APPLIED AND ENVIRONMENTAL MICROBIOLOGY, Mar. 1986, p. 539-545 Vol. 51, No. 3 0099-2240/86/030539-07$02.00/0 Copyright C 1986, American Society for Microbiology Growth and Metabolism of Lactic Acid Bacteria during and after of at Different pH C. R. DAVIS, D. J. WIBOWO, T. H. LEE,t AND G. H. FLEET* School ofFood Science and Technology, The University of New South Wales, Kensington, New South Wales, 2033 Received 16 September 1985/Accepted 9 December 1985

Commercially produced red wines were adjusted to pH 3.0, 3.2, 3.5, 3.7, or 4.0 and examined during and after malolactic fermentation for growth of lactic acid bacteria and changes in the concentrations of carbohydrates, organic acids, amino acids, and acetaldehyde. With one exception, Leuconostoc oenos conducted the malolactic fermentation in all wines and was the only species to occur in wines at pH below 3.5. Malolactic fermentation by L. oenos was accompanied by degradation of malic, citric, and fumaric acids and production oflactic and acetic acids. The concentrations of arginine, histidine, and acetaldehyde also decreased at this stage, but the behavior of hexose and pentose sugars was complicated by other factors. Pediococcus parvulus conducted the malolactic fermentation in one containing 72 mg of total sulfur dioxide per liter. Fumaric and citric acids were not degraded during this malolactic fermentation, but hexose sugars were metabolized. P. parvulus and species of Lactobacillus grew after malolactic fermentation in wines with pH adjusted above 3.5. This growth was accompanied by the utilization of wine sugars and production of lactic and acetic acids.

Lactic acid bacteria make a significant contribution to the changes in the concentration of constituents which affect the quality of wines by conducting malolactic fermentation and sensory quality of wines. The metabolism of wine compo- by causing spoilage at later stages of vinification (3, 17, 18, nents by lactic acid bacteria after malolactic fermentation is 20). Wine composition and, consequently, wine quality are largely unstudied. Quantitative studies that correlate the altered by these bacteria, because some wine components kinetics of bacterial growth in wines with changes in the are utilized as substrates for growth and metabolic end concentration of specific wine constituents have not been products are excreted into the wine. conducted. Consequently, the biochemical mechanisms by Many qualitative studies (for a review, see references 3, which lactic acid bacteria grow in wines are still poorly 17, and 18) and more recent quantitative data (9, 11, 19) have understood. shown that Leuconostoc oenos is the species most fre- In this paper we report the enumeration, isolation, and quently occurring in wine and is mainly responsible for the identification of lactic acid bacteria from several wines at malolactic fermentation. Species of Pediococcus and close time intervals during and after malolactic fermentation. Lactobacillus have been isolated to a lesser extent and are The growth of isolated species is correlated with their more likely to occur at stages after malolactic fermentation utilization of wine sugars, organic acids, and amino acids. in wines of higher pH (3.5 to 4.0). A major weakness of the The effect of wine pH on both the species isolated and their ecological studies so far reported has been the infrequency utilization of wine components is also reported. and irregularity of analysis of wines during the critical periods when lactic acid bacteria grow. Such a problem has a-risen because from which wine samples are ob- MATERIALS AND METHODS tained are generally located at considerable distances from research laboratories (9). Consequently, a complete ecolog- Wine samples. wine was taken from wineries A and ical description of the growth of lactic acid bacteria in a B (50 liters from each) immediately after completion of particular wine at close time intervals during vinification is alcoholic fermentation (about 7 days after the crushing of the still lacking; this ecology would be particularly complex in ) and was transported to the University of New South the higher-pH wines, in which the successive growth and Wales laboratory for storage and analysis. The two wineries death of several species may occur (9, 11). are located in the Hunter Valley district of New South Wales, The chemical changes in the composition of wine that are Australia. At the laboratory, the wines were racked off the caused by lactic acid bacteria are not well defined, except for lees and dispensed as 4-liter volumes in sterile glass flasks that the conversion of malic acid to lactic acid during malolactic were fitted with air locks. Duplicate volumes of each wine fermentation by L. oenos and to a lesser degree by species of were adjusted to the desired pH by the addition of either Pediococcuts and Lactobacillus (3, 17, 18, 20, 27). Isolated sterile 50% tartaric acid or 5 M potassium hydroxide. The pH studies suggest that wine carbohydrates (10, 24, 28) and of one set of duplicates remained unadjusted. Wines were amino acids (3, 23, 34) may be utilized by these bacteria stored at 20°C and were sampled regularly for analysis of during the malolactic fermentation and that this metabolism lactic acid bacteria, organic acids, monosaccharides, pH, and as well as that of organic acids (3, 17, 20, 27, 28) can lead to (at some sampling times) acetaldehyde and amino acids. Samples (30 ml) for analysis were taken from the center ofthe were * Corresponding author. flask after thorough mixing of the contents. The flasks t Present address: The Research Institute, Glen flushed with CO2 at each sampling. Microbiological and pH Osmond, South Australia, Australia 5064. analyses were conducted immediately after sampling. The 539 540 DAVIS ET AL. APPL. ENVIRON. MICROBIOL.

