APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Apr. 1983, p. 1261-12700099-2240/83/041261-10$02.00/0Copyright C 1983, American Society for Microbiology

Vol. 45, No. 4

Colorless Sulfur Bacteria, Beggiatoa spp. and Thiovulum spp.,

in 02 and H2S MicrogradientsBO B. J0RGENSEN* AND NIELS P. REVSBECH

Institute of Ecology and Genetics, University ofAarhus, Ny Munkegade, DK-8000 Aarhus C, Denmark

Received 6 July 1982/Accepted 8 December 1982

The interactions between colorless sulfur bacteria and the chemical microgra-dients at the oxygen-sulfide interface were studied in Beggiatoa mats from marinesediments and in Thiovulum veils developing above the sediments. The gradientsof 02, H2S, and pH were measured by microelectrodes at depth increments of 50pm. An unstirred boundary layer in the water surrounding the mats and veilsprevented microturbulent or convective mixing of 02 and H2S. The two substratesreached the bacteria only by molecular diffusion through the boundary layer. Thebacteria lived as microaerophiles or anaerobes even under stirred, oxic water.Oxygen and sulfide zones overlapped by 50 ,um in the bacterial layers. Bothcompounds had concentrations in the range of 0 to 10 ,umol liter-' and residencetimes of 0.1 to 0.6 s in the overlapping zone. The sulfide oxidation was purelybiological. Diffusion calculations showed that formation of mats on solid sub-strates or of veils in the water represented optimal strategies for the bacteria toachieve a stable microenvironment, a high substrate supply, and an efficientcompetition with chemical sulfide oxidation. The continuous gliding movement ofBeggiatoa cells in mats or the flickering motion of Thiovulum cells in veils wereimportant for the availability of both 02 and H2S for the individual bacteria.

Beggiatoa and Thiovulum species are repre-sentatives of the gradient-type colorless sulfurbacteria which live in aquatic environments atthe transition between 02 and H2S. They areoften observed as isolated patches on organic-rich sediments along protected shores of lakesand of the sea. They may also cover the decay-ing remains of dead animals and plants fromwhich large amounts of H2S are produced (11,16). In the presence of oxygen, the bacteriaoxidize H2S to elemental sulfur, So, which accu-mulates as sulfur droplets inside the cell wall.The elemental sulfur can be further oxidized tosulfate by the bacteria (19, 24). Under anaerobicconditions, So may serve as an alternative elec-tron acceptor in Beggiatoa species and be re-duced back to H2S (19).When oxygen and sulfide coexist, they react

spontaneously with a half-life in the order of 1 h(1, 6, 8). The sulfur bacteria living at the 02-H2Sinterface must therefore always compete withthe autocatalytic oxidation of sulfide. Both Beg-giatoa and Thiovulum species have adapted tothis requirement by growing as sheets at thetransition between oxygen and sulfide. Beggia-toa cells, which are filamentous and glidingorganisms, form thin mats over solid substratesfrom which H2S is released (11, 16, 24). Thiovu-lum cells, which are very large, spherical bacte-ria with peritrichous flagella, form fragile veils

floating in the water, only loosely attached to theunderlying substrate (3, 17, 28). These typicalgrowth forms demonstrate a highly developedchemotactic behavior of the bacteria, whichenables them to adjust rapidly to the unstablechemical gradients.Although the general habitat of the colorless

sulfur bacteria has been known for a long time(15, 16), the microenvironmental factors whichdetermine growth are not well understood. BothBeggiatoa and Thiovulum species excrete slimeand form mucilagenous sheets (17, 24, 28). Thismay perhaps provide a special adaptation to thelife in dynamic gradients between two spontane-ously reacting substrates. How do these bacteriainteract with these gradients, and what is thechemical microenvironment immediately sur-rounding the bacterial cells? It has not previous-ly been possible to answer these questions dueto insufficient resolving power of the availablechemical techniques.

Microelectrodes were recently introduced inthe study of 02, H2S, and pH gradients inmicrobial communities (13, 22; unpublisheddata). The electrodes have dimensions whichallow chemical measurements to be made with aspatial resolution of 50 ,um or even less. Theyprovide an ideal tool for investigating the steepgradients in which colorless sulfur bacteria live.The microelectrodes were used in the present

1261

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1262 J0RGENSEN AND REVSBECH

study to analyze the oxygen and sulfide environ-ment of Beggiatoa and Thiovulum cells growingon mud cores from coastal sediments. Diffusionfluxes and reaction rates of 02 and H2S werecalculated from the results to explain the inter-actions between the gradient bacteria and thedynamic chemocline.

MATERIALS AND METHODSColorless sulfur bacteria were obtained from sedi-

ment cores collected in shallow water of Danish la-goons (Limfjorden, Aarhus Bay, Kal0 Lagoon) duringthe period July to October 1981. The sediments con-sisted of organic-rich muds, which at the time ofsampling were covered by a whitish film of Beggiatoaspp. The in situ temperature was 12 to 18°C, and thesalinity was in the range of 20 to 25o/oo. The coreswere kept in the laboratory at 20°C and submerged inslowly circulating, aerated seawater from the biotope.After a few days to a few weeks, the sediments werecovered by dense mats of Beggiatoa spp. In somecores, Thiovulum veils also had developed. The fol-lowing studies were made either on the whole, prein-cubated cores or on small samples of the sulfur bacte-ria collected directly from the enrichments.Measurements of oxygen, sulfide, and pH were

