Aerobic Conversion of Dimethyl Sulfide and Hydrogen Sulfide by Methylophaga Sulfidovorans: Implications for Modeling DMS Convers

Home , Dimethyl sulfide, Hydrogen sulfide

MICR8810bOGY ECOLOGY ELSEVIER FEMS Microb~ologyEcology 22 (1997) 155-165

Aerobic conversion of dimethyl sulfide and hydrogen sulfide by

Methylophaga sulfidovouans: implications for modeling DMS Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 conversion in a microbial mat

Jolyri M.M. de Zwart *, J. Gijs Muenen Department of Microbiology and Enzymology, Deqt University of Technology, Julianalaan 67, 2628 BC Delft, The Netherlands Received I August 1996; revised 7 November 1996; accepted 8 November 1996

Abstract

Methylophaga su~fidovorans is an obligately methylotrophic, DMS-oxidizing organism, isolated from microbial mat sediment. DM§ and H2S, both present in marine microbial mats, can be used as energy sources by this organism. In batch cultures of M. sulfidovorans, sequential H2S and DMS utilization occurred. In energy-limited continuous cultures, with DMS, methanol and H2S as substrates, mixotrophic growth of M. sulfidovorans was observed, showing that at low concentrations these substrates can be used simultaneously. Oxygen and H2S uptake experiments showed that the critical concentration at which sulfide inhibiticn of EMS oxidation occurred was between 15 and 40 p~!!-I. '*.!so ir, crude enrichmezts of D?dS oxidizers a decrease of 50% in DMS-oxidizing capacity for about 200 pmol l-' Hz§ was observed. The new physiological data obtained with the pure cultures of &I. ssu~idovoranswere incorporated in a compartment model of a microbial mat and gave improved predictions of DMS profiles and DMS emissions from the mat, both when phototrophic activity is present (day) and when it is absent (night).

Keyrrjords: Microbial mat; Dimethyl sulfide oxidation; Hydrogen sulfide oxidation; Compartment rnodei; Methylophaga .sul$dovorans

1. Introduction convert organic sulfur compounds. The different groups of microorganisms produce and consume sev- In marine microbiai mats, different functional eral sulfur compounds, of which H2S and DMS are groups of microorganisms form a cooperative com- important representatives. H2S is often produced in munity [I]. The quantitatively most important func- large amounts by sulfate-reducing bacteria. h sul- tional groups are well-described, and consist of fate-reducing activity of 10 pmol ml-' day-' was phototrophs, colorless sulfur bacteria, heterotrophs measured in a Microcoleus chthonoplastes-dominated and sulfate-reducing bacteria. A quantitatively less microbial mat [2]. These high production rates may important group is the population of bacteria that result in sulfide concentrations as high as 200-500 ~mol1-' in the lower, anoxic layers of a mat [3]. HzS is oxidized to sulfate by the (aerobic) colorless -- * Corresponding author. Fax: +3S (IS) 278 2355; and (anoxygenic) phototrophic sulfur bacteria. The e-mail: [email protected] resulting concentration gradient leads to upward dif-

0168-6496197/517.00 Copyright C 1997 Federation of European Mic~.obiologicaSSocieties. Published by Eisevier Science B.V. PIIS01 68-6496(96)00086-4 156 J M. ;M. de Zbvnrt, J. 6. Kuenen l FEiMS Microbiology Ecology 22 11997; 155-165

fusion of H2S in the mat. DMS occurs in much were used as a validation for the assumption that smaller amounts than sulfide. In marine sediments, kinetic and physiological data of M. su&dovorans the main source of DMS is biological degradation of can be used in simulations during a diurnal cycle. dimethyl sulfoniopropionate (DMSP) 141, an osmo- lyte present in many phoiotrophs [5]. The production rate of DM§ during a bloom period of a marine 2. Materials and methods microbial mat was estimated to be + 20 nmol mi-' day-I (a factor 500 lower than sulfide production) 2.1. Bacterial culture and natural samples and coupled to photosynthesis [6]. DMS degradation in sedirncnts can take place under oxic and anoxic Microbial mat sediment samples were obtained in

