Pure culture fermentations conducted over a 35-day period with the primary pitching rate showed more time was needed for most of the strains. B. bruxellensis (BSI-Drie) was the only initial strain with an attenuation level comparable to what would be observed from fermentations conducted with Saccharomyces cerevisiae.
Primary Pitching Rate
During initial fermentations with the primary pitching rate of 12×106 cells/ml, it was observed that final pH had a correlated effect on the apparent attenuation of each strain. The drop in pH to nearly 4.0 suggests acids other then acetic acid were produced by the Brettanomyces strains during anaerobic fermentations. Acetic acid is not produced in great quantities by anaerobic cells and the role of oxygen, necessary for the production of acetic acid has previously been demonstrated (Custers 1940; Ciani and Ferraro, 1997; Aguilar Uscanga, 2003; Garcia Alvarado et al., 2007). The observed levels of ethyl acetate in some of the fermentations would suggest that acetic acid was formed in small quantities before being further esterfied, although levels were usually below what is typically found in ales (Meilgaard, 1975a). Spaepen et al. (1978) found Brettanomyces yeast to produce high concentrations of the higher fatty acids C6-C12, which could be responsible for the observed drop in pH. It seems plausible that strains, which had a greater degree of fermentation during anaerobic conditions, excreted larger amounts of acids into the medium due to substantial TCA-cycle activity, coupled with the need for amino acid synthesis. Organic acids and similar compounds excreted into the medium could have multiple functions, while lowering the acidity, they also might promote growth and fermentation by helping to restore the redox balance, similar to Sheffers (1966) findings that pyruvic acid was capable of stimulating anaerobic fermentation.
The sugar reduction of eight Brettanomyces strains observed in this study during anaerobic fermentation with the primary pitching rate showed different strains exhibited multiple forms of sugar utilization. The monosaccharides glucose and fructose were completely consumed during semi-aerobic batch culture propagations along with the disaccharide sucrose. During anaerobic fermentation B. bruxellensis (WY5112) produced glucose above levels present in the initial wort, while also having slightly more fructose present. This is believed to be the first time this phenomenon has been observed quantitatively in Brettanomyces spp. during anaerobic fermentations. The presence of an external and internal a-glucosidase enzyme was found by Shantha Kumara et al. (1993) to be present in B. lambicus and was capable of hydrolyzing malto-dextrins with up to 9 degrees of polymerization. It was their opinion that the enzyme was responsible for the over attenuation seen during Lambic maturation, as they found a-glucosidase to produce glucose and the next lower malto-oligosaccharide. The observed increase in glucose would confirm the presence and activity of the a-glucosidase enzyme in this strain, and the extra presence of fructose shows that it is the major mechanism for sugar metabolism, as it appears to have even hydrolyzed sucrose into glucose and fructose. The strain B. lambicus (WLP653) also had measurable levels of glucose, fructose and sucrose, while present in solution slightly below the levels measured in the initial wort. Significant utilization of maltose was concurrently observed and it seems likely that the a-glucosidase enzyme was responsible for a majority of the sugar utilization in B. lambicus (WLP653) during anaerobic fermentations. While no glucose, fructose, or sucrose was measured in fermentations conducted with B. bruxellensis (BSI-Drie & CMY001) their degree of sugar utilization could be better explained by the presence of the a-glucosidase enzyme and quick utilization of glucose, fructose, and sucrose during fermentation. Future screening of Brettanomyces strains for a-glucosidase activity could yield further information regarding strains that are suitable for primary fermentation.
