Omega Yeast is temporarily closed during the COVID-19 shut down. Stay safe, and we’ll see you on the other side!

Improved Functional Assays and Risk Assessment for STA+ Strains of Saccharomyces cerevisiae

040918 Omega  Tr14941 Web

Download a copy here.

ABSTRACT

Diastatic yeasts are a major contamination risk to packaged beer. Brewery quality control programs rely on selective microbial growth media and molecular detection of the STA1 gene, however there is a wide functional range of diastatic activity that remains unresolved by these current methods. Herein, we provide a comprehensive analysis of diastatic yeast selection media using our collection of STA1+ strains. For the preparation of microbial selection media, we highlight several factors to control including plate age, autoclave cycling conditions, cupric sulfate and dipotassium phosphate concentrations. We further present several functional assays that confirm the variation in diastatic activity across STA1+ strains and show a clear correlation with earlier onset and increasing strength in strains that contain an intact STA1 promoter. Furthermore, we have developed robust and simple plate-based assay to confirm diastatic activity of potential contaminants. Collectively, these functional assays provide improved risk assessment of re-fermentation in packaged product with the potential for determining contaminant thresholds for different STA1+ strains and weighing in on the decision to dump beer or issue a product recall.

INTRODUCTION

Diastatic strains of Saccharomyces cerevisiae are widely used industrial brewing strains, yet represent one of the major contamination threats to packaged beer. These strains are characterized by the ability to secrete a glucoamylase protein that degrades dextrins. Dextrins are mixtures of D-glucose polymers derived from grain starches in brewing. The dextrin composition of wort is influenced by the use of different malts and adjuncts and the mashing conditions; thus, dextrin levels can vary significantly between beer styles. The glucoamylase enzyme produced by diastatic strains breaks dextrins down into individual glucose molecules, which can then be consumed by diastatic and non-diastatic yeasts, leading to issues described below. Cross contamination of traditional ale or lager beers with a diastatic strain of S. cerevisiae poses risks of hyperattenuation and refermentation resulting in unwanted increased production of alcohol and CO2. In packaged product, diastatic contamination can result in gushing beer and exploding bottles/cans. Diastatic strains of S. cerevisiae are the traditional yeast used in Belgian beer styles such as Saison, Belgian Golden Strong, and Biere de Garde. Among 1,011 S. cerevisiae isolates sequenced from wild, clinical and industrial sources, diastatic yeast were only found within a subset of industrial beer yeast and small group of isolates in French Guiana (Krogerus et al. 2019). Thus, these diastatic strains represent a rare sub-population of S. cerevisiae that are ubiquitous in the brewing industry, but are far less common among the overall S. cerevisiae population. These strains have mistakenly been called S. diastaticus, which incorrectly indicates that they are a non-cerevisiae species. These strains are often referred to as S. cerevisiae var. diastaticus, which inappropriately designates them as a single variant, and not a collective grouping of diastatic yeasts with a wide range of genotypic and phenotypic traits. This nomenclature issue is analogous to flocculent brewing strains, which also have the shared phenotype of flocculation among genetically and phenotypically diverse strains. For simplicity and clarity, we refer to these diastatic strains as STA1+ strains of S. cerevisiae.

The genetic determinants for diastatic activity in S. cerevisiae are the STA1, STA2, and STA3 genes. These genes are unlinked and reported positions are on Ch. V, II, and XIV, respectively (Pretorius et al. 1988). Absence of STA1 correlates with loss of diastatic activity, and disruption of the STA1 promoter decreases diastatic activity, both suggesting that STA1 encodes the major, functional diastatic enzyme (Krogerus et al. 2019). However, another study has demonstrated that disruption of STA2 also ablates diastatic activity, highlighting that this phenotype is complex and incompletely understood (Pretorius et al. 1986). The STA1/2/3 genes arose from a gene fusion event between the FLO11 and SGA1 (Yamashita et al. 1987; Lo and Dranginis 1996). SGA1 is non-essential, intracellular ⍺-1,4 glucoamylase expressed under sporulation conditions and is typically inactive in industrial brewing strains (Pugh and Clancy 1990). FLO11 encodes a GPI-anchored cell surface glycoprotein that is induced in late stages of fermentation and plays a major role in flocculation. The STA genes contain the upstream regulatory sequence of FLO11 along with the FLO11 signal sequence for secretion fused to the ⍺-1,4 glucoamylase of SGA1 (Gagiano et al. 1999, Kim et al. 2004; Adam et al. 2004). The chimeric STA1 gene encodes a secreted ⍺-1,4 glucoamylase that is expressed during fermentation. This results in the degradation of dextrins in wort and hyperattenuated beers with low residual extract and elevated ethanol.

