anaerobic methaneoxidation: occurrence and ecology · trations in lake sediments and digested...

Vol. 39, No. 1APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Jan. 1980, p. 194-2040099-2240/80/01-0194/11$02.00/0

Anaerobic Methane Oxidation: Occurrence and EcologyALEXANDER J. B. ZEHNDER* AND THOMAS D. BROCK

Department ofBacteriology, University of Wisconsin, Madison, Wisconsin 53706

Anoxic sediments and digested sewage sludge anaerobically oxidized methaneto carbon dioxide while producing methane. This strictly anaerobic processshowed a temperature optimum between 25 and 37°C, indicating an activemicrobial participation in this reaction. Methane oxidation in these anaerobichabitats was inhibited by oxygen. The rate of the oxidation followed the rate ofmethane production. The observed anoxic methane oxidation in Lake Mendotaand digested sewage sludge was more sensitive to 2-bromoethanesulfonic acidthan the simultaneous methane formation. Sulfate diminished methane formationas well as methane oxidation. However, in the presence of iron and sulfate theratio of methane oxidized to methane formed increased markedly. Manganesedioxide and higher partial pressures of methane also stimulated the oxidation.The rate of methane oxidation in untreated samples was approximately 2% of theCH4 production rate in Lake Mendota sediments and 8% of that in digestedsludge. This percentage could be increased up to 90% in sludge in the presence of10 mM ferrous sulfate and at a partial pressure of methane of 20 atm (2,027 kPa).

Measurements of methane in freshwater andmarine environments made by several geo-chemists (2, 12, 19) have suggested that an an-aerobic consumption of methane might occur.Although aerobic methane oxidation is a well-known process (9, 18), the existence of any or-ganisms which could oxidize methane anaerobi-cally has been controversial (7, 13, 17, 24, 27).Recently, Zehnder and Brock (30) have shownthat all methane-forming bacteria tested are alsoable to oxidize a small amount of methane an-aerobically and that this oxidation occurs at thesame time that methane is produced. With mostof the methanogens, the only product formedfrom methane is carbon dioxide, but with Meth-anosarcina methanol and acetate are alsoformed and with the "acetate organism" (30) themajor product of methane oxidation is acetate.Zehnder and Brock (30) presented evidence thatmethane oxidation by methanogens was not asimple back reaction, but involved intermediatesdifferent than those involved in methane pro-duction. In the present paper, we extend thestudies on anaerobic methane oxidation to an-aerobic habitats where methanogens are active.We show that anaerobic methane oxidation oc-curs routinely in such habitats and that themethane-producing bacteria might be activelyinvolved in the process. However, a net con-sumption of methane, such as is required toexplain the geochemical data (2, 12, 19), was notobtained, although in the presence of appropri-ate additions and under high pressures of meth-ane, the oxidation rate was 0.9 of the productionrate. In addition to our general studies describing

the nature of the anaerobic oxidation process,we have also carried out experiments in whichvarious alternative electron acceptors [sulfate,iron (III), manganese (IV)] were added to deter-mine whether they might be involved in theprocess. It can be seen from Table 1 that thefree energy values for reaction of methane withsome of these electron acceptors are almost asfavorable as for reaction with 2. Although someof these additions markedly affect the process,we were unable to determine whether their in-fluence was direct (on the methanogens) or in-direct (on other organisms in the system). Thepresent study should serve as a basis for furtherexperiments on the nature and geochemical im-portance of the anaerobic methane oxidationprocess.

MATERIALS AND METHODSLocation and sampling. (i) Freshwater sedi-

ments. Freshwater sediments were obtained fromLake Mendota, a hard-water eutrophic dimictic lakewith a mud bottom. From July until turnover in Oc-tober, the water overlying the sediments is anoxic andcontains hydrogen sulfide. These sediments activelyproduce methane during the entire year (32). Sampleswere taken in the deepest section of the lake with awater column depth of 23 m. Grab samples were takenwith an Eckman dredge (Wildlife Supply Co., Saginaw,Mich.) and were immediately transferred into 200-mlMason jars (ordinarily used for canning). The jarswere filled up to the top with special attention to avoidentrapping air. The ferrous sulfide (5) in the sedimentsprevented harm to anaerobes by the brief exposure tooxygen while the sediments were dispensed (29). Sed-iment samples could be stored at 4°C for at least 2weeks without any loss of methanogenic activity.

194

on February 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

ANAEROBIC METHANE OXIDATION 195

TABLE 1. Gibbs free energy' for methane aselectron donor and various inorganic electron

acceptors atpH 7 and 25oCbAG

Electron acceptorkcal/e- kJ/e-

CO2 0 0SO42-c -0.67 -2.8FeOOH(s)d -5.4 -22.6MnO2(s)' -19.46 -81.48N03-f -22.77 -95.3402 -24.6 -103.00

aCalculated from values in Wagman et al. (26) andStumm and Morgan (25).

b These equilibria have been calculated assumingthe following concentrations: [H+] = 10-7 M; [HCO3-]= 10-3 M; [HS-] = 10-3 M; [SO42-] = 10-2 M; [NO3-]= 10-2 M; PCH, = 1 atm; Po2 = 0.2 atm.

C Reduced form: H2S/HS- at pH 7.dAssuming the amorphous form in the oxidized

state and FeCO3(s) (siderite) in the reduced state. Incase of Fe(OH)2(s) or a-FeS(s) as the reduced state,the energy changes are 0.92 kcal/e- (3.9 kJ/e-) and-11.34 kcal/e- (47.5 kJ/e-), respectively.

' Manganate IV, "a-MnO2"(s), in the oxidized stateand rhodochrosite MnCO3(s), in the reduced state.Between pe8 and 10.5 (470 and 620 mV) y-MnOOH(s)(manganite) and Mn3O4(s) (hausmannite) may be sta-ble and act as possible electron acceptors. The Gibbsfree energy for manganite reduction is -17.56 kcal/e-(73.5 kJ/e-); the reduction of hausmannite yields-18.15 kcal/e- (78.0 kJ/e-).

f Reduced form: N2.g Reduced form: H20.

