|Home | About | Journals | Submit | Contact Us | Français|
Nineteen hyperthermophilic heterotrophs from deep-sea hydrothermal vents, plus the control organism Pyrococcus furiosus, were examined for their ability to grow and produce H2 on maltose, cellobiose, and peptides and for the presence of the genes encoding proteins that hydrolyze starch and cellulose. All of the strains grew on these disaccharides and peptides and converted maltose and peptides to H2 even when elemental sulfur was present as a terminal electron acceptor. Half of the strains had at least one gene for an extracellular starch hydrolase, but only P. furiosus had a gene for an extracellular β-1,4-endoglucanase. P. furiosus was serially adapted for growth on CF11 cellulose and H2 production, which is the first reported instance of hyperthermophilic growth on cellulose, with a doubling time of 64 min. Cell-specific H2 production rates were 29 fmol, 37 fmol, and 54 fmol of H2 produced cell−1 doubling−1 on α-1,4-linked sugars, β-1,4-linked sugars, and peptides, respectively. The highest total community H2 production rate came from growth on starch (2.6 mM H2 produced h−1). Hyperthermophilic heterotrophs may serve as an important alternate source of H2 for hydrogenotrophic microorganisms in low-H2 hydrothermal environments, and some are candidates for H2 bioenergy production in bioreactors.
Sugar catabolism and H2 production were examined previously in Pyrococcus furiosus (7, 12, 21), a hyperthermophile from a shallow marine geothermal beach. However, the ability of other Pyrococcus and closely related Thermococcus species from deep-sea hydrothermal vents to hydrolyze various sugar polymers and produce H2 is poorly understood. These data are important for modeling carbon and energy flow and biogeochemical processes in deep-sea hydrothermal vent environments and for assessing the utility of hyperthermophiles for industrial bioenergy applications.
Pyrococcus and Thermococcus are generally considered to be elemental sulfur reducers that catabolize peptides, but some species can also hydrolyze starch using amylopullulanase (4) and cyclodextrin glucanotransferase (22) (Fig. 1). P. furiosus and Pyrococcus horikoshii also have β-1,4-endoglucanase genes whose recombinant protein products can hydrolyze cellulose (Fig. 1) (1, 2), but previous attempts at growing these organisms on cellulose were unsuccessful (1, 2). Pyrococcus abyssi likewise has a β-1,4-endoglucanase gene (see Table S1 in the supplemental material), and both it and P. horikoshii were isolated from deep-sea hydrothermal vent environments. This raises the question of the likelihood of hyperthermophile growth on β-1,4 sugar polymers in these environments. Following starch and putative β-1,4 glucan hydrolysis, Pyrococcus and Thermococcus catabolize the resulting glucose to acetate and CO2 in a process that involves ferredoxin-dependent redox enzymes (Fig. 1) (28). Then, in lieu of sulfur reduction, a membrane-bound, ferredoxin-dependent hydrogenase reduces H+ to produce H2 and translocates Na+ across the cytoplasmic membrane to form a Na+ motive force that is used for ATP synthesis via oxidative phosphorylation (16, 20). However, cell-specific and community rates of H2 production by these organisms on various substrates are very limited.
The goal of this study was to characterize α-1,4- and β-1,4-linked glucan catabolism and H2 production in 6 Pyrococcus and 13 Thermococcus strains from deep-sea hydrothermal vent sites along the Juan de Fuca Ridge in the northeastern Pacific Ocean (10, 17, 26). These strains were screened for their ability to grow on maltose, cellobiose, and peptides; for their ability to produce H2; and for the presence of genes for amylopullulanase, cyclomaltodextrin glucanotransferase, cyclomaltodextrinase, and β-1,4-endoglucanase (families 5 and 12). Cultures were grown with and without sulfur to determine how well these strains grow without sulfur (they may have a sulfur requirement), whether they produce H2 in the presence of sulfur, and what percentage of the terminal-electron-accepting metabolite was H2 relative to H2S. We also adapted P. furiosus to grow on CF11 cellulose without sulfur and measured its cell-specific and total community H2 production rates on cellulose, starch, their respective dimers (cellobiose and maltose), and peptides to determine the relative amounts and rates of H2 produced on different substrates.