Within the STA1+ industrial brewing strains, there is considerable variation in attenuation levels. A recent study identified and characterized a 1,162 bp deletion upstream of the STA1 gene (Krogerus et al. 2019). This deletion removes key upstream activation sequences (UAS) and results in a decrease in STA1 gene expression and also diastatic activity. For simplicity, we will further refer to strains containing an intact STA1 gene as STA1+, and strains lacking the STA1 gene as sta1-. We will refer to strains containing the 1,162 bp deletion as uasΔ, and strains without this deletion as UAS+. This UAS deletion is easily detected by PCR and can be used as a screening tool for brewer’s when characterizing a house diastatic yeast or an unknown diastatic contaminant. Though PCR genotyping provides a valuable characterization and screening tool for brewers, there are limitations. First, there is considerable variability in diastatic activity among STA1+ strains that is not explained solely by the 1,162 bp deletion in the STA1 promoter. Second, PCR genotyping is susceptible to user error and extraneous variables that can lead to false-positive and false-negative results. Third, PCR results can only confirm the presence of DNA and does not determine whether the source of the DNA is a living cell. Lastly, there is the potential for a brewing contaminant to produce a dextrin-degrading enzyme unrelated to the STA genes. For example, a Brettanomyces bruxellensis strain was isolated from an overattenuated beer and found to produce both extracellular and intracellualar enzymes capable of degrading dextrins (Kumara et al. 1993).

This study set out to provide better tools for brewers to use in the detection and risk assessment of diastatic yeast contaminants. First, selection media for diastatic S. cerevisiae were evaluated by screening our STA1+ collection on commercially available LCSM and FPDM. Based on observed variability, we developed an Omega-optimized LCSM recipe. We next developed two independent functional assays to test for diastatic activity in the brewery environment, and confirmed a range of diastatic activities. The onset and strength of diastatic activity both increased in STA1+/UAS+ strains relative to STA1+/uasΔ strains.

Finally, we also developed a simple plate-based assay that clearly distinguished STA1+/UAS+ and STA1+/uasΔ strains. Each of these screening assays can be used to determine the diastatic activity of an unknown contaminant or as a complementary approach to verify PCR results. Lastly, we provide evidence that these assays correlate well with re-fermentation studies in packaged product and can be used to assess risk of a potential diastatic contaminant. Taken together, this work presents three new, validated methods to detect diastatic activity to aid the brewing industry with identifying and preventing contamination of their products.

EXPERIMENTAL METHODS

Yeast Strains
Industrial brewing strains used in this study are referenced by their commercial name supplied by Omega Yeast labs, LLC. A full summary of the strains can be found in Table 1.

Table 1

PCR Genotyping of STA1 and the 1,162 bp Deletion in the STA1 promoter
Multiplex PCR was performed using the previously published primers SD-5A and SD-6B for STA1 gene 868 bp fragment (Yamauchi et al. 1998) and STA1_UAS_Fw and STA1_UAS_Rv for the STA1 UAS 599 bp promoter fragment (Krogerus et al. 2019). Genomic DNA was prepared using an LiAc/SDS extraction (Looke et al. 2011). Briefly, roughly one colony was resuspended in 50 μL of 0.2M LiAC/1%SDS buffer solution and heated to 70°C for 5 minutes. The solution was precipitated with 200 μl of 100% EtOH at -20°C for 10 minutes. DNA was pelleted at 13,000 r.p.m. for 3 minutes, rinsed with 70% EtOH and air dried. The pellet was resuspended in 50 μl of DNAse/RNAse free water (Gibco). PCR reactions included 2 μl of DNA, 1 µM each primer, and 2x Gotaq mastermix (Promega). Touchdown PCR was performed (Korbie and Mattick 2008). Resulting PCR status was reported as UAS+ and STA1+ if the corresponding 599 bp and 868 bp bands were detected.

Preparation of Omega-optimized LCSM Selection Medium and Serial Dilution Plating for Comparing Commercial Formulations
Media formulations selected for comparison: Omega-optimized LCSM (0.4% yeast extract, 0.2% dried malt extract, 0.2% yeast peptone, 0.1% dipotassium ortho-phosphate, 0.05% ammonium sulfate, 0.06% cupric sulfate, 1% dextrose, 2% agar), commercial LCSM (Weber Scientific), and FPDM (Weber Scientific). YPD agar (1% yeast extract, 2% peptone, 2% glucose, 2% agar) was used as a nutrient medium to compare growth and recovery of the serial diluted cultures. Plates were autoclaved together using a solid load setting for 20 minutes (Hiyarama HVA-110) and poured 24 hours, 48 hours or 96 hours prior to plating serial dilutions of the indicated strains. Serial dilutions of 1:6 were performed from a starting concentration of 1 million cells/ml with 5 μl of diluted cells spotted. Plates were imaged after 3 days at 30°C.


