Overexpression of RAD51 Enables PCR-Based Gene Targeting in Lager Yeast

Lager beer fermentations rely on specific polyploid hybrids between Saccharomyces cerevisiae and Saccharomyces eubayanus falling into the two groups of S. carlsbergensis/Saaz-type and S. pastorianus/Frohberg-type. These strains provide a terroir to lager beer as they have long traditional associations and local selection histories with specific breweries. Lager yeasts share, based on their common origin, several phenotypes. One of them is low transformability, hampering the gene function analyses required for proof-of-concept strain improvements. PCR-based gene targeting is a standard tool for manipulating S. cerevisiae and other ascomycetes. However, low transformability paired with the low efficiency of homologous recombination practically disable targeted gene function analyses in lager yeast strains. For genetic manipulations in lager yeasts, we employed a yeast transformation protocol based on lithium-acetate/PEG incubation combined with electroporation. We first introduced freely replicating CEN/ARS plasmids carrying ScRAD51 driven by a strong heterologous promoter into lager yeast. RAD51 overexpression in the Weihenstephan 34/70 lager yeast was necessary and sufficient in our hands for gene targeting using short-flanking homology regions of 50 bp added to a selection marker by PCR. We successfully targeted two independent loci, ScADE2/YOR128C and ScHSP104/YLL026W, and confirmed correct integration by diagnostic PCR. With these modifications, genetic alterations of lager yeasts can be achieved efficiently and the RAD51-containing episomal plasmid can be removed after successful strain construction.


Introduction
Lager beer is the most popular alcoholic beverage worldwide. It is distinguished from top-fermented ale beers by the use of dedicated bottom fermenting lager yeasts. Pure culture lager yeasts were introduced in the late 19th century and the first lager yeast strain became known as Saccharomyces carlsbergensis [1,2]. Lager yeasts are particularly suited for cold fermentations below 15 • C, providing a specific and clean taste [3]. Lager yeast strains are hybrids of two Saccharomyces species, namely S. cerevisiae and S. eubayanus [4][5][6][7][8]. The S. cerevisiae parent is related to ale yeasts while the S. eubayanus parent has recently been isolated from South America and Asia [9,10]. The low temperature fermentation capabilities of lager yeast hybrids may have originated from the parental S. eubayanus mitochondrial DNA that is invariably present in today's lager yeasts [11].
Genetic manipulations in lager yeast are hampered by the allopolyploid nature of these hybrids but also by their low transformation and homologous recombination (HR) efficiency [12]. This is demonstrated by the very few reports on targeted gene alterations in lager yeast [13][14][15]. These shortcomings of lager yeasts are in sharp contrast to the high efficiency and ease with which S. cerevisiae