A total of 19 Pyrococcus and Thermococcus strains from hydrothermal vent fields on the Juan de Fuca Ridge in the northeastern Pacific Ocean (see Table S2 in the supplemental material) were used for this study, along with a control strain of P. furiosus. P. furiosus DSM3638 was purchased from the Deutsche Sammlung von Mikrooganismen und Zellkulturen (Braunschweig, Germany).
The base medium was derived from DSM medium 141 (www.dsmz.de) and was composed of the following in distilled water (per liter): 18 g NaCl, 4 g MgCl2·6H2O, 3.45 g MgSO4·7H2O, 0.335 g KCl, 0.25 g NH4Cl, 0.14 g CaCl2·2H2O, 0.14 g K2HPO4, 0.5 g yeast extract (Difco; vitamin B12 fortified), 1 g NaHCO3, 10 ml of DSM medium 141 mineral solution, 10 ml of DSM medium 141 vitamin solution, 1 ml of 0.2% (NH4)2Fe(SO4)2-0.2% (NH4)2Ni(SO4)2 solution, 0.10 ml of 100 mM Na2WO4·2H2O-100 mM Na2SeO4 solution, and 0.25 mg resazurin. The primary carbon source was 0.5% (wt vol−1) of maltose, cellobiose, starch, CF11 cellulose (Millipore), Avicel microcrystalline cellulose, SolkaFloc cellulose, or casein hydrolysate (Difco). To reduce the media, 0.025% (wt vol−1) cysteine-HCl·H2O and Na2S·9H2O each was added. The pH of the medium was adjusted to 6.80 ± 0.05 (room temperature) unless otherwise stated. After reduction and pH adjustment, 1 ml of 1 M potassium phosphate solution (pH 6.8) was added as a pH buffer. For growth on sulfur, 0.1% (wt vol−1) of USP-grade sublimed elemental sulfur (Sigma) was added to the media. The headspace was flushed with N2-CO2 (80%:20%).
Each strain was transferred three successive times on each growth medium prior to scoring for growth or experimentation. Adaptation of P. furiosus to growth on CF11 cellulose was performed using a culture adapted to growth on cellobiose and growing it on 0.5% CF11 cellulose plus 0.05% cellobiose. With each successive transfer, the amount of cellobiose was reduced 0.005% until cellobiose could be omitted from the medium. P. furiosus was grown on cellulose with and without stirring in serum bottles and with stirring in a 2-liter bioreactor to determine whether this organism was suitable for use in a cellulolytic bioreactor for H2 production. Stirred cultures were contained in 120-ml serum bottles holding 50 ml of medium with a stir bar. The medium was stirred at approximately 100 rpm in the incubator using a magnetic stir plate. Experiments were run in duplicate with and without stirring with time points every 2 h for cell counts. Growth of P. furiosus in a 2-liter bioreactor was performed by inoculating 1.6 liters of medium with a logarithmic-growth-phase culture. The medium was stirred and maintained at 95°C (± 0.1°C) and pH 7.3 (± 0.2) using temperature and pH controls and had a gas flow of Ar-CO2 (80%:20%) at 30 ml min−1.
