches of the human body. Finally, we showed that disrupting genes by site-directed mutagenesis in C. concisus is possible, and we hope this report will provide researchers with new genetic tools to study this emerging pathogen.

Experimental Procedures

Bacterial strains and plasmids

E. coli and C. concisus strains and plasmids used in this study are listed in Table S1. Genomic DNA from either C. concisus ATCC-51562 or C. concisus ATCC-BAA-1457 (13826) was used as template for all PCR amplifications. All plasmids and PCR products were sequenced at the Georgia Genomics Facility, University of Georgia, Athens, GA.

Growth conditions

Campylobacter concisus was routinely grown on Brucella agar (Becton Dickinson, Sparks, MD) plates supplemented with 10% defibrinated sheep blood (Hemostat, Dixon, CA) (BA plates). Chloramphenicol (Cm, 8 µg/ml) was added as needed. Cells were grown at 37 °C, in sealed pouches filled with anaerobic mix, a commercial gas mixture containing 10% H2, 5% CO2 and 85% N2 (Airgas, Athens, GA). For liquid cultures, 165-ml sealed bottles were filled with 10 mL of Brain-Heart Infusion (BHI, Becton Dickinson) supplemented with 10% fetal calf serum (FCS, Gibco Thermo Fisher). To study the effect of H2 and O2 on C. concisus growth, bottles were first flushed with N2 for 15 min, then CO2 (5% headspace partial pressure, h.p.p.) was injected in every bottle, followed by H2 (20% h.p.p.) and/or O2 (5% h.p.p.), as indicated. To study the effect of tetrathionate or thiosulfate under anaerobic conditions, bottles were first sparged with N2 for 15 min, followed by anaerobic mix for 15 min. Additional H2 (10%) was added, then sodium tetrathionate (10 mM) or sodium thiosulfate (15 mM) were added as indicated. For all liquid growth experiments, the inoculum was prepared as follows: C. concisus cells grown on BA plates for less than 24 h were harvested, resuspended in BHI and standardized to the same OD600, before being inoculated (1:100) into 10 mL of BHI-FCS. The starting OD600 was between 0.03 to 0.04 (corresponding to 6.7 × 107 to 9 × 107 CFU/mL, respectively). The actual bacterial concentration at the time of inoculation was determined by plating serial dilutions and counting CFU. Cells were grown (triplicate or quadruplicate for each strain and condition) for 24 h at 37 °C under vigorous shaking (200 rpm). Growth yield (CFU/mL) was estimated as follows: samples from each bottle were serially diluted (up to 10−7) in PBS and 5 µL of each dilution was spotted in triplicate on BA plates. CFU were counted after 48 h of incubation under H2-enriched microaerophilic conditions. E. coli cells were grown aerobically in Luria-Bertani (LB) medium or plates at 37 °C, unless indicated otherwise. Ampicillin (100 µg/mL) and chloramphenicol (25 µg/mL) were added as needed.

Identification of H2-induced proteins by MALDI-MS

C. concisus (strain 51562) was grown on BA plates under H2-enriched microaerophilic conditions or in BHI-FCS liquid broth, under four different gas atmospheres, as described above. After 24 h, cells were harvested, broken by sonication and spun down. Cell-free extracts were isolated and the protein concentration was determined with the BCA protein kit (Thermo Fisher Pierce, Rockford, IL, USA). Samples were processed using in-gel digestion with trypsin. MALDI was performed at the University of Georgia Proteomics and Mass Spectrometry (PAMS) Facility, on a a Bruker Autoflex using reflectron mode and 2,5-dihydrocybenzoic acid as matrix. Results were analyzed using the Mascot MS/MS ion search (Matrix Science, Boston, MA) and searches were performed on the National Center for Biotechnology Information (NCBI) non-redundant database (against genome sequences of C. concisus strains 51562 and 13826).

Construction of C. concisus mutants

The construction of each mutant followed the same 3-step strategy: (1) generation of DNA constructs used for mutagenesis; (2) methylation of DNA constructs using C. concisus cell-free extracts and S-Adenosyl Methionine (SAM); (3) Transformation of C. concisus with the (purified) methylated DNA and selection on antibiotics-containing plates. In the first step, a splicing-by-overlap-extension (SOE) PCR method was used. Briefly, two DNA sequences ranging from 0.5 to 1-kb in size and flanking each target sequence (hydB, hyfB, hypE or ttrA, respectively) were amplified by PCR (iProof polymerase, Bio-Rad, Hercules, CA), using genomic DNA from strain 51562 and specific primers for each target (Table S1). Each set of two PCR products was then combined with a 740 bp-long cat (chloramphenicol resistance) cassette that has its own promoter43 and the final SOE PCR step yielded a product containing both flanking sequences with the cat cassette in the middle. In the second step, each tripartite PCR product was purified and methylated, following a modified method previously used to generate mutants in H. pylori36,44,59,60. Briefly, approximately 25 µg of DNA was incubated for 2 h at 37 °C with 150–250 µg of (cell-free extract) total protein from C. concisus (either strain 13826 or 51562, depending on final recipient strain) in presence of 0.4 mM of SAM (New England Biolabs, Ipswich, MA). After methylation, each PCR product was purified again (Qiaquick purification kit, Qiagen, Valencia, CA) and used to transform C. concisus; each strain was transformed by natural transformation or electroporation (BTX Transporator Plus, 2,500 V/pulse) with its corresponding strain-methylated DNA (1–5 µg). Transformed cells were first plated on BA plates and incubated (H2-enriched microaerophilic conditions) for 8–12 h before being transferred onto BA supplemented with 8 µg/mL Cm. Colonies appeared after 3 to 5 days. The concomitant deletion in the gene of interest (hypE, hyfB or ttrA) and the insertion of cat was confirmed by PCR, using genomic DNA from mutants as template and appropriate primers.

