The potential of positively-charged cellulose sponge for malolactic fermentation of wine, using Oenococcus oeni

Enzyme and Microbial Technology 28 (2001) 415 419 www.elsevier.com/locate/enzmictec The potential of positively-charged cellulose sponge for malolactic fermentation of wine, using Oenococcus oeni Sergi Maicas*,1, Isabel Pardo, Sergi Ferrer Departament de Microbiologia i Ecologia, Facultat de Biologia, Universitat de València, Burjassot, Spain Received 30 May 2000; received in revised form 31 October 2000; accepted 9 November 2000 Abstract Malolactic fermentation (MLF) is a secondary bioconversion developed in some wines involving malic acid decarboxylation. The induction of MLF in wine by cultures of free and immobilized Oenococcus oeni cells was investigated. This work reports on the effect of surface charges in the immobilization material, a recently described fibrous sponge, as well as the ph and the composition of the media where cells are suspended. A chemical treatment provided positive charge to the sponges (DE or DEAE) and gave the highest cell loadings and subsequent resistance to removal. Preculture media to grow the malolactic bacteria before the immobilization procedure were also evaluated. We have established favorable conditions for growth (Medium of Preculture), suspension solution (Tartrate-Phosphate Buffer), suspension ph (3.5 5.5) and immobilization matrix (DE or DEAE cellulose sponge) to induce MLF in red wine. The use of a semi-continuous system permitted a high-efficiency malic acid conversion by 2 10 9 cfu sponge 1 in at least four subsequent batch fermentations. 2001 Elsevier Science Inc. All rights reserved. Keywords: Cellulose sponge; Immobilization; Lactic acid bacteria; Leuconostoc oenos; Malic acid; MLF; Red wine 1. Introduction Malolactic fermentation (MLF) is a secondary process that occurs in red wine after alcoholic fermentation and consists of the conversion of L-malate to L-lactate and carbon dioxide. As a consequence of this reaction, the total acidity decreases and the organoleptic properties and biologic stability of the wine are generally improved [1,2]. Oenococcus oeni is the major bacterial species found in wines during the MLF, and is well adapted to the low ph and high ethanol concentration of wine [2,3]. Under natural conditions, the MLF is developed in weeks or months and satisfactory results are not always achieved because of physicochemical or nutritional conditions in the wine [4 6]. The probability of obtaining rapid and complete MLF is significantly enhanced by inoculating the wine with high levels of selected strains of O. oeni [7]. Immobilization may increase productivity due to greater * Corresponding author. Tel.: 34-963900022; fax: 34-963636301. E-mail address: sergi.maicas@iata.csic.es (S. Maicas). 1 Present address: Departament de Biotecnologia, Institut d Agroquímica i Tecnologia d Aliments (CSIC), P.O. Box 73. Burjassot, València, Spain 46100-E. packing density or provide a more protective environment, and also improve subsequent cell separation. The improvement of immobilization techniques for wine deacidification has long been studied using alginates [8,9], polyacrylamide [6], -carrageenan [10] and oak chips [11] but, nobody has yet tried to immobilize malolactic cells in ion sponge material for this purpose. For beverages, cell immobilization that involves little preparation and no additional chemicals is ideal. By using external surface adsorption, such as on to DEAE cellulose beads, support volume to biomass ratio is high and risk of attrition can restrict process conditions. With encapsulation, mass transfer limitations of nutrients in and/or metabolites out, can lead to pronounced biomass gradients, often with dead or inactive cells in the center. The sponge we have used, which is cellulose based, can be conveniently cut to size and used in packed or fluidized bed configurations. It is also autoclaved without apparent loss of activity [12] and surface properties can be tailored (i.e. made acidic or basic) by addition of functional groups [13]. Scott and O Reilly [12] used it to co-immobilize Saccharomyces cerevisiae and Lactobacillus plantarum to induce alcoholic and malolactic fermentation in cider. To succeed by inoculating O. oeni in wine to accomplish malolactic fermentation an additional requirement must be satisfied: 0141-0229/01/$ see front matter 2001 Elsevier Science Inc. All rights reserved. PII: S0141-0229(00)00339-2

