Valery N. Urakov

Engineering wine yeasts to combine efficient extracellular polysaccharide degradation with reduced urea accumulation may support improvements in winemaking. This study aimed to replace the arginase-encoding CAR1 locus in the genome of the triploid industrial wine yeast Saccharomyces cerevisiae I-328 with an expression cassette for the endopolygalacturonase gene PGU1, placed under the control of the PGK1 gene promoter and terminator. CRISPR/Cas9-mediated genome editing was performed using an acetamidase-selectable plasmid followed by phenotypic screening and molecular genetic verification. Whole genome sequencing confirmed the replacement of all three CAR1 alleles by the PGU1 expression cassettes. The modified strain I-328-Y21 showed enhanced degradation of polygalacturonic acid relative to the parental strain. Consistently, RT-qPCR analysis showed that the transcription level of the PGU1 gene in strain I-328-Y21 was about 300 times higher than in I-328. Disruption of arginase—the enzyme that generates urea from arginine—reduced urea content in the growth medium by about ninefold. As ethyl carbamate is formed from ethanol and yeast-excreted urea, this strain could help lower ethyl carbamate in wine. Fermentation performance and other technological characteristics remained unaffected by the genome editing. The new strain, in which the arginase gene is replaced by an endopolygalacturonase expression cassette, is a promising candidate for winemaking.

Engineering wine yeasts to combine efficient extracellular polysaccharide degradation with reduced accumulation of ethyl carbamate, a group 2A carcinogen, may support improvements in winemaking. This study aimed to replace the arginase-encoding CAR1 locus in the genome of the triploid industrial wine yeast Saccharomyces cerevisiae I-328 with an expression cassette for the endopolygalacturonase gene PGU1, placed under the control of the PGK1 gene promoter and terminator. CRISPR/Cas9-mediated genome editing was performed using an acetamidase-selectable plasmid followed by phenotypic screening and molecular genetic verification. Whole genome sequencing confirmed the replacement of all three CAR1 alleles by the PGU1 expression cassettes. The modified strain I328-Y21 showed enhanced degradation of polygalacturonic acid relative to the parental strain. Consistently, RT-qPCR analysis showed that the transcription level of the PGU1 gene in strain I328-Y21 was about 300 times higher than in I-328. Inactivation of arginase, which forms urea through the breakdown of arginine, resulted in about nine-fold decrease in the content of urea in the growth medium. Since ethyl carbamate is formed as a result of the interaction between ethanol and urea excreted by yeast cells, this strain could be used to reduce its content in wine. Winemaking experiments show that genome editing did not affect fermentation performance and other technological characteristics. The new strain with the replacement of the arginase gene with the endopolygalacturonase expression cassette is promising for use in winemaking.

The prion-forming regions (PFRs) of yeast prion proteins are usually located at either the N- or C-terminus of a protein. In the Sup35 prion, the main prion structure contains 71 N-terminal residues. Here, we investigated the importance of the terminal PFR location for prion properties. Two prionogenic sequences of 29 and 30 residues and two random sequences of 23 and 15 residues were added to the Sup35 N-terminus, making the original PFR internal. These proteins were overproduced in yeast with two variants of the Sup35 prion. Mapping of the prion-like structures of these proteins by partial proteinase K digestion showed that in most cases, the extensions acquired an amyloid fold, and, strikingly, the prion structure was no longer present or was substantially altered at its original location. The addition of two to five residues to the Sup35 N-terminus often resulted in prion instability and loss when the respective genes were used to replace chromosomal SUP35. The structures of yeast prions Mot3, Swi1, Lsb2, candidate prions Asm4, Nsp1, Cbk1, Cpp1, and prions based on scrambled Sup35 PFRs were mapped. The mapping showed that the N-terminal location of a QN-rich sequence predisposes to, but does not guarantee, the formation of a prion structure by it and that the prion structure located near a terminus does not always include the actual terminus, as in the cases of Sup35 and Rnq1.