Cores (6.6-cm diameter) were ordinarily taken by ascuba diver. After the cores were removed, they weresealed with black rubber stoppers at the lake bottomand immediately taken to the laboratory in their liners.Sediment samples from the different depths weretaken with a syringe through holes which were drilledat 1-cm intervals into the liners before core samplingin the lake (the holes were sealed with adhesive tapeto avoid loss of liquid during transport). In some casesa Benthos core sampler was used, taking into accountthat the uppermost 20 cm of the sediment was consid-erably compressed by this method. Subsamples fromthese cores were taken as described above.

(ii) Marine sediments. A 20-cm core of sediment,black and smelling strongly of hydrogen sulfide, wasused in the marine sediment study. The core wascollected by Michael Klug in the Izembek Bay whichis adjacent to Cold Bay, Alaska. The bay supports oneof the most productive sea grass beds (Zostera mar-ina) in the world. The core originated from an areadevoid of grasses, but having a heavy detrital mat overthe sediments. The area is about 2 m deep at high tideand never completely exposed at the lowest tides.Syringes could not be used to take subsamples, be-cause broken shells and coarse sand caused cloggingeven when large-bore needles were used. Therefore,we applied the following method: the lower part of thecore liner was sealed with a one-hole rubber stopper.Oxygen- and sulfate-free ocean water (see below for

composition) was added slowly through the hole witha slight overpressure to force the sediment core to thetop of the liner. Subsamples were taken by carefullypushing a glass tube (1-cm inner diameter) 1 or 2 cminto the sediment. To avoid compression, a slightvacuum was applied to the tube. The subsamples weretransferred to serum vials which were immediatelysealed and made anaerobic (see below). This proce-dure allowed only minimal exposure of a minor part ofthe sediment to the air. Traces of oxygen which never-theless difused into the sediment were immediatelyremoved in the serum vials (see below).

(iii) Digested sewage sludge. Digested sewagesludge was obtained from the digestor of the Madisonmetropolitan sewage treatment plant. This slightlyoverloaded digestor runs on an average loading of 3kg/m3 (as volatile solids) with a mean retention timeof 8 to 10 days.

Incubation procedure. All incubations were madein the dark in 35-ml serum vials with 20 ml of liquid.The vials were closed with black-lip rubber stoppers,sealed with an aluminum seal (1), and made anaerobicby evacuating and flushing alternately several timeswith the gas mixture desired. Additions of nonvolatilesubstances were made before the flushing procedure.To dilute samples, we used either a mineral salt me-dium or synthetic ocean water.

Mineral salt medium consisted of the following (ingrams per liter of distilled water): KH2PO4, 0.41;Na2HPO4, 0.43; NH2Cl, 0.48; NaCl, 0.48; CaCl2-2H20,0.18; MgCl2.6H20, 0.16. The free bicarbonate concen-trations in lake sediments and digested sludge differgreatly. Therefore, sodium bicarbonate was added insuch amounts that the final concentration in the in-cubation vials was the same as in the natural sample.The pH was kept at neutrality by means of carbondioxide in the headspace. The anaerobic metabolismof different substrates supplied often increased thepH, which was detected by measuring the carbondioxide concentration in the headspace. When neces-sary, the pH was readjusted by injecting carbon diox-ide into the headspace.

Synthetic ocean water. The amounts of thechemicals in grams per liter of distilled water werecalculated from the total ion composition of seawatergiven by Riley and Chester (23): NaCl, 23.8; MgCl2.6H20, 11; Na2SO4, 4; CaCl2.2H20, 1.5:, KCI, 0.76;NaHCO3, 0.2; NaBr, 0.082; SrCl2.6H20, 0.024; KF.2H20, 0.0066; H3BO3, 0.0037. For most of our studies,we omitted the sodium sulfate. To keep the ionicstrength constant, 3 g/liter of NaCl were added in-stead. The pH was kept at neutrality as describedabove.

Colloidal MnO2. To precipitate colloidal man-ganese dioxide, the procedure described by Morganand Stumm (15) was used. The "B-MnO2" (manganateIV) (25) formed was washed five times with and sub-sequently stored in distilled water.

Determination of methane oxidation. Radioac-tive 14CH4 was injected into the headspace with unla-beled methane (2 ml). Radioactive methane was madefrom ['4C]bicarbonate or [2-'4C]acetate by means ofmethanogenic organisms (30). This methane was rig-orously tested for radiochemical purity, and no con-taminations have been detected (30). To analyze for

VOL. 39, 1980


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

196 ZEHNDER AND BROCK

methane oxidation products, 2 ml of liquid was re-moved with a syringe and injected through a rubberseptum into a 10-ml serum vial containing 1 ml of 1.5N NaOH. Subsequently the vial was vigorously shakento absorb all C02 or volatile thiols. To fractionate thesample, first a short needle was inserted into the gasspace of the serum vial. To obtain volatile amines, airwas bubbled through the vial and sequentially bubbledthrough two scintillation vials, in series, which con-tained 0.1 N H2SO4 (23). Scintillation vials were sub-sequently replaced by three others in series. The firstvial contained 4 ml of an acid (pH < 2) 3% (wt/vol)HgCl2 solution to absorb volatile thiols such as meth-ane thiol and hydrogen sulfide. The next two vialseach contained 2 ml of phenethylamine (scintillationgrade) and 2 ml ofmethanol to trap the carbon dioxide.To liberate these acid-volatile compounds, 2 ml of a 6N H2SO4 solution was carefully injected into the serumvial containing the alkaline sample. After the C02 wasbubbled off, the acidified sample was filtered througha glass fiber filter (Whatman glass microfiber paper,grade GF/C, Scientific Products), and the filtrate wascollected. Then, 4 ml of the filtrate, the 0.1 N H2S04trapping solution, and the HgCl2 solution were eachmixed with Aquasol (New England Nuclear Corp.,Boston, Mass.). The scintillation vials with pheneth-ylamine-methanol solution received 10 ml of a toluenefluor mixture with 0.375 g of PPO (2,5-diphenyloxa-zole, Beckman Instruments, Inc., Fullerton, Calif.) and0.1 g of dimethyl-POPOP (1,4-bis[2(4 methyl-5-phen-yloxazolyl)]-benzene, Packard Instrument Co., Inc.)per 1,000 ml of toluene. All 14C radioactivity wascounted with a Tri-Carb 3375 scintillation spectrome-ter (Packard) with the window set at 40 to 1,000 andthe gain set at 12%. Quench corrections were made bythe channels ratio method.

