Helicobacter hepaticus Hydrogenase Mutants Are Deficient in Hydrogen-Supported Amino Acid Uptake and in Causing Liver Lesions in A/J Mice
http://www.100md.com
感染与免疫杂志 2005年第9期
Department of Microbiology
Department of Pathology, College of Veterinary Medicine, University of Georgia, Athens, Georgia 30602
ABSTRACT
Helicobacter hepaticus, a causative agent of chronic hepatitis and hepatocellular carcinoma in mice, expresses a nickel-containing hydrogen-oxidizing hydrogenase enzyme. Growth of a hyaB gene-targeted mutant was unaffected by the presence of hydrogen, unlike the wild-type strain, which showed an enhanced growth rate when supplied with H2. Hydrogenase activities in H. hepaticus were constitutive and not dependent on the inclusion of H2 during growth. Addition of nickel during growth significantly stimulated both urease (for wild-type and hyaB) and hydrogenase (for wild-type) activities. In a 5-h period, the extent of 14C-labeled amino acid uptake by the wild type was markedly enhanced in the presence of hydrogen and was >5-fold greater than that of the hyaB mutant strain. In the presence of H2, the short-term whole-cell amino acid uptake Vmax of the parent strain was about 2.2-fold greater than for the mutant, but the half-saturation affinity for amino acid transport was the same for the parent and mutant strain. The liver- and cecum-colonizing abilities of the strains was estimated by real-time PCR quantitation of the H. hepaticus-specific cytolethal distending toxin gene and showed similar animal colonization for the hyaB mutant and the wild type. However, at 21 weeks postinoculation, the livers from mice inoculated with wild type exhibited moderate lobular lymphoplasmacytic hepatitis with hepatocytic coagulative necrosis, but the hydrogenase mutants exhibited no histological evidence of lobular inflammation or necrosis.
INTRODUCTION
In recent years, more than three different Helicobacter species have been recovered from rodents (24, 25, 31), with H. hepaticus being the most well-studied enterohepatic Helicobacter species. First isolated in 1992 from untreated A/JCr control mice, H. hepaticus is a gram-negative, microaerophilic bacterium (6) that occurs naturally in many strains of inbred mice (30). Mice infected with this bacterium develop chronic hepatic lesions and are more prone to developing hepatocellular carcinoma (25, 26), which has made H. hepaticus an excellent model for studying mechanisms of bacterium-associated liver carcinogenesis. Although H. hepaticus can be isolated from the liver of infected mice, it is more consistently recovered from the intestinal tract since the primary site of colonization is the lower bowels of mice (6, 7). Recently, Helicobacter species DNA was reported to be associated with primary hepatocellular carcinomas in human liver samples (9, 23).
At the same time, a hydrogen uptake hydrogenase enzyme well studied for its roles in nonpathogenic bacteria was demonstrated to be important for colonization of animals by some human pathogens (i.e., by Salmonella enterica serovar Typhimurium and Helicobacter pylori [14, 21]); this sparked our interest in studying the physiological role of this enzyme in H. hepaticus (16). The colonization deficiency of hydrogenase structural gene mutants of H. pylori was attributed to their inability to utilize hydrogen as an energy substrate. H. hepaticus hydrogenase is also a hydrogen uptake hydrogenase that can oxidize molecular hydrogen to yield protons and electrons; the low-potential electrons can enter the electron transport chain and thus help in energy conservation (15, 29). The hydrogen content in the liver of live mice is comparable to that in the stomach (50 μM), but in the small intestine the hydrogen content is almost fourfold higher than in the stomach (14). The whole-cell Km of H. hepaticus hydrogenase for hydrogen is ca. 2.5 μM, making it a high-affinity enzyme (16). The source of hydrogen in the animal tissues is attributed largely to the fermentation reactions leading to formation of acetate and butyrate with hydrogen as a by-product (13) due to the necessity for electron disposal.
To investigate whether the energy-conserving role of hydrogenase is useful for colonization of H. hepaticus, a gene-targeted mutant strain was generated in the structural gene (hyaB) encoding for the large subunit of hydrogenase. This mutant strain was compared to the wild type for physiological and colonization abilities. Histopathological studies of the infected tissues were also carried out to compare the pathological effects caused by the wild type and the mutant. A transposon-generated mutant strain of H. hepaticus was described recently (32). Therefore, the present study represents only the second description of a mutant strain of H. hepaticus and the first non-transposon-mediated targeted mutagenesis of H. hepaticus so that possible virulence factors can now be readily assessed.
MATERIALS AND METHODS
Bacterial strains and growth conditions. H. hepaticus strain ATCC 51449 and Escherichia coli DH5 (BRL) were used for genetic manipulations. H. hepaticus was grown on brucella agar plates (Difco) supplemented with 10% defibrinated sheep blood (BA) with either chloramphenicol for the hyaB mutant (30 μg/ml) or without any antibiotics. Plates were incubated microaerobically at 37°C in an incubator (5% CO2 and 1% O2). Broth experiments (growth and amino acid uptake studies) were performed in Mueller-Hinton broth (MHB) with either 2 or 5% fetal bovine serum (FBS) in bottles with gas-controlled environment. The atmosphere contained 85% N2, 5% CO2, and 10% H2. E. coli was grown on Luria-Bertani medium supplemented with ampicillin (100 μg/ml) or chloramphenicol and grown at 37°C aerobically. Fecal and tissue samples were plated on BA medium with amphotericin B (10 μg/ml), vancomycin (10 μg/ml), and cefoperazone (20 μg/ml).
Construction of H. hepaticus hyaB (HH 0057) mutant strain by insertional mutagenesis. The hyaB gene, which encodes for the large subunit of hydrogenase (26, 28), lies between hyaA (which encodes for the small subunit of hydrogenase) and hyaC (which encodes for a cytochrome-containing portion of the hydrogenase-specific respiratory pathway) (Fig. 1A). The downstream genes hyaD and HH0060 (i.e., hydE) are also a part of the hydrogenase-specific operon, since they are orthologues of the hydD gene and hydE of H. pylori and have been shown to be required for synthesis of hydrogenase (3). Primers hyaBF (5'-TTCGTGGTATGAGGATAATCAGCC-3') and hyaBR (5'-AATAAAGCACAACTCCCGTGAGAG-3') were used to PCR amplify a 1,303-bp fragment with wild-type H. hepaticus ATCC 51449 genomic DNA as a template. This fragment contained partial sequences of hyaB and its adjacent gene, hyaC. The PCR fragment was ligated into pGEM-T vector (Promega) according to the manufacturer's instructions. The cloned construct was obtained by transforming the ligation mixture into E. coli DH5. Subsequently, a chloramphenicol (Cm) resistance cassette (30) was inserted into a unique AvaI site within hyaB giving the construct, pGEM-T:hyaB:Cm (Fig. 1). The recombinant plasmid was introduced into H. hepaticus by electrotransformation (pulse of 2.5 kV in a Transporator Plus apparatus; BTX). As a result of allelic exchange, the hyaB mutant strain was obtained by plating transformants on chloramphenicol-containing medium.
Hydrogenase assays of wild type and hyaB mutant strain. Since the mutation was within the large subunit of hydrogenase, the strain was assayed for hydrogenase activity by using an amperometric method (15). Nickel has been shown to be required for the posttranslational processing of the large subunit of hydrogenase in E. coli, Alcaligenes eutrophus, Bradyrhizobium japonicum, and other bacteria since the catalytic activity of hydrogenases depends on the presence of nickel in its active site (10, 3, 12). H. hepaticus wild-type and hyaB mutant strains were grown on BA medium supplemented with different concentrations of NiCl2 (0, 1, 2, 6, 8, 10, or 20 μM) to address possible differences in the hydrogenase (15) and urease activities (21) since the nickel concentration (supplementation) was varied. Plates were incubated at 37°C in an anaerobic jar sparged with anaerobic gas mixture (10% H2 and 5% CO2 balance N2). Whole cells were used for the assay, with O2 provided as the final electron acceptor (Table 1).
Growth of wild type and hyaB mutant in the presence or absence of hydrogen. In order to study the effects of hydrogen on the growth, wild-type and hyaB mutant strains were grown in the presence of 10 μM NiCl2 and with 10% argon versus with 10% hydrogen added. Serum bottles containing MHB were sparged with N2 to expel all of the air. CO2 (5%), O2 (2%), and H2 (10%) or argon (10%) were injected into the bottles (percent of the headspace). After autoclaving, FBS and NiCl2 solution (10 μM) were injected. In some experiments 2% serum was used and in others 5% was used, as indicated in the text or figure legends. All of the growth and amino acid uptake experiments were performed with 5% serum initially, since that level was used in previous H hepaticus studies (11, 18). After it was determined that use of a lower serum level (2%) facilitated H2-dependent studies (see below), all (growth and amino acid transport) experiments were repeated with that serum level as shown. The increase in growth was measured spectrophotometrically (A600) and is expressed as absorbance units (optical density [OD]) (Fig. 2). OD was converted to cells/ml by using a standard curve based on a direct (microscopic) counting approach (16). From this curve, an OD of 1.0 is approximately equal to 1.6 x 109 cells/ml of the broth medium. The growth abilities of both the wild type and the mutant strain were reduced by supplementing the medium with 20 μM nickel (data not shown).
14C-labeled amino acid uptake ability of hyaB mutant and wild type. In order to investigate whether membrane-bound hydrogenase can use hydrogen as an energy substrate for uptake of amino acids, H. hepaticus wild-type and hyaB mutant cells were grown overnight on a shaker at 37°C in MHB with 2% FBS in 160-ml serum bottles, with four bottles per strain. These tightly sealed bottles were previously sparged with N2 to expel all of the air and injected with 2% O2 and 5% CO2 (vol/vol of headspace). After the OD at 600 nm (OD600) reached 0.08, 10% H2 (vol/vol of headspace) was injected into two of the bottles for each strain, and 10% argon (vol/vol of headspace) was injected into the other two. After 2 h, 14C uniformly labeled amino acids (Amersham Biosciences CFB104; specific activity of 52 mCi/milliatom) were injected into the bottles so that the final specific activity in the medium within the serum bottles was 1.0 μCi/ml. For the 0-h reading, 2 ml of the culture suspension was immediately withdrawn from the bottles (after injection of the label and mixing the solution) by using a sterile syringe, and four separate samples (2 ml) were added to wells of a filtering manifold under vacuum (17). The filter used was a 0.22-μm-pore-size membrane (Millipore GSWP 02500). After the cell suspension was filtered, the membrane was washed with two separate 1-ml aliquots of sterile phosphate-buffered saline (PBS; pH 7.2) to wash off the 14C-labeled amino acids externally adhering to the cells, and then the membranes were transferred into individual scintillation vials. Then, 5 ml of the fluor (ScintiVerse; Fischer Scientific) was added. Scintillation spectrometry was done as described previously (17). The background level uptake of 14C-labeled amino acids (at 0 h) was estimated after averaging the four separate 2-ml samples. After background sampling, the bottles were allowed to shake at 37°C for 2 and 5 h, and 2-ml samples (total of eight replicate samples) of the culture suspension were taken and filtered as described above. The average cpm/2 ml of filtered cells at 0 h was subtracted from the averaged 2- and 5-h time points, resulting in the data shown (Table 2).