made by the use of microelectrodes (2, 25). Details ofthe application of these electrodes to the study ofmicrobial mats are described elsewhere (13; unpub-lished data). The oxygen electrodes had a sensing tipdiameter of 1 to 5 pm. The electrode current was readon a picoamperemeter with a calomel electrode as thereference. Linear calibration was made between zerooxygen measured below the bacterial mats and airsaturation maintained in the aerated water above themat. In microaquaria, where the water was stagnant,air saturation was read at the air-water interface. Thesilver-silver sulfide electrode had a sensing tip 25 ,umin diameter, and the electrode potential was read on amillivoltmeter with expanded scale. Calibration wasmade at a range of sulfide concentrations in seawaterbuffered at pH 7.0. The pH electrode had a diameter of50 ,um, and the length of the sensitive part was 100 to200 p.m. Calibration was made with standard pHbuffers in seawater at pH 7.0 and 9.2. The electrodeswere attached to a micromanipulator with which theycould be moved at defined increments of 10 to 100 ,um.Measurements of the microgradients in whole cores

were done under a dissection microscope. At 25 to 50times magnification, the exact position of the electrodetip could be determined relative to visible Beggiatoafilaments or Thiovulum cells. This was important forcorrelating the zones of sulfur bacteria with the chemi-cal gradients at a spatial accuracy of less than ±25 ,um.All measurements were made in semidarkness becauseof the presence of purple sulfur bacteria below theBeggiatoa mats. Even at normal room light thesephotosynthetic bacteria caused a detectable loweringof the sulfide concentration in the mat. When theelectrode position was found in the light, the coreswere kept dark for 10 to 15 min until a steady state wasreached and a profile could be measured.

Isolated Beggiatoa tufts or pieces of Thiovulum veilwere also studied in a microaquarium made from aglass slide and a cover slip and sealed on three sideswith petrolatum. The chemical gradients were mea-

sured while observing the bacteria in a light micro-scope at high magnification.Measurements of mat topography and spatial distri-

bution of cells were made with the micromanipulator,with the electrode as the pointer.

RESULTS

Beggiatoa mats. The vertical gradients of oxy-gen, sulfide, and pH were measured in a denseBeggiatoa mat covering the black mud of asediment core. The Beggiatoa filaments com-prised three size groups, having diameters of 1.5to 3, 5 to 6, and 15 to 30 ,um, respectively. Thesize groups were mixed within the mat structure,with the smallest filaments concentrated mainlyin the denser part of the mat and the larger onesprotruding in loops at the surface. The wholeBeggiatoa layer was 500 to 700 ,um thick. Dia-toms were present in patches in the uppermost 0to 100 ,um, and filamentous cyanobacteria (Os-cillatoria sp.) lay scattered on the surface. At300 to 700 ,um depth (i.e., below the 02-H2Sinterface), purple sulfur bacteria formed smallcolonies of oval cells.The three chemical gradients were measured

consecutively at exactly the same point in themat (Fig. 1). Oxygen was uniformly distributedin the aerated and slowly circulating seawaterabove the mat. Between 500 ,um (0.5 mm) abovethe mat and the mat surface the water wasunstirred, and the oxygen concentration sudden-ly dropped linearly from air saturation to nearzero. Oxygen penetrated through this unstirredlayer by molecular diffusion and reached 100 p.minto the bacterial mat. Hydrogen sulfide waspresent at high concentration, 500 to 1,000 ,umolliter-, in the black mud. In the uppermost 2mm, it diffused upward along a steep, lineargradient which ended just at the lower boundaryof oxygen. In the expanded view in Fig. 1, theoverlap between oxygen and hydrogen sulfide isseen to be only about 50 p.m. In this narrowzone, all the oxidation of the sulfide evidentlytook place. Due to heterotrophic processes inthe mud, the pH dropped steeply through theboundary layer from the normal value of 8.2 inthe seawater to 7.4 in the mud. There was a pHminimum within the Beggiatoa mat, probablydue to acidic products from the sulfide oxida-tion.

Chemical gradients, such as those presentedin Fig. 1, were measured at three differentpositions of the same mat and with very similarresults. The overlap between oxygen and sulfidewas less than 100 pum in all three profiles. Thedata shown in Fig. 1 represent only one set ofmeasurements, as average values would tend toblur the sharpness of the oxygen-sulfide inter-face. Judging from the shape of the 02 gradientand from the electrode sensitivity of 2 to 5 ,umol

APPL. ENVIRON. MICROBIOL.

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

SULFUR BACTERIA IN 02 AND H2S MICROGRADIENTS 1263

02 (yjmol literr-l)0 50 100 150 200 250 300

0 100 200 300 400 500 600

H2S ( pmol liter-l )

1i/I

//

I

02 (jPm[l liter1 )0 50 100 150 200 250 300

Beggictocmat .0.2

7//I, / ~~~~~~~~~0.4H2S

Mud -0.6"IM

X _k aQBIa

0 S0 10 300

H25 ( mol {i tW-l)

FIG. 1. Vertical distribution of 02, H2S, and pH through a Beggiatoa mat growing on a mud surface (A). Theinterface between 02 and H2S is shown in expanded scale (B). An unstirred boundary layer in the water at 0.0 to0.5 mm created a steep diffusion gradient of 02 from the circulating water above.

of 02 liter-1, it is unlikely that the oxygen fluxpenetrated below 100 ,um depth, allowing asignificant bacterial respiration here at an unde-tected trace concentration of 02.