conditions [7-91. Since it is formed in the upper, May 1996 from the Oosterschelde, an estuarine inter- Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 genera!!y oxic, layer of the mat, it is ass~medthat tidal region on the south-west coast of the DMS will be converted primarily by aerobic bacter- Netherlands. Methylophaga suljidovorans (L,MD ia. As already found, H2S and DMS coexist near the 95.210) had been previously isolated from microbial oxic/anoxic interface of the mat [8]. H2S is consid- mat sediment samples obtained from the same site ered to be an intermediate in the aerobic DMS me- PI. tabolism of known DMS-oxidizing bacteria [9-131, and can be used as a supplementary energy source 2.2. Culture media for growth. Little is known, however, of possible diauxic or inhibitory effects when DMS and H2S Mineral medium contained per liter: 15 g NaCI, are supplied together, or whether the presence of 0.5 g (NH4)2S04, 0.33 g GaC12.6H20, 0.2 g KCI, 1 g H2S affects DMS degradation. These factors may MgS04.7H20, 0.02 g KHzP04, 2 g Na2C03, 1 mg have a considerable effect on the prediction of the FeS04-7H20,1 ml trace element solution [9j, 1 ml release of DM§ from a microbiai mat into the atmo- vitamin solution [9]. pH was maintained at 7.5 sphere. (20.3) with 1 N HC1. Liquid medium containing Using the highest dilution to extinction method, a DMS was kept in glass bottles, which were sealed specialist in aerobic DMS oxidation has been iso- with butyi rubber stoppers to prevent DMS loss. lated: Methylophaga sulJidovorans [9]. This bacterium The purity of the cultures was routinely checked is an obligate methylotroph that can only use metha- by streaking on brain heart infusion (BHI) plates nol, methylamine, dimethyl amine or DM§ as energy supplemented with 1.5% NaCl. and carbon sources. Hydrogen sulfide can be used as --".-.. "I--,. -..I$-+- ' an energy somce. TI-:---.I$1 ~l~vaulla~e + 1 aLii&l Lllail aullaLb i3 tne end product of sulfide oxidation. As M, sulJido- vorans is considered to be a representative bacterium DMS from the headspace of the cultures was for aerobic DMS oxidation in marine microbial measured with a gas chromatograph wlth a sulfur- mats, its kinetic properties could be used to predict specific, fiame photometric detector, equipped with a the effect of Has on DMS degradation in sit=, using Hayesep R column [9j. Thc detection limitwas 0.5 a previously published (mathematical) compartment ppm in the gas phase. This detection method also model [9]. served lor semi-quantitative H2S measurements in In this paper we report on the extension of the serum bottles with M. suljidovorans and with sedi- model with physiological data on H2S inhibition ment samples. The detection limit was 100 pmol on DM§ oxidation. The breakdown of DMS in the I-' in the liquid phase at pH = 7. presence of hydrogen sulfide was studied in batch Oxygen consumption rates by bacterial suspen- and continuous cultures of M. sulJidovorans. The in- sions were measured with a polarographic Clark- hibitory effect of Hz§ on DM§ oxidation in the pure type electrode in a well-mixed Biological Oxygen culture was compared with observations obtained Monitor at 28°C. The oxygen concentration m the with crude enrichments of fresh microbial mat sam- headspace of serum bottles filled with sediment samples. These results of the crude enrichment cultures ples was measured with a Fisons Instruments (8000 J. M. IM. de Zxurt, J G. Kuenen l FEMS 1Micrcbiology Ecologj~22 (1997) 155--165 157 series) gas chromatograph equipped with a molecu- optical denslty (OD, 430 nm) was correlated linearly lar sieve 5A packed column. with the dry weight (DW) and thc organlc carbon Thiosulfate was determined according to Sorbo analys~s(TOC) of these cultures. ?he correlation [14].Sulfide was determined by iodometric titration was. DW (mg 1-1)=194xOD+14 (r=096, n= 15) for sulfide concentrations higher than 1 mmol I-' (detection limit 0.1 mmol I-') [l5]; and by the meth- 2.5. Oxygen and N2S consumption by ylene blue xethod for sul5de concentrations lower ?d. sulJidovorans with and without DMS than 1 mmol 1-"detection limit 1 pmol I-') 1161. Biomass was determined as organic carbon using Cells used for oxygen and H2Suptake experiments total organic carbon anajysis with a Tocamaster type were cultivated in continuous cultures with miiicral 815-B. The carbon content of M. su@dovorans was medium with DMS (2-3 mmol 1-I) as the sole sub- Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 50% (w) [9]. Biomass was also measured by dry strate. H2S uptake 1.vas detemined in vials fi!!ed with weight determinations [9]. 5 ml cell suspension and sealed with butyl rubber At the pH of the experiments, sulfide will be pres- septa. Cells were centrifuged (1 0000 X g) and resus- ent as HS- and H2S. These two forms will be re- pended in 25 :5 mmol I-' carbonate:HEPES buffer in ferred to as 'sulfide' or 'H2S7. Sulfide was always 15 g I-' NaCI. DMS (100 ymol I-') was added with added to media as a solution of Na2S.8H20. a syringe. At time=O, 25 ymol 1-' Na2S.8F120 CpH=8) was added. This low concentration of sulfide was chosen io prevent cheniical oxidaiioii (see Section 2.6). The disappearance rate of sulfide was M. sulJidovorans was cultivated in batch and chem- then measured by stopping the reaction with zinc ostat cultures. Batch cultures in mineral medium acetate (in acetic acid) solution at different time in- supplemented with 10 mmol 1 ' methanol were tervals. The oxygen uptake rate of cell suspension made at 25OC at a pH of 7.5 k0.5 in flasks on a taken directly from the chemostat was measured in rotary shaker at 100 rpm. Batch cultures with a biological oxygen monitor, using 1-12S concentra- DMS and H2S were made in closed serum bottles tions ranging from 0 to 130 pmol I-', with or with- with a butyl rubber septum. T'ne bottles were filled out DMS (260 pmol 1-lj. with 20 : 80% medium:air. At the concentrations of DM§ and H2S used, less than 30% of the available 2.6. Measurement of DMS oxidation capacit~ of oxygen would be required for their complete oxida- crude enrichments at dzferent H2S concent~.ations tion. The cell suspension used for these batch experiments contained only -4 mg I-' (dry weight) to Sedimcnt samp!cs ycrz d:lutcd 125-fold in mineral ensure low but detectable oxidation rates that would medium. The (diluted) slurry (30 ml) was kept sta- make measurements at different time intervals accu- tionary in the dark In 100 mi serum bottles, sealed rate. These bottles were kept stationary in the dark wlth butyl rubber septa. In order to obtain a crude at 25°C at a pH of 7.5 k0.5. Fermentors for contin- enrichment of DMS utilizing bacteria, the diluted uo~~cultivation were made of glass and polycarbo- sed~mentsuspensions wcrc first supplicd wlth 0.5- nate, instead of the more conventional glass and 1.0 mmol 1-I DMS One sample was heated to stainless steei. This experimentai setup was essentiai 9U"G for -2 mm. 4h1s sample showed no DMS for the prevention of severe corrosion problems. disappearance, confinnlng ~tsbiological nature. After Continuous cultures with mineral medium and DMS had been consumed in the unheated bottles growth-limiting methanol (10 mmol I-l) or with with an average rate (determined over the whole growth-limiting mixtures of methanol (10 mmol period of incubation) of -- 5 pmol I-' ml slurry I-'), sulfide (10 mmol I-') and dimethyl sulfide (4 day-', the samples were aerated for 2 min Carbo- mmol I-') as substrates were grown at 28"C, a pO2 nate buffer (-- 10 mmol I-') was added and the pH of 50% air saturation, pH 7.6 and a dilution rate of was adjusted to 7.5 wlth 1 N NC1 DM§ was added 0.03 h-I. For the cultures grown mixotrophical1y on to the bottles (-0.5 mmol I-') and wlthin 24 h, methanol, sulfide and DMS, it was found that the DMS oxadat~onhad resumed Only after 24 h was 158 J ,W. M de Zivart, J C. Kuenen 1 FEMS hfiilicrohiology Ecology 22 (1997) 155-16.5

