Vol. 59, No. 11APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Nov. 1993, p. 3803-38150099-2240/93/113803-13$02.00/0Copyright © 1993, American Society for Microbiology

Heterogeneous Distribution of Microbial Activity inMethanogenic Aggregates: pH and Glucose MicroprofilesPIET N. L. LENS,1'2 DIRK DE BEER,3 CAREL C. H. CRONENBERG,3 FRANS P. HOUWEN,l

SIMON P. P. OTFL'ENGRAF,3 AND WILLY H. VERSTRAETEl 2*

Laboratory of Microbial Ecology' and Centre for Environmental Sanitation,2 University of Gent, Coupure L.653, 9000 Gent, Belgium, and Laboratory for Chemical Technology, University ofAmsterdam,

Nieuwe Achtergracht 166, 1018 WVAmsterdam, The Netherlands3

Received 4 March 1993/Accepted 3 September 1993

Methanogenic aggregates, harvested from an upflow anaerobic sludge blanket reactor treating potato starchwastewater, were acclimatized to either glucose or a mixture of sugars and organic nitrogen compounds (i.e.,diluted molasses). Both types of granules exhibited internal pH and substrate concentration gradients inmineral medium (pH 7.0, 30°C) as was measured with microelectrodes. Glucose-acclimatized granulessuspended in a mineral medium lacking glucose exhibited a distinct internal pH decrease of about 1 U withinthe granule, suggesting strong metabolism by the acidogenic bacteria. Molasses-acclimatized and aged granulessuspended in mineral medium did not exhibit such a pH decrease, suggesting the importance of the metabolicstate of these acidogens. The pH gradient did not occur in deactivated granules and was not observable instrongly buffered media (mineral medium containing 33 mM phosphate or reactor liquid). When glucose (0.5to 5.0 mM) was added to the mineral medium, granules exhibited a convex pH profile. Glucose consumptionwas located exclusively in the outer 200 to 300 .m of the aggregates (mean diameter = 1.5 mm). The additionof 20 mM 2-bromoethanesulfonic acid to the mineral medium indicated that the higher pH levels in the centreof the granule appeared to be related to the activity of methanogens. It is suggested that acidogenic activityoccurs predominantly in the outer 200 to 300 ,um of the aggregate and methanogenic activity occurs

predominantly in the center of the investigated granules.

The microbial community involved in anaerobic digestionprocesses in natural habitats as well as in manufactureddigestion systems is known to be quite complex, comprisinghydrolytic, fermentative, acidogenic, and methanogenic bac-teria. Complete anaerobic degradation of organic matter tomethane (CH4) and carbon dioxide (CO2) requires the con-certed action of these different microbial species (4, 13, 63).Anaerobic wastewater treatment relies on the control ofthese microbial interactions by applying appropriate reactorconditions and reactor configurations (22, 40). The upflowanaerobic sludge blanket (UASB) reactor concept (29, 30)uncouples solid and hydraulic retention times by biomassaccumulation in compact, well-settling aggregates. Thisgranular configuration has several engineering advantagessuch as a maximal microorganism-to-space ratio and facili-tated cell separation (15).UASB granule formation is mediated by autoimmobiliza-

tion processes, though factors governing granulation are stillunclear. It has been shown that inorganic nuclei (e.g., clayminerals) and high levels of divalent cations (e.g., Ca2+ andMg2+) can initiate granule formation (24, 50). Besides,microbial factors also can mediate granule formation, e.g.,extracellular polymer formation (48), the presence of Meth-anothrix sp. microcolonies as microbial nuclei (15), and theinteractions between methanogens and syntrophic acetogens(56).An important aspect in the study of granulation is the

ultrastructure of the granules. Light and electron micros-copy, sometimes in combination with histological and immu-nofluorescent techniques, have been widely applied (Table

* Corresponding author. Electronic mail address: LME@FLAND.RUG.AC.BE.

1). In addition, measurements of substrate removal rates ofintact and disrupted granules, as well as thermodynamiccalculations, have been used to study granular ultra-structures. These techniques gave evidence for both a ho-mogeneous and a layered distribution of individual cells ormicrocolonies in granules grown under different reactorconditions.Although the techniques mentioned above are useful to

visualize population distributions, they do not measure insitu activities within granules. Recently, pH microsensorswere introduced to study acetate conversion kinetics inmethanogenic granules (8, 9). The conversion of acetate(pKa = 4.75 at 25°C) into the weaker carbonic acid (pKa =

6.36 at 25°C) and methane is accompanied by a pH increase.This pH increase can be measured at micrometer level bymeans of microelectrodes. In this study, pH and glucosemicroelectrodes were used to measure in situ activities ofboth acidogenic and methanogenic bacteria in UASB aggre-gates. The application of such microprofiles to map thespatial distribution of these populations within the granulesis also discussed.

MATERIALS AND METHODS

Source of UASB aggregates and acclimatization procedure.Granular methanogenic sludge was obtained from a full-scaleUASB reactor (58) operated at S.A. Van Den Broeke foodfactory (Leuze, Belgium). Granules were harvested at asampling point 2 m above the bottom of the reactor andacclimatized during a short period (10 days) to two differentgrowth media containing only soluble substrates. The firstgrowth medium consisted of a glucose solution with noadditional nitrogen source. Acclimatization was performedin a 0.25-liter vessel by a fed-batch feeding procedure at 33°C

3803

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

TABLE 1. Reported ultrastructures of UASB granules grown under different conditions

Aggregate structure Analytical technique(s) Wastewater/substrate(s) Temp (0C)0 Reference

Homogeneous distributionNo internal organization SEM/TEMb Sugar refinery 32 14No internal organization SEM Sugar beet/volatile fatty acids 35 1Propionate-grown granules consisted TEM/Immunocytochemical Sugar refinery/propionate/ethanol 35 18

of microcolonies; no specific dis- methodstribution in ethanol-grown gran-ules

Structured distributionDifferent morphotypes in periphery SEM/TEM Starch+sucrose 25 21and core

Different morphotypes in periphery SEM Sugar beet/maize starch 35 1and core

Three morphologically distinct lay- SEM/TEM Sucrose 35 37ers

Three morphologically distinct lay- TEM/Light microscopy Acetate 60 3ers

Thermodynamic evidence for tro- Thermodynamic calcula- Saccharose/sucrose 35 43phic microniches tions of propionate deg-

radationMetabolic activity is partitioned in Activity tests on abrased Sucrose 35 19

separate layers granulesBasic texture of core and peripheral Histochemical and immu- Acetate+propionate+butyrate 38, 55 59

zone; methanogens grown in clus- nohistochemical sec-ters, layers, and lawns tions, microscopy