The presence of volatile compounds produced by Brettanomyces spp. in pure culture fermentations has only briefly been observed by previous studies (Van Oevelen, et al. 1976; Spaepen, et al. 1978; 1982). Of the five ethyl esters quantified during fermentations with the primary pitching rate, ethyl caproate and ethyl caprylate were the only esters produced by most of the strains in high concentrations. The observation of these two esters confirms the presence of caproic and caprylic acid among other higher fatty acids previously suspected for lowering pH during fermentations. Ethyl caprylate was observed at levels as high as 4.5 times the threshold level in beer, as found by Meilgaard (1975b) and would appear to be a dominant organoleptic compound along with ethyl caproate. These finding are in agreement with Spaepen et al. (1978) who observed in Brettanomyces the production of higher fatty acids while observing measurable increases in ethyl esters C6-C12 during the period which Brettanomyces yeasts were most active in spontaneous Lambic fermentations. While ethyl caprate was not measured in our study it’s possibly another important ester produced by Brettanomyces yeasts as Spaepen et al. (1978) found the fatty acid capric acid to be produced by Brettanomyces spp. along with significant amounts of ethyl caprate in Lambic and gueuze beers. Ethyl caproate produced by seven of the eight strains in this study has an organoleptic odor described as sweet, fruity, pineapple with an estery green nuance while ethyl caprylate has an organoleptic odor described as waxy, sweet, musty, pineapple and fruity. The ethyl esters of higher fatty acids were reported by Meilgaard (1975b) as additive, with the concentrations found in this study not having been observed in beer before. The levels of ethyl acetate observed in fermentations were highly variable and suggest that the ability of each strain to produce acetic acid during anaerobic fermentations is strain specific. This is in agreement with the finding of Freer (2002) who found similar variability in acetic acid. Produced by various strains of Brettanomyces.
The lack of measurable isoamyl acetate in this study is in agreement with previous findings where very low levels of isoamyl acetate was observed in beers fermented with Brettanomyces yeasts (Spaepen et al. 1982). Spaepen et al. (1982) further found that esterases present in Brettanomyces spp. were capable of hydrolyzing isoamyl acetate while synthesis stopped at 30% of the expected equilibrium constant.
Higher alcohols were not major by-products of the Brettanomyces strains observed in this study. Studies have shown the synthesis of higher alcohols to be directly linked to amino acid metabolism, which in turn is related to cell growth (Hazelwood et al. 2008). Therefore it is possible that low growth rates during anaerobic conditions could be a major factor into the low levels of higher alcohols present. Fewer than 30 mg/l total higher alcohols were formed by B. bruxellensis (BSI-Drie), which underwent the highest degree of fermentation out of any of the strains. Martens (1996) did not observe detectable amounts of n-propanol formed during pure culture fermentations with Brettanomyces spp. Those findings are not in agreement with this study as all eight strains were able to produce varying amounts of n-propanol. The overall level of higher alcohols was strikingly low, as Boulton and Quain (2006 p.118) reported 100-200 mg/l was normally found in beer. Quain and Duffield (1985) proposed that higher alcohols were metabolized as cellular redox control helping to restore the NAD+ to NADH ratio. If higher alcohol production does indeed mediate the oxidation of NADH, then screening strains for higher alcohol production could be beneficial in finding strains that could be used for primary fermentation.
The phenolic compound 4-vinylguaiacol was produced at low levels by each strain during pure culture fermentation while 4-vinylphenol was not measured at detectable levels. Recent studies have shown the vinyl reductase enzyme to rapidly convert 4-vinylphenol to 4-ethylphenol during the late exponential phase of growth and during stationary cell phase (Dias et al., 2003). The reason for the lack of measurable 4-vinylphenol in this study could very well be due to the rapid conversion of the compound into 4-ethylphenol, given fermentations lasted 35 days. The low levels of 4-vinylguaiacol still present could suggest the vinyl reductase enzyme may have a lower affinity for this phenolic compound. Godoy et al. (2008) found the coumarate decarboxylase enzyme to have 80% relative activity towards ferulic acid the precursor to 4-vinylguaiacol, and it’s very possible that the affinity of the vinyl reductase enzyme acts similarly with 4-vinylguaiacol. The likelihood that all eight strains had vinyl reductase activity seems suspicious, as this would not be consistent with the findings of Godoy et al. (2009) or Conterno et al. (1996) who detected vinylphenol reductase activity in a variable percentage of Brettanomyces bruxellensis isolates. Given the low levels of 4-vinylguaiacol observed, it’s possible that no 4-vinylphenol was produced during pure culture fermentations. A study conducted by Hernanz et al., (2001) showed ferulic acid to be roughly 75% more abundant then p-coumarate acid in malted barely. Given the all pale malt grain bill used for producing wort in this study, a low to negligible p-coumarate acid content is a possibility and can’t be ruled out. Further studies involving vinyl and ethyl phenol production during pure culture anaerobic fermentation could yield valuable information on the organoleptic qualities associated with Brettanomyces beers.