NazS (0--2 mmol I-') added in order to ensure low oxygen tension in the medium. These samples were all incubated at 25°C. DMS removal was quantitatively determined with gas chromatography. HzS could only be deternilled qualitatively with gas chromatography, since the proportion of FIzS in the gas phase was dependent on the changing pH as thiosulfate was produced from DMS and H2S. The rate of chemical oxidation of sulfide in miner- a! medium was measured with a (bio!c;gical) oxygen monitor. It was found that both for sulfide concen- Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 trations < 100 pmo! 1-l and for p02< 50% air saturation, chemical oxidation was insignificant ( < 0.1 pmol O2 min-'1. In the stationary batch cultures with crude enrichments or with pure cultures of M. su&dovorans with -0.5 mmol 1-I DMS, the oxygen concentration, measured after 24 h of incubation, Fig. 1. Schematic representation of a microbial mat, adapted from [6]. Heterotrophs are situated in the first. second and third was low to instead of only the third compartment, oxidation.

2.7. Compartment model of microbial mat and base production by the microorganisms, and diffusion into the different compartments. The model The compartment model for a microbial mat has is used for simulation of concentration gradients and been described in detail elsewhere [6],and is there- fluxes of these components. In the model, the five fore only described in general terms here. The math- functional groups of microorganisms of the coopera- ematical model is based on biological production tive community are considered to be distributed over and consumption of organic carbon, dimethyi sul- diiTerent compartments, carrying out the reactions fide, hydrogen sulfide and oxygen. These compo- listed in Table 1. These five groups are phototrophs, nents are transported between the different compart- rnethylotrophs (i.e. DMS-oxidizing bacteria), color- ments of the mat by diffusion. The pH in the layers less sulfur bacteria, heterotrophs and, in the lowest of the mat is dominated by a bicarbonate buffer, acid compartment, sulfate-reducing bacteria. The empha- sis of the model was oii DMS emissions and fiiixes. The bacterium Methyloplzaga sul$dovorans (provi- Table 1 sionally named Methylopila suEfovoran,r in a previous List of the functional groups of micro-organisms with their metabolic reactions, used in the mathematical model publication on the model) was (and is) used as a ~ ~-model organism for the DMS-oxidizing community. Functional group Metabolic reaction Thc metabolic activity of the microorganisms in the Oxygenic phototrophs Light+H20+C02-t mat is driven by photosynthesis in the daylight which CH20+02+(CH3)2S" serves as the sole source of organic carbon (Table 1). Chemolithoautotrophs H2S+202+ H2S04

Methylotrophs (CH:I)~S+SO~-t 2COz+H2S04+2H20 The minimum amount of biomass required for sta- H2S+202 -> H~SO~~ tionary growth was used as initial biomass values for Heterotrophs CH20+02-+ C02+H20 the simulations. For the colorless sulfur bacteria, Sulfate reducers HzS04+2CH20-, H2S+2C02+2H20