It was an unexpected observation that thebacteria which were exposed to a current ofaerated seawater actually lived under anoxic ormicrooxic conditions. (The word "microoxic" issuggested here as a term for environments ofvery low oxygen concentration. The microoxicenvironment is the natural habitat of microaero-philic organisms.) This Was due to the unstirredboundary layer, which was therefore studied inmore detail in a sequence of oxygen micropro-files. Ten vertical profiles were measured in arow at 1-mm intervals (Fig. 2). The figure clearlyillustrates how the unstirred boundary layercovered the Beggiatoa mat as a 1-mm-thickblanket, which prevented the bacteria from be-ing exposed to the high oxygen concentrations inthe water. As in Fig. 1, oxygen penetrated only100 ,um into the mat where the zero oxygenisopleth was found. Small hollows into the matwere also unstirred as shown in Fig. 2. Beggia-toa cells, which were present in the hollows,grew in loose tufts, a structure which was dis-tinctly different from the smooth surface of thesurrounding mat.Thiovulum veils. Thiovulum cells developed

abundantly in some of the sediment cores stud-ied. Often the cells were swimming freely amongthe Beggiatoa filaments, but in a few cores,coherent veils formed floating just above thesediment surface. It was possible to measure theoxygen gradient just above the veils. Precisemeasurements of the sulfide gradient below theveil was more difficult because the veil tended toadhere to the electrode tip. A weak current ofaerated seawater flowed above the veils. Theywere very fragile but often redeveloped within afew hours after they had been disturbed. Smallpieces of a veil were transferred to a petri dishwith oxic seawater. Here the Thiovulum cellsrapidly formed spheres of veils with diametersfrom a few hundred micrometers to a few milli-meters. The smaller spheres floated freely in thewater, whereas the larger ones adhered to theglass wall. The oxygen gradient above one of thespheres was measured.The results are shown in Fig. 3. The sphere

was 2 mm in diameter and the cells were 8 to 16,um wide. The oxygen concentration was uni-form in the air-saturated water around thesphere. About 600 ,um from the surface of theThiovulum veil the oxygen concentration de-creased linearly and reached zero in the middleof the veil, which was about 100 pm thick. TheThiovulum sphere thus enclosed a small volume

pH7.0 7.2 7.4 7.6 7.8 8.0 8.2

If I I ----.

BWater

02

A

3 Water current

02 pH

2

3 Unstirred lor

\\\i\% %

, 'f3-4

2 - H-H2

0.6

0.6

0.4

0.2

mm m

VOL. 45, 1983

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1264 J0RGENSEN AND REVSBECH

FIG. 2. Vertical section through a 0.5-mm-thick Beggiatoa mat on a mud surface. Beggiatoa formed a tuftwithin the anoxic water in the central hollow. Isopleths of oxygen concentration expressed as the percentage ofair saturation show how the unstirred boundary layer covered the mat like a blanket and created microoxicconditions at the mat surface. The three insets show actual oxygen profiles from which isopleths wereconstructed. Arrows indicate the surface of the Beggiatoa mat.

of anoxic water. The anoxic conditions werecreated and maintained by the high oxygenconsumption of the veil. Since sulfide could notbe detected inside the sphere, the bacteria mayhave used internal storage material such as S ororganic reserves as substrate. The Thiovulumcells experienced only a small percentage of theoxygen concentration of the surrounding water.This was again due to the unstirred boundarylayer of water which surrounded the veil andthrough which oxygen was transported by diffu-sion.A similar situation was found in the sheets of

veil floating on the sediment surface (Fig. 3B).The boundary layer was only 400 pm thick dueto the water current above, and yet the Thiovu-lum cells were living as microaerophiles. Theveil very sharply separated the oxic and anoxicregions of the water column.Measurements of pH and H2S concentration

around the Thiovulum cells were made only inone case in which Thiovulum cells were swarm-ing freely in a 200-Rm-thick horizon within ahollow of a Beggiatoa mat (Fig. 4). The cellswere swimming rapidly around the depth where

the oxygen concentration just reached zero.About one-third of the cells were on the oxicside and two-thirds were on the anoxic side atany time. Sulfide was detectable only up to 200,um above the oxic-anoxic interface. The con-centration was about 30 ,umol liter-' just at theinterface and about 75 ,umol liter- at 200 ,umbelow. The H2S concentrations were, however,fluctuating, as there was no veil to stabilize thegradients, and the numbers are therefore notaccurate. The data do show that the swarmingThiovulum cells were positioned just at the tran-sition between oxygen and sulfide. The pHdecreased over the same depth intervals from7.62 to 7.40. This corresponded to a pH gradientof 0.5 pH unit per mm in the water.

Microaquaria. The diameter of the oxygenmicroelectrodes was a few micrometers at thesensing tip, i.e., of the same dimensions as thefilaments of Beggiatoa or Thiovulum cells. Itshould therefore be possible to study the con-centrations of oxygen immediately surroundingindividual sulfur bacteria, provided they can beobserved at sufficient magnification. This wouldallow more detailed investigations of their che-

...-, APPL. ENVIRON. MICROBIOL.