In the course of incubation some samples producedmethane, and the ['4C]methane was therefore contin-uously diluted. The specific activity of methane wasdetermined for each time point by the method de-scribed by Zehnder et al. (31). The total amount ofCH4 oxidized was calculated by the method of Zehnderand Brock (30). In a time course experiment it wasnecessary to determine the distribution of the carbon-ate species to calculate the total amount of "'4C02"[CO2 refers to the total of C02(gas), C02(dissolved),HC03-, and CO32-] formed from methane. For thispurpose, control samples were prepared exactly asthose in which the methane oxidation had to be fol-lowed, but instead of radioactive methane 0.5 uCi ofNaH'4C03 (50 mCi/mmol) was added. These controlswere sampled simultaneously with the other vials todetermine the distribution of the carbonate species.The radioactive bicarbonate controls were not neces-sary if an entire incubation mixture was made alkalineto trap all of the CO2 in the liquid phase. In this casesamples for the amine analysis were taken directlyfrom the headspace and injected into vials containing0.1 N H2SO4. In all cases, the dissolved radioactivemethane was bubbled off with air. Freshly prepared['4C]bicarbonate controls were taken to check thereproducibility and efficiency of the radioactive CO2recovery. Before the analysis the vials were shaken for30 min to allow the carbonate species to equilibrate.The standard deviation of the C02 measurements did

not exceed 5% under the condition used (e.g., verycarefully bubbling and scintillation counting at leastup to 2,000 cpm).Tubes for higher pressures. For pressures higher

than 3 atm, the methods of Balch and Wolfe (1) weremodified by using autoclavable stainless steel tubesand a special device for pressurizing them (Fig. 1). Thetubes were shaken during pressurization to allow gasesto dissolve.Gas analysis. Oxygen and carbon dioxide were

measured with a Packard model 419 gas chromato-graph equipped with a thermal conductivity detector.Oxygen was quantified with a column (70 cm long, 2.5mm inside diameter) held at room temperature andpacked with molecular sieve (100/120 mesh). Heliumas carrier gas had a flow rate of 30 ml/min. For thequantitation of carbon dioxide, the same settings wereused, except that Poropack QS (80/100 mesh) wasused as stationary phase. The detection linits for

5

3.

FIG. 1. High pressure tube and device for pressur-ization. All parts in stainless steel if not otherwisestated. Part 1, Cap 5/8 inch (ca. 15.9 mm) with holein the center (Swagelok). Part 2, Black rubber stopperno. 0 with the top cut off. Part 3, Tube 18 cm long,outer diameter 5/8 inch, with a 0.65-inch (ca. 16.5mm) wall. Part 4, Same cap as under I but withouthole. This tube should hold up to 250 atm (25,331kPa). Part 5, Tube with 1/4-inch (ca. 6.4-mm) outerdiameter filled with cotton to use as sterile filter. Part6, Female connector (Swagelok). Part 7, Connector toLuer Lock made out ofKel-F (Hamilton, Co., part no.86536). Part 8, Disposable needle.

APPL. ENVIRON. MICROBIOL.


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

VOL. 39, 1980

oxygen were 0.1 nmol and those for carbon dioxidewere 2.1 umol. Methane was assayed with a gas chro-matograph with a flame ionization detector (31). Athigh methane partial pressures it was necessary todilute samples before injection.

Control experiments. Two different "killed" con-trols were made for each experiment and addition. Inthe one, high temperature (900C) was used to preventbiological activity. The other control received 10 mgof mercury bichloride per ml of solution. This fairlyhigh concentration was needed because mercury formsan insoluble salt with sulfide which is present in con-siderable amounts in sediments and digested sludge.These two control experiments were chosen based on

the consideration that a system should undergo theleast changes while all biologically mediated reactionswere inactivated (4, 30).

Chemicals, radiochemicals, and gases used. 2-Bromoethanesulfonic acid sodium salt was obtainedfrom Eastman Kodak Co., Rochester, N.Y. All chem-icals were of reagent grade. ['4C]Sodium bicarbonate(50 mCi/mmol), [2-'4C]sodium acetate (2 mCi/mmol),['4C]sodium formate (3.9 mCi/mmol), ['4C]methanol(0.85 mCi/mmol), [14C]methanethiol (14.5 mCi/mmol), and [methyl-'4C]methionine (10 mCi/mmol)were purchased from New England Nuclear, Boston,Mass. Gases and gas mixtures were purchased fromMatheson Gas Products, Joliet, Ill. in anaerobic purity.We never detected in these gases oxygen contamina-tion.

RESULTS

Methane formation and methane oxida-tion in anaerobic habitats. When anaerobicsediments or digested sludge are incubated an-

aerobically under a headspace of nitrogen andcarbon dioxide which contains ['4C]methane, aproduction of 14CO2 can be observed. Theamount of 14C02 formed increases with time inparallel with the build-up of "endogenous"methane (Fig. 2). A typical time course of suchan experiment is shown in Fig. 2. The oxidationproduct from methane was >99% carbon dioxide.The rest of the radioactivity was acid soluble,and no labeled amines or thiols were found. Toensure that the observed formation of carbondioxide from methane was not due to oxygenleaking into the vials through the black rubberstoppers, serum vials with anaerobic black mudfrom Lake Mendota were either incubated justas they were or immersed upside down in asealed anaerobe jar (BBL Microbiology Sys-tems, Cockeysville, Md.) filled to the top with10mM neutralized titanium (III) citrate solution(33) (water containing sulfide as an anaerobicbarrier is not suitable because the aluminumseals of the serum vials react with sulfide andare subsequently dissolved). This setup made itunlikely that oxygen might penetrate into thevials containing the sediments. The results ofthis test are summarized in Table 2 and clearly


0 4 8 12 16 20 0 2 6 10

DAY DAY6

FIG. 2. Typical time course of methane formationand simultaneous anaerobic methane oxidation byanoxic Lake Mendota surface sediment collected at

the end of the summer stagnation period and bydigested sewage sludge. With both materials, 10 mlwas diluted with an additional 10 ml ofmineral saltsmedium. Initial specific activity of methane, 39 ,iCi/mmol. A heat inactivated and HgCl2 killed sampleserved as controls.

