• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of iaiPermissionsJournals.ASM.orgJournalIAI ArticleJournal InfoAuthorsReviewers
Infect Immun. Nov 2004; 72(11): 6294–6299.
PMCID: PMC523013

Respiratory Hydrogen Use by Salmonella enterica Serovar Typhimurium Is Essential for Virulence


Based on available annotated gene sequence information, the enteric pathogen salmonella, like other enteric bacteria, contains three putative membrane-associated H2-using hydrogenase enzymes. These enzymes split molecular H2, releasing low-potential electrons that are used to reduce quinone or heme-containing components of the respiratory chain. Here we show that each of the three distinct membrane-associated hydrogenases of Salmonella enterica serovar Typhimurium is coupled to a respiratory pathway that uses oxygen as the terminal electron acceptor. Cells grown in a blood-based medium expressed four times the amount of hydrogenase (H2 oxidation) activity that cells grown on Luria Bertani medium did. Cells suspended in phosphate-buffered saline consumed 2 mol of H2 per mol of O2 used in the H2-O2 respiratory pathway, and the activity was inhibited by the respiration inhibitor cyanide. Molecular hydrogen levels averaging over 40 μM were measured in organs (i.e., livers and spleens) of live mice, and levels within the intestinal tract (the presumed origin of the gas) were four times greater than this. The half-saturation affinity of S. enterica serovar Typhimurium for H2 is only 2.1 μM, so it is expected that H2-utilizing hydrogenase enzymes are saturated with the reducing substrate in vivo. All three hydrogenase enzymes contribute to the virulence of the bacterium in a typhoid fever-mouse model, based on results from strains with mutations in each of the three hydrogenase genes. The introduced mutations are nonpolar, and growth of the mutant strains was like that of the parent strain. The combined removal of all three hydrogenases resulted in a strain that is avirulent and (in contrast to the parent strain) one that is unable to invade liver or spleen tissue. The introduction of one of the hydrogenase genes into the triple mutant strain on a low-copy-number plasmid resulted in a strain that was able to both oxidize H2 and cause morbidity in mice within 11 days of inoculation; therefore, the avirulent phenotype of the triple mutant is not due to an unknown spurious mutation. We conclude that H2 utilization in a respiratory fashion is required for energy production to permit salmonella growth and subsequent virulence during infection.

Together, enteric pathogens are responsible for an estimated 2 million deaths annually (reference 3 and the World Health Organization site at http://www.who.int/health-topics/index.html) and cause millions more cases of diarrheal illness annually, even in developed countries (see Centers for Disease Control site at http://www.cdc.gov/health/default.htm). Based on annotated whole-genome sequences, intestinal disease-causing bacteria such as Salmonella spp., Escherichia coli, Shigella spp., Yersinia spp., and Campylobacter spp. all contain homologous hydrogenases (5). These are membrane-associated H2-splitting enzymes that carry out the relatively simple H2-oxidizing reaction H2 → 2e + 2H+. Another pathogenic bacterium with a predicted (but unstudied) NiFe uptake-type hydrogenase is Actinobacillus pleuropneumoniae; it is in the Pasteurellaceae family, is able to persist in tonsils or necrotic lung tissue, and is the causative agent of porcine pleuropneumoniae (see National Center for Biotechnology Information site at http://www.ncbi.nlm.nih.gov/ and then access strain 4074).

The membrane-bound hydrogenases associated with H2 oxidation typically split molecular H2 via a unique NiFe metal center, with the release of protons and low-potential electrons (16). The H2-splitting reaction by hydrogenase does not yield energy as ATP per se, but the two protons released from H2 can contribute to a proton gradient across the membrane (16). Importantly, the NiFe hydrogenase enzymes are membrane associated whereby the electrons generated from splitting molecular hydrogen are sequentially passed to heme-containing or quinone-reactive proteins. Therefore, the total generated proton gradient is thought to result from a combination of the “sidedness” of the H2-splitting reaction along with the sequential reduction of other redox enzymes within the membrane. The potential energy thus generated can be used for ATP production via oxidative phosphorylation or to drive carbon transport systems (8). For some bacteria, molecular hydrogen can represent the entire energy source used for growth, provided that a terminal electron acceptor is available to allow disposal or energy harvesting of the large amount of generated protons and electrons. It was only recently demonstrated that this process of energy generation from H2 could be important for bacterial pathogenesis within animal hosts (12).