For studying whole-cell kinetic parameters, a series of limiting substrate levels (14C-labeled amino acids) were added to 10-ml volumes of cells in sealed (60-ml) vials. The level of activity (substrate) added ranged from 1.0 to 0.04 μCi/ml (nine different concentrations were used) of cell suspension. The cells had been pregrown to an OD600 of 0.08 as described above, and then H2 was added for the 2-h period (10% [vol/vol] of headspace) in tightly sealed bottles previously sparged with N2 to expel all of the air and also injected with 2% O2 and 5% CO2 (vol/vol of headspace). The cells were transferred to the 60-ml vials (containing the same atmosphere as described above) by using argon-sparged syringes. Therefore, the cells never suffered an O2 shock or a change in gas (H2 or argon) composition exposure during the (20 min) uptake assay. For kinetics, the transport assay was performed within the linear uptake activity time frame of incubation with substrate. The incubation with substrate proceeded for 20 min, and four 2-ml samples were individually filtered and 14C counted as described above. Filtered samples from the medium alone with the individual label activity were averaged and subtracted from the experimental (cell) samples. The half-saturation affinity and the maximum velocity of the uptake were estimated from a linear transformation of the data in the form of a double-reciprocal plot yielding a line equation of y = bx + a. The y-intercept parameter (variable a in the equation) for each strain is reported in the text. A saturation constant in terms of Km cannot be assigned to the data, since the serum already contains amino acids at an unknown level, but relative half-saturation affinities and Vmax for the wild type compared to the mutant could be accurately estimated.
Mouse infection. Six-week-old inbred male A/J mice, certified to be Helicobacter-free (i.e., routinely tested for the absence of H. hepaticus, H. bilis, and H. muridarum), were obtained from Jackson Laboratories (Bar Harbor, Maine). The animals were housed with autoclaved food, water, and bedding and were quarantined from other experimental animals. Cage changes were performed in a laminar flow hood. Animals were housed in groups of four per microisolator cage.
H. hepaticus wild-type or mutant strains were grown for 72 h on BA plates and harvested by centrifugation with one wash step in PBS (pH 7.2). The OD of the inoculum was adjusted to 0.32 at 600 nm, which is approximately equivalent to 5 x 108 cells/ml (16). The suspension was maintained in argon-sparged buffer to minimize oxygen exposure. Two doses (0.2 ml each) of this suspension of the wild-type or the hyaB mutant culture were administered to each mouse by oral gavage on alternating days. Ten mice were inoculated for each bacterial strain. Six control mice were sham inoculated with 0.2 ml of sterile PBS on alternating days. Fecal pellets were sampled from the cages after completion of 20 weeks. After 21 weeks the mice were euthanized, and the liver and cecum were excised for qualitative, quantitative, and histopathological analysis. The proximal colon was excised only for histopathological testing.
Qualitative analysis of liver, cecum, and fecal samples. A total of 200 mg of fecal pellets collected from the cages at 20 weeks postinoculation was homogenized in 5 ml of sterile PBS. The suspension (200 μl) was plated on BA plates with triple antibiotic (amphotericin B [10 μg/ml], vancomycin [10 μg/ml], and cefoperazone [20 μg/ml]). The plates were kept in a microaerobic incubator (1% O2, 5% CO2) at 37°C for 72 to 96 h.
Mice were euthanized 21 weeks postinoculation, and the liver and cecum were excised. Liver (left, right, caudate, and part of the median lobe) and cecum (the entire tissue after the removal of a sample for histopathology) were first homogenized by using a tissue tearor (Biospec Products, Inc., Bartlesville, OK). The homogenized tissue suspensions (200 μl) were plated on BA plates with triple antibiotic to check for the presence or absence of H. hepaticus. For verifying H. hepaticus growth from tissues or suspensions (see above), microscopy, catalase test, and PCR approaches were all used. Part of the median lobe of the liver was used for histopathology.
Quantitative analysis of liver and cecum. It is commonly acknowledged that counting colonies of H. hepaticus is a challenging protocol, since the cells (even as colonies) tend to grow as a continuous spreading lawn. Hence, we performed a real-time PCR to quantify H. hepaticus in the tissues (7, 8). In addition, real-time PCR is a rapid and sensitive technique which has been used to amplify and detect H. hepaticus DNA from tissue isolates of infected mice (1, 2, 32). After homogenization of the liver and cecum, the DNA was extracted from 25 mg of the tissue by using the DNeasy tissue kit (QIAGEN). Quantitative analysis was performed by real-time PCR (iCycler Thermal Cycler; Bio-Rad) with the H. hepaticus-specific cdtB primers (cdtBF 5'-GGCTAGATACAAGAATCGCTAAT-3' and cdtBR 5'-CTACCTACTACCGCATAATCAAG-3'), which produce a 109-bp amplicon. A cdtB probe (5'-DFAM-CCGTATTACTGCTCCAAACTCTGCTACAC-DBH1-3') was also used. Specific primers and probe were designed by using the Beacon Designer Software (version 3.0; Premier Biosoft by Bio-Rad). Primers and probe for the cdtB gene (encoding for subunit B of the cytolethal distending toxin) were shown to be highly specific for H. hepaticus by doing a BLAST search against known genomes. The standard curve of threshold cycle versus log DNA was linear in the range of cycles from 18 to 29 and included 18 points from log DNA concentration levels between 3.2 and 7. The tissue homogenates were evaluated within this linear range, where 2 fg equaled one genome.
Real-time PCR analysis was performed by using a 50-μl mixture containing 25 μl of iQ Supermix (Bio-Rad), 200 nM concentrations each of cdtF and cdtR primers, cdtB probe, and 10 μl of tissue DNA (equivalent to 1.25 mg of tissue). Conditions for real time were 1 cycle at 95°C for 3 min and 40 cycles each of 95°C for 30 s and 58°C for 30 s. Samples ranging from 102 to 107 fg of H. hepaticus genomic DNA were used to generate a standard curve (not shown). A linear standard curve with a correlation coefficient of 0.982 was obtained by plotting CT values versus log DNA concentration in femtogram ranges. The CT values (threshold cycle) ranged between 18 and 29 for 18 points, encompassing log DNA concentrations between 3.2 and 7.0. The tissue sample DNA was estimated and converted to the number of copies of H. hepaticus genome by using the following calculations: the genome size of H. hepaticus is 1.8 Mbp which is equivalent to 11.88 x 108 g/mol (since 1 bp = 660 g/mol). According to Avogadro's number, 1 mol of any substance contains 6 x 1023 molecules. Thus, 11.88 x 108 g will contain 6 x 1023 molecules, or 1 g will contain 5 x 1014 molecules; therefore, 2 fg of DNA will be equivalent to one molecule or one copy of the H. hepaticus genome.
Histopathology studies. The tissue samples measuring about 1 cm by 1 cm (median lobe of liver, cecum, and proximal colon) were fixed overnight in 10% buffered formalin and were routinely processed and embedded in paraffin. Next, 5-μm-thick sections were stained with hematoxylin-eosin stain and were graded semiquantitatively for different histological parameters as follows: portal inflammation (0, no inflammation; 1, mild infiltrate in minority of portal tracts; 2, mild to moderate infiltrate in almost all portal tracts); biliary lesions/oval cell changes (0, no lesions; 1, occasional or mild cholangitis); and lobular inflammation and/or hepatocytic necrosis (0, absence; 1, mild; 2, moderate). The lobular inflammation scores (including hepatocytic coagulative necrosis) that were positive were further subranked for subjection to a Wilcoxon rank test based on their inflammation ranking. For example, the first 8 rankings (all wild-type samples) were listed first in the sequential rank, and then the remaining 12 rankings (no inflammation) were equally weighted among the two wild-type and ten mutant samples, as is permitted by the Wilcoxon test (19). Among the 20 rankings, the 10 assigned to the wild type are compared against the 10 mutant strain assignments for statistical ranking comparison. This statistical test, which clearly demonstrated a difference in lobular inflammation between the wild type and the hyaB strain, was not applied to the other histopathology data, and no statistical differences are claimed for that data.
RESULTS
Insertional mutagenesis of hyaB gene. The disruption of the hyaB gene was confirmed by PCR as an increased size of the PCR product on an agarose gel (2,150 bp for the hyaB mutant versus 1,303 bp for the wild type). The chloramphenicol resistance (Cmr) cassette was inserted in the same direction as the hyaB gene; the organization of the genes involved and the Cmr construct are depicted in Fig. 1. To establish that these two mutations are nonpolar, an ideal method would be to carry out complementation studies, but there are limitations to the genetic techniques currently available to manipulate H. hepaticus. Nevertheless, based on the annotated sequence (27) the only genes expected to be interrupted even if the insertion is polar, are ones related to hydrogenase-specific enzyme maturation (3; see also Materials and Methods).
H. hepaticus colonies tend to grow as a continuous lawn; thus, it is quite rare to obtain isolated colonies upon plating. However, the number of transformants obtained was extremely low for introducing hyaB mutations: only two or three transformant colonies were obtained per 108 to 109cells. However, since the number of transformants was very low, the margins of each colony could be easily differentiated from one another on the BA plate.
Nickel affects on hydrogenase and urease activities in the wild type and the hyaB mutant. Nickel is required for the posttranslational processing of the large subunit of hydrogenases of H2-oxidizing bacteria, and their catalytic activity depends on the presence of nickel at the unique active site (12, 20, 29). Initial experiments indicated that growth of the strains in liquid medium was facilitated by the addition of (up to 10 μM) nickel, but such stimulation was variable and could not be attributed to the Ni-hydrogenase enzyme; this was because some of the metal-dependent stimulation was observed in both the parent and the hyaB strain. We observed more than a 25-fold increase in urease activity in the presence of 10 μM added nickel versus conditions without nickel supplementation for both (hyaB and wild-type) bacterial strains (Table 1). Even a 1 μM nickel supplementation increased urease activity sevenfold over no supplementation, and this was a statistically significant difference (Table 1). Similar nickel stimulation of urease activity values was observed for the hyaB mutant as for the wild type. However, no statistically significant effects of Ni supplementation on growth could be observed (doubling times varied between 7.5 and 11.5 h in the absence of H2). Perhaps the variability is due to variations in the nickel content of the (purchased) medium.
Hydrogenase activities of the wild-type H. hepaticus strain also increased significantly with increasing concentrations of supplemented nickel in the medium (Table 1). Activity decreased (for both hydrogenase and urease) at more than 10 and 8 μM nickel, respectively, perhaps due to toxicity at the higher nickel concentrations. Similarly, growth was inhibited at nickel concentrations of greater than 10 μM NiCl2 supplementation. For subsequent growth experiments (see below) and to study the hydrogenase-specific effects, subsequent experiments (such as amino acid transport) always included 10 μM nickel in the medium within the closed gas atmosphere. The hyaB strain completely lacked hydrogenase activity even in the presence of externally added nickel (see footnote to Table 1), as expected due to the disruption of the (hyaB) structural gene.