416 S. Maicas et al. / Enzyme and Microbial Technology 28 (2001) 415 419 cells must be adapted to wine stress before the inoculation to avoid cell death [1]. A new nutritive medium (MP) [14] has been studied to enhance bacterial survival and ability to perform malolactic fermentation in wine. 2. Materials and methods 2.1. Bacterial strain and its cultivation The lactic acid bacteria used in this study was Oenococcus oeni (syn. Leuconostoc oenos) strain M42 [15]. It was propagated in Medium for Leuconostoc oenos (MLO) broth [16] at 28 C and ph 4.8 or Medium of Preculture (MP) [14] at 28 C and ph 4.5 for 48 72 h. 2.2. Immobilization media A plain cellulose based sponge (neutral, no surface treatment) was used along with, PRODUCTIV DE (basic, N (CH 2 CH 3 ) 2 H functional group) and DEAE (basic, (CH 2 ) 2 N (CH 2 CH 3 ) 2 H), supplied by BPS Separations, Spennymoor, UK. The sponges were obtained in flat sheets, approximately 0.8 cm thick. For cell uptake measurements and malolactic fermentation assays, the sponge was cut into 0.8 cm sided cubes. After drying sponge samples in an oven at 60 C for 48 h, recorded weight loss was 86%, 83% and 84% for DEAE, DE and plain sponges respectively. 2.3. Immobilization studies Late log phase cells, grown in MLO broth or MP broth up to OD 600nm 1.0 were harvested by centrifugation (18,000 g), washed twice in tartrate-phosphate buffer (TPB) [17] and suspended in 250 ml shake flasks containing 100 ml of fresh ph adjusted (ranging from 3.5 to 5.5) buffer and 2 g (wet weight) of sponge. The flasks were shaken at 100 rpm in an incubator (28 C) and duplicate samples were periodically taken to determinate OD of cells remaining in suspension. Cells were 10-fold diluted to measure into the linear range, when required. Calibration was done plotting OD 600nm against cfu ml 1. 2.4. Fermentation studies After immobilization, the DE sponges were washed twice in TPB to discard remaining cells in suspension. TPB was immediately replaced with 100 ml of sterilized Monastrell red wine containing in g l 1 : glucose, 0.9; fructose, 1.0; malic acid, 3.5 and 11% (v/v) ethanol at ph 3.5; then fermentation was started. To act as a control, fermentation was also set up with a flask that did not contain immobilization media. At the start of the fermentation, bacteria were added to the medium of that flask at an initial concentration equivalent to the total amount of cells immobilized on the sponge. The temperature of the flasks was maintained at Fig. 1. Oenococcus oeni M42 uptake by plain, ; basic (DE), and basic (DEAE), F sponge. Cells were suspended in phosphate-tartrate buffer at ph 4.5. Results represent the means of two independent experiments. 28 C. The semi-continuous fermentation process was achieved by replacing the spent wine with fresh wine every 24 h. Whenever there was a decrease in malic acid consumption, the immobilized cells were discarded. 2.5. Analytical techniques Periodically, one ml samples were aseptically collected from the flasks; bacterial densities were determined by plate counting on MLO agar at 28 C for 48 72 h and remaining malic acid quantities were determined as previously described [7]. All the data were obtained from duplicated assays. 3. Results 3.1. Rate of O. oeni cell uptake by sponges In order to check the effectiveness of the material at holding the cells, comparison was made between plain, acidic and basic (DE and DEAE) sponges of similar characteristics (i.e. pore size distribution and surface area). With O. oeni cells suspended in buffer, due to the net negative charge of cell walls, when exposed to negatively-charged sponge (PRODUCTIV CM (-COO )), uptake was virtually zero (data not shown). Uptake from TPB by plain sponge (no surface treatment) almost mirrored that of acidic sponge and only around 10 7 cfu (g sponge) 1 were adsorbed onto the matrix. The lower cell loading with the plain sponge was explained on the basis of the neutral charge. However, significant uptake was exhibited with the positively-charged sponges (Fig. 1). These results, recorded in the experiments with the positively charged materials, demonstrate the significant advantage the DE and DEAE sponges have for accumulating O. oeni cells compared with the acidic and plain sponges.

