Evaluation of Malolactic Bacteria Isolated from Oregon Winest

Transcription

1 APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Aug. 1989, p Vol. 55, No /89/821-7$2./ Copyright C) 1989, Amerian Soiety for Mirobiology Evaluation of Malolati Bateria Isolated from Oregon Winest T. HENICK-KLING,t W. E. SANDINE,* AND D. A. HEATHERBELL Department of Mirobiology and Department of Food Siene and Tehnology, Oregon State University, Corvallis, Oregon Reeived 7 Deember 1988/Aepted 12 May 1989 Oregon is a ool wine-produing region where grapes harateristially ontain high onentrations of organi aids. To redue the natural aidity and inrease the mirobiologial stability and flavor omplexity of the wine, malolati fermentation is enouraged. In this study, strains of Leuonosto oenos indigenous to Oregon wines were evaluated for their suitability to ondut malolati fermentation in Oregon wines. Tests determined the malolati ativity of the Oregon isolates in omparison with ommerial strains ML-34, PSU-1, MLT-kli, and ens 44-4 under various temperature and ph onditions. Sensitivities to sulfur dioxide, ethanol, and fumari aid also were determined. Two Oregon strains, Er-la and Ey-2d, were seleted for ommerial winemaking tests beause they had greater malolati ativity under onditions of low ph (3.) and low temperature (15 and 8 C), respetively. Oregon's vitiulture areas are similar in limate to the ooler wine-growing regions of Europe (2). Also, the Oregon wine industry, exept for fruit wine prodution, is based on the use of European grape varieties of the speies Vitis vinifera. Oregon musts and wines similarily have low ph and high titratable aidity values. The average ph of white musts from different loations in Oregon is 3.2, and the average titratable aidity (TA) as tartari aid is 9.7 g/liter; for red musts, the average ph is 3.25 and the average TA is 9.5 g/liter (31). An important fator affeting the quality of higher-aidity wines is the malolati fermentation (MLF). In this fermentation, lati aid bateria (LAB) onvert L-mali aid to L-lati aid and CO2 (8, 15, 22, 24, 25, 31). Mali aid onstitutes between and 5% of the wine aidity, the remainder being mostly tartari aid (23). High malate onentration in musts and wines is assoiated with unripe grapes, sine as the grapes ripen, the mali aid is metabolized in the grape berry and may disappear ompletely. In ool limates, however, grapes may not ripen ompletely and are then proessed with high amounts of malate, resulting in high total aidity in the wine. The onversion of the diarboxyli mali aid to the monoarboxyli lati aid with the release of CO2 redues the aidity of the wine; other baterial metabolites further modify wine flavor (7). Indution of MLF offers mirobiologial stability by ensuring that the degradation of mali aid does not our in the bottle, where the growth of LAB and the formation of CO2 are onsidered spoilage. MLF ours readily in high-ph wines if not atively disouraged by the use of inhibitory onentrations of S2 or sterilization by heat or filtration (31). In low-ph wines, spontaneous MLF by indigenous LAB has been unreliable and indution of MLF by inoulation with ommerially available strains of Leuonosto oenos has shown little suess (19, 31). With the assumption that LAB adapted to fermentation * Corresponding author. t Tehnial Paper no. 8713, Oregon Agriultural Experiment Station. t Present address: Department of Food Siene and Tehnology, Cornell University, Geneva, NY. Present address: Division of Hortiulture and Proessing, Division of Sientifi and Industrial Researh, Aukland, New Zealand. 21 onditions and omposition of Oregon wines might degrade malate more rapidly, Izuagbe et al. (14) isolated and haraterized several strains of L. oenos from different wine samples. Fifteen of these isolates were ompared in the present study with ommerial strains of malolati bateria to determine whether or not any were better suited for indution of MLF under Oregon winemaking onditions of low ph and low temperature. MATERIALS AND METHODS Cultures. The malolati strains used in this study were isolated from wines undergoing spontaneous MLF in two Oregon wineries. These strains, lassified as L. oenos (9, 1; T. P. Dohman, M. S. thesis, Oregon State University, Corvallis, 1982), were Ey-la, Ey-2a, Ey-2b, Ey-2, Ey-2d, and