Plasmid Design and Constructions
The plasmids used in this study are listed in Table 2. The primers required for plasmid constructions or for PCR-based gene targeting are listed in Table 3. Plasmid DNAs were amplified in E. coli and prepared using a Plasmid Midi DNA Purification Kit (Genaxxon, Ulm, Germany). PCR fragments were cloned into pGEM (Promega, Madison, USA). The ScRAD51/YER095W open reading frame was amplified from BY4741 genomic DNA using primers 260 and 261 and the Arthroascus (=Saccharomycopsis) schoenii TEF1 promoter was amplified from A. schoenii genomic DNA using primers 106 and 107, and both were cloned into pGEM. To place LacZ under the control of the AsTEF1-promoter, AsTEF1p was amplified form pGEM-AsTEF1p (E026) with primers 108/109, which added 45 bp of flanking homology region to E025-pRS417-AgTEFp-LacZ (here LacZ is under the control of the Ashbya gossypii TEF-promoter). The AgTEF-promoter was removed from E025 by restriction digestion with KpnI/XhoI, and the vector and AsTEF1p were cotransformed into yeast to combine them via in vivo recombination, yielding plasmid E065. All restriction enzymes were obtained from Thermo Fisher (Asse, Belgium). All primers were obtained from Sigma Aldrich (Overijse, Belgium).   E150/pRAD51 (GEN3) was constructed by in vivo recombination of E065 vector backbone with fragments carrying ScRAD51-ORF and HYG3 marker. To this end, E065 was linearized by digesting with EcoRV/SacI, which removed most of the LacZ-ORF, and then the band was gel purified. Flanking adaptor regions, to guide in vivo recombination, were added to ScRAD51-ORF and HYG3 marker fragments using primers 258/259 and 253/254, respectively. Then BY4741 was transformed with the two PCR products and the cut vector. In order to use a G418 resistance marker for PCR-based gene targeting in lager yeast, the GEN3 marker in E150 was replaced by LEU2 in a PCR-based gene targeting approach using BY4741 as a host. To this end, ScLEU2 was amplified from pRS415 with primer 344/345 by PCR and used for transformation, generating E160.
Novel synthetic marker genes YES1 and YES2 were also constructed using in vivo recombination in S. cerevisiae using plasmid pRS415 as the vector backbone. YES1 and YES2 contain the AsTEF1-promoter that controls kanamycin or hygromycin resistance gene-ORFs, respectively. For construction, pRS415 was linearized with SmaI, the AsTEF1p was amplified from E026 using primers 226/227, the kanR-ORF was amplified with primers 228/229 and the hygR-ORF with primers 230/231 to add suitable adaptors for homologous recombination. Vector, promoter and resistance gene ORFs were then cotransformed in S. cerevisiae and the resulting transformants were selected on G418 or hygromycin, respectively. All plasmids obtained by in vivo recombination in S. cerevisiae were verified and shuttled into E. coli for propagation. A. schoenii promoters were used as heterologous promoters in lager yeast and S. cerevisiae.
HYG3 is a synthetic selectable marker that consists of the 694 bp A. schoenii PGK1 promoter, the 1026 bp hygromycin resistance gene ORF and the Candida albicans URA3 terminator, which was synthesized by GenScript (GenScript, Piscataway, NJ, USA). Promoter sequences were derived from the A. schoenii CBS 7425 genome, which can be accessed via GenBank under accession JNFU00000000 CBS 7425 [43].

Yeast Transformation
For the transformation of S. cerevisiae, the lithium acetate/single strand DNA/polyethylene glycol 4000 protocol was used as described [44]. For lager yeast transformation, we adapted a protocol from Thompson et al. [39]. WS34/70 was grown overnight to stationary phase in 50 mL YPD. Then the culture was diluted to OD 600 = 0.3 in 100 mL fresh YPD. Cells were harvested at mid-log growth with OD 600 = 0.7-0.8 (after approximately 3 h of incubation). The culture was centrifuged and washed once with sterile water and resuspended in 25 mL of 0.1 M lithium acetate/10 mM dithiothreitol/10 mM TE solution (Tris HCl: EDTA = 10:1) and incubated for 1 h at room temperature. The cells were washed in 25 mL ice-cold distilled sterile water twice and once with 10 ml of ice-cold 1 M sorbitol, then resuspended in 100 µL ice-cold sorbitol. A 100 µL aliquot of the cell suspension was used for transformation. Transforming DNA (15 µL) was mixed with the cells, incubated for 5 min on ice and electroporated with 1.8 kV in 0.2 cm cuvettes. Then, cells were resuspended in 1 ml cold sorbitol and transferred to a tube with 300 µL YPD. This suspension was incubated for 3 h at 30 • C prior to plating on selective plates that did not contain sorbitol.

Verification of Lager Yeast Transformants
The Weihenstephan strain WS34/70 was used for gene targeting experiments. Target genes were derived from the S. cerevisiae parental subgenome. Sequences for verification primers (and also for the S1 and S2 primers) were deduced from the genome sequence that is available at DDBJ/EMBL/GenBank under the accession AZAA00000000. PCR-based gene targeting and diagnostic PCR were performed as described [45].

Design and Testing of Synthetic Marker and Reporter Genes
We were interested in developing novel synthetic marker and reporter genes for dual use in non-conventional yeasts, including lager yeasts and S. cerevisiae. To this end, we tested heterologous promoters derived from Arthroascus schoenii (Saccharomycopsis schoenii). Saccharomycopsis species are non-conventional yeasts used in diverse brewing settings, e.g., to generate nuruk [46,47]. We recently obtained several Saccharomycopsis genome sequences including that of A. schoenii [43]. Here we fused the AsTEF1 promoter with the LacZ reporter gene derived from Streptococcus thermophilus [48]. Additionally, we generated a completely new synthetic marker gene based on the AsPGK1 promoter controlling a hygromycin resistance gene ORF. Both genes were introduced into BY4741 on episomal CEN6/ARSH4 plasmids ( Figure 1). The functionality of both the synthetic marker and the reporter were demonstrated in vivo in S. cerevisiae: HYG3 was functional as we obtained hygromycin-resistant transformants that carried the plasmids, and the LacZ reporter was active as transformants were able to convert colorless X-Gal into a blue dye ( Figure 1B). This indicates that both A. schoenii promoters derived from AsTEF1 and AsPGK1 were functional in the heterologous host S. cerevisiae and we thus went on to employ these genes in lager yeast.