S. Maicas et al. / Enzyme and Microbial Technology 28 (2001) 415 419 417 Fig. 2. Oenococcus oeni M42 uptake by basic sponges (DE or DEAE). Cells were suspended in either phosphate-tartrate buffer (TPB) or MLO broth at ph 4.0. DEAE sponge in MLO, ; DE sponge in MLO, ; DEAE sponge in TPB, F and DE sponge in TPB, E. Results represent the means of two independent experiments. Fig. 4. Malic acid consumption in wine (ph 3.5) by Oenococcus oeni M42, free or immobilized onto DE sponge, previously grown in MLO or MP. Free cells cultured in MP, ; free cells cultured in MLO, ; immobilized cells cultured in MP, F and immobilized cells cultured in MLO, E. Results represent the means of two independent experiments. The ph of the medium in which the cells were resuspended (3.5 5.5) did not significantly influence cell adhesion (data not shown). Nevertheless a negative impact was detected in immobilizations done from culture medium (Fig. 2). Better results were achieved in TPB, which provided less competition for binding sites. This suggests coating of surface binding sites by medium constituents interfering with cell attachment. As a consequence, in order to maximize loading, particularly in terms of exposure time, for subsequent immobilization studies, cells were always recovered from their growth medium by centrifugation, washed and resuspended in TPB prior to exposure to the matrix. This enabled an uptake of 75% of available bacteria. Cell leakage into the medium was also determined in a column system operated at different fluxes (Fig. 3). DEAE and especially DE material capabilities were also demonstrated in tests to assess resistance to cell removal. At a flow rate of up to 30 ml min 1, no removal was recorded, after which a leakage of cells occurred, attributed to wash-out of cells not attached to matrix surfaces but loosely to other cells. As the flow rate increased, up to 150 ml min 1, there was a continuous rise in cell number in the circulating medium but always lower than 50% of previously immobilized bacteria. This indicated cells more closely associated with support surfaces shearing off and, therefore, a practical operational limit for the system. In practice, however, malolactic activity rates made us fix a low flow to ensure appropriate malic acid degradations. In semi-continuous flask fermentations the numbers of cells in suspension within the flasks were always 10 6 cfu ml 1 due to the attraction to the surface charge of sponge material. As the DE sponge, with functional groups giving a weakly positive surface, was significantly better at accumulating the O. oeni cells, it was selected for semi-continuous fermentation trials to induce malolactic fermentation in wine. 3.2. Semi-continuous process for malolactic fermentation Fig. 3. Effect of flow rate on bacterial removal (in TPB) from DE sponge. Results represent the means of three independent measurements. O. oeni M42 cells grown in MLO broth or MP broth and immobilized on DE sponges were incubated with red wine for 24 h (Fig. 4). Malic acid consumption by cells growing in MP was significantly better than with MLO. To assess the effectiveness of an experimental semi-continuous system, every 24 h the spent wine was replaced with fresh wine (each cycle took 24 h) and, at the end of the fourth cycle, the system was discarded because of the decrease in malolactic activity. In a traditional MLF, substantial malic acid consumption can only be observed after 3 4 weeks. In an immobilized system this period can be shortened and a period of 24 h enables about 50% consumption which is considered to be satisfactory for industrial production (Fig.

