Acetate Production from Whey LactoseUsing Co-Immobilized Cells ofHomolactic and Homoacetic Bacteria in aFibrous-Bed Bioreactor

Yan Huang, Shang-Tian Yang

Department of Chemical Engineering, The Ohio State University, 140 West19th Avenue, Columbus, Ohio 43210; telephone: 614-292-6611; fax:614-292-3769; e-mail: yang.15@osu.edu

Received 14 August 1997; accepted 1 May 1998

Abstract: Acetate was produced from whey lactose inbatch and fed-batch fermentations using co-immobilizedcells of Clostridium formicoaceticum and Lactococcuslactis. The cells were immobilized in a spirally woundfibrous sheet packed in a 0.45-L column reactor, withliquid circulated through a 5-L stirred-tank fermentor. In-dustrial-grade nitrogen sources, including corn steep li-quor, casein hydrolysate, and yeast hydrolysate, werestudied as inexpensive nutrient supplements to wheypermeate and acid whey. Supplementation with either2.5% (v/v) corn steep liquor or 1.5 g/L casein hydrolysatewas adequate for the cocultured fermentation. The over-all acetic acid yield from lactose was 0.9 g/g, and theproductivity was 0.25 g/(L h). Both lactate and acetate athigh concentrations inhibited the homoacetic fermenta-tion. To overcome these inhibitions, fed-batch fermenta-tions were used to keep lactate concentration low and toadapt cells to high-concentration acetate. The final ac-etate concentration obtained in the fed-batch fermenta-tion was 75 g/L, which was the highest acetate concen-tration ever produced by C. formicoaceticum. Even atthis high acetate concentration, the overall productivitywas 0.18 g/(L h) based on the total medium volume and1.23 g/(L h) based on the fibrous-bed reactor volume. Thecells isolated from the fibrous-bed bioreactor at the endof this study were more tolerant to acetic acid than theoriginal culture used to seed the bioreactor, indicatingthat adaptation and natural selection of acetate-tolerantstrains occurred. This cocultured fermentation processcould be used to produce a low-cost acetate deicer fromwhey permeate and acid whey. © 1998 John Wiley & Sons,Inc. Biotechnol Bioeng 60: 498–507, 1998.Keywords: acetic acid fermentation; whey; fibrous-bedbioreactor

INTRODUCTION

Acetic acid is an important organic chemical, with an annualproduction rate of∼ 4.68 billion pounds in 1995 in theUnited States. Current commercial production of glacial

acetic acid is exclusively by petrochemical routes (Ghoseand Bhadra, 1985). One important industrial use of aceticacid is to produce acetate deicers, including calcium mag-nesium acetate (CMA) as a noncorrosive road deicer (Chol-lar, 1984) and potassium acetate and sodium acetate as air-port runway deicers. These acetate deicers are environmen-tally friendly (Fritzsche, 1992), but are expensive to use(Bryan, 1992). Thus, there has long been an interest inproducing low-cost acetic acid and acetate from fermenta-tion of glucose (Marynowski et al., 1985; Parekh andCheryan, 1990, 1991; Wiegel et al., 1991), woody biomass(Wise and Augenstein, 1988), and cheese whey (Yang et al.,1992a).

Whey is a byproduct from the manufacture of cheese andcasein. It contains about 45 g/L lactose, 10 g/L protein, 10g/L salts, and 0.1–0.8 g/L lactic acid. The biological oxygendemand (BOD) of whey is high,∼ 40,000 mg/L. Some largedairies spray-dry whey to whey powder, produce whey pro-tein concentrate from sweet whey by ultrafiltration, andproduce lactose from whey permeate by crystallization.These whey products are used in human foods and animalfeed; however, they account for less than∼ 50% of the totalwhey solids produced in the United States (Yang and Silva,1995). Currently, there are still large amounts of whey by-products in various forms, including acid whey, sweet wheypermeate, and de-lactose whey permeate for which new usesmust be found, otherwise these byproducts will be treated aspollutants.

Whey has been extensively studied as a fermentation sub-strate for producing lactic, propionic, and acetic acids. How-ever, its use for anaerobic acetic acid fermentation has beenlimited because no homoacetogen can directly ferment lac-tose to acetate. Recently, the feasibility of producing acetatefrom whey fermentation using a coculture consisting of ho-molactic and homoacetic bacteria has been studied in free-cell batch cultures (Tang et al., 1988) as well as in immo-bilized-cell continuous cultures (Yang et al., 1992). Thehomolactic and homoacetic bacteria sequentially convertlactose to lactate and then to acetate, with an overall aceticacid yield from lactose of∼ 0.9 g/g. However, in these

Correspondence to:S.T. YangContract grant sponsors: Department of Transportation, Federal High-

way Administration; New York State Energy Research and DevelopmentAuthority

Contract grant number: DTFH61-93-C-00013

© 1998 John Wiley & Sons, Inc. CCC 0006-3592/98/040498-10

studies, whey permeate was supplemented with largeamounts of yeast extract and trypticase. Without these nu-trient supplements, the fermentation was slow, and conver-sion of whey lactose to acetate was also low. Yeast extractand trypticase are expensive to use in large-scale productionof acetate from whey; thus, inexpensive nutrient supple-ments, such as corn steep liquor and industrial-grade yeastand casein hydrolysates, were studied in this work.