Cell concentrations were measured using a Petroff-Hausser counting chamber and a phase-contrast light microscope. H2 concentrations were measured using a Shimadzu GC-8A gas chromatograph equipped with a thermal conductivity detector and a Supelco 60/80 Carboxen 1000 column (15 ft by 1/8 in. Stainless steel [SS]) with argon as the carrier gas. H2S concentrations were measured spectrophotometrically using the methylene blue method (6). For growth and H2 production rate measurements, 12 60-ml serum bottles containing 25 ml of medium were inoculated with a logarithmic-growth-phase culture grown at 95°C. At 0, 2, 4, 6, 8, and 10 h after inoculation, a pair of bottles was removed from the incubator and set aside for analysis. The growth rate was calculated using a best-fit curve through the logarithmic portion of the growth curve. The cell-specific H2 production rates were calculated by measuring the slope of the amount of H2 per bottle plotted against the total number of cells per bottle. The assumption was that each cell produces the same amount of H2 per cell doubling for a given growth condition. The growth and cell-specific H2 production rates were each compared using a linear regression analysis, analysis of covariance, and a Tukey test (α = 0.05) as described previously (29).
DNA was extracted from each of the 20 heterotroph strains using a genome extraction kit (Genome Wizard; Promega) and quantified using a Nanodrop ND-1000 spectrophotometer. Degenerate PCR primers specific to amylopullulanase, cyclomaltodextrin glucanotransferase, cyclomaltodextrinase, and β-1,4-endoglucanase (families 5 and 12) were designed in this study from the highly conserved regions of Thermococcales amino acid sequence alignments (see Table S3 in the supplemental material). For cyclomaltodextrin glucanotransferase and cyclodextrinase, two sets of forward and reverse primers were designed. PCR was performed with Taq DNA polymerase using the following conditions: initial denaturation at 94°C for 5 min followed by 35 cycles of denaturation at 94°C for 40 s, annealing at 50°C for 40 s, and an extension at 72°C for 1 min with additional extension at 72°C for 10 min in the final cycle. When the product signal was weak, the annealing temperature was decreased to 45°C. PCR products whose sizes matched the predicted size of the product were sequenced to verify that the correct gene fragment had been amplified (see Tables S4 to S6 in the supplemental material).
All 19 deep-sea strains plus P. furiosus grew on maltose, cellobiose, and peptides when sulfur was present and most to abundances exceeding 108 cells ml−1 (see Table S2). However, only 9 of the 20 strains grew without sulfur, and P. furiosus and Thermococcus strains ES1 and Du06 were the only ones that grew to more than 3 × 107 cells ml−1. When grown on maltose plus sulfur, 10 of the 20 strains produced H2 to as much as 51% of the total respiratory metabolite (H2 plus H2S). The remaining strains produced trace amounts of H2 (<1%). All but Thermococcus strains SM06 and 9N3 produced H2 when grown on peptides plus sulfur.
Using PCR screening, 8 deep-sea strains had amylopullulanase gene homologs, 5 had cyclomaltodextrin glucanotransferase gene homologs, and 9 had cyclomaltodextrinase gene homologs (see Table S2), suggesting at least some capacity for growth on α-1,4-linked polymers. Only P. furiosus had a family 12 β-1,4-endoglucanase gene homolog, and none of the strains had a family 5 β-1,4-endoglucanase gene homolog. These findings generally match the pattern found in the three Pyrococcus and five Thermococcus genomes that have been sequenced (see Table S1). Each genome contains the genes for the membrane-bound hydrogenase and at least one α-amylase, α-glucosidase, and β-glucosidase. Six of the genomes encode amylopullulanase, four encode cyclomaltodextrin glucanotransferase, and three encode one of two β-1,4-endoglucanases.