~ ..- - methylotrophs and sulfate-reducing bacteria, these The stoichiometric relations of the reactions are presented, except carbon-based concentrations were 10, 0.02 and 25 for the DMS production (a) by the phototrophic bacteria. The rnmOl c 1-1, respectively. 1 c-molbiomass was de- DMS formation rate is assumed to be coupled to the average DMSP content of photosynthetic bacteria and the photosynthetic fined as ~~~.800.5N0.3=24 g [6]. The concentration rate [6]. In the present model (b) the methylotrophs can convert of heterotrophs was estimated to be 15 mmol C I-'. both DMS and H2S. The pH in the mat rises sharply during the light 3"". . tlon curve (Monod-llke, wlth afinlty constants of *a %g ,--- 1 pmol I-' 02 and maxlmum oxldatlon rates in the 7 zk 3 1 5 same range as the values for the substrate uptake I kmetlcs). In the earher version of the model a ll~~ear -_ ii I I,; - I zg I relation between the oxygen concentration and max- "- -- I + / - 1 imum specific uptake rate was assumed Thlrdly, pH - cj L 3 dependence of substrate conrxnptlon ratzs of the colorless sulfur bactena. the heterotrophs and the -r 5 -c~:t?2kva- , i -= @A~L+&.S sulfate-reducmg bactena was modeled with a sec- 5" I B ond-order function; the upi~mlrmpH was 7.5 and X 0 a decrease or increase of 2.5 pH unlts reduced these --1 cscj Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 3 1 rates to zero. The pH dependence of the photo- g 2- 25 I Jr- trophic activ~tyand DMS convers~onwas already - o - implemented in the first model Fmally, H2S oxida- :: A 812 2s tion by both the colorless sulfur bacterla and the T nti -ia,s: methylotrophs in compartment 2 is now recognized (Fig. 1, Table I), whereas in the previous model, Fig. 2. Results of continuous cultivation of 1M. sulfidovorans grown in mineral medium with methanol (l0+0.1 mmol 1-11, sulfide was oxidized solely by the colorless sulfur - - UMS (4.5t 1 mmoi I-') and H~S(9i-0.5 mmol 1-I). 011 day o bacteria. T'ne affinity constant used for the coiorless and day 1 the dilution rates were 0.01 and 0.02 h-l, respectively. sulfur bacteria (K,. sulfide) was 1 urn01 1-I sulfide r51.6 \ " J ~romday 3 onwards, the d~lut~onrate was 0 03 h-I The dry For the methylotrophs, ~t was assumed to be we~ghtwas 'constant' for the first 8 days, and was near the value cal to the affinity constant of Methylophugu suljido- of blomass that would be expected ~f only methanol was used as sole energy and carbon source The product of DMS and H2S, vorans, i.eo 5 pmol 1-I sulfide (unpublished results). th~osulfatewas present m the culture after 2 days The expected The mathematical model was developed using the b~onlassy~elds on the three substrates, for these cond~t~ons,are programming meta-language PSI, based on the Ian- ~ndlcatedby the three bars (mg dry I-'), based on the sage C [6] and is available on floppy disk for inter- concentrations of DMS and sulfide In the medlum ?upply The ested readers. rheoretlcal growth curve IS ~nd~catedby the drawn ilne

2.8. Oxygen consumption during dark periods period [6], because of the depletion of the natural (modeling considerations) carbonate buffer (i.e. due to carbon dioxide fixation), and slowly decreases diiir'lng the night, di~eto difhi- ~?n.~rlr.- simulations, reported earlier [5i, were madp sion of acid (H+) and base (OH-.). for a 12 h light period. The simulations have now Four major adjustments to the earlier model were been run for 36 h, in 12 h light and dark periods, made in order to make the simulations in light and instead of only a 12 h light period. This necessitated dark periods realistic. Firstly, the heterotrophic com- a closer look into the metabolic activity of the fui~c- munity, previously situated in the third layer of the tional groups in the dark, when ihere is no oxygen mat, has now been situated in the first, second and production. The sulfate-reducing bacteria produce third layers (Fig. I), with initial biomass concentra- H2S independently of the light or dark [I]. DMS is tions of 2.5, 2.5 and 5 mmol C I-', respectively. excreted by the phototrophs in the top layers of the Compartment 2 now contains three functional mat. The oxygen shortage at night (= substrate ex- groups: the methylotrophs, heterotrophs and color- cess) gives rise to the following two questions. (I) less sulfur bacteria. The presence of oxygen-consum- How is the available oxygen divided over the three ing heterotrophs proved essential in the model to functional groups, the heterotrophs, colorless sulfur allow the oxygen concentration in the lower parts bacteria and methylotrophs? (2) Which substrate, of the mat to decrease to practically zero (i.e. H2S or DM§, will be preferred under O2 limitation < 0.5 pmol 1-I). Secondly, the specific growth rates by the methylotrophs? With respect to the first ques- for oxygen are now simulated according to a saturation, the affinities for oxygen of the different func- 160 J. M.M. de Zvvart, 1. G. Ktienen 1 FEMS LMicrobiology Ecology 22 (1997) 135-165