Uptake of 14 C-labeled amino acids by the wild type and the hyaB mutant. The amino acid uptake abilities after 2 and 5 h by the wild-type cells in MHB with 2% fetal bovine serum and in an H2-containing atmosphere were 7- and 6.5-fold greater, respectively, than for the wild type in argon (all data summarized after subtracting the 0-h reading). The amino acid uptake levels at 2 and 5 h by the wild-type cells in hydrogen were 6- and 5.5-fold greater, respectively, than for the hyaB mutant strain in the presence of hydrogen (Table 2). The mutant strain exhibited the same (low) uptake ability in the presence of hydrogen or argon, as expected if the transport rate was markedly facilitated by hydrogen oxidation. The difference between the wild type and mutant or between the H2 versus argon condition (for the wild type) was less pronounced when the same experiment as the 2% serum condition was done but with 5% serum in the medium. Only an approximate twofold difference (H2 stimulation when H2 versus an argon atmosphere are compared) was observed in amino acid uptake (2-h time point) when the cells were grown in the higher serum content medium (see footnote for Table 2). The total amount of amino acid uptake by the wild type in the two (different serum concentrations) media cannot be compared based on our (radioactive label approach) results since the serum itself contains amino acids. Nevertheless, it is possible that H2 use may be most beneficial when the cells exhaust other available substrates.
Maximum uptake rates and kinetics of amino acid transport. The whole-cell amino acid uptake kinetic parameters were compared between the parent strain and the hyaB mutant with both strains; this was determined for both strains incubated with 10% H2 provided during the assay. In short-term assays of amino acid uptake with a series of substrate concentrations (including limiting substrate concentrations), the double-reciprocal linear transformation of the data yielded a line equation in the form of y = bx + a, where a, the inverse of the Vmax, was determined to be 0.279 for the parent strain and 0.605 for the mutant. This difference was statistically significant (P < 0.05) since the linear transformation plot was based on four replicate samples at each of nine different substrate concentrations for each strain. Therefore, the Vmax for the wild type for amino acid transport is 2.2-fold more than for the mutant strain. However, the half-saturation affinities calculated from the slope parameter (variable b in the equation above) were similar for both strains. From this result it appears that H2 augments the transport rate, but it likely does not cause synthesis of new amino acid transport components with a higher affinity for the carbon- and/or nitrogen-containing substrates (i.e., amino acids).
Growth of the wild type and hyaB mutant in the presence or absence of hydrogen. As was described for H. pylori (21), we determined that H. hepaticus hydrogenase activity is constitutively expressed in rich medium (MHB plus serum), irrespective of whether the cells are grown with (10% partial pressure H2 added to gas phase) or without added H2 (argon substituted for H2, data not shown). For H. hepaticus, activities achieved in MHB plus serum were the same (in nmol/min/109 cells, mean ± the standard deviation for five replicates) with H2 (41 ± 4) as when cells received 10% argon instead of H2 (42 ± 5). These whole-cell activities are considerably greater than we reported previously (16); this discrepancy was investigated and found to be due to the combination of using liquid cultures and MHB medium rather than a blood-based medium used previously (16). Nevertheless, the growth rate of the wild-type H. hepaticus in the presence of hydrogen was significantly better (lower doubling time) than when argon replaced H2 as 10% of the atmosphere (Fig. 2). This difference in growth was observed in three separate experiments, with each time point assayed three times for each experiment. For all three separate experiments, the maximum growth rate achieved in H2 for the wild type was approximately two times (doubling time in hours of 4.2 ± 0.7 for H2 and 8.4 ± 1.4 for argon) that of the same culture condition but lacking H2 (argon replaced hydrogen); this growth rate was statistically significantly different (P < 0.05) among the two values (argon compared to H2) for the wild type and also between the wild type in H2 compared to either of the two atmospheric conditions applied to the hyaB strain. Not surprisingly, growth of the mutant strain was not affected by H2 and was lower than for the parent strain (Fig. 2). As was observed for the amino acid uptake studies, this growth stimulation affect by H2 was less pronounced when 5% serum was used (data not shown). Nevertheless, it was observed that the serum level (2% versus 5%) used did not significantly affect the whole-cell hydrogenase activities achieved (data not shown). The maximum doubling times for the hyaB strain with or without H2 (Fig. 2) was about 12 ± 3 h; this was not statistically a significantly different result than the result for the parent strain without hydrogen.
Qualitative analysis of H. hepaticus in feces, liver, and cecum. H. hepaticus was recovered from liver and cecum homogenates (at 21 weeks after inoculation), as well as from fecal suspensions (at 20 weeks) from each mouse inoculated with the wild-type or hyaB mutant strain. Growth observed on BA plus three antibiotics was confirmed to be that of H. hepaticus by microscopy, catalase test, and PCR analysis (data not shown). Tissue homogenates and fecal suspensions from control (uninoculated) mice did not show growth of H. hepaticus on any of the plates.
Quantitative analysis of liver and cecum. The colonization efficiency of the wild-type and the hyaB mutant strain was approximated by using real-time PCR. A standard curve was used for estimating the femtograms of DNA, which was later converted to genome copy numbers (data not shown). The results of real-time analysis showed comparable genome copy numbers in the liver and cecum from the mice inoculated with either strain at 21 weeks. The average genome copy (from 10 mice for each strain) among all samples (each determination from 1.25 mg of tissue) ranged from about 1 x 103 to 2 x 103 (liver) or from about 1 x 106 up to 2.5 x 106 (cecum). However, even with many replicate samples (two from each mouse for each tissue source for a total of 20 for each tissue) the results were not statistically different among the (two) test strains. These results indicated that expression of hydrogenase may not be important for H. hepaticus colonization of the mouse liver or the cecum. No H. hepaticus DNA was detected in tissues of control mice.
Histopathology results. At necropsy, gross examination of liver, intestines and other visceral organs revealed no significant lesions in any of the mice tested. Fibrosis, lymphoid nodules, or tumors were absent in the liver of all of the inoculated mice. Biliary epithelial and oval cell changes were absent in most mice (10 animals for each bacterial strain), but one mouse inoculated with the wild type showed mild biliary hyperplasia and oval cell changes. Mild portal infiltration consisting of lymphocytes and plasma cells in some of the portal region or areas within the hepatic lobules was observed in six wild-type-inoculated mice, and moderate infiltration in almost all portal areas was observed in one (wild-type-inoculated) mouse. Mild portal infiltration in a few of the portal areas was observed in seven hyaB-inoculated mice but also in three of the six uninoculated animals. The latter (uninoculated controls) had no evidence of degeneration, inflammation, or necrosis, or neoplasm, and all were negative for H. hepaticus by real-time PCR analysis of liver and cecum. Portal infiltration differences among the stains was not significant.
A clear difference between the mice inoculated with the wild type and the hyaB mutant was seen in lobular inflammation characteristics. Most (8 of 10) of the wild-type-inoculated mice showed mild or moderate lobular inflammation (lymphoplasmacytic hepatitis) with hepatocytic coagulative necrosis, and this was statistically highly significantly different (at P < 0.01 [Wilcoxon rank analysis]) than for the hydrogenase mutant, for which no mice exhibited those characteristic hepatic lesions; none of the 10 mice inoculated with the hyaB mutant showed any evidence of lobular inflammation or necrosis. Representative photomicrographs of liver sections are shown in Fig. 3. There was no histological evidence of lesions in the cecum or the proximal colon for any test group (data not shown), and control mice had no histological evidence of lesions in the liver (Fig. 3D), cecum, or the proximal colon.
DISCUSSION
The vast majority of Helicobacter research involves the gastric pathogen H. pylori. However, many intriguing Helicobacter species exist in animals (see references 5 and 25), and physiological and virulence studies of H. hepaticus have become increasingly important due to the association of this enterohepatic Helicobacter species with chronic hepatitis in mice and due to the increased incidences of hepatocellular carcinoma in these infected animals (31, 26). H. hepaticus thus serves as an excellent model for studying liver carcinogenesis, and such studies might be relevant to human liver adenocarcinoma development (23).
H. hepaticus hydrogenase is a hydrogen uptake hydrogenase that can oxidize molecular hydrogen to yield protons and electrons which are key players in energy conserving processes of all cells. In order to investigate the role of hydrogenase in the colonization and virulence of H. hepaticus, a mutant strain was generated in the gene encoding the large subunit of hydrogenase (hyaB). The wild-type strain possessed hydrogenase activity, and the activity increased when the cells were grown in the presence of supplemented nickel. A battery of nickel transport and metal accessory proteins facilitate nickel insertion into the NiFeS center (10, 12). Similarly, urease is a nickel-containing enzyme, requiring Ni-dependent maturation machinery. Nickel supplementation was shown to facilitate urease activity in H. pylori (4, 28), but hydrogenase enzyme activity did not increase significantly in response to nickel supplementation of the medium (22). Nickel significantly stimulated the H. hepaticus activity for both key Ni enzymes (urease and hydrogenase), and this stimulation effect is greater than is observed for H. pylori. This difference in the two Helicobacter spp. could be related to the histidine rich nickel-binding protein present in H. pylori (Hpn and Hpn-like protein), which would be expected to serve as storage reservoirs for nickel; both of these histidine-rich proteins are absent in H. hepaticus. Therefore, H. hepaticus may depend greatly on exogenous nickel. As expected due to the absence of the hydrogenase structural gene, nickel supplementation could not restore the hydrogenase activity to the hyaB mutant, and yet nickel supplementation stimulated urease activity in the mutant as in the parent strain.
In vitro studies showed that liquid growth in the presence of hydrogen (doubling time) of the wild-type strain was better than that for the hyaB mutant; this was most likely due to the ability of the wild type to use hydrogen as an additional energy substrate for growth. It is noteworthy that in the initial isolation of H. hepaticus from liver, a microaerophilic atmosphere containing hydrogen was used (6, 7). Growth stimulation due to supplying H2 to diverse H2 oxidizing bacteria has been observed previously (17, 29) but has not been reported for pathogenic bacteria. For H. hepaticus, hydrogenase activity was constitutive (not dependent on the inclusion of H2 during growth), like that described for H. pylori previously (21). However, in contrast to H. pylori, H. hepaticus activity did not increase when cells were grown with H2.
A clue to the role of H2 oxidation was provided by amino acid uptake assays. Most H2-oxidizing bacteria are chemoautotrophs, using the energy from H2 to fuel the energy input needed for carbon dioxide fixation (29). The CO2 fixation enzymes are highly conserved. However, no such CO2-fixing enzymes exist in the Helicobacter spp., based on the annotated genome sequences of (two) H. pylori or (one) H. hepaticus strains. Therefore, it is expected the energy from H2 may be used to assimilate non-CO2 sources of carbon. Wild-type cells had much increased ability for uptake of 14C-labeled amino acids in the presence than in the absence of hydrogen or when compared to the hyaB mutant cells either in the presence or absence of hydrogen. This is likely due to the ability of wild-type cells to glean energy from hydrogen, and this energy form is coupled to amino acid transport. There is precedent for hydrogen oxidation providing energy for bacterial carbon transport (17). From the H. hepaticus sequence it is clear that this bacterium, like H. pylori, has abundant amino acid transport systems (27). H2 stimulation of both growth and amino acid uptake was most evident when 2% serum was used rather than 5%; this could indicate that hydrogenase may be most important to the bacterium as an alternative energy source when nutrient conditions are diminished.