It was suggested that H2-using hydrogenase enzymes might enable enteric bacteria to glean energy from the splitting of molecular hydrogen (1). The high-energy gas is produced by colonic flora within animals (5), and because it is freely diffusible, the gas can be measured within both intestinal and nonintestinal tissue (5, 6, 12). The gastric pathogen Helicobacter pylori contains only a single membrane-associated hydrogenase (7), and it was demonstrated that use of H2 by this enzyme is important for the bacterium's ability to colonize the stomach (12). H. pylori thrives in a nutrient-poor environment (the gastric mucosa), so its use of H2 was proposed to be a clever way for the pathogen to be independent of the host for its prime energy source. Here we address the importance of H2 use for the pathogenicity of Salmonella enterica serovar Typhimurium, a common food poisoning bacterium closely related to the typhoid fever-causing bacterium S. enterica serovar Typhimurium; we assess the role of enteric H2 use in the mouse model of typhoid fever (13). Strains with mutations at STM3147 (for the large subunit), STM1538, and STM1787 were generated; these designations correspond to the genes hybC (STM3147) and hydB (STM1538 and STM1787). Also, as the hydrogenases of S. enterica serovar Typhimurium have been studied little, we first had to address whether the predicted enzymes were functional in a respiratory (H2-oxidizing) fashion. This was also assessed by a mutant analysis approach, after obtaining conditions where high hydrogenase activity was expressed by the parent strain.


Amperometric hydrogenase assays.

The conditions for obtaining hydrogenase activity involved the growth of cells on a blood-containing medium (11) in a microaerobic H2-containing atmosphere (see Table Table1).1). S. enterica serovar Typhimurium cells grown for 1 day on the blood agar plates were suspended in phosphate-buffered saline (PBS) (11), and 8-ml samples at cell concentrations of 8 × 108 per ml were assayed for H2 and O2 uptake activities simultaneously. This assay was accomplished by using the same sample in a stirred and sealed amperometric dual-electrode chamber (9). Hydrogen and oxygen were added as needed from gas-saturated solutions of PBS. H2 uptake rates were linear until the substrate reached levels of about 3 to 5 μM. For methylene blue-dependent rates, the chamber lacked oxygen but contained methylene blue at 200 μM, and the cells were permeabilized with Triton X-100 (6) before assay. Cell numbers were determined by performing dilutions and plate counts with MacConkey medium. For determining the affinity of whole cells for H2, the H2-O2 uptake assay was performed with a series of limiting H2 levels between 1.2 and 10 μM. The O2 was at a saturating level for all these assays but was maintained below 55 μM, as the affect of high O2 levels on the three separate (hydrogenase) enzymes is not known. The double-reciprocal plot yields a line equation as indicated in the text, for which kinetic parameters can be calculated. The Km from these data is referred to herein as the half-saturation affinity for H2, because in our case, the kinetic constants are for a whole-cell system rather than for the pure enzyme (the latter being the conventional system for such determinations).

Variations of growth conditions for obtaining respiratory H2-oxidizing activity

Mutant strain construction.

