Comparison between Immobilized Klu yverorn yces fragilis and Saccharom yces cerevisiae Coimmobilized with P-Galactosidase, with Respect to Continuous Ethanol Production from Concentrated Whey Permeate

B. Hahn-Hagerdal Applied Microbiology, Chemical Center, University of Lund, P.O. Box 124 S-227 00 Lund, Sweden

Accepted for publication September 27, 1984

Kluyveromyces fragilis immobilized in calcium alginate gel was compared to Saccharornyces cerevisiae coim- mobilized with p-galactosidase, for continuous ethanol production from whey permeate in packed-bed-type col- umns. Four different whey concentrations were studied, equivalent to 4.5, 10, 15, and 20% lactose, respectively. In all cases the coimmobilized preparation produced more ethanol than K. fragilis. The study went on for more than 5 weeks. K. fragilis showed a decline in activity after 20 days, while the coimmobilized preparation was stable during the entire investigation. Under experimental con- ditions theoretical yields of ethanol were obtained from 4.5 and 10% lactose substrates with the coimmobilized system. Using 15% lactose substrate, theoretical yields were only obtained when a galactose-adapted immobilized S. cerevisiae column was run in series with the coim- mobilized column. Then a maximum of 71 g/L ethanol was produced with a productivity of 2.5 g/L h. The coim- mobilized column alone gave a maximum ethanol con- centration of 52 g/L with a productivity of 4.5 g/L h, whereas immobolized K. fragilis only produced 13 g/L ethanol with a productivity of 1.1 g/L h. It was not possible to obtain theoretical yields of ethanol from the highest substrate concentration.

INTRODUCTION

There is a growing interest in using whey or whey permeate as a substrate for large-scale ethanol pro- duction. The fermentation product is then used in the wine industry or as supplementary liquid fuel.

Cheese whey permeate is obtained by ultrafiltration of cheese whey, a process by which highly nutritious whey proteins are recovered as a protein concentrate. The resulting permeate contains approximately 4.4% lactose and 0.5% salts. Biotechnology and Bioengineering. Vol. XXVII, Pp. 914-916 (1985) 0 1985 John Wiley & Sons, Inc.

The fermentation of whey permeate constitutes some problems in that the fermentable sugar is lactose. Or- ganisms like Kluyveromyces sp. ferment lactose but, on the other hand, seem to be inhibited by moderate concentrations of sugar and salt.4

There also seems to be some difficulty in obtaining high alcohol yields when using Kluyveromyces sp., probably due to the inactivation of the lactose activity at ethanol concentrations above 4%. However, recently it was reported that strains of Kluyveromyces fragilis were adapted to produce up to 10% (v/v) of ethanol from a concentrated whey permeate containing 22.5% l a c t o ~ e . ~

It has previously been reported that p-galactosidase coimmobilized with Saccharomyces cerevisiue in cal- cium alginate gel could continuously ferment whey permeate to ethanol.637 The present article deals with a comparison between immobolized K . frugilis and S . cerevisiae coimmobilized with P-galactosidase with respect to their capacity to continuously ferment whey permeate and concentrated whey permeate to ethanol.

MATERIALS AND METHODS

Saccharomyces cerevisiae, Y 1, was a generous gift from Professor A. L. Demain, MIT, Cambridge, MA. Kluyveromyces fragilis, NRRLY 24 15, was a generous g& from Professor F. Kosikowski, Cornell University, Ithaca, NY. p-Galactosidase was generously supplied by Miles KaliChemie GmbH & Co., KG, Hannover, Germany.

CCC 0006-3592/85/060914-03$04.00

Sodium alginate (practical grade, type IV), N-hy- droxy-succinimide, 1-ethyl-3-( 3-dimethyl-amino-propy 1 - 1)-carbodi-imid HC 1 (EDC), alcohol dehydrogenase from equine liver, and p-nicotinamide adenine dinu- cleotide were purchased from Sigma Chemical Co. St. Louis, MO.