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

SULFUR BACTERIA IN 02 AND H2S MICROGRADIENTS 1265

mm 02( 6a0ir sot.)40 60 1

mmmm

FIG. 3. Spherical Thiovulum veil which in oxic seawater enclosed an anoxic pocket. The bacterial oxygenuptake created a steep diffusion gradient in the 0.6-mm-thick unstirred layer surrounding the veil (A). The oxygenprofile over a flat Thiovulum veil (sheet) is shown for comparison (B).

motactic behavior and of the interrelations be-tween the bacterial populations and their chemi-cal environment.A preliminary experiment of this type is re-

ported here. To study the bacteria at high magni-fication, a simple microaquarium was built andfixed on a microscope stage (Fig. 5). The aquari-um was filled with anoxic water from the Beg-giatoa mats, and a Beggiatoa tuft was placednear the open side. A calomel reference elec-trode was connected to the aquarium via a saltbridge. The oxygen microelectrode was attachedto a micromanipulator in a horizontal position sothat the electrode tip and bacteria could beobserved simultaneously at 100 to 400 timesmagnification.

Within 1 to 2 h, most Beggiatoa filaments hadoriented themselves in a zone 1.5 mm from theair-water interface through which oxygen dif-fused into the aquarium (Fig. 5). The filamentswere lying on the lower glass wall in a narrowband which was sharply bounded toward theoxic side. No Beggiatoa filaments occurred atoxygen concentrations above 10% air satura-tion. There was a less sharp boundary on theanoxic side where filaments lay scatteredthroughout the preparation. That observationindicated that the gliding Beggiatoa cells active-ly avoided higher oxygen concentrations. This isin accordance with their growth pattern on thesediments where a smooth Beggiatoa mat creat-ed microoxic or anoxic conditions for the wholepopulation (cf. Fig. 2).

Similar observations were made with Thiovu-

lum cells. A small piece of a veil was placed inthe anoxic water of the microaquarium, and thebacteria soon formed a new veil parallel to thewater surface at a distance of 1 to 2 mm.Initially, the cells swam around seemingly atrandom, but once they were established in theveil the amplitude of their movements perpen-dicular to the water surface did not exceed about100 ,um. The oxygen microelectrode showedthat they were positioned just around the bound-ary where oxygen reached zero. The individualcells in the veil moved rapidly within an oxygengradient from 0 to 1 to 3% air saturation.

DISCUSSIONThe present results have shown that Beggia-

toa and Thiovulum species live under very low02 and H2S concentrations of a few micromolesper liter, irrespective of the high concentrationsin the surrounding macroenvironment (abovethe 1-mm scale). The bacteria live as true mi-croaerophiles when they grow in mats or veils.They are then surrounded by unstirred boundarylayers which separate them and protect themfrom the surrounding high and fluctuating 02and H2S concentrations of the macroenviron-ment. Dissolved substrates reach the bacteriathrough these boundary layers by moleculardiffusion along steep gradients.The presence of unstirred boundary layers in

water adjacent to solid surfaces is a phenome-non well known from fluid dynamics (26). Theirimportance has not been fully appreciated inaquatic ecology or microbiology, mainly be-

VOL. 45, 1983

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1266 J0RGENSEN AND REVSBECH

02 ( cair sat.)

BEGGIATOA /

THIOVULUM iBLAC' 'C ~.MUD8LBLCK

----.- pH 7.52 -30 pM

H2S

H2S --

pH 7.40-75 pM H2S

I

1.0

0.5

0.0

mm

0.5

FIG. 4. Thiovulum cells swarming, without veil formation, at the oxygen-sulfide interface within a hollow in aBeggiatoa mat. Microelectrode measurements of 02, H2S, and pH were made along the broken line in the center.

cause measuring techniques have had insuffi-cient resolving power to demonstrate that theyexisted. The unstirred boundary layers are cre-ated by the viscous forces (internal friction) inthe water, which cause a thin water film to stickto the solid surface. The viscous forces in thisfilm dominate the inertial forces which drive theturbulent movement of the surrounding water.The thickness of the unstirred boundary layerdepends, among other things, on the roughnessof the solid surface and on the degree of turbu-lence of the stirred water.We will discuss how the gradient bacteria,

Beggiatoa and Thiovulum species, have adaptedto a life governed by diffusion through suchunstirred boundary layers.

Substrate diffusion and consumption. The mi-crogradients of oxygen and pH above the Beg-giatoa mat shown in Fig. 1 indicated a sharptransition between a stagnant boundary layeradjacent to the mat and the stirred bulk waterabove. The thickness of the boundary layer wasfound to vary with the rate of stirring. Thelinearity of the oxygen gradient at 0 to 0.5 mmindicated that there was a constant diffusioncoefficient and no oxygen consumption withinthe layer. The sulfide gradient was also linear at0.1 to 1.5 mm below the oxygen zone, whichshowed that it was governed by diffusion withlittle influence from production or consumptionof H2S in this layer. Thus, the whole consump-tion of oxygen must have taken place in theuppermost 100 j±m of the Beggiatoa mat, where-as sulfide was consumed in the zone from 50 to100 ,um depth where it met and coexisted withoxygen in a dynamic steady state.The transport of oxygen and sulfide by molec-

ular diffusion is a rapid process in dimensionsbelow 1 mm. The diffusion time of an oxygenmolecule through the unstirred layer can becalculated from the Einstein-Smoluchowski re-lation:

t = L2/2D = 1 min

where t is the diffusion time, L is the diffusionpath (500 ,um), and D is the diffusion coefficientof 02 at 20°C (2.06 x 10-5 cm2 S-1 [4]).