All mutations were in the structural genes for hydrogenase and were constructed in a way that would not disrupt downstream genes. The three hydrogenase genes targeted were ones that are homologous to genes encoding membrane-associated NiFe uptake hydrogenases (a fourth putative hydrogenase in S. enterica serovar Typhimurium is homologous to the hydrogenase, termed HycE, of E. coli that is proposed to be associated with electron transfer reactions within a formate hydrogen lyase complex) (1). Although the three targeted enzymes are predicted to be membrane-bound NiFe types and were homologous to the E. coli hydrogenases, assigning them to genes based on E. coli nomenclature (like hya genes) is ambiguous. The genes belonging to gene groups I, II, and III are analogous to hyb and to two different but similar versions of hyb, respectively, (8a, 13a). We describe the specific genes that were disrupted by number. S. enterica serovar Typhimurium strain ATCC 14028 (JSG210) was used as the parent for the construction of all the mutants. The Lambda Red system was used to construct deletion mutations in the hydrogenase genes (2). An antibiotic cassette located on plasmid pKD4 was amplified by PCR. Primers were designed (at their 5′ ends) to contain homologous sequences to the DNA outside the fragment assigned for deletion. Primers were designed to delete a 4,972-bp region of group I genes (STM3147 through STM3150), represented by coordinates 3313762 to 3308790 in the Institute for Genomic Research (TIGR) comprehensive microbial resource for S. enterica serovar Typhimurium. Deletions in the group II genes, including STM1538 and STM1539, were created by deleting a 2,905-bp fragment (coordinates 1614904 to 1611999). For group III, a fragment of 2,905 bp was deleted (coordinates 1884828 to 1887733, including genes STM1786 and STM1787). Genes within the deleted regions encoded large and small hydrogenase subunits. Each of the PCR fragments was transformed by electroporation into a strain of S. enterica serovar Typhimurium containing the Red helper plasmid, allowing the uptake of linear DNA and recombination. The antibiotic resistance cassette in the mutants was eliminated by transforming the strains with the FLP (flippase site-specific recombinase active at the FLP recognition target [FRT]) synthesis-inducing plasmid, pCP20 (2). FRT-flanked resistance genes, as well as the FLP helper temperature-sensitive plasmid, were both lost at 43°C. Double mutants were obtained by P22HTint-mediated transduction of an antibiotic-marked strain with a single-gene deletion into the appropriate strain with a single-gene deletion by antibiotic selection, followed by transformation with the pCP20 plasmid and elimination of the antibiotic cassette. The triple mutant was constructed in a similar manner, using the appropriate double deletion mutant (gene group I and II negative) as the recipient. All deletions were confirmed by PCR with primers designed using sequences outside of the deleted DNA regions. Each deletion left a few base “scars” in place of the deleted DNA. The scar region contained a ribosome-binding site at the 5′ end as well as a start codon, which allowed reengagement of the ribosome for the downstream gene transcription. Also, reverse transcription-PCR of genes directly downstream of the deletion (for group I, STM3145; for group II, STM1536; and for group III, STM1789) was performed in order to confirm the lack of polar effects. Additionally, the expression of STM3142 (ferrichrome-binding periplasmic protein) was measured by reverse transcription-PCR in the triple mutant strain and was found to be the same as for the wild-type strain. Therefore, the expression of genes downstream of the deleted regions was unaffected, so the results presented are due to hydrogenase deficiencies only. Strain numbers corresponding to the introduced double deletions and the triple mutant are given below (see Fig. Fig.2),2), and two other strains (containing a low-copy-number plasmid) are described in the following paragraph.

FIG. 2.
Virulence of S. enterica serovar Typhimurium strains for mice. The data shown are for a total of 30 mice each for the wild type and the triple mutant strain, based on the combined data from two separate experiments. Values on the y axis indicate the percent ...

Primers JG915 (5′ GCTCTAGAAAAAATACGCGTTATG 3′) and JG916 (5′CCCAAGCTTAGCGTACCTGGACGGC 3′) were used to generate a PCR product containing STM1786 and STM1787. These two genes correspond to those missing in the group III deletion strain. The PCR fragment was cloned into pWSK29 (17) by using the XbaI and HindIII sites engineered into the 5′ ends of JG915 and JG916, respectively. STM1786 and STM1787 are expressed from the lac promoter of the vector. The resultant plasmid pG3229 was transformed by electroporation into strain JSG321, the triple hydrogenase gene mutant, creating strain JSG2495. The plasmid vector pWSK29 was also transformed into strain JSG321, creating JSG2497.

Mouse experiments.