The whey permeate substrates holding 45, 100, 150, and 200 g/L lactose, respectively, were prepared by redissolving a spray-dried sweet whey permeate powder kindly supplied by the department of Food Engineering, Lund University, Alnarp, Sweden. The pH was adjusted to 4.5 as it had been found previously that acid whey was more readily fermented than sweet whey.7 Calcium chloride was added to 0.01M to stabilize the alginate gel used as immobilization matrix.

Three different immobilized preparations were used: S . cerevisiue was grown on glucose and coimmo-

bilized with p-galactisidase as previously described for S . cerevisiue coimmobilized with P-glucosidase.* The enzyme (250 mg) was bound to 50 mg sodium alginate using 52 mg N-hydroxy-succinimide and 46 mg EDC.9 The alginate-P-galactosidase complex was then coen- trapped with 160 mg (dry wt) yeast cells in another 100 mg sodium alginate”,” in the shape of beads, 2 mm diameter. Calcium chloride (0.1M) was used for bead formation. This recipe will typically result in 3.7 g (wet wt) beads having a density of 1.05 g/mL.

K . frugilis grown on lactose and S . cerevisiue grown on galactose were also entrapped in alginate, but with the enzyme excluded.

The immobilized preparations were filled in small columns made from disposable syringes. A bed volume of 7 mL corresponding to 7.33 g (wet wt) beads was used. The flowrate was 0.6 mL/h, corresponding to a dilution rate of 0.09 h-’. The columns were operated at 30°C in a thermostated cabin.

Ethanol concentrations in the column effluents were determined either enzymatically using alcohol dehydrogenase” or by gas chromatography as described elsewhere. l 3

RESULTS AND DISCUSSION

Three experimental set-ups were run in parallel: the first was a column containing alginate-immobilized K . frugilis; the second, a column containing S. cerevisiue coimmobilized with p-galactosidase; the third, one column containing S . cerevisiue coimmobilized with p-galactosidase followed by a second column containing immobilized S . cerevisiue, which had been grown on galactose instead of glucose. The three experimental set-ups were fed the same whey permeate substrate holding increasing concentrations of lactose at the same flowrate. The ethanol concentrations in the effluents from the columns were recorded. (Fig. 1).

At all concentrations of lactose in the whey permeate S . cerevisiue coimmobilized with p-galactosidase pro-

duces higher amounts of ethanol than immobilized K . frugilis. This particular strain of K . frugilis was adapted to produce high concentrations of ethanol from con- centrated whey ~ e r m e a t e . ~ However, the fermentation process was slow. It took 15 days to produce 10% (v/v) ethanol, which might explain the poor performance in continuous ethanol fermentation, even though a low dilution rate of 0.09 h-’ was used.

Even at the lowest lactose concentration, K . frugilis cannot convert lactose to ethanol completely, and after 3 weeks of operation the productivity declines signif- icantly. A maximum of 57% of the theoretical yield is obtained at the lowest substrate concentration, 45 g/L, with a productivity of 1.1 g/L h.

In contrast, both experiments using S . cerevisiue coimmobilized with P-galactosidase produced theo- retical yields of ethanol both from 45 and 100 g/L substrate concentrations. At a substrate concentration of 150 g/L lactose only the experimental set-up using the coimmobilized preparation followed by a column with immobilized S . cerevisiue grown on galactose was able to produce theoretical yields of ethanol.

In a previous study comparing the continuous fer- mentation of sweet and acid whey using S . cerevisiue coimmobilized with p-galactosidase, theoretical yields of ethanol were not obtained, probably due to the fact that too high a dilution rate, 0.3 h-’, was used.7 How- ever, in both cases the yields were higher than 50% of theoretical, indicating that galactose was fermented as well. In a packed-bed reactor, as the presently used columns might be regarded, the glucose formed on hydrolysis is most likely fermented in the lower part of the column. Therefore, the yeast in the top of the column will gradually be adapted to galactose, which has been reported to be necessary for ethanol production from galactose with S . ~ e r e v i s i u e . ~ Adding a second column containing immobilized galactose-adapted S . cerevisiae is one way of extending this concept.