The economical use of a fermentation route for acetateproduction also depends on the reactor productivity and thefinal product concentration achievable from the fermenta-tion (Busche et al., 1982). A higher fermentation produc-tivity usually can be achieved with an immobilized-cell bio-reactor; however, conventional immobilized-cell bioreac-tors usually suffer from productivity loss over time, whenthe cells are used continually or repeatedly in continuous orfed-batch fermentations. To overcome these problems, wehave previously developed a fibrous-bed, immobilized-cellbioreactor for propionic acid and lactic acid fermentations(Silva and Yang, 1995; Yang et al., 1994; 1995). The fi-brous-bed bioreactor gave stable, high-rate production ofthe fermentation product for long periods because of thehigh density of active cells maintained in the fibrous bed.Good fermentation results were also obtained with nonster-ile whey permeate and acid whey without nutrient supple-ments when the fermentation was performed in the fibrous-bed bioreactor, but not in the conventional free-cell batchfermentation. A higher product concentration was also ob-tained with the fibrous-bed bioreactor operated in a recyclefed-batch mode (Yang et al., 1995). The fibrous-bed biore-actor has also been studied for continuous production ofacetate from lactate and lactose in synthetic media (Yang etal., 1992a). However, the highest acetate concentration everattained in previous studies was only∼ 40 g/L, which isconsiderably lower than the 80 g/L or higher attained inaerobic vinegar fermentation (Ebner and Follmann, 1983;Kondo and Kondo, 1996) and the anaerobic homoaceto-genic fermentation of glucose by an acetate-tolerant strainof C. thermoaceticum(Parekh and Cheryan, 1994a).

The main objective of this work was to study the feasi-bility of producing a high concentration of acetate brothfrom whey fermentation with minimal nutrient supplementsby using the fibrous-bed bioreactor. The fermentation ki-netics was also studied to elucidate the effects of lactate andacetate on the cocultured fermentation. Fed-batch operationwas studied to evaluate its effectiveness in adapting cells totolerate and produce a high acetate concentration. The im-portance of controling lactate formation from lactose andmaintaining a low lactate concentration during the cocul-tured fermentation was also studied and is discussed in thisarticle.

MATERIALS AND METHODS

Cultures and Media

The homolactic acid bacterium,Lactococcus lactis(OSUstock culture #588) and the homoacetogenic bacterium,

Clostridium formicoaceticum(ATCC 27076) were used inthis study. The stock cultures of these bacteria were main-tained in synthetic media containing either lactose or lactateas the carbon source. The compositions of these media canbe found elsewhere (Tang et al. 1988).

Fresh sweet whey permeate (WP) and frozen concen-trated acid whey used in this study were obtained fromBrewster Dairy, Inc. (Brewster, OH) and Kraft Foods, Inc.(Chicago, IL), respectively. They were stored in a cold roomat 4°C until use. The storage time was usually less than 1week. WP contained 30 to 45 g/L lactose, 1.5 to 4.2 g/Llactate, and less than 0.5 g/L protein. Unless otherwisenoted, WP was filter-sterilized using a 0.2-mm (pore size)membrane filter. The concentrated acid whey (∼ 40% totalsolids) was diluted with tap water to a lactose concentrationof ∼50 g/L or lower, and was then sterilized by autoclavingat 121°C for 20 min. The diluted acid whey also contained∼10 g/L lactic acid and∼3 g/L protein.

To promote the fermentation, WP and acid whey weresupplemented with some nitrogenous sources, includingyeast extract (Difco), trypticase (Difco), casein hydrolysate(UBC and Red Star), yeast hydrolysate (Red Star), and cornsteep liquor (CSL). The corn steep liquor was received in aconcentrated form (∼ 45% total solids) from Cargill’s cornwet-milling plant in Eddyville, Iowa. Before use, it wasdiluted with tap water by fourfold, heat sterilized at 121°Cfor 20 min, and the supernatant was then removed for use asa nutrient supplement to WP and acid whey in various vol-ume fractions. CSL also contained small amounts of glu-cose and lactic acid.

Formulation of Whey Media

Previous studies have shown that whey permeate was nu-tritionally sufficient forL. lactis,but not forC. formicoace-ticum (Tang et al., 1988). To develop a proper mediumformulation for whey permeate as a feedstock for acetateproduction, the effect of various nutrient supplements on thegrowth of C. formicoaceticumwas studied in serum tubes.Ten milliliters of whey media (pH 7.6) containing 10 g/Llactate and various amounts and types of nutrient supple-ments were placed in each serum tube. After inoculation,cell growth was followed by measuring the optical densityof the culture tube in a spectrophotometer (Sequoia-Turner,Model 340) at 660-nm wavelength. The final medium pHand acetate content were also analyzed. After appropriatenutrient supplementation was identified, the whey mediawere tested in the cocultured fermentations.

Kinetic Studies

Batch cultures ofC. formicoaceticumgrown in syntheticlactate media containing various amounts of lactate and ac-etate were studied to evaluate the effects of lactate andacetate on cell growth. These cultures were grown in serumtubes and cell growth was followed by measuring the opti-cal density at 660 nm with a spectrophotometer.