sample was then filtered through a membrane (pore size, 0.45 ,um) and frozen until required for chemical analysis. Wines from several other wineries were obtained and stored as described above, but without pH adjustment. Enumeration and identification of lactic acid bacteria. Lac- tic acid bacteria were enumerated by spread-inoculating 0.1 ml of the wine sample (diluted if necessary) over the surface of plates of MRS agar (Oxoid Ltd., London, England) and tomato juice agar (9, 25). Plates were incubated in an atmosphere of 5% CO2 for 7 days at 30°C. Plating media contained 100,ug of cycloheximide per ml to suppress the O A ~~~~A--pH Z 1-0 * growth of yeasts. Colonies that grew on the surface of the 13.. .. plates and which were gram positive and catalase negative (9, 25) were counted and identified as lactic acid bacteria. M Several representatives of each colony type were subcul- 0 tured and identified to genus and species level by the tests 320 c oo described previously for L. oenos (13), for Pediococcus A3,,, parvulus (2, 14), and for isolates of Lactobacillus (29, 30). Carbohydrate fermentation tests were conducted in API 50 160 - ]400 CH galleries (La Balmes les Gottes-38390, Montalieu Vercien, France). Chemical analyses. Hexose and pentose sugars were sep- arated and determined by gas-liquid chromatography (GLC) 10 5 0 >%40 -d of their per-o-acetylaldonitrile derivatives. Glucitol and 0 myo-inositol were also determined by GLC but as acetate Procedures for the derivatization of the sugars derivatives. 20 have been previously described (35). The derivatized sugars - A were separated by injection into a 2100 Aerograph chromato- A graph (Varian Associates, Palo Alto, Calif.) equipped with a glass column (1 m by 3 mm inner diameter) packed with 10% 0 A-A DEGS on Chromosorb W/AW 80/100; nitrogen was the 10 50 90 130 carrier gas at a flow rate of 60 ml/min. Injector and detector Time after crushing (days) were 190 and 250°C, respectively, and the oven temperatures FIG. 1. Effect of growth of lactic acid bacteria on pH and temperature was programmed from 140 to 190°C at 1°C/min. concentration of organic acids and carbohydrates during and after Before derivatization, wine samples were treated to remove the malolactic fermentation of Shiraz wine (pH 3.2) from A. substances that interfered with the resolution of the sugars (a) L. oenos (0). (b) Malic acid (0), lactic acid (0), citric acid (E), by GLC. This involved passage of 1.0 ml of the wine sample acetic acid (U), fumaric acid (A), and pH (A). (c) Glucose (0), first through a small column (1.0 ml) containing mixed-anion fructose (0), Ul (LI), myo-inositol (A), and glycerol (0). (d) (Bio-Rad AG-1 x 8, acetate form, 100/200 mesh; Bio-Rad Rhamnose (0), ribose (OI), arabinose (-), xylose (A), and mannose Laboratories, Richmond, Calif.) and cation (Dowex-1 x 8, (A). The concentration of galactose (80 mg/liter) and that of glucitol H+ form, 100/200 mesh) resins and then through a C18 (80 mg/liter) remained constant throughout storage and are not Sep-Pak (Waters Associates, Inc., Milford, Mass.). The shown. columns were eluted with distilled water, and the eluate containing the sugars was concentrated by freeze-drying. The concentration of malic acid was also determined by an The dried residue was used for derivatization. Studies with enzymatic method with a kit from Boehringer. spiked wine samples indicated that the pretreatment and The free amino acids were separated by HPLC and derivatization procedures gave greater than 85% recovery of estimated by post-column derivatization with ortho- each sugar. The identification of wine carbohydrates was phthalaldialdehyde (16). The technology and equipment used confirmed by combined GLC-mass spectrometry on a for these analyses were from Waters Associates. The chro- Finnigan 4021 instrument at the Australian Wine Research matograph was equipped with an amino acid analysis column Institute. (Waters Associates) that was gradient eluted with sodium GLC performed by the method described above did not eluent A and sodium eluent B buffers (Waters Associates). satisfactorily separate either glucose from mannitol or man- Acetaldehyde was measured enzymatically by using a kit nose from 2-deoxyglucose. Consequently, the concentra- from Boehringer. tions of glucose and mannose, as well as those of fructose and glycerol, were determined enzymatically by using kits RESULTS from Boehringer GmbH, Mannheim. Shiraz wine from winery A. The alcohol and total sulfur Organic acids were measured by high-pressure liquid dioxide (SO2) concentrations of the Shiraz wine from winery chromatography (HPLC) with