BALB/c female mice (obtained from the National Cancer Institute, Frederick, Md.) were inoculated orally as described previously (15) with 0.1-ml volumes of washed cells (containing 106 bacteria) suspended in PBS. The mice were observed twice daily, and morbidity was recorded. The organ bacterial burdens postinoculation were obtained by euthanizing mice (96 h after the inoculation of four mice with each bacterial strain). The livers and spleens were immediately removed from the euthanized mice, and the organs were homogenized in PBS. Dilutions were plated onto MacConkey agar, a medium selective for gram-negative, lactose-negative bacteria, and colonies were counted the next day. No colonies were observed in homogenized organs from uninoculated mice (two to a cage).

Measurements of H2 levels within the small intestine of mice were performed by making a small incision into the intestinal wall with a razor blade and inserting a 50-μm-diameter-tip-size H2 microelectrode probe less than 0.5 mm into the intestine. For splenic H2 determinations, the probe was placed 0.2 to less than 1.0 mm into the spleen tissue as described previously for H2 measurements in liver tissues from live mice (6). These determinations, including instrument calibrations, were performed like those described in detail previously for H2 measurements in other tissues (6, 12). Care was taken to keep all the organs as intact as possible during surgery and for microelectrode measurements, and the mice were kept under anesthesia during the procedure. Acquiring a stable H2 signal within the tissue sometimes required no probe movement for up to 12 s. Twelve independent measurements were made (four each for three separate mice) for each tissue, and the mean of these measurements is reported below.


Characteristics of H2 oxidation activity.

S. enterica serovar Typhimurium hydrogenase activity has been ascribed to at least two distinct but similar membrane-associated hydrogenases (4, 14), and possible roles for these enzymes in anaerobic energy metabolism have been proposed (1). The complete genome sequence of S. enterica serovar Typhimurium LT2 indicates that the bacterium contains genes for three putative homologous membrane-associated, H2-utilizing-type hydrogenases (see http://www.tigr.org/tigr-scripts/CMR2/GenomePage3.spl?database=ntst01). Although previously reported S. enterica serovar Typhimurium hydrogenase activity was always determined anaerobically, the gene-annotated sequence reveals that S. enterica serovar Typhimurium has several O2-binding oxidases that could perhaps allow for the complete respiratory oxidation of electrons from H2 all the way to O2 reduction. If the reductant H2 could be used simultaneously with O2 as the acceptor (i.e., via respiration), then it is expected that a high-efficiency energy yield would be available to allow H2-mediated growth of cells (5, 16). This is a common role for NiFe hydrogenases in many aerobic bacteria (16). Therefore, we measured H2 oxidation coupled to O2-dependent respiration in the parent strain in various gas atmospheres and culture medium conditions, including blood agar and microaerobic atmospheres (Table (Table1).1). Previous studies of enteric bacteria hydrogenase used cells grown on either a glucose-peptone medium (14) or on Luria Bertani (LB) medium (4) under strictly anaerobic conditions. H2 oxidation was monitored herein simultaneously with O2-dependent respiration by use of H2 and O2 electrodes on the same (sealed and continuously stirred) samples.

In addition to LB, we tested blood agar as a possible high-nutrient growth medium to enable hydrogenase expression. This proved to be highly beneficial to hydrogenase expression (Table (Table1).1). The parent strain was able to readily oxidize H2 at rates observed for another H2-oxidizing pathogenic bacterium (H. pylori) under the same incubation conditions (i.e., blood agar plus a microaerobic H2-containing atmosphere) (12). Salmonella activities on blood agar were four times that on LB when both were incubated with an anaerobic gas mixture (Table (Table1,1, compare growth conditions 1 and 6). Oxygen repressed hydrogenase expression, as seen by comparing condition 2 or 4 with condition 1 in Table Table11 (also compare condition 7 or 9 with condition 6). This phenomenon of O2 repression on hydrogenase expression is common for respiratory hydrogenases (16). Also, incubation with H2 augmented expression, as demonstrated by comparing conditions 1 and 5 and conditions 6 and 10.

Additional characteristics of H2 oxidation.

