Problems with Yeast Contamination during Wort Souring

MSc Project > Results & Dicussion: Problems with Yeast Contamination

This post is part of a series detailing the findings of my MSc research project which looked at the effects of different fermentation parameters on wort souring with Lactobacillus. If you haven’t already, take a look at the MSc project page for a full overview.

At the start of my project I actually had access to two different Lactobacillus species. I had both the L. brevis strain WLP672 from White Labs (San Diego, USA) as previously mentioned, in addition to the L. buchneri strain 5335 from Wyeast (Hood River, USA). To test the concept of my study and familiarise myself with all the procedures and equipment I decided to carry out a limited number of lactic fermentation trials with both strains. To initially save time I made the decision to propagate both bacteria by sampling directly from the packaged products provided by the manufacturers as opposed to isolating each strain separately. At the time the strains were packaged in a 35 ml vial from White Labs and 100 ml foil pouch from Wyeast. The following section describes my method for propagating each strain:

Method for direct propagation from commercial culture

A 1.040 SG wort was prepared as previously described (see Materials and Methods). The hot break material present after autoclaving was immediately removed by centrifuging the hot wort in sterile 50 ml centrifuge tubes for 3 min at 40,000 rpm. The centrifuged wort was re-combined to give 200 ml in a sterile 250 ml screw cap laboratory bottle. The wort was then cooled to 20°C and inoculated with 5 ml (2.5% v/v) of liquid suspension taken directly from the commercially produced cultures. The bottle lids were sealed and the cultures of both strains were incubated under static conditions for 24 h at 30°C before use.

Yeast contamination

Once both bacteria had been propagated I used them in two sets of identical fermentations with variables of temperature and wort SG, exactly as detailed earlier with the isolated L. brevis WLP672. The problem and the reason for not posting the results, was that in all of the samples checked with a microscope (a random selection of ten were picked between 48-72 h), yeast was discovered (Fig. 1). Considering the reputation of Lactobacillus as a beer spoilage microorganism this was somewhat ironic. It also reminded me about a few discussions I’d read concerning reports of 100% Lactobacillus fermentations going to completion and yeast being the far more likely explanation.

Fig. 1. Example of yeast contamination identified with a microscope in a 30°C fermentation sample after 48 h. The sample had been inoculated with a single L. brevis WLP672 culture. Image taken with smartphone of wet mounted sample.

Before the presence of yeast was confirmed, there was an inkling that something had gone awry based on a few other observations. Visually there was pronounced effervescence and foam formation on the surface of each sample vial when the screw-cap lids were removed (Fig. 2). This was so severe that each sample had to be degassed for several minutes before they could be analysed (Fig. 3).

Fig. 2. Excessive foaming and gas release from yeast contaminated fermentation samples after 24 h.

Fig. 3. Foam formation in yeast contaminated fermentation sample while measuring TA. This photo was taken after an unsuccessful attempt to degas the sample.

I later took great pains with some isolated L. brevis WLP672 samples to try and capture some of the gas release on camera to see the difference. Almost no effervescence was ever visible until samples were chilled and even then only a brief release of very small bubbles was seen (Fig. 4).

Fig. 4. Gas release from lactic fermentation inoculated with an isolated L. brevis WLP672 culture. Photo taken after 10 days.

Density and apparent attenuation

Apparent attenuation (AA) is used to express the change in wort density during fermentation and provides some insight into the extent of carbohydrate metabolism and the concentration remaining. Using the acquired SG results, the AA was calculated for all lactic fermentations as follows:

(3)   \begin{equation*}   AA = \frac{[(OG-FG)\times100]}{OG}$ \end{equation*}

AA = Apparent attenuation %
OG = Original wort specific gravity before fermentation
FG = Final wort specific gravity after fermentation

The AA of fermentations carried out with directly propagated cultures of L. brevis WLP672 and L. buchneri 5335 showed a marked difference to those using an isolated culture of L. brevis WLP672. Not only were the values much greater but there was a large variance between them. As an example, considering only 72 h fermentation results, the highest AA recorded for directly propagated L. brevis WLP672 was 46.4 ± 7.4 % at 30°C while the lowest was 12.7 ± 0.6 % at 20°C. In stark contrast, the maximum AA value with isolated L. brevis WLP672 was 6.1 ± 0.2 % for an inoculation rate of 2×107 CFU/ml (LF4) and the lowest was 2.0 ± 0.1 % at 20°C (LF1).

Fig. 5. Mean apparent attenuation (%) for all samples in fermentations with directly propagated L. brevis WLP672 (), directly propagated L. buchneri 5335 () and isolated L. brevis WLP672 – LF1-4 ().

Plotting the combined mean AA (Fig. 5.) illustrated the significant disparity in density changes between fermentations using an isolated L. brevis WLP672 culture and those with directly propagated cultures of L. brevis WLP672 and L. buchneri 5335. While the fermentations and mean AA values are not entirely comparable, as additional variables were introduced in LF3 & LF4, they serve to demonstrate how yeast contamination had a profound effect on wort gravity. It’s also worth noting the sample sizes represented by the mean AA results. For the isolated L. brevis WLP672 culture, SG measurements were gathered from 108 independent fermentations compared to 54 each for directly propagated L. brevis WLP672 and L. buchneri 5335 cultures.

