The budding yeast Saccharomyces cerevisiae is one of the most widely known and used microbes. Recently, its global phylogenomic diversity has been uncovered to an unprecedented scale (Peter et al. 2018; Pontes et al. 2020) . Two main clades of the species have adapted for ale beer brewing, namely the Ale 1 and the Mosaic (diastatic) beer (Gallone et al. 2018; Peter et al. 2018; Paraíso et al. 2023) . The Ale 1 yeasts have several described subgroups, while the Mosaic beer clade includes diastatic strains (Krogerus and Gibson 2020) . A third group of traditional farmhouse ale yeasts has recently been described (Preiss et al. 2018; 2024) . In bottom-fermented lager beer production, the Saaz or the Frohberg type of S. pastorianus , a hybrid of S. cerevisiae and S. eubayanus is used, and several other hybrids have also been described (Langdon et al. 2019; Gallone et al. 2019) .

Beer, through much of its history, still contained live yeasts following fermentation and was stored in wooden barrels and casks where a sufficient residual extract allowed for refermentation (Flavin et al. 2023; Štulíková et al. 2020), leading to carbonation. Since the end of the 19 ^th century, bottled beers became prevalent, followed by many technological improvements, more standardized and eventually most often pasteurized products (Raihofer et al. 2022) . Traditional abbey ales, along with some specialty products such as lambic beers, have retained the method of bottle refermentation (Derdelinckx et al. 1995) . Unfiltered and unpasteurized beers have also retained a share in the market. Currently, the emerging craft beer scene and customer demand for innovation led to bottle or cask refermentation to be more widely used once again (Štulíková et al. 2020; Marconi et al. 2016; Baiano 2021). In addition, live yeasts may also be introduced to bottled products inadvertently (Meier-Dörnberg et al. 2018; Krogerus and Gibson 2020) . Other traditional fermented beverages have experienced a craft revival, too, e.g. natural and unfiltered wines and pét-nats (Wei et al. 2023) .

Scientific and industry interest in locally selected beer strains increased recently (Bonatto 2021; Kerruish et al. 2024) . Isolates both from primary fermentation and from bottled beer have been sequenced by various research groups, and the home brewing community has also seen more and more interest in isolates sourced from bottles.

Here, we describe a simple set of experiments aimed at raising student engagement in brewing microbiology that revolves around the topic of beer bottle isolates. The experiments were successfully applied in the BSc-level Biotechnology program at the Faculty of Science and Technology of the University of Debrecen. Inspired by the BREWMOR initiative (Bridging Research and Education with Model Organisms – www.brewmor.org) a network of teaching and research faculty dedicated to propagating experiential learning for biology students, and by published student projects such as the yEvo (Moresi et al. 2024) our Department organized two consecutive single-semester courses for students designed to (1) teach scientific literature and database search and analysis of biomedical patents, and then to (2) teach the isolation and analysis of yeasts with the purpose of assessing Saccharomyces diversity in bottled beverages.

During the first (seminar-type) class, database searches, scientometrics, and the basics of intellectual property in science were covered. Topics included phylogenomic studies on beer yeasts, innovations in brewing, the largest producers of yeast starters, and patents (e.g. Farber, 2019) of the companies. Students prepared project works on the topics and discussed trends, marketing strategies in brewing, and legal perspectives – specifically, the Nagoya protocol (European Commission, 2021) . Details on topics are listed in a supplementary document (doi: 10.6084/m9.figshare.28407530). Based on experience during the class, we found that student engagement was helped by the interlinked nature of topics.

In the second semester of first-year students, a “General and Applied Microbiology” lab course was organized where students learnt how to isolate pure cultures of yeasts from the sediment of commercially obtained beer bottles of various origin, along with a single natural wine sample, all purchased locally (Table 1). Multiplex PCR-fingerprinting analysis (as described in Imre et al. 2019) was applied to the isolates, to compare Saccharomyces samples (Table 2). The method includes a control ITS primer pair, resulting in a control band for fungi (around 750 bp for Saccharomyces but markedly shorter in most other yeasts), S. cerevisiae -specific microsatellite primer pairs, and primers specific to S. cerevisiae delta sequences. Thus, this fingerprinting results in a single ITS band in the case of non- Saccharomyces samples, and multiple bands for S. cerevisiae ( Figure 1 a).