A direct amperometric recording of whole cells of S. enterica serovar Typhimurium using both H2 and O2 as dissolved gases in the buffer is shown in Fig. Fig.1.1. Nevertheless, some H2 oxidation occurred after the O2 was exhausted. As shown, when O2 was exhausted (at about the 5-min mark in Fig. Fig.1),1), H2 uptake slowed but continued at a low rate, i.e., approximately 8% of the aerobic rate, and this diminished rate was maintained for about 7 to 10 min. After this period of diminished activity, over the next 3 to 4 min, H2 uptake diminished more and ceased entirely (data not shown). This result was observed in over 10 different assays and was not observed previously when other H2-oxidizing pathogenic bacteria, such as H. pylori or Helicobacter hepaticus, were assayed (6). We attribute the “anaerobic” H2 oxidation either to endogenous acceptors (perhaps organic acids like fumarate) still present within the bacterium or perhaps to residual electron acceptors present in the medium. These predictions are supported by the observation that the incubation of cell suspensions (cells removed from the blood agar medium into PBS) in an H2-containing atmosphere for 10 to 30 min at room temperature resulted in cells that no longer exhibited the anaerobic H2 respiration activity (i.e., the terminal substrate is presumably exhausted). Normally, cells were suspended in PBS and assayed immediately, so a low rate of H2 oxidation occurred without O2. By performing hydrogenase assays in the absence of oxygen, along with the use of mutant strains, we conclude that some, but not all, of the endogenous or anaerobic activity (which is the minor H2 respiration activity) could be assigned to the function of hydrogenase produced by gene group I, but gene group I is also responsible for O2-dependent H2 oxidation and respiration (Table (Table22).

FIG. 1.
Dual-channel amperometric recording of simultaneous H2 and O2 use by whole cells of S. enterica serovar Typhimurium. H2 and O2 were injected into the (5.5-ml-volume) amperometric chamber from gas-saturated solutions, and the gases were monitored continuously ...
Aerobic H2 oxidation activity by S. enterica serovar Typhimurium strainsa

In Fig. Fig.1,1, it can also be observed that the stoichiometry of H2-O2 respiration for washed cells in PBS is approximately 2 mol of H2 oxidized per mol of O2 consumed. After the anaerobic H2 oxidation ceased and O2 was supplied again, the stoichiometry of H2 uptake to O2 uptake was measured more precisely at 2.0, as expected for the complete oxidation of H2 by O2. The fact that the bulk of H2 oxidation (even in the first 5 min of the assay) occurs via respiration to O2 was corroborated by cyanide inhibition experiments performed as follows. The addition of 0.1 mM cyanide to the S. enterica serovar Typhimurium cell suspension prior to the start of the H2 uptake assay (15-min incubation with sodium cyanide in an argon-sparged atmosphere) inhibited 52% of the hydrogenase activity compared to the activity in a cell suspension with no inhibitor added, and the addition of 1.0 mM cyanide inhibited 90% of the H2 uptake activity (compared to the no-inhibitor-added cell suspension). The cyanide additions did not affect the methylene blue-dependent H2 uptake activity (i.e., the H2-splitting hydrogenase reaction), so the inhibitor must be acting at the level of the O2-binding heme-containing proteins, as expected.

Hydrogenases that consume molecular H2 typically have high affinities for the substrate. By performing H2 uptake assays amperometrically (with O2 as the terminal acceptor), including assays with limiting H2 levels, we determined the half-saturation affinity for H2 by wild-type S. enterica serovar Typhimurium to be 2.1 μM. This was determined from the activities of whole cells at nine separate H2 concentrations and from a linear transformation of the data in the form of a double-reciprocal plot yielding the line equation y = bx + a, where b and a are 1.1475 and 0.5545, respectively.

Mutant strain characterizations.

Strains with mutant NiFe hydrogenases were made by creating mutations in the STM3147, STM1538, and STM1786 genes. The mutated genes are referred to as hydrogenase gene groups I, II, and III, respectively (see Materials and Methods). Table Table22 provides the STM numbers for the disrupted large subunit of the hydrogenase gene for the various mutants, the TIGR-assigned gene designation for each hydrogenase, and our measured O2-dependent H2 oxidation activities of the wild-type and mutant strains. Individual strains with single mutations in each of the three hydrogenase gene groups all had decreased O2-dependent H2 uptake activity compared to that of the parent strain (Table (Table2);2); this finding indicates that each of the three enzymes contributes to respiratory H2 oxidation. Still, one of the three enzymes (encoded by a gene in group I, designated hybC) is a lesser contributor to the overall activity (under laboratory conditions) than the other two hydrogenases. All double mutant combinations showed further reduced activity compared to that of the parent or the single mutant strains. Only the mutant strain lacking all three hydrogenases failed to oxidize H2. The growth rates in LB liquid medium for the wild-type strain and the triple mutant were the same (data not shown).