Bromocresol Green Maltodextrin (BGM) Medium Preparation
Liquid BGM (bromocresol green maltodextrin medium) is used to assess dextrin degradation based on using maltodextrin as a carbon source coupled with a pH indicator dye that changes color as yeast cells metabolize the maltodextrin. Components include maltodextrin at 2% (w/v; Sigma-Aldrich, #419672), yeast nitrogen base with ammonium sulfate at 0.67% (w/v, Sunrise Science Products, #1501-100), K2HPO4 at 0.014% (w/v; J.T. Baker, #4012-05), and bromocresol green at 0.022% (w/v; BBL, #04-198). The medium was then brought to a pH value of 5.8 - 6.0 using 1M HCl. Medium was then vacuum filter sterilized with a 0.2 µm PES filter. The ability of strains to degrade dextrin was tested using BGM. Isolated colonies grown on YPD agar were directly inoculated into 500 µl of BGM medium in a sterile 1.5 mL microcentrifuge tube. Tubes were capped and rotated at 70 r.p.m. at room temperature for up to 20 days. Color change was assessed in two different ways. First, a visual estimation was performed to determine whether the color was blue, green, or yellow. This estimation was used for table displays as shown in Figures 2 and 3. Second, color change was quantified by collecting cells by centrifugation, then removing 100 µL of the media to a 96-well plate and measuring absorbance at 620 nm in a UV-Vis plate reader (Biotek Epoch2). Biological triplicates were performed using a separate colony for each replicate. To test the lower limit of detection in pure culture, 10-fold serial dilutions of three STA1+ strains with varying diastatic activity (OYL-026, OYL-027 and OYL-019) were prepared to inoculate BGM with 100 - 1,000,000 cells. Similarly, the lower limit of detection was evaluated in mixed culture by mixing these STA1+ strains in different ratios with a sta1- strain (OYL-006, British Ale I). In both cases, visual and quantitative assessment of color change was performed as described above.


Quantification of Dextrin Type III Degradation Using Iodine Staining
This assay is based on absorbance of iodine-stained starch. Individual colonies from YPD plates were inoculated into YPEG liquid medium (1% yeast extract, 2% peptone, 2% ethanol, 3% glycerol) + 2% Dextrin type III (MP Biomedicals) and incubated overnight at 23°C with agitation. Cultures were removed from the shaker and incubated without agitation at 23°C for 4 weeks. Dextrin degradation was measured at day 7, 14, 21 and 28. Each week, 200 µL of the culture was removed and briefly centrifuged to remove cells. 20 µL of the supernatant was transferred to a 1 cm cuvette with 1 mL of 0.04% Lugol’s iodine solution (10% potassium iodide, 5% iodine). Measurements for absorbance were collected at 680 nm using a UV-vis spectrophotometer (Thermo Scientific GENESYS 180). The UV-vis spectrophotometer was blanked with 0.04% Lugol’s iodine. A sta1- strain (OYL-088) was used as a negative control to normalize the STA1+ strain measurements and determine the percentage of Dextrin type III remaining. The results are averaged from 8 biological replicates. Reported for each strain is the week at which Dextrin type III levels was reduced to <50% of starting Dextrin type III (as determined by the negative control OYL-088).


Maltodextrin Cross-feeding Assay to Detect Diastatic Activity
Maltodextrin plates were prepared using a synthetic complete minimal media formulation (Sunrise Science; 0.67%YNB+nitrogen, 0.2% SC-Complete mix, 2% agar) and 2% maltodextrin (Spectrum Chemical). Strains grown on YPD agar were patched to the maltodextrin plates adjacent to W303. The plates were incubated at 30°C aerobically for 1 month of observation. Strains were scored based on the following criteria; - (no difference to compared to OYL-088), + (growth of STA1+ strain), ++ (growth of STA1+ strain and supported growth of W303), and +++ (growth of STA1+ strain with red pigment in W303). Limit of detection assays were performed with a spread plate of W303 and spotting of 20 µL test culture. The test cultures were from saturated propagations. The STA1+/UAS strain (OYL-026) was serially diluted 1:10 into the sta1- control strain (OYL-088). The corresponding concentration of cells is 2x106 cells/spot STA1+/UAS strain (100) down to 2-3 cells/spot STA1+/UAS strain and 2x106 cells sta1- strain (10-6).


Bottle Refermentation and Analysis
Strains from YPD agar were inoculated into 2 ml YPEG liquid media and incubated overnight at 23°C with agitation. Saturated cultures were inoculated into 355 mL bottles of Lagunitas IPA at 3 million cells/bottle. Crowns were removed from bottles, the culture was inoculated and the bottles were recrowned. Negative controls included a sta1- strain (OYL-088) and a mock inoculation with water. A set of 10-fold dilutions was also performed with OYL-026, OYL-056, OYL-027 and OYL-042. The starting 3 million cells/bottle was serially diluted 1:10 in sterile water and the resulting dilutions were inoculated into bottles. Bottles were incubated at 30°C for one month. The residual extract and alcohol by volume were measured using a DMA™ 4500 M and Alcolyzer ME (Anton Paar). Each inoculation was performed in biological triplicate.