Construction of hypE::cat mutant

Primers CchypE-1 and CchypE-2 (Table S2) were designed to amplify a 500 bp-long DNA sequence corresponding to the first half of the hypE open reading frame (ORF) (13826_1093 or 51562_1332) as well as to incorporate the 5′ end of the cat marker. Primers CchypE-3 and CchypE-4 were designed to amplify a 520 bp-long sequence corresponding to the 3′ end of cat and the second half of the hypE ORF. The final SOE amplification step using CchypE-1 and CchypE-4 generated a 1,730 bp-long hypE::cat DNA sequence, with the cat cassette located in the middle of the hypE gene. The hypE::cat construct was either methylated and used to transform C. concisus, or it was cloned into plasmid pBluescript KS (pBS-KS). In this case, pBS-KS was digested with SmaI and ligated with the hypE::cat PCR product that had been previously blunt-ended with T4 polymerase. The newly generated plasmid (plasmid pSB624, Table S1) was then methylated as described above prior to transformation.

Construction of hydB::cat mutant

Primers CchydB-1 and CchydB-2 (Table S2) were designed to amplify a 625 bp-long DNA sequence corresponding to the beginning of the hydB ORF (13826_0100 or 51562_1311) and the 5′ end of cat. Primers CchydB-3 and CchydB-4 were designed to amplify a 675 bp-long sequence corresponding to the 3′ end of cat and the 3′ end of the hydB ORF. The final SOE amplification step with both PCR products, the cat cassette and primers CchydB-1 and CchydB-4 generated a 2,000 bp-long hydB::cat DNA sequence, in which approximately 415 bp (of the 1,720 bp-long hydB ORF) is missing and replaced by cat.

Construction of hyfB::cat mutant

Primers CchyfB-1 and CchyfB-2 (Table S2) were designed to amplify a 700 bp-long DNA sequence corresponding to the beginning of the hyfB ORF (13826_1914 or 51562_0661) and the 5′ end of cat. Primers CchyfB-3 and CchyfB-4 were designed to amplify a 650 bp-long sequence corresponding to the 3′ end of cat and the 3′ end of the hyfB ORF. The final SOE amplification step with both PCR products, the cat cassette and primers CchyfB-1 and CchyfB-4 generated a 2,050 bp-long hyfB::cat DNA sequence, in which approximately 620 bp of the 1,925 bp-long hyfB ORF is missing and replaced by cat.

Construction of ttrA::cat mutant

Primers CcttrA1 and CcttrA2 (Table S2) were designed to amplify a 1,070 bp-long DNA sequence corresponding to the beginning of the ttrA ORF (13826_2089 or 51562_0690) and the 5′ end of cat. Primers CcttrA3 and CcttrA4 were designed to amplify a 920 bp-long sequence corresponding to the 3′ end of cat and the 3′ end of the ttrA ORF. The final SOE step with both PCR products, the cat cassette and primers CcttrA-1 and CcttrA-4 generated a 2680 bp-long ttrA::cat DNA sequence, in which approximately one third (970 bp out of 2,988 bp) of the core sequence of ttrA ORF is replaced by cat.

Hydrogenase assays

Whole cells H2-uptake hydrogenase assays

H2-uptake was assayed using a previously described amperometric method19. Briefly, C. concisus cells from strain 13826 or 51562 were grown for 24 h on BA plates under H2-enriched microaerophilic conditions, harvested and resuspended in phosphate buffered saline (PBS). Cell density (OD600) was measured to evaluate cell concentration (1 unit of OD600 corresponds to approximately 2.25 × 109 cells/mL, as determined in this study). A known volume of cells was injected into a 1.8 mL chamber containing H2-saturated PBS and H2 disappearance (H2 uptake by live C. concisus cells) was monitored as previously described19. Activities are reported as nmoles of H2 used per min per 109 cells, and represent 3 to 4 independent measurements.

H2-evolving hydrogenase assays

H2-synthesis was measured by monitoring the oxidation of dithionite-reduced methyl viologen (MV, ε604 = 1.39 mM−1 cm−1) at 604 nm61. Briefly, C. concisus 13826 WT and 13826 ΔhyfB mutant strains were grown on BA plates under H2-enriched microaerophilic conditions. After 24 h, cells were harvested in N2-saturated HEPES-NaOH (50 mM) pH 7.5, broken by sonication and spun down for 5 min at 15,000 × g. Cell-free extracts were isolated and the protein concentration was determined using the BCA protein kit. MV (10 mM final concentration) was added to N2-saturated HEPES in 1.8 mL glass cuvettes closed with rubber stoppers. Freshly prepared sodium dithionite was injected to reduce MV and give a dark blue color with OD604 of approximately 1, and the reaction was initiated by adding 5 µl (20 to 40 µg) of cell-free extracts from either WT or ΔhyfB mutant cells. Activities are expressed as µmoles of H2 produced per min per mg of protein. Results represent means and standard deviations of three independent growth experiments, with assays done in triplicate.

Genome sequence analysis

Gene and protein sequences of C. concisus strains 13826 (BAA-1457) and 51562 were obtained from the integrated microbial genomes (IMG) website of the Joint genome Institute (https://img.jgi.doe.gov). The following databases and prediction tools were also used: Genbank (www.ncbi.nlm.nih.gov), Uniprot (www.uniprot.org), and STRING (www.string-db.org).