HUANG AND YANG: ACETATE PRODUCTION FROM WHEY 499

Fermentation

Bioreactor Construction and Operation

The immobilized-cell bioreactor was made of a glass col-umn packed with spiral-wound terry cloth. Detailed descrip-tion of the reactor construction has been given elsewhere(Yang et al., 1992a). The column reactor itself had a work-ing volume of ∼ 450 mL, and was connected to a 5-Lstirred-tank fermentor (Marubishi MD-300) through a recir-culation line. A positive pressure (3∼ 5 psig) of N2/CO2 gasmixture was applied to the fermentor head space to maintainanaerobic condition in the system. Unless otherwise noted,the entire reactor system contained∼3.5 L of the mediumand was maintained at 37°C and pH 7.5 by adding6NNaOH solution. The reactor was operated with a high re-circulation rate of∼300 mL/min in either batch or fed-batchmode. The reactor was first inoculated withC. formicoace-ticumand allowed to grow for several weeks in the syntheticmedium containing lactate as the carbon source. During thisperiod, a high cell density of the homoacetogen was immo-bilized in the fibrous bed. The reactor was then inoculatedwith L. lactisand provided with yeast extract supplementedwhey permeate as the substrate. This sequential inoculationprocedure was used to ensure that the slower-growingC.formicoaceticumwould not be outgrown by the faster-growing L. lactis in the cocultured fermentation. After twobatches, both bacteria were properly immobilized in thefibrous bed.

Batch Fermentation

Batch fermentations of whey permeate were studied first toevaluate the feasibility of using corn steep liquor and otherindustrial-grade nitrogen sources as an economical nutrientsupplement to promote the fermentation. In the batch study,3 L of whey medium were used in the 5-L fermentor. Whenone batch fermentation was completed, the fermented me-dium was replaced with a fresh medium to start a new batch.The immobilized cells in the fibrous bed were repeatedlyused in subsequent batch fermentations. Liquid sampleswere taken from the fermentor at proper time intervals andwere stored frozen until analysis.

Fed-Batch Fermentation

Fed-batch fermentations also were performed to adapt theimmobilized cells to high acetate concentrations and to pro-duce a maximum acetate concentration from whey lactoseby the cocultured fermentation. Lactose, in either a concen-trated solution or in powder form, was added to the fermen-tor when both lactose and lactate in the whey medium werenearly depleted. Lactose addition was repeated for severaltimes until the fermentation ceased to utilize the substrate orto produce the final product, acetate. The immobilized cellsin the fibrous bed were repeatedly used in subsequent fed-batch fermentations.

Assay Methods

The concentrations of lactose, lactate, and acetate in thefermentation broth were analyzed with high performanceliquid chromatography (HPLC). The HPLC system con-sisted of an injector (Waters U6K), a pump (Waters 6000A),an organic acid analysis column (Bio-Rad HPX-87H), acolumn heater (Bio-Rad), a reflective index detector (Wa-ters 410), and an integrator (Varian 4270). Details can befound elsewhere (Yang et al., 1992a). The cell density in thefibrous-bed bioreactor at the end of the study was deter-mined by removing the fibrous bed and washing off all thecells from the fibrous matrix. The OD value and the liquidvolume of the washing water were then measured to calcu-late the total cell dry weight.

RESULTS

Whey Medium Formulation

For C. formicoaceticum,whey permeate without nutrientsupplements gave poor cell growth. Preliminary screeningexperiments indicated that only yeast extract (Difco)supplementation was needed for good cell growth in wheypermeate, whereas all other nutrients (e.g., trypticase, vita-mins, trace metals, and mineral salts) used in the syntheticmedium did not affect cell growth in whey permeate (datanot shown). Yeast extract contained large amounts of or-ganic nitrogen (amino acids and peptides), vitamins, min-erals, and other growth factors, but is also relatively expen-sive. Therefore, several industrial-grade yeast hydrolysates(at 5 g/L), casein hydrolysate (at 5 g/L), and corn steepliquor (at 10% v/v) were also tested for their effect on cellgrowth. The results showed that all these supplements gavecomparable cell growth rates to that from the control ex-periment using the more expensive yeast extract (Difco) inthe synthetic medium.

Casein hydrolysate (CA) and corn steep liquor (CSL)were further evaluated to determine their appropriatesupplement levels to whey permeate and their effects on thegrowth of C. formicoaceticum.As shown in Table I, bothCA (at 1.5 g/L) and CSL (at 0.8% v/v) were as good as yeastextract (at 2 g/L) in providing good cell growth and acetateproduction in whey permeate. In general, the fermentationwas better at higher supplement levels.

The results were further confirmed with batch fermenta-tions of acid whey, which was prefermented with homolac-tic bacteria to convert lactose to lactate. About the samefermentation rate and acetate yield were obtained from thesynthetic medium and acid whey supplemented with 2 g/Lyeast extract and 2 g/L trypticase, 2.5% (v/v) corn steepliquor, or 1.5 g/L casein hydrolysate (data not shown). Onthe other hand, acid whey without nutrient supplementsyielded slow growth and low acetate production.

Batch Fermentation

Figure 1 shows two typical batch fermentations of wheypermeate supplemented with 1.5 g/L casein hydrolysate and

500 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 60, NO. 4, NOVEMBER 20, 1998

2.5% (v/v) corn steep liquor, respectively. As shown in thisfigure, lactose was converted to lactate and then to acetateby the coculture, with an overall acetic acid yield fromlactose of∼ 0.9 g/g. The reactor productivity was∼ 0.24g/(L h) based on the total medium volume in the fermenta-

tion. As shown in Table II, similar results were also ob-tained with other whey media, including supplemented acidwhey. In all cases studied, supplementation with either 1.5g/L CA or 2.5% CSL was sufficient to achieve fast fermen-tation and high conversion. The fermentation was poorwhen only 1% CSL was added to whey permeate. The op-timal supplement level must be determined with more ex-periments and based on cost analysis. However, these ex-periments showed that efficient acetate production fromwhey lactose could be achieved with sweet whey permeateand acid whey supplemented with small amounts of inex-pensive corn steep liquor or industrial-grade yeast and ca-sein hydrolysates using the fibrous-bed bioreactor operatedin batch mode.

