A culture apparatus for maintaining H at sub-nanomolar 2 concentrations

We devised a microbial culture apparatus capable of maintaining sub-nanomolar H concentrations. This apparatus 2 provides a method for study of interspecies hydrogen transfer by externally fulﬁlling the thermodynamic requirement for low H concentrations, thereby obviating the need for use of cocultures to study some forms of metabolism. The culture vessel is 2 constructed of glass and operates by sparging a liquid culture with puriﬁed gases, thereby removing H as it is produced. We 2 used the culture apparatus to decouple a syntrophic association in an ethanol-consuming, methanogenic enrichment culture, allowing ethanol oxidation to dominate methane production. We also used the culture apparatus to grow pure cultures of the ethanol-oxidizing, proton-reducing Pelobacter acetylenicus (WoAcy 1), and to study the bioenergetics of growth. (cid:211) Science B.V. All reserved. grow 2 3 (10 ) was again serially diluted and allowed to grow. This procedure was repeated three additional times until consistent growth was achieved.


Introduction
degrade the organic substrates to respiring organisms. Due to thermodynamic constraints, the organic The syntrophic degradation of organic material is substrates can only be consumed in this fashion an environmentally and economically important pro-when the concentration of H is low. Respiring 2 cess which occurs during anaerobic digestion organisms utilize the H , and maintain low con-2 (Schink, 1997). Syntrophic degradation, also called centrations so that the syntrophic oxidation of the secondary fermentation, involves the cooperation of organic substrates is sufficiently exergonic. two or more organisms to consume a single sub-Many anaerobic microorganisms are capable of strate; the substrate (organic acids, alcohols, amino acting as syntrophic partners during the degradation acids, and aromatics) is generally a product of of organic material. Many of these organisms are primary fermentation. Hydrogen (H ) is thought to found within the Genus Syntrophomonas, though 2 be the key intermediary in this process, transferring others include sulfate reducers, species of Pelobacreducing equivalents from the organisms which ter, benzoate degraders, and others. Many such organisms are available in pure culture as they are also capable of growth on substrates which do not require syntrophic coupling. Previous attempts to grow these organisms on 'syntrophic' substrates in  the absence of partner organisms have met with only limited success (Mountfort and Kaspar, 1986; Stams subsequently flow through a purifier to remove H , 2 et al., 1993;Schink, 1997). CO, and O . The gas then flows through a stirred 2 This study describes the design, construction, and glass culture vessel, where biologically produced H 2 use of a flow-through culture apparatus capable of is rapidly transferred from the liquid to the gas growing monocultures of 'syntrophs' by externally phase. The resultant H -containing gas flows through 2 maintaining the thermodynamic requirement for low a series of traps to remove water and hydrogen H concentration. We further describe the growth of sulfide. The analytical portion of the apparatus is 2 both pure and enrichment cultures of H -producing located downstream and is used to measure the 2 ethanol oxidizers. concentrations of gases entering and exiting the culture vessel. The location of the analytical portion of the apparatus allows for passive sampling of gas 2. Materials and methods metabolism from the culture. A schematic diagram of the entire apparatus is shown in Fig. 1 H contamination of the vessel must be minimized, the vessel is purged for at least 24 h prior to addition 2 (3) strict anaerobic conditions must be maintained, of medium, (2) the vessel is kept under positive and (4) all experiments must be performed aseptical-pressure to prevent air contamination, (3) a reducing ly. agent, generally cysteine, is added in minor quan-To ensure that gas exchange between phases tities to maintain reduced conditions, (4) minor occurs rapidly, the bottom of the culture vessel amounts of resazurin are added to the medium as a contains a glass frit which produces fine bubbles visible redox indicator, (5) all gases are of the (estimated size 10-100 mm) which give the solution highest purity available, and (6) gases flow through a a milky white appearance. The glass frit allows for heated column which removes traces of O in 2 an even distribution of bubbles, though when the addition to H and CO (see Section 2.3.). In addition 2 impeller is not used organisms may accumulate near to cysteine, sufide can also be used as a reducing the surface of the frit. The vessel is also equipped agent, though sulfide is lost as gaseous hydrogen with a glass stirrer to mix the liquid medium and sulfide at a rate which is pH dependent. Other maintain uniform conditions. The screw-shaped stir-reducing agents, including thiosulfate, are also comrer was fashioned from a piece of pyrex plate (60 3 patible with the culture apparatus. 1 1/20 3 1/80); the top and bottom halves are To ensure that no contamination is introduced to threaded in opposite senses, minimizing vortex for-the culture vessel the entire vessel is cleaned and mation, shearing, and disruption of cells. The stirrer autoclaved prior to use (30 min, 1218C), sterile plugs is driven by a variable speed power head (model consisting of glass wool are located directly up-RZR-1, Caframo Ltd., Wiarton, Ontario, Canada), stream and downstream of the vessel, sterile techand is generally operated between 200 and 600 rpm.