P. furiosus was the only organism examined that had a gene for extracellular β-1,4-endoglucanase and grew to >108 cells ml−1 on cellobiose in the absence of sulfur. It was selected for cellulose growth studies and serially adapted through stepwise cellobiose weaning in successive transfers to grow on 0.5% (wt vol−1) CF11 cellulose plus 0.01% yeast extract without cellobiose or sulfur. At the outset of our experiments, P. furiosus was incapable of growth on 0.05% (wt vol−1) yeast extract alone, but in the process of adapting it to growth on CF11 cellulose we also adapted it to grow on low yeast extract (down to 0.01%). However, the growth rate and maximum cell concentration for growth on 0.5% CF11 cellulose plus 0.01% yeast extract were significantly higher than those for growth on 0.01% yeast extract alone (Fig. 2A). There was no growth on either 0.5% Avicel microcrystalline cellulose or 0.5% SolkaFloc cellulose, which are highly recalcitrant forms of commercial cellulose. Stirring the cultures during growth on CF11 cellulose had no effect on growth rate relative to static cultures, but it extended the length of logarithmic growth, doubled the maximum cell concentration, and tripled the maximum H2 produced. Growth of P. furiosus on CF11 cellulose was scaled up to a 2-liter bioreactor, with a doubling time of 84 min and H2 detected in the exhaust gas.
Cell-specific H2 production rates for P. furiosus grown on various substrates varied significantly with substrate (P < 0.05), with starch and maltose yielding the lowest production rates, CF11 cellulose and cellobiose yielding intermediate rates, and casein hydrolysate yielding the highest rate (Fig. 2B). The doubling time on CF11 cellulose was significantly shorter (P < 0.05) than the doubling time on all the other substrates except starch, but maximum cell concentrations were highest on α-1,4-linked sugars, being 10-fold higher than those on casein hydrolysate (Table 1). Therefore, while the cell-specific H2 production rates are highest on peptides, the total amount of H2 produced by a community of P. furiosus is three to five times higher when a population is growing on α-1,4-linked polymers than on peptides.
Deep-sea Pyrococcus and Thermococcus species from hydrothermal vents grow on α-1,4- and β-1,4-linked glucose dimers, and many have the genetic capacity to hydrolyze α-1,4 polymers. However, they generally lack the ability to hydrolyze β-1,4 polymers. Strains lacking extracellular α-1,4- and β-1,4-linked polysaccharide hydrolases are apparently dependent upon allochthonous extracellular hydrolases in order to grow on maltose and cellobiose. All of the strains in this study produced H2, but most still have some requirement for sulfur, and only half produced more than trace amounts of H2 when sulfur was present.
Bacterial biofilm and the mucus produced by alvinellid polychaete worms are potential sources of polysaccharides and sulfur in hydrothermal vent environments that could support hyperthermophilic heterotrophs. Hyperthermophiles, mesophilic bacteria, and alvinellid polychaetes often live within centimeters of each other in vents of the northeast Pacific Ocean atop and within porous hydrothermal mineral deposits where seawater mixes with superheated hydrothermal fluid and forms steep thermal gradients. Paralvinella sulfincola polychaete worms from the Juan de Fuca Ridge live in dense assemblages on hot surfaces of active hydrothermal sulfide deposits at temperatures up to 55°C (8, 25). They secrete copious amounts of mucus that is continuously sloughed off their epidermal surface, which may protect them from toxic metals and sulfide (11). This mucus contains a high proportion of neutral sugar carbohydrates, especially glucose, and contains up to 60% elemental sulfur on a dry weight basis (11, 23). A 16S rRNA gene survey of P. sulfincola tubes and adjacent substratum showed the presence of Thermococcaceae (15), and five of the Thermococcus strains used in this study (CX4, CL1, CL2, ES1, and FTW06) were isolated from P. sulfincola (10, 17, 26). Interestingly, these strains produced the highest proportion of H2 when grown with maltose and sulfur. Furthermore, some mesophilic hydrothermal vent bacteria, such as Alteromonas macleodii and Pseudoalteromonas species, also secreted exopolysaccharides composed mainly of glucose, galactose, and mannose with minor amounts of β-glycosidic linkages (19). Therefore, glucose-based polysaccharides, most likely with primarily α-1,4 linkages, appear to be readily available for hyperthermophilic heterotrophs in deep-sea vent environments.