RESULTS AND DISCUSSION

Evaluation and Optimization of Selection Media for STA1+ Strain Detection
The most widely used STA1+ strain selection media, Lin’s Cupric Sulfate Medium (LCSM), is permissive for strains that have elevated copper resistance (Lin Y. 1981). This strongly correlates to the STA1+ strains within the Beer II Group, as these are more closely related to industrial wine strains that have adapted resistance to copper-based fungicides (Galone et al. 2016). The recently introduced “Farber Pham Diastaticus Media” (FPDM) is suggested to be superior to LCSM for the detection of STA1+ strains and additionally allows for differentiation of STA1+ strains by starch degradation and a zone of clearance. The exact composition of this medium has not been reported.

We set out to optimize LCSM with the goal of having high recovery for STA1+ strains and minimal growth of sta1- strains. We first prepared LCSM according to the original published formulation (Lin Y. 1981), and then performed a variety of tests to determine the ideal concentrations of cupric sulfate and dipotassium ortho-phosphate by comparing growth of strains in our catalog of STA1+ and sta1- brewing strains (data not shown). Concentrations of 0.06% cupric sulfate and 0.01% dipotassium ortho-phosphate were ideal for selective growth of our STA1+ strains (Figure 1). To denote the difference compared to the original recipe, we named our recipe Omega-optimized LCSM.

To assess the recovery rates and background growth of sta1- strains, serial dilutions were performed. All strains were plated on YPD agar to represent non-selective conditions controlling for equal plating and comparable growth rates (Figure 1A). STA1+ and UAS+ genotyping of the OYL strains was determined by PCR (Table 3). We grouped the strains in STA1+/UAS+ (Group I), STA1+/uasΔ (Group II) and strains that have high levels of background when copper sulfate concentrations are reduced (Group III). Figure 1B shows the comparison of commercial LCSM, FPDM and the Omega-optimized LCSM. Commercial LCSM was the most stringent selection media with recovery of <50% of the Omega STA1+ strains (OYL-026, OYL-033, OYL-112, OYL-500, OYL-501 and OYL-042) and no background growth of sta1- strains. Commercial LCSM also resulted in a lower limit of detection with 0.5-20% recovery rates in OYL-033, OYL-112, OYL-500, OYL-501. FPDM exhibited the greatest effect of plate aging with ~36-fold less cell growth from 24 to 48 hours, with no permissible growth after 96 hrs. Manufacturers advise using plates within 24 hours of pouring. FPDM also showed the highest background growth with 100% of the Group III sta1- strains exhibiting growth at the 1 million/ml plated cell density. With our Omega-optimized LCSM, we were able to recover all Omega STA1+ strains. With the exception of OYL-040, we observed 100% recovery to the last serial dilution spot (~3 cells) if used within 48 hrs. The media minimizes background growth, but there is low-level, leaky growth of sta1- OYL-024 and OYL-030 Belgian strains. The Omega-optimized LCSM is effective after 96 hours, but shows slightly higher stringency.

With these results we conclude that while LCSM is a suitable selection media for STA1+ containing strains, attention must be paid to the cupric sulfate and dipotassium ortho-phosphate concentrations, autoclave cycling, age of the plates and the potential for background observed in a limited number of sta1- strains. Of the formulations tested, the Omega-optimized LCSM had the best performance. Of note, these STA1+ selection media showed no difference in growth between UAS+ and uasΔ strains (Figure 1). These results are not unexpected, as LCSM is based on the observation that copper resistance has been linked to presence of the STA1 gene, but the mechanism of copper resistance is unrelated to the STA1 gene or its product (Strope et al. 2015, Gallone et al. 2016). Thus, copper resistance doesn’t necessarily predict diastatic activity, and screening with this medium could result in identification of a strain that lacks diastatic activity (sta1-, false positive) or displays significantly reduced diastatic activity, as is the case for some of the STA1+/uasΔ strains. We therefore explored further methods to characterize differences between these STA1+ strains.

Figure 1
Table 3
Table 3A

Methods for Functionally Assessing Diastatic Risk of STA1+ Strains
Few methods exist for the direct phenotypic evaluation of the diastatic activity of STA1+ strains (Meier-Dörnberg et al. 2018, Krogerus et al. 2019). Diastatic activity varies widely among STA1+ strains and recently has been shown to strongly correlate with a naturally occurring 1,162 bp deletion within the STA1 promoter region (Krogerus et al. 2019). Brewers are faced with a decision to recall or destroy beer that has a confirmed STA1+ PCR result or a contaminant that shows copper sulfate resistance, though because of this strain-to-strain variability the real risk of this result leading to refermentation and exploding bottles is unknown. Within our own observations and repeated accounts of false positive and negative rates from our brewery customers, we set out to develop robust, easy-to-use methods that breweries could use to routinely and accurately test yeast strains for diastatic (dextrin-degrading) activity. Our goal was to develop a number of different assays, providing flexibility so that brewers could use an assay that works best within their current production scheme and accurately assess spoilage risk. Ultimately, we developed three independent methods as described below.