Since the microgradients show a pure diffu-sion transport of both 02 and H2S, it is possibleto calculate the flux and the consumption ratefor both substrates. The flux is calculated fromFick's first law:

F = D * dC/dx

where F is the flux, + is the porosity of thesystem (+ is 1.0 in the water phase and isestimated to be 0.9 in the Beggiatoa mat), D isthe diffusion coefficient at 20°C (2.06 x 10-5 cm2s-' for 02 [4], 1.52 x 10-5 cm2 s-1 for H2S-HS-as calculated from radiotracer measurements byJ0rgensen et al. [13] from a cyanobacterial mat,1.56 x 10-5 cm2 s-' for H2S gas alone [4]), anddC/dx is the vertical gradient of 02 or H2Sconcentration.As the thickness of the consumption layer is

known as well as the substrate concentrations,in this layer, it is possible to calculate theconsumption rate per volume as well as theturnover time (residence time) of the substratepool. As the substrate concentration has a gradi-ent within the layer, the calculated turnover timerepresents a mean value for the total pool. Themean turnover time is simply the total pool size

mm

0.0

mm

0.5

APPL. ENVIRON. MICROBIOL.

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

SULFUR BACTERIA IN 02 AND H2S MICROGRADIENTS

MICRO - AOUARIUM

02 (1. air sat.)40 60 80

FIG. 5. Illustration of Beggiatoa cells in a 1-mm-deep microaquarium with anoxic seawater. Oxygen diffused1.5 mm into the seawater from the ambient air and reached zero concfntration within the densest Beggiatoa zone(A). The filaments formed a band on the bottom glass slide. This created both vertical and horizontal gradientswithin the aquarium, and the oxygen gradient, therefore, appeared nonlinear (B).

within the consumption layer divided by the fluxinto the layer, under the reasonable assumptionthat the diffusion gradients are in steady state.Diffusion steady states in these dimensions areestablished within a few minutes. The results areshown in Table 1 together with similar calcula-tions for a Thiovulum veil based on data fromFig. 2 (sheet). The thickness of the 02 consump-tion zone in the Thiovulum veil was determinedfrom microscopic observations of the cells con-current with measurements of 02 microgra-dients. A thickness of 50 ,um was in accordancewith the lower curved part of the 02 gradient inFig. 3, which indicated consumption.The fluxes of oxygen and sulfide into the thin

Beg iatoa mat were 0.99 x 10-10 and 0.44 x10-0°mol cm-2 s-1, respectively. The part ofthe oxygen consumption which was due to sul-fide oxidation will depend on the oxidation prod-uct. An oxidation only to elemental sulfur wouldconsume 0.22 x 10-10 mol of 02 cm-2 S-1,whereas a complete oxidation to sulfate wouldconsume 0.88 x 10-10 mol of 02 cm-2 s-1. Theactual value must lie somewhere in between,depending on how fast the Beggiatoa cells accu-mulated elemental sulfur. The remaining oxygenconsumption could be used to oxidize othersubstrates such as dissolved organic molecules.Oxygen penetrated only 100 pum into the Beg-

giatoa mat before it was all consumed. In this 0to 100 p.m layer, the average 02 concentrationwas 6 ,umol liter-. Due to the small pool size,the turnover time was very short, only 0.6 s.This shows how extremely dynamic the chemi-cal microenvironment of the gradient bacteria

can be. Similar calculations for the 50-p.m-thicklayer of 02-H2S coexistence showed a turnovertime for the H2S pool of 0.6 s. Oxygen had aturnover time of only 0.1 s in the Thiovulum veil.The calculated fluxes of oxygen and sulfide

are in good agreement with directly measuredrates of oxygen uptake and of sulfide productionfrom sulfate reduction measured in organic-richcoastal sediments, similar to those on which theBeggiatoa and Thiovulum species were growing(12).

Beggiatoa mats. It was the early work ofWinogradsky (27) at the end of the last centuryon enrichment cultures of Beggiatoa cells whichled to the concept of chemoautotrophic metabo-lism in these organisms. Since then, the autotro-phic growth of Beggiatoa cells, as well as theirability to use H2S as an energy source, has beenquestioned repeatedly (5, 20). Recent studies onpure cultures ofBeggiatoa cells have shown thatmost strains grow heterotrophically (18, 20, 24),and in the one B. alba strain investigated, thetypical autotrophic C02-fixing system is absent(23). A report on chemolithotrophic hetero-trophy in chemostat-grown B. alba (10) was laterquestioned by Kuenen and Beudeker (1Sa).Such a metabolism would, however, agree wellwith a habitat at the oxic-anoxic interface insediments and water.Although the chemolithotrophy on H2S re-

mains questionable, the present study givesstrong indication that Beggiatoa cells growing inmats have the ability to oxidize all sulfide as-cending from the sediment very fast and effi-ciently.

VOL. 45, 1983 1267

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1268 J0RGENSEN AND REVSBECH

TABLE 1. Diffusion flux and consumption of 02 and H2S in a Beggiatoa mat and a Thiovulum veil

Flux (mol Uptake Uptake rate Avg concn TurnoverGenus Substrate cm-2(5-l) zone ((mmolliter- s-') AvgMon time (s)cms) Zfl lamllter'm) liter-') tms

Beggiatoa 02 0.99 x 1o-l 100 9.9 6 0.6H2S 0.44 x 10-10 50 8.8 5 0.6

Thiovulum 02 1.3 x 10-10 50 26 3 0.1

Bacteria which oxidize H2S with oxygen mustcompete with the spontaneous autooxidationwhich takes place concurrently. The chemicaloxidation of H2S in seawater and freshwater hasbeen studied in the laboratory as well as in thechemocline of stratified water bodies. The half-life of sulfide is generally in the range of 1 to 3 hat air saturation of oxygen and 20°C (1, 6, 8).Almgren and Hagstrom (1) found a half-life of 38to 81 min for sulfide at I to 5 ,umol liter-1, i.e., atthe same concentrations as in the O2-H2S inter-face of the Beggiatoa mat. The residete time ofH2S in the mat was only 0.6 s. Thus, the sulfideoxidation in the mat was 10,000- to 100,000-foldfaster than in water. Although chemical cata-lysts are able to speed up the chemical oxidation10- to 100-fold (7), the spontaneous reactionwould still be less than 1% of that in the Beggia-toa mat. It is therefore concluded that the proc-ess was biologically catalyzed. As Beggiatoaconstituted the completely dominating bacterialbiomass at the O2-H2S interface, it is suggestedthat this organism was chiefly responsible forthe rapid oxidation of H2S.