nique is used in handling any components of the The rod of the stirrer is fitted to a bore in the vessel, and sampling ports located on top of the TeflonE plug, and is lubricated with a small amount vessel are sterilized before each use. of grease (KrytoxE, Dupont, Deepwater, NJ). The snug fit of the glass rod through the hole in the 2.3. Gases TeflonE plug, coupled with the use of grease, is sufficient to create a seal under slight positive Mass flow controllers (model 8100, Unit Instrupressures. We have not observed biofilm formation ments, Yorba Linda, CA) are employed to precisely during experiments.
control the flow-rate and mixing ratios of gases. The Because metal surfaces are known to produce H MFCs are controlled by a digital power supply 2 in the presence of water, metal has been eliminated (model DX-5, Unit Instruments, Yorba Linda, CA) completely from portions of the vessel which contact which is capable of simultaneously controlling severwater. Though the vessel consists primarily of glass, al channels. Each tank of gas is connected to an minor amounts of TeflonE, PFA (perfluoroalkoxy), individual MFC, and flow-rates are confirmed by use and TeflonE-coated rubber are also present. Portions of a bubble flow meter (The Gilibrator, Gilian of the vessel constructed using PFA are the Instruments Corp., W. Caldwell, NJ). The following SwagelokE fittings and the tubing leading from the gases have been used with the culture apparatus: (1) vessel, while the plug located on the top of the vessel UHP N (So-Cal Airgas, Lakewood, CA), (2)  use of a gas chromatograph equipped with a flame Discrete liquid samples were taken during growth ionization detector. The sampling port located at the for analysis of acetate, pH, and growth yield. Growth top of the vessel allows for removal of discrete liquid yields were determined in duplicate by harvesting samples for other analyses.
cells at the end of the experiment, centrifuging 35 ml of the culture (4000 3 g for 1 h), desiccating the 2.5. Operation pellet and measuring the resulting mass. Acetate was measured with an HPLC using an organic acids The culture vessel is sterilized, assembled, and column (Alltech, IOA-1000) and a UV/ VIS detector purged with H -free gas beginning more than 1 day set at 210 nm (0.5 mM H SO mobile phase set at 2 2 4 21 before inoculation. Upstream and downstream H 0.6 ml min ).

Calculation of DG9
Free energy yields (DG9) were calculated using standard thermodynamic equations. Values for 2 CH COO and pH were interpolated from measured 3 concentrations, while H and temperature were mea-2 sured for each calculation. Values for ethanol were calculated from initial conditions by subtracting 2 CH COO production; assimilation of ethanol de-3 rived carbon into cell mass was not considered, and is not likely to be significant for thermodynamic calculations. Several factors are involved in calculating DG9. Temperature is important through the effect of entropy on DG89 ( 2 TDS) as well as its effect on the deviation from equilibrium (RT 3 lnhQj). The pH is important through its effect on DG89 as well as 2 through its effect on the speciation of CH COO / 3 CH COOH. We assumed that all CH COOH was in The H -stripping culture system has been used to mol , and DG89-H 525.69 kJ per pH unit (0 at 2 1 analyze H production from several different cultures pH 0). The measured concentrations of H , H , and 2 2 2 including pure cultures of Methanobacterium strain CH COO were assumed to be equal to the con-3 Marburg, Methanosaeta thermophila strain CALS-1, centrations apparent to the organism, and all ac-P. acetylenicus strain WoAcy 1, and ethanol-oxitivities were assumed to equal 1.
dizing methanogenic enrichment cultures similar to the classical 'Methanobacillus omelianskii' (Bryant et al., 1967). Fig. 3 demonstrates the net production 3. Results of H , CH and acetate in an ethanol utilizing 2 4 methanogenic enrichment culture grown in the cul-The culture vessel is capable of achieving gas ture vessel with a defined mineral salts medium phase H levels below our analytical detection limit containing 20 mM ethanol. The metabolic activity of 2 23 (10 Pa), corresponding to an equilibrium con-the H -producing organisms far exceeded the 2 centration below 10 picomolar in the liquid phase. methanogenic activity when H was held low, there-2 Fig. 2 demonstrates the flushing of H from an by uncoupling the 'syntrophic' association with the 2 empty culture vessel which is given an initial pulse methane producers (Fig. 3). of H . The residence time calculated from Fig. 2 Pure cultures of P. acetylenicus were grown in a 2 (15.6 min) closely matches the expected residence mineral salts medium containing 20 mM ethanol. time (16 min) based on calculations using flow-rate The evolution of H was monitored as a function of 2 and total volume. When the vessel contains liquid, time in the exhaust gas of the culture apparatus the residence time of H is about half as long during several experiments. Fig. 4 demonstrates a 2 because the total volume of gas in the system is typical H production profile, while Fig. 5 dem-2 about half as large. Because H is relatively insolu-onstrates the net production of H (calculated from a 2 2 ble and the culture is constantly being sparged, the production profile) and acetate (which builds up in the culture apparently act to maintain a consistent 25 methane (10 ) was again serially diluted into the same medium free energy yield for the catabolic pathway (Fig. 6), and allowed to grow. After 2 weeks, the lowest dilution to grow counteracting thermodynamic changes caused by 23 (10 ) was again serially diluted and allowed to grow. This changes in pH, temperature (Fig. 7), and in the procedure was repeated three additional times until consistent relative proportions of ethanol and acetate.
growth was achieved.