Kinetic assays to determine half-saturation affinity and Vmax are normally applied to substrate use by enzymes, but such determinations have been useful to correlate whole-cell substrate use characteristics of H. pylori to in vivo virulence characteristics (21). Also, with regard to H2 oxidation-mediated affects, it was shown that H2 use by Azotobacter vinelandii increased the bacteriums Vmax for mannose uptake without altering the bacterium's affinity for the substrate (17). The Vmax value expressed by the H. hepaticus wild type (in H2) for amino acid transport was 2.2-fold greater than for the mutant strain. This would be consistent with H2 providing an energy source (perhaps ATP) to augment uptake rates. Still, the half-saturation affinities for amino acid uptake were similar for the parent and the mutant. From this latter result it appears that, although H2 stimulates transport, it likely neither causes synthesis of new amino acid transport components with a higher affinity for the carbon substrates nor preferentially enables higher affinity uptake systems to be used.
It may appear from our real-time analysis to estimate genome number (and thereby comparative colonization) by the wild type and hyaB mutant that there is no growth advantage due to H2 use within the liver or cecum. This is surprising in light of the other results and suggests that hydrogen may be only one among many growth substrates utilized by H. hepaticus within the tissues. If the tissues (especially liver) are well supplied with carbon and energy sources, this would be consistent with the idea proposed above that H2 use may be most important under low-nutrient conditions. Nevertheless, a clear difference was seen in the photomicrographs of the liver of mice inoculated with the hyaB mutant versus the wild type; these results showed complete absence of lobular inflammation by the hyaB mutant (and similar to uninoculated control mice results), whereas the livers from 8 of 10 mice inoculated with the wild type showed moderate inflammation with necrosis. Although hydrogenase may not be directly responsible for the pathogenesis seen in the liver tissue of mice inoculated with the wild type, the highly diffusible high-energy reductant is present in liver tissue of live mice (16), and energy released by hydrogen oxidation may be used to augment synthesis of virulence-related proteins or enzymes responsible for causing inflammation and necrosis. Possibly, high-efficiency solute uptake systems (such as for amino acids) that are supported by H2 oxidation may be coupled to virulence.
The animal colonization attributes of many Helicobacter spp. (including by H. hepaticus) have been well documented and reviewed (5, 25). However, this is the first evidence of the involvement of molecular hydrogen use via hydrogenase in the development of characteristics related to bacterial hepatitis, namely, inflammation and necrosis of the liver tissue. Therefore, the present study adds H. hepaticus to the small list of bacteria (i.e., H. pylori and S. enterica serovar Typhimurium) that use hydrogen to augment their virulence in animals. Further long-term studies are warranted to address the mechanisms by which H. hepaticus hydrogenase is involved in the pathogenesis of hepatitis and perhaps in carcinoma development. Finally, it is noteworthy that insertion cassette mutagenesis was successful for H. hepaticus specific gene disruption, and it is hoped this procedure can be used by many researchers for studying roles of other genes in this bacterium.
ACKNOWLEDGMENTS
This study was supported by NIH NCI grant RO3-CA103095.
We thank Sue Maier (University of Georgia) for assistance with the 14C experiments and A. Olczak for advice on protocols.
REFERENCES
1. Avenaud, P., B. Le. Bail, K. Mayo, A. Marais, R. Fawaz, P. Bioulac-Sage, and F. Megraud. 2003. Natural history of Helicobacter hepaticus infection in conventional A/J mice, with special reference to liver involvement. Infect. Immun. 71:3667-3673.
2. Battles, J. K., J. C. Williamson, K. M. Pike, P. L. Gorelick, J. M. Ward, and M. A. Gonda. 1995. Diagnostic assay for Helicobacter hepaticus based on nucleotide sequence of its 16S rRNA gene. J. Clin. Microbiol. 33:1344-1347.
3. Benoit, S., N. Mehta, G. Wang, M. Gatlin, and R. J. Maier. 2004. Requirement of hydD, hydE, hypC, and hypE genes for hydrogenase activity in Helicobacter pylori. Microb. Pathog. 36:153-157.
4. Contreras, M., J-M Thiberge, M-A Mandrand-Bertholet, and A. Labigne. 2003. Characterization of and the roles of NikR, a nickel-responsive pleiotropic autoregulator of Helicobacter pylori. Mol. Microbiol. 49:947-963.
5. Ferrero, R. L., and J. G. Fox. 2001. In vivo modeling of Helicobacter-associated gastrointestinal diseases, p. 565-582. In H. L. Mobley, G. L. Mendz, and S. L. Hazell (ed.), Helicobacter pylori: physiology and genetics. ASM Press, Washington, D.C.
6. Fox, J. G., F. E. Dewhirst, J. G. Tully, B. J. Paster, L. Yan, N. S. Taylor, M. J. Collins, Jr., P. L. Gorelick, and J. M. Ward. 1994. Helicobacter hepaticus sp. nov., a microaerophilic bacterium isolated from livers and intestinal mucosal scrapings from mice. J. Clin. Microbiol. 32:1238-1245.
7. Fox, J. G., X. Li, L. Yan, R. J. Cahill, R. Hurley, R. Lewis, and J. C. Murphy. 1996. Chronic proliferative hepatitis in A/J Cr mice associated with persistent Helicobacter hepaticus infection: a model of Helicobacter-induced carcinogenesis. Infect. Immun. 64:1548-1558.
8. Ge, Z., D. A. White, M. T. Whary, and J. G. Fox. 2001. Fluorogenic PCR-based quantitative detection of a murine pathogen, Helicobacter hepaticus. J. Clin. Microbiol. 39:2598-2602.
9. Huang, Y., X. G. Fan, Z. M. Wang, J. H. Zhou, X. F. Tian, and N. Li. 2004. Identification of Helicobacter species in human liver samples from patients with primary hepatocellular carcinoma. J. Clin. Pathol. 57:1273-1277.
10. Jacobi, A., R. Rossmann, and A. Bock. 1992. The hyp operon gene products are required for the maturation of catalytically active hydrogenase isoenzymes in Escherichia coli. Arch. Microbiol. 158:444-451.
11. Kornilov'ka, I., I. Nilsson, M. Utt, A. Ljungh, and T. Wadstrom. Immunogenic proteins of Helicobacer pullorum, Helicobacter bilis, and Helicobacter hepaticus identified by two-dimensional gel electrophoresis and immunoblotting. Proteomics 2:775-778.
12. Magalon, A., and A. Bock. 2000. Dissection of the maturation reactions of the [NiFe] hydrogenase 3 from Escherichia coli taking place after nickel incorporation. FEBS Lett. 473:254-258.
13. Maier, R. J. 2003. Availability and use of molecular hydrogen as an energy substrate for Helicobacter species. Microbes Infect. 5:1159-1163.
14. Maier, R. J., A. Olczak, S. Maier, S. Soni, and J. Gunn. 2004. Respiratory hydrogen use by Salmonella enterica serovar Typhimurium is essential for virulence. Infect. Immun. 72:6294-6299.
15. Maier, R. J., C. Fu, J. Gilbert, F. Moshiri, J. Olson, and A. G. Plaut. 1996. Hydrogen uptake hydrogenase in Helicobacter pylori. FEMS. Microbiol. Lett. 141:71-76.
16. Maier, R. J., J. Olson, and A. Olczak. 2003. Hydrogen-oxidizing capabilities of Helicobacter hepaticus and in vivo availability of substrate. J. Bacteriol. 185:2680-2682.
17. Maier, R. J., and J. Prosser. 1988. Hydrogen-mediated mannose uptake in Azotobacter vinelandii. J. Bacteriol. 170:1986-1989.
18. Myles, M. H., R. S. Livingston, B. A. Livingston, J. M. Criley, and C. L. Franklin. 2003. Analysis of gene expression in ceca of Helicobacter hepaticus-infected A/JCr mice before and after development of typhlitis. Infect. Immun. 71:3885-3893.
19. Noether, G. 1971. Introduction to statistics: a fresh approach. Houghton-Mifflin, Boston, Mass.
20. Olson, J. W., C. Fu, and R. J. Maier. 1997. The HypB protein from Bradyrhizobium japonicum can store nickel and is required for the nickel-dependent transcriptional regulation of hydrogenase. Mol. Microbiol. 24:119-128.
21. Olson, J. W., and R. J. Maier. 2002. Molecular hydrogen as an energy source for Helicobacter pylori. Science 298:1788-1790.
22. Olson, J. W., N. S. Mehta, and R. J. Maier. 2001. Requirement of nickel metabolism proteins HypA and HypB for full activity of both hydrogenase and urease in Helicobacter pylori. Mol. Microbiol. 39:176-182.
23. Pellicano, R., V. Mazzaferro, W. F. Grigioni, M. A. Cutufia, S. Fagoonee, L. Silengo, and R. M. Ponzetta. 2004. Helicobacter species sequences in liver samples from patients with and without hepatocellular carcinoma. World J. Gastroenterol. 10:598-601.
24. Phillips, M. W., and A. Lee. 1983. Isolation and characterization of a spiral bacterium from the crypts of rodent gastrointestinal tracts. Appl. Environ. Microbiol. 45:675-683.
25. Solnick, J. V., and D. B. Schauer. 2001. Emergence of diverse Helicobacter species in the pathogenesis of gastric and enterohepatic disease. Clin. Microbiol. Rev. 14:59-97.
26. Schauer, D. B. 2001. Enterohepatic Helicobacter species, p. 533-548. In H. L. T. Mobley, G. L. Mendz, and S. T. Hazell (ed.), Helicobacter pylori: physiology and genetics. ASM Press, Washington, D.C.
27. Suerbaum, S., C. Josenhans, T. Sterzenbach, B. Drescher, P. Brandt, M. Bell, M. Droge, B. Fartmann, H. P. Fischer, Z. Ge, A. Horster, R. Holland, K. Klein, J. Konig, L. Macko, G. L. Mendz, G. Nyakatura, D. B. Schauer, z. Shen, J. Weber, M. Frosch, and J. G. Fox. 2003. The complete genome sequence of the carcinogenic bacterium Helicobacter hepaticus. Proc. Natl. Acad. Sci. USA 100:7901-7906.
28. van Vliet, H., E. J. Kuipers, B. Waidner, B. J. Davies, N. de Vries, C. W. Penn, C. M. Vandenbroucke-Grauls, M. Kist, S. Bereswill, and J. G. Kusters. 2001. Nickel-responsive induction of urease expression in Helicobacter pylori is mediated at the transcriptional level. Infect. Immunol. 69:4891-4897.
29. Vignais, P. M., B. Billoud, and J. Meyer. 2001. Classification and phylogeny of hydrogenases. FEMS Microbiol. Rev. 25:455-501.
30. Wang, Y., and D. E. Taylor. 1990. Chloramphenicol resistance in Campylobacter coli: nucleotide sequence, expression and cloning vector construction. Gene 94:23-28.
31. Ward, J. M., J. G. Fox, M. R. Anver, D. C. Haines, C. V. George, M. J. Collins, Jr., P. L. Gorelick, K. Nagashima, M. A. Gonda, R. V. Gilden, et al. 1994. Chronic active hepatitis and associated liver tumors in mice caused by a persistent bacterial infection with a novel Helicobacter species. J. Natl. Cancer Inst. 86:1222-1227.
