Vol. 58, No. 9APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Sept. 1992, p. 2751-27570099-2240/92/092751-07$02.00/0Copyright © 1992, American Society for Microbiology

Examination of Bacterial Characteristics of AnaerobicMembrane Bioreactors in Three Pilot-Scale Plants for

Treating Low-Strength Wastewater by Application of theColony-Forming-Curve Analysis Method

NAOAKI KATAOKA,1 YUTAKA TOKIWA,2* YASUO TANAKA,1 KIICHI FUJIKI,3HIROYUKI TARODA,4 AND KIYOSHI TAKEDA2

Development and Planning Department, Ebara Corporation, 1-6-27, Khonan, Minato-ku,Tokyo 108,1 Chemical Ecology Division, Fermentation Research Institute, Agency of Industrial Science

and Technology, MITI, 1-1-3, Higashi Tsukuba, Ibaraki 305,2 Biotechnology and EnvironmentalEngineering Department, Research Institute, Ishikawajima-Harima Heavy Industries Co., Ltd.,

Isogo-ku, Yokohama-shi, Kanagawa 235,3 and Research and Development Department,DIC-Degremont Co., Ltd., 631, Sakado, Sakura-shi, Chiba 285,4 Japan

Received 2 January 1992/Accepted 10 June 1992

Characteristic sludge ecosystems arising in anaerobic membrane bioreactors of three pilot-scale plantstreating low-strength (less than 1 g of biological oxygen demand per liter) sewage or soybean-processingwastewater were examined by analysis of the colony-forming-curves (CFC) obtained by counting colonies atsuitable intervals. The wastewaters, containing high amounts of suspended solids (SS) (SS/chemical oxygendemand ratio, 0.51 to 0.80), were treated by using two types of bioreactors: (i) a hydrolyzation reactor forsolubilization and acidification of SS in wastewater and (ii) a methane fermentation reactor for producingmethane. The colony counts for the two sewage treatment plants continued to increase even after 3 weeks ofincubation, whereas those for soybean-processing wastewater reached an approximately constant level within3 weeks of incubation. The CFCs were analyzed by correlating the rate of colony appearance on roll tubes withthe physiological types of bacteria present in the bioreactors. It was found that there were large numbers ofslow-colony-forming anaerobic bacteria within the bioreactors and that the viable populations consisted of afew groups with different growth rates. It is considered that the slow-growing colonies appearing after 10 daysof incubation were the dominant microflora in the sewage treated by hydrolyzation reactors. In particular,highly concentrated sludge (30.0 g of mixed-liquor volatile SS per liter) retained by the membrane separationmodule contained a large number of such bacteria. Slow-growing colonies of these bacteria could be countedby using a sludge extract medium prepared from only the supernatant of autoclaved sludge. In addition, thehighest colony counts were almost always obtained with the sludge extract medium, meaning that most of theanaerobic bacteria in these sludges have complex nutrient requirements for growth. This report also indicatesthe usefulness of application of the CFC analysis method to the study of bacterial populations of anaerobictreatment systems.

The anaerobic degradation of complex organic matter tomethane and carbon dioxide involves the interaction of fourmetabolic groups of bacteria, namely, hydrolytic, acetoge-nic, homoacetogenic, and methanogenic bacteria (35). Aconsiderable amount of research into the microbiology ofanaerobic digesters, using techniques such as immunologicanalysis (21-23), enzyme activity measurement (2, 7, 8), lipidand poly-p-hydroxybutyric acid analysis (13), and micro-scopic observation (10, 24, 27, 29), has recently been re-ported. Although the bacterial colony-forming method usinga roll tube (15) has frequently been applied to enumeration ofthe total viable number after a period of incubation and toisolation of the individual physiological bacteria groups inanaerobic digestion ecosystems (9, 16, 31-33), little attentionhas so far been given to the rate of colony appearance on theroll tubes.

It has been reported that the colony-forming-curves (CFC)of soil bacteria obtained by counting the number of visiblecolonies on a plate at suitable intervals give some informa-tion, such as the ratio of the population size and the growth

* Corresponding author.

rate of bacteria by the superposition of the first-order-reaction model proposed by Hattori (11, 12, 17, 18). Therelationship between the number of visible colonies on aplate and the incubation time was formulated on the basis ofthe first-order-reaction model. Ishikuri and Hattori (18) havepreviously reported that each tier of the CFC was composedof a group having a similar growth rate and that the retarda-tion time until the appearance of visible colonies of thepopulation on a plate directly reflected the growth rate ofbacteria from natural environments. In methanogenic en-richment populations, the time required for formation ofcolonies by the pure isolates was similar to that required forthe formation of a colony by that culture during its initialisolation (19). Thus, the growth characteristics of anaerobicbacterial populations can be represented by their CFCs.However, there are no reports of the CFC analysis methodbeing applied to the analysis of bacterial populations ofanaerobic treatment systems.The present research relates to obtaining the growth and

physiological characteristics of highly concentrated anaero-bic bacterial populations arising in membrane bioreactors bythe application of the CFC analysis method. The microbiol-