TABLE 2. Methane formation and oxidation by lakesedimentsa

CH4 CH4 oxidized (dpm)bSample formed Acid Sol-

(ml) 14C02 uble'

Immersedd37°C 7.31 ± 0.52 62,400 + 4,800 180 ± 48

Not immersed370C 7.25 ± 0.61 63,500 ± 5,300 210 ± 41370C + HgCl2 <0.05 450 ± 63 120 ± 50

(10 mg/ml)90°C <0.05 720 ± 82 88 ± 43

'Sediment (10 ml) 1:1 diluted and incubated for 7 daysunder an atmosphere of nitrogen and carbon dioxide. Standarddeviation of mean of three assays.bA 2.3-ACi amount (5.1 x 106 dpm) of CH4 added per vial

in 2 ml of CH4.'Acid-soluble radioactivity was measured after filtration

and removal of C02. Background counts are already sub-tracted from these values. The acid-soluble counts in the killedsamples are at a 95% confidence interval not different fromthe background counts.

dIncubated in an anaerobic jar, filled to the top with a 10mM neutralized titanium III citrate solution (33).

show that a possible leak of traces of oxygen isnot responsible for the methane oxidized. Thevery low conversions of ['4C]methane in thepoisoned and heat-treated samples strongly sug-gest that the observed C02 formation from CH4is due to biological activity. The active vialsproduced unlabeled methane and therefore con-tinuously lowered its specific activity (mean spe-cific activity of CH4, 11.1 ,uCi/mmol). In thekilled samples, however, the mean specific activ-ity of methane remained constant and higher


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/


(mean specific activity of CH4, 25.6 ,uCi/mmol).This fact, together with the low amount of[I4C]methane converted to '4CO2 in the killedsamples, let us conclude that a possible physicalor chemical process which is not biologicallycatabolyzed accounts only for a quantitativelyunimportant portion of the total amount of CH4converted to C02. A dilution of the sedimentinoculum resulted in a slower methane produc-tion; the rate of methane oxidation was loweredas well. The ratio of methane oxidized to meth-ane formed, however, remained relatively con-stant: 0.016 with 1:1 (vol/vol) diluted sedimentsand 0.019 with a 1:10 (vol/vol) dilution. This isa further evidence that the observed methaneoxidation is due to the presence of the anaerobichabitat. The dilution experiments and thosesummarized in Table 2 were also performed withdigested sludge. Qualitatively there was no dif-ference between the results obtained with sludgeand those obtained with lake sediments. Allexperiments presented below were done withsediments and digested sludge, but we report theresults obtained with digested sludge only iftheymarkedly differ from those of the sediments orif they help to clarify the findings from thesediments.Effect of oxygen on methane formation

and oxidation. By transferring the sediment orthe digested sludge into the serum vials, airmight have oxidized some reduced compoundswhich can serve subsequently as electron accep-tors for methane oxidation. To test this hypoth-esis, a series of anaerobic sediment samples(taken during the stagnation period) were di-luted and shaken intensively and aerobically ona rotary shaker at 370C in an open Erlenmeyerflask for 6 h. The black mud turned to brownduring this treatment. This brown sedimentslurry was then incubated in sealed serum vialswith air as headspace. No positive effect of ox-ygen on methane oxidation was observed (Fig.3), and the oxygen was consumed (presumablyby the facultative anaerobic microorganisms).Methane formation and methane oxidationstarted simultaneously only after the oxygentension was reduced. The vials were not shakenduring incubation; therefore, strict anaerobicmicroenvironments could be formed and meth-ane could be produced in a vial where someoxygen was still present. The addition of acetateresults in a faster oxygen consumption and con-sequently a shorter lag phase for methane for-mation and oxidation (data not shown). In di-gested sludge, oxygen was used up after only 1day of incubation, but qualitatively the resultsobtained with sludge were similar to those inFig. 3. The inhibition of methane oxidation by

oxygen in these sediments suggests that oxygencontamination is not responsible for the meth-ane oxidation observed in our vials.Temperature optimum for methane for-

mation and oxidation. The demonstration ofa temperature optimum provides strong evi-dence that one step in a reaction observed isbiologically mediated (4). Methane oxidation insediments showed a distinct temperature opti-mum. For comparison, the methane formationoptima are shown as well (Fig. 4). When thesediments are incubated for 2 weeks at highertemperatures (45 to 7000) a thermophilic pop-ulation develops which forms methane and oxi-dizes it at an optimum of 550C. Under thenno-philic conditions methane is forned both fromcarbon dioxide reduction as well as from acetatedecarboxylation (data not shown).

E

FIG. 3. Effect ofoxygen on methane formation andanaerobic methane oxidation in previously anoxicsediments ofLakeMendota. Initially, 2 ml ofmethane(specific activity 27 ,uCi/mmol) was added to 15 ml ofheadspace. Incubation was at 15°C.

TEMPERATURE (°C)

FIG. 4. Temperature optimum for endogenousmethane formation and anaerobic methane oxidationby anoxic surface sediments of Lake Mendota.Mesophilic optimum (solid lines) measured after 7days of incubation and the thermophilic (dashedlines) optimum after a total of 14 days.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

VOL. 39, 1980

Depth distribution of methane oxidationin anoxic sediments of Lake Mendota andIzembek Bay. Methane oxidation as a functionof the depth in a core of sediment from LakeMendota was monitored for 8 weeks. Theamount of methane produced and oxidized in-creased steadily during the time observed. Thevalues for week 6 are given in Fig. 5. This coreshows two distinct peaks of active methane for-mation. Due to the softness of the sediments,the two peaks can only be seen in Lake Mendotasamples if cores were taken by a scuba diver. Inthe case of gravity core samples, maximummethane formation and oxidation occur in thefirst centimeter ofthe sediment. This is probablydue to the compression of the very loose top 8cm of the sediment. As an example of a marinehabitat, Izembek Bay sediments were investi-gated over a period of 125 days (Fig. 6). Thedifferent sections of this core show a very longlag before methane fornation and oxidationstarts. Here, as in Fig. 5, methane oxidation isparalleled by methane fornation.