The second question was which substrate will be preferred by the rncthylotrophs under oxygen lirnita- tion (DMS andlor H2S). Factors that determine this are the affinity constant for the substrate, the maximum specific growth rate and the energy generation per electron (the K,, plf~~and AGeO). A low affinity constant, a high specific growth rate and a high energy (per electron) generation cause a substrate to be favored. The kinetic parameters of Methylophaga su@d~vorans for D?dS are 1 ~;mol1-I (&) and 0.055 h-' (yhIAx).For H2S the affinity constant is Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021

5 pxx! !F1 (K,). The apparent maximurr, specific growth rate was estimated from the H2S consumption rate of - 112 nmol mg biomass-' min-l, observed in batch cultures and the increase of -4 g biomass mol-l H2S observed in continuous cultures with methanol and H2S [6].This would translate into a maximum specific growth rate of 0.026 h-I (pMAX). The energy generaiioii per electririi for sulfide when Time (mid it is combusted fully to sulfate is 21 k4 e-mol-l. For formaldehyde (intermediate in DM§ oxidation) this Fig. 3. H2S uptake experiments in the presence (0)or absence is 45 kJ e-moi-l [17]. Considering these three factors (a) of 100 pnol I-' DMS. Washed cell suspension of 40 mg I-' M. sulJidovorans from a DM§-limited cslture was used. The ac- it may, therefore, be expected that DM§ conversion tivity of the ceils in buffer was -70 nmoi HZ§ min-' mg would be preferred over H2S conversion. biomass-'. The arrow indicates the time when sulfide was added. tionai groups need to be known. in generai, oxygen consumption rates are described by Monod kinetics (the same as for substrate uptake, see 161). The oxygen uptake rate is described by the affinity constant for oxygen, the maximum oxygen uptake rate and +LA ---..lo+:,.- -.-- LlIC: p~pUILILIUII>;LC;. In the mode!, thc value for thc affinity constant for oxygen is assumed to be similar for all three groups (1 ymol 1-I 02).The maximum oxygen uptake rates of the different functional groups are assumed to be in the same range as the maximum substrate uptake rates. In thc present model, sulfide (major substrate) can also be oxidized in addltion to DMS (mmor substrate) by the rnethylotrophs. The ratio of the colorless sulfur bacteria: methylotrophs will be approximately 1 : 1. The actual concentration of both groups in the new model was 5 mmol 1-' C. The total heterotrophic population (compartments 1, 2 and 3) was I0 mmol 1-I C. It can be concluded that oxygen will be used by all three groups in the model as the parameters for the oxygen uptake rate are all of the same order of mag- Fig. 4. DM§ oxidation rate in slurries of microbial mat sediment, nitude. against different arnounts of W2S added under oxic conditions. .I M M de Zcc)arl, J G. Kuenen / FEMS Microbiology Ecollogj 22 (1997) 155---I65 161