a chromatograph (Waters A were 14.2% and 11.2 mg/liter, respectively. Samples were Associates) equipped with an ion exclusion, cation exchange adjusted to pH 3.0, 3.2, 3.7, or 4.0, and one sample remained column (Bio-Rad HPX-87H) and a detector operating at 210 unadjusted at pH 3.5. The growth of bacteria in the wines at nm. The column was eluted with 0.1% phosphoric acid at pH 3.2 and 3.7 is shown in Fig. 1 and 2 along with some 55°C at a flow rate of 0.5 ml/min. Before HPLC, wine changes in wine composition. samples were filtered through a membrane (pore size, 0.45 L. oenos was present at initial levels of 10 to 100 cells per ,um) and passed through a C18 Sep-Pak. This procedure ml, and after a lag period of 1 to 2 days, grew to a maximum yielded >95% recovery of the organic acids found in wine. population of 107 to 108 cells per ml in wines at pH 3.2, 3.5, VOL. 51, 1986 LACTIC ACID BACTERIA IN WINE 541

identification studies are required. Increases in the concen- trations of glucose and fructose and decreases in the con- centrations of mannose and arabinose were also observed in IL. the wine at pH 3.0, in which there was no bacterial growth and no malolactic fermentation. The microbiological and chemical changes that occurred 0 in the wines after malolactic fermentation depended upon pH. Data for the wines at pH 3.2 and 3.7 are presented in 1 and L. oenos survived at a level of - Fig. 2, respectively. _J cn E approximately 106 cells per ml in the wine at pH 3.2 and was the only species isolated from this wine. The concentrations .0 of glucose and fructose continued to increase, and those of ._i mannose and arabinose continued to decrease. The concen- 10 C_ .W m trations of to a lesser rhamnose E ribose, xylose, and, extent, _J D) increased slightly. L. oenos remained at a level of 106 cells per ml in the wine at pH 3.5 the an N-- during postmalolactic period. However, c' unidentified species of Lactobacillus was also isolated from this wine at 60 days after crushing, and it subsequently grew to a population of 105 cells per ml to coexist with L. oenos. cN 00 G 0 The growth of this coincided with small but C species signifi-

0 cant decreases in the concentrations of glucose, fructose, c: myo-inositol, and ribose, indicating that they were being utilized as growth substrates. The concentrations of arabinose and mannose continued to decrease in the same fashion as did the wine at pH 3.2 and did not specifically correlate with the growth of the Lactobacillus sp. Both P. parvulus and Lactobacillus cellobiosus were isolated from the wine at pH 3.7 during the postmalolactic fermentation period. These species grew to populations of 106 to 107 cells per ml, and when this occurred, L. oenos was no longer isolated from the wine (Fig. 2). The growth of P. parvulus and Lactobacillus cellobiosus was accompanied by Time after crushing (days) substantial changes in the chemical composition of the wine. FIG. 2. Effect of growth of lactic acid bacteria on pH and Lactic and acetic acid concentrations increased from 1.8 and concentration of organic acids and carbohydrates during and after 0.4 g/liter, respectively, at the end of malolactic fermentation the malolactic fermentation of Shiraz wine (pH 3.7) from winery A. to 3 and 1.65 g/liter, respectively, and wine pH decreased (a) L. oenos (0), P. parvulus (A), and Lactobacillus cellobiosus (O). from 3.95 to 3.72. The concentrations of glucose, fructose, Symbols for panels b, c, and d are the same as for Fig. 1. myo-inositol, glycerol, ribose, xylose, and Ul decreased markedly (Fig. 2). The concentration of arabinose also 3.7, and 4.0. The growth rate increased as the wine pH was decreased, although not in coincidence with bacterial increased. The maximum population was achieved at 44 growth. Mannose, after virtually complete utilization, later days after crushing in wine at pH 3.2 and at 17 days after increased to a level between 25 and 30 mg/liter. crushing in wine at pH 4.0. No growth occurred in wine at P. parvulus and an unidentified species of Lactobacillus pH 3.0. The malolactic fermentation occurred in conjunction grew in the wine at pH 4.0 during the period of postmalo- with the growth of L. oenos and was indicated both by the lactic fermentation and achieved populations of 107 cells per complete degradation of malic acid and by a parallel increase ml by 30 and 45 days, respectively, after crushing. in concentration of lactic acid to about 1.8 g/liter. The pH of However, both L. oenos and the Lactobacillus sp. disap- the wines increased by approximately 0.3 U during this peared by 71 days, leaving P. parvulus as the