Ey-4b, isolated from a 1978 Merlot; Ey-42, Ey-42a, and Ey-42b, isolated from a 1979 Chardonnay; Ey-, isolated from a 1978 Pinot Noir; Er-la, Er-lb, and Er-i, isolated from a 1979 Pinot Noir; and Er-2, Er-2a, Er-3, and Er-4a, isolated from a 1978 Chardonnay. L. oenos ML-34 (13) was obtained on a nutrient agar slant from R. E. Kunkee, University of California, Davis. L. oenos PSU-1 (4) was reeived as a freeze-dried preparation (Leuostart) from Tri Bio Laboratories, State College, Pa. It was rehydrated by following the manufaturer's instrutions. Strain 44-4 (28) was obtained from BioLogials In., Berkeley, Calif., as a freeze-dried preparation and rehydrated in the resuspension medium supplied by the manufaturer. Strain MLT-kli, a Swiss isolate, was on a nutrient agar slant made available by Mirolife Tehnis, Sarasota, Fla. All ultures were maintained on MRV-8 agar slants (14). Temperature trials. Wine used was prepared in the experimental winery of the Department of Food Siene and Tehnology, Oregon State University. It was made from Pinot Noir grapes grown on the Agriultural Experiment Station at Medford, Oreg. Grape juie and wine were analyzed for degree Brix (soluble solids by refratometer, in grams per kilogram as surose), TA (in grams per liter as tartari aid), volatile aidity (VA; by Cash steam distillation as grams per liter of aeti aid), ph, and alohol (ebulliometer volume perent) (Table 1) (1). The wine was dispensed (2 ml) into 25-ml Erlenmeyer flasks and autolaved for 12 min at 121 C. After being ooled (25 C), the wine was analyzed for ph, TA, and VA (Table 1). Downloaded from on November 9, 218 by guest

2 VOL. 55, 1989 TABLE 1. Analytial data on Pinor Noir grape must and new wine used in temperature trials SamplepH TA 'Brix ~% (Vol/Vol) VA Sample ph TA Brix alohol (g/liter) Grapes New wine A Winea a New wine heated (A) in an autolave for 12 min at 121 C. After being autolaved, the wine ontained no ethanol and no VA. Erlenmeyer flasks with the uninoulated heated wines were plaed into three onstant temperature inubators at 2, 15, and 8 C overnight to equilibrate before inoulation. The baterial ultures were grown at 3 C in MRV-8 broth (ph 5.5) in 16-mm loosely srew-apped glass tubes inubated under CO2 (GasPak system; BBL Mirobiology Systems, Cokeysville, Md.). After 35 h, the tubes had reahed maximum turbidity and were refrigerated at 6 C. Approximately 2 h later, the wines were inoulated with 1% (vol/vol) of these ultures and plaed bak into the inubators. An uninoulated heated wine sample served as a ontrol at eah temperature. Viable ells in the inoulated heated wine were determined by plating on MRV-8 medium at ph 4.6. Free and total sulfur dioxide and total soluble solids were determined by the methods given by Amerine and Ough (1). Wines were periodially analyzed by paper hromatography (16) for the disappearane of malate and for the formation of latate. Ativity of the malolati ultures was followed by measuring the size of the malate spot and reording its hange. Variation in the size of the malate spot of dupliate samples was minimized by dividing its size (width times height) by the size of the malate spot of the uninoulated ontrol. When the malate spot had disappeared on the paper hromatogram, samples were taken for plate ount analysis as desribed above. Malolati fermentation was onsidered omplete when malate ould no longer be deteted on the paper hromatograph. ph trials. Srew-apped (16 mm) tubes were filled with 1 ml of MRV-8 broth ontaining 2 g of L-malate per liter and adjusted to ph values ranging from 2.8 to 4. with 8 N tartari aid. The media were sterilized by autolaving (121 C for 15 min) and the desired ph values were verified before inoulation. Cultures were grown at 3 C in MRV-8 broth (ph 4.6) for 4 days to maximal ell density and inoulated at a rate of 1% (vol/vol). The onentration of L-malate was determined enzymatially (21). Malolati fermentation was onsidered omplete when the onentration of malate had reahed less than 5 mg/liter. Malolati ativity. On the basis of the results of the ph trials, omparative rates of MLF were alulated for eah strain in a manner similar to that of Lafon-Lafourade (17). Ativity was expressed as the milligrams of malate degraded per log of viable ells per milliliter in 1 ml of medium in 24 h at 