To assess the ability of the strains to cause disease, a common mouse model was used (15). This assay uses death as the end point, as the bacterium is highly invasive in mice, resulting in typhoid fever-like symptoms. The results are shown in Fig. Fig.2.2. As observed in other S. enterica serovar Typhimurium virulence studies, the wild-type strain caused most of the mice to die within 10 days of oral administration. All double mutant strain combinations were either as virulent as the parent strain (Fig. (Fig.2)2) or, for the strains containing only hydrogenase II or III, somewhat less virulent than the parent. Nevertheless, the presence of any one of the three hydrogenases is sufficient for the bacterium to cause severe disease, as the virulence characteristics of all three double mutant strain combinations showed that all three enzymes are individually sufficient for virulence (i.e., at least 50% of inoculated mice were dead at day 11 postinoculation with all double mutant strains). The strain containing only hydrogenase I retained nearly full virulence, but the importance of hydrogenases II and III are shown by the result that a hydrogenase I mutant had nearly the same virulence capacity as the wild type (data not shown).

The triple mutant strain JSG321 was clearly less virulent than the parent strain (for the statistical analysis used, see the legend to Fig. Fig.2);2); indeed, out of 30 inoculated mice with that strain, none died. The expression levels of the three hydrogenase enzymes within the animal are not known, but from the in vivo results, it is clear that the hydrogenase of group I, a minor contributor to the overall activity of lab-grown bacteria, is an important enzyme for virulence. The tissue-specific expression of the individual enzymes will be an important area of study in the future, as will comparisons of the intraperitoneal versus oral routes of infection.

Organ bacterial burden.

The colonization numbers of the triple mutant and wild-type strains in the liver and spleen were determined 4 days postinoculation (106 cells introduced orally into each of the four mice for each bacterial strain) with the result that viable S. enterica serovar Typhimurium bacteria were recovered from the organs of mice inoculated with the parent (H2-using) strain, but no cells were recovered from mice inoculated with the triple mutant. The ranges of colonization numbers (S. enterica serovar Typhimurium bacteria recovered 96 h after inoculation) among four mice inoculated with the wild type were 5.0 × 104 to 1.9 × 105 CFU per liver and 3.0 × 104 to 1.8 × 105 CFU per spleen. Therefore, it is possible that the mutant strain is eliminated from the intestine or during transit to the mesenteric lymph nodes.

Complementation of the triple mutant.

To rule out the possibility that an unknown mutation (unrelated to H2 metabolism) occurred in the triple mutant that severely affects its virulence, we reintroduced one of the three hydrogenases into the triple mutant strain. Strain JSG2495 is the triple mutant strain containing the genes corresponding to the group III deletion on a low-copy-number vector; its H2 oxidation and virulence characteristics were studied in comparison to those of the triple mutant strain containing the vector alone (strain JSG2497) and of the wild-type strain and the group I and II double mutant strain JSG315 (i.e., the latter contains only hydrogenase corresponding to gene group III). In an experiment like that described for Table Table2,2, the H2 oxidation activity of the partially complemented triple mutant strain (JSG2495) was (in nanomoles of H2 per minute per 109 cells; mean ± standard deviation for four replicates) 4.1 ± 0.5, about twice that of strain JSG315 (1.9 ± 0.4) and about one-third that of the wild-type strain (14.4 ± 2.0). No activity could be detected when four replicate samples of the triple mutant strain containing the vector alone (strain JSG2497) were assayed. The morbidity of mice associated with these strains was determined as described above for the other mutant strains. Out of five mice inoculated with either the wild type, JSG2495, JSG315, or JSG2497 (the mutant strain containing the vector only), zero, two, two, and five mice, respectively, were alive at day 17 postinoculation. Two of the three mice that died within 15 days due to inoculation with strain JSG2495 died within 11 days of the oral administration. Therefore, the introduction of one of the hydrogenases into the triple mutant restored significant virulence capacity to a degree that was similar to that of the strain that is able to synthesize only group III hydrogenase.