Effects of Lactate

Lactate is the substrate forC. formicoaceticum.When thelactate concentration was lower than 20 g/L, the specificgrowth rate of the homoacetogen increased with increasingthe lactate concentration in the medium. However, at higherlactate concentrations (>20 g/L), lactate inhibited the cellgrowth in the serum-tube study (data not shown).

Figure 2 shows the effect of lactate accumulation on ac-etate production in the cocultured fermentation. In general,lactate production by the homolactic bacteria was muchfaster than acetate formation from lactate by the homoace-togen. Therefore, when the initial lactose concentration washigher, more lactate accumulation occurred during the co-cultured fermentation. The high concentration of accumu-lated lactate negatively affected acetate production in thecocultured fermentation. As can be seen in Figure 2, theacetate production rate decreased significantly when largeamounts of lactate accumulated in the medium, indicating asubstrate inhibition toC. formicoaceticum.The inhibitioneffect of lactate on cell growth was also observed at con-centrations higher than 20 g/L in the cell growth kineticstudy. Therefore, it is important to keep the lactate concen-tration lower than 20 g/L to allow good acetate productionrate. This would limit the initial lactose concentration in thewhey media unless balanced lactate formation and con-sumption rates could be otherwise maintained throughoutthe cocultured fermentation.

Lactate accumulation and inhibition were more severewhen there was less nutrient supplementation (Fig. 2C).This was because the homoacetic fermentation dependedmore on additional nutrients than the homolactic fermenta-tion. At low supplement levels, the homoacetic fermentationwould be much slower than the homolactic fermentation,resulting in more lactate accumulation in the coculturedfermentation. Substrate (lactate) inhibition was also foundwith continuous cultures (Yang et al., 1992) and other ho-moacetic fermentation studies withC. thermoaceticum(Brumm, 1988). Therefore, to attain higher acetate concen-trations, fed-batch fermentation was recommended andstudied.

Table I. Growth test ofC. formicoaceticumin whey permeate with vari-ous nutrient supplements.

Nutrient supplement Growtha Final pH Acetate (g/L)

None +/− 5.85 0.12

Yeast extract, 2 g/L ++ 5.38 1.08

Corn steep liquor,0.1% (v/v) + 5.80 0.210.2% + 5.61 0.370.4% ++ 5.36 0.830.8% ++ 5.49 1.081.6% ++ 5.37 1.21

Casein hydrolysate,0.25 g/L +/− 5.32 0.100.5 g/L + 5.27 0.101.0 g/L ++ 5.20 0.541.5 g/L ++ 5.38 1.50

a+/− 4 cell growth was not significant; +4 significant cell growth; ++4 good cell growth.

Figure 1. Kinetics of acetate production from lactose by the coculturedfermentation at pH 7.5, 37°C.(A) whey permeate supplemented with 1.5g/L casein hydrolysate,(B) whey permeate supplemented with 2.5% (v/v)corn steep liquor (40% T.S.).

HUANG AND YANG: ACETATE PRODUCTION FROM WHEY 501

Fed-Batch Fermentation

Figure 3 shows two fed-batch fermentations of whey per-meate supplemented with 2.5 % corn steep liquor. As shownin this figure, steady acetate production to reach a highacetate concentration was achieved in the fed-batch fermen-tation. Lactate accumulation and inhibition were reduced toa minimal level, if any, by periodic addition of lactose. Thehighest acetate concentration reached in the fermentationwas∼75 g/L when the fermentation stopped, probably be-cause of product (acetate) inhibition. However, this aceticacid concentration is the highest level ever produced byC.formicoaceticumin either pure culture using lactate as thesubstrate or coculture withL. lactisusing whey or lactose asthe substrate.

The volumetric productivities and acetate yields were de-termined for batches corresponding to each of the five lac-tose addition periods during the fedbatch fermentation. Asalso shown in Figure 3, acetate productivity gradually de-creased with time as the acetate concentration increased to70 g/L. The productivity decreased dramatically at the ac-etate concentration higher than 70 g/L, indicating a strongproduct inhibition at this level. The overall productivitybased on the total liquid volume was 0.176 g/(L h) for the3.5-L run and 0.245 g/(L h) for the 2-L run. Acetate yieldremained almost constant throughout the fed-batch fermen-tation. The overall acetic acid yield from lactose was∼ 0.9g/g, similar to the ones from batch fermentations.

The second fedbatch fermentation (Fig. 3B) was fasterthan the first one (Fig. 3A), because the total medium vol-ume was smaller in the second one. Based on the fibrous-bed bioreactor volume, the overall acetate productivity in

the fed-batch fermentation was 1.23 g/(L h) for the first oneand 1.13 g/(L h) for the second one. The similar reactorproductivities indicate that most reactions occurred in thefibrous bed bioreactor, instead of in the fermentor. This isconsistent with the fact that most cells were immobilized inthe fibrous bed. It was found at the end of this study that lessthan 20% of the total cell population was present as sus-pended cells in the fermentation broth.