32. Young, V. B., K. A. Knox, J. S. Pratt, J. S. Cortez, L. S. Mansfield, A. B. Rogers, J. G. Fox, and D. B. Schauer. 2004. In vitro and in vivo characterization of Helicobacter hepaticus cytolethal distending toxin mutants. Infect. Immunol. 72:2521-2527.(Nalini S. Mehta, Stephane)
Department of Pathology, College of Veterinary Medicine, University of Georgia, Athens, Georgia 30602
ABSTRACT
Helicobacter hepaticus, a causative agent of chronic hepatitis and hepatocellular carcinoma in mice, expresses a nickel-containing hydrogen-oxidizing hydrogenase enzyme. Growth of a hyaB gene-targeted mutant was unaffected by the presence of hydrogen, unlike the wild-type strain, which showed an enhanced growth rate when supplied with H2. Hydrogenase activities in H. hepaticus were constitutive and not dependent on the inclusion of H2 during growth. Addition of nickel during growth significantly stimulated both urease (for wild-type and hyaB) and hydrogenase (for wild-type) activities. In a 5-h period, the extent of 14C-labeled amino acid uptake by the wild type was markedly enhanced in the presence of hydrogen and was >5-fold greater than that of the hyaB mutant strain. In the presence of H2, the short-term whole-cell amino acid uptake Vmax of the parent strain was about 2.2-fold greater than for the mutant, but the half-saturation affinity for amino acid transport was the same for the parent and mutant strain. The liver- and cecum-colonizing abilities of the strains was estimated by real-time PCR quantitation of the H. hepaticus-specific cytolethal distending toxin gene and showed similar animal colonization for the hyaB mutant and the wild type. However, at 21 weeks postinoculation, the livers from mice inoculated with wild type exhibited moderate lobular lymphoplasmacytic hepatitis with hepatocytic coagulative necrosis, but the hydrogenase mutants exhibited no histological evidence of lobular inflammation or necrosis.
INTRODUCTION
In recent years, more than three different Helicobacter species have been recovered from rodents (24, 25, 31), with H. hepaticus being the most well-studied enterohepatic Helicobacter species. First isolated in 1992 from untreated A/JCr control mice, H. hepaticus is a gram-negative, microaerophilic bacterium (6) that occurs naturally in many strains of inbred mice (30). Mice infected with this bacterium develop chronic hepatic lesions and are more prone to developing hepatocellular carcinoma (25, 26), which has made H. hepaticus an excellent model for studying mechanisms of bacterium-associated liver carcinogenesis. Although H. hepaticus can be isolated from the liver of infected mice, it is more consistently recovered from the intestinal tract since the primary site of colonization is the lower bowels of mice (6, 7). Recently, Helicobacter species DNA was reported to be associated with primary hepatocellular carcinomas in human liver samples (9, 23).
At the same time, a hydrogen uptake hydrogenase enzyme well studied for its roles in nonpathogenic bacteria was demonstrated to be important for colonization of animals by some human pathogens (i.e., by Salmonella enterica serovar Typhimurium and Helicobacter pylori [14, 21]); this sparked our interest in studying the physiological role of this enzyme in H. hepaticus (16). The colonization deficiency of hydrogenase structural gene mutants of H. pylori was attributed to their inability to utilize hydrogen as an energy substrate. H. hepaticus hydrogenase is also a hydrogen uptake hydrogenase that can oxidize molecular hydrogen to yield protons and electrons; the low-potential electrons can enter the electron transport chain and thus help in energy conservation (15, 29). The hydrogen content in the liver of live mice is comparable to that in the stomach (50 μM), but in the small intestine the hydrogen content is almost fourfold higher than in the stomach (14). The whole-cell Km of H. hepaticus hydrogenase for hydrogen is ca. 2.5 μM, making it a high-affinity enzyme (16). The source of hydrogen in the animal tissues is attributed largely to the fermentation reactions leading to formation of acetate and butyrate with hydrogen as a by-product (13) due to the necessity for electron disposal.
To investigate whether the energy-conserving role of hydrogenase is useful for colonization of H. hepaticus, a gene-targeted mutant strain was generated in the structural gene (hyaB) encoding for the large subunit of hydrogenase. This mutant strain was compared to the wild type for physiological and colonization abilities. Histopathological studies of the infected tissues were also carried out to compare the pathological effects caused by the wild type and the mutant. A transposon-generated mutant strain of H. hepaticus was described recently (32). Therefore, the present study represents only the second description of a mutant strain of H. hepaticus and the first non-transposon-mediated targeted mutagenesis of H. hepaticus so that possible virulence factors can now be readily assessed.
MATERIALS AND METHODS
Bacterial strains and growth conditions. H. hepaticus strain ATCC 51449 and Escherichia coli DH5 (BRL) were used for genetic manipulations. H. hepaticus was grown on brucella agar plates (Difco) supplemented with 10% defibrinated sheep blood (BA) with either chloramphenicol for the hyaB mutant (30 μg/ml) or without any antibiotics. Plates were incubated microaerobically at 37°C in an incubator (5% CO2 and 1% O2). Broth experiments (growth and amino acid uptake studies) were performed in Mueller-Hinton broth (MHB) with either 2 or 5% fetal bovine serum (FBS) in bottles with gas-controlled environment. The atmosphere contained 85% N2, 5% CO2, and 10% H2. E. coli was grown on Luria-Bertani medium supplemented with ampicillin (100 μg/ml) or chloramphenicol and grown at 37°C aerobically. Fecal and tissue samples were plated on BA medium with amphotericin B (10 μg/ml), vancomycin (10 μg/ml), and cefoperazone (20 μg/ml).
Construction of H. hepaticus hyaB (HH 0057) mutant strain by insertional mutagenesis. The hyaB gene, which encodes for the large subunit of hydrogenase (26, 28), lies between hyaA (which encodes for the small subunit of hydrogenase) and hyaC (which encodes for a cytochrome-containing portion of the hydrogenase-specific respiratory pathway) (Fig. 1A). The downstream genes hyaD and HH0060 (i.e., hydE) are also a part of the hydrogenase-specific operon, since they are orthologues of the hydD gene and hydE of H. pylori and have been shown to be required for synthesis of hydrogenase (3). Primers hyaBF (5'-TTCGTGGTATGAGGATAATCAGCC-3') and hyaBR (5'-AATAAAGCACAACTCCCGTGAGAG-3') were used to PCR amplify a 1,303-bp fragment with wild-type H. hepaticus ATCC 51449 genomic DNA as a template. This fragment contained partial sequences of hyaB and its adjacent gene, hyaC. The PCR fragment was ligated into pGEM-T vector (Promega) according to the manufacturer's instructions. The cloned construct was obtained by transforming the ligation mixture into E. coli DH5. Subsequently, a chloramphenicol (Cm) resistance cassette (30) was inserted into a unique AvaI site within hyaB giving the construct, pGEM-T:hyaB:Cm (Fig. 1). The recombinant plasmid was introduced into H. hepaticus by electrotransformation (pulse of 2.5 kV in a Transporator Plus apparatus; BTX). As a result of allelic exchange, the hyaB mutant strain was obtained by plating transformants on chloramphenicol-containing medium.
Hydrogenase assays of wild type and hyaB mutant strain. Since the mutation was within the large subunit of hydrogenase, the strain was assayed for hydrogenase activity by using an amperometric method (15). Nickel has been shown to be required for the posttranslational processing of the large subunit of hydrogenase in E. coli, Alcaligenes eutrophus, Bradyrhizobium japonicum, and other bacteria since the catalytic activity of hydrogenases depends on the presence of nickel in its active site (10, 3, 12). H. hepaticus wild-type and hyaB mutant strains were grown on BA medium supplemented with different concentrations of NiCl2 (0, 1, 2, 6, 8, 10, or 20 μM) to address possible differences in the hydrogenase (15) and urease activities (21) since the nickel concentration (supplementation) was varied. Plates were incubated at 37°C in an anaerobic jar sparged with anaerobic gas mixture (10% H2 and 5% CO2 balance N2). Whole cells were used for the assay, with O2 provided as the final electron acceptor (Table 1).
Growth of wild type and hyaB mutant in the presence or absence of hydrogen. In order to study the effects of hydrogen on the growth, wild-type and hyaB mutant strains were grown in the presence of 10 μM NiCl2 and with 10% argon versus with 10% hydrogen added. Serum bottles containing MHB were sparged with N2 to expel all of the air. CO2 (5%), O2 (2%), and H2 (10%) or argon (10%) were injected into the bottles (percent of the headspace). After autoclaving, FBS and NiCl2 solution (10 μM) were injected. In some experiments 2% serum was used and in others 5% was used, as indicated in the text or figure legends. All of the growth and amino acid uptake experiments were performed with 5% serum initially, since that level was used in previous H hepaticus studies (11, 18). After it was determined that use of a lower serum level (2%) facilitated H2-dependent studies (see below), all (growth and amino acid transport) experiments were repeated with that serum level as shown. The increase in growth was measured spectrophotometrically (A600) and is expressed as absorbance units (optical density [OD]) (Fig. 2). OD was converted to cells/ml by using a standard curve based on a direct (microscopic) counting approach (16). From this curve, an OD of 1.0 is approximately equal to 1.6 x 109 cells/ml of the broth medium. The growth abilities of both the wild type and the mutant strain were reduced by supplementing the medium with 20 μM nickel (data not shown).
14C-labeled amino acid uptake ability of hyaB mutant and wild type. In order to investigate whether membrane-bound hydrogenase can use hydrogen as an energy substrate for uptake of amino acids, H. hepaticus wild-type and hyaB mutant cells were grown overnight on a shaker at 37°C in MHB with 2% FBS in 160-ml serum bottles, with four bottles per strain. These tightly sealed bottles were previously sparged with N2 to expel all of the air and injected with 2% O2 and 5% CO2 (vol/vol of headspace). After the OD at 600 nm (OD600) reached 0.08, 10% H2 (vol/vol of headspace) was injected into two of the bottles for each strain, and 10% argon (vol/vol of headspace) was injected into the other two. After 2 h, 14C uniformly labeled amino acids (Amersham Biosciences CFB104; specific activity of 52 mCi/milliatom) were injected into the bottles so that the final specific activity in the medium within the serum bottles was 1.0 μCi/ml. For the 0-h reading, 2 ml of the culture suspension was immediately withdrawn from the bottles (after injection of the label and mixing the solution) by using a sterile syringe, and four separate samples (2 ml) were added to wells of a filtering manifold under vacuum (17). The filter used was a 0.22-μm-pore-size membrane (Millipore GSWP 02500). After the cell suspension was filtered, the membrane was washed with two separate 1-ml aliquots of sterile phosphate-buffered saline (PBS; pH 7.2) to wash off the 14C-labeled amino acids externally adhering to the cells, and then the membranes were transferred into individual scintillation vials. Then, 5 ml of the fluor (ScintiVerse; Fischer Scientific) was added. Scintillation spectrometry was done as described previously (17). The background level uptake of 14C-labeled amino acids (at 0 h) was estimated after averaging the four separate 2-ml samples. After background sampling, the bottles were allowed to shake at 37°C for 2 and 5 h, and 2-ml samples (total of eight replicate samples) of the culture suspension were taken and filtered as described above. The average cpm/2 ml of filtered cells at 0 h was subtracted from the averaged 2- and 5-h time points, resulting in the data shown (Table 2).