The present results show that at a dilution rate of 0.09 h- ’ S . cerevisiue coimmobilized with p-galacto- sidase can alone produce theoretical yields of ethanol from up to 100 g/L lactose with a miximum productivity of 4.5 g/L h. Under the conditions chosen it is first at a substrate concentration of 150 g/L that the gal- actose-adapted immobilized S . cerevisiue column is needed to give theoretical yields of ethanol. However, this experimental set-up then results in a lower pro- ductivity, 2.5 g/L h, because the bed size has been doubled. The results also indicate the possibility of using the coimmobilized preparation solely, but at an even lower dilution rate, in order to ferment the 150 g/L substrate completely.

When the substrate concentration was increased to 200 g/L lactose, no further increase in product con- centration was achieved. Whether this is due to “aging” of the immobilized preparations or to the high osmolality of the substrate (except for the sugar content whey salts have been concentrated to the equivalent of 23

HAHN-HAGERDAL: CONTINUOUS ETHANOL PRODUCTION 91 5

I -

-

-

-

200

150

- - I

-100 I L

- 2 c

0

W v)

0

- v

- 2 - 4

50

r -----------

80

I _-___________-I

-

L

I I di P

J 10 20 30 DAYS

Figure 1. Continuous production of ethanol from increasing concentrations of whey permeate (dashed line). (B) Calcium alginate immobilized K. fragilis; (0) S . cerevisiae coimmobilized with p-galactosidase; (0) S. cerevisiae coimmobilized with p-galactosidase followed by a column of equal size containing immobilized S. cerevisiae grown on galactose. Column size 7 mL, containing 7.3 g (wet wt) alginate beads. pH 4.5, 30"C, flowrate 0.6 mL/h.

g/L), which might have an inhibitory effect on ethanol fermentation, remains to be established.

In conclusion, S. cerevisiue coimmobilized with p- galactosidase is a much more efficient way to contin- uously ferment concentrated whey permeate than using K. frugilis, even if this organism can ferment lactose directly. It might be possible that K. fragilis ferments whey permeate to better yields of ethanol if the substrate is continously supplemented with certain limiting nu- trients. Such a continuous supplementation should then be economically evaluated with respect to the cost of coimmobilizing the enzyme with S. cerevisiue, which in this report has been shown to be a technical solution with high operational stability.

Maria Nilsson, Peter Blom, and Lars Hagerman have given skillful technical assistance. This study was supported by the Swedish Natural Science Research Council and the Swedish Board for Tech- nical Development.

References

1. 2. 3. 4.

5. 6. 7. 8. 9.

10.

11. 12.

13.

L. Reesen and R. Strube, Proc. Biochem. 13(11), 21 (1978). E. J. Mann, Dairy 2nd. Int . , 45(3) 47 (1980). H. C. Chen, and R. R. Zall, Proc. Biochem., 17(1), 20 (1982) V. S. O'Leary, C. Sutton, M. Bencivengo, B. Sullivan, and V. H. Holsinger, Biotechnol. Bioeng., 19, 1689 (1977). J. Gawel and F. V. Kosikowski, J. Food Sci., 43, 1717 (1978). B. Hagerdal, Acta Chem. Scan., B, 34(8), (1980). B. Hahn-Hagerdal, Appl . Biochem. Biotechnol., 7 , 43 (1982). B. Hahn-Hagerdal, Biotechnol. Bioeng. 26, 771 (1984). P. Cuatrecasas and I. Parikh, J . Biochem., 11(12), 2291 (1972). U. Hackel, J. Klein, R. Magnet, and F. Wagner, Eur. J . Appl. Microbio., 1, 291 (1975). M. Kierstan and C. Bucke, Biotechnol. Bioeng., 19,387 (1977). N. 0. Kaplan and M. M. Ciotti, in Methods in Enzymology, Vol. 3, S . P. Colowick, and N. 0. Kaplan, Eds. (Academic Press, New York, 1957). B. Hahn-Hagerdal, and B. Mattiasson, Eur. J. Appl. Microbiol. Biotechnol., 14, 140 (1982).

916 SIOTECHNOLOGY AND BIOENGINEERING, VOL. 27, JUNE 1985