Two of the isolates only showed the single ITS control ( Figure 1 a) and were identified as Aureobasidium pullulans and Pichia kluyveri subsequently (Table 1). A further 17 isolates (Table 1) showed Saccharomyces band patterns ( Figure 1 a). Bands were scored and compared, and students were asked to interpret the results. Students successfully completing the course were included as group authors in this publication.

As shown in Figure 1 a, two groups, consisting of seven and four isolates, were identified with identical band patterns, indicating strain-level identity or very close relatedness. The first (Fingerprinting Group I) consisted of six Belgian ale isolates and a single UK isolate, all from bottle conditioned/refermented beers. Four of the seven isolates originated from a single brewery’s products ( Figure 1 a,c, Table 1). The yeasts collected from these may represent a commonly used strain or closely related group of strains, or alternatively, a common contaminant. The second group’s samples (Fingerprinting Group II) are from products with no statement on bottle refermentation. They were found in beer styles normally associated with their specific yeasts (IPA, pilsner IPA, kveik ale, Weißbier/Hefeweizen) ( Figure 1 a,c; Table 1). It is thus highly unlikely that the yeasts strain isolated from them was responsible for primary fermentation; it is plausibly a common contaminant strain. The remaining yeasts displayed unique patterns, with the most different being the Czech lager and natural wine isolates ( Figure 1 a).

Students were tasked to interpret the results and discuss them in their lab notes. For the discussion, the publications on beer yeast diversity and the clades used during the previous seminar proved useful. Students also had to discuss how a more definitive subtyping of the isolates could be achieved to identify the clades the yeasts belong to. The possible approach, phylogenomics, was shortly discussed in the light of the above-mentioned earlier studies.

Although phylogenomics was out of the scope of the class, we later chose four isolates for sequencing and a phylogenomics ( Figure 1 b). Based on coverage data, only the isolate from the unfiltered unpasteurized Czech lager, UDeb-LY0002, proved to be a hybrid. Its S. cerevisiae subgenome was grouped into the Lager: Frohberg clade (coverage graph doi: 10.6084/m9.figshare.27988580). To our knowledge, the survival of lager yeasts after bottling has not been shown using molecular genetic data previously. The phylogenomic dendrogram also showed that the natural wine isolate UDeb-WY0018 belonged to the Wine/European – Georgian subclade, while the UDeb-AY0002 isolate, representing the Fingerprinting Group II, was a member of the diastatic Mosaic Beer clade. The ale isolate UDeb-BY0015 from the Fingerprinting Group I, however, belonged to the Wine/European clade and might represent a bottle conditioning yeast ( Figure 1 b).

Our results highlight that care must be taken when bottle isolates are examined in context of traditional beer styles and yeast clades as they may not be the original strains responsible for primary fermentation. Breweries rarely publish the identities of their strains and bottle refermentation may or may not involve the use of specific commercially available bottle conditioning yeasts (such as Fermentis Safbrew T-58, S-33, or LalBrew CBC-1, a repurposed Champagne yeast) (Štulíková et al. 2020). Diastatic yeast contaminants may further complicate the picture (Paraíso et al. 2023) . Other species may also be found in bottled beer. These aspects all present multi-faceted and engaging opportunities to be discussed with students according to our experience, as they encounter a gradually unfolding complexity of industrial microbes. Costs associated with such a beer-yeast focused laboratory class are also manageable. Genome sequencing and bioinformatics may enhance the findings subsequently, as shown here.

Further studies may extend the scope of bottled products. For example, lambic beers are known to have a diverse microbiota which is only known to the level of species, not clades (Bongaerts et al. 2024) , providing opportunities for more studies especially in the case of bottled products that harbor live microbes easily accessible to consumers.

Yeast isolation. Beer and natural wine samples were commercially obtained in Hungary for the purpose of isolation. After decanting the product, sediments were suspended in sterile YPD (yeast extract, peptone, dextrose, VWR Chemicals, Solon, OH, USA) and spread to YPD agar plates and incubated for 2–3 days at 22°C. Colonies were further purified with sterile inoculation loops until single-cell derived colonies were obtained. Isolates were deposited into our culture collection at –70°C in YPD medium supplemented with 30% v/v glycerol.