2.9. DMS production (modelmg considerations) the thiosulfate concentration rose and became constant aker 2 days. The thiosulfate concentration was It is not yet clear whether DMS production is approximately 6 mmol I-'. This agrees well with the directly coupled to photosynthesis. For DMS pro- expected 6-7 lnmol 1-I for complete oxidation of duction and consumption during the day and night DM§ and H2S. A large increase in the biomass con- in the model, therefore, three different situations (A, centration could be observed after 8 days (5.8 vol- B and C) were considered. In situation A, DM§ ume changes). production is fully coupled to photosynthesis (absent at night). In situation R: DMS production is inde- 3.3. Oxygen and H2S consumption with nizd without pendent of day c;r night, and the average of the prc;- Dm duction in situation A. In the third situation, DMS Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 prodnction is sti!! independent of day or night (as in The sulfide untakeY experiments were carried eut B), but sulfide inhibition (see Section 3) is taken into with cells from a continuous culture with DM§ as account. a growth-limiting substrate (Fig. 3). The 332s uptake rate was not influenced by the presence of DIMS. This experiment was essential for the interpretation 3. Results of the DMS-dependent oxygen uptake experiments with FIBS, i.e. that oxygen uptake at 'high' sulfide 3.1. Batch growth of M. sidj4uJuvorans oii DMS iiiid concentratioii could be (fully) attribilied to the oxi- ff2 S dation of sulfide. It was found that DMS-dependent oxygen uptake was fully inhibited for H2S concen- Batch cultures with both substrates showed se- trations exceeding 40 pmol 1-' (Table 2). quential use of H2S and DMS (diauxy). Only after H2S had been completely oxidized (i.e. < 100 pmol 3.4. Efect of NzS concentration on DMS degradation I-') was DMS consumed. This occurred at all H2S in crude sediment enrichments concentrations tested (0.5-6 mmol I-'). The EI2S oxidation rare for all sarnpies was i i2 k 13 nmoi rnin-I The reiationship between the DMS disappearance mg biomass-'. DMS was oxidized at a rate of rate in crude enrichments of DMS-consuming sedi- 100 + 20 nmol min-' mg biomass-'. Table 2 3.2. continuous cultivation of M. suFdovorans On Resuits of oxygen uptake experiments with Increasing H2S amounts, w:th =r :5.:th=.;: EMS DAfS. i~ethaiioliiiid 1Y2S - H2S (~pmolI-') qO2 (+DMS) qOz (-DMS) Ratlo ------Chemostat cultivation with a growth-limiting mix- 216233 0 ture of methanol, DMS and HZ§ in the medium 8 690k 51 390 2 30 1.8 supply eventually resulted in stable, multiple sub- 15 621 F 88 427 i. 3 15 strate-limited cultures (Eg. 2). On day 0, continuous 28 559 + 130 A~O+15 12 cultivation of a culture pregrswn in batch on rnetha- 36 382 + 99 345 + 5 11 40 342 i- 82 250 + 5 14 no1 started with all three substrates m the medlurn 64 257 + 5 245 i- 24 10 supply at a dilution rate of 0.01 K1.On day I, the 128 240+ 13 250 -1- 5 10 - - rate was increased 0'02 and lhen to The cells were obtained from a DMS-limited continuous culture. 0.03 h-' from the third day onwards. Over a period Cell suspensions were washed in (iron-free) mineral medium. The of 22 days, the optical density (430 nm), total organ- concentration of M, su@dovorcms was 30f 2 mg 1" The DMS concentration was 260 pmol I-'. All data points for the oxygen is carbon., drv. weight.", thiosulfate concentration in the cell suspension, sulfide concentration in the me- uptake with H2S and DMS were measured in triplicate. Hz§ measurements were done in duplicate. q02 stands for specific oxygen dium su~~i~and and the 'On- Uptake rate (iImo[ min-.' mg &ionass-l), The ratio of oxygen centration in the cell suspension were determined at uptake is defined as q02(+DMs)/~o~ (-DM§), F~~ rates > 1, regular intervals. After addition of DMS and Hz$ DMS is (partly) oxidized next to full oxidation of sulfide. Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 J IM 1M de Zctart, J G Kuenenl FEWS ~Wzcroh~ologyEcology 22 11997) 155-16.5 163

tions in all compartments. When the light period starts again, organic carbon is again replenished by photosynthesis. Three different situations of DMS production in a mat were considered (see Section 2.9). For all three cases DMS is initially completely oxidized, during a light period. Eowever, already after a few hears, DMS oxidation stops due to the pH rise (shown in detail in [6]),and the concentration in compartment 2 .--;L,,reases. At night, in the first case (A, Fig. 81, the DMS concentration is nil because DMS is not pro- Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 d~ced.In the second case (B, Fig. I?), DMS conversion takes place after the pH in the mat has settled below 8.5 (after 15 h), during the dark period. This leads to a decrease in the DMS concentration. Full oxidation of DMS does not take place due to the - 3 limited availability of oxygen. In the third simulation 07 (Fig. 8C) the presence of sulfide in compartment 2 0 4 8 12 16 20 24 28 32 36 (Fig. 7) resulis iii the iiihibition of EMS oxidation, and thus leads to a significant concentration of DMS Time (h) during the dark period. The DMS concentration in Fig. 7. Simulation of sulfide concentrations in all four compart- compartment 2 is directly related to the emission of ments of the microbial mat during a 3 X 12 h, alternating light/ DM§ from the sediment to the atmosphere as diffu- darkllight period. files in the four compartments (Figs. 6 and 7). Dur- ing the iight period the pH increases due to photosynthesis. The decrease in pH in compartment 4 at -,--. I- 250 compartment 2 the start of a light period is caused by sulfate reduction. This decrease is overruled within 2 h by the large increase in pH (i.e. OH- production) by photo- sv-nfL--:;- n..:-- &L- -:- -,-.-LLG~~\ LJUL~LL~LLIC ~ll~ht,the pH in a!! !ayers of the mat drops to the initial value of -7.5, due to diffusion. The sulfide concentration increases from almost zero to 75 and 125 pmol 1-"n compartments 1 and 2, respectively, during the night. This can be explained by the 95% decrease in the specific growth rate of the colorIess sulfur bacteria during oxygen limitation (i.e. the dark period). In compartments 3 and 4, H2S is not oxidized. The profiles here are fully determined by production and diffusion of H2S. During the night, the concentrations of H2S will in- O'i' ' ' I crease in compartments 3 and 4, as the concentration 0 4 8 12 16 20 24 28 32 36 gradient (the driving force of the diffusion process) is Time (h) less than during the day. At the end of the dark Fig. 8. Simulation of DMS concentration in compartment 2. A: period, the sulfide production in compartment 4 de- DMS production is coupled to photosynthetic activity; B: with- creases, due to the depletion of organic carbon, out the inhibition effect of HzS on DMS oxidation; G: with this which leads to a decrease in the sulfide concentra- effect. 164 JMM de Zttart, J C KuenenIFEMS lbfrcroblology Ecologj~22 (1997) 155-165

sion is the only DMS-related process, taking place in of -4200 pol 1-l.The sulfide inhibition constant compartment 1. of the methyl mercaptan oxidase of H~yhornicrobium These simulations. based on experimental work EG was 90 vmol 1-I, and thus in the same range [11] with Methyloplzaga sulj5dovorans, show that sulfide as observed in the sediment and the cultures of M. profiles in the sediment are strongly related to sulfidovorans. The concentration of 40 pmol 1-I sul- DM§ emissions from the sediment. fide, found to fully inhibit DMS oxidation by M. sul$dovorans, was used in the simulations. In the 36 h simulation of the physical and biolo- 4. Discussion gical processes in a mat, the fact that there is less oxygen mailable at night (Fig. 5) results in an in-