The low levels of acetaldehyde observed after pure culture anaerobic fermentation may be of relevance. This observation might suggest the metabolism of Brettanomyces spp. is affected during anaerobic conditions, and a rate-limiting step is connected with acetaldehyde production. Conterno et al. (2006) found through physiological testing that exogenous sources of biotin and thiamine were needed for growth. The effect of limited biotin and thiamine could hinder the production of acetaldehyde, as thiamine pyrophosphate is a co-enzyme needed for pyruvate decarboxylase to act on pyruvate, creating acetaldehyde an intermediate in the production of ethanol from pyruvate during fermentation. Thiamine dependent enzymes catalyze further reactions during anaerobic conditions and further studies should observe the effects biotin and thiamine have on Brettanomyces yeast during anaerobic fermentation. SO2 has also been shown to greatly affect acetaldehyde levels and is produced by various Saccharomyces spp. (Frivik and Ebeler, 2003). Research into the presence or absence of the sulfite reduction sequence in the genome of Brettanomyces spp. could be beneficial in understanding if SO2 affects the levels of acetaldehyde during fermentation. Also of interest were the low levels of diacetyl and 2,3-pentanedione. This could be due to their counteracting effect in restoring the NAD+/NADH balance observed by Scheffers (1961, 1966, 1979) and their ability to stimulate anaerobic fermentation in the yeast.
Effects of Pitching Rate
A change in the initial cell concentration at the start of fermentation had variable effects on the fermentation performance of four Brettanomyces strains over a 35-day period. Results were strain dependent with limited general patterns emerging.
The connection between initial cell concentration and apparent attenuation was found to be strain specific, given that both increases and decreases in attenuation were observed as the pitching rate increased. This suggests that cell concentration is linked to fermentation performance as no other variables were changed. Brendam (2009) found Brettanomyces bruxellensiscells in stationary phase were capable of fermenting glucose to produce ethanol and acetic acid, while further observing that for a given temperature the rate of production was linked to the biomass quantity in each fermentation. While previous research conducted with industrial lager strains has shown that fermentations were sped up when the pitching rate was increased (Verbelen et al., 2009), that does not appear to be the case with all strains of Brettanomyces.
Quantitative analysis found volatile compound production was variably affected by the change in the initial cell concentration. For both Wyeast strains (WY5526 & WY5151) an increase in esters and higher alcohols was generally observed which did not correlate with their levels of attenuation observed. It was also observed that increasing the cell concentration increased the production of volatile compounds in these two strains. The opposite was observed with B. bruxellensis (BSI-Drie) as cell count increased the amount of esters and higher alcohols produced decreased, this was coupled with the decrease in attenuation and appears that a lower pitching rate has a positive effect on the metabolism and fermentation of this strain. Volatile compound production in B. bruxellensis (CMY001) was unique in that ester and higher alcohol production increased from the lowest pitching rate to the middle pitching rate, but then decreased generally to the same levels as was produced with the lowest pitching rate. This observation is not coupled with the observed apparent attenuation, as this was the only strain that saw an increase in attenuation when the initial cell concentration was further increased.