Transformation of Lager Yeast
Previous analyses already indicated that lager yeasts' transformation efficiency is drastically lower than that of S. cerevisiae. In S. cerevisiae lithium acetate/ss DNA/PEG protocols are commonly used [44]. We compared this standard protocol with modified versions, which were developed to deal with yeast strains that are non-responsive to standard lithium acetate or electroporation protocols [39,40]. Our modified protocol included parts of both protocols: a lithium acetate treatment (without DMSO) and an electroporation step (Li-Ac/Ep). To allow time for the expression of the antibiotic resistance genes used in this study, we incubated the cell suspensions after electroporation in an osmotically stabilized medium before plating on selective plates. To compare both protocols, we employed a freely replicating plasmid (E160) with HYG3 as the selectable marker.
In our hands, the lithium acetate/ss DNA/PEG protocol (with DMSO) yielded only a few transformants per µg of DNA, demonstrating the poor transformation efficiency of lager yeast with this method, even after lowering the temperature of the heat shock step at 40 • C, which we tested as lager yeasts are known to be temperature sensitive and do not grow at elevated temperatures above 34 • C. The Li-Ac/Ep protocol, however, resulted in much higher transformation efficiencies (1-2 orders of magnitude, Figure 2).

PCR-Based Gene Targeting is Enhanced by RAD51 Overexpression
Plasmid transformation is regularly far more efficient than integrative transformation. PCR-based gene targeting approaches are convenient because in one PCR reaction short-flanking homology regions can be added to a selection marker. We used standard S1 and S2 primers that added 50 bases of flanking homology region. However, in our hands, we failed to obtain stable lager yeast transformants with these PCR-based disruption cassettes, even when using the Li-Ac/Epo protocol ( Figure 3A). After prolonged incubation on selective plates (>2d) small colonies may appear. However, these colonies did not continue to grow upon restreaking on new selective plates or in selective liquid media and represent background growth ( Figure 3B). In contrast, the transformation of a Weihenstephan lager yeast strain that harbors a plasmid-encoded ScRAD51 expressed from a strong TEF promoter yielded a large number of transformants on primary selective plates using disruption cassettes for ScHSP104 ( Figure 3C). A randomly selected set of colonies continued to grow upon restreaking and also grew well in liquid YPD supplemented with hygromycin ( Figure 3D and data not shown).

192
Plasmid transformation is regularly far more efficient than integrative transformation. PCR-193 based gene targeting approaches are convenient because in one PCR reaction short-flanking 194 homology regions can be added to a selection marker. We used standard S1 and S2 primers that 195 added 50 bases of flanking homology region. However, in our hands, we failed to obtain stable lager 196 yeast transformants with these PCR-based disruption cassettes, even when using the Li-Ac/Epo 197 protocol ( Figure 3A). After prolonged incubation on selective plates (>2d) small colonies may appear.

198
However, these colonies did not continue to grow upon restreaking on new selective plates or in 199 selective liquid media and represent background growth ( Figure 3B). In contrast, the transformation Verification of these transformants that grew upon restreaking and growth in liquid culture was done by standard diagnostic PCR, amplifying the novel joints at the borders of marker integration [45]. This confirmed targeted gene disruption (Figure 4). Overall, between 60-75% of primary transformants could be cultivated further in a liquid culture, and all for all of those diagnostic PCRs indicated correct integration of the marker cassette at the target locus (n > 20). However, we did not obtain any transformants in the same strain without the overexpression of ScRAD51, demonstrating the usefulness of this approach. To verify that the marker integration and PCR-based gene targeting success was not solely locus dependent, we targeted a second gene, ScADE2, in the same manner. Transformation efficiencies were found to be similar for both loci and diagnostic PCR was successfully employed for the verification of the deletion of an ADE2 allele ( Figure 4C).
plates supplemented with hygromycin and G418 (100 µg/mL final concentration each). Restreaking transformants could be cultivated further in a liquid culture, and all for all of those diagnostic PCRs 217 indicated correct integration of the marker cassette at the target locus (n > 20). However, we did not 218 obtain any transformants in the same strain without the overexpression of ScRAD51, demonstrating 219 the usefulness of this approach. To verify that the marker integration and PCR-based gene targeting 220 success was not solely locus dependent, we targeted a second gene, ScADE2, in the same manner.