2751

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

2752 KATAOKA ET AL.

TABLE 1. Characteristics of wastewater fed to three plants

Amt (mg/liter) of:Plant

BOD COD SS TOC

A 240 490 300 NaB 214 353 283 138C 852 1,357 689 555a NT, not tested.

ogy of these populations is investigated by correlating thetime required for formation of a visible colony on roll tubeswith the physiological types of anaerobic bacteria present inthe membrane bioreactors. This paper also shows that theCFC analysis method is useful for studying the microbiologyof anaerobic sludge ecosystems.

MATERIALS AND METHODS

Source of samples. Anaerobic sludge samples were takenfrom the hydrolyzation and methane fermentation reactorsof three pilot-scale anaerobic wastewater treatment plants.Screw bottles (100 ml) were filled to the top with mixed-liquor samples and immediately closed tightly with stoppers.Sludge samples transported to the laboratory were stored at4°C for a maximum of 5 days until subjected to chemical andbacterial analysis.

Operating conditions of the three plants. The analyticaldata on wastewaters fed to the three pilot-scale plants (A, B,and C) are shown in Table 1. The three wastewaters had lowbiological oxygen demands (BOD) and high suspended-solid(SS) concentrations. The SS/chemical oxygen demand(COD) ratio was high, in the range of 0.51 to 0.80. Plant Aand B wastewaters were sewage with high levels of cellulosicmaterials. Plant C was soybean-processing wastewater dis-charged from an edible-oil refinery plant and containedrelatively high amounts of SS composed mainly of proteins(544 mg/liter) and carbohydrates (234 mg/liter) from soybean(34).

Plant A was for domestic sewage and had a designcapacity of 20 m3/day. The plant consisted of pretreatmentequipment (a screen [pinhole opening, 20 ,um] for separatingSS from the raw sewage), a hydrolyzation reactor (totalvolume, 2.0 mi3) for solubilization of the separated SS, anupflow anaerobic sludge blanket (UASB)-type reactor (totalvolume, 5.4 mi3; reaction volume, 1.6 mi) for methanefermentation, and a membrane separation unit. The hydro-lyzation reactor was maintained at a temperature of 35°C anda pH of 6. Enzyme for cellulose hydrolysis was added at 3%per influent SS (9.5 g/liter). The hydraulic retention time was3 days. The UASB reactor, combined with an external-pressure capillary-type ultrafiltration membrane module(molecular weight cutoff, 15,000; membrane area, 100 mi2)made of polysulfone and polyvinylalcohol, was operated atordinary temperature; the effective volumetric loading ratewas 2.8 kg of BOD per m3/day.

Plant B was for municipal sewage and had a designcapacity of 10 m3/day. The plant consisted of pretreatmentequipment (1 m3) (a polyurethane sponge for removing theSS), a hydrolyzation reactor (total volume, 0.5 m3) forsolubilization of the concentrated SS (more than 10 g/liter), a

fluidized-bed-tyge reactor (total volume, 1.0 mi3; reactionvolume, 0.66 m ) for methane fermentation, and membraneseparation units. The hydrolyzation reactor, combined withan internal-pressure tubular-type ultrafiltration membrane

TABLE 2. Efficiency of treatment processes in three plants

Gas productionFeed rate % Removal' of: (Normal liters ofPlant (m3/day) gas/kg of COD)

BOD COD TOC Total Methane

A 20.0 86.6 83.0 NT 510 18B 11.5 91.4 89.7 89.3 22 11C 7.18 71.8 77.7 74.4 241 201

a (Total influent concentration - effluent concentration)/total influent con-centration. NT, not tested.

module (molecular weight cutoff, 13,000; membrane area,0.94 m2) made of polyacrylonitrile, was maintained at atemperature of 30°C and a pH of 5.5 to 6.5. The hydraulicretention time was 5 days. The fluidized-bed reactor, com-bined with an external-pressure hollow-fiber-type microfil-tration membrane module (pore size, 0.1 ,um; membranearea, 54 mi2) made of polyethylene, was operated at ordinarytemperature; effective volumetric loading rate was 1.1 kg ofBOD per m3/day.