2 C. This expression was hosen in order to estimate the performane of a baterial strain in wine, that is, how atively it would degrade the mali aid from a ertain inoulum size. We were interested not in the amount of obtainable growth but in how atively a given ulture would degrade the mali aid. Malate was determined enzymatially as desribed above for the ph trials. Ethanol tolerane. The effet of various onentrations of ethanol on ell growth and malate metabolism was studied MALOLACTIC BACTERIA IN OREGON WINES 211 for strains ML-34 and Er-la. The Chardonnay wine used had a ph of 3.3 and ontained 9.4% (vol/vol) alohol, 11.1 g of TA per liter, residual sugar (Clinitest; Ames Division, Miles Laboratories, Elkhart, Ind.), 5.6 g of malate per liter, and 8 mg of free and 13 mg of total SO2 per liter. No attempt was made to sterilize the wine, sine the experiment was meant to simulate a winery environment in whih a ommerial starter ulture is inoulated into wines of different alohol onentrations. The wine ontained 9.4% (vol/vol) ethanol, and appropriate amounts of 95% (vol/vol) ethanol were added to ahieve alohol onentrations of 1, 12, and 14% (vol/vol). The ph of the wine was adjusted to 3.5 with 5 N NaOH. Sterile, 16-mm srew-apped glass tubes were filled ompletely with the wines to minimize the growth of aerobi spoilage organisms. Inoula were grown for 4 days at 3 C in 16-mm srew-apped glass tubes in MRV-8 broth (ph 4.6), and the wine was inoulated with 1% (vol/vol) ML-34 (1.5 x 19 CFU/ml) or 1% (vol/vol) Er-la (1.2 x 19 CFU/ml). An uninoulated ontrol was maintained for eah treatment of the wine. The wines were inubated at 2 C. The number of viable ells was determined by plate ount on MRV-8 medium (ph 4.6). Cyloheximide (7 mg/liter; Sigma Chemial Co., St. Louis, Mo.) was added to the plate ount medium to inhibit yeast growth. The mali aid was determined enzymatially (21). Fumarate and sulfur dioxide. The same wine used in the test of ethanol tolerane was used. It was divided into two lots; to one lot 3 mg of SO2 (potassium metabisulfite) per liter was added, and the other lot was left as a ontrol. The wine without added SO2 ontained 8 mg of free and 13 mg of total SO2 per liter, and the wine with the added SO2 ontained 2 mg of free and 4 mg of total SO2 per liter after 1 day of equilibration. Eah lot was divided into three sublots to whih,.5, or 1.5 g of fumari aid per liter was added. The ph of the wine in all lots was then raised to 3.5 with 5 N sterile NaOH. Sterile, 16-mm srew-apped glass tubes were filled ompletely with the wines with the different treatments. One tube with eah treatment was inoulated at 1% (vol/vol) with strain ML-34 (1.9 x 19 CFU/ml) or strain Er-la (1.5 x 19 CFU/ml), produing three tubes for eah inoulum. The inoula had been grown in MRV-8 broth (ph 4.6) at 3 C in 16-mm srew-apped glass tubes for 4 days. The wines were then inubated at 2 C. The number of viable ells was determined as desribed above. Sulfur dioxide and ph. The same Chardonnay wine was used; 1 liter was adjusted to ph 3.3 with 1 N NaOH, and another liter was left at ph 3.3. To both wines, various additions of SO2 were made. The final onentrations of free and total SO2 were determined 24 h after the addition to allow equilibration of the two forms. The amounts of free and total SO2 were determined by the method of Amerine and Ough (1). Sine no differene in the S2-binding apaity of the wine at different ph values was observed, the same additions of SO2 were made to both wines. Samples (64 ml) of eah wine were put into srew-apped glass bottles, and SO2 was added from a stok solution of potassium metabisulfite (17.7 g/liter). Additions of 2, 3, and 4 mg of SO2 per liter resulted in 13, 2, and 26 mg of free and 3, 4, and 49 mg of total S2 per liter. No addition was made to one set of wine samples; it ontained 8 mg of free and 13 mg of total SO2 per liter. Sterile, 2-mm srew-apped glass tubes were ompletely filled with the wines adjusted to the various phs and SO2 onentrations. After 24 h, one tube of eah treatment was inoulated with 1% (vol/vol) eah of ML-34 or Er-la. The Downloaded from on November 9, 218 by guest