For studying whole-cell kinetic parameters, a series of limiting substrate levels (14C-labeled amino acids) were added to 10-ml volumes of cells in sealed (60-ml) vials. The level of activity (substrate) added ranged from 1.0 to 0.04 μCi/ml (nine different concentrations were used) of cell suspension. The cells had been pregrown to an OD600 of 0.08 as described above, and then H2 was added for the 2-h period (10% [vol/vol] of headspace) in tightly sealed bottles previously sparged with N2 to expel all of the air and also injected with 2% O2 and 5% CO2 (vol/vol of headspace). The cells were transferred to the 60-ml vials (containing the same atmosphere as described above) by using argon-sparged syringes. Therefore, the cells never suffered an O2 shock or a change in gas (H2 or argon) composition exposure during the (20 min) uptake assay. For kinetics, the transport assay was performed within the linear uptake activity time frame of incubation with substrate. The incubation with substrate proceeded for 20 min, and four 2-ml samples were individually filtered and 14C counted as described above. Filtered samples from the medium alone with the individual label activity were averaged and subtracted from the experimental (cell) samples. The half-saturation affinity and the maximum velocity of the uptake were estimated from a linear transformation of the data in the form of a double-reciprocal plot yielding a line equation of y = bx + a. The y-intercept parameter (variable a in the equation) for each strain is reported in the text. A saturation constant in terms of Km cannot be assigned to the data, since the serum already contains amino acids at an unknown level, but relative half-saturation affinities and Vmax for the wild type compared to the mutant could be accurately estimated.
Mouse infection. Six-week-old inbred male A/J mice, certified to be Helicobacter-free (i.e., routinely tested for the absence of H. hepaticus, H. bilis, and H. muridarum), were obtained from Jackson Laboratories (Bar Harbor, Maine). The animals were housed with autoclaved food, water, and bedding and were quarantined from other experimental animals. Cage changes were performed in a laminar flow hood. Animals were housed in groups of four per microisolator cage.
H. hepaticus wild-type or mutant strains were grown for 72 h on BA plates and harvested by centrifugation with one wash step in PBS (pH 7.2). The OD of the inoculum was adjusted to 0.32 at 600 nm, which is approximately equivalent to 5 x 108 cells/ml (16). The suspension was maintained in argon-sparged buffer to minimize oxygen exposure. Two doses (0.2 ml each) of this suspension of the wild-type or the hyaB mutant culture were administered to each mouse by oral gavage on alternating days. Ten mice were inoculated for each bacterial strain. Six control mice were sham inoculated with 0.2 ml of sterile PBS on alternating days. Fecal pellets were sampled from the cages after completion of 20 weeks. After 21 weeks the mice were euthanized, and the liver and cecum were excised for qualitative, quantitative, and histopathological analysis. The proximal colon was excised only for histopathological testing.
Qualitative analysis of liver, cecum, and fecal samples. A total of 200 mg of fecal pellets collected from the cages at 20 weeks postinoculation was homogenized in 5 ml of sterile PBS. The suspension (200 μl) was plated on BA plates with triple antibiotic (amphotericin B [10 μg/ml], vancomycin [10 μg/ml], and cefoperazone [20 μg/ml]). The plates were kept in a microaerobic incubator (1% O2, 5% CO2) at 37°C for 72 to 96 h.
Mice were euthanized 21 weeks postinoculation, and the liver and cecum were excised. Liver (left, right, caudate, and part of the median lobe) and cecum (the entire tissue after the removal of a sample for histopathology) were first homogenized by using a tissue tearor (Biospec Products, Inc., Bartlesville, OK). The homogenized tissue suspensions (200 μl) were plated on BA plates with triple antibiotic to check for the presence or absence of H. hepaticus. For verifying H. hepaticus growth from tissues or suspensions (see above), microscopy, catalase test, and PCR approaches were all used. Part of the median lobe of the liver was used for histopathology.
Quantitative analysis of liver and cecum. It is commonly acknowledged that counting colonies of H. hepaticus is a challenging protocol, since the cells (even as colonies) tend to grow as a continuous spreading lawn. Hence, we performed a real-time PCR to quantify H. hepaticus in the tissues (7, 8). In addition, real-time PCR is a rapid and sensitive technique which has been used to amplify and detect H. hepaticus DNA from tissue isolates of infected mice (1, 2, 32). After homogenization of the liver and cecum, the DNA was extracted from 25 mg of the tissue by using the DNeasy tissue kit (QIAGEN). Quantitative analysis was performed by real-time PCR (iCycler Thermal Cycler; Bio-Rad) with the H. hepaticus-specific cdtB primers (cdtBF 5'-GGCTAGATACAAGAATCGCTAAT-3' and cdtBR 5'-CTACCTACTACCGCATAATCAAG-3'), which produce a 109-bp amplicon. A cdtB probe (5'-DFAM-CCGTATTACTGCTCCAAACTCTGCTACAC-DBH1-3') was also used. Specific primers and probe were designed by using the Beacon Designer Software (version 3.0; Premier Biosoft by Bio-Rad). Primers and probe for the cdtB gene (encoding for subunit B of the cytolethal distending toxin) were shown to be highly specific for H. hepaticus by doing a BLAST search against known genomes. The standard curve of threshold cycle versus log DNA was linear in the range of cycles from 18 to 29 and included 18 points from log DNA concentration levels between 3.2 and 7. The tissue homogenates were evaluated within this linear range, where 2 fg equaled one genome.
Real-time PCR analysis was performed by using a 50-μl mixture containing 25 μl of iQ Supermix (Bio-Rad), 200 nM concentrations each of cdtF and cdtR primers, cdtB probe, and 10 μl of tissue DNA (equivalent to 1.25 mg of tissue). Conditions for real time were 1 cycle at 95°C for 3 min and 40 cycles each of 95°C for 30 s and 58°C for 30 s. Samples ranging from 102 to 107 fg of H. hepaticus genomic DNA were used to generate a standard curve (not shown). A linear standard curve with a correlation coefficient of 0.982 was obtained by plotting CT values versus log DNA concentration in femtogram ranges. The CT values (threshold cycle) ranged between 18 and 29 for 18 points, encompassing log DNA concentrations between 3.2 and 7.0. The tissue sample DNA was estimated and converted to the number of copies of H. hepaticus genome by using the following calculations: the genome size of H. hepaticus is 1.8 Mbp which is equivalent to 11.88 x 108 g/mol (since 1 bp = 660 g/mol). According to Avogadro's number, 1 mol of any substance contains 6 x 1023 molecules. Thus, 11.88 x 108 g will contain 6 x 1023 molecules, or 1 g will contain 5 x 1014 molecules; therefore, 2 fg of DNA will be equivalent to one molecule or one copy of the H. hepaticus genome.
Histopathology studies. The tissue samples measuring about 1 cm by 1 cm (median lobe of liver, cecum, and proximal colon) were fixed overnight in 10% buffered formalin and were routinely processed and embedded in paraffin. Next, 5-μm-thick sections were stained with hematoxylin-eosin stain and were graded semiquantitatively for different histological parameters as follows: portal inflammation (0, no inflammation; 1, mild infiltrate in minority of portal tracts; 2, mild to moderate infiltrate in almost all portal tracts); biliary lesions/oval cell changes (0, no lesions; 1, occasional or mild cholangitis); and lobular inflammation and/or hepatocytic necrosis (0, absence; 1, mild; 2, moderate). The lobular inflammation scores (including hepatocytic coagulative necrosis) that were positive were further subranked for subjection to a Wilcoxon rank test based on their inflammation ranking. For example, the first 8 rankings (all wild-type samples) were listed first in the sequential rank, and then the remaining 12 rankings (no inflammation) were equally weighted among the two wild-type and ten mutant samples, as is permitted by the Wilcoxon test (19). Among the 20 rankings, the 10 assigned to the wild type are compared against the 10 mutant strain assignments for statistical ranking comparison. This statistical test, which clearly demonstrated a difference in lobular inflammation between the wild type and the hyaB strain, was not applied to the other histopathology data, and no statistical differences are claimed for that data.
RESULTS
Insertional mutagenesis of hyaB gene. The disruption of the hyaB gene was confirmed by PCR as an increased size of the PCR product on an agarose gel (2,150 bp for the hyaB mutant versus 1,303 bp for the wild type). The chloramphenicol resistance (Cmr) cassette was inserted in the same direction as the hyaB gene; the organization of the genes involved and the Cmr construct are depicted in Fig. 1. To establish that these two mutations are nonpolar, an ideal method would be to carry out complementation studies, but there are limitations to the genetic techniques currently available to manipulate H. hepaticus. Nevertheless, based on the annotated sequence (27) the only genes expected to be interrupted even if the insertion is polar, are ones related to hydrogenase-specific enzyme maturation (3; see also Materials and Methods).
H. hepaticus colonies tend to grow as a continuous lawn; thus, it is quite rare to obtain isolated colonies upon plating. However, the number of transformants obtained was extremely low for introducing hyaB mutations: only two or three transformant colonies were obtained per 108 to 109cells. However, since the number of transformants was very low, the margins of each colony could be easily differentiated from one another on the BA plate.
Nickel affects on hydrogenase and urease activities in the wild type and the hyaB mutant. Nickel is required for the posttranslational processing of the large subunit of hydrogenases of H2-oxidizing bacteria, and their catalytic activity depends on the presence of nickel at the unique active site (12, 20, 29). Initial experiments indicated that growth of the strains in liquid medium was facilitated by the addition of (up to 10 μM) nickel, but such stimulation was variable and could not be attributed to the Ni-hydrogenase enzyme; this was because some of the metal-dependent stimulation was observed in both the parent and the hyaB strain. We observed more than a 25-fold increase in urease activity in the presence of 10 μM added nickel versus conditions without nickel supplementation for both (hyaB and wild-type) bacterial strains (Table 1). Even a 1 μM nickel supplementation increased urease activity sevenfold over no supplementation, and this was a statistically significant difference (Table 1). Similar nickel stimulation of urease activity values was observed for the hyaB mutant as for the wild type. However, no statistically significant effects of Ni supplementation on growth could be observed (doubling times varied between 7.5 and 11.5 h in the absence of H2). Perhaps the variability is due to variations in the nickel content of the (purchased) medium.
Hydrogenase activities of the wild-type H. hepaticus strain also increased significantly with increasing concentrations of supplemented nickel in the medium (Table 1). Activity decreased (for both hydrogenase and urease) at more than 10 and 8 μM nickel, respectively, perhaps due to toxicity at the higher nickel concentrations. Similarly, growth was inhibited at nickel concentrations of greater than 10 μM NiCl2 supplementation. For subsequent growth experiments (see below) and to study the hydrogenase-specific effects, subsequent experiments (such as amino acid transport) always included 10 μM nickel in the medium within the closed gas atmosphere. The hyaB strain completely lacked hydrogenase activity even in the presence of externally added nickel (see footnote to Table 1), as expected due to the disruption of the (hyaB) structural gene.