Growth yields measured for P. acetylenicus (WoAcy1) grown on ethanol are low, 2.2 6 0.5 21 the liquid phase) during a separate experiment. The g mol acetate-dry weight, corresponding to the low observed reaction stoichiometry shown in Eq. (1): amount of free energy available for the entropically driven oxidation of ethanol. These yields are, how- ever, similar to those estimated in coculture studies with the same organism (Seitz et al., 1990a). agrees well with the expected stoichiometry. Hydrogen production typically began within minutes of inoculation, and increased for several hours until stabilizing at a critical level corresponding to the 4. Discussion minimum thermodynamic yield (Fig. 4). The partial pressure of H in the exhaust gas of the culture Anaerobic microorganisms, particularly those in-2 vessel typically ranged from 30-85 Pa during this volved in terminal degradation of organic material, growth in pure culture. The key ability which allows are able to grow from very small quantities of P. acetylenicus to conserve energy presumably lies in energy. It is generally accepted that some anaerobic its use of a transmembrane ion pump to drive the microorganisms are able to grow on a 'biological endergonic production of H from NADH (Haus-2 energy quantum' equivalent to the extrusion of one child, 1997). Recent estimates indicate that P. ion from the cytoplasm (Schink, 1997). Other an-acetylenicus utilizes 2 / 3 of ATP production to drive aerobes, like P. acetylenicus, are thought to conserve an electrochemical gradient which in turn drives the energy through substrate level phosphorylation even endergonic production of H (Schink, 1997). The 2 though the thermodynamic yield for the catabolic growth yield and free energy yields observed in the 21 process is lower than the |70 kJ mol required for present pure culture study lend further support to this irreversible synthesis of ATP (Schink, 1997). Our hypothesis. calculations indicate that the amount of energy Calculating thermodynamic yields from cultures available to P. acetylenicus in these studies ranged grown in the apparatus assumes that equilibrium is 21 from 26 to 33 kJ mol , equivalent to the irreversible rapidly achieved between the environment surroundformation of about one third of an ATP per mol of ing the cell, and the gas phase. The small size of the ethanol oxidized (Fig. 6); such an energy yield is bubbles produced by the glass frit and the use of an near the absolute minimum for energy metabolism.
impeller, help to facilitate rapid gas transfer. During Similar energetics and growth yields have been growth, the cultures constantly produce H , and 2 estimated in coculture experiments involving P. therefore maintain an H flux from the cell into the 2 acetylenicus with various H -oxidizing syntrophic surrounding liquid. Each cell is surrounded by a 2 partners (Seitz et al., 1990a,b), though never during diffusive boundary layer in which diffusion is the have a profound influence on the thermodynamics of H production (Conrad and Wetter, 1990). Tempera-2 ture affects H production through its effect on 2 entropy (DG89 5 DH 2 TDS), which influences the standard Gibbs free energy (DG89), as well as through its influence on the deviation from standard conditions (DG9 5 DG89 1 RT lnhQj). Results shown in Fig. 7 demonstrate the tightly coupled relationship between temperature, free energy yield, and H production. The general result for H produc-2 2 ing reactions, holding all other factors constant, is that higher temperatures allow for higher H con-2 centrations. The converse is true for lower temperatures. Changes in pH can influence the free energy when there is a net production or consumption of protons during metabolism, as is often the case during syntrophic degradation. For example, acetic acid production caused the pH of the liquid culture in the lines of constant free energy.
The culture apparatus described here shows potential for study of other forms of metabolism besides those already discussed. Suitable substrates dominant mixing process (Fenchel et al., 1998). may include additional alcohol, substituted aromat-Each cell experiences a microenvironment of higher ics, acetate, glycolate, and amino acids. The culture localized H concentrations so that use of gas phase apparatus also shows potential for enrichment and 2 H concentrations to calculate thermodynamic yields isolation of other 'syntrophs'. The advantage of a 2 consistently overestimates the actual energy available culture apparatus such as this is that it mimics to the organism. The net effect is that an H natural conditions and fulfills the thermodynamic 2 producing organism within the culture vessel is requirement for low H ; this capability may obviate 2 living from less energy than calculations indicate. the need for use of cocultures in studying many Such factors may explain the small differences forms of H metabolism. 2 between free energy yields calculated with P. acetylenicus, and those calculated in coculture studies (Seitz et al., 1990b).