3 212 HENICK-KLING ET AL. APPL. ENVIRON. MICROBIOL. o a) C) Co a1) a) a: _; ranged from 1.2 x 14 (Ey-2b) to 6.6 x 16 (Ey-42a) per ml; at 15 C, they ranged from <1 (PSU-1, ML-34, Ey-42, Er-2, and ontrol) to 1. x 17 (Ey-42b) per ml; at 8 C, they ranged from <1 (PSU-1, ML-34, Ey-2a, Ey-42b, Er-2, and ontrol) to 2.3 x 15 (Ey-2b) per ml. Representative results of the relative malate utilization rates at 2, 15, and 8 C are given in Fig. 1 and 2. The fastest strains to omplete MLF were Ey-2d, Ey-, and Er-la: eah ompleted MLF within 21 days. These strains were followed by Ey-42 and Er-2 at 36 days, Ey-2b at 41 days, Ey-2a at 52 days, Ey-41 and PSU-1 at 83 days, Ey-42a at 87 days, ML-34 at 144 days, and Ey-42b, whih had not ompleted MLF,,_,_._. after 228 days when the trial was terminated At 15 C, the differenes in temperature tolerane beame Melativeonentratonths ofmat lac ishow very large. Three strains, PSU-1, ML-34, and Ey-42, did not Months at 2'C any malolati ativity within 12 months. Seven FIG. 1. R mented withielative onentrations of malate in Pinot Noir fer- strains, Ey-2a, Ey-2d, Ey-42a, Ey-42b, Ey-41, Ey-, and L. oenos ML-34 at 2 C as determined by measuring Er-2, ompleted MLF within this time, and two strains, w the size of thie mali aid spot on the paper hromatogram, dividing ' '. it by the size of the mali aid spot of the uninoulated ontrol, and Er-la and Ey-2b, fermented the malate partially. Of the multiplying tby 8 for graphial demonstration. seven strains that ompleted MLF most quikly, Ey-2d finished at 145 days, followed by Ey-42b at 158 days, Ey-41 at 187 days, Er-2 at 225 days, Ey-42a at 256 days, Ey-2a at inoula had[been grown in MRV-8 broth (ph 4.6) at 3 C for 39 days, and Ey- at 34 days. 3 days. At 8C, none of the strains ompleted MLF within 12 The number of viable ells was determined as desribed months; however, three strains, Ey-2d, Ey-2b, and Ey-42a, above. The onentration of malate was determined enzy- did ferment the mali aid to very low onentrations. Three matially (221). other strains, Ey-, Ey-42, and Ey-41, fermented small amounts of mali aid. All other strains were inative at this RESULTS temperature. ph trials. The results of trials at ph 4., 3.5, 3.2, and 3. Tempera' ture trials. The initial ell density in the inou- are given in Table 2. For the seletion of strains suitable for lated, heate-d wines was approximately 16 ells per ml, with MLF in Oregon low-ph wines, nine Oregon strains were the exepti on of ML-34, whih ontained 3.5 x 17 ells per evaluated in omparison to the four ommerial strains ml, and Ey -2d, Ey-, and Er-la, ontaining 3 x 15 ells per PSU-1, ML-34, MLT-kli, and ens It may be seen that, ml. No vialble ells were present in the uninoulated ontrol in ontrast to the ommerial strains, several isolates, espe- and end of the experiment. At the onlu- ially Er-la, performed better at ph 3. and 3.2. At ph 4., at the begiinning sion of eah trial, viable ell ounts in wine stored at 2 C all strains grew rapidly to high ell densities, between 4 x A n B 1._o 11 v Months at 2C D NM. I. I I.. I. I. I I I Months at 1 5C 1 12 Downloaded from on November 9, 218 by guest C i S ;.) D o 6 I 4-2 Months at 8'C FIG. 2. Relative onentrations of malate in Pinot Noir fermented with L. oenos Ey-2d at 2 C (A), 15 C (B), or 8 C (C) as determined by measuring the approximate size of the mali aid spot on the paper hromatogram, dividing it by the approximate size of the mali aid spot of the uninoulated ontrol, and multiplying by 8 for graphial demonstration.

4 VOL. 55, 1989 MALOLACTIC BACTERIA IN OREGON WINES 213 TABLE 2. Malate onentration found when various L. oenos strains were inubated in MRV-8 broth at different ph values for the days indiated at 2 C Malate onn (mg/liter) after no. of days Strain ph 4. ph 3.5 ph 3.2 ph PSU-1 3,14 1, , ,82 3,188 3,91 2,58 1,998 2,26 ML-34 3, ,117 3,43 1,14 3,139 3,114 3,191 2,141 1,38 2,219 MLT-kli 3,333 2, , ,157 2,124 2, , , ,11 2,499 2,463 2,125 2,1 2,11 Er-la 3, , ,111 2,466 1,9 2, Er-lb 3, , ,143 2,27 1,633 2,58 1,181 1,266 Er-l 3, , ,24 2,499 2,428 2, Er-3b 3, ,152 1, ,85 2,621 2,592 2, Er-4a 3, , ,14 2,537 2,341 2, Ey-la 3,243 1, , ,291 2,362 2,36 2,19 2,99 2,17 Ey-2 3, ,98 2, ,136 2,72 2,76 2,77 1,425 1,365 Ey-2d 3,21 1,117 3,133 3,78 2,85 2,921 1,938 1,616 2,96 2,26 2,143 Ey-4b 3,349 1, ,165 3,262 2,711 3,146 2,22 1,591 2,135 2,311 2,34 18 and 2 x 19 CFU/ml, and MLF was ompleted in most ultures before maximum ell density was reahed after 3 