Thermophilic conditions promote Histochemical and immu- Acetate+propionate+butyrate 38, 55 36cluster formation of methanogenic nohistochemical sec-subpopulations tions, microscopy

Compartmentalization of acidogenic pH and glucose microsen- Glucose/molasses 30 This study(outer 200-300 p.m) and methano- sorsgenic (center) activity

a Temperature at which granules were grown.I SEM, scanning electron microscopy; TEM, transmission electron microscopy.

with a volumetric loading rate of about 0.25 g of chemicaloxygen demand (COD) per g of volatile suspended solids(VSS) per day. Besides fresh granules, granules that hadbeen stored anaerobically for 5 months at 10°C were alsoacclimatized to glucose (referred to as aged granules). Thesecond growth medium was diluted distillery molasses, afull-growth medium with a C/N ratio (wt/wt) of 2.5, contain-ing 1 hg of COD per m3 of which about 50% was sugars, andall essential minerals. Acclimatization to molasses was car-ried out as with glucose, except that a laboratory UASBreactor (2.5 liters) with an upstream velocity of 1 m/h wasused.

Control experiments were performed in aggregates deac-tivated by exposure either to a N2-purged medium (seebelow) supplemented with 2% glutaraldehyde (5 days), 8%formaldehyde (5 days), or to 0.2% HgCl2 (3 days). 2-Bro-moethanesulfonic acid (BES) was used to inhibit selectivelymethane-producing bacteria. Fresh glucose-acclimatizedgranules were preincubated (3 days) in medium containing 10mM BES. The same concentration of BES was also presentduring microprofile measurements. Methanogenic activity inaged glucose-acclimatized granules was inhibited by 50 mMBES during the measurement, omitting preincubation. Afteraddition of the inhibitors, the pH of the medium wasadjusted to pH 7.0.

Microelectrode preparation and microprofile measure-ments. Microprofiles of pH were measured by using liquidion-exchange microelectrodes with a tip diameter of 1 ,um(9). For the liquid membrane, Hydrogen Ionophore II(Fluka, Switzerland) was used. The preparation of oxygen-and pH-independent enzyme microelectrodes for glucose

with tip sizes of less than 10 ,um has been describedpreviously (5-7).The medium used in the microprofile measurements (8)

contained 60 g of NaH2PO4 per m3 (0.5 mM) unless specifiedotherwise. It was prepared anaerobically by purging withnitrogen gas. Aggregates were preincubated statically over-night at 30°C in substrate-free medium. Before measure-ments were started, aggregates were equilibrated in themeasuring cell for at least 1 h to obtain steady-state profiles.Steady-state conditions were verified by measuring micro-profiles within short time intervals. For each granule, at leastduplicate (penetration and withdrawal) microprofiles wererecorded at two different locations. Measurements wereperformed at 30 +- PC.

Modelling. A mathematical model is proposed for theprediction of glucose and pH microprofiles in the aggregates.In this model, the conversion of glucose to methane wassimplified to two sequential reactions:

VC6H1206 -1-- 1.8 CH3COOH + 0.9 CH4 +

0.9 CO2 + 0.1 [C6H1206] biomass

CH3COOH -k CO2 + CH4

(1)

(2)where V1 and V2 are the reaction rates of the glucoseconversion and the acetate degradation, respectively. Forreaction 1, acetate production from glucose was coupledwith hydrogenotrophic methanogenesis (either directly fromglucose or through propionate oxidation). A biomass yield of0.1 on glucose was assumed (64). The mass balances for

3804 LENS ET AL.

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

pH AND GLUCOSE MICROPROFILES IN UASB GRANULES 3805

glucose, acetate, and CO2 (dissolved as bicarbonate) weredescribed by the following equations, respectively:

Glucose: (Deifr2)d(r2 dC1/dr)Idr = VIAcetate: (De2fr2)d(r2 dC2/dr)/dr = V2 - 1.8 VIC02: (De3/r2)d(r2 dC3Idr)Idr = -0.9 VI - V2

where Dei are the effective diffusion coefficients (squaremeters per second) inside the aggregate and Ci are thesubstrate concentrations (in moles per cubic meter). Itshould be noted that the Dei were assumed independent ofthe ionization state of the compound (33). The substrateconsumption rates V1 and V2 (in moles per cubic meter persecond) of glucose and acetate degradation, respectively,were assumed to be of mixed order according to the follow-ing equations:

VI = Vm. CI/(Ksl + Cl)V2 = Vm2 C21(Ks2 + C2)

with Vm, and Vm2 the respective maximum conversion rates(in moles per cubic meter per second) and Kj1 and K,2 thesaturation constants (in moles per cubic meter). To evaluatethe effect of the spatial distribution of glucose-degradingactivity, reaction 1 was assumed to be restricted to theperiphery (Ro < r <R) of the aggregate. For positions insidethe aggregate smaller than Ro, no glucose consumption isassumed:

VI = V,, C,/(K,+ C,)

V,= 0

Ro<r <R0 <Ro

If Cli is the glucose interface concentration, then the bound-ary conditions are as follows:

r = R; C, = C,i;r = RO; dC,/dr = 0

The resulting set of equations cannot be solved analytically;therefore, a numerical iteration procedure was used.

(i) Numerical iteration procedure. The iterations werestarted with a pH of 7.0 assumed throughout the aggregate.Then, glucose, acetate and bicarbonate profiles were calcu-lated using Euler's method (44), dividing the aggregate into32 segments. Simultaneously, pH profiles were calculated bysolving the alkalinity balance for each segment, taking alsothe buffer ions present in the medium into account (33). Thefollowing equilibrium constants of the buffering ions wereused i61): H3P04, 6.75 x 10-3; H2PO4-, 6.47 x 10-8;HP04 -, 2.41 x 10-13; H2C03, 4.71 x 10-7; HC03-, 5.13 x10-11; acetic acid, 1.76 x 10-5; and H20, 1.35 x 10-14. Thecalculated local pH values were used to adjust the local V1and V2, and a new iteration cycle was started. The procedurewas repeated until steady state was achieved.