It should be mentioned that high acetate concentrationsalso inhibited the homolactic bacteria, as can be seen fromthe decreased lactose consumption rate at increased acetateconcentrations (Fig.3). However, this inhibition should nothave any negative effect on the overall fermentation ratebecause the homolactic acid fermentation was still fasterthan the homoacetic acid fermentation. In fact, this inhibi-tion might have helped to balance the formation and con-sumption of lactate in the cocultured fermentation by slow-ing down the homolactic fermentation to a rate closer to thehomoacetic fermentation.

DISCUSSION

Nutrient Supplements

As shown in this study, either CA or CSL can be used tosupplement whey permeate or acid whey for their efficientuse as economical feedstocks for acetate production by thecocultured fermentation. To our best knowledge, this is thefirst successful demonstration of anaerobic acetate produc-tion with a high yield from an industrial-grade feedstock.All previous reported anaerobic homoacetic fermentationsused synthetic media containing a high concentration of

Table II. Fermentations with various nutrient supplements.

MediumInitial lactose/lactate

conc. (g/L)Final acetateconc. (g/L)

Acetate yielda

(g/g)Productivity

g/(L h))

Synthetic lactate medium2 g/L YE (Difco) 37.5 30.2 0.95 0.2075 g/L YE (Amberexb) 28.8 25.4 0.90 0.221

Whey permeate with:1.5 g/L CA 36.7 30.5 0.92 0.2181% CSL 51.2 21.8 0.84 0.1352.5% CSL 34.3 29.8 0.87 0.2522.5% CSL 27.0 25.5 0.92 0.2642.5% CSL + 1.5 g/L CA 37.3 30.3 0.86 0.309

Acid whey with:2.5% CSL 21.0 20.4 0.90 0.231

Fed-batch fermentationsWP w/2.5% CSL; 3.5 L 34.3 74.8 0.89 0.1762 L 27.0 73.5 0.91 0.245

aAcetate yield was based on the net production of acetate divided by net consumption of substrate(lactose and lactate).

bAmberex 1009 (Red Star, Milwaukee, WI) is a Brewer’s yeast extract.Note.YE: yeast extract; CA: Casamino acid or hydrolyzed casein powder; CSL: corn steep liquor

(with 40% total solids); WP: whey permeate.

502 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 60, NO. 4, NOVEMBER 20, 1998

yeast extract (Shah and Cheryan, 1995), which is cost-prohibitive for industrial fermentation. The low-cost nutri-ent supplements and feedstocks are important to the eco-nomical production of a low-cost acetate deicer.

However, there are cautions in using the relatively crudeindustrial feedstocks. In this work, we encountered somecontamination problems with corn steep liquor as a supple-ment to whey permeate. Frozen-stored CSL was difficult tosterilize because it contained a high number of bacterialspores. Incomplete sterilization of the medium resulted inpoor fermentations due to contamination by a butyrate pro-

ducer, which outgrew the homoacetogen in the contami-nated batch fermentation and resulted in a low acetate yieldof only ∼ 0.65 g/g (data not shown). The contaminant wasprobably a spore-formingClostridium, because its sporeswere hard to inactivate in autoclaving. The butyrate pro-duced by this organism also inhibited the homoacetic bac-terium. In this situation, CSL must be autoclaved twice,with a 1 dperiod between the two autoclavings to allow thespores to germinate. However, this stringent sterilizationprocedure was not required when fresh CSL was used inpreparing the whey media.

Fibrous-Bed Bioreactor

Comparison to Other Studies

There are many advantages to use the coculture and thefibrous-bed bioreactor in the fermentation. Under fed-batchfermentation conditions, a high acetate concentration wasproduced from the immobilized-cell fermentation. Thehighest acetate concentration previously obtained from freecells ofC. formicoaceticumgrown on lactate was only∼ 40g/L (Yang et al., 1992a). In this work, 75 g/L of acetateconcentration and a reactor productivity of 1.23 g/(L h)were obtained. Even higher productivity is possible with

Figure 2. Effects of lactate formation and nutrient concentration on ac-etate production in the cocultured fermentation at pH 7.5, 37°C.(A) WP(27 g/L lactose) supplemented with 2.5% (v/v) CSL,(B) WP (41 g/Llactose) supplemented with 2.5% (v/v) CSL,(C) WP (50 g/L lactose)supplemented with 1% (v/v) CSL.

Figure 3. Kinetics of fed-batch fermentations of whey permeate supple-mented with 2.5% (v/v) corn steep liquor at pH 7.5, 37°C.(A) 3.5-L totalmedium volume,(B) 2-L total medium volume.

HUANG AND YANG: ACETATE PRODUCTION FROM WHEY 503

higher nutrient supplements and a higher cell density in thefibrous-bed bioreactor.