Hydrogen levels in tissues.

It was proposed that H2 produced by colonic bacteria might reach tissues within animals by a combination of cross-epithelium diffusion and vascular-based transport processes (5, 12). Molecular hydrogen levels ranged from 118 to 239 μM in the small intestine of live mice (the mean value for 12 determinations was 168 μM), and spleen tissue H2 levels were similar (approximately 43 μM) to what was reported previously for liver tissue (6). In either case, these H2 levels are higher than the level needed to essentially saturate whole-cell hydrogenase, based on affinities of the lab-grown bacteria for H2. For the intestine and for the liver or spleen, the measured levels were about 80- and 20-fold, respectively, above the cells' half-saturation value of about 2 μM. We should thus expect rapid turnover of the H2-using hydrogenases, especially in intestinal environments where enteric bacteria can thrive.

Based on our knowledge of H2 respiration (16), the use of H2 in an O2-dependent respiratory pathway by Salmonella would be expected to result in ATP production to bolster cell growth. The animal results described here demonstrate the importance of H2 use by an enteric bacterium for survival and growth in vivo. It is likely that this is a common mechanism of energy generation by enteric pathogens within the host. The intestinal flora is the presumed source of H2 (5), and the fermentation reactions they carry out would be expected to provide a continuous supply of molecular hydrogen. In addition, the host does not use this high-energy substrate, so even under conditions in which the host is nutrient poor in terms of sugars or peptides in the serum, the pathogen can grow. With the uniformly high affinity for H2 by studied uptake hydrogenases, the pathogen should readily oxidize the high-energy reductant, even under conditions of low H2 in the bloodstream (lower than we have measured). The advantage to the pathogen is obvious, as the host cannot use this substrate.

Some of the pathogenic enteric bacteria can now be considered to use molecular hydrogen as a critical growth substrate in the host animal; this finding adds an important group of bacteria to our very limited range of pathogens (like H. pylori) able to use H2 as a growth substrate to colonize the host (12). H. pylori contains a single H2-utilizing, membrane-bound hydrogenase with a high H2 affinity (12). Another gastrointestinal pathogen that is physiologically similar to H. pylori and contains a NiFe H2 uptake hydrogenase is Campylobacter jejuni. For studies of any pathogen that uses H2, further virulence experiments using germ-free mice, and therefore lacking the presumed source of the substrate H2, would be informative. The reason that enteric bacteria have three similar and all active H2-using enzymes is unknown but could be related to the different environments they encounter. It will therefore be interesting to determine the affinities for H2 of each of the hydrogenases separately, as well as to determine the tissue-specific expression of each H2-utilizing enzyme; it would be expected that the high-affinity enzymes would be most useful to the bacterium in the tissues with the lowest H2 levels (i.e., liver and spleen) and that the lowest-affinity enzymes could function well in tissues with abundant H2 levels (i.e., the intestine). H. pylori contains a single H2-utilizing hydrogenase, also with a low half-saturation affinity (1.8 μM) for the gas (12), but the bacterium has a very limited colonization range in the host. The identification of agents that selectively inhibit bacterial hydrogenases (with their unique active centers containing Ni, Fe, CN, and CO) may represent potential therapeutic strategies for the elimination of Salmonella-based and other enteric infections.