(ii) Parameter definition. The values of the kinetic param-eters K1 and K2 were obtained from the literature (see Table2), and Vmi and Vm2 were determined experimentally (seebelow). The pH dependence of Vmi and Vm2 was obtained byan empirical fit of experimental data in combination withdata from literature (24, 55). The effective diffusion coeffi-cient (De) of glucose was also determined in this study (seebelow). The De of all other reactants was assumed to be 65%of their molecular diffusion coefficient, in accordance withthe experimentally determined values of glucose.

(iii) Determination of Vmi and Vm2. Sludge samples (0.5 to1.0 g of dry matter) were centrifuged (10 min at 10.000 x g),washed twice in oxygen-free phosphate buffer (100 mM, pH

7.2), and transferred anaerobically to 200-ml vials containing150 ml of phosphate buffer (25 mM) supplemented with 20mM BES. After 20 min, tests were initiated by the additionof glucose to a final concentration of 15 mM. Duplicate vialswere stirred at 30°C in the dark. The pH of the buffer wasmaintained by addition of 0.1 M NaOH by means of anautomatic titrator. Acid production rates from glucose weredetermined as described by Lievense et al. (32). The pHdependence of the glucose conversion was determined in 25mM phosphate buffers. At the end of the test, liquid sampleswere analyzed for volatile fatty acids and lactic acid.Methane production rates from acetate were measured in

triplicate at 30°C as described elsewhere (28).(iv) Determination of the effective diffusion coefficient for

glucose. Selected granules with a diameter of 1.58 (+0.04)mm were deactivated by either heating to 80°C for 10 min,shaking in pure chloroform for 10 min, or exposure for 1 h to0.2% (wt/vol) HgCl2 or to 3% (wt/vol) glutaraldehyde. Afterbeing rinsed, the deactivated aggregates were mounted in ameasuring cell containing 100 mM phosphate buffer, and aglucose microelectrode was positioned in the center of theaggregates. After preincubation for at least 30 min at 30°C,glucose was added to the well-mixed solution to give a finalconcentration of 1 mM. The evolution of the glucose con-centration in the center of the aggregates was followed overtime, and De was calculated as described previously (5). Forthese measurements, the oxygen-independent glucose mi-croelectrode could not be used because of its long responsetime (1 min). Therefore, a simple glucose oxidase-coatedmicroelectrode (5) was used. This type of electrode can onlybe applied under aerobic conditions and has a 90% responsetime of less than 1 s.Other methods, chemicals, and gases. Scanning electron

microscopy (SEM) was done as described by Vandenabeeleet al. (54). Lactate was extracted, derivatized, and analyzedas described by Littmann et al. (34). Extraction (23) andanalysis (57) of volatile fatty acids was done as describedelsewhere. Methane was analyzed according to the methodof Van Meenen et al. (57).

All chemicals used were analytical grade and were ob-tained from commercial sources. Nitrogen gas was obtainedfrom Air Liquide Antwerpen (Belgium) and Hoekloos (Am-sterdam, The Netherlands).

RESULTS

General characteristics of the aggregates. The diameter ofthe investigated aggregates ranged from 0.5 to 2.0 mm, theaverage being 1.5 mm. Glucose-acclimatized granules con-verted glucose at pH 7.0 (in the presence of BES) at a rate of12.5 mmol of glucose per g of VSS per day. Propionate andacetate, in the ratio 3:1, were the main products; no lacticacid was measured. Glucose conversion rates at pH 6.0 and5.5 were 67 and 29% of the rate at pH 7.0, respectively.Methane was produced from acetate at a rate of 4.5 mmol ofCH4 per g of VSS per day at pH 7.0. At pH 6.0 and 5.5 theserates were, respectively, 140 and 110% of the rate at pH 7.0.

Determination of the effective diffusion coefficient (De) forglucose was possible with glucose-acclimatized granulesdeactivated by either of the four procedures used. Noglucose gradients were observed under steady-state condi-tions in the absence of glucose in the bulk liquid. When 1mM glucose was added to the medium, reproducible tran-sient state responses were measured (data not shown) in allaggregates investigated. De appeared to be influenced by themethod of deactivation and ranged from 5.1 x 10-10 m2/s

VOL. 59, 1993

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

8.0

7.0

Q

6.0

5.0-400 -200 0 200 400 600 800

Depth (Am)

FIG. 1. Steady-state pH microprofiles measured in glucose-acclimatized UASB aggregates in medium (0.5 mM NaH2PO4) withoutsubstrate. 0, deactivated aggregates; L, fresh aggregates; V, aged aggregates; V, aged aggregates + 50 mM BES.

(chloroform treatment), 4.6 x 10-10 m2/s (heat treatment),and 3.9 x 10-10 m2/s (glutaraldehyde treatment) to 3.8 x10-10 m2/s (HgCl2 treatment).pH microprofiles. Fresh glucose-acclimatized aggregates

showed a pH drop of 0.8 to 1.5 U corresponding to a protongradient of 7.7 x 10-3 mol/m4 when exposed to mediumwithout substrate (Fig. 1). Since no pH microgradients wereobserved in deactivated aggregates, active cellular metabo-lism appeared responsible for the generation of protons inthe granule. No volatile fatty acids or lactate could bedetected in disintegrated glucose-acclimatized granules inoxygen-free phosphate buffer (50 mM) without glucose.Aged granules, on the other hand, showed a slight internalpH increase, which did not occur when methanogenic activ-ity was inhibited by BES (Fig. 1).

In the presence of acetate, smaller pH decreases weremeasured in the aggregates (Fig. 2). Almost identical pHprofiles were observed whether 5 or 10 mM acetate waspresent in the bulk liquid. The internal pH decrease of about0.4 U was accompanied with a proton gradient of 6.3 x 10-4molIm4. Using the model, it was calculated that the smallerpH decreases could not be attributed to the increased buffercapacity of the medium upon acetate addition only (data notshown). The same phenomenon was observed when propi-onate was added to the medium (data not shown). Theseresults indicated that acetoclastic methanogenic activitycompensates (partly) for the pH decrease due to endogenousactivity of fermentatives. Therefore, the measured pH mi-croprofiles represent the sum of these activities.