Parekh and Cheryan (1994b) reported that 83∼ 100 g/Lof acetate was produced from glucose by a mutant strain ofC. thermoaceticumin a fed-batch fermentation. The acetateyield was only 0.74∼ 0.8 g/g glucose and the productivitywas 0.6∼ 0.85 g/(L h). Improved productivity was alsoreported by the same research group with a two-stage con-tinuous fermentation process with cell recycle and at alower acetate concentration of∼35 g/L (Shah and Cheryan,1995). However, a complex synthetic medium with excessnutrients and cell recycle with a hollow-fiber membranefilter were used to maintain a high viable-cell density intheir study. Similarly, high acetic acid concentration (up to90 g/L) and productivities were also produced from aerobicfermentation of ethanol byAcetobacter acetiin a repeatedfed-batch culture with cell recycle (Park et al., 1991). How-ever, cell viability was very low at acetic acid concentra-tions higher than 60 g/L. A two-stage fermentation processwas thus suggested to partially overcome this problem (Itoet al., 1991). However, vinegar fermentation has a low ace-tic acid yield from glucose,∼0.6 g/g or lower, and requiresextensive aeration. Furthermore, using a hollow-fiber mem-brane filter for cell recycle to achieve a high cell density andreactor productivity in these studies could be a major prob-lem for long-term operation and process scale-up becausedead cells accumulate and filtrate flux declines with time.The fibrous-bed bioreactor gave a comparable fermentationperformance and would not be subjected to membrane foul-ing and other problems associated with the hollow fiber.

The good fermentation performance of the fibrous-bedbioreactor can be attributed to its abilities to maintain a highcell density (∼30 g/L) in the bioreactor and to adapt the cellsto tolerate a high acetate concentration. With the high celldensity in the fibrous bed, fermentation no longer dependson fast cell growth in the batch reactor to achieve a highproductivity. Consequently, a lower nutrient supplementlevel can be used compared to that required for the free-cellfermentation system. A previous study by Tang et al. (1988)used 8 g/L yeast extract to supplement whey permeate, butcould only get∼20 g/L acetate with 50% substrate conver-sion. Also, the high cell density allowed the reactor to betterresist contamination. As mentioned before, improper steril-ization of CSL could seriously reduce acetate production.However, the reactor with a high cell density in the fibrousbed could be easily recovered from the contamination bychanging the contaminated medium to a new, properly ster-ilized medium in the fermentor. The fermentation rate andacetate yield returned to normal levels after replacing themedium (data not shown). This indicated that the fibrous-bed bioreactor can tolerate a low contaminant level. Previ-ous studies with this type of bioreactor have also indicatedthat nonsterile media could be used as the feed to the reactoras long as the contaminant level in the medium was not highand would not take over the bioreactor (Silva & Yang, 1995;Yang et al., 1995).

Acetate Tolerance

Another advantage of the fibrous-bed bioreactor is its abilityto quickly adapt cultures to tolerate a high acetate concen-tration. At the end of this study, cells removed from thefibrous-bed bioreactor were tested for their growth toleranceto acetate under suspended-cell culture condition and theresults were compared with those from the original cultureused to seed the bioreactor. As shown in Figure 4, cells fromthe fibrous-bed bioreactor had a much higher tolerance toacetate than the original culture, as indicated by their higherspecific growth rate and lower sensitivity to acetate inhibi-tion. A previous study also showed that the highest acetateconcentration allowing for cell growth was∼ 50 g/L (Tanget al., 1989). These results suggest that the fibrous-bed bio-reactor provided an environment suitable for adapting andenriching acetate-tolerant strains. Similar results were alsofound with propionic acid fermentation performed in thefibrous-bed bioreactor (Yang et al., 1995).

It is desirable to use acetate-tolerant mutants to produce ahigh acetate concentration in the fermentation. Mutantsused in fermentation were usually induced with chemicalmutagenic agents or ultraviolet light, and then selected withselective enrichment procedures (Parekh and Cheryan,1991). However, this procedure is tedious and could take along time to find a useful mutant. Reed et al. (1987) devel-oped a continuous selection process in a bioreactor envi-ronment incorporated with UV irradiation and a chemicalmutagen, which killed most of cells in the bioreactor. In thiswork, acetate tolerant strains were obtained in the bioreactorwithout using mutagens. Only a selection pressure (that is ahigh acetate concentration) was needed to obtain the mutantin a relatively short period. It is speculated that the envi-ronment provided in the fibrous bed may be an importantfactor in getting the acetate-tolerant mutant.

To further demonstrate the ability of the fibrous-bed bio-

Figure 4. Effects of acetic acid concentration on the specific growth ratesof the adapted cells from the bioreactor and the original culture ofC.formicoaceticum.(m: data from Tang et al., 1989;n: data from this work;detailed experimental procedures can be found in Tang et al., 1989)

504 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 60, NO. 4, NOVEMBER 20, 1998

reactor to adapt cells to high acetate concentrations, a newbioreactor was set up and inoculated with the original cul-ture of C. formicoaceticum.The synthetic lactate mediumwas used in the fermentation to grow the cells in the fed-batch mode. Concentrated lactate solution with yeast extractwas added to the fermentor when lactate was about to bedepleted. As shown in Figure 5, after the third batch, acetateconcentration reached 60 g/L. At this point, more lactateand some acetate was added to the medium to raise theacetate concentration to∼75 g/L. As also shown in Figure 5,the acetate formation at this high acetate level was slow, andthe highest acetate concentration reached was∼ 80 g/L.Figure 6 shows the effects of acetate concentration on thereactor productivity and acetate yield from lactate, and onthe specific growth rate obtained from the growth kineticsstudy. It is clear that cells had good tolerance to acetate untilthe acetate concentration was higher than 7%. These resultsare consistent with those from the cocultured fermentationwith lactose as the substrate. The acetic acid yield fromlactate was∼ 0.95 g/g lactic acid and also was not affectedby the acetate concentration.