Method #1: Monitoring pH as a Readout of Maltodextrin-Supported Growth
Instead of using correlation with copper resistance as a proxy for presence of the STA1 gene, we sought to develop methods that directly assess diastatic activity. The first assay relies on pH change as an indicator of growth in medium with maltodextrin supplied as the only carbon source. Initial formulation trials suggested a liquid assay would perform better than a plate-based assay, as the plate-based assays had significant background growth even for strains lacking STA1 (data not shown). This medium contains bromocresol green, which starts as dark blue, but changes to green and eventually yellow as growing yeast cells acidify the medium. A schematic illustrating how to perform this assay for detection of diastatic S. cerevisiae is shown in Figure 2A. A panel of OYL commercial brewing strains was tested using this assay and analyzed both visually and spectrophotometrically (Figure 2B, 2C). The sta1- strains, OYL-006 and OYL-088, both failed to acidify the medium as expected (Figure 2B, 2C). In contrast, we observed three general behaviors: strong diastatic activity (medium was turned completely yellow within 3 days), moderate diastatic activity (medium was turned completely yellow within 8 days), and weak diastatic activity (color change was incomplete) (Figure 2B, 2C). While most of the STA1+/UAS+ strains had strong diastatic activity, OYL-033 exhibited moderate diastatic activity (Figure 2B, 2C). STA1+/uasΔ strains exhibited highly variable diastatic activity, and could be found with high, moderate, and weak activity (Figure 2C). Of note, the OYL-040 strain exhibited the weakest diastatic activity, requiring incubation of >12 days to observe color change of the medium. These results highlight that there are likely genetic determinants beyond the presence of STA1 and the uasΔ that determine diastatic activity.

Finally, we sought to determine how well this assay would perform in mixed culture to estimate our limit of detection for identifying a diastatic strain in a population of non-diastatic yeast (as might be found in a brewing situation). Again, results were assessed both visually and spectrophotometrically (Figure 3). Three strains were tested at different cell concentrations, both alone and in mixed culture with a sta1- strain (British Ale, OYL-006). The STA1+/UAS+ French Saison (OYL-026) strain was able to acidify the medium even with 100 cells, though when mixed with a non-diastatic strain it was undetectable if diluted less than 1:10 (Figure 3). Two STA1+/uasΔ strains were also tested (Belgian Ale OYL-019 and Belgian Saison OYL-027). Both strains performed moderately when alone, though with slower kinetics than OYL-026 (Figure 3). As with OYL-026, they performed less well in mixed culture. These results suggest that this assay is not ideal for testing mixed culture, unless there is a very high ratio of diastatic cells. Rather, this assay is ideal for testing pure culture isolates to confirm whether or not they possess diastatic activity.

Figure 2
Figure 3

Method #2: Quantification of Dextrin Type III Degradation Using Iodine Staining
The second assay characterizes the potential risk of the different diastatic S. cerevisiae strains based on direct quantification of dextrin degradation. The Omega STA1+ collection along with a sta1- control (OYL-088) were assayed for starch degradation over time. Colonies were cultured overnight with agitation in YPGE + 2% Dextrin III media to promote growth while relieving any glucose repression of STA1 transcription (Kartasheva et al. 1996). The cultures were then left still on the benchtop for up to one month to monitor Dextrin III degradation. Iodine was used to detect the remaining Dextrin type III in each sample using absorbance at 680 nm. A schematic illustrating this assay is shown in Figure 4A. Summarized in Figure 4B is the timing of starch breakdown in the various STA+ strains. Several highly diastatic strains such as OYL-026, OYL-056, OYL-205, OYL-500 and OYL-501 were able to degrade >50% of the Dextrin III by week 2. These strains also contain an intact STA1 promoter region (UAS+), and thus are predicted to have higher STA1 expression and diastatic activity. This wasn’t always the case with STA1+/UAS+ strains, as OYL-033 and OYL-039 typically took up to 3 weeks to degrade >50% of the Dextrin III. Nonetheless, when assessing STA1+/UAS+ strains relative to STA1+/uasΔ strains after 4 weeks, the STA1+/UAS+ strains exhibited significantly more Dextrin type III degrading activity (Figure 4D). Variation in diastatic activity was observed in STA1+/uasΔ strains, with OYL-042 showing starch breakdown by 3 weeks and other strains such as OYL-019, OYL-200 and OYL-040 rarely reaching >50% dextrin breakdown by week 4. In fact, when the assay was terminated the average amount of remaining starch for OYL-040 was 93% (Figure 4C). OYL-019, OYL-040 and OYL-200 measurements were carried out to 8 weeks to investigate this further. Eventually OYL-019 and OYL-200 reached 50-60% dextrin breakdown, where 80-100% of the original dextrin levels remained in OYL-040 (data not shown). The STA1 gene from OYL-040 was sequenced to identify potential genetic changes that could explain how a STA+/UAS+ strain exhibited such low diastatic activity. Multiple SNPs were observed in OYL-040 STA1 that were distinct from reported STA1 sequences of STA1+/UAS+ and STA1+/uasΔ (Krogerus et al. 2019), including two amino acid changes (S310L and V668P) (Figure 4E). Further experimentation is needed to determine whether these unique OYL-040 SNPs could lead to an altered Sta1 protein and lessened diastatic activity. This highlights the potential for other strain-specific genetic alterations that further distinguish STA1+ strains and their respective diastatic risk.