The combined mean AA for all LF1-4 fermentations after 72 h was 4.2 ± 1.2 %, corresponding to a change in density of 1.0018 ± 0.0006 SG (0.47 ± 0.15°P). This is a drastic difference from directly propagated L. brevis WLP672 fermentations with an average AA of 29.8 ± 10.9 % at 72 h and a density drop of 1.0132 ± 0.0046 SG (3.36 ± 1.18°P). Even after just 24 h the AA (5.6 ± 6.3 %) of fermentations with directly propagated L. brevis WLP672 had exceeded the 72 h AA average (4.2 ± 1.2 %) for the isolated L. brevis WLP672 fermentations by 33%.

Monitoring bacteria growth and specific gravity

To further investigate the relationship between Lactobacillus growth and changes in wort SG, as well as provide additional confirmation that the isolated L. brevis WLP672 culture was not infected with yeast, a simple growth test was conducted as follows:


The cell growth of L. brevis WLP672 was approximated by measuring the optical density at 600 nm (OD600) of cultures using a Helios Epsilon UV-Vis Spectrophotometer (Thermo Fisher Scientific, Waltham, USA). Autoclaved and centrifuged 1.042 SG wort was prepared as desribed earlier (see Materials and Methods) and 45 ml was added to twelve sterile 50 ml centrifuge tubes. Each tube was inoculated with 1 ml (2.2% v/v) of the isolated L. brevis WLP672 stock culture, sealed and left to ferment under static conditions at 30°C. Destructive samples were taken at 0, 24, 48 and 72 h to measure OD600 and specific gravity. Triplicate samples were produced and analysed for each time point.

Fig. 6. Growth of L. brevis WLP672 at 30°C in 1.042 SG wort with an inoculation rate of 2.2% (v/v). Specific gravity () and OD600 (). Error bars represent SD.

Monitoring the growth of L. brevis WLP672 in 1.042 SG wort (Fig. 6) the largest increase in OD600 was observed in the first 24 h, consistent with typical bacterial growth phase behaviour.1 The drop in wort SG after 72 h was 1.0014 ± 0.00005 (0.37 ± 0.012°P), reasonably close to the 72 h mean of all the isolated L. brevis WLP672 fermentations (Fig. 5). The AA was also very similar, reaching 3.42 ± 0.11 % after 72 h.  The absence of yeast was also confirmed with a microscope for the isolated L. brevis WLP672 stock cultures and checked again before use in fermentation trials.

Isolated colonies

The discovery of yeast contamination when using directly propagated cultures meant that out of necessity only cultures propagated from selective media could be used. It was found that both L. brevis WLP672 and L. buchneri 5335 strains grew equally well on both MRS and Raka-Ray media, with and without 10 mg/L cycloheximide. Colonies were visible within 24 h and distinguishable morphological traits could be discerned after 48-72 h. As shown below (Fig. 7), L. brevis WLP672 appeared as raised white circular colonies exhibiting smooth edges and a shiny surface.

Fig. 7 Morphology of L. brevis WLP672 on MRS agar spread plate after 10 days incubation at 30°C.

After determining that isolation and propagation were viable, the decision was made to proceed with a single strain of LAB for future fermentation trials in order to simplify the process. The L. brevis WLP672 strain was chosen based tentatively on it’s superior souring performance even with yeast contamination and because slightly more literature data could be found regarding wort souring with the L. brevis species.


Yeast as a spoilage microorganism in lactic fermentations

Yeast contamination was clearly identified in samples from lactic fermentations carried out with directly propagated cultures of L. brevis WLP672 and L. buchneri 5335. Following this discovery, no attempt was made to identify the yeast or the source of contamination as it was deemed beyond the scope of the project. While the results of these trials couldn’t be used to draw any meaningful conclusions about Lactobacillus fermentation performance, they did highlight the dramatic consequences of infection by yeast and the threat that yeast contamination presents to wort souring with pure cultures of Lactobacillus.

Wort is both a nutrient rich substrate and with an initial pre-fermentation pH of around pH 5-5.5, a reasonably favourable environment for a variety of organisms.2 This is especially true at the elevated temperatures employed when wort souring (typically 30 – 40°C or 96-104°F).3 In general though there are very few microorganisms able to tolerate the low pH environment that results from organic acid production by lactic acid bacteria and a number of studies have detailed both their antimicrobial and antifungal properties.4, 5, 6, 7 The detrimental and inhibitory effects on yeast from LAB derived organic acids are also well documented.8 Many yeast are nevertheless still able to function at low pH, with a number of species demonstrating very high intrinsic tolerance for acidic conditions.9, 10 Some strains, including Saccharomyces cerevisiae brewing strains, also have the ability to quickly adapt to a previously terminal acidic environment.11