days. At ph 3.5, growth of all strains was slower but maximum ell densities were similar to those obtained in media at ph 4.. Also, the rate of malate degradation was slower. The rate of malate atabolism was fastest in the ultures inoulated with the highest ell density (PSU-1 and ens 44.4). The medium with ph 3.2 was inoulated with a large number of bateria; all wines ontained, after inoulation, approximately 17 CFU of viable ells per ml. At this ph, the ell density did not inrease above the initial ell density. MLF was fastest in the ultures whih maintained the highest density of viable ells (>16 CFU/ml). The number of viable ells of PSU-1 and ML-34 dereased rapidly after inoulation, and MLF did not ommene. Only six Oregon isolates were able to metabolize malate at ph 3.. These ultures ontained the highest number of viable ells over the inubation period, with the exeption of MLT-kli and Ey-2d. At ph 3.2, strain MLT-kli was by far the most ative of all strains, with a malolati ativity of This strain nearly ompleted MLF within 24 days. Others whih were more ative than the majority of the group (average of all 13 strains was 5.67; standard deviation, 4.75) were strains Ey-4b, Ey-2d, and Er-la. ML-34 and PSU-1 were less tolerant to low ph; they did not show any malolati ativity at ph 3.2 after 24 days. At ph 3., the most ative strain was Er-la. Five other Oregon isolates also showed signifiant malolati ativities: Er-4a, Er-i, Er-3b, Er-lb, and Ey-2. These six strains also had some malolati ativity at ph 2.8 after 44 days (data not shown). Overall, the Oregon isolates degraded malate more atively at low ph than did the ommerial strains used for omparison. Ethanol tolerane. ML-34, Er-la, and the indigenous baterial flora in Chardonnay wine showed the same sensitivity towards ethanol. Their growth and onurrent degradation of mali aid were similarly inhibited with inreasing ethanol onentration (data not shown). Malate degradation was omplete within 44 days at 1% (vol/vol) ethanol, whereas it was 5% slower at 12% (vol/vol) and ompletely inhibited at 14% (vol/vol). Following inoulation, the number of viable bateria dereased rapidly by 2 to 4 logs with larger dereases at the higher ethanol onentrations. The ell numbers of ML-34 dereased less than the ell numbers of Er-la at all three ethanol onentrations. However, following this initial derease in ell numbers, Er-la regained its viability more rapidly than ML-34 in all three wines. The number of viable bateria in the uninoulated ontrol inreased from less than log 2 CFU/ml at the start of the experiment to log 6.83 at 1% ethanol, log 6.11 at 12% (vol/vol) ethanol, and log 6.4 at 14% (vol/vol) ethanol by day 44. Thus at day 44 the number of viable ells in the uninoulated ontrol wines were almost idential with that of the inoulated wines at 1 and 12% (vol/vol) ethanol. The rates of mali aid degradation were almost idential in the wine adjusted to 1% ethanol. At 12% ethanol, the mali aid degradation rates of the inoulated wines were faster than in the uninoulated ontrol. At 14% (vol/vol) ethanol, however, the number of viable ells in the ontrol wine was 1 log higher at days 44 and 59 than the number of viable ells in the same wine inoulated with ML-34 or Er-la. At 14% ethanol, the mali aid degradation was fastest in the uninoulated wine whih had the 1-fold-higher number of viable ells at day 44, when the degradation of mali aid started. Fumarate and sulfur dioxide. The presene of.5 and 1.5 g of fumarate per liter in wine of ph 3.3 inhibited MLF by ML-34 and Er-la for at least 7 days, at whih point the experiment was stopped (Fig. 3A). The number of viable ells dereased in these wines from 5 x 17 CFU/ml at inoulation to less than 1 x 12 CFU/ml after day 31 (Fig. 3B). The ell numbers dereased more sharply for ML-34 than for Er-la. In the wine with no fumarate added, both strains degraded all malate within 72 days. In the uninoulated ontrol wines, with and without fumarate added, no signifiant redution in the onentration of mali aid ourred over 72 days. In the presene of 26 mg of free and 49 mg of total S2 per liter at ph 3.3 (3 mg of added SO2 per liter, ph 3.3), MLF was inhibited in all wines with or without fumari aid added (data not shown). The number of viable ells was redued to 1 X 13 CFU/ml and less. In the presene of fumarate, the number of viable ells dereased to less than 1 CFU/ml by day 69; without the addition of fumarate, the number of viable ells dereased rapidly from log 1.6 x 17 CFU/ml to 1.6 x 14 CFU/ml by day 4; thereafter it slowly dereased to 2 x 13 CFU/ml by day 69 (data not shown). The number of viable ells for the wines inoulated with Er-la, with 26 mg of free and 49 mg of total SO2 per liter at ph 3.3, with or Downloaded from on November 9, 218 by guest