METHANE OXIDIZED (,CMOL)0 Ql 02 0.3 Q4 0.5 0.6 07

METHANE FORMED (,MOL)

E

0

FIG. 5. Profile ofmethane formed and oxidized ina core ofLake Mendota sediment sampled 26 August1977 and measured after 6 weeks of incubation at10°C. The values are given for 10 cm3 of wet sedi-ments. The dry weight values for each depth given as

milligrams per milliliter oforiginal sediments are as

follows: 0 cm, 0.5; 1 cm, 12.2; 2 cm, 8.0; 3 cm, 15.2; 5cm, 33.2; 7 cm, 26.8; 10 cm, 134.8; 15 cm, 161.0; 22 cm,

199.2; and 32 cm, 286.0.


0 0.2

_0 20

METHANE OXIDIZED (vMOL)04 0.6 0.8 1.0 1.2 1.4

FIG. 6. Depth profile of methane formed and oxi-dized by a core from the eelgrass beds ofthe IzembekBay Lagoon, Alaska. Incubated at 15°C for 125 days.The insert shows the time course of the surface of thesediments. The values are given for 10 cm3 of wetsediments.

Inhibition of methane formation and ox-idation with 2-bromoethanesulfonic acid. 2-Bromoethanesulfonic acid is a specific inhibitorfor methane bacteria (1). Figure 7 gives a dose-response curve of digested sludge and Lake Men-dota sediments for 2-bromoethanesulfonic acid.As shown, the anaerobic methane oxidationprocess shows an even greater sensitivity to theinhibitor than methane formation. Methane for-mation is only inhibited at relatively high con-centrations compared with pure culture studies(8, 30), which show 50% inhibition at about a106 M. In the complex system investigated here,methanogenesis had a slower response to theinhibitor and a 24-h preincubation with 2-bro-moethanesulfonic acid had to precede the actualexperiments to observe the full effect of theinhibitor. After preincubation with the inhibitorthe headspace was made free of methane, andsubsequently 2 ml of 14CH- of (44.8,uCi/mmol)was added. The amount of methane oxidizedwas determined after 48 h.

In Table 3 the 50% inhibition concentrationsfor 2-bromoethanesulfonic acid are listed foreach individual substrate capable of supportingmethanogenesis. Methane oxidation is by far the


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/


-6 -5 -4 -3 `2 -e -7 -6 -5 -4LOG [Br CH2 CH2 SO3 Na] (MOLAR)

FIG. 7. Inhibition of methane formation and oxi-dation by 2-bromethane-sulfonic acid in anoxic LakeMendota sediments (B) and digested sewage sludge(A). Samples of both habitats were diluted 1:2 withmineral salts medium. Incubation was at 37and 15°Cfor sludge and sediment, respectively. Values ob-tained in absence of2-bromoethanesulfonic acid rep-resent 0%o inhibition.

TABLE 3. Inhibition of endogenous methaneproduction, anaerobic methane oxidation, andmethane formation from various precursors by 2-

bromoethanesulfonic acid in Lake Mendotasedimentsa

Substrate or process Br CH2CH2bi3NaAcetate ([2-l4C]acetate)b. 7.1 x 10-3cMethionine (L-[methyl-'4C]methio-

nine) .. ...... ...... 3.2 x 1O-3Endogenous CH4 production .. 2.8 x 1O-3Methane thiol ([14C]methane thiol) 2.1 x 1o-3CO2 reduction (['4C]bicarbonate) 1.0 x 10-3Methanol ([14C]methanol) ......... 6.3 x 10'Methane oxidation ([14C]methane) 4.5 x 105

a Twenty-four hours preincubated with the inhibi-tor and then 48 h incubated with the labeled substrateat 15°C. In case of total methane formation the head-space was made free of methane after this 24 h. Theinhibitor concentrations tested were 0, 10-6 10-5, 10-4,10-3, and 10-2 M.

b The nature of labeled substrate is given in paren-theses.

c These values are means of duplicates. They mayvary in the same order of magnitude, but keep theiroverall consecutive order.

most sensitive of all the processes to 2-bromo-ethanesulfonic acid. This inhibition study, to-gether with the above results, suggest that theremight be a direct link between methane forma-tion and anaerobic methane oxidation. The fol-lowing experiments were initiated to testwhether in our systems sulfate reducers mighttake part in an active anaerobic methane oxi-dation process. Various amounts of sodium sul-fate were added to diluted (1:1) sediment sam-

ples from Lake Mendota, and methane forma-tion and oxidation were measured (Fig. 8). Meth-ane formation is clearly reduced by the presenceof sodium sulfate, and so is methane oxidation.Neither in Lake Mendota sediments nor in di-gested sludge is methane oxidation stimulatedby the presence of increasing amounts of sodiumsulfate. Under our experimental conditions, it isimportant to know not only the total amount ofCH4 oxidized but also the consumption rate ofmethane relative to its production rate. There-fore we have plotted in Fig. 8 the ratio of meth-ane oxidized to methane formed. If a net meth-ane oxidation occurred, this ratio would becomegreater than 1. It can be seen that sodium sulfateor a metabolic product of sulfate inhibits notonly methane production but to an even greaterextent methane oxidation.

Effects of various additions on anaerobicmethane oxidation. The above results showthat sulfate alone cannot increase the relativeamount of methane oxidized compared with theamount of methane formed. Therefore, we in-vestigated whether other additions might in-crease this ratio (Table 4). The values in thisand the following tables originate from timecourse experiments. The 14-day time point waschosen because the ratio of methane oxidized tomethane formed levels off at this time (Fig. 8).The addition of either acetate or hydrogen assubstrates increased the amount of methane ox-idized but simultaneously resulted in a highermethane production, and hence did not changethe ratio significantly. Methane oxidation overmethane production is stimulated to the sameextent by iron (III) chloride and iron (II) chlo-ride. Since iron (III) is likely to be convertedimmediately to iron (II) in our strongly reducing

0

0

C:

a0

I

UC)x0

L. MENDOTA SEDIMENT

NO SOADDED M = =

+10"I SO2

4 a 12 16 20 24 28DAY

FIG. 8. Ratio of methane oxidized to methaneformed as a function of various amounts of sodiumsulfate added to the sediments of Lake Mendota.Incubation was at 150C.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