surviving period. No species other than L. oenos was detected in the organism. The chemical changes that occurred in this wine wines during the malolactic fermentation. were similar to those of the wine at pH 3.7 except that Other chemical changes that occurred during malolactic galactose was also utilized. fermentation in wines at pH 3.2 and 3.7 are shown in Fig. 1 Acetaldehyde, present at 15 mg/liter at the end of alcoholic and 2, respectively. The trends were the same for the wines fermentation, was partially degraded in all wines during at pH 3.5 and 4.0, although the time scales were different. malolactic fermentation. The proportion degraded increased Citric and fumaric acids were completely degraded, and the from 13.6 to 29% as the wine pH increased from 3.2 to 4.0. concentration of acetic acid increased by about 0.1 g/liter. Table 1 shows the concentration of some amino acids in Glucose concentration doubled from 90 to approximately the wine at pH 3.7 just before malolactic fermentation, at the 160 mg/liter by the end of the fermentation, and fructose end of this reaction (62 days) and at a later stage (100 days) concentration also increased, although to a lesser extent. of storage. With the exception of histidine and arginine, Both arabinose and mannose concentrations decreased dur- which were completely utilized during malolactic fermenta- ing malolactic fermentation. However, the concentration of tion, the concentrations of all individual amino acids in- galactose, ribose, xylose, rhamnose, glycerol, myo-inositol, creased slightly during this fermentation and storage. Similar glucitol (data not shown), and an unidentified monosac- data were found for the wine at pH 3.2, except that the charide (Ul) remained unchanged. Mass spectral data indi- increases in concentrations were slightly smaller. cated that Ul is probably a hexose glycoside, but further Shiraz wine from winery B. Shiraz wine from winery B 542 DAVIS ET AL. APPL. ENVIRON. MICROBIOL. contained 11.9% alcohol and 72 mg of total SO2 per liter. -J Samples were adjusted to pH 3.0, 3.2, 3.7, and 4.0, and one E sample remained unadjusted at pH 3.5. L. oenos, P. parvulus, and Lactobacillus buchneri were isolated from the U- wine at a level of 10 to 100 cells per ml at the end of alcoholic Li fermentation. Thereafter, no lactic acid bacteria were de- 0 tected in the wines at pH 3.0, 3.2, 3.5, or 3.7 at any stage -i during storage for 150 days. P. parvulus commenced growth in the wine at pH 4.0 after a lag phase of approximately 40 days (Fig. 3). It achieved a maximum population of almost per 107 cells ml and survived in the wine until the end of 10 sz>b A ~~~~~4-1/;PHp2.e sampling. Malolactic fermentation occurred concomitantly Y A 3.9 with the growth ofP. parvulus. Citric and fumaric acids were 0 ~~~~~~ H 0 e not degraded during this period, and acetic acid was not produced; the concentrations of acetic and lactic acids, however, increased substantially during storage after 05t~~~ 160 A AA =| malolactic fermentation (Fig. 3). The concentrations of glu- A cose and fructose in the wine increased before malolactic 0l A ___ { 600 fermentation, but then with decreased along those of galac- . 600 ' tose and Ul during the fermentation. The concentrations of '16 J -. 3- 01/0A arabinose, xylose, and glucitol remained constant during and X~~~~~~~~~~~~A-. ,, 400. after malolactic fermentation, but those of ribose and myo- 0~~~~~~~~~~~~~ inositol decreased slightly at the time that acetic acid was produced. Mannose, which was not detected in the wine do 10 30 50 789 4 20 before malolactic fermentation, subsequently increased to 26-0 0-0 ' about 40 mg/liter. A 0 Other wines. L. oenos was the only species isolated from Shiraz and wines obtained from seven A~~~~ other wineries throughout Australia. This species conducted A the malolactic fermentation during laboratory storage of 01 lo~~~~~~~~~~~~~~~~ these wines at their unadjusted pH (3.5 to 3.6). In all wines, ()L. oeo10_ (O,P30 avls(O,adLcoailu50 70 9 '140 uhei()150 (O),arabnos (A,adanse() the malolactic fermentation was accompanied by the degra- Time after crushing (days) dation of citric and fumaric acids and production of acetic FIG. acid (Fig. 1). Changes in the concentration of sugars were 3. Effect of growth of lactic acid bacteria on pH and concentration of organicTh acidsocnrtoand carbohydratesfrbs,aaioeduring and after the malolactic20ut 40talafermentation of Shiraz wine (pH 4.0) from winery B. (a) L. oenos (0), P. parvulus (Li), and Lactobacillus buchneri (A). (b) Malic acid (0), citric acid (El), lactic acid (0), acetic acid (U), TABLE 1. Effect of malolactic