It was only 10% of the 500-,um-thick Beggia-toa mat which was within the 50-,um-thick 02-

H2S interface at any time. Since the filamentsglide around continuously at the speed of 100 to200 p.m min-' (9), they may at intervals havepassed through the interface zone and rapidlyaccreted sulfur droplets. These were then slowlyoxidized to sulfate above the sulfide zone orperhaps reduced back to H2S below the oxygenzone (19).

Thiovulum veils. The formation of Thiovulumveils serves the same purpose as the formationof Beggiatoa mats, namely to position the bacte-ria exactly at the interface between the oxic andthe sulfide zones and to stabilize the interface.The veils are sufficiently rigid to withstand slowmovements of the water, and they enable anexpansion of the anoxic zone into the water. Inthis way, Thiovulum veils can outcompete thesediment-bound Beggiatoa filaments, whichhave sometimes been observed to glide up onthe top side of Thiovulum veils.The establishment of a new veil was observed

under the microscope. Initially, swarming Thio-vulum cells aggregated exactly at the oxic-anox-ic interface. As the cells reached the interface,their rapid swimming changed into a vibrating

motion as they were seemingly caught in thedeveloping veil. The amplitude of their move-ments along the oxygen gradient was then onlyabout 100 ,um. An agglutination of slime threadssecreted from the aggregating cells was suggest-ed by Wirsen and Jannasch (28) as the mecha-nism of veil formation.The chemotactic response of Thiovulum cells

to traces of 02 must be very sensitive, as wasillustrated by the following observation. In amicroscope preparation of anoxic porewaterfrom a sulfuretum, a phyto-flagellate (Euglenasp.) was observed to swim around vividly, pur-sued by a flock of 5 to 10 Thiovulum cells. Thebacteria were evidently able to trace the oxygenwhich, like a comet's tail, was trailing after theflagellate. When the microscope light wasswitched off for a few seconds, the Thiovulumcells immediately lost track of the flagellate andswam around at random, but in the light againthey soon renewed the pursuit.

In the following, we will analyze the veil as amechanism by which Thiovulum cells can growin the free-water phase with an optimal utiliza-tion of sulfide and oxygen as substrates. Thiovu-lum cells may possibly utilize other reducedcompounds as well, but the ability to growautotrophically or mixotrophically at the ex-pense of sulfide as an energy source was at leaststrongly indicated by the studies of Wirsen andJannasch (28).Even the invisible, gelatinous Thiovulum veil

of 100 ,um thickness was able to create unstirredboundary layers (Fig. 3). A main function of theveil must be to serve as a solid interface betweenthe oxic and the sulfide zones of the water. Theveil prevents mixing of oxygen and sulfide bymicroturbulence or by convection.The diffusion time for an 02 molecule to go

through the 100-Rxm-thick veil is only 1 s, and yetoxygen reached only halfway through the veilbefore it was all consumed (Fig. 3). The overlapbetween the oxygen and sulfide zones musttherefore have been extremely narrow, onlyabout 50 i,m, and the residence time of thecompounds at the interface was correspondinglyshort, 0.1 s for oxygen. The sulfide turnovertime must be approximately the same, and therapid sulfide oxidation must therefore be biologi-cally catalyzed.The creation of a stable, microoxic environ-

APPL. ENVIRON. MICROBIOL.

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

SULFUR BACTERIA IN 02 AND H2S MICROGRADIENTS 1269

ment requires the concerted action of a largepopulation of Thiovulum cells growing in a veil.It could not be achieved by individual cellsswimming freely in the unstable environment.This may explain why Thiovulum cells oftenform minute spheres of veils in oxic water. Aslong as the oxygen consumption of the bacteriacan balance the oxygen diffusion through theboundary layer they can maintain anoxic condi-tions inside the sphere. Continuous measure-ments made in the sphere in Fig. 3 showed thatanoxic conditions were maintained for morethan 1 h. With time, however, individual cellsbegan to detach from the veil and swarm out intothe oxic water.

Diffusion limitation of Thiovulum cells. Carefulobservations on microgradients of oxygen inThiovulum veils showed that the cells on theinner side of the veil were in anoxic water. Forthese cells to reach oxygen, it was necessary tomove. Cell movements within the veil seemedrandom, with an amplitude of only 50 to 100 ,um.This, however, would be sufficient to bring cellsinto contact with both oxygen and sulfide withinseconds. The bacteria would not outrun diffu-sion by this movement since the diffusion time ofoxygen over the same distance is also 1 s. Theoxygen molecules, however, had a residencetime of only 0.1 s and, therefore, they did notdiffuse that far. A rapid, vibrating movement,therefore, could quite possibly help to bring thebacteria into contact with both oxygen and sul-fide.