Plant C was for soybean-processing wastewater with ahigh soybean solid concentration and had a design capacityof 7.5 m3/day. The plant consisted of a fixed-bed-typereactor (total volume, 3 mi3), which had a two-phase systemcomposed of acid fermentation (total volume, 1.0 mi3; reac-tion volume, 0.54 mi3) and methane fermentation (totalvolume, 2.0 mi3; reaction volume, 1.6 mi3) sections, and amembrane separation unit. Both sections were packed withnonwoven fabric carriers. The reactor, combined with anexternal-pressure capillary-type ultrafiltration membranemodule (molecular weight cutoff, 15,000; membrane area, 50mi2) made of polysulfone and polyvinylalcohol, was main-tained at a temperature of 30°C, and the pHs of the acid andmethane fermentation sections were maintained at 6.0 and7.5, respectively. The hydraulic retention times of the acidand methane fermentation sections were 3.3 to 3.5 and 6.7 to7.0 h, respectively. The volumetric loading rate was 2.0 kg ofBOD per m3/day.The treatment performances of the three plants are shown

in Table 2. The SS in the plant effluent was less than 1mg/liter.

Procedures for bacterial enumeration. (i) Sample prepara-tion. Strict anaerobic techniques used for medium prepara-tion were essentially the same as those of Hungate (15) asmodified by Azuma and Suto to develop the gas jet method(3). Traces of 02 in the gases were removed by passing thegases through a reduced-copper column. The composition ofmedium for serial dilution was as follows: (NH4)2HP04, 1.0g; KH2PO4, 0.2 g; K2HPO4, 1.6 g; MgSO4 7H20, 0.2 g;NaCl, 0.1 g; CaC12 2H20, 0.02 g; FeSO4- 7H20, 0.01 g;Na2MoO4 2H20, 0.5 mg; Na2WO4. 2H20, 0.5 mg;MnSO4, 0.5 mg; cysteine. HCl1 H20, 0.25 g; Na2S. 9H20,0.25 g; resazurin, 1 mg in 1 liter of distilled water. The pHwas adjusted to 7.0. The mixed-liquor sludge samples werediluted 1:10 in the dilution medium and homogenized with aNissei AM-77 homogenizer at 18,000 rpm for 1 min underice-cold conditions. The homogenates were then diluted 1:10in the dilution medium and sonicated with a Tomy SeikoUR-20R sonicator at 20-W output for 2 min under ice-coldconditions. The homogenization and sonication procedureswere carried out in a Hirasawa ANX-1L anaerobic glove boxunder an atmosphere of 80% N2, 10% CO2, and 10% H2.Serial 10-fold dilutions were made with butyl rubber stopper-

APPL. ENvIRON. MICROBIOL.

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ANAEROBIC MEMBRANE BIOREACTORS 2753

TABLE 3. Characteristics of anaerobic sludges studied

Plant andtypeH MLVSS TOC VFAb (mM)orcP(g/liter) (mg/liter) Acetate Propionate Butyrate Valerate

AHR 6.2 6.7 640 6.9 7.1 1.2 0.9MFR 6.5 39.6 30 0.2 <0.1 ND ND

BHR 5.7 30.0 1,045 3.6 4.4 0.9 1.0MFR 6.6 8.5 43 0.1 ND ND ND

C (MFR) 6.1 2.0 307 3.9 1.3 0.8 0.7

a HR, hydrolyzation reactor; MFR, methane fermentation reactor.b ND, not detected.

sealed tubes containing 9.0 ml of the dilution medium andprepared to a dilution of 10-' by using 1-ml gassed sterileplastic syringes.

(ii) Culture media and colony counts. Three kinds of mediawere used for the colony counts by the roll tube method (15).Balch-1 medium was the same as medium number 1 of Balchet al. (4), which contained 0.25% sodium formate, 0.25%sodium acetate, 0.2% yeast extract, 0.2% Trypticase pep-tone, mineral salts, trace vitamins, trace metals, and reduc-ing agents. m-VL medium, which was modified from thatdescribed by Azuma and Suto (3) and Ueki et al. (32), wasprepared by omitting the supematant of sewage digestorfluid from VLS agar (32). It contained 0.025% each ofglucose, cellobiose, maltose, and soluble starch plus 0.5%Trypticase peptone, 0.25% yeast extract, 0.01% beef ex-tract, mineral salts, and reducing agents. The sludge extract(SE) medium contained 0.25 g each of cysteine. HCl H20and Na2S. 9H20 plus 1 mg of resazurin and 22 g of agar perliter of the supematant of autoclaved sludge. SE for enumer-ation was made for each sludge and prepared by autoclavingat 121°C for 20 min and centrifuging at 12,000 x g for 30 min.Total organic carbon (TOC) values in the prepared SE werein the range of 600 to 1,050 mg/liter. The gas phases and finalpHs of these three media were N2 and 6.8 to 7.0, respec-tively.For colony counts, samples (0.5 ml) diluted within the

range of 10-5 to 10-8 were each inoculated into duplicate32-ml pressure glass tubes containing 9.0 ml of medium, andthen the roll tubes were formed. Cultures were incubated at30°C. Visible colonies on the roll tube were counted atsuccessive intervals. By plotting the mean of colony countsfrom duplicate tubes for each incubation time, CFCs wereobtained. The final counts for roll tubes differed by less than9% from the mean. Population numbers were expressed asCFU per milliliter of sludge or CFU per gram of mixed-liquor volatile SS (MLVSS; described below).