Effect of glucose concentration and buffer capacity. Figure 3shows glucose microprofiles in glucose-acclimatized aggre-gates as a function of glucose concentration and the buffer

capacity of the bulk liquid. The presence of a diffusiveboundary layer of about 100 ,um reduced the glucose con-centrations at the interface of the aggregate. In all cases,glucose conversion occurred almost exclusively in the outer200 to 300 ,um of the aggregates. With a low buffer capacityin the medium (0.5 mM NaH2PO4) and a glucose concentra-tion of 2.5 mM (Fig. 3A), an interfacial glucose gradient of2.8 x 103 mol/m4 was measured. No glucose gradient wasobserved in the core of these granules. Increasing the buffercapacity to 10 mM NaH2PO4 in medium with 2.5 mMglucose resulted in a steeper glucose gradient (5.7 x 103mol/m4) at the interface (Fig. 3B). The corresponding glu-cose fluxes over the interface were 1.9 x 106 mol/m2/s and3.4 x 10-6 mol/m2/s in aggregates in medium containing 0.5and 10.0 mM NaH2PO4, respectively. An interfacial glucoseflux identical to that with 2.5 mM glucose was observedwhen 5.0 mM glucose was applied to well-buffered medium(Fig. 3B). At a low glucose concentration (0.5 mM), allglucose was consumed in the outer 150 to 200 ,um indepen-dent of the buffer capacity of the medium (Fig. 3).The pH microprofiles in weakly buffered medium showed

a typical convex shape (Fig. 4A). The pH dropped to about5.5 inside the aggregate and in some aggregates values as lowas 5.2 were recorded in the presence of 5 mM glucose,though the pH of the bulk liquid was 7.0. The minima in thepH profiles, around 150 and 250 ,um from the granule surfacefor 0.5 and 2.5 mM glucose, respectively (Fig. 4A), coin-cided with the depths at which the glucose gradients becamezero (Fig. 3A). BES-inhibited aggregates in medium with 2.5mM glucose did not show the typical pH increase in thedeeper layers of the aggregates (Fig. 4B). Therefore, meth-anogens must contribute to this pH increase. Aggregates

3806 LENS ET AL.

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

pH AND GLUCOSE MICROPROFILES IN UASB GRANULES

7

Q

6

5-400 -200 0 200 400 600 800

Depth (,um)Steady-state pH microprofiles in glucose-acclimatized UASB aggregates in medium containing 0 (L), 1 (-), 5 (V), and 10 (*) mM

further showed a diffusive boundary layer of 100 to 200 ,umfor protons. No pH microgradients were observed when thebuffer capacity of the medium was increased to 33 mMNaH2PO4 at pH 7.0 in the presence of 5 mM glucose (datanot shown).

Microprofile simulations. Figure 5 shows calculated andmeasured pH and glucose microprofiles as a function of thespatial distribution of the acidogenic populations. If fermen-tative activity was assigned to be located exclusively in theouter 300 ,um of the aggregate, the modelled glucose micro-profiles agreed well with those experimentally obtained (Fig.5A). The latter assumption gave also a qualitatively good fitfor modelled pH microprofiles (Fig. SB), though there was adistinct discrepancy between measured and calculated pHprofiles in the periphery of the aggregate. In contrast, whenglucose-converting activity was assumed to occur through-out the aggregate, calculated glucose concentrations in thecenter of the aggregate were substantially lower than themeasured values (Fig. SA). Also, the pH microprofiles didnot show the typical convex shape (Fig. SB).

Molasses-acclimatized UASB aggregates. In the absence ofsubstrate, molasses-acclimatized aggregates showed a slightinternal pH increase of about 0.3 U (Fig. 6). This wasdifferent from the microprofiles measured in fresh aggregatesacclimatized to glucose (Fig. 1). Again, the presence of 2.5mM glucose in the medium resulted in the typical convexmicroprofile (Fig. 6). Moreover, even with a slightly lowerpH in the bulk medium, higher pH values were measured inthe center of molasses- versus glucose-acclimatized aggre-gates (Fig. 6 and Fig. 4A).

Aggregates in medium of pH 6.8 containing 2.5 mMglucose showed an internal pH drop of about 0.8 U (Fig. 6).Increasing the pH of the bulk liquid to 7.8 enlarged the pH

drop inside the granule with 0.7 U. The thickness of the layerover which the pH drops occurred (200 to 300 ,um), however,did not increase. No pH microgradients were observed inmolasses-acclimatized aggregates in reactor liquid (dilutedmolasses) at pH 6.8. Addition of bicarbonate to increase thepH to 7.8 did not give rise to the development of a pHgradient (data not shown).

In both molasses- and glucose-acclimatized granules, longthin filaments resembling Methanothrix spp. (Fig. 7A) werepresent throughout the granules and were even observed onthe granule surface. Besides these methanogens, sarcina-shaped cells were also present in the core of the molasses-acclimatized granules but not in the core of the glucose-acclimatized granules. Peripheral populations in bothmolasses- and glucose-acclimatized granules consisted ofdifferent rods and cocci (Fig. 7B). Granules consisted of anumber of concentrically arranged segments (Fig. 7C).

DISCUSSION

Microzonation of microbial activity. Measurements of pHand glucose microprofiles showed that metabolic activitiesinvolved in the methanogenesis of glucose were spatiallydistributed within the granules investigated. An outer zoneof 200 to 300 p,m with dominant glucose-converting activityand an inner zone where methanogenic activity was domi-nant could be distinguished. This in situ activity distributiondid not coincide with the multiple concentric layers observedin sliced granules (Fig. 7C). Hence, morphological distribu-tion within a granule (Table 1) does not necessarily corre-

spond with compartmentalized activity.In the deeper layers of the glucose-acclimatized aggre-

gates, glucose gradients were absent (Fig. 3) although glu-

FIG. 2.acetate.

VOL. 59, 1993 3807

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

-V-V

*'s.

V

17-V"'I V

Ak - Vp _'

-200 0 200 400 600

Depth (Am)

FIG. 3. Steady-state glucose microprofiles measured in glucose-acclimatized UASB aggregates in medium containing 0.5 (-), 2.5 (V), and5 (*) mM glucose. (A) 0.5 mM NaH2PO4. (B) 10 mM NaH2PO4.

cose concentrations were at least 10 times the reported Ksvalues for glucotrophic microorganisms (Table 2). This canbe explained either by pH inhibition of the glucose conver-sion or simply by the absence of glucotrophic microorgan-isms. As shown by the activity measurements, glucosedegradation rates were strongly inhibited at low pH. There-

fore, in weakly buffered medium, the low interfacial glucosegradient and the high glucose concentration in the center ofthe aggregate (Fig. 3A) could be attributed to low local pHs(Fig. 4A). The higher internal pHs in aggregates exposed tostrongly buffered medium (10 mM NaH2PO4) allowed glu-cose conversion at a higher rate. This resulted in increased