Estimates of cell density in the veils weredone under a microscope by observing veilsforming in microaquaria as well as by observingintact veils under the dissection microscope. Anocular grid was used for the cell counts. Thedensities varied between 105 to 106 cells cm-2,with 5 x 105 as an average value (cf. reference11). The density corresponded approximately toone compact cell layer, which was evidentlyenough to completely consume the high influx ofoxygen.The low concentration and rapid turnover of

substrate in the veil may lead to diffusion limita-tion of the metabolic rate in individual cells. Toevaluate this possibility, it is necessary to knowthe rate of substrate uptake per cell. If weassume an average cell density of 5 x 105 percm2, an oxygen uptake of 1.3 x 10-10 mol of 02cm-2 s- as calculated in Table 1, and an H2Sflux of 0.44 x 10-10, similar to that reaching theBeggiatoa mat in the same sediment cores as theThiovulum veil was overgrowing, we obtain thefollowing rates: 02 uptake = 2.6 x 10-16 mol of02 cell-1 s-1; H2S uptake = 0.9 x 10-16 mol ofH2S cell-' s-1.The estimated rate of H2S uptake can be

compared with the autotrophic CO2 fixation

rates measured by Wirsen and Jannasch (28) inenrichments of Thiovulum cells. When growingunder optimal conditions in veils, the bacteriaassimilated up to 1.22 ,ug of C02 per 106 cells perh. With a complete oxidation of H2S via elemen-tal sulfur to sulfate and with an energy yield of10% comparable to that of the thiobacilli (R. F.Beudeker, J. C. Gottshal, and J. G. Kuenen,Antonie van Leeuwenhoek J. Microbiol. Serol.,in press), the rate of sulfide oxidation in theseThiovulum enrichments would have been 0.4 x10-16 mol of H2S cel-1 s-1. If the H2S wasoxidized to elemental sulfur only, the rate wouldbe 1.6 x 10-16 mol of H2S cel-1 s-1. Ourestimate above of 0.9 x 10-16 is between thesetwo rates.The substrate concentration at which a non-

moving, spherical cell with a radius, R, and asubstrate consumption rate, v, becomes diffu-sion limited is (14):

5= v

where S is the substrate concentration and D isthe diffusion constant of the substrate. With theabove calculated values for oxygen, v = 2.6 x10-16 mol of 02 cell-1 s-1, D = 2.06 x 10-s cm2s-1, and R = 6 x 10-4 cm, the limiting concen-tration for oxygen is 1.7 ,umol of 02 liter-1.Thus, at 1.7 ,umol of 02 liter7', these Thiovulumcells should deplete the surrounding watersphere of oxygen to an extent where the concen-tration just reaches zero at the cell surface. Atlower concentrations of oxygen in the surround-ing water, the oxygen uptake of the bacteria willdecrease because the availability of the oxygenis limited by the rate of molecular diffusionthrough the cells' own diffusion sphere. Thiovu-lum cells in the veil lived at oxygen concentra-tions between 0 and 5 to 10 ,umol liter-', i.e., ator slightly above diffusion limitation. A similarcalculation for the uptake of H2S shows that thecells should become diffusion limited at 0.8 ,umolof H2S liter-1. These calculations indicate thatthe Thiovulum cells not only create the steepmicrogradients through the stagnant boundarylayer, but they assimilate oxygen and sulfide soefficiently that even the individual cells comeclose to diffusion limitation in their uptake rate.This must be the highest efficiency a bacteriumof that size can reach in the competition with thechemical reaction between 02 and H2S.The substrate limitation can be reduced a little

because the cells swim. With a swimming rate,,u, of 50 ,um s-1, typical for bacteria, the sub-strate availability of the diffusion limited cell isincreased by a factor, G, which is (14):

G = 1 2 2p = 1.3-7w-D

VOL. 45, 1983

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from

1270 J0RGENSEN AND REVSBECH

With the same values of the parameters as thoseused above, the increase in oxygen availabilityfor the cell will be 30%. This is hardly enough tobe the important explanation for cell movementsin the veil. It can also be calculated that the cellsdo not gain more substrate by stirring the waterwithin the veil. At a stirring rate of 50 ,um s-1over the 100-pum-thick veil, the substrate trans-port by diffusion would still be faster than bystirring (cf. reference 21).The large size of Thiovulum, 10 to 20 p.m in

diameter, is unusual for bacteria in general, butit is typical for several colorless sulfur bacteriawhich have specialized to grow between theopposed gradients of oxygen and sulfide. Theneed for these bacteria to establish and maintaina physical barrier between the oxygen and sul-fide zones, which can withstand a certain degreeof water turbulence, could possibly be a factorwhich selects for this large cell size. Steepdiffusion gradients of the energy sources maythen provide a sufficient substrate availability toallow the large cell size without significant diffu-sion limitation of individual cells. Competitionfrom smaller bacteria with a more efficient sub-strate uptake may, e.g., in Thiovulum veils,force these to abandon the old slime web atintervals and establish a new one (27).

In conclusion, Beggiatoa and Thiovulum spe-cies represent optimal adaptations to life at the02-H2S interface. Further studies of the physiol-ogy and chemotaxis of these bacteria should bemade in pure cultures in combination with mi-croelectrode measurements of the gradients tobetter understand the details of this adaptation.

ACKNOWLEDGMENTWe thank J. Gijs Kuenen for critically reading the manu-

script.

LITERATURE CITED1. Ahngren, T., and I. Hagstrom. 1974. The oxidation rate of

sulphide in sea water. Water Res. 8:395-400.2. Baumgartl, H., and D. W. Lubbers. 1973. Platinum needle

electrodes for polarographic measurements of oxygen andhydrogen, p. 130-136. In M. Kessler et al. (ed.), Oxygensupply. Urban & Schwarzenberg, Munich.