(iii) Counts of sulfate-reducing bacteria. Sulfate-reducingbacteria were enumerated by the roll tube method withm-Butlin medium, which was a modification of medium C ofButlin et al. (6) and contained 0.35% sodium lactate, 0.1%yeast extract, mineral salts, and reducing agents. The gasphase was N2, and the final pH was 7.1 to 7.4. The methoddescribed by Ueki (31) was used to count black colonies,which were considered sulfate-reducing bacteria.

(iv) Counts of methanogenic bacteria. Acetate- or formate-utilizing methanogenic bacteria were enumerated by themost-probable-number (MPN) technique with three tubesper dilution. The enumeration was performed in the pressureglass tubes containing 9.0 ml of the Balch-1 medium de-scribed above. The gas phase was N2. Samples (1.0 ml)

diluted within the range of 10-' to 10-' were inoculated intothree MPN tubes. After 4 weeks of incubation at 30°C, thegas phase of each tube was examined for the presence ofmethane, and acetate- or formate-utilizing methanogenicbacteria were enumerated by the MPN technique on thebasis of the numbers of tubes positive for methane produc-tion.

Analytical procedures. The weight of the MLVSS in sludgewas determined according to Standard Methods (1). Theweight of MLVSS indicates the amount of organic matter inthe solid fraction of the sludge. TOC was measured by aTOC analyzer (TOC-500; Shimadzu Co.), using potassiumhydrogenphthalate as standard. The methane content of thegas phase in the culture tube was analyzed by a ShimadzuGC-4C gas chromatograph with a stainless steel column(WG-100; 6.4 mm by 1.8 m) and a thermal conductivitydetector. The column, inlet, and detector temperatures were50, 80, and 80°C, respectively. Argon was used as a carriergas. Gas samples (0.2 ml) were collected from the culturetubes by inserting the needle of a pressure-lock syringethrough the butyl rubber stoppers, and each collected gaswas injected into a gas chromatograph. Volatile fatty acids(VFA; acetic acid, propionic acid, butyric acid, and valericacid) were analyzed by a Shimadzu GC-14A gas chromato-graph with a packed glass column (FAL-M 10%; 3 mm by 2.1m) with Shimalite TPA (60/80 mesh) support and a flameionization detector. The column, inlet, and detector temper-atures were 150, 190, and 190°C, respectively. N2 was usedas a carrier gas. Samples (0.9 volume) were acidified with 0.1volume of 1 N HCl, and then the acidified samples (1.0 ,u)were injected into a gas chromatograph.

RESULTS

Chemical characteristics of sludge samples. The chemicalcharacteristics of sludge samples used for analysis of bacte-rial populations are shown in Table 3. Sludge from theUASB reactor in plant A, which was granular-type sludge,and the hydrolyzation reactor in plant B had considerableamounts of MLVSS. Dissolved-TOC and -VFA values werehigh in both of the hydrolyzation reactors for sewage be-cause of the increase of products by solubilization of sewagesolids. On the other hand, VFA values in the methanefermentation reactors were much lower because methaniza-tion of the solubilized solution and the pretreated filtrateproceeded without the accumulation of VFA.CFCs from hydrolyzation reactors. The CFCs obtained

with the Balch-1, m-VL, and SE media in two hydrolyzationreactors (plants A and B) for sewage solids are shown in Fig.1. The CFCs from both of the reactors showed a very slow

VOL. 58, 1992

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

2754 KATAOKA ET AL.

60

50

c 40

50. - 30CE

200 10

10

c

0I-

C00 o:. :o U.

'o _k'3O

Incubation time (day) Incubation time (day)

FIG. 1. CFCs of anaerobic bacteria in hydrolyzation reactors. Colonies were counted with three kinds of media: 0, SE; A, m-VL; andC, Balch-1. The means of colony counts from duplicate roll tubes are plotted.

rate of colony formation. Especially in the case of plant B,the considerable amount of highly concentrated sludge re-tained by the combined ultrafiltration membrane modulecontained large numbers of slow-growing colonies. Most ofthe colonies counted with the SE and m-VL media began toappear after 1 week of incubation, and the counts continuedto increase for up to 5 weeks of incubation. However, inBalch-1 medium, late-appearing colonies were very few. SEmedium gave the highest colony counts, at 2.1 x 108 CFU/ml(3.2 x 1010 CFU/g of MLVSS) and 5.4 x 108 CFU/ml (1.8 x1010 CFU/g of MLVSS) after 5 weeks of incubation inreactors of plants A and B, respectively. This suggests thatslow-growing bacteria are the dominant microflora in hydro-lyzation reactors.CFC from methane fermentation reactors. Figure 2 shows

the CFCs from three kinds of methane fermentation reac-tors, namely, UASB (plant A), fluidized-bed (plant B), andfixed-bed (plant C) types of reactors. Plant C (for soybean-processing wastewater) showed the highest colony counts,whereas plants A and B (for sewage) showed much lower

10

C

0I -

.E_E- -.