418 S. Maicas et al. / Enzyme and Microbial Technology 28 (2001) 415 419 authors [6,8,11], who explained it on basis of alcohol lethality. However, our data show higher efficiency with immobilized cells than with free cells due to the enhanced microenvironment that helped bacterial survival. That is, an increase in malic acid consumption rate was obtained by immobilization of cells in addition to advantages due to avoiding centrifugation and resuspension of cells. Moreover, O. oeni cells immobilized in DE or DEAE sponges retained their abilities to induce MLF in wine after storage at 4 C for 21 days without an apparent loss of activity. Substantial malic acid consumption (ranging from 40% to 50%) can be observed by using the stored cells, after 24-h incubation periods. 4. Discussion Fig. 5. Malic acid consumption in wine (ph 3.5) by Oenococcus oeni M42 previously grown in MP: free, F or immobilized onto DE sponge,, in four subsequent cycles. Assays were performed at 20 C. Results represent the means of two independent experiments. 5). This is due to the high concentration of cells that can be adsorbed to the matrix (about 1.5 2.0 10 9 cfu (g sponge) 1 ). A control fermentation was also performed with a nonsponge system (all cells remained in free suspension), to make comparison between processes with a high number of cells free or immobilized onto sponges. Results obtained in a free cell system, showed that stable rate of degradation (around 40 45%) could be achieved by centrifugation of cells after a batch fermentation and resuspension in fresh wine (Fig. 5). Similar experiments were carried out with DE sponge immobilized bacteria. Satisfactory results were recorded in the first 24 h batch, where the initial 3.5 g l 1 of malic acid were almost completely metabolized (Fig. 5). The sponge may be adsorbing sugars and other nutrients that provide a comfortable environment for the cells. This is not possible in other immobilization techniques that employ internal association/entrapment, such as alginate beads or glass spheres, where nutrient mass transfer limitations may occur [18]. However, subsequent batches with immobilized cells showed a strong drop in malolactic activity and only 50% of the initial malic acid was consumed in each 24 h period for 4 6 cycles. This can be due to cell mortality, rather than leakage, indicated by the low level of cells found in suspension. The decrease in activity of immobilized O. oeni cells after repeated use was also reported by some The aim of this work was to assess the possibility of using cells of O. oeni (syn. L. oenos), immobilized by adsorption to cellulose sponges, for the MLF of wine. Previous reports describing the use of immobilized bacteria for this purpose have shown the possibility of immobilizing O. oeni by entrapment in polyacrylamide [6], -carrageenan [10,19] or alginate [8,9,20]. Most of these materials are rejected by the wine producers, due to pre-fermentation preparation and requirements of additional chemicals. The use of these materials may mitigate commercial implementation and demands the development of a new matrix, which is easy to prepare and use. In this paper we show the utilization of a recently described fibrous sponge [12] to immobilize O. oeni for MLF in wine which can meet these requirements. A chemical surface treatment provided basic characteristics to the sponges (DE or DEAE) and gave the highest cell loadings and subsequent resistance to removal. Results using basic sponges were significantly better than with plain sponge material, suggesting ionic attraction between cell envelopes and matrix. In terms of loading DE sponge, it has been shown that ph (ranging from 3.5 to 5.5) exhibited no influence in uptake but suspension media had important effects. Particulates from the nutritive media coated the basically charged DE sponge and suggested that uptake from buffered solutions (TPB) which provided least competition for binding sites than proteins in synthetic media [13,18]. In terms of bioconversion of malate, our findings show that it can be efficiently achieved using a semi-continuous system based on immobilization of cells in DE cellulose sponges without loss of their activity using repeated batch fermentations even after storage at 4 C. Rates of malolactic activity were enhanced when O. oeni cells were precultured in nutritive MP broth autonomously of its free or immobilized state. Nevertheless, in the first cycle of fermentation, MLF was accomplished in less than 3hinflasks containing cells immobilized onto sponge while free cells required 24 h to metabolize 3.5 g l 1 of malic acid. Unfortunately, the process was not so satisfactory when the immobilized cells