VOL. 39, 1980

systems by a pure chemical reaction, only thelatter is listed in Table 4. Nitrate (10 mM)entirely stopped methane formation and meth-ane oxidation. It was therefore not further in-vestigated. A combination of acetate or hydro-gen with sulfate resulted in higher methane pro-

duction and oxidation, but the ratio still re-

mained lower than the one obtained without anyaddition (data not shown).Effects of elevated partial pressure of

methane on the anaerobic methane oxida-tion. As seen in Table 5 higher partial pressures

ofmethane give a considerable increase in meth-ane oxidation, especially in the presence ofsomepotential electron acceptors. Interestingly, ace-


tate eliminates the stimulation of both iron sul-fate and manganese dioxide. The standard de-viation between duplicates in Table 4 and 5 wasat most 14%. In the same assay, however, theoxidation compared with the formation did notshow such a variation; that means at a lowerCH4 production methane oxidation was alsosmaller. Experiments with 20 atm (2,027 kPa) ofmethane were performed in stainless steel tubes(Fig. 1). The initial high partial pressure ofmethane renders a direct measurement of meth-ane formation impossible. Therefore, for eachoxidation experiment an identical tube contain-ing [14C]bicarbonate (50 mCi/mmol) instead of14CH4 was incubated as well. The methane

TABLE 4. Methane formation and methane oxidation by Lake Mendota sediments and digested sludge as afunction of various additional substrates under an atmosphere which initially contained 0.12 atm (12.2 kPa)

of methanea

CH4 formed (uxmol) CH4 oxidized (,umol) formed)Added sub- Amt

formed)

strate Sedimenaltb Digested Digested Seiet DigestedstmvSedinen sludge Sediment sludge Sediment sludge

Noned 333 1,770 5.48 136.3 0.016 0.077H2 1.5 490 2,050 8.33 153.7 0.017 0.075Acetate 0.2 530 1,920 9.01 149.8 0.017 0.078Na2SO4 0.2 138 1,530 0.41 91.8 0.003 0.060FeCl2 0.2 410 1,790 20.5 143.2 0.05 0.080FeSO4 0.2 130 1,550 16.9 170.5 0.13 0.11MnO2 0.2 110 1,710 9.9 171.0 0.09 0.10

a Fourteen days of incubation. Means of duplicates. Variations between each of a pair: -20%; with a resultingmaximal standard deviation of + 14%. Controls for each addition: heat inactivated and killed with HgCl2. Allvalues represent the measured values niinus the controls.

bLake Mendota sediment, 1:1 diluted. Collected during summer stagnation. Incubated at 150C.'Digested sludge, 1:1 diluted. Incubated at 370C.d Endogenous OH4 formation.

TABLE 5. Methane formation and methane oxidation by Lake Mendota sediments and digested sludgewhen various substrates are added, under an atmosphere with a methane partialpressure of2 atm (203

kPa)a

CH4 formed (Lmol) CH4 oxidized (umol) Ratio(CH4 oxidized/Amt (mmol/ OH4 formed)

Added substrate vial)

Sediment' Digeste Sediment Digested Sediment Digestedsludge' sludge sludge

Noned 294.0 1,800 10.3 290.0 0.035 0.164Acetate 0.2 483.1 1,980 17.4 333.0 0.036 0.168Na2SO4 0.2 73.4 1,570 0.37 169.0 0.005 0.108FeCl2 0.2 108.2 1,630 9.74 319.0 0.09 0.190FeSO4 0.2 22.0 1,535 3.28 224.0 0.149 0.264FeSO4 + acetate 0.2 + 0.2 214.0 1,715 10.70 288.0 0.05 0.168MnO2 0.2 7.6 1,680 2.77 405.0 0.364 0.241MnO2 + acetate 0.2 + 0.2 245.3 1,795 11.0 290.8 0.045 0.162a Fourteen days of incubation. Means of duplicates. Variations between each of a pair: <20% with a resulting

maximal standard deviation of +14%. Controls for each addition: Heat inactivated and killed with HgCl2. Allvalues represent the measured values minus the controls.

bLake Mendota sediment 1:1 diluted collected during summer stagnation. Incubation at 15°C.c Digested sludge diluted 1:1. Incubated at 37°C.d Endogenous CH4 formation.


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/


formed through carbon dioxide reduction couldthus be determined very accurately (the C02pool size was not altered with the 20 ,tCi ofradioactive bicarbonate). Total methane fonna-tion was assumed to be three times the amountof methane formed from carbon dioxide (10, 11,14, 22). To assure that stainless steel does nothave an effect on methane oxidation, incuba-tions were also made with glass and stainlesssteel tubes with an atmosphere containing 1 atm(203 kPa) of methane. In both kinds of con-tainers similar results were obtained. As com-pared to Table 5 (no additional substrate), theratio of methane oxidized over methane formedincreased to 0.11 and 0.43 for the sediments anddigested sludge, respectively. High nitrogen par-tial pressures did not stimulate methane oxida-tion over methane formation. To investigate theobserved pressure effect further, we incubated1:15 by volume diluted sludge under 20 atm(2,027 kPa) ofmethane with different substrates,and 2-bromoethanesulfonic acid as inhibitor. Inthis experiment the sludge was diluted to reducethe organic carbon supply and consequently thesubstrate concentration for methane formation.The inhibitor should give us an indication ofwhether the additional substrate activates a dif-ferent group of organisms not sensitive to thiscoenzyme M analog. As seen, 2-bromoethane-sulfonic acid inhibits, in relative terms, methaneoxidation equally under all conditions used (Ta-ble 6).

DISCUSSION

Factors controlling the distribution and activ-ity of aerobic methane-oxidizing bacteria havebeen studied extensively (21), but whether a netoxidation of methane occurs in anoxic environ-

ments is still uncertain. In our study we usedradiotracers and could clearly show that meth-ane is oxidized anaerobically by microorganisms.In unpublished work we have also found anaer-

obic methane oxidation in activated sludge andsoil samples. However, in the environments andduring the time intervals that we studied, moremethane was formed than consumed. In otherinvestigations in which anaerobic methane oxi-dation was observed or deduced, sulfate was theonly obvious oxidant in these systems (2, 12, 19).In our experiments, sulfate alone inhibits meth-ane oxidation not only in absolute but also inrelative terms (oxidation-production ratio). Sul-fate is readily reduced to sulfide in Lake Men-dota sediments (29). We did not further investi-gate whether the inhibition is caused by sulfateor sulfide. However, since free sulfide slowsmethanogenesis markedly in these sediments(29) and the presence of iron neutralizes thenegative effect of sulfate addition, we think it israther the free sulfide which inhibits our system(iron forms insoluble iron sulfide and leads so toa sulfide detoxification). With none of the dis-cussed electron acceptors were we able to enrichfor methane oxidation, even with incubationtimes over hundreds of days. Methanogenic ac-tivity was always an absolute prerequisite foranaerobic methane oxidation.