CU.o C)o

, _

C o

E -.

bo EU.

co -

0

40

counts. The highest counts after 5 weeks of incubation were5.4 x 107 CFU/ml (1.4 x 109 CFU/g of MLVSS), 7.9 x 107CFU/ml (9.3 x 109 CFU/g of MLVSS), and 4.8 x 108CFU/ml (2.4 x 1011 CFU/g of MLVSS) in plants A, B, andC, respectively. SE medium had the highest counts amongthe three media (not including plant A, which was not testedwith SE medium). The CFC pattern for plant C differedgreatly from those of plants A and B. The counts with thethree media for plant C reached an approximately constantlevel after 3 weeks of incubation; that is, 95% of the totalcounts appeared within 3 weeks of incubation (Fig. 2C). Onthe other hand, the CFCs from plants A and B resembledthose from the hydrolyzation reactors except for the minorcounts of late-appearing colonies. The counts with SE andm-VL media increased smoothly until the second to thirdweek of incubation, whereas counts with Balch-1 mediumcontinued to increase for up to 5 weeks of incubation (Fig.2A and B). The counts with m-VL medium for enumeratingmainly acidogenic bacteria were low in plants A and B (4.8 x107 and 2.0 x 107 CFU/ml at 5 weeks of incubation)

c

5E'0 E

U.Co

o

C)

40 40

Incubation time (day) Incubation time (day) Incubation time (day)

FIG. 2. CFCs of anaerobic bacteria in methane fermentation reactors. Colonies were counted with three kinds of media: 0, SE; A, m-VL;and 0, Balch-1. The means of colony counts from duplicate roll tubes are plotted.

APPL. ENVIRON. MICROBIOL.

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ANAEROBIC MEMBRANE BIOREACTORS 2755

TABLE 4. Enumeration of sulfate reducers and methanogens in three plants

No. of bacteria/ml ina:

Bacteria Plant A Plant B Plant C

HR MFR HR MFR (MFR)

Sulfate-reducersb 3.6 x 107 7.6 x 106 3.0 x 105 1.4 x 106 2.7 x 107Methanogensc 4.0 x 104 1.1 X 107 4.0 x 104 1.1 X 107 NT

a HR, hydrolyzation reactor; MFR, methane fermentation reactor; NT, not tested.b Black colonies were counted on the roll tube with m-Butlin medium after 4 weeks of incubation.c Acetate- or formate-utilizing methanogens were enumerated by the MPN technique with three tubes per dilution; they were cultured with Balch-1 medium.

Values were determined from the presence of methane in sets of three tubes after 4 weeks of incubation.

compared with those for hydrolyzation reactors. The supe-riority of acidogenic bacteria in hydrolyzation reactors isreasonable enough, considering the function of these reac-tors. However, the methane fermentation reactor of plant Ccontained large numbers of acidogenic bacteria (3.4 x 108CFU/ml). In plants A and B, counts with Balch-1 mediumcontaining formate and acetate as a carbon source werealmost the same (5.6 x 107 to 7.6 x 107 CFU/ml) after 5weeks of incubation.The CFCs for the three plants show a pattern of relatively

slow development. The colony-forming rates of anaerobicbacteria in the reactors of plants A and B differ from that inthe methane fermentation reactor of plant C. Most of thecolonies from the hydrolyzation reactors showed a muchslower growth than those from the methane fermentationreactor of plant C. The CFCs also indicate that the bacterialpopulations of the hydrolyzation and methane fermentationreactors differ in their nutritional requirements.Enumeration of sulfate-reducing and methanogenic bacte-

ria. The counts of sulfate-reducing bacteria (sulfate reduc-ers) and methanogenic bacteria (methanogens) in the threeplants are shown in Table 4. Sulfate reducers were enumer-ated by counting black colonies on roll tubes containinglactate as an electron donor. The black colonies began toappear on 10-5- or 10-6-dilution roll tubes after 1 to 2 weeksof incubation, and their numbers reached a maximum after 3weeks of incubation. The proportion of sulfate reducercounts compared with total colony counts on the same rolltube ranged from 0.5 to 28%. The hydrolyzation reactor ofplant A contained considerable numbers of sulfate reducers(3.6 x 107 CFU/ml [5.4 x 109 CFU/g of MLVSS]). Of thethree methane fermentation reactors, that of plant C had thehighest sulfate reducer count (2.7 x 107 CFU/ml [1.3 x 1010CFU/g of MLVSSJ).