Uptake of 14 C-labeled amino acids by the wild type and the hyaB mutant. The amino acid uptake abilities after 2 and 5 h by the wild-type cells in MHB with 2% fetal bovine serum and in an H2-containing atmosphere were 7- and 6.5-fold greater, respectively, than for the wild type in argon (all data summarized after subtracting the 0-h reading). The amino acid uptake levels at 2 and 5 h by the wild-type cells in hydrogen were 6- and 5.5-fold greater, respectively, than for the hyaB mutant strain in the presence of hydrogen (Table 2). The mutant strain exhibited the same (low) uptake ability in the presence of hydrogen or argon, as expected if the transport rate was markedly facilitated by hydrogen oxidation. The difference between the wild type and mutant or between the H2 versus argon condition (for the wild type) was less pronounced when the same experiment as the 2% serum condition was done but with 5% serum in the medium. Only an approximate twofold difference (H2 stimulation when H2 versus an argon atmosphere are compared) was observed in amino acid uptake (2-h time point) when the cells were grown in the higher serum content medium (see footnote for Table 2). The total amount of amino acid uptake by the wild type in the two (different serum concentrations) media cannot be compared based on our (radioactive label approach) results since the serum itself contains amino acids. Nevertheless, it is possible that H2 use may be most beneficial when the cells exhaust other available substrates.
Maximum uptake rates and kinetics of amino acid transport. The whole-cell amino acid uptake kinetic parameters were compared between the parent strain and the hyaB mutant with both strains; this was determined for both strains incubated with 10% H2 provided during the assay. In short-term assays of amino acid uptake with a series of substrate concentrations (including limiting substrate concentrations), the double-reciprocal linear transformation of the data yielded a line equation in the form of y = bx + a, where a, the inverse of the Vmax, was determined to be 0.279 for the parent strain and 0.605 for the mutant. This difference was statistically significant (P < 0.05) since the linear transformation plot was based on four replicate samples at each of nine different substrate concentrations for each strain. Therefore, the Vmax for the wild type for amino acid transport is 2.2-fold more than for the mutant strain. However, the half-saturation affinities calculated from the slope parameter (variable b in the equation above) were similar for both strains. From this result it appears that H2 augments the transport rate, but it likely does not cause synthesis of new amino acid transport components with a higher affinity for the carbon- and/or nitrogen-containing substrates (i.e., amino acids).
Growth of the wild type and hyaB mutant in the presence or absence of hydrogen. As was described for H. pylori (21), we determined that H. hepaticus hydrogenase activity is constitutively expressed in rich medium (MHB plus serum), irrespective of whether the cells are grown with (10% partial pressure H2 added to gas phase) or without added H2 (argon substituted for H2, data not shown). For H. hepaticus, activities achieved in MHB plus serum were the same (in nmol/min/109 cells, mean ± the standard deviation for five replicates) with H2 (41 ± 4) as when cells received 10% argon instead of H2 (42 ± 5). These whole-cell activities are considerably greater than we reported previously (16); this discrepancy was investigated and found to be due to the combination of using liquid cultures and MHB medium rather than a blood-based medium used previously (16). Nevertheless, the growth rate of the wild-type H. hepaticus in the presence of hydrogen was significantly better (lower doubling time) than when argon replaced H2 as 10% of the atmosphere (Fig. 2). This difference in growth was observed in three separate experiments, with each time point assayed three times for each experiment. For all three separate experiments, the maximum growth rate achieved in H2 for the wild type was approximately two times (doubling time in hours of 4.2 ± 0.7 for H2 and 8.4 ± 1.4 for argon) that of the same culture condition but lacking H2 (argon replaced hydrogen); this growth rate was statistically significantly different (P < 0.05) among the two values (argon compared to H2) for the wild type and also between the wild type in H2 compared to either of the two atmospheric conditions applied to the hyaB strain. Not surprisingly, growth of the mutant strain was not affected by H2 and was lower than for the parent strain (Fig. 2). As was observed for the amino acid uptake studies, this growth stimulation affect by H2 was less pronounced when 5% serum was used (data not shown). Nevertheless, it was observed that the serum level (2% versus 5%) used did not significantly affect the whole-cell hydrogenase activities achieved (data not shown). The maximum doubling times for the hyaB strain with or without H2 (Fig. 2) was about 12 ± 3 h; this was not statistically a significantly different result than the result for the parent strain without hydrogen.
Qualitative analysis of H. hepaticus in feces, liver, and cecum. H. hepaticus was recovered from liver and cecum homogenates (at 21 weeks after inoculation), as well as from fecal suspensions (at 20 weeks) from each mouse inoculated with the wild-type or hyaB mutant strain. Growth observed on BA plus three antibiotics was confirmed to be that of H. hepaticus by microscopy, catalase test, and PCR analysis (data not shown). Tissue homogenates and fecal suspensions from control (uninoculated) mice did not show growth of H. hepaticus on any of the plates.
Quantitative analysis of liver and cecum. The colonization efficiency of the wild-type and the hyaB mutant strain was approximated by using real-time PCR. A standard curve was used for estimating the femtograms of DNA, which was later converted to genome copy numbers (data not shown). The results of real-time analysis showed comparable genome copy numbers in the liver and cecum from the mice inoculated with either strain at 21 weeks. The average genome copy (from 10 mice for each strain) among all samples (each determination from 1.25 mg of tissue) ranged from about 1 x 103 to 2 x 103 (liver) or from about 1 x 106 up to 2.5 x 106 (cecum). However, even with many replicate samples (two from each mouse for each tissue source for a total of 20 for each tissue) the results were not statistically different among the (two) test strains. These results indicated that expression of hydrogenase may not be important for H. hepaticus colonization of the mouse liver or the cecum. No H. hepaticus DNA was detected in tissues of control mice.
Histopathology results. At necropsy, gross examination of liver, intestines and other visceral organs revealed no significant lesions in any of the mice tested. Fibrosis, lymphoid nodules, or tumors were absent in the liver of all of the inoculated mice. Biliary epithelial and oval cell changes were absent in most mice (10 animals for each bacterial strain), but one mouse inoculated with the wild type showed mild biliary hyperplasia and oval cell changes. Mild portal infiltration consisting of lymphocytes and plasma cells in some of the portal region or areas within the hepatic lobules was observed in six wild-type-inoculated mice, and moderate infiltration in almost all portal areas was observed in one (wild-type-inoculated) mouse. Mild portal infiltration in a few of the portal areas was observed in seven hyaB-inoculated mice but also in three of the six uninoculated animals. The latter (uninoculated controls) had no evidence of degeneration, inflammation, or necrosis, or neoplasm, and all were negative for H. hepaticus by real-time PCR analysis of liver and cecum. Portal infiltration differences among the stains was not significant.
A clear difference between the mice inoculated with the wild type and the hyaB mutant was seen in lobular inflammation characteristics. Most (8 of 10) of the wild-type-inoculated mice showed mild or moderate lobular inflammation (lymphoplasmacytic hepatitis) with hepatocytic coagulative necrosis, and this was statistically highly significantly different (at P < 0.01 [Wilcoxon rank analysis]) than for the hydrogenase mutant, for which no mice exhibited those characteristic hepatic lesions; none of the 10 mice inoculated with the hyaB mutant showed any evidence of lobular inflammation or necrosis. Representative photomicrographs of liver sections are shown in Fig. 3. There was no histological evidence of lesions in the cecum or the proximal colon for any test group (data not shown), and control mice had no histological evidence of lesions in the liver (Fig. 3D), cecum, or the proximal colon.
DISCUSSION
The vast majority of Helicobacter research involves the gastric pathogen H. pylori. However, many intriguing Helicobacter species exist in animals (see references 5 and 25), and physiological and virulence studies of H. hepaticus have become increasingly important due to the association of this enterohepatic Helicobacter species with chronic hepatitis in mice and due to the increased incidences of hepatocellular carcinoma in these infected animals (31, 26). H. hepaticus thus serves as an excellent model for studying liver carcinogenesis, and such studies might be relevant to human liver adenocarcinoma development (23).
H. hepaticus hydrogenase is a hydrogen uptake hydrogenase that can oxidize molecular hydrogen to yield protons and electrons which are key players in energy conserving processes of all cells. In order to investigate the role of hydrogenase in the colonization and virulence of H. hepaticus, a mutant strain was generated in the gene encoding the large subunit of hydrogenase (hyaB). The wild-type strain possessed hydrogenase activity, and the activity increased when the cells were grown in the presence of supplemented nickel. A battery of nickel transport and metal accessory proteins facilitate nickel insertion into the NiFeS center (10, 12). Similarly, urease is a nickel-containing enzyme, requiring Ni-dependent maturation machinery. Nickel supplementation was shown to facilitate urease activity in H. pylori (4, 28), but hydrogenase enzyme activity did not increase significantly in response to nickel supplementation of the medium (22). Nickel significantly stimulated the H. hepaticus activity for both key Ni enzymes (urease and hydrogenase), and this stimulation effect is greater than is observed for H. pylori. This difference in the two Helicobacter spp. could be related to the histidine rich nickel-binding protein present in H. pylori (Hpn and Hpn-like protein), which would be expected to serve as storage reservoirs for nickel; both of these histidine-rich proteins are absent in H. hepaticus. Therefore, H. hepaticus may depend greatly on exogenous nickel. As expected due to the absence of the hydrogenase structural gene, nickel supplementation could not restore the hydrogenase activity to the hyaB mutant, and yet nickel supplementation stimulated urease activity in the mutant as in the parent strain.
In vitro studies showed that liquid growth in the presence of hydrogen (doubling time) of the wild-type strain was better than that for the hyaB mutant; this was most likely due to the ability of the wild type to use hydrogen as an additional energy substrate for growth. It is noteworthy that in the initial isolation of H. hepaticus from liver, a microaerophilic atmosphere containing hydrogen was used (6, 7). Growth stimulation due to supplying H2 to diverse H2 oxidizing bacteria has been observed previously (17, 29) but has not been reported for pathogenic bacteria. For H. hepaticus, hydrogenase activity was constitutive (not dependent on the inclusion of H2 during growth), like that described for H. pylori previously (21). However, in contrast to H. pylori, H. hepaticus activity did not increase when cells were grown with H2.
A clue to the role of H2 oxidation was provided by amino acid uptake assays. Most H2-oxidizing bacteria are chemoautotrophs, using the energy from H2 to fuel the energy input needed for carbon dioxide fixation (29). The CO2 fixation enzymes are highly conserved. However, no such CO2-fixing enzymes exist in the Helicobacter spp., based on the annotated genome sequences of (two) H. pylori or (one) H. hepaticus strains. Therefore, it is expected the energy from H2 may be used to assimilate non-CO2 sources of carbon. Wild-type cells had much increased ability for uptake of 14C-labeled amino acids in the presence than in the absence of hydrogen or when compared to the hyaB mutant cells either in the presence or absence of hydrogen. This is likely due to the ability of wild-type cells to glean energy from hydrogen, and this energy form is coupled to amino acid transport. There is precedent for hydrogen oxidation providing energy for bacterial carbon transport (17). From the H. hepaticus sequence it is clear that this bacterium, like H. pylori, has abundant amino acid transport systems (27). H2 stimulation of both growth and amino acid uptake was most evident when 2% serum was used rather than 5%; this could indicate that hydrogenase may be most important to the bacterium as an alternative energy source when nutrient conditions are diminished.