3. Boer, W. E., J. W. M. laRiviere, and A. L. Houwink.1961. Observations on the morphology of Thiovulummajus Hinze. Antonie van Leeuwenhoek J. Microbiol.Serol. 27:447-460.

4. Broecker, W. S., and T. H. Peng. 1974. Gas exchangerates between air and sea. Tellus 26:21-35.

5. Burton, S. D., and R. Y. Morita. 1964. Effects of catalaseand culture conditions on growth of Beggiatoa. J. Bacteri-ol. 88:1755-1761.

6. Chen, K. Y., and J. C. Morris. 1972. Kinetics of oxidationof aqueous sulfide by 02. Environ. Sci. Technol. 6:529-537.

7. Chen, K. Y., and J. C. Morris. 1972. Oxidation of sulfideby 02-catalysis and inhibition. J. Sanit. Eng. Div. Proc.Am. Soc. Civ. Eng. 98:215-227.

8. Cline, J. D., and F. A. Richards. 1969. Oxygenation ofhydrogen sulfide in seawater at constant salinity, tempera-ture, and pH. Environ. Sci. Technol. 3:838-843.

9. Crozier, W. J., and T. J. B. Stier. 1926. Temperaturecharacteristics for speed of movement of thiobacteria. J.Gen. Physiol. 10:185-193.

10. Gude, H., W. R. Strohl, and J. M. Larkin. 1981. Mixotro-phic and heterotrophic growth of Beggiatoa alba in con-tinuous culture. Arch. Microbiol. 129:357-360.

11. J0rgensen, B. B. 1977. Distribution of colorless sulfurbacteria (Beggiatoa spp.) in a coastal marine sediment.Mar. Biol. (N.Y.) 41:19-28.

12. J0rgensen, B. B. 1982. Mineralization of organic matter inthe sea bed-the role of sulphate reduction. Nature (Lon-don) 296:643-645.

13. Jorgensen, B. B., N. P. Revsbech, T. H. Blackburn, andY. Cohen. 1979. Diurnal cycle of oxygen and sulfidemicrogradients and microbial photosynthesis in a cyano-bacterial mat sediment. Appi. Environ. Microbiol. 38:46-58.

14. Koch, A. L. 1971. The adaptive responses of Escherichiacoli to a feast and famine existence, p. 147-217. In A. H.Rose and J. F. Wilkinson (ed.), Advances in microbialphysiology, vol. 6. Academic Press, Inc., London.

15. Kuenen, J. G. 1975. Colourless sulfur bacteria and theirrole in the sulfur cycle. Plant Soil 43:49-76.

lSa.Kuenen, J. G., and R. F. Beudeker. 1982. Microbiology oftheobacilli and other sulfur-oxidizing autotrophs, mixo-trophs and heterotrophs. Phil. Trans. R. Soc. London B298:473-497.

16. Lackey, J. B., E. W. Lackey, and G. B. Morgan. 1965.Taxonomy and ecology of the sulfur bacteria, p. 1-23.University of Florida, College of Engineering, Bull. Ser.119, vol. 19. University of Florida Press, Gainesville, Fla.

17. laRiviere, J. W. M. 1963. Cultivation and properties ofThiovulum majus Hinze, p. 61-72. In C. H. Oppenheimer(ed.), Symposium on marine microbiology. Charles CThomas, Publisher, Springfield, Ill.

18. Nelson, D. C., and R. W. Castenholz. 1981. Organic nutri-tion of Beggiatoa sp. J. Bacteriol. 147:236-247.

19. Nelson, D. C., and R. W. Castenholz. 1981. Use of re-duced sulfur compounds by Beggiatoa sp. J. Bacteriol.147:140-154.

20. Pringshehm, E. G. 1967. Die Mixotrophie von Beggiatoa.Arch. Mikrobiol. 59:247-254.

21. Purcell, E. M. 1977. Life at low Reynolds numbers. Am.J. Phys. 45:3-11.

22. Revsbech, N. P., B. B. Jorgensen, and T. H. Blackburn.1980. Oxygen in the sea bottom measured with a micro-electrode. Science 207:1355-1356.

23. Strohl, W. R., G. C. Cannon, J. M. Shively, H. Gude,L. A. Hook, C. M. Lane, and J. H. Larkin. 1981. Hetero-trophic carbon metabolism by Beggiatoa alba. J. Bacteri-ol. 148:572-583.

24. Strohl, W. R., and J. M. Larkin. 1978. Enumeration,isolation, and characterization of Beggiatoa from fresh-water sediments. Appl. Environ. Microbiol. 36:755-770.

25. Thomas, R. C. 1978. Ion-sensitive intracellular microelec-trodes, how to make and use them. Academic Press, Inc.,London.

26. White, F. M. 1974. Viscous fluid flow. McGraw-Hill BookCo., New York.

27. Winogradsky, S. 1888. Beitrage zur Morphologie undPhysiologie der Bacterien. I. Zur Morphologie und Phy-siologie der Schwefelbacterien. Arthur Felix, Leipzig.

28. Whrsen, C. O., and H. W. Jannasch. 1978. Physiologicaland morphological observations on Thiovulum sp. J. Bac-teriol. 136:765-774.

APPL. ENVIRON. MICROBIOL.

on February 1, 2020 by guest

http://aem.asm

.org/D

ownloaded from