fermentation and subsequent fumaric acid (A), and pH (A). (c) Glucose (0), fructose (0), conservation on the concentration of some amino acids of a Shiraz galactose (U), Ul (El), and myo-inositol (A). (d) Xylose (0), ribose wine at pH 3.7 (o),arabinose (p),and mannose (A). Conc (mmol/liter) measured for a Shiraz wine from the McLaren Vale district, After AminoAminoacid Before(10 completion(62After storage South Australia. The concentration of ribose, arabinose, BeMFore1(1 of MLF and of Ul decreased daySdas" (62 dayS)b 10dy) myo-inositol, particularly during malolactic fermentation; the concentration of the other sug- Aspartic acid 0.015 0.055 0.045 ars, including glucose and fructose, remained constant. L. Threonine + asparaginec 0.085 0.125 0.17 oenos (106 cells per ml), P. parvulus (105 cells per ml), and Serine + glutamined 0.085 0.175 0.09 Lactobacillus brevis (103 cells per ml) were isolated from one Glutamic acid 0.08 0.16 0.18 Shiraz wine at pH 3.85 on completion of the malolactic Glycine 0.085 0.16 0.205 fermentation. Alanine 0.15 0.235 0.21 Valine 0.005 0.06 0.105 DISCUSSION Methionine NDe 0.01 0.01 Isoleucine 0.005 0.035 0.055 The growth cycle of lactic acid bacteria in wines is Leucine 0.015 0.115 0.135 complex and consists of distinct phases that occur during Tyrosine 0.03 0.056 0.06 alcoholic fermentation, malolactic fermentation, and conser- Phenylalanine ND 0.045 0.055 vation after malolactic fermentation (11, 19, 28). In this a-Aminobutyric acid 0.085 0.145 0.195 study we give a detailed, quantitative description of the Lysine 0.01 0.095 0.075 ecological and chemical behavior of lactic acid bacteria Histidine 0.035 ND ND during and after malolactic fermentation. Overall, our data Arginine 0.02 ND ND strengthen and extend the conclusions of previous studies (3, a MLF, Malolactic fermentation. 17, 18, 20). b Days after crushing (Fig. 2). The effect of pH on the growth rate of lactic acid bacteria ' Acids not resolved on chromatogram; concentration expressed as milli- moles of threonine per liter. in wines is well demonstrated in the literature (4, 5, 22) and d Acids not resolved on chromatogram; concentration expressed as milli- by the data in Fig. 1 and 2. Generally, the rate of bacterial moles of serine per liter. growth and malolactic fermentation increased as wine pH e ND, Not detected. was increased from 3.0 to 4.0. In addition, pH had a VOL. 51, 1986 LACTIC ACID BACTERIA IN WINE 543 profound, selective effect upon the species that grow in malolactic fermentation was not accompanied by the specific wines. Usually, L. oenos was the only species isolated from utilization of any hexose or pentose sugars. Strains of L. wines with a pH below 3.5. In such wines, L. oenos grew oenos isolated from these wines fermented glucose and exponentially to conduct the malolactic fermentation, and fructose as determined by the API tests, but, contrary to thereafter remained in a viable but nonproliferating state for expectation, the concentrations of these two sugars in- long periods (Fig. 1). The long-term survival of L. oenos creased significantly during malolactic fermentation. These under practical winery conditions, however, is determined increases were not directly related to growth, as might by the addition of SO2 (19). As noted by ourselves and others appear from Fig. 1 and 2, because they occurred in other (28), Pediococcus and Lactobacillus spp. rarely grow in wines (e.g., the same Shiraz wine at pH 3.0 and the Shiraz wines with a pH below 3.5, but do occur in wines of pH 3.5 wine described in Fig. 3), in which there was no growth ofL. and above, particularly after the malolactic fermentation has oenos. Other researchers have also noted increases in the been completed by L. oenos (Fig. 2). The growth of these concentrations of some wine sugars during malolactic fer- bacteria at this stage seems antagonistic to the survival of L. mentation (8, 10), but the mechanism of this behavior has not oenos (Fig. 2); we have noted this behavior previously (9). yet been satisfactorily explained. Residual enzymatic activ- The molecular basis of this antagonistic interaction and its ities from grapes and yeasts could be involved. The hydro- enological significance warrant further study. lysis of sucrose would produce both glucose and fructose, The concentration of SO2 may also selectively influence but sucrose was not found in our wine samples even before the species of lactic acid bacteria that grow in wines. This is malolactic fermentation (data not shown). Trehalose is an- well illustrated by the Shiraz wine from winery B, which had other important wine disaccharide (32) whose hydrolysis a total SO2 concentration of 72 mg/liter and did not support would yield glucose. We did not specifically analyze this the growth of any lactic acid bacteria until the pH was sugar, but some preliminary measurements indicated that its adjusted to 4.0. Despite the initial presence of L. oenos, concentration decreased during malolactic fermentation. Lactobacillus buchneri, and P. parvulus in this wine, P. The hydrolysis of phenolic glycosides of wine (18) could also parvulus was the only species able to grow, and it conducted lead to an increase in the concentration of monosaccharide the malolactic fermentation (Fig. 3). According to other sugars. However, the utilization of glucose and fructose by studies (18, 33), a total SO2 concentration of >50 mg/liter L. oenos still remains a possibility, since they may be generally restricts the growth of lactic acid bacteria in wines, generated at a rate faster than they are utilized. especially at the lower pH values, when a greater proportion Decreases in the concentration of mannose and arabinose of the SO2 is in the undissociated, antimicrobial form. The during malolactic fermentation (Fig. 1 and 2) were not frequent occurrence ofPediococcus spp. in Australian wines specifically related to the growth of L. oenos, as they also (9, 25) may be indicative of wines with a high concentration occurred under conditions at which no bacterial growth was of SO2 and a high pH (e.g., 3.5 to 4.0). evident (e.g., in the same wine at pH 3.0 and in the Shiraz The populations of 106 to 107 cells per ml formed by lactic wine described in Fig. 3). Moreover, the strains of L. oenos acid bacteria either during or after malolactic fermentation isolated from the wines at this stage were not able to ferment were quantitatively significant and produced measurable arabinose, as measured by the API tests. We are not able to changes in the concentrations of some wine components. provide a satisfactory explanation for these decreases; fur- Although published evidence is not strong, it is generally ther studies are needed. Increases in the concentration of considered that wine sugars are utilized as carbon substrates mannose at later stages of vinification (Fig. 2 and 3) are also for the growth of lactic acid bacteria (18, 24, 28). Our data difficult to interpret, but could possibly arise from the provide experimental evidence to support this conclusion, hydrolysis of mannan polysaccharide of the yeast cell wall but also show that the behavior of sugars in some wines is during autolysis (1). quite complex. As expected from what is now a well-established fact (3, The concentrations of glucose, fructose, and the uniden- 17), malic acid was degraded by L. oenos and P. parvulus tified monosaccharide Ul decreased in direct proportion to during malolactic fermentation, with the concomitant pro- the growth of P. parvulus during the malolactic fermentation duction of lactic acid. The degradation of citric and fumaric of the Shiraz wine at pH 4 from winery B (Fig. 3), and it may acids and production of acetic acid were important second- be concluded that these carbohydrates were metabolized by ary reactions during the malolactic fermentation by L. oenos this species. Similar conclusions can be made for the utili- but not during the malolactic fermentation by P. parvulus. zation of glucose, fructose, Ul, myo-inositol, glycerol, Other researchers (6, 7, 10, 28, 31, 36) have also observed ribose, and xylose during the growth of pediococci and these secondary reactions; there is probably direct metabo- lactobacilli after malolactic fermentation of the Shiraz wine lism of citric acid to acetic acid. The metabolic pathway of at pH 3.7 from winery A (Fig. 2). Metabolism of the hexose this reaction has been studied for dairy lactic acid bacteria and pentose sugars (18, 28) probably accounted for the large (15). The metabolism of citric acid in wine is also related to increases in the concentration of acetic and lactic acids also an increase in the concentration of diacetyl and acetoin (12, observed at this stage. The metabolism of glycerol by lactic 28, 37). acid bacteria was related to production of the bitter sub- Significant changes in the concentration of wine amino stance acrolein (10, 18). The utilization of myo-inositol by acids have been reported to occur during malolactic fermen- wine lactic acid bacteria is not well documented in the tation. Some amino acids decrease in concentration, literature, but was reported earlier by Peynaud and Domercq whereas others increase, but no consistent trends have (26). Further studies are required to