Yeast identification. Colony DNA for PCR tests was isolated according to Lõoke et al. (Lõoke, Kristjuhan, and Kristjuhan 2011) from the colonies and stored in 1×TE. Briefly, a part of yeast colony was suspended in 100 µl lysis buffer (200 mM LiOAc 1% SDS solution), and incubated at 70°C for 15 min, then 300 μl 96% ethanol was added, the samples were mixed by brief vortexing. DNA was collected by centrifugation at 4°C with 15,000× g for 10 min. The precipitate was dissolved in 100 μl TE, then cell debris was spun down by brief centrifugation (15,000× g for 1 min), and 1 μl supernatant was used for a PCR reaction of 50 μl end volume. These colony DNA samples were used to differentiate Saccharomyces samples from other yeasts. We used the interdelta and microsatellite fingerprinting multiplex PCR combining interdelta, microsatellite ( YLR177w , YOR267c ), and as a control, ITS 1–4 primer pairs (Imre et al. 2019) with the GoTaq G2 polymerase (Promega, Madison, WI, USA) (Tables 2,3). We performed gel electrophoresis of the products (2% TBE agarose, 90 min, 100 V), Gene Ruler 1 kb size marker (Thermo Fischer Scientific, Waltham, MA, USA) was used for each run. We identified Saccharomyces samples based on band patterns, the non- Saccharomyces yeasts were subjected to PCR targeting the 26S ribosomal large subunit rRNA gene. The GoTaq G2 polymerase was used with primers NL1 and NL4 (Tables 2,3). The PCR products were subjected to capillary sequencing after PCR cleanup with the E.Z.N.A. cycle pure kit (Omega Bio-Tek, Norcross, GA, USA) by the sequencing core facility of the University of Debrecen. Sequenograms were manually inspected and the NCBI BLAST service was used for species identification, whereby the species with the closest hit in the NCBI GenBank was considered as a putative species identification, and then the sequence of the hit species’ type strain was once again aligned to the query sequence using BLAST. A similarity of >99% with the type was considered a definitive species identification. Sequences were deposited in GenBank (accession numbers: UDeb-Abp1: PQ381248; UDeb-Pk1: PQ381247).

Fingerprinting analysis. The Saccharomyces isolates identified as described above, were subjected to another round of multiplex PCR and gel electrophoresis for fingerprinting. In this case, genomic DNA was isolated using the glass bead method as described in Hanna and Xiao (2006). Concentration was set to 100 ng/µl in TE buffer. This better-quality genomic DNA is adviseable for a more reproducible and clearer fingerprinting band pattern. Multiplex fingerprinting was performed (Imre et al. 2019) with the GoTaq Flexi Hot Start polymerase (Table 2), gel electrophoresis conditions were the same as described above (Table 3). Band patterns were analyzed using GelJ (Heras et al. 2015) implementing the UPGMA clustering method with Dice coefficient. Original gel photos are uploaded to FigShare (doi: 10.6084/m9.figshare.28365200).

Whole genome sequencing and phylogenomics. We chose four isolates to be subjected to whole-genome sequencing using Illumina technology and a phylogenomic analysis involving members of previously described clades was performed after mapping to a concatenated reference of the species of the Saccharomyces genus. Genomes used are detailed in a supplementary table at FigShare (doi: 10.6084/m9.figshare.27988601). Mapping was followed by allele calling and generating a neighbor-joining tree as described in detail in the supplementary methods (doi: 10.6084/m9.figshare.27988613). The NEWICK file of the dendrogram was uploaded to FigShare (doi: 10.6084/m9.figshare.27988586). Coverage was visualized for the species of the genus from the mapping .bam files in bins of 10kb with a sliding window approach and this was used to check whether any of the isolates were hybrids. Raw sequencing reads were deposited in BioProject PRJNA1195563 of the NCBI Sequence Read Archive.

Table 1. Yeast isolates analyzed in this study.

Table 2. Primers used in this study

Table 3. PCR program used in this study.