Kinetic assays to determine half-saturation affinity and Vmax are normally applied to substrate use by enzymes, but such determinations have been useful to correlate whole-cell substrate use characteristics of H. pylori to in vivo virulence characteristics (21). Also, with regard to H2 oxidation-mediated affects, it was shown that H2 use by Azotobacter vinelandii increased the bacteriums Vmax for mannose uptake without altering the bacterium's affinity for the substrate (17). The Vmax value expressed by the H. hepaticus wild type (in H2) for amino acid transport was 2.2-fold greater than for the mutant strain. This would be consistent with H2 providing an energy source (perhaps ATP) to augment uptake rates. Still, the half-saturation affinities for amino acid uptake were similar for the parent and the mutant. From this latter result it appears that, although H2 stimulates transport, it likely neither causes synthesis of new amino acid transport components with a higher affinity for the carbon substrates nor preferentially enables higher affinity uptake systems to be used.
It may appear from our real-time analysis to estimate genome number (and thereby comparative colonization) by the wild type and hyaB mutant that there is no growth advantage due to H2 use within the liver or cecum. This is surprising in light of the other results and suggests that hydrogen may be only one among many growth substrates utilized by H. hepaticus within the tissues. If the tissues (especially liver) are well supplied with carbon and energy sources, this would be consistent with the idea proposed above that H2 use may be most important under low-nutrient conditions. Nevertheless, a clear difference was seen in the photomicrographs of the liver of mice inoculated with the hyaB mutant versus the wild type; these results showed complete absence of lobular inflammation by the hyaB mutant (and similar to uninoculated control mice results), whereas the livers from 8 of 10 mice inoculated with the wild type showed moderate inflammation with necrosis. Although hydrogenase may not be directly responsible for the pathogenesis seen in the liver tissue of mice inoculated with the wild type, the highly diffusible high-energy reductant is present in liver tissue of live mice (16), and energy released by hydrogen oxidation may be used to augment synthesis of virulence-related proteins or enzymes responsible for causing inflammation and necrosis. Possibly, high-efficiency solute uptake systems (such as for amino acids) that are supported by H2 oxidation may be coupled to virulence.
The animal colonization attributes of many Helicobacter spp. (including by H. hepaticus) have been well documented and reviewed (5, 25). However, this is the first evidence of the involvement of molecular hydrogen use via hydrogenase in the development of characteristics related to bacterial hepatitis, namely, inflammation and necrosis of the liver tissue. Therefore, the present study adds H. hepaticus to the small list of bacteria (i.e., H. pylori and S. enterica serovar Typhimurium) that use hydrogen to augment their virulence in animals. Further long-term studies are warranted to address the mechanisms by which H. hepaticus hydrogenase is involved in the pathogenesis of hepatitis and perhaps in carcinoma development. Finally, it is noteworthy that insertion cassette mutagenesis was successful for H. hepaticus specific gene disruption, and it is hoped this procedure can be used by many researchers for studying roles of other genes in this bacterium.
ACKNOWLEDGMENTS
This study was supported by NIH NCI grant RO3-CA103095.
We thank Sue Maier (University of Georgia) for assistance with the 14C experiments and A. Olczak for advice on protocols.
REFERENCES
1. Avenaud, P., B. Le. Bail, K. Mayo, A. Marais, R. Fawaz, P. Bioulac-Sage, and F. Megraud. 2003. Natural history of Helicobacter hepaticus infection in conventional A/J mice, with special reference to liver involvement. Infect. Immun. 71:3667-3673.
2. Battles, J. K., J. C. Williamson, K. M. Pike, P. L. Gorelick, J. M. Ward, and M. A. Gonda. 1995. Diagnostic assay for Helicobacter hepaticus based on nucleotide sequence of its 16S rRNA gene. J. Clin. Microbiol. 33:1344-1347.
3. Benoit, S., N. Mehta, G. Wang, M. Gatlin, and R. J. Maier. 2004. Requirement of hydD, hydE, hypC, and hypE genes for hydrogenase activity in Helicobacter pylori. Microb. Pathog. 36:153-157.
4. Contreras, M., J-M Thiberge, M-A Mandrand-Bertholet, and A. Labigne. 2003. Characterization of and the roles of NikR, a nickel-responsive pleiotropic autoregulator of Helicobacter pylori. Mol. Microbiol. 49:947-963.
5. Ferrero, R. L., and J. G. Fox. 2001. In vivo modeling of Helicobacter-associated gastrointestinal diseases, p. 565-582. In H. L. Mobley, G. L. Mendz, and S. L. Hazell (ed.), Helicobacter pylori: physiology and genetics. ASM Press, Washington, D.C.
6. Fox, J. G., F. E. Dewhirst, J. G. Tully, B. J. Paster, L. Yan, N. S. Taylor, M. J. Collins, Jr., P. L. Gorelick, and J. M. Ward. 1994. Helicobacter hepaticus sp. nov., a microaerophilic bacterium isolated from livers and intestinal mucosal scrapings from mice. J. Clin. Microbiol. 32:1238-1245.
7. Fox, J. G., X. Li, L. Yan, R. J. Cahill, R. Hurley, R. Lewis, and J. C. Murphy. 1996. Chronic proliferative hepatitis in A/J Cr mice associated with persistent Helicobacter hepaticus infection: a model of Helicobacter-induced carcinogenesis. Infect. Immun. 64:1548-1558.
8. Ge, Z., D. A. White, M. T. Whary, and J. G. Fox. 2001. Fluorogenic PCR-based quantitative detection of a murine pathogen, Helicobacter hepaticus. J. Clin. Microbiol. 39:2598-2602.
9. Huang, Y., X. G. Fan, Z. M. Wang, J. H. Zhou, X. F. Tian, and N. Li. 2004. Identification of Helicobacter species in human liver samples from patients with primary hepatocellular carcinoma. J. Clin. Pathol. 57:1273-1277.
10. Jacobi, A., R. Rossmann, and A. Bock. 1992. The hyp operon gene products are required for the maturation of catalytically active hydrogenase isoenzymes in Escherichia coli. Arch. Microbiol. 158:444-451.
11. Kornilov'ka, I., I. Nilsson, M. Utt, A. Ljungh, and T. Wadstrom. Immunogenic proteins of Helicobacer pullorum, Helicobacter bilis, and Helicobacter hepaticus identified by two-dimensional gel electrophoresis and immunoblotting. Proteomics 2:775-778.
12. Magalon, A., and A. Bock. 2000. Dissection of the maturation reactions of the [NiFe] hydrogenase 3 from Escherichia coli taking place after nickel incorporation. FEBS Lett. 473:254-258.
13. Maier, R. J. 2003. Availability and use of molecular hydrogen as an energy substrate for Helicobacter species. Microbes Infect. 5:1159-1163.
14. Maier, R. J., A. Olczak, S. Maier, S. Soni, and J. Gunn. 2004. Respiratory hydrogen use by Salmonella enterica serovar Typhimurium is essential for virulence. Infect. Immun. 72:6294-6299.
15. Maier, R. J., C. Fu, J. Gilbert, F. Moshiri, J. Olson, and A. G. Plaut. 1996. Hydrogen uptake hydrogenase in Helicobacter pylori. FEMS. Microbiol. Lett. 141:71-76.
16. Maier, R. J., J. Olson, and A. Olczak. 2003. Hydrogen-oxidizing capabilities of Helicobacter hepaticus and in vivo availability of substrate. J. Bacteriol. 185:2680-2682.
17. Maier, R. J., and J. Prosser. 1988. Hydrogen-mediated mannose uptake in Azotobacter vinelandii. J. Bacteriol. 170:1986-1989.
18. Myles, M. H., R. S. Livingston, B. A. Livingston, J. M. Criley, and C. L. Franklin. 2003. Analysis of gene expression in ceca of Helicobacter hepaticus-infected A/JCr mice before and after development of typhlitis. Infect. Immun. 71:3885-3893.
19. Noether, G. 1971. Introduction to statistics: a fresh approach. Houghton-Mifflin, Boston, Mass.
20. Olson, J. W., C. Fu, and R. J. Maier. 1997. The HypB protein from Bradyrhizobium japonicum can store nickel and is required for the nickel-dependent transcriptional regulation of hydrogenase. Mol. Microbiol. 24:119-128.
21. Olson, J. W., and R. J. Maier. 2002. Molecular hydrogen as an energy source for Helicobacter pylori. Science 298:1788-1790.
22. Olson, J. W., N. S. Mehta, and R. J. Maier. 2001. Requirement of nickel metabolism proteins HypA and HypB for full activity of both hydrogenase and urease in Helicobacter pylori. Mol. Microbiol. 39:176-182.
23. Pellicano, R., V. Mazzaferro, W. F. Grigioni, M. A. Cutufia, S. Fagoonee, L. Silengo, and R. M. Ponzetta. 2004. Helicobacter species sequences in liver samples from patients with and without hepatocellular carcinoma. World J. Gastroenterol. 10:598-601.
24. Phillips, M. W., and A. Lee. 1983. Isolation and characterization of a spiral bacterium from the crypts of rodent gastrointestinal tracts. Appl. Environ. Microbiol. 45:675-683.
25. Solnick, J. V., and D. B. Schauer. 2001. Emergence of diverse Helicobacter species in the pathogenesis of gastric and enterohepatic disease. Clin. Microbiol. Rev. 14:59-97.
26. Schauer, D. B. 2001. Enterohepatic Helicobacter species, p. 533-548. In H. L. T. Mobley, G. L. Mendz, and S. T. Hazell (ed.), Helicobacter pylori: physiology and genetics. ASM Press, Washington, D.C.
27. Suerbaum, S., C. Josenhans, T. Sterzenbach, B. Drescher, P. Brandt, M. Bell, M. Droge, B. Fartmann, H. P. Fischer, Z. Ge, A. Horster, R. Holland, K. Klein, J. Konig, L. Macko, G. L. Mendz, G. Nyakatura, D. B. Schauer, z. Shen, J. Weber, M. Frosch, and J. G. Fox. 2003. The complete genome sequence of the carcinogenic bacterium Helicobacter hepaticus. Proc. Natl. Acad. Sci. USA 100:7901-7906.
28. van Vliet, H., E. J. Kuipers, B. Waidner, B. J. Davies, N. de Vries, C. W. Penn, C. M. Vandenbroucke-Grauls, M. Kist, S. Bereswill, and J. G. Kusters. 2001. Nickel-responsive induction of urease expression in Helicobacter pylori is mediated at the transcriptional level. Infect. Immunol. 69:4891-4897.
29. Vignais, P. M., B. Billoud, and J. Meyer. 2001. Classification and phylogeny of hydrogenases. FEMS Microbiol. Rev. 25:455-501.
30. Wang, Y., and D. E. Taylor. 1990. Chloramphenicol resistance in Campylobacter coli: nucleotide sequence, expression and cloning vector construction. Gene 94:23-28.
31. Ward, J. M., J. G. Fox, M. R. Anver, D. C. Haines, C. V. George, M. J. Collins, Jr., P. L. Gorelick, K. Nagashima, M. A. Gonda, R. V. Gilden, et al. 1994. Chronic active hepatitis and associated liver tumors in mice caused by a persistent bacterial infection with a novel Helicobacter species. J. Natl. Cancer Inst. 86:1222-1227.
32. Young, V. B., K. A. Knox, J. S. Pratt, J. S. Cortez, L. S. Mansfield, A. B. Rogers, J. G. Fox, and D. B. Schauer. 2004. In vitro and in vivo characterization of Helicobacter hepaticus cytolethal distending toxin mutants. Infect. Immunol. 72:2521-2527.(Nalini S. Mehta, Stephane)