5

4

3k

A

EQ)U~)0u

0& 2

1

05

4

3KE

o

cD 2 k

o l-400 800

3808 LENS ET AL.

1

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

pH AND GLUCOSE MICROPROFILES IN UASB GRANULES 3809

7

Qa-

6

5

7

Q

6

5-1-400 -200 0 200 400 600 800

Depth (,um)FIG. 4. pH microprofiles in glucose-acclimatized UASB aggregates in the presence of glucose. (A) Steady-state pH microproffles

measured in medium (0.5 mM NaH2PO4) containing 0 (0), 0.5 (-), 2.5 (V), and 5 (*) mM glucose. (B) Influence of BES on the pHmicroprofiles: (0) 0 mM glucose; (V) 2.5 mM glucose; (V) 2.5 mM glucose + 10 mM BES. Notice that the microprofiles for 0 and 2.5 mMglucose presented in panels A and B are the same.

interfacial glucose fluxes and in lower glucose concentra-tions in the center of the aggregate (Fig. 3B). However, alsoat a high buffer capacity of the medium, gradients wereabsent, indicating that no glucose degradation-occurred in

the core of the granules. Simulations of the pH and glucosemicroprofiles (Fig. 5), including pH dependence of reactions1 and 2, strongly indicated the absence of glucotrophicmicroorganisms in the core of the aggregates as well, since

VOL. 59, 1993

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

5

4

Ea)co0:3

6)

3

2

0

7

C)

6

5

-400 -200 0 200 400 600 800

Depth (jim)FIG. 5. Comparison of modelled (lines) and measured (symbols) microprofiles of glucose (A) and pH (B) in glucose-acclimatized granules.

Glucose conversion was assumed to occur either exclusively in the outer 300 ,um of the granule ( ) or homogeneously throughout thegranule (--- -). Calculations and measurements were done with 2.5 mM glucose (V, V) or 5 mM glucose (*), in the presence of 0.5 mMNaH2PO4 (open symbols) or 10 mM NaH2PO4 (closed symbols) in the medium.

the best fits were obtained when fermentative activity wasrestricted to the outer 300 pm of the aggregates. Increasingin situ pH levels in molasses-acclimatized granules resultedin a higher glucose-degrading activity, as indicated by the

larger pH drop (Fig. 6). However, the thickness of the layerover which the pH drop occurred did not change, indicatingthat also in molasses-acclimatized granules the glucose-converting microorganisms were mainly located in the outer

3810 LENS ET AL.

1

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

pH AND GLUCOSE MICROPROFILES IN UASB GRANULES 3811

8

7

a! X ~~~~~~~~~~~~~~VV-v--vvV-

6

-400 -2'00 0 200 400 600 800

Depth (Atm)FIG. 6. Influence of the pH of the medium on the shape of pH microprofiles in molasses-acclimatized UASB aggregates. O, 0mM glucose;

0, 2.5 mM glucose, pH 6.8; V, 2.5 mM glucose, pH 7.8.

200 to 300 ,um. These results are in agreement with thefindings of Guiot et al. (20), who found the highest glucose-converting activity in the outer 14% volume fraction ofabrased mesophilic granules grown on sucrose. SimulatedpH microprofiles did not fit well in the periphery of theaggregates (Fig. 5B). This is mainly due to the simplificationsand assumptions made in the model, such as the represen-tation of methanogenesis of glucose by only two reactions.Methanogenic activity could be spatially assigned to the

center of the aggregates by pH microprofile measurements incombination with a specific inhibitor (BES) (Fig. 4B). How-ever, microscopic observations and measurements of theunique coenzyme F420 of methanogenic bacteria (data notshown) indicated a homogeneous distribution of methano-gens. This has also been reported for other types of granules(14, 18, 59). The pH microprofiles do not allow us to estimatethe role of methanogenic activity in the periphery of theaggregates since pH microprofiles reflect the combined effectof acid-producing and acid-consuming activities. Both themicroprofiles and the activity measurements with acetateshowed that methanogenic activity occurred at pHs as lowas 5.2 to 5.5. This is somewhat surprising since the optimalpH for methanogens ranges between 6.3 and 7.3 for mostmethanogens (60). Growth of methanogenic bacteria hasnevertheless been reported at pH values of 3.5 (38).

Effect of the acclimatization procedure. Different acclima-tization procedures resulted in differently shaped pH micro-profiles measured in medium without substrate (Fig. 1 andFig. 6). This might be due to different metabolic behaviorinduced by the growth media used in the acclimatizationprocedure or to different species in the bacterial population,as evidenced by microscopic observations. In granules that

had been stored 5 months prior to glucose acclimatization,methanogenic activity was dominant over endogenous fer-mentative metabolism. The reduced activity of the acidifyingbacteria may be explained by their higher die-off ratescompared with those of methanogens (16, 22). Conse-quently, although granules can be stored unfed for more than6 months at temperatures of 10 to 20°C without seriousdeterioration (10, 62), their microbial composition and activ-ity can change considerably over time.

Effective diffusion coefficient for glucose. The measuredeffective diffusion coefficient for glucose in the aggregateswas 55 to 75% of the molecular diffusion coefficient (Do) ofglucose in water at 300C (6.8 x 10-10 m2/s) (45). Similarratios for DJIDo have been reported for the diffusivity ofglucose in aerobic flocs and biofilms (31, 39, 42). Thedeactivation procedure influenced the De value: heat treat-ment or exposure of the aggregates to pure chloroform gaverise to somewhat higher De values compared with deactiva-tion by exposure to glutaraldehyde or HgCl2. There is stillsome controversy about the effect of the deactivationmethod on the diffusivity. Heat treatment (53) and exposureto HgCl2 (46) have been reported to affect the diffusionalcharacteristics of cell aggregates, whereas others (11, 39)found no effect of exposure to glutaraldehyde or HgCl2.Methods where no deactivation of the aggregates is neces-sary, e.g., the use of nonmetabolizable tracers like lithiumsalts (25) and noninvasive nuclear magnetic resonance tech-niques (52), may be good alternatives to investigate diffu-sional properties.