5 214 HENICK-KLING ET AL. A mg/l 4 mg 5 X3-1 - Omg/L v B I Days at 2C; ph mg--l 2 -l1oomg/l 1 5 mg/l Days at 2C; ph 3.5 FIG. 3. Effet of three onentrations of fumari aid on the rate of malate redution (A) and number of viable ells (B) of L. oenos ML-34 in Chardonnay wine at 2 C, ph 3.5, 8 and 13 mg of free and total SO2, respetively, per liter, and 9.7% (vol/vol) alohol. without fumarate, also dereased dramatially, from 1.6 x 17 CFU/ml at inoulation to between 1 x 13 and 1 x 14 CFU/ml by day 4, and remained nearly onstant until the end of the experiment at day 72. The number of viable ells in the wine with 3 mg of SO2 per liter but no fumarate added started inreasing after day 4 but never beame high enough to degrade any large amounts of mali aid. Sulfur dioxide and ph. At ph 3.3 in the Chardonnay wine with no additional S2 (8 mg of free and 13 mg of total SO2 per liter), ML-34 ompleted MLF by day 46 (Fig. 4A). Er-la ompleted MLF within 37 days. The addition of 2, 3, and 4 mg of S2 per liter 24 h before inoulation delayed MLF by both strains by between 6 and 95 days (Fig. 4A). Without the addition of SO2, the number of viable ells was redued between.5 log/ml for Er-la and 1 log/ml for ML-34 in the wine without addition of SO2. Additions of 2 and 3 mg of S2 per liter redued the number of viable ells to approximately log 4 and log 2 CFU/ml, respetively; the addition of 4 mg of SO2 per liter effetively eliminated all viable ells in the wine (Fig. 4B). After 6 to 9 days, MLF ourred in all wines exept in the wine with 3 mg of added SO2 per liter inoulated with Er-la, where no MLF ourred within 12 days (Fig. 4). DISCUSSION All strains in the temperature trial had rather long fermentation times, onsidering that the wine in whih they were A" 5 '4 e )~ ~ ~ m Iz o 2-7 %^4 mg/l 1 - mg/l 2 mg/l B Days at 2C; ph OmgI O mg/l Daysat22;pH3.3 APPL. ENVIRON. MICROBIOL. FIG. 4. Effet of three additions of SO2 at ph 3.3 on the rate of malate redution (A) and number of viable ells (B) of L. oenos ML-34 in Chardonnay wine at 2 C, 8 and 13 mg of free and total SO2, respetively, per liter, and 9.7% alohol. SO2 additions were made to the wine 24 h before inoulation with the bateria; the onentrations of free and total SO2 in the wines at inoulation were 8 and 13, 13 and 3, 2 and 4, and 26 and 49 mg/liter for additions of, 2, 3, and 4 mg of SO2, respetively, per liter. studied was relatively high in ph and low in aidity and did not ontain any alohol or free S2 after being autolaved. Autolaving the wine might have altered its omposition so that it beame less fermentable; that is, nutrients and growth fators ould have been destroyed or inhibitory substanes ould have been produed. Wines produed from heattreated juie (thermovinifiation) have been found to be less suitable for growth of LAB than those produed without heat treatment (5, 15, 19). However, for this omparison we assumed that all bateria tested were affeted to the same degree. At ph 4., near-optimal onditions for survival and growth of L. oenos allowed all strains to omplete MLF within 3 to 5 days; also, at ph 3.5, most of the strains ompleted MLF within 9 days. For a suessful inoulation whih would initiate MLF within days after inoulation, it has been reported that a wine should be inoulated with a minimum of 16 CFU of viable bateria per ml (3, 11, 2, 26, 29). Beelman and Gallander (3) found that an atively fermenting ulture in wine ontained more than 17 viable ells per ml. This effet, of a minimum number of viable ells for effetively onduting MLF, is also evident from the data reported here. The amount of malate degraded in a ulture Downloaded from on November 9, 218 by guest