221
Transformation efficiencies were found to be similar for both loci and diagnostic PCR was  lager yeast relying on HR [13][14][15]. There are other issues beyond the mere technical difficulties.

239
Secondly, due to the hybrid nature, sporulation is severely crippled due to failure to proceed through 240 meiotic divisions, resulting in hybrid sterility, which is more pronounced in triploid Group I strains 241 than in tetraploid Group II strains [49,50]. Thirdly, even though lager yeasts are the workhorses of

Discussion
Lager yeasts are the workhorses of the beer industry. Detailed strain characterizations and molecular genetics-assisted yeast breeding, however, are hampered by the lack of proof-of-concept technologies. A major obstacle in lager yeast research is the surprisingly inefficient gene targeting in lager yeast. Thus, there are only a few reports of the successful molecular genetic manipulation of lager yeast relying on HR [13][14][15]. There are other issues beyond the mere technical difficulties. Primarily, these are complicated by the hybrid nature of lager yeast. Lager yeasts are allopolyploid hybrids between S. cerevisiae and S. eubayanus parents [9]. Group I/Saaz strains are triploid, while Group II/Frohberg strains, to which Weihenstephan 34/70 belongs, are tetraploid [2]. Additionally, aneuploidies could further change allele frequencies, complicating gene knockout experiments. Secondly, due to the hybrid nature, sporulation is severely crippled due to failure to proceed through meiotic divisions, resulting in hybrid sterility, which is more pronounced in triploid Group I strains than in tetraploid Group II strains [49,50]. Thirdly, even though lager yeasts are the workhorses of the beer industry, there is a reluctance to employ genetically modified yeast strains, a notion that has rather been strengthened over recent years.
In addition to their industrial importance, lager yeast hybrids present excellent model systems to study hybrid biology, including hybrid vigor, hybrid sterility, adaptive evolution and the analysis of hybrid protein complexes [51,52]. Major advances in hybrid yeast breeding resolved the F1-sterility problem represented in lager yeast [53].
Our study indicates that RAD51 overexpression in lager yeast opens the tool-box for all genetic manipulation previously only available in S. cerevisiae. Although transformation efficiencies in lager yeasts are far lower than in S. cerevisiae, rational strain design based on targeted gene replacements has become feasible. We have successfully employed this strategy already to other loci, suggesting that potential locus dependent variations in gene targeting are not inhibitory to successful gene targeting. Furthermore, by selecting specific homology regions, even closely related alleles of the S. cerevisiae and S. eubayanus parental genomes can be distinguished. The overexpression of RAD51 has been shown to improve HR in other systems, so it may also be advantageous in other non-conventional yeasts in which molecular genetic studies are hampered by the preferential ectopic integration of gene targeting cassettes.
Two additional advances to improve HR in resilient strains in recent years include the use of Ku70 mutant strains and the establishment of CRISPR/Cas9 methodologies [12]. Deletion of Ku70 inactivates the NHEJ pathway, thus favouring HR and targeted gene alterations [33]. However, in S. cerevisiae, the deletion of Ku70 has been shown to generate defects in telomere maintenance and cell cycle regulation [54,55]. Thus, this may not be favorable in lager yeast as breeding efforts to restore a wildtype Ku70 status after genetic engineering are also quite laborious.
Recent reports established CRISPR/Cas9 in lager yeast, making this a promising tool for strain engineering [12,56]. CRISPR/Cas9 will be particularly useful for simultaneous alterations of multiple alleles in a single step, even more so when the number of alleles may vary due to aneuploidies. In a recent report, simultaneous single and double deletions of SeATF1 and SeATF2 were performed in lager yeast, demonstrating the power of this tool [12]. It will be interesting to see if a combined approach, overexpressing RAD51 and the use of CRISPR/Cas9, will further improve gene replacement efficiencies.