Practical implications of microprofiles. Microprofilesshowed a diffusive boundary layer of 100 to 200 ,um for bothglucose (Fig. 3) and pH (Fig. 4). As the average fluid velocity

VOL. 59, 1993

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

3812

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

pH AND GLUCOSE MICROPROFILES IN UASB GRANULES 3813

FIG. 7. SEM micrographs of the UASB aggregates. Long, thin filaments, most probably Methanothrix-like bacteria present in the core (A)and different microbial species observed in the outer 200 to 300 pLm (B) of molasses-acclimatized aggregates. (C) Low-magnification view ofa cleaved glucose-acclimatized aggregate showing the multilayered structure of the granule.

in the flow cell (0.5 mm/s) was comparable to the upstreamvelocity in the reactor (0.3 mm/s), a comparable diffusiveboundary layer will prevail in the reactor. The presence ofsuch a layer results in decreased mass transfer from the bulkliquid to the aggregate. The resistance to outdiffusion ofacids (protons) resulted in low internal pHs in medium withlow buffer capacity (0.5 mM NaH2PO4) (Fig. 4A), whichinhibited glucose conversion (Fig. 3). It is generally assumedthat internal mass transfer resistance is not significant inmethanogenic biofilms (22). Verification experiments byDolfing (12) and Schmidt and Ahring (49) were performed inwell-buffered media (5 mM NaH2PO4). This study shows,however, that this assumption is only valid in well-bufferedmedia.The pH is known to influence fermentation pathways (35),

the inhibition of methane-producing bacteria by sulfide (26)and volatile fatty acids (27), and the kinetic parameters Fm.and Ks (17, 55). It therefore can be expected that theexistence of pH microproffles in aggregates at reactor con-ditions will influence the reactor performance. In this study,internal pH microgradients were observed when aggregateswere exposed to a weakly buffered (0.5 mM NaH2PO4)medium but not in highly buffered medium (33 mMNaH2PO4) or reactor liquid (molasses). The magnitude ofpHgradients at reactor conditions was estimated by consideringthe following operational parameters of the UASB reactor(58): a volumetric loading rate of 7 kg of glucose per m3 perday and an average sludge concentration of 25 kg of cell dryweight per m3 with an average granule diameter of 1.5 mm.Assuming that 1 mol of glucose is converted to 2 mol ofacetate, complete metabolization of the feed would result ina flux of 0.004 meq of organic acid per cm2 of granule surfaceper day. With an average effective diffusion coefficient of 1.2x 10' m2/s, this flux equals an interfacial substrate gradientof about 0.4 mol eq/m3/mm. Hence, an internal pH gradient

TABLE 2. Parameters used for the simulation of theconcentration profiles

Unit ofParameterA measure- Magnitude Reference

ment"

Conversion ratecVm., mol/m3/s 0.019 This studyVma2 mol/M3/s 0.014 This studyKs mol/m3 0.05 64

Ks2 moIm3 0.5 12Diffusion coefficientDei m2/s 4.42 x 10-10 This studyDe2 m2/s 1.20 x 10 9 2dDe3 m2/s 1.13 x 10-9 2d

Layer thicknessLeft7e m 0.3 x 10-6 This studya Parameters given for the following (subscripts): 1, glucose; 2, acetate; 3,

methane.b Vm,, expressed as moles per cubic meter (wet weight) per second; Ks

expressed as moles per cubic meter (dry weight).I The data of Vm.,. were calculated assuming that 1 g VSS = 7.4 g (wet

weight) (8).d Adjusted for 30'C according to the Stokes-Einstein relation (51).'e In case that glucose is only converted in the periphery of the aggregate.

will only occur in reactors treating weakly (a few equivalentsper cubic meter) buffered wastewaters. Although the phos-phate concentration of most wastewaters is usually wellbelow 1 mM, other buffer systems can give rise to a buffercapacity of 10 to 100 eq/m3 (41). It is therefore unlikely thatpH microprofiles generally occur at reactor conditions.However, in case of overloading or disturbance of themethanogenic population, acid production might consumeall the alkalinity present in the wastewater. In such cases,internal pH drops accompanied with reduced conversionrates (8) can be expected. This implies that the alkalinity ofthe reactor liquid is a far more important control parameterthan the pH. Further research combining on-line alkalinitymonitors (47) with microsensors can optimize control strat-egies for anaerobic digesters based on a thorough under-standing of microlevel phenomena in UASB granules.

ACKNOWLEDGMENTS

We thank Biotim N.V., Antwerpen, Belgium, for the gift of thegranules. We are grateful to Geert Boucneau and JullapongThaveesri for technical assistance, to Bert van Groen for his helpwith the glucose microsensor, and to W. Bohyn for his help with theSEM. H. van den Heuvel and G. Maestrojuan are gratefullyacknowledged for valuable discussions during the preparation of themanuscript.

This study was supported by a grant of the Instituut voorWetenschappelijk Onderzoek in Nijverheid en Landbouw (I.W.O.N.L.), Brussels, Belgium.

REFERENCES1. Alibhai, K. R. K., and C. F. Forster. 1986. Physicochemical and

biological characteristics of sludges produced in anaerobic up-flow sludge blanket reactors. Enzyme Microb. Technol. 8:601-606.

2. Andrussow, L., and B. Schramm. 1969. Diffusion in flus-sigkeiten. Binare systeme, p. 639-645. In K. Schafer (ed.),Landolt-Bornstein, Zahlenwerte und Funktionen aus Physikusw. 5. Teil, Bandteil A, Transportphanomene I. SpringerVerlag, Berlin.

3. Bochem, H. P., S. M. Schoberth, B. Spey, and P. Wengler. 1982.Thermophilic biomethanation of acetic acid: morphology andultrastructure of a granular consortium. Can. J. Microbiol.28:500-510.

4. Bryant, M. P. 1979. Microbial methane production-theoreticalaspects. J. Anim. Sci. 48:193-201.

5. Cronenberg, C., and J. C. van den Heuvel. 1991. Determinationof glucose diffusion coefficients in biofilms with micro-elec-trodes. Biosens. Bioelectronics 6:255-262.

6. Cronenberg, C., B. van Groen, D. de Beer, and H. van denHeuvel. 1991. Oxygen-independent glucose microsensor basedon glucose oxidase. Anal. Chim. Acta 242:275-278.

7. Cronenberg, C. C. H., J. C. van den Heuvel, and S. P. P.Ottengraf. 1993. Direct measurement of glucose profiles inimmobilized yeast gels with a pH-insensitive micro-electrodeunder anaerobic conditions. Biotechnol. Tech. 7:237-242.

8. de Beer, D., J. W. Huisman, J. C. van den Heuvel, and S. P. P.Ottengraf. 1992. The effect of pH profiles in methanogenicaggregates on the kinetics of acetate conversion. Water Res.26:1329-1336.

9. de Beer, D., and J. C. van den Heuvel. 1988. Gradients inimmobilized biological systems. Anal. Chim. Acta 213:259-265.