6 VOL. 55, 1989 depends on the number of viable ells and the metaboli ativity of the individual ells. In all experiments, the arrest of MLF ould be attributed to loss of viability of the baterial ulture. At ell densities below log 5 to log 6 CFU/ml, the amount of malate degraded an be so small that it is not deteted in the samples taken. Beause of logarithmi inreases in the number of viable ells, the degradation of malate appears to be more rapid at ell densities above 16 CFU/ml. For the seletion of bateria most suited to fermenting malate at low ph, the most important results are in the data for ph 3.2 and 3.. Bateria seleted for indution of MLF in low-ph wines must be able to grow and ferment malate over the ph range from 2.8 to 3.5, whih is typial for wines from ool limates (26, 29). In this range, ph 3.2 is a ritial point below whih it is very rare to see a spontaneous MLF and also very diffiult to indue a MLF by baterial inoulation (6, 12, 18). Several Oregon isolates were onsistently more ative at low ph than the ommerial strains. Table 2 shows the superior ability of several Oregon isolates to ferment mali aid at low phs of 3.2 and 3.. The most ative strain at these phs was Er-la. Comparisons of the malolati ativity of individual strains indiate strain-speifi differenes in malate degradation. These differenes might be due to different viability of the ells or to different ativity of malate transporting and degrading systems. Strain ML-34 differs from Ey-2d and Ey-4b in that it degrades mali aid very rapidly one it begins MLF, while Ey-2d and Ey-4b are slower in degrading mali aid one they begin fermenting. It was notied that the Ey isolates were generally more old tolerant than the Er isolates, while the Er isolates were more low-ph tolerant than the Ey isolates. The ellar temperature at winery EY, from whih the Ey isolates were isolated, is normally lower than the ellar temperature at winery KE, from whih the Er isolates were isolated. The onsistantly lower ellar temperatures at winery EY apparently seleted for bateria that are more tolerant to low temperature. Information as to whether the wines at winery KE are usually lower in ph than the wines at winery EY was not available. Silver and Leighton (27, 28) also found that.4 to.7 g of fumarate per liter inhibited growth and MLF by L. oenos at phs below 4.. The addition of 3 mg of SO2 per liter 24 h before inoulation, giving 2 mg of free and 4 mg of total SO2 per liter, was suffiient to inhibit MLF by both strains of L. oenos, with or without the addition of fumarate, whih onfirms other published data (31). The bateriidal effets of SO2 and fumarate seemed to be additive for both strains. No spontaneous MLF ourred in any of the ontrol wines, whih does support the view that it is advantageous to inoulate wines, even those at ph 3.5, but with low onentrations of SO2 (8 mg of free and 13 mg of total SO2 per liter) and alohol (9.7%) and at a favorable temperature. The results of the S2-pH trial demonstrate (i) the inreasing bateriidal effet of inreasing onentrations of SO2, (ii) the ph dependene of the bateriidal effet of SO2, (iii) that the prevention of MLF by one addition of SO2 is not permanent, and (iv) that strains ML-34 and Er-la are equally sensitive to additions of SO2. These observations are in agreement with previous reports (3), suggesting that most or all strains of L. oenos are equally sensitive to even small amounts of SO2. From the present work, strains Er-la and Ey-2d were seleted over ML-34, PSU-1, ens 44-4, and MLT-kli for MALOLACTIC BACTERIA IN OREGON WINES 215 ommerial trials beause of their improved MLF abilities under low temperatures and low phs. ACKNOWLEDGMENTS We thank Barney Watson for helpful disussions and tehnial assistane. We further aknowledge the tehnial assistane of Liz Dodd, Cathy Haskell, Dave Lattin, Dave Ceryl, and Jennie Wade. Appreiation also is expressed to ooperating Oregon wineries and to R. Avedoveh, Jr. T. Henik-Kling was a Tartar award reipient. LITERATURE CITED 1. Amerine, M. A., and C. S. Ough Methods for analysis of musts and wines. John Wiley and Sons, In., New York. 2. Aney, W. W Oregon limates exhibiting adaptation potential for vinifera. Am. J. Enol. Viti. 25: Beelman, R. B., and J. F. Gallander The effet of grape skin treatments on indued malolati fermentation in Ohio wines. Am. J. Enol. Viti. 21: Beelman, R. B., A. Gavin, and R. M. Keen A new strain of Leuonosto oenos for indued malo-lati fermentation in