The batch experiments with M. .su@dovorans show crease in H2S in all layers of the mat, with the effect Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 the sequential use of HzS and DMS indicating that being most pronounced in compartme~~ts1 and 2. above 100 pmoi 1-I H2S diauxy occurs. In continu- This is of interest, since the concentration rises ous culture, simultaneous oxidation of methanol, from almost zero to significant values (70-130 DMS and H2S occurs. This shows that at low con- pmol I-.') that may inhibit DMS oxidation. An im- centrations of sulfide and DMS, mixed substrate util- portant outcome of the modeling was that this inhib- ization is possible. However, a delayed increase in ition did indeed have a significant effect on the biomass concentration occurred. DMS and H2S predicted DMS emissions from the mat. If continu- were both completely oxidized to thiosulfate and ous DMS production during day and night is con- methanol was fully converted from the first day of sidered, the DMS concentration remains high if H2S the experiment. The biomass increase observed after inhibition of DMS oxidation is taken into account. six volume changes followed the theoretical growth The DMS emission from the mat, corresponding curve assuming that all cells present would be cap- with these 'high' DMS concentration, is 5 pmol able of obtaining energy from the oxidation of DMS m-2 h-l, whereas it is ~2 pmol m--2 hPi if this and H2S. This excluded the possibility of a strain inhibition is not taken into account. The inhibition mutation, since, in that case, the biomass increase of DMS oxidation found in crude enrichments oe- would have been less steep. The morphoiogy and curs at higher sulfide concentrations, but would still physiological characteristics of the cells (oxygen up- result in an increased DMS emission. take kinetics for DMS) remained the same. This phe- There is little published infomation on daytnight nomenon also occurred for mixed substrate contin- variations in sulfur emissions from mats that can be uous cultures on methanol and ISMS and/or H2S used for validation of the predictions made with this ..--.I-- 7-- -&..-.l,.-2 ---.:---- -&,.I ,I:+:--- mode;; It has iin L -- L LL-L rr 0 LLIIUC~It~r: SL~LILC.I~~IU G~~V~LVLLI~~~L~L~L L;U~UILIUIIS used. .__WGVGI, V~GI~ S~iiiw!~ LIM? -n-2a GILLLS- Apparently, this is due to uncoupling of energy sions from a Danish estuary were related to the day/ transduction and biomass growth. There is, as yet, night cycle with detectable emissions during the night no satisfactory explanation for this. The observation [18]. DMS emission varied strongly, with an average had no implications for the model described in this of 1 pmol m-' day-l, with peaks of 335 pmol m-2 paper, as DMS was consurned without a lag time. dayu"= 14 pin01 DMS mP2 bP'j, which is in the The relevant observation here is that DMS and H2S range of the predicted emissions. It was found that can be used simultaneously at low H2S concentra- DMS emission took place during the day and the tions. early night. There is currently no clear relationship In cultures of M. sulJidovorans, sullide oxidation between DMS production from DMSP and the diur- was not affected by the presence of DMS (Fig. 3). In nal cycle [19-211. On the basis of the measured [I81 contrast, DMS oxidation was inhibited by sulfide. and predicted emissions it might be hypothesized DMS oxidation was 100% inhibited at sulfide > 40 that DMS production will be coupled to photosyn- pmo1 I-'. Experiments with 10 different crude en- thetic activity and will, thus, be absent at night. This richments indicated that this inhibition is apparently would lead to the observed absence of DMS emis- common, although at different levels. On average, sions at night. 50% inhibition was found for H2S concentrations The model was developed to serve as an explana- J M IVI de Zwnrt, .I G Kuenen 1 FEMS ,%4!rrohroiogy Ecology 22 (1997) 155-165 165 tory tool For the observations made from a marine model for biological conversions of DMS in a microbial microbial mat. This work tries to bridge the gap mat; effect of pH on DMS fluxes. FEMS Microbiol. Ecoi. 18. 247-255. between physiology and modeling by collecting phys- [7] Kiene, R.P. and Capone, D.G. (1988) Microbial transforma- iological data of a representative of the aerobic tions of methylated sulfur compounds in anoxic salt marsh DMS-oxidizing bacteria and the implementation of sediments. Microbioi. Ecol. 15, 275-291. that in a mathematical model. By usiing this kind of [8] Visscher, P.T., Quist, P. and van Gemerden, H. (1991) Meth- general data, extrapolations may be made about ylated sulfur compounds in microbial mats: In situ concentrations and metabolism by colorless sz!fur bacterium. Appl. DMS emissions from a mat. In general, the model Environ. Microbiol. 57, 1758-1763. leads to a better insight into the relevant processes in [9] de Zwart, J.M.M., Ne!isse, P.N. and Kuenen. J.G. (1 996) an ecosystem and additionally may serve as a help%! Isolation and charac:crization of iVfethy1oi;haga su@doi;oii2ns, sp. nov.; an obligately methylotrophic, aerobic, dirnethylsul-