Anaerobiosis

Homoacetogens are strict anaerobes; they cannot grow wellin the presence of trace oxygen. In pure culture,C. formi-

coaceticumand other homoacetogens, such asC. thermo-aceticum,must be cultivated in a medium in which theoxygen content has been purged with nitrogen gas and theredox potential reduced to a negative value with a reducingagent, such as cysteine. This anaerobiosis growth require-ment could be a serious problem in large-scale homoaceticfermentations. This problem, however, was overcome byusing the cocultured fermentation. Co-immobilization with

Figure 5. Kinetics of homoacetic fermentation of lactate byC. formicoaceticumimmobilized in the fibrous bed bioreactor. Fermentations were performedin the sequence of gradually increasing acetate concentration (from A to D) to adapt the cells to tolerate and produce high acetate concentrations from lactate.

Figure 6. Effects of acetic acid concentration on acetic acid yield, pro-ductivity, and the specific growth rate of the homoacetogen grown in thefibrous-bed bioreactor. (Yield and productivity data were calculated fromdata shown in Figure 5; specific growth rate data were from Figure 4.)

HUANG AND YANG: ACETATE PRODUCTION FROM WHEY 505

the homolactic bacterium in the fibrous bed removed therequirement for gas purging or chemically reducing thewhey medium. The homolactic acid bacterium, which is notas sensitive to oxygen as the homoacetogen, might haveremoved oxygen, thus reducing the redox potential to a levelsuitable for the strict anaerobe, homoacetogen.

Scanning electron micrographs (Fig. 7) show that thehomolactic bacteria (cocci) and the homoacetic bacteria(long rods) co-lived in the fibrous bed. Most of them colo-nized together on the surface of the fiber. Although it cannotbe easily seen in Figure 7, other SEM photos showed thatmost ofC. formicoaceticumcells were attached directly tothe fiber surface and were then covered byL. lactis cells,which did not attach to the fiber surface as well. This kindof colonization is consistent with the immobilization se-quence used in the reactor start up and explains why thehomoacetogens can survive in the medium without reducingits redox potential. Apparently, these two bacteria have de-veloped an intimate relationship for both to survive well inthe bioreactor.

CONCLUSIONS

Whey permeate and acid whey can be used as feedstocks toproduce acetate using the coculture immobilized in the fi-brous-bed bioreactor with corn steep liquor, casein hydro-lysate, or industrial-grade yeast extracts as nutrient supple-ments. About 0.9 g of acetic acid was produced from eachgram of lactose consumed in the cocultured fermentation. Ahigh acetate concentration of 75 g/L was attained in fed-batch fermentation. Fed-batch operation adapted the cultureto tolerate high acetate concentrations and induced and en-riched acetate-tolerant strains in the bioreactor. The fermen-tation rate decreased only slightly with increasing acetateconcentration up to∼ 70 g/L, then dropped dramaticallybeyond 70 g/L. At 75 g/L acetate, the overall productivitywas∼ 0.23 g/(L h) based on the total medium volume and1.13 g/(L h) based on the fibrous-bed reactor volume. Themaximum acetate concentration that the bacteria could tol-erate was∼ 80 g/L. The homoacetic bacteria were alsoinhibited by a high lactate concentration (> 20 g/L) accu-mulated in the cocultured fermentation. The lactate inhibi-tion problem was also overcome by using fed-batch fermen-tation to maintain a balanced lactate formation and con-sumption.

In this work, lactose was used as additional substrate inthe fed-batch fermentation study. For industrial production,concentrated acid whey or de-lactose whey permeate, whichhave high total solids and lactose concentration (up to∼ 200g/L), can be used in the fed-batch fermentation process. Ourpreliminary economic analysis indicates that acetate deicerscan be produced from whey lactose using this fermentationprocess at a cost significantly lower than that of the presentcommercial CMA deicer produced from glacial acetic acid(Yang and Chollar, 1996).

Technical assistance and SEM photos provided by Dr. Hui Zhuare gratefully acknowledged.

References

Brumm, P. J. 1988. Fermentation of single and mixed substrate by theparent and an acid-tolerant, mutant strain ofClostridium thermoace-ticum.Biotechnol. Bioeng.32: 444–450.

Bryan, W. L. 1992. Research to reduce the cost of calcium magnesiumacetate, Ch. 18. In: F. M. D’Itri (ed.), Chemical deicers and the envi-ronment. Lewis Publishers, Boca Raton, FL.

Busche, R. M., Shimshick, E. J., Yates, R. A. 1982. Recovery of acetic acidfrom dilute acetate solution. Biotechnol. Bioeng. Symp.12: 249–262.

Chollar, B. H. 1984. Federal Highway Administration research on calciummagnesium acetate—an alternative deicer.Public Roads47: 113.

Ebner, H., Follmann, H. 1983. Acetic acid. In: H. J. Rehm and G. Reed(eds.), Biotechnology, Vol 3, Chapter 3b. Verlag Chemie GmbH,Weinheim.

Fritzsche, C. J. 1992. Nonpoint source pollution, Calcium Magnesium Ac-etate deicer, an effective alternative for salt-sensitive areas. WaterEnvironment & Technology, January 1992, 44–51.

Ghose, T. K., Bhadra, A. 1985. Acetic acid. In: Moo-Young, M. (ed.),Comprehensive biotechnology, vol 3, chapter 36. Pergamon Press,New York.

Ito, L., Sota, H., Honda, H., Shimizu, K., Kobayashi, T. 1991. Efficient

Figure 7. Scanning electron micrographs ofC. formicoaceticum(longrods) andL. lactis (cocci) co-immobilized on the fiber surface in thefibrous-bed bioreactor.