2-Bromoethanesulfonic acid is a very specificinhibitor for methane-producing bacteria in purecultures (8). It is thus very important to notethat this compound also affects methane oxida-tion. In samples from natural habitats the oxi-dation process is more sensitive to this inhibitorthan is methane production (Fig. 7). These find-ings suggest also that there is a relationshipbetween methane formation and methane oxi-dation in the environments studied and that the

TABLE 6. Methane formation and methane oxidation in digested sludge (diluted 1:15) under increasedpartial pressure of methane (PCH4 = 20 atm [2,027 kPa]), various substrates, and 2-bromoethanesulfonic

acid acting as inhibitor.

Additional Amt Inhibitorb O o Ratio (CH4 oxidized/substrate (mmol/ (mM) CH4formedO(umol) CH4 oxidizedc (umol) CH4 formed)tube)'~None 20.4 ± 2.1 9.78 ± 1.4 0.480None 0.1 20.0 ± 1.8 2.28 ± 0.3 0.112None 1.0 6.4 ± 0.6 0.76 ± 0.13 0.118FeSO4 0.15 3.9 ± 0.41 3.53 ± 0.53 0.908FeSO4 0.15 0.1 3.9 ± 0.42 0.98 ± 0.14 0.253FeSO4 0.15 1.0 1.3 ± 0.2 0.54 ± 0.1 0.400MnO2 0.15 11.4 ± 1.5 5.24 ± 0.7 0.459MnO2 0.15 0.1 11.2 ± 1.1 1.38 ± 0.21 0.122MnO2 0.15 1.0 3.6 ± 0.2 0.54 ± 0.1 0.148

a A 20-ml tube containing 15 ml of liquid phase.b Added as 2-bromoethanesulfonic acid sodium salt.c Liquid phase (15 ml) incubated 14 days at 37°C. Methane formation was calculated from parallel experiments

containing '4C02 instead of 14CH4. It was assumed that C02 reduction accounts for 1/3 of the total CH4 production.Standard deviation of mean of two assays.



http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/


methane bacteria might at least partially beinvolved in the observed oxidation. These inhi-bition experiments lead to a paradoxical situa-tion, namely that in a digestor the actual amountof methane formed could be slightly increasedby adding 2-bromoethanesulfonic acid. At aninhibitor concentration of 10' M, methane for-mation is not yet affected, but the oxidation isblocked 50% (Fig. 7). Since approximately 8% ofthe methane formed is oxidized in the sludge(Table 4), the overall methane production couldbe increased up to 4% with the above inhibitorconcentration. The fact that the relative stimu-lation by iron sulfate, manganese dioxide, andelevated methane pressures are all inhibitedequally by 2-bromoethanesulfonic acid suggestthat at least one step in the overall oxidationprocess is sensitive to this coenzyme M analogand remains the same regardless of the sub-strates.One explanation for our findings could be that

in the habitats investigated methane is oxidizedto a significant extent only through a coupledtwo-step mechanism. In the first step, the meth-anogenic bacteria activate methane and formintermediates, e.g., acetate and methanol (31).In the second step these compounds are subse-quently oxidized to carbon dioxide by a non-methanogenic population which is able to utilizemanganese dioxide or sulfate as an electron ac-ceptor. This reaction catalyzed by methanogensis under standard conditions endergonic. It canshift into exogonic range if its products are con-tinuously removed and kept at a very low con-centration by organisms of the second step. Thestimulation of the oxidation found with man-ganese dioxide and iron sulfate point in thisdirection. Sulfate alone might not be able tolower the concentration of the intermediatesenough, and in addition the free sulfide formedmay.be toxic. In presence of iron, however, theconcentration of free sulfide from sulfate reduc-tion is very small and hence may result in amore efficient substrate utilization. Table 5 givesa further evidence that a dual step mechanismmight work in that increasing the pool size ofacetate, a possible intermediate (30), inhibits thestimulating effect of manganese dioxide and ironsulfate. The likely involvement of acetate-cleav-ing methanogens rather than carbon dioxide re-ducers is indicated by the fact that acetate is themost important substrate for methane formationin sludge (22) and sediments (29).We do not pretend that our hypothesis pro-

vides a complete explanation for the geochemi-cal observations in anoxic marine environments,but our results provide reasonable evidence thatmethane is not biologically inert in strict anaer-obic ecosystems.

ACKNOWLEDGMENTSWe thank Michael Klug for providing us with Izembek Bay

sediment cores and R. D. Fallon for collecting sediment coresin Lake Mendota with scuba equipment.

Financial support was provided by the Department ofEnergy (contract EY-76-S-02-2161), by the College of Agri-cultural and Life Sciences, and by the Wisconsin AlumniResearch Foundation. A.J.B.Z. was supported in part by apostdoctoral fellowship from the Swiss National Foundationfor Scientific Research.

LITERATURE CITED1. Balch, W. E., and R. S. Wolfe. 1976. New approach to

the cultivation of methanogenic bacteria: 2-mercapto-ethanesulfonic acid (HS-CoM)-dependent growth ofMethanobacterium ruminantium in a pressurized at-mosphere. Appl. Environ. Microbiol. 32:781-791.

2. Barnes, R. O., and E. D. Goldberg. 1976. Methaneproduction and consumption in anoxic marine sedi-ments. Geology 4:297-300.

3. Belyaev, S. S., and K. S. Laurinavichus. 1978. Micro-biological formation of methane in marine sediments,p. 327-337. In W. E. Krumbein (ed.), Environmentalbiogeochemistry and geomicrobiology. Ann Arbor Sci-ence, Ann Arbor.