tool in formulating relevant (experimental) research Downloaded from https://academic.oup.com/femsec/article/22/2/155/541232 by guest on 27 September 2021 fide o.-'~:-.Aiuxing bacteriuini from a microbial mat. FEMS PPfiiro- qnestions. &fortunately, only a !imited set of data biol. Ecol. 20, 261 -270. (kinetics, physiology) is presently available, and the [IO] Suylen, G.M.H. (1988) Microbial Metabolism of Dimethylsul- assun~ptionsthat have to be made in the model are phide and Related Compounds. PhD Thesis, Technical Uni- considerable. The further development of this kind versity Delft. of model (for any ecosystem) will, therefore, be de- [ll] Suylen, G.M.H., Large, P.J., van Dijken, J.P. and Kuenen, pendent on the ample availability of kinetic and J.G. (1987) Methylmercaptan oxidase, a key enzyme in the metabolism of methylated sulfur compounds in Hyphornicro- physiological data of representatives of the different hium EG. J. Gen. Microbiol. 133, 2989-2997. functional groups in the ecosysiem. [12] de Bont, J.A.M., van Dijken, J.P. and Harder, W. (1981) Dimethyl sulfoxide and DMS as a carbon, sulphur and energy source for growth of Hyphomicrobium S.J. Gen. Microbiol. Acknowledgments 1227, 315-323. 1131 Visscher, P.T. and Taylor, B.F. (1993) A new mechanisms for the aerobic catabolism of dimethylsulfide. Appl. Env. Micro- This research was supported by the Dutch Foun- biol. 59, 3784-3789. dation for Scientific Research (NWO, NOP WA), [14] Sorbo, R. (1957) A colorimetric method for the determination Grant No. 91-163. We gratefully acknowledge L. de of thiosulfate. Biochim. Biophys. Acpa 23, 412416, Vries for technicai assistance and Dr. L.A. Robert- 1151 !+and, M.C., Greenberg, A.E. and Taras, M.J. (1985) Stand- ard ~ethodsfor the Examination of Water and Waste -water, son for helpful discussions and correction of the 16th edn. APHA, AWA, WPCF, New York. English text. 1161 Triiper, H.G. and Schtegel, H.G. (1964) Sulfur metabolism in 'Thiorhodaceae. I. Quantitative measurements on growing cells of Chromatium okenii. Antonie van Leeuwenhoek 30, 225-238. [17] Thatier, R.K., Sungerman, K. and Decker, K. (1977) Energy conservation in chemoirophic anaerobic bacteria. Bacterial. Rev. 41: 100-180. [I] van Gemerden, H. (1993) Microbial mats; A joint venture. 1181 Jorgenson, B.B. and Okholm-Hansen, B. (1985) Emissions of Marine Geol. 113, 3-25. biogenic sulfur gases from a Danish estuary. Atm. Env~ron. [2] Jorgenson, B.B. (1994) Sulfate reduction and thiosulfate trans- 19, 1737-1749. formation in a cyanobacterial mat during a die1 oxygen cycle. [19] Stal, L.J. (1995) Physiological ecology of cyanobacteria in FEMS Microb~ol.Ecol. 13, 303-312. microbial mats and other communities. New Phytol. 131, 1.- [3] Visscher, P.T. (1992) Microbial Sulfur Cycling in Laminated 32. Ecosystems. PhD Thesis, university of Groningen. [20] Stefels, J. and van Boekei, W.W.M. (1993) Product~onof [4] Visscher, P.T. and van Gemerden, H. (1991) Production and DMS from dissolved DMSP in axenic cultures of the marine consumption of dimethylsulfo propionate in marine microbial phytoplankton species Phaeocystis sp. Mar. Prog. Ser. 97, 11- mats. Appl. Environ. Microbiol. 57, 3237-3242. 18. [S] Keller, M.D., Bellows, W.K. and Guillard, R.R.L. (1989) Di- [21] Groene, T. (1995) Biogenic production and consumption of methylsulfide production in marine phytoplankton. In: Bio- dimethylsulfide (DMS) and dimethylsulfoniopropionate genic Sulfur in the Environment (Salzman, E.S. and Cooper, (DMSP) in the marine epipelagic zone: a review. J. Mar. W.J., Eds.), pp. 167-181. American Chemical Society (Sym- Syst. 6, 191-209. posium Series 393), Washington, DC. [6] de Zwart, J.M.M. and Kuenen, J.G (1995) Compartment