506 BIOTECHNOLOGY AND BIOENGINEERING, VOL. 60, NO. 4, NOVEMBER 20, 1998

acetic acid production by repeated fed-batch fermentation using twofermentors. Appl. Microbiol. Biotechnol.36: 295–299.

Kondo, T., Kondo, M. 1996. Efficient production of acetic acid fromglucose in a mixed culture ofZymomonas mobilisandAcetobactersp.J. Fermentation Bioeng.81: 42–46.

Marynowski, C. W., Jones, J. L., Tuse, D., Boughton, R. L. 1985. Fermen-tation as an advantageous route for the production of an acetate salt forroadway deicing., Ind. Eng. Chem. Res. Dev.24: 457–465.

Parekh, S. R., Cheryan, M. 1990. Acetate production from glucose byClostridium thermoaceticum.Process Biochem. Intl.25: 117–121.

Parekh, S. R., Cheryan, M. 1990. Fed-batch fermentation of glucose toacetate by an improved strain ofClostridium thermoaceticum.Bio-technol. Lett.12: 861–864.

Parekh, S. R., Cheryan, M. 1991. Production of acetate by mutant strains ofClostridium thermoaceticum.Appl. Microbiol. Biotechnol. 36:384–387.

Parekh, S. R., Cheryan, M. 1994a. High concentrations of acetate with amutant strains ofC. thermoaceticum.Biotechnol. Lett.16: 139–142.

Parekh, S. R., Cheryan, M. 1994b. Continuous production of acetate byClostridium thermoaceticumin a cell-recycle membrane bioreactor.Enzyme Microb. Technol.16: 104–109.

Park, Y. S., Toda, K., Fukaya, M., Okumura, H., Kawamura, Y. 1991.Production of a high concentration acetic acid byAcetobacter acetiusing a repeated fed-batch culture with cell recycling. Appl. Microbiol.Biotechnol.35: 149–153.

Park, Y. S., Fukaya, M., Okumura, H., Kawamura, Y., Toda, K. 1991.Production of acetic acid by a repeated batch culture with cell recycleof Acetobacter aceti.Biotechnol. Lett.13(4):271–276.

Reed, W. M., Keller, F. A., Kite, F. E., Bogdan, M. E., Ganoung, J. S.1987. Development of increased acetic acid tolerance in anaerobichomoacetogens through induced mutagenesis and continuous selec-tion. Enzyme Microb. Technol.9: 117–120.

Shah, M. M., Cheryan, M. 1995. Improvement of productivity in aceticacid fermentation withClostridium thermoaceticum.Appl. Biochem.Biotechnol.51/52: 413–422.

Silva, E. M., Yang, S. T. 1995. Continuous production of lactic acid from

acid whey byLactobacillus helveticusin a fibrous-bed bioreactor. J.Biotechnol.41: 59–70.

Tang, I. C., Yang, S. T., Okos, M. R. 1988. Acetic acid production fromwhey lactose by the Coculture ofStreptococcus lactisandClostridiumformicoaceticum.Appl. Microbiol. Biotechnol.28: 138–143.

Tang, I. C., Okos, M. R., Yang, S. T. 1989. Effects of pH and acetic acidon homoacetic fermentation of lactate byClostridium formicoaceti-cum.Biotechnol. Bioeng.34: 1063–1074.

Wiegel, J., Carreira, L. H., Garrison, R. J., Rabek N. E., Ljungdahl, L. G.1991. Calcium magnesium acetate (CMA) manufacture from glucoseby fermentation with thermophilic homoacetogenic bacteria, Chapter16. In: Calcium magnesium acetate, an emerging bulk chemical forenvironmental applications. Elsevier, Amsterdam, Netherlands.

Wise, D. L., Augenstein, D. C. 1988. An evaluation of the bioconversion ofwoody biomass to calcium acetate deicing salt. Solar Energy41:453–463.

Yang, S. T., Huang, Y., Hong, G. 1995. A novel recycle batch immobilizedcell bioreactor for propionate production from whey lactose., Biotech-nol. Bioeng.45: 379–386.

Yang, S. T., Silva, E. M. 1995. Novel products and new technologies foruse of a familiar carbohydrate, milk lactose. J. Dairy Sci.78:2563–2583.

Yang, S. T., Tang, I. C., Zhu, H. 1992a. A novel fermentation process forcalcium magnesium acetate (CMA) production from cheese whey.Appl. Biochem. Biotechnol.34/35: 569–583.

Yang, S. T., Zhu, H., Lewis, V. P., Tang, I. C. 1992b. Calcium magnesiumacetate (CMA) production from whey permeate: Process and economicanalysis, resources conservation and recycling7: 181–200.

Yang, S. T., Zhu, H., Li, Y., Hong, G. 1994. Continuous propionate pro-duction from whey permeate using a novel fibrous bed bioreactor.Biotechnol. Bioeng.43: 1124–1130.

Yang, S. T., Chollar, B. H. 1996. Production of low-cost acetate deicersfrom biomass and industrial wastes, proceedings for the 4th interna-tional symposium on snow removal and ice control technology, Reno,Nevada, August 12–16, 1996. Transportation Research Board, Na-tional Research Council, Vol. 1, p. C-6.

HUANG AND YANG: ACETATE PRODUCTION FROM WHEY 507