The only significant trend observed in each of the strains was a decrease in 4-vinylguaiacol correlated with an increase in the pitching rate. This observation that higher cell concentrations produced less volatile compounds is contrary to previous studies that found volatile phenol content was not directly linked to growth or cell concentrations (Barbin et al., 2008; Godoy et al., 2009). Volatile phenolic production by Brettanomyces yeast has not been studied in wort before, making these findings intriguing.
The carbonyl compound acetaldehyde and vicinal diketones, diacetyl, and 2,3-pentanedione were not affected by the change in pitching rates which is contrary to the finding of Verbelen et al., (2009) who found small increases in acetaldehyde and substantial increases in diacetyl as pitching rates of Saccharomyces pastorianus were increased. The multitude of changes observed in fermentation performance over the range of three pitching rates makes it difficult to draw any conclusions as to what an ideal pitching rate would be. It would seem that other factors are more important in fermentation performance then initial cell concentration. More studies conducted with a larger range of pitching rates and more strains could further the understanding of pitching rate and its impact on fermentation performance with Brettanomyces yeast.
Initial Lactic Acid Concentration
Low pH environments are the conditions which Brettanomyces yeasts are most commonly associated with, yet no previous studies have been conducted to observe the effects of low pH on the combined fermentation performance by this yeast. For this study, wort was acidified with lactic acid before inoculating to an initial cell concentration of 12×106 cells/ml.
The effects on apparent attenuation over a 35-day period produced significant results in most of the strains. In fermentations conducted with initial lactic acid concentrations of 3,000 mg/l, four of the eight strains finished with 70% or greater attenuation. This was a large increase compared to the apparent attenuations observed in non-acidified wort fermentations. Two of the well attenuating strains finished just short of 90% apparent attenuation. The observed over attenuation appears to have occurred due to the low pH. Shantha Kumara et al. (1993) found the optimal pH for both the internal and external a-glucosidase enzyme to be 6.0 – 6.2, with stability greatly decreasing as pH decreased. The findings of this study show that a pH around 3.2 appears to enhance the activity of these enzymes in a variable amount of strains, or it might be proposed that another set of enzymes exist which lead to over attenuation and are enhanced at a low pH. Even at 1,000 mg/l fermentations showed greater attenuation then without the addition of lactic acid. In a study by Garcia Alvarado et al. (2007), the effects of acetic acid during anaerobic conditions showed that levels of acetic acid at 1,000 mg/l and 2,000 mg/l had no inhibitory effect on glucose consumption and ethanol production. During the same study acetic acid levels at 3,000 mg/l and above were found to inhibit glucose consumption and ethanol production. While earlier studies have shown high levels of acetic acid are toxic to yeast cells (Pinto et al., 1989), lactic acid does not appear to be toxic at similar levels and instead a stimulatory effect was observed in almost all the strains. The only exception to this was with a lactic acid concentration of 3,000 mg/l fermentations containing B. bruxellensis (WY5112) appeared to be inhibited possibly by the high levels of lactic acid. It seems interesting that B. claussenii (WLP645) a strain that would have originally been cultured from English stock ale, not traditionally having a low pH, did not show a significant change in apparent attenuation as the levels of lactic acid were increased. However, B. claussenii (WY5151) seemingly cultured from a similar origin did have an observed increase in attenuation as lactic acid concentration increased.
Quantitative analysis showed the increase in lactic acid concentration had noticeable effects on ester production. An overall increase in ethyl acetate was measured for most of the strains, with variable levels of production observed. Throughout the study B. bruxellensis (WY5112) was found to be the lowest ethyl acetate producing strain with little effect observed as initial concentrations of lactic acid increased. Other strains production increased with levels of ethyl acetate comparable to Saccharomyces cerevisiae (13 – 20 mg/l, Meilgaard, 1975a), still not being a major component of the organoleptic attributes. The B. bruxellensis strains (BSI-Drie & CMY001) were observed to produce high levels of ethyl acetate relative to the other strains with B. bruxellensis (BSI-Drie) producing the highest levels throughout the study, except for a slight decrease, which was observed when the initial lactic acid concentration increased from 1,000 to 3,000 mg/l. The observed increases in ethyl acetate as lactic acid concentrations were increased appears to coincide with the increase in attenuation as it appears acetic acid production while at a lower rate occurs side by side with ethanol production during anaerobic fermentation.