Acetate- or formate-utilizing methanogens were enumer-ated by the MPN method. The numbers of acetate-utilizingmethanogens in the methane fermentation reactors werehigh (1.1 x 107 cells per ml), whereas those in the hydrolyz-ation reactors were low (4.0 x 104 cells per ml).Using a method studied previously (19), we observed the

roll tubes with Balch-1 medium from the methane fermenta-tion reactor of plant A directly under an epifluorescencemicroscope (Olympus model BHS-RFK) with a 2x objectiveand a V (BP-405) excitation filter. However, no fluorescentcolony of methanogens was found, although the lower-dilution tubes (10-4 and 10-5) produced trace amounts ofmethane (less than 2.6% in volume).

DISCUSSION

This investigation was an attempt to analyze the bacterialpopulations in characteristic sludge ecosystems arising in

anaerobic membrane bioreactors for treating low-strengthwastewater. We observed the time required for bacterialcolonies to appear on roll tubes and tried to correlate thatwith the physiological characteristics of bacteria present inthe bioreactors. The results of CFC analysis show that thetype of wastewater and the operating conditions of bioreac-tors significantly influence bacterial biomass, bacterial flora(as regards, for example, the proportion of slower-growingbacteria), and nutritional requirements. The present workalso shows that the CFC analysis method is an effective wayof studying the microbiology of anaerobic treatment sys-tems.The CFCs showed a characteristic pattern that differed

depending on the medium. SE medium containing only thesupematant of autoclaved sludge gave the highest counts,with many slower-growing colonies appearing after 10 daysof incubation. Large numbers of bacteria in the sludges,especially the slower-growing bacteria, probably have com-plex requirements for growth. Iannotti et al. (16) stated thatthe long time for development of visible colonies appeared tobe characteristic of anaerobic digesters and that the im-provements in recovery (viable count per total microscopiccount) were due to slower-growing organisms not found inother media. The results of our research support this idea.Our findings show that the anaerobic bacterial populations

for treating sewage (plants A and B) are mainly slow-growingcompared with those for treating soybean-processing waste-water (plant C). The large numbers of slower-growing bac-teria in hydrolyzation reactors seem to arise from the diffi-culty of effecting hydrolysis of sewage solids containinglarge amounts of cellulosic materials. Pfeffer (28) and Klass(20) have reported that in domestic wastewater sewagedigestors, the first-step reaction, namely, hydrolysis of or-ganic solids, is the rate-limiting step in the overall processbecause of the long retention times required for hydrolysis ofthese solids. It has been reported that in sewage sludgedigestors, the number of hydrolytic and fermentative bacte-ria was large (108 to 109/ml of mesophilic sewage sludge) (25,30), whereas that of the cellulolytic bacteria was very small(1.6 x 104 to 9.7 x 105/ml of sewage sludge) (26). Likewise,Hobson and Shaw (14) reported that the total numbers ofanaerobic bacteria and cellulolytic bacteria in domesticdigestor sludge were 2.4 x 107 and 4 x 103/ml, respectively.It is reasonable to consider that the population size ofcellulolytic bacteria is very small in hydrolyzation reactorsfor treating sewage solids. The hydrolyzation reactors em-ployed were operated with a 3- or 5-day hydraulic retentiontime for obtaining sufficient solubilization of sewage solids.The long retention time probably influences aspects of thebacterial flora such as the proportion of slower-growingbacteria. We can further explain the predominance of theslower-growing bacteria in the hydrolyzation reactors by the

VOL. 58, 1992

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

2756 KATAOKA ET AL.

interactions between hydrolytic bacteria and their dependentassociates regarding substrate incorporation. That is to say,the hydrolytic and fermentative bacteria serve as a primaryconsumer of various substrates in influent solids, whereasnonhydrolytic bacteria such as the slower-growing bacteriagrow by utilizing organic compounds and growth factors, forexample, vitamins and amino acids, derived from cellularcomponents produced by death and lysis of cells and metab-olites produced by viable cells. The higher counts obtainedwith SE medium also suggest the presence of syntrophicbacteria. Both the long retention time and the high reactorbiomass concentration by membrane separation seem to berelated to the advance of syntrophic interaction among thespecies.For efficient methane fermentation of wastewaters, it is