Editor: J. B. Bliska


1. Böck, A., and G. Sawers. 2002. Fermentation, p. 262-282. In F. C. Neidhardt et al. (ed.), Escherichia coli and Salmonella: cellular and molecular biology, 2nd ed., vol. 1. ASM Press, Washington, D.C.
2. Datsenko, K. A., and B. L. Wanner. 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. USA 97:6640-6645. [PMC free article] [PubMed]
3. de Bruyn, G. 2000. Infectious disease: diarrhea. West. J. Med. 172:409-412. [PMC free article] [PubMed]
4. Jamieson, D. J., R. G. Sawers, P. A. Rugman, D. H. Boxer, and C. F. Higgins. 1986. Effects of anaerobic regulatory mutations and catabolite repression on regulation of hydrogen metabolism and hydrogenase isoenzyme composition in Salmonella typhimurium. J. Bacteriol. 168:405-411. [PMC free article] [PubMed]
5. Maier, R. J. 2003. Availability and use of molecular hydrogen as an energy substrate for Helicobacter species. Microbes Infect. 12:1159-1163. [PubMed]
6. Maier, R. J., J. Olson, and A. Olczak. 2003. Hydrogen-oxidizing capabilities of Helicobacter hepaticus and in vivo availability of the substrate. J. Bacteriol. 185:2680-2682. [PMC free article] [PubMed]
7. Maier, R. J., C. Fu, J. Gilbert, F. Moshiri, J. W. Olson, and A. G. Plaut. 1996. Hydrogen uptake hydrogenase in Helicobacter pylori. FEMS Microbiol. Lett. 141:71-77. [PubMed]
8. Maier, R. J., and J. Prosser. 1988. Hydrogen-mediated mannose uptake in Azotobacter vinelandii. J. Bacteriol. 170:1986-1989. [PMC free article] [PubMed]
8a. Menon, N. K., C. Y. Chatelus, M. Dervartanian, J. C. Wendt, K. T. Shanmugam, H. D. Peck, Jr., and A. E. Przybyla. 1994. Cloning, sequencing, and mutational analysis of the hyb operon encoding Escherichia coli hydrogenase 2. J. Bacteriol. 176:4416-4423. [PMC free article] [PubMed]
9. Merberg, D., E. B. O'Hara, and R. J. Maier. 1983. Regulation of hydrogenase in Rhizobium japonicum: analysis of mutants altered in regulation by carbon substrates and oxygen. J. Bacteriol. 156:1236-1242. [PMC free article] [PubMed]
10. Noether, G. E. 1971. Introduction to statistics: a fresh approach. Houghton Mifflin, Boston, Mass.
11. Olczak, A. A., J. W. Olson, and R. J. Maier. 2002. Oxidative-stress resistance mutants of Helicobacter pylori. J. Bacteriol. 184:3186-3193. [PMC free article] [PubMed]
12. Olson, J. W., and R. J. Maier. 2002. Molecular hydrogen as an energy source for Helicobacter pylori. Science 298:1788-1790. [PubMed]
13. Salyers, A., and D. D. Whitt. 2002. Bacterial pathogenesis: a molecular approach. ASM Press, Washington, D.C.
13a. Sargent, F., S. P. Ballentine, P. A. Rugman, T. Palmer, and D. H. Boxer. 1998. Reassignment of the gene encoding Escherichia coli hydrogenase 2 small subunit: identification of a soluble precursor of the small subunit in a hypB mutant. Eur. J. Biochem. 255:746-754. [PubMed]
14. Sawers, R. G., D. J. Jamieson, C. F. Higgins, and D. H. Boxer. 1986. Characterization and physiological roles of membrane-bound hydrogenase isoenzymes from Salmonella typhimurium. J. Bacteriol. 168:398-404. [PMC free article] [PubMed]
15. Tamayo, R., S. S. Ryan, A. J. McCoy, and J. G. Gunn. 2002. Identification and genetic characterization of PmrA-regulated genes involved in polymyxin B resistance in Salmonella enterica serovar Typhimurium. Infect. Immun. 70:6770-6778. [PMC free article] [PubMed]
16. Vignais, P. M., B. Billoud, and J. Meyer. 2001. Classification and phylogeny of hydrogenases. FEMS Microbiol. Rev. 25:455-501. [PubMed]
17. Wang R. F., and S. R. Kushner. 1991. Construction of versatile low-copy-number vectors for cloning, sequencing and gene expression in Escherichia coli. Gene 100:195-199. [PubMed]

Articles from Infection and Immunity are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...