Ethyl lactate, previously shown to be an important flavor active component of Lambic beer, observed an exponential increase as initial lactic acid concentrations were increased. The levels observed in this study at 3000 mg/l were not as high as what was observed in Lambics as measured by Van Oevelen, et al. (1976). While a lactic acid content of 3,000 mg/l is representative of levels found in typical Lambic, it would appear the relative activity of the esterase is slow in synthesizing ethyl lactate. These findings are in agreement with Spaepen and Verachtert (1982) who found the esterase activity with ethyl lactate to be lower then with other esters. Ethyl lactate synthesis was found to be similar between strains when lower levels of lactic acid were present. As the lactic acid concentration increased ethyl lactate production became variable between the strains with low, medium and high synthesizing strains observed.
The increasing levels of lactic acid had an inhibitory effect on the synthesis of ethyl caproate and ethyl caprylate during fermentations. The enzymes responsible for the esterification of these two compounds have not been studied before and it seems possible that either product inhibition occurred or at low pH the relative activity of ester synthesis is lowered. Throughout the study no detectable amounts of ethyl caproate were measured during fermentations conducted with B. claussenii (WLP645) while low levels of ethyl caprylate were. These findings suggest two separate enzymes are responsible for the synthesis of these esters with one being inactive in this strain. While the strains observed in this study showed variability in the amounts of ethyl caproate and ethyl caprylate produced, the levels found in most of the fermentations were significantly higher then previously observed in beer and the enzymes responsible for their production should be an emphasis in further studies. The characterization of these enzymes would aid in the understanding of their capacity to influence the esters produced by Brettanomyces yeast.
Throughout the entire study no detectable amounts of isoamyl acetate or isobutyl acetate were observed as products of fermentation, regardless of lactic acid concentrations. It appears the Brettanomyces strains used in this study are incapable of synthesizing these esters. The ability of the Brettanomyces strains to produce ethyl butyrate during fermentation was found to be variable with very low production levels observed in half the strains. The levels produced during fermentation decreased slightly as a result of increasing lactic acid concentrations, and do not appear to be an important fraction of the esters produced. More research conducted on the genome of Brettanomyces yeast could yield valuable information on the genes that are responsible for ester synthesis.
Further decreases were observed in the production of 4-vinylguaiacol during anaerobic fermentations containing high concentrations of lactic acid. Initially an increase in 4-vinylguaiacol was observed when non-acidified fermentations were compared to fermentations containing 100 mg/l of lactic acid. At 100 mg/l the highest concentration of 4-vinylguaiacol was observed with production then decreasing as lactic acid concentration was increased. Edlin et al. (1998) found the coumarate decarboxylase enzyme present in B. anomalus to have optimal conversion at a pH of 6.0. This is contrary to the findings in this study were under the given conditions a pH of around 4.0 was optimal for 4-vinylguaiacol production. These findings are in agreement with Godoy et at. (2008) who found enzymatic activity was stable at pH 3.4 with only a 25% decrease in activity over a 24 hours period. The same study found an increased ethanol content of 10% and 12% (v/v) drastically decreased the enzymes activity by 50% after 10 minutes. Further research into volatile phenolic compounds produced by Brettanomyces yeast during wort fermentation would lend a greater knowledge of the coumarate decarboxylase and vinyl reductase enzymes responsible for such compounds. These and other studies on Brettanomyces yeast could aid brewers in choosing low producing ethyl phenol strains and controlling unwanted flavors and aromas during fermentation.
The findings in this study appear to have application in the brewing industry and further research into the metabolism of Brettanomyces yeast and the effects of low pH will advance the understanding of the optimal conditions for anaerobic fermentations with this yeast.