understand the metab- emerged for any individual amino acid, except for a decrease olism of this substance as well as that of Ul. Myo-inositol, in arginine, which is probably converted to ornithine (3, 23, Ul, ribose, and arabinose, but not glucose or fructose, were 28, 34). We were able to confirm the utilization of arginine by also metabolized by L. oenos during malolactic fermentation L. oenos during malolactic fermentation, and in addition, we of the Shiraz wine from the McLaren Vale district. noted the utilization of histidine. It is possible that histidine Data for the Shiraz wines (winery A, pH 3.2 or 3.7) (Fig. was decarboxylated to histamine, which is also known to 1 and 2) suggest that the growth of L. oenos during occur in wines (18). An increase in the concentration of all 544 DAVIS ET AL. APPL. ENVIRON. MICROBIOL. other amino acids could be explained by the degradation of 11. Fleet, G. H., S. Lafon-Lafourcade, and P. Ribereau-Gayon. 1984. wine proteins by proteases produced by lactic acid bacteria, Evolution of yeasts and lactic acid bacteria during fermentation but the production of such enzymes by these bacteria has not and storage of Bordeaux wines. Appl. Environ. Microbiol. yet been studied. Dairy lactic acid bacteria are well known 48:1034-1038. 12. Fornachon, J. C. M., and B. Lloyd. 1965. Bacterial production for their protease production (21). Increases in the concen- of diacetyl and acetoin in wine. J. Sci. Food Agric. 16:710- tration of amino acids might also occur through yeast 716. autolysis (1). Amino acid metabolism by wine lactic acid 13. Garvie, E. I. 1974. Genus II. Leuconostoc van Tieghem 1878, bacteria has not been studied. By analogy to dairy fermen- 198, emend. mut. char. Hucker and Pederson 1930, 66, p. tation (21), such metabolism is likely to affect wine quality. 510-513. In R. E. Buchanan and N. E. Gibbons (ed.), Bergey's The HPLC technology that now permits rapid measurement manual of determinative bacteriology, 8th ed. The Williams & of free amino acids will facilitate further study of the Wilkins Co., Baltimore. behavior of amino acids during vinification and should 14. Garvie, E. I. 1974. Nomenclatural problems of the pediococci. enable a better understanding of their contribution to wine Request for an opinion. Int. J. Syst. Bacteriol. 24:301-306. 15. Kempler, G. M. 1983. Production of flavor compounds by quality. microorganisms. Adv. Appl. Microbiol. 29:29-51. The degradation of acetaldehyde by L. oenos is signifi- 16. Klapper, D. C. 1982. New low cost fully automated amino acid cant, since SO2, which is strongly bound to this component, analyses using gradient HPLC, p. 509-515. In M. Elzinga (ed.), is freed to become active against further bacterial growth Methods in protein sequence analysis, vol. 25. The Humana (33). Press, Inc. Clifton, N.J. The data described in this paper, together with those 17. Kunkee, R. E. 1974. Malolactic fermentation and . reported by others (18, 20, 33), have provided further Adv. Chem. Ser. 137:151-170. evidence of the important influence of pH and S02 concen- 18. Lafon-Lafourcade, S. 1983. Wine and brandy, p. 81-163. ln tration upon the conduct of the malolactic fermentation H.-J. Rehm and G. Reed (ed.), Biotechnology, vol. 5, Verlag and, Chemie, Weinheim, Federal Republic of Germany. consequently, upon the quality of red wines. They also 19. Lafon-Lafourcade, S., E. Carre, and P. Ribereau-Gayon. 1983. indicate the potential of bacteria to grow in wines after Occurrence of lactic acid bacteria during the different stages of conclusion of the malolactic fermentation and the need for vinification and conservation of wines. Appl. Environ. Micro- sound winemaking practices to prevent spoilage at this biol. 46:874-880. stage. Furthermore, our findings suggest that the metabolism 20. Lafon-Lafourcade, S., and P. Ribereau-Gayon. 1984. Develop- of wine components, especially the carbohydrates, during ments in the microbiology of wine production, p. 1-45. In M. E. and after malolactic fermentation is more complex than Bushell (ed.), Progress in industrial microbiology. Modern ap- previously thought and requires more detailed investigation. plications of transitional biotechnologies, vol. 1. Elsevier Bio- medical Press, Amsterdam. ACKNOWLEDGMENTS 21. Law, B. A., and J. Kolstad. 1983. Proteolytic systems in lactic acid bacteria. Antonie van Leeuwenhoek J. Microbiol. 49:225- We thank the Australian Wine Research Institute for financial 245. support and the Australian Government for Commonwealth 22. Liu, J., and J. F. Gallander. 