essential to enhance the solubilization and methanization ofhigh amounts of solids. In domestic sewage, approximately75% of the organic materials are in the suspended fraction(5). The present investigation employed two different hydro-lyzation processes: (i) a reactor adding enzyme for hydroly-sis of cellulose (plant A) and (ii) a reactor retaining highlyconcentrated sewage solids and bacteria by a membraneseparation unit (plant B). The solubilization rates of the twoprocesses were low, at 40.9% for the volatile SS standard(plant A) and 25% for the TOC standard (plant B). Eachreactor had the following bacterial properties: (i) large num-ber of sulfate reducers (3.6 x 10 /ml; plant A); (ii) largenumbers of slower-growing bacteria (plant B). Municipalsewage usually contains 50 to 200 mg of sulfate per liter, andsulfate reduction can thus proceed during the solid solubili-zation process. In sewage digestor fluids, sulfate reducersare present at a concentration of 107 to 108/ml (31). We cansay from the sulfate reducer counts that sulfate reductionoccurs simultaneously with hydrolysis of sewage solids andprobably favors anaerobic decomposition of the solids. Inorder to improve the low solubilization rate of sewage solids,the introduction of not only solubilization and acidificationbut also of an efficient methanization step with sulfatereduction into the hydrolyzation reactors, like anaerobicdigestion of sewage sludge, is probably effective. In practice,the methane fermentation of highly concentrated sewagesolids within the hydrolyzation reactors combined with amembrane separation module increased the solubilizationrate from 60 to 77% and from 45 to 62% as the volatile SSstandard in plants A and B, respectively (data not shown).We found that the bioreactors studied make characteristic

anaerobic sludge ecosystems containing large numbers ofslow-growing bacteria. The physiological and ecologicalunderstanding of slow-growing bacteria within the highlyconcentrated sludge will likely be increasingly important forreactor efficiency. As for the CFC analysis method, we cansay that it is a useful technique for ascertaining the physio-logical characteristics of populations in anaerobic treatmentsystems, although there are some restrictions due to destruc-tion from complex interactions among species in bacterialaggregates, the presence of a number of noncultivable bac-teria in artificial culture media, and the difficulty of achievingoptimum conditions for diverse bacteria. This techniquerequires further testing in other anaerobic ecosystems toconfirm our proposal.

ACKNOWLEDGMENTS

This research was carried out as part of the National Research andDevelopment Program for New Wastewater Treatment System(Aqua Renaissance '90 Project) being conducted by the Ministry ofInternational Trade and Industry. This study was also supported by

the New Energy and Industrial Technology Development Organiza-tion.

REFERENCES1. American Public Health Association. 1985. Standard methods for

the examination of water and wastewater, 16th ed. AmericanPublic Health Association, Washington, D.C.

2. Ashley, N. V., and T. J. Hurst. 1981. Acid and alkaline phos-phatase activity in anaerobic digested sludge: a biochemicalpredictor of digester failure. Water Res. 15:633-638.

3. Azuma, R., and T. Suto. 1970. Validity of transfer of thetaxonomical position of Corynebacterium pseudopyogenes fromgenus Corynebacterium to genus Actinomyces, p. 493-505. InH. Iizuka and T. Hasegawa (ed.), Proceedings of the FirstInternational Conference on Culture Collections. University ofTokyo Press, Tokyo.

4. Balch, W. E., G. E. Fox, L. J. Magrum, C. R. Woese, and R. S.Wolfe. 1979. Methanogens: reevaluation of a unique biologicalgroup. Microbiol. Rev. 43:260-296.

5. Barbosa, R. A., and G. L. Sant'Anna, Jr. 1989. Treatment ofraw domestic sewage in an UASB reactor. Water Res. 23:1483-1490.

6. Butlin, K. R., M. E. Adams, and M. Thomas. 1949. The isolationand cultivation of sulfate-reducing bacteria. J. Gen. Microbiol.3:46-59.

7. Chung, Y.-C., and J. B. Neethling. 1988. ATP as a measure ofanaerobic sludge digester activity. J. Water Pollut. Control Fed.60:107-112.

8. Chung, Y.-C., and J. B. Neethling. 1989. Microbial activitymeasurements for anaerobic sludge digestion. J. Water Pollut.Control Fed. 61:343-349.

9. De Haast, J., and T. J. Britz. 1987. Characterization of someanaerobic bacteria from the liquid phase of a mesophilic anaer-obic digester fed with a prefermented cheese whey substrate.Microb. Ecol. 14:167-177.

10. Dolfing, J., A. Griffioen, A. R. W. van Neerven, and L. P. T. M.Zevenhuizen. 1985. Chemical and bacteriological composition ofgranular methanogenic sludge. Can. J. Microbiol. 31:744-750.

11. Hattori, T. 1982. Analysis of plate count data of bacteria innatural environments. J. Gen. Appl. Microbiol. 28:13-22.

12. Hattori, T. 1988. The analysis of colony-forming curves, p.57-63. In T. D. Brock (ed.), The viable count: quantitative andenvironmental aspects. Science Tech Publishers, Madison,Wisc.

13. Henson, J. M., P. H. Smith, and D. C. White. 1985. Examinationof thermophilic methane-producing digesters by analysis ofbacterial lipids. Appl. Environ. Microbiol. 50:1428-1433.