10. de Zeeuw, W. J. 1988. Granular sludge in UASB reactors, p.132-145. In G. Lettinga, A. J. B. Zehnder, J. T. C. Grotenhuis,and L. W. Hulshoff Pol (ed.), Proceedings of the workshop on

VOL. 59, 1993

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

APPL. ENVIRON. MICROBIOL.

granular sludge; microscopy and technology. Pudoc Publishers,Wageningen, The Netherlands.

11. Dibdin, G. H. 1981. Diffusion of sugars and carboxylic acidsthrough human dental plaque in vitro. Arch. Oral Biol. 26:515-523.

12. Dolfing, J. 1985. Kinetics of methane formation by granularsludge at low substrate concentrations. Appl. Microbiol. Bio-technol. 22:77-81.

13. Dolfing, J. 1988. Acetogenesis, p. 417-468. In A. J. B. Zehnder(ed.), Biology of anaerobic microorganisms. John Wiley &Sons, Inc., New York.

14. Dolfing, J., A. Griffioen, A. van Neerven, and L. Zevenhuizen.1985. Chemical and bacteriological composition of granularmethanogenic sludge. Can. J. Microbiol. 31:744-750.

15. Dubourguier, H. C., G. Prensier, and G. Albagnac. 1988. Struc-tural and microbial activities of granular anaerobic sludge, p.18-33. In G. Lettinga, A. J. B. Zehnder, J. T. C. Grotenhuis,and L. W. Hulshoff Pol (ed.), Proceedings of the workshop ongranular sludge; microscopy and technology. Pudoc Publishers,Wageningen, The Netherlands.

16. Ghosh, S., and F. G. Pohland. 1974. Kinetics of substrateassimilation and product formation in anaerobic digestion. J.Water Pollut. Control Fed. 46:748-758.

17. Graef, S. P., and J. F. Andrews. 1973. Mathematical modellingand control of anaerobic digestion. AICHE Symp. Ser. 136:101-131.

18. Grotenhuis, J. T. C., M. Smith, C. M. Plugge, X. Yuansheng,A. A. M. van Lammeren, A. J. M. Stams, and A. J. B. Zehnder.1991. Bacteriological composition and structure of granularsludge adapted to different substrates. Appl. Environ. Micro-biol. 57:1942-1949.

19. Guiot, S. R., F. A. Macleod, and A. Pauss. 1989. Thermodynam-ical and microbiological evidence of trophic microniches forpropionate degradation in a methanogenic sludge-bed reactor,p. 173-183. In J.-P. Belaich, M. Bruschi, and J. L. Garcia (ed.),Microbiology and biochemistry of strict anaerobes involved ininterspecies transfer. Plenum Press, New York.

20. Guiot, S. R., A. Pauss, and J. W. Costerton. 1991. A structuredmodel of the anaerobic granule consortium. Water Sci. Technol.25:1-10.

21. Harada, H., G. Endo, Y. Tohya, and K. Momonoi. 1988. Highrate performance and its related characteristics of granulatedsludges in UASB reactors treating various wastewaters, p.1011-1020. In A. Tilche and A. Rozzi (ed.). Proceedings of thefifth international symposium on anaerobic digestion. Bologna,Italy.

22. Henze, M., and P. Harremoes. 1983. Anaerobic treatment ofwastewater in fixed film reactors-a literature review. WaterSci. Technol. 15:1-101.

23. Holdeman, L. V., E. P. Cato, and W. E. C. Moore (ed.). 1977.Anaerobe laboratory manual, p. 130. Virginia Polytechnic Insti-tute and State University, Blacksburg.

24. Hulshoff Pol, L. W. 1989. The phenomenon of granulation ofanaerobic sludge. Dissertation, Wageningen Agricultural Uni-versity, Wageningen, The Netherlands.

25. Kitsos, H. M., R. S. Roberts, W. J. Jones, and T. G. Tornabene.1992. An experiment study of mass diffusion and reaction rate inan anaerobic biofilm. Biotechnol. Bioeng. 39:1141-1146.

26. Koster, I. W., A. Rinzema, A. L. de Vegt, and G. Lettinga. 1986.Sulfide inhibition of the methanogenic activity of granularsludge at various pH-levels. Water Res. 20:1561-1567.

27. Kus, F., and U. Wiesmann. 1991. Kinetics of the anaerobicdegradation of acetic and propionic acid. p. 521-524. In H.Verachtert and W. Verstraete (ed.), Proceedings of the interna-tional symposium environmental biotechnology. KoninklykeVlaamse Ingenieursvereniging, Antwerp, Belgium.

28. Lens, P., M.-P. De Poorter, C. C. Cronenberg, and W. Verstra-ete. Sulfate reducing and methane producing bacteria in aerobicwastewater treatment systems. Submitted for publication.

29. Lettinga, G., R. Roersma, and P. Grin. 1983. Anaerobic treat-ment of raw domestic sewage at ambient temperature using agranular bed UASB reactor. Biotechnol. Bioeng. 25:1701-1723.

30. Lettinga, G., R. Roersma, P. Grin, W. de Zeeuw, L. Huishoff, L.

Van Velsen, S. Hobma, and G. Zeeman. 1981. Anaerobic treat-ment of sewage and low strength wastewaters, p. 271-291. InProceedings of the second international symposium on anaero-bic digestion. Travemunde, Germany.

31. Libicki, S. B., P. M. Salmon, and C. R. Robertson. 1988. Theeffective diffusive permeability of a nonreacting solute in micro-bial cell aggregates. Biotechnol. Bioeng. 32:68-85.

32. Lievense, L. C., K. van't Riet, and A. Noomen. 1990. Measure-ment and modelling the glucose-fermenting activity of Lactoba-cillus plantarum. Appl. Microbiol. Biotechnol. 32:669-673.

33. Liou, J. K., and I. Rousseau. 1986. Mathematical model forinternal pH control in immobilized enzyme particles. Biotech-nol. Bioeng. 28:1582-1589.

34. Littmann, S., E. Schulte, and L. Acker. 1982. Beurteilung deshygienischen Zustandes von Eiprodukten von der Pasteuris-ierung uber das Muster der organischen Sauren. Z. Lebensm.Unters. Forsch. 175:101-105.

35. Lowe, S. E., and J. G. Zeikus. 1991. Metabolic regulation ofcarbon and electron flow as a function of pH during growth ofSarcina ventriculi. Arch. Microbiol. 155:325-329.