P. furiosus grew better than any of the deep-sea strains without sulfur and was the only organism that possessed extracellular enzymes for degrading both α-1,4 and β-1,4 glucans. It was serially adapted to grow on cellulose, which is the first reported observation of hyperthermophilic cellulose-to-H2 conversion. Its doubling time was 64 min, which exceeds any bacterial rate reported previously for growth on cellulose and supports the idea that growth rates on cellulose increase with temperature (13). This may be due in part to increased heat loosening and exposing individual cellulose polymers within its crystalline array for easier enzymatic degradation. The higher rate of cell-specific H2 production on peptides may be due to a higher demand for H2 production to meet ATP demand via oxidative phosphorylation, since ATP yield per peptide is lower than ATP yield per glucose molecule. Nevertheless, total H2 production rates were much higher when P. furiosus was grown on sugars, due to higher cell concentrations and generally higher growth rates.
Sugar-to-H2 conversion by hyperthermophilic heterotrophs may have ecological significance, since the organisms could provide an alternate source of H2 for hydrogenotrophic microorganisms in hydrothermal systems. Low H2 concentrations in basalt-hosted hydrothermal vents are predicted to limit the growth of methanogens (9), and previous studies showed that hyperthermophilic methanogens can be supported by the H2 produced by hyperthermophilic heterotrophs (3, 5, 14). The hyperthermophilic methanogen Methanocaldococcus strain JH146 produced CH4 at a cell-specific rate of 0.17 pmol cell−1 doubling−1 over a wide range of temperatures, pH, and salt concentrations (27). Assuming that 4 mol of H2 are consumed for each mole of CH4 produced and that the H2 production rate by deep-sea hyperthermophilic heterotrophs is half that of P. furiosus due to a sulfur requirement, we estimate that the H2 produced by the doubling of 26 to 47 hyperthermophilic heterotroph cells would support the doubling of a single Methanocaldococcus cell, depending on the substrate used by the heterotroph. This is in agreement with a previous study where hyperthermophilic methanogens comprised 3% of a coculture population with hyperthermophilic heterotrophs (3). If methanogens are dependent upon interspecies H2 transfer in some hydrothermal systems, then their growth is likely to be spatially constrained by the availability of allochthonous organic substrates and the presence of hyperthermophilic heterotrophs.
It was suggested that cellular maintenance energies double with every 9°C increase in growth temperature independent of species type (24). If this is so, a hyperthermophile growing at 95°C would need to produce 250 times more energy for growth than an organism growing at 23°C. This would require an increase in metabolic activity, and thus higher cell-specific rates of metabolite production, to meet this energy demand. The combination of high maintenance energy costs, high growth rates, and the use of a thermodynamically favorable electron donor (ferredoxin instead of NADH) for H2 production leads to high H2 production rates by Pyrococcus and Thermococcus. These points are favorable from an industrial standpoint. Furthermore, H2 is a low-solubility gas that can be combusted to gain about the same amount of energy as burning CH4 (18). It can also be used as a low-pollution fuel for conventional fuel cells and combustion-free electricity generation with an efficiency that is 50% higher than that of combustion driven approaches (18). In conclusion, the conversion of α-1,4- and β-1,4-linked sugars and peptides to H2 by hyperthermophilic heterotrophs may be a valuable energy source for H2-limited autotrophs in deep-sea hydrothermal environments and could be exploited to convert organic feedstock into high levels of H2 for industrial bioenergy applications.
Special thanks go to Anna Eboch for her DNA extractions and Valeriy Znakharchuk for his culture screening efforts. We thank Craig Moyer (Western Washington University) for kindly providing us with seven of the Thermococcus strains used for this study.
This work was supported by grants to J.F.H. from the NSF Division of Ocean Sciences (OCE-0732611), the Northeast Sun Grant Institute of Excellence (NE07-030), and the USDA CSREES (MAS00945).
†Supplemental material for this article may be found at http://aem.asm.org/.
Published ahead of print on 18 March 2011.