Eastern wines. Am. J. Enol. Viti. 28: Beelman, R. B., F. J. MArdle, and G. R. Duke Comparison of Leuonosto oenos strains ML-34 and PSU-1 to indue malo-lati fermentation in Pennsylvania red table wines. Am. J. Enol. Viti. 31: Castino, M., L. Ussegio-Tomasset, and A. Gandini Fators whih affet the spontaneous initiation of the malo-lati fermentation in wines. The possibility of transmission by inoulation and its effets on organolepti properties, p In J. G. Carr, C. V. Cutting, and G. C. Whiting (ed.), Lati aid bateria in beverages and food. Aademi Press, In., New York. 7. Davis, C. R., D. Wibowo, R. Eshenbruh, T. H. Lee, and G. H. Fleet Pratial impliations of malolati fermentation: a review. Am. J. Enol. Viti. 36: Dittrih, H. H Mikrobiologie des Weines. Handbuh der Getraenketehnologie. Verlag Eugen Ulmer, Stuttgart, Federal Republi of Germany. 9. Garvie, E Leuonosto oenos sp. nov. J. Gen. Mirobiol. 48: Garvie, E., and J. A. E. Farrow The differentiation of Leuonosto oenos from non-aidophili speies of Leuonosto and identifiation of five strains from the Amerian Type Culture Colletion. Am. J. Enol. Viti. 31: Hayman, D. C., and P. R. Monk Starter ulture preparation for the indution of malo-lati fermentation in wine. Food Tehnol. Aust. 34: Henik-Kling, T Improving malolati fermentation, p In R. E. Smart, R. J. Thornton, S. B. Rodriguez, and J. E. Young (ed.), Proeedings of the Seond International Symposium for Cool Climate Vitiulture and Oenology. New Zealand Soiety for Vitiulture and Oenology, Aukland, New Zealand. 13. Ingraham, J. L., R. H. Vaughn, and G. M. Cooke Studies on the malo-lati organisms isolated from California wines. Am. J. Enol. Viti. 11: Izuagbe, Y. S., T. P. Dohman, W. E. Sandine, and D. A. Heatherbell Charaterization of Leuonosto oenos isolated from Oregon wines. Appl. Environ. Mirobiol. 5: Kunkee, R. E Malo-lati fermentation. Adv. Appl. Mirobiol. 9: Kunkee, R. E Simplified hromatographi proedure for detetion of malo-lati fermentation. Wines and Vines 49(3): Lafon-Lafourade, S Fators of the malo-lati fermentation of wines, p In J. G. Carr, C. V. Cutting, and G. C. Whiting (ed.), Lati aid bateria in beverages and food. Aademi Press, In., New York. 18. Lafon-Lafourade, S., E. Carre, A. Lonvaud-Funel, and P. Ribereau-Gayon Indution de la fermentation malola- Downloaded from on November 9, 218 by guest

7 216 HENICK-KLING ET AL. tique d'une biomasse industrielle ongellee de L. oenos apres reativation. Connaiss. Vigne Vin 17: Martiniere, P., J. C. Sapis, and P. Ribereau-Gayon Evolution de nombre de bateries latiques vivantes au ours de la vinifiation et de onservation des vins. C. R. Seanes Aad. Agri. Fr. 6: Mayer, K Neue Erkenntnisse of dem Gebiet des biologishen Saureabbaus. Shweiz. Z. Obst-Weinbau 112: MCloskey, L. P Enzymati assay for mali aid and malao-lati fermentations. Am. J. Enol. Viti. 31: Radler, F Die mikrobiologishen Grundlagen des Saureabbaus im Weim. Zentralbl. Bakteriol. Parasitenkd. Abt. 2 12: Radler, F Die organishen Sauren im Wein and ihr mikrobieller Stoffwehsel. Dtsh. Rundsh. 71: Radler, F Mirobial biohemistry. Experientia 42: Radler, F., and K. Broehl The metabolism of several arboxyli aids by lati aid bateria. Z. Lebensm.-Unters.- Forsh. 179: Rie, A. C., and R. L. Mattik Natural malo-lati APPL. ENVIRON. MICROBIOL. fermentation in New York State wines. Am. J. Enol. Viti. 21: Silver, J., and T. Leighton Control of malo-lati fermentation in wine. 2. Isolation and haraterization of a malolati organism. Am. J. Enol. Viti. 32: Silver, J., and T. Leighton Control of malolati fermentation in wine. 1. Mehanisms of fumari aid inhibition and onsiderations onerning malate metabolism by lati aid bateria, p In D. Webb (ed.), Proeedings of the University of California, Davis, Grape and Wine Centennial Symposium. University of California, Davis. 29. Vetsh, U Untersuhungen zur Vermehrung von Bakterium graile (Leuonosto oenos) waehrend des biologishen Saeureabbaus in Wein. Shweiz. Z. Obst- Weinbau 19: Watson, B. T., Jr., E. E. Dodd, and D. Heatherbell Maturity data and wine analysis for some premium Oregon grape varieties. Pro. Oreg. Horti. So. 71: Wibowo, D., R. Eshenbruh, C. R. Davis, G. H. Fleet, and T. H. Lee Ourrene and growth of lati aid bateria in wine: a review. Am. J. Enol. Viti. 36: Downloaded from on November 9, 218 by guest