Product quality and consistency could potentially be impacted by yeast contamination during wort souring but one of the more severe ramifications, particularly if kettle souring, is that any alcohol produced via yeast fermentation will be largely evaporated during the boiling step. Another issue is that reports comparing single and mixed fermentations with a variety of LAB and yeast show that in all cases LAB growth is repressed, especially during the early stages of fermentation and in some cases by a substantial amount, depending on the initial inoculation rates.12 Therefore, as pH and TA production normally correlate closely with Lactobacillus growth in wort it’s reasonable to assume that yeast contamination could lead to lower acid production and final beer sourness than if uncontaminated.13

SG as an indicator of yeast contamination

It was notable from the results that the pure culture of L. brevis, when used to ferment wort, was only capable of reducing the density by a small fraction. The average across lactic fermentations LF1-4 being an AA of 4.2 ± 1.3 %, equivalent to a decrease of 1.0018 ± 0.0006 SG (0.47 ± 0.15°P). The largest single drop, an AA of 6.1 ± 0.2 % or 1.0025 ± 0.0001 SG (0.64 ± 0.02°P) was recorded in LF4 at 72 h using a pitch rate of 2×107 (Fig. 12). Similar results have also been reported elsewhere for a variety of Lactobacillus species.14

The exact change in wort density due to the activity of Lactobacillus will be influenced by a host of variables (wort properties, bacteria strain, fermentation conditions etc.). From the results reported in this study it does however seem safe to conclude that if a significant decrease in specific gravity is observed, it’s unlikely that LAB alone are responsible. The measured density, when used during lactic fermentation, can therefore provide a fast, simple and reliable indicator of an issue such as yeast contamination, requiring further investigation by the brewer.


  1. Rezvani, F., Ardestani, F., Najafpour, G. (2017) Growth kinetic models of five species of lactobacilli and lactose consumption in batch submerged culture, Braz. J. Microbiol., 48, 251-258.
  2. Bokulich, N. A., Bamforth, C. W. (2013) The microbiology of malting and brewing, Microbiol. Mol. Biol. Rev., 77, 157-172.
  3. Peyer, L. C., Zarnkow, M., Jacob, F., De Schutter, D. P., Arendt, E. K. (2017) Sour brewing: Impact of Lactobacillus amylovorus FST2.11 on technological and quality attributes of acid beers, J. Am. Soc. Brew. Chem., 75, 207-216.
  4. Vriesekoop, F., Krahl, M., Hucker, B., Menz, G. (2012) 125th Anniversary review: Bacteria in brewing: The good, the bad and the ugly, J. Inst. Brew., 118, 335-345.
  5. O’Mahony, A. (2000) Characterisation of antimicrobial producing lactic acid bacteria from malted barley, J. Inst. Brew., 106, 403-410.
  6. Peyer, L. C., Axel, C., Lynch, K. M., Zannini, E., Jacob, F., Arendt, E. K. (2016) Inhibition of Fusarium culmorum by carboxylic acids released from lactic acid bacteria in a barley malt substrate, Food Control, 69, 227-236.
  7. Lowe, D. P., Arendt, E. K. (2004) The use and effects of lactic acid bacteria in malting and brewing with their relationships to antifungal activity, mycotoxins and gushing: A review, J. Inst. Brew., 110, 163- 180.
  8. Thomas, K. C., Hynes, S. H., Ingledew, W. M. (2002) Influence of medium buffering capacity on inhibition of Saccharomyces cerevisiae growth by acetic and lactic acids, Appl. Environ. Microbiol., 68, 1616–1623.
  9. Osburn, K., Amaral, J., Metcalf, S. R., Nickens, D. M., Rogers, C. M., Sausen, C., Caputo, R., Miller, J., Li, H., Tennessen, J. M., Bochman, M. L. (2018) Primary souring: A novel bacteria-free method for sour beer production, Food Microbiol., 70, 76-84.
  10. Roels, S. P., Van Nieuwerburg, F., Van Landschoot, A., De Vuyst, L., Snauwaert, I., Vandamme, P. (2016) Microbial diversity and metabolite composition of Belgian red-brown acidic ales, Int. J. Food Microbiol., 221, 1-11.
  11. Rogers, C. M., Veatch, D., Covey, A., Staton, C., Bochman, M. L. (2016) Terminal acidic shock inhibits sour beer bottle conditioning by Saccharomyces cerevisiae, Food Microbiol., 57, 151-158.
  12. Hübbe, T. (2016) Effect of mixed cultures on microbiological development in berliner weisse, Institut für lebensmittel- und lebensmittelchemie, Technische Universität Berlin, Berlin, Germany.
  13. Peyer, L. C., Zannini, E., Jacob, F., Arendt, E. K. (2015) Growth study, metabolite development, and organoleptic profile of a malt based substrate fermented by lactic acid bacteria, J. Am. Soc. Brew. Chem., 73, 303-313.
  14. Tenhovirta, S. (2019) The effects of lactic acid bacteria species on properties of sour beer, Department of Food and Nutrition, University of Helsinki, Helsinki, Finland.

Leave a Comment

Your email address will not be published. Required fields are marked *