Figure 4

Method #3: Maltodextrin Cross-feeding Assay to Detect Diastatic Activity
A major risk that diastatic strains pose to packaged beer is the potential for a small amount of diastatic cells to secrete the glucoamylase enzyme, which then degrades starch over time and provides simple sugars for yeast to ferment. This premise was used to design a cross- feeding assay to score diastatic activity. STA1+ strains were patched adjacent to a sta1- indicator strain on plates containing maltodextrin as the sole carbon source and scored for STA1+ supported growth of the sta1- indicator strain. We predicted that if the STA1+ strain secreted high levels of Sta1 protein, then the enzymatic breakdown of maltodextrin would support the growth of a sta1- indicator strain (Figure 5A). The sta1- indicator strain used was the lab strain, W303, containing the ade2-1 allele, which results in the accumulation of the intermediate metabolite P-ribosylaminoimidazole carboxylate (red pigment) into vacuoles and the development of red colonies. This strain choice helped differentiate between robust growth (red), intermediate growth (red-white gradient) and slow growth (white) of the sta1- strain (Figure 5B, 5C). The Omega STA1+ strains were assayed with the criteria illustrated in Figure 5C and the results are summarized in Table 2. In just one week of growth, the STA1+/UAS+ strains were able to support intermediate to robust growth of the W303 strain. Only a subset of STA1+/uas- supported intermediate growth, others exhibited slow growth without the cross-feeding of the W303 strain, and the OYL-040 strain was indistinguishable from the sta1- control, OYL-088 (Figure 5B). Growth was scored again after 30 days and still several of the STA1+/uas- strains (OYL-019, OYL-025 and OYL-040) were unable to provide enough usable glucose to cross-feed and promote color change in the W303 sta1- indicator strain (Table 2). These results correlate very well with Methods 1 and 2 and again we observed OYL-040 was the least diastatic relative to the STA1+/uasΔ strains. Within a week of growth, this assay strongly differentiates between STA1+/UAS+ and STA1+/uasΔ strains. Therefore, this assay can be used similar to PCR as a follow up for any suspected contaminant to further differentiate between strains that contain intact promoters and the 1,162 bp deletion. In addition, the varying diastatic phenotypes within STA1+/UAS+ and STA1+/uasΔ genotypes can be used to help assess risk of hyperattenuation.

The limit of detection for this assay was assessed as a method for screening yeast slurries, propagations or enriched beer samples. The culture/slurry was spotted onto a lawn of W303 cells and color change was scored at the perimeter of the spot. Unfortunately, not all strains exhibited robust enough diastatic activity to resolve contaminants below 1 million cells/ml. An enrichment approach in liquid LCSM media was employed to test this further. Overnight saturated YPGE cultures were diluted into liquid LCSM for a 24-hour enrichment. OYL-026 (STA1+/UAS+) was serially diluted 1:10 into OYL-088 (sta1-) inoculated cultures. The lowest dilution volume 10-6 corresponds to 200/ml cells of OYL-026 and 200x106 cells/ml of OYL-088. 10 ul of the LCSM enrichments were spotted onto the YNB-SC maltodextrin plates with a lawn of the indicator strain W303. A color change in W303 indicating the presence of OYL-026 was observed down to the 10-6 dilution (Figure 5D). The turnaround of this assay was 7-9 days and may not be suitable for screening brewery propagations as this yeast would already be in production.