14. Hobson, P. N., and B. G. Shaw. 1974. The bacterial populationof piggery-waste anaerobic digesters. Water Res. 8:507-516.

15. Hungate, R. E. 1969. A roll tube method for cultivation of strictanaerobes, p. 117-132. In J. R. Norris and D. W. Ribbons (ed.),Methods in microbiology, vol. 3B. Academic Press, Inc., NewYork.

16. Iannotti, E. L., J. R. Fischer, and D. M. Sievers. 1978. Mediumfor the enumeration and isolation of bacteria from a swine wastedigester. Appl. Environ. Microbiol. 36:555-566.

17. Ishikuri, S., and T. Hattori. 1985. Formation of bacterial colo-nies in successive time intervals. Appl. Environ. Microbiol.49:870-873.

18. Ishikuri, S., and T. Hattori. 1987. Analysis of colony formingcurves of soil bacteria. Soil Sci. Plant Nutr. 33:355-362.

19. Kataoka, N., Y. Tokiwa, and K. Takeda. 1991. Improvedtechnique for identification and enumeration of methanogenicbacterial colonies on roll tubes by epifluorescence microscopy.Appl. Environ. Microbiol. 57:3671-3673.

20. Klass, D. L. 1984. Methane from anaerobic fermentation. Sci-ence 223:1021-1028.

21. Kobayashi, H. A., E. Conway de Macario, R. S. Williams, andA. J. L. Macario. 1988. Direct characterization of methanogensin two high-rate anaerobic biological reactors. Appl. Environ.Microbiol. 54:693-698.

22. Macario, A. J. L., and E. Conway de Macario. 1988. Quantita-tive immunologic analysis of the methanogenic flora of digestors

APPL. ENvIRON. MICROBIOL.

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ANAEROBIC MEMBRANE BIOREACTORS 2757

reveals a considerable diversity. Appl. Environ. Microbiol.54:79-86.

23. Macario, A. J. L., J. F. K. Earle, D. P. Chynoweth, and E.Conway de Macario. 1989. Distinctive patterns of methanogenicflora determined with antibody probes in anaerobic digestors ofdifferent characteristics operated under controlled conditions.Syst. Appl. Microbiol. 12:216-222.

24. MacLeod, F. A., S. R. 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.

25. Mah, R. A., and C. Sussman. 1968. Microbiology of anaerobicsludge fermentation. I. Enumeration of the nonmethanogenicanaerobic bacteria. Appl. Microbiol. 16:358-361.

26. Maki, L. R. 1954. Experiments on the microbiology of cellulosedecomposition in a municipal sewage plant. Antonie van Leeu-wenhoek 20:185-200.

27. Morgan, J. W., L. M. Evison, and C. F. Forster. 1991. Theinternal architecture of anaerobic sludge granules. J. Chem.Tech. Biotechnol. 50:211-226.

28. Pfeffer, J. T. 1968. Increased loadings on digesters with recycleof digested solids. J. Water Pollut. Control Fed. 40:1920-1933.

29. Robinson, R. W., D. E. Aidn, R. A. Nordstedt, M. V. Thomas,and H. C. Aldrich. 1984. Light and electron microscopic exam-

inations of methane-producing biofilms from anaerobic fixed-bed reactors. Appl. Environ. Microbiol. 48:127-136.

30. Toerien, D. F., and M. L. Siebert. 1967. A method for theenumeration and cultivation of anaerobic acid-forming bacteriapresent in digesting sludge. Water Res. 1:397-404.

31. Ueki, A., H. Minato, R. Azuma, and T. Suto. 1980. Enumerationand isolation of anaerobic bacteria in sewage digestor fluids:enumeration of sulfate-reducers by the anaerobic roll tubemethod. J. Gen. Appl. Microbiol. 26:25-35.

32. Ueki, A., E. Miyagawa, H. Minato, R. Azuma, and T. Suto. 1978.Enumeration and isolation of anaerobic bacteria in sewagedigestor fluids. J. Gen. Appl. Microbiol. 24:317-332.

33. Varel, V. H. 1984. Characteristics of some fermentative bacteriafrom a thermophilic methane-producing fermenter. Microb.Ecol. 10:15-24.

34. Yushina, Y., J. Hasegawa, and H. Satoh. 1991. Comparisonstudies on two-phase methane fermentation processes com-bined with membrane separation. Suishitsu Odaku Kenkyuu14:312-321. (In Japanese.)

35. Zeikus, J. G. 1979. Microbial populations in digesters, p. 61-89.In D. A. Stafford, B. I. Wheatley, and D. E. Hughes (ed.),Anaerobic digestion. Applied Science Publishers, Barking, Es-sex, England.

VOL. 58, 1992

on March 23, 2020 by guest

http://aem.asm

.org/D

ownloaded from