36. Macario, A. J. L., F. A. Visser, J. B. van Lier, and E. Conway deMacario. 1991. Topography of methanogenic subpopulations ina microbial consortium adapting to thermophilic conditions. J.Gen. Microbiol. 137:2179-2189.

37. MacLeod, F. A., R. S. Guiot, and J. W. Costerton. 1990.Layered structure of bacterial aggregates produced in an upflowanaerobic sludge bed and filter reactor. Appl. Environ. Micro-biol. 56:1598-1607.

38. Maestrojuan, G. M., and D. R. Boone. 1991. Characterization ofMethanosarcina barkeni MST and 227, Methanosarcina mazeiS-6T, and Methanosarcina vacuolata Z-761T. Int. J. Syst. Bac-teriol. 41:267-274.

39. Matson, J. V., and W. G. Characklis. 1976. Diffusion intomicrobial aggregates. Water Res. 10:877-885.

40. McCarty, P. L. 1982. One hundred years of anaerobic treat-ment, p. 3-22. In D. E. Hughes, D. A. Stafford, B. F. Weatley,W. Baader, G. Lettinga, E. J. Nyns, W. Verstraete, and R. L.Wentworth (ed.), Anaerobic digestion. Elsevier Biochemical,Amsterdam.

41. Metcalf and Eddy, Inc. 1984. Wastewater engineering: treat-ment, disposal, reuse, 2nd ed. G. Tchobanoglous (ed.) McGraw-Hill, New York.

42. Onuma, M., T. Omura, T. Umita, and J. Aizawa. 1985. Diffusioncoefficient and its dependency on some biochemical factors.Biotechnol. Bioeng. 27:1533-1539.

43. Pauss, A., R. Samson, and S. Guiot. 1990. Thermodynamicevidence of trophic microniches in methanogenic granularsludge-bed reactors. Appl. Microbiol. Biotechnol. 33:88-92.

44. Pearson, C. E. 1986. Numerical methods in engineering andscience, p. 165. Van Nostrand Reinhold, New York.

45. Perry, R. H., and D. Green. 1984. Chemical engineers' hand-book, 6th ed., p. 3-285-3-287. McGraw-Hill, New York.

46. Revsbech, N. P. 1989. Diffusion characteristics of microbialcommunities determined by use of oxygen microsensors. J.Microbiol. Methods 49:111-122.

47. Rozzi, A., A. C. Di Pinto, and A. Brunetti. 1985. Anaerobicprocess control by bicarbonate monitoring. Environ. Technol.Lett. 6:594-601.

48. Sam-soon, P. A. L. N. S., R. E. Loewenthal, P. L. Dold, andG. V. R. Marais. 1987. Hypothesis for pelletization in the upflowanaerobic sludge blanket reactor. Water S A 13:69-80.

49. Schmidt, J. E., and B. K. Ahring. 1991. Acetate and hydrogenmetabolism in intact and disintegrated granules from an acetate-fed, 55'C, UASB reactor. Appl. Microbiol. Biotechnol. 35:681-685.

50. Schmidt, J. E., A. J. L. Macario, B. K. Ahring, and E. Conwayde Macario. 1992. Effect of magnesium on methanogenic sub-populations in a thermophilic acetate-degrading granular con-sortium. Appl. Environ. Microbiol. 58:862-868.

51. Skelland, A. H. P. 1974. Diffusional mass transfer, p. 54-82.John Wiley & Sons, Inc., New York.

52. Stilbs, P. 1987. Fourier transform pulsed-gradient spin-echostudies of molecular diffusion. Prog. NMR Spectr. 19:1-45.

3814 LENS ET AL.

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from

pH AND GLUCOSE MICROPROFILES IN UASB GRANULES 3815

53. Tatevossian, A. 1985. The effect of heat inactivation, tortuosity,extracellular polyglucan and ion-exchange sites on the diffusionof ["4C]-sucrose in human dental plaque residue in vitro. Arch.Oral Biol. 30:365-371.

54. Vandenabeele, J., D. de Beer, R. Germonpre, and W. Verstraete.1992. Manganese oxidation by microbial consortia from sandfilters. Microb. Ecol. 24:91-108.

55. van den Berg, L., G. B. Patel, D. S. Clark, and C. P. Lentz. 1976.Factors affecting rate of methane formation from acetic acid byenriched methanogenic cultures. Can. J. Microbiol. 22:312-319.

56. Vanderhaegen, B., E. Ysebaert, K. Favere, M. Van Wambeke, T.Peeters, V. Panic, V. Vandenlangenbergh, and W. Verstraete.1991. Acidogenesis in relation to in-reactor granule yield. WaterSci. Technol. 25:21-30.

57. Van Meenen, P., J. Vermeulen, and W. Verstraete. 1988. Fra-gility of anaerobic SSF consortia, p. 345-356. In E. R. Hall andP. N. Hobson (ed.), Anaerobic digestion 1988. IAWPRC, Lon-don.

58. Van Wambeke, M., S. Grusenmeyer, W. Verstraete, and R.Longry. 1990. Sludge bed growth in an UASB reactor treatingpotato processing wastewater. Process Biochem. 25:181-186.

59. Visser, F. A., J. B. van Lier, A. J. L. Macario, and E. Conway deMacario. 1991. Diversity and population dynamics of methano-genic bacteria in a granular consortium. Appl. Environ. Micro-biol. 57:1728-1734.

60. Vogels, G. D., J. Keltjens, and C. Van der Drift. 1988. Biochem-istry of methane production, p. 707-770. In A. J. B. Zehnder(ed.), Biology of anaerobic microorganisms. John Wiley &Sons, Inc., New York.

61. Weast, R. C., and M. J. Astle. 1980. CRC handbook of chem-istry and physics, p. D-161-D-167. CRC Press, Inc., BocaRaton, Fla.

62. Wu, W., J. Hu, and X. Gu. 1985. Properties of granular sludgein upflow anaerobic sludge blanket (UASB) reactor and itsperformance, p. 339-351. In China State Biogas Association(ed.), Anaerobic digestion 1985. Guangzhou, China.

63. Zehnder, A. J. B. 1978. Ecology of methane formation, p.349-376. In R. Mitchell (ed.), Water pollution microbiology,vol. 2. Wiley, New York.

64. Zoetemeyer, R. J., J. C. van den Heuvel, and A. Cohen. 1982. pHinfluence on acidogenic dissimilation of glucose in an anaerobicdigestor. Water Res. 16:303-311.

VOL. 59, 1993

on March 5, 2021 by guest

http://aem.asm

.org/D

ownloaded from