Figure 5
Table 2

Bottle Refermentations Assess Diastatic Activity in Packaged Product
The phenotypic assays described here are in strong agreement with previous observations that there are high risk and low risk STA1+ strains. To test this in packaged beer, STA1+ strains were inoculated into a Lagunitas IPA bottles to mimic contamination at packaging. Lagunitas IPA was chosen as a representative ale, with a 2.6°P residual extract, moderate hopping rate and 6.2% abv. One-month post-inoculation, OYL-026, OYL-112 and OYL-056 strains lowered the residual extract to 1.3-1.6°P (Figure 6A). This represents an addition of >2 volumes of CO2 in the packaged product, which at room temperature is above the pressure rating for a glass beer bottle and a significant hazard. The remaining STA1+/UAS+ strains, excluding OYL-501, were >0.6°P below the OYL-088 sta1- control. In agreement with the other assays presented in this paper, OYL-501 displays lower diastatic activity relative to the other STA1+/UAS+ strains. Collectively these results suggest OYL-501 may pose less risk to packaged product than other STA1+/UAS+ strains. Interestingly, the residual extracts in OYL-019, OYL-025, OYL-031 and OYL-042 were also lowered by 0.4-0.6P (Figure 6A) despite the 1,162 bp promoter deletion. At this high rate of inoculant (10,000 cells/ml) these STA1+/uasΔ strains still pose a re-fermentation risk. The mean Plato and ABV of STA1+/UAS+ and STA1+/uasΔ were compared to the sta1- control OYL-088 and mock water inoculation (Figure 6B and 6C). The STA1+/UAS+ strains were significantly lower in Plato than the negative controls, whereas the STA1+/uasΔ strains were not. Several strains resulted in increased ABVs (0.4-0.6%) that are above the TTB tolerance for reported ABV (0.3%). These bottle experiments confirm that overall STA1+/UAS+ strains are more hyperattenuative, however STA1+/uasΔ strains still pose a risk to packaged product.


To determine the minimum dose (CFU/ml) that results in re-fermentation of the packaged beer within a month, bottles were inoculated with 10-fold dilutions of the most potent STA1+/UAS+ (OYL-026 and OYL-056) and STA1+/uasΔ (OYL-027, OYL-042) strains (Figure 6D). After 1 month, the minimum dose of 10 cells/ml of OYL-026 and OYL-056 was sufficient to re-ferment 0.6°P and 0.2°P, respectively, whereas 10 cells/ml of OYL-027 and OYL-042 were not sufficient for re-fermentation. We predict that over extended periods of time, lower doses would result in re-fermentation. Additionally, we observed a dose response for increasing re-fermentation with increased number of cells, suggesting that thresholds for CFU/ml contamination levels each diastatic strain can be determined. It is important to note that the composition of the beer will have an impact on this as well. Factors such as the presence of additional viable sta1- yeast and the amount of dextrins/residual extract at packaging must be considered.

Figure 6

CONCLUSIONS

Herein, we provide a comprehensive analysis of selection media for diastatic strains of S. cerevisiae using our OYL collection of STA1+ strains. We highlight several factors to control for including plate age, autoclave cycling conditions, cupric sulfate and dipotassium phosphate concentrations. Secondly, we confirm the variation in diastatic activity across STA1+ strains and show a clear correlation with earlier onset and increasing strength in strains that contain an intact promoter (STA1+/UAS+). We also provide multiple robust and simple liquid- or plate-based assays for the functional risk assessment of diastatic S. cerevisiae contamination. Overall, we propose that quality control for diastatic S. cerevisiae contaminations in breweries should include detection by a combination of verifying STA1/UAS status with PCR and/or functional assays described here. This information can be used to assess the risk of re-fermentation in packaged product with the potential for determining contaminant thresholds for different STA1+ strains and weighing in on the decision to dump beer or issue a product recall.

Overall, the functional assays presented throughout this paper corelate medium-high diastatic activity with the presence of an intact promoter (STA1+/UAS+) as described previously (Krogerus et al 2019). In fact, the OYL collection of STA1+/UAS+ scored medium-high diastatic activity across the board (Table 3). Our customers commonly report hyperattenuation and continued fermentation in package product with OYL-026, OYL-056 and OYL-500, which were among the highest scoring strains in these assays. One notable exception within the STA1+/UAS+ strains, was OYL-501 which scored medium-high in our functional assays, but in the bottle refermentation experiment lacked diastatic activity. This result was surprising to our team and we repeated this several times to verify. However, consistent with reports from our customers, once OYL-501 finishes fermentation it is much less likely to result in re-fermentation in packaged product than its parent strain OYL-026. OYL-501 is a segregant from a genetic hybrid of OYL-026 (STA1+/UAS+) and the OYL-005 Irish ale yeast (sta1-). We predict that as a result, this strain is missing an unknown/undescribed genetic determinant important for maintaining diastatic activity in finished beer. Possibilities include surviving in finished beer, continuing to express STA1, or reactivating fermentation.