4. Brock, T. D. 1978. The poisoned control in biogeochem-ical investigations, p. 717-725. In W. E. Krumbein (ed.),Environmental biogeochemistry and geomicrobiology.Ann Arbor Science, Ann Arbor.

5. Brock, T. D., and K. O'Dea. 1977. Amorphous ferroussulfide as a reducing agent for culture of anaerobes.Appl Environ. Microbiol. 33:254-256.

6. Cappenberg, T. E., and R. A. Prins. 1974. Interrelationsbetween sulfate-reducing and methane-producing bac-teria in bottom deposits of a fresh-water lake. HI. Ex-periments with "C-labeled substrates. Antonie vanLeeuwenhoek J. Microbiol. Serol. 40:457-469.

7. Davis, J. B., and H. F. Yarbrough. 1966. Anaerobicoxidation of hydrocarbons by Desulfovibrio desulfuri-cans. Chem. Geol. 1:137-144.

8. Gunsalus, RI H., D. Eirich, J. Romesser, W. Balch, S.Shapiro, and R. S. Wolfe. 1976. Methyl transfer andmethane formation, p. 191-197. In H. G. Schlegel, G.Gottschalk, and N. Pfennig (ed.), Microbial productionand utilization of gases (H2, CH4, CO). E. Goltze, Goet-tingen.

9. Higgins, I. J., and J. R. Quale. 1970. Oxygenation ofmethane by methane-grown Pseudomonas methanicaand Methanomonas methanooxidans. Biochem. J. 118:201-208.

10. Jeris, J. S., and P. L McCarty. 1965. The biochemistryof methane fermentation using '4C-tracers. J. WaterPoll. Control. Fed. 37:178-192.

11. Kaspar, H. F., and K. Wuhrmann. 1978. Kinetic param-eters and relative turnovers ofsome important catabolicreactions in digesting sludge. Appl. Environ. Microbiol.36:1-7.

12. Martens, C. S., and R. A. Berner. 1977. Interstitialwater chemistry of anoxic Long Wland Sound sedi-ments. I. Dissolved gases. Limnol. Oceanogr. 22:10-25.

13. Mason, I. 1977. Methane as a carbon source in biologicaldenitrification. J. Water Poll. Control Fed. 49:855-857.

14. McCarty, P. L. 1964. The methane fernentation, p. 314-343. In H. Heukelekian and C. Dondero (ed.), Principlesand applications in aquatic microbiology. Wiley, NewYork.

15. Morgan, J. J., and W. Stumm. 1964. Colloid chemicalproperties of manganese dioxide. J. Colloid. Sci. 19:347-359.

16. Oremland, RI S., and B. F. Taylor. 1975. Inhibition ofmethanogenesis in marine sediments by acetylene andethylene: validity of the acetylene reduction assay foranaerobic microcosms. Appl. Microbiol. 30:707-709.

VOL. 39, 1980


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/


17. Oremland, R. S., and B. F. Taylor. 1978. Sulfate reduc-tion and methanogenesis in marine sediments. Geo-chim. Cosmochim. Acta 42:209-214.

18. Patt, T. E., G. C. Cole, J. Bland, and R. S. Hanson.1974. Isolation and characterization of bacteria thatgrow on methane and organic compounds as solesources of carbon and energy. J. Bacteriol. 120:955-964.

19. Reeburgh, W. S. 1976. Methane consumption in CariacoTrench waters and sediments. Earth Planet. Sci. Lett.28:337-344.

20. Riley, J. P., and R. Chester. 1971. Introduction to ma-rine chemistry, p. 80. Academic Press Inc., New York.

21. Rudd, J. W. M., and R. D. Hamilton. 1978. Methanecycling in a eutrophic shield lake and its effects onwhole lake metabolism. Limnol. Oceanogr. 23:337-348.

22. Smith, P. H., and R. A. Mah. 1966. Kinetics of acetatemetabolism during sludge digestion. Appl. Microbiol.14:368-371.

23. Smith, D. W., C. B. Fliermans, and T. D. Brock. 1972.Technique for measuring '4CO2 uptake by soil micro-organisms in situ. Appl. Microbiol. 23:595-600.

24. Sorokin, Y. L. 1957. Ability of sulfate reducing bacteriato utilize methane for reduction of sulfate to hydrogensulfide. Dokl. Akad. Nauk. SSSR 115:816-818.

25. Stumm, W., and J. J. Morgan. 1970. Aquatic chemistry.Wiley, New York.

26. Wagman, D. D., W. H. Evans, V. B. Parker, I. Halow,S. M. Bailey, and R. H. Schumm. 1968. Selected


values of chemical thermodynamic properties. Tablesfor the first thirty-four elements. Technical note 270-3.U.S. Department of Commerce, National Bureau ofStandards, Washington, D.C.

27. Wertlieb, D., and W. Vishniac. 1977. Methane utiliza-tion by a strain of Rhodopseudomonas gelatinosa. J.Bacteriol. 93:1722-1724.

28. Winfrey, M. R., D. R. Nelson, S. C. Klevickis, and J.G. Zeikus. 1977. Association of hydrogen metabolismwith methanogenesis in Lake Mendota sediments. Appl.Environ. Microbiol. 33:312-318.

29. Winfrey, M. R., and J. G. Zeikus. 1977. Effect of sulfateon carbon and electron flow during microbial methan-ogenesis in freshwater sediments. Appl. Environ. Micro-biol. 33:275-281.

30. Zehnder, A. J. B., and T. D. Brock. 1979. Methaneformation and methane oxidation by methanogenic bac-teria. J. Bacteriol. 137:420-432.

31. Zehnder, A. J. B., B. Huser, and T. D. Brock. 1979.Measuring radioactive methane with the liquid scintil-lation counter. Appl. Environ. Microbiol. 37:897-899.

32. Zeikus, J. G., and M. R. Winfrey. 1976. Temperaturelimitation of methanogenesis in aquatic sediments.Appl. Environ. Microbiol. 31:99-107.

33. Zehnder, A. J. B., and K. Wuhrmann. 1976. Titanium(III) citrate as a nontoxic oxidation-reduction bufferingsystem for the culture of obligate anaerobes. Science194:1165-1166.


http://aem.asm

.org/D

ownloaded from

http://aem.asm.org/

anaerobic methaneoxidation: occurrence and ecology · trations in lake sediments and digested...

Documents