1983. Effect of pH and sulfur Postgraduate Awards to C.R.D. and D.W. dioxide on the rate of malolactic fermentation in red table wines. Am. J. Enol. Vitic. 34:44-46. LITERATURE CITED 23. Mayer, K., G. Pause, and U. Vetsch. 1973. Aminosaurengahalte 1. Arnold, W. N. 1981. Autolysis, p. 135-137. In W. N. Arnold in verlauf der vinification einiger rotweine. M. H. (ed.), Yeast cell envelopes. Biochemistry, biophysics, and ultra- Klosterneuburg 23:331-340. structure. CRC Press, Inc., Boca Raton, Fla. 24. Melamed, N. 1962. Determination des sucres residuels des vins, 2. Back, W. 1978. Zur taxonomie der gattung Pediococc us. leur relation avec la fermentation malolactique. Ann. Technol. Phanotypische und genotypische Abgrenzung der bisher bekan- Agric. 11:5-32. nten Arten sowie beschreibung einer neuen Bierschadlichen 25. Pan, C. S., T. H. Lee, and G. H. Fleet. 1982. A comparison of Art: Pediococcus inopinatus. Brauwiss. 31:237-250, 312-330, five media for the isolation of lactic acid bacteria from wines. 338-340. Aust. Grapegrower Winemaker 19:42-46. 3. Beelman, R. B., and J. F. Gallander. 1979. Wine deacidification. 26. Peynaud, E., and S. Domercq. 1961. Etudes sur les bacteries Adv. Food Res. 25:1-53. lactiques des vins. Ann. Technol. Agric. 10:43-60. 4. Bousbouras, G. E., and R. E. Kunkee. 1971. Effect of pH on 27. Pilone, G. J., R. E. Kunkee, and A. D. Webb. 1966. Chemical malolactic fermentation in wine. Am. J. Enol. Vitic. 22:121-126. characterization of wines fermented with various malo-lactic 5. Castino, M., L. Usseglio-Tomasset, and A. Gandini. 1975. Fac- bacteria. Appl. Microbiol. 14:608-615. tors which affect the spontaneous initiation of malolactic fer- 28. Ribereau-Gayon, J., E. Peynaud, P. Ribereau-Gayon, and P. mentation in wines. The possibility of transmission by inocula- Sudraud. 1975. Traite d'oenologie. Sciences et techniques du tion and its effect on organoleptic properties, p. 139-148. In J. vin. Tome 2. Caracteres des vins. Maturation du raisin. Levures G. Carr, C. V. Cutting, and G. C. Whiting (ed.), Lactic acid et bacteries. Dunod, Paris. bacteria in beverages and food. Academic Press, Inc. (London), 29. Rogosa, M. 1974. Genus I. Lact'obacillus Beijerinck 1901, 212. Ltd., London. Nom. Cons. Opin. 38, Jud. Comm. 1971, 104, p. 576-593. In R. 6. Charpentie, Y. 1954. Contribution a l'etude biochemique des E. Buchanan and N. E. Gibbons (ed.), Bergey's manual of facteurs de l'aciditite des vins. Ann. Technol. Agric. 3:89-167. determinative bacteriology, 8th ed. The Williams & Wilkins 7. Cofran, D. R., and J. Meyer. 1970. The effect of fumaric acid on Co., Baltimore. malolactic fermentation. Am. J. Enol. Vitic. 21:189-192. 30. Sharpe, M. E. 1981. The genus Lactobacillus, p. 1653-1679. In 8. Costello, P. J., P. R. Monk, and T. H. Lee. 1985. An evaluation M. P. Starr, H. Stolp, H. G. Truper, A. Balows, and H. G. of two commercial strains for induction of malolactic fermenta- Schleger (ed.), The prokaryotes. A handbook on habitats, tion under winery conditions. Food Technol. Aust. 37:21-23. isolation, and identification of bacteria, vol. 2. Springer-Verlag 9. Costello, P. J., G. J. Morrison, T. H. Lee, and G. H. Fleet. 1983. KG, Berlin. Numbers and species of lactic acid bacteria in wines during 31. Shimazu, Y., and M. Watanabe. 1979. Malolactic fermentation vinification. Food Technol. Aust. 35:14-18. in . J. Ferment. Technol. 57:512-518. 10. Dittrich, H. H., W. R. Sponholz, B. Wunsch, and M. Wipfler. 32. Smedt, de P., P. A. P. Liddle, B. Cresto, and A. Bossard. 1981. 1980. Zur Veranderung des Weines durch den bakteriellen The analysis of non-volatile constituents of wine by glass Saureabbau. Wein-Wiss. 35:421-429. capillary gas chromatography. J. Inst. Brew. 87:349-351. VOL. 51, 1986 LACTIC ACID BACTERIA IN WINE 545

33. Somers, T. C., and L. G. Wescombe. 1982. quality: the 77:223-227. critical role of SO2 during vinification and conservation. Aust. 36. Webb, R. B., and J. L. Ingraham. 1960. Induced malolactic Grapegrower Winemaker 19:68-74. fermentations. Am. J. Enol. Vitic. 11:59-63. 34. Temperli, A., and U. Kuensch. 1976. Die veranderungen der 37. Zeeman, W., J. P. Snyman, and C. J. Van Wyke. 1982. The gehalte an freien Aminosauren wahrend der Vinifikation. Qual. influence of yeast strain and malolactic fermentation on some Plant. Plant Foods Hum. Nutr. 26:141-148. volatile bouquet substances and on quality of , p. 35. Varma, R., R. S. Varma, and A. H. Nardi. 1973. Separation of 79-90. In A. D. Webb (ed.), Symposium Proceedings, Grape aldononitrile acetates of neutral sugars by gas-liquid chromatog- and Wine Centennial, 1980. University of California, Davis, raphy and its application to polysaccharides. J. Chromatogr. California.