S. Maicas et al. / Enzyme and Microbial Technology 28 (2001) 415 419 419 were transferred to fresh wine and the process was started again after 4 6 experimental cycles. Cells began to show significant signs of aging, since there was a decrease in malic acid consumption. Results, although better than those obtained with free cells, reflected the damage to the bacteria when exposed to the wine containing ethanol, which leads to a decrease in the viability of the cells in the system [2,4]. Overall, our findings extend the conclusions of previous studies [11,19,20] showing that O. oeni cells can be immobilized onto different supports to perform MLF in wine. This work helps in the implementation of an immobilized reactor for practical wine fermentation by the use of a recently developed food grade sponge [12,13]. Acknowledgments This work has been partially supported by grants from the Comisión Interministerial de Ciencia y Tecnología (ALI93-0246) and by grants from the M.E.C. (Spanish Government) and Servei d Investigació (Universitat de València) to S.M. Dr. J.A. Scott is gratefully acknowledged for providing us sponge material and tuition. Finally, we would like to thank P. González-Cabo for her technical assistance. References [1] Davis CR, Wibowo D, Eschenbruch R, Lee TH, Fleet GH. Practical implications of malolactic fermentation: a review. Am J Enol Vitic 1985;35:290 301. [2] Wibowo D, Eschenbruch R, Davis CR, Lee TH. Occurrence and growth of lactic acid bacteria in wine: a review. Am J Enol Vitic 1985;36:302 13. [3] Ramos A, Lolkema JS, Konings WN, Santos H. Enzyme basis for ph regulation of citrate and pyruvate metabolism by Leuconostoc oenos. Appl Environ Biotechnol 1995;61:1303 10. [4] Kunkee RE. Control of malo-lactic fermentation induced by Leuconostoc citrovorum. Am J Enol Vitic 1967;18:71 7. [5] Beelman RB, Gallander JF. The effect of grape skin treatments on induced malo-lactic fermentation in Ohio wines. Am J Enol Vitic 1970;21:193 200. [6] Rossi J, Clementi F. L-malic acid catabolism by polyacrylamide gel entrapped Leuconostoc oenos. Am J Enol Vitic 1984;36:100 2. [7] Maicas S, González-Cabo P, Ferrer S, Pardo I. Production of Oenococcus oeni biomass to induce malolactic fermentation in wine by control of ph and substrate addition. Biotechnol Lett 1999a;21:349 53. [8] Spettoli P, Nuti MP, Crapisi A, Zamorani A. Technological improvement of malolactic fermentation in wine by immobilized microbial cells in a continuous flow reactor. Ann NY Acad Sci 1987;501: 386 9. [9] Cuenat P, Villetaz JC. Essais de fermentation malolactique des vins par bactéries lactiques immobilisées du genre Leuconostoc. Rev Suisse Vitic Arboric Hortic 1984;16:145 51. [10] Crapisi A, Spettoli P, Nuti MP, Zamorani A. Comparative traits of Lactobacillus brevis, Lactobacillus fructivorans and Leuconostoc oenos immobilized cells for the control of malo-lactic fermentation in wine. J Appl Bacteriol 1987;63:513 21. [11] Janssen DE, Maddox IS, Mawson AJ. An immobilized cell reactor for the malolactic fermentation of wine. Aust New Zealand Wine Ind J 1993;8:161 5. [12] Scott JA, O Reilly A. Co-immobilization of selected yeast and bacteria for controlled flavour development in an alcoholic cider beverage. Proc Biochem 1995;31:111 7. [13] Scott JA, O Reilly AM, Kirkhope S. A fibrous sponge matrix to immobilise yeast for beverage fermentations. Biotechnol Tech 1995; 9:305 10. [14] Maicas S, Pardo I, Ferrer S. Continuous malolactic fermentation in red wine using free Oenococcus oeni. World J Microbiol Biotechnol 1999b;15:737 9. [15] Pardo I, Zúñiga M. Lactic acid bacteria in Spanish red rosé and white musts and wines under cellar conditions. J Food Sci 1992;57:392 405. [16] Caspritz G, Radler F. Malolactic enzyme of Lactobacillus plantarum. J Biol Chem 1983;258:4907 10. [17] Henick-Kling T. Control of malolactic fermentation. In: Henick- Kling T, editor. Technical review. Adelaide: The Australian Wine Research Institute, 1986. p. 3 6. [18] O Reilly AM, Scott JA. Use of an ion-exchange sponge to immobilise yeast in high gravity apple based (cider) alcoholic fermentations. Biotechnol Lett 1993;15:1061 666. [19] McCord JD, Ryu DDY. Development of malolactic fermentation process using inmobilized whole cells and enzymes. Am J Enol Vitic 1985;36:214 8. [20] Spettoli P, Bottacin A, Nuti MP, Zamorani A. Immobilization of Leuconostoc oenos ML 34 in calcium alginate gels and its application to wine technology. Am J Enol Vitic 1982;33:1 5.