Within the group of STA1+/uasΔ strains, there was considerable variability from mild-medium diastatic activity in our functional assays. Without an intact promoter, the expression of STA1 may be more variable/stochastic and result in both low and medium-risk diastatic strains. It is also possible that other unknown genetic determinants are at play and the low level of STA1 expression is consistent within these strains, but the ability to survive and thrive allows for the range of activities described. An outlier within the STA1+/uasΔ strains was OYL-040, which had the lowest scoring diastatic activity in each of our assays. With this observation, we hypothesized that there was an additional inactivating mutation within the STA1-coding sequence in addition to the promoter deletion. We did not observe any deleterious mutations such as an insertion/deletion, frameshift or premature stop codon, however we did identify two single nucleotide polymorphisms (SNPs) that result in amino acid substitutions. Importantly, these OYL-040 STA1 SNPs are novel and have not been reported in the STA1 sequences that are publicly available (Krogerus et al. 2019). It is possible that these amino acid substitutions alter the activity of Sta1, however additional experimentation is necessary to confirm these SNPs result in the weakened diastatic activity observed in OYL-040.

In summary, these functional assays are cheap and accessible for brewery QC labs and provide two added benefits. First, they can easily be used to rule out false positive/negative results from traditional PCR/microbiological methods. Second, these functional assays can be used to assess the risk of potential contamination and factor in on the decision to destroy or recall product.

ACKNOWLEDGMENTS

Great Central Brewing Company for the use of the Anton Paar DMA 4500 and Alcolyzer.

REFERENCES

Adam, A. C., Latorre-García, L. and Polaina, J. (2004) 'Structural analysis of glucoamylase encoded by the STA1 gene of Saccharomyces cerevisiae (var. diastaticus)', Yeast, 21(5), pp. 379-88.

Gagiano, M., Van Dyk, D., Bauer, F. F., Lambrechts, M. G. and Pretorius, I. S. (1999a) 'Divergent regulation of the evolutionarily closely related promoters of the Saccharomyces cerevisiae STA2 and MUC1 genes', J Bacteriol, 181(20), pp. 6497-508.

Gallone, B., Steensels, J., Prahl, T., Soriaga, L., Saels, V., Herrera-Malaver, B., Merlevede, A., Roncoroni, M., Voordeckers, K., Miraglia, L., Teiling, C., Steffy, B., Taylor, M., Schwartz, A., Richardson, T., White, C., Baele, G., Maere, S. and Verstrepen, K. J. (2016) 'Domestication and Divergence of Saccharomyces cerevisiae Beer Yeasts', Cell, 166(6), pp. 1397-1410.e16.

Kartasheva, N. N., Kuchin, S. V. and Benevolensky, S. V. (1996) 'Genetic aspects of carbon catabolite repression of the STA2 glucoamylase gene in Saccharomyces cerevisiae', Yeast, 12(13), pp. 1297-300.

Kim, T. S., Kim, H. Y., Yoon, J. H. and Kang, H. S. (2004) 'Recruitment of the Swi/Snf complex by Ste12-Tec1 promotes Flo8-Mss11-mediated activation of STA1 expression', Mol Cell Biol, 24(21), pp. 9542-56.

Kim, T. S., Lee, S. B. and Kang, H. S. (2004) 'Glucose repression of STA1 expression is mediated by the Nrg1 and Sfl1 repressors and the Srb8-11 complex', Mol Cell Biol, 24(17), pp. 7695-706.

Korbie, D. J. and Mattick, J. S. (2008) 'Touchdown PCR for increased specificity and sensitivity in PCR amplification', Nat Protoc, 3(9), pp. 1452-6.

Krogerus, K., Magalhães, F., Kuivanen, J. and Gibson, B. (2019) 'A deletion in the STA1 promoter determines maltotriose and starch utilization in STA1+ Saccharomyces cerevisiae strains', Appl Microbiol Biotechnol, 103(18), pp. 7597-7615.

Kumara, H.M.C.S., De Cort, S., Verachtert, H. (1993) ‘Localization and characterization of alpha-glucosidase activity in Brettanomyces lambicus’, Applied and Environmental Microbiology, 59(8), pp. 2352-2358.

Lo, W. S. and Dranginis, A. M. (1996) 'FLO11, a yeast gene related to the STA genes, encodes a novel cell surface flocculin', J Bacteriol, 178(24), pp. 7144-51.

Meier-Dörnberg, T., Kory, O. I., Jacob, F., Michel, M. and Hutzler, M. (2018) 'Saccharomyces cerevisiae variety diastaticus friend or foe?-spoilage potential and brewing ability of different Saccharomyces cerevisiae variety diastaticus yeast isolates by genetic, phenotypic and physiological characterization', FEMS Yeast Res, 18(4).

Pretorius, I. S. and Marmur, J. (1988) 'Localization of yeast glucoamylase genes by PFGE and OFAGE', Curr Genet, 14(1), pp. 9-13.

Yamauchi, H., Yamamoto, H., Shibano, Y., Amaya, N., Saeki, T. 1998. ‘Rapid methods for detecting Saccharomyces diastaticus, a beer spoilage yeast, using the polymerase chain reaction.’ Journal of the American Society of Brewing Chemists. 56(2), pp. 58-63.