In winemaking, malolactic fermentation (MLF), which converts L-malic acid to L-lactic acid, is often applied after the alcoholic fermentation stage to improve the sensory properties of the wine and its microbiological stability. MLF is usually performed by lactic acid bacteria, which, however, are sensitive to the conditions of alcoholic fermentation. Therefore, the development of wine yeast strains capable of both alcoholic fermentation and MLF is an important task. Using genome editing, we engineered a modified variant of the triploid wine yeast strain Saccharomyces cerevisiae I-328, in which the CAR1 arginase gene was replaced by the malate permease gene from Schizosaccharomyces pombe and the malolactic enzyme gene from Oenococcus oeni. Genome-wide transcriptional profiling confirmed the expression of the introduced genes and revealed a limited effect of the modification on global gene expression. Winemaking experiments show that genome editing did not affect fermentation activity and ethanol production, while use of the modified strain resulted in a tenfold reduction in malate content with simultaneous formation of lactate. The resulting wines had a softer and more harmonious taste compared to wine obtained using the parental strain. Inactivation of arginase, which forms urea and L-ornithine through the breakdown of arginine, also resulted in a twofold decrease in the content of urea and the carcinogenic ethyl carbamate in wine. Thus, the new strain with the replacement of the arginase gene with the MLF gene cassette is promising for use in winemaking.

In winemaking, malolactic fermentation (MLF), which converts L-malic acid to L-lactic acid, is often applied after the alcoholic fermentation stage to improve the sensory properties of the wine and its microbiological stability. MLF is usually implemented by lactic acid bacteria, which, however, are sensitive to the conditions of alcoholic fermentation. Therefore, the development of a wine yeast strains capable of both alcoholic fermentation and MLF is an important task. Using genome editing, we engineered a modified variant of the triploid wine yeast strain Saccharomyces cerevisiae I-328, in which the CAR1 arginase gene was replaced with the malate permease gene from Schizosaccharomyces pombe and the malolactic enzyme gene from Oenococcus oeni. Genome-wide transcriptional profiling confirmed the expression of the introduced genes and revealed a limited effect of the modification on global gene expression. Winemaking experiments show that genome editing did not affect the fermentation activity and the production of ethanol, while the use of the modified strain ensured a tenfold reduction in malate content with the simultaneous formation of lactate. The resulting wines had a softer and more harmonious taste compared to the wine obtained using the parental strain. Inactivation of arginase, which forms urea and L-ornithine through the breakdown of arginine, also led to a twofold decrease in the content of urea and the carcinogenic ethyl carbamate in wine. Thus, the new strain with the replacement of the arginase gene with the MLF gene cassette is promising for use in winemaking.

The prion-forming regions (PFRs) of yeast prion proteins are usually located at either the N- or C-terminus of a protein. In the Sup35 prion, the main prion structure contains 71 N-terminal residues. Here we investigated the importance of the terminal PFR location for prion properties. Two prionogenic sequences of 29 and 30 residues and two random sequences of 23 and 15 residues were added to the Sup35 N-terminus, making the original PFR internal. These proteins were overproduced in yeast with two variants of the Sup35 prion. Mapping of the prion-like structures of these proteins by partial proteinase K digestion showed that in most cases the extensions acquired an amyloid fold and, strikingly, the prion structure was no longer present or was substantially altered at its original location. The addition of two to five residues to the Sup35 N-terminus often resulted in prion instability and loss when the respective genes were used to replace chromosomal SUP35. The structures of yeast prions Mot3, Swi1, Lsb2, candidate prions Asm4, Nsp1, Cbk1, Ybr16 and prions based on scrambled Sup35 PFRs were mapped. The mapping showed that the N-terminal location of a QN-rich sequence predisposes to, but does not guarantee, the formation of a prion structure by it, and that the prion structure located near a terminus does not always include the actual terminus, as in the cases of Sup35 and Rnq1.

Amyloid formation is associated with many incurable diseases. For some of these, sporadic cases are much more common than familial ones. Some reports point to the role of somatic cell mosaicism in these cases via origination of amyloids in a limited number of cells, which can then spread through tissues. However, specific types of sporadic mutations responsible for such effects are unknown. In order to identify mutations capable of increasing the de novo appearance of amyloids, we searched for such mutants in the yeast prionogenic protein Sup35. We introduced to yeast cells an additional copy of the SUP35 gene with mutated amyloidogenic domain and observed that some nonsense mutations increased the incidence of prions by several orders of magnitude. This effect was related to exposure at the C-terminus of an internal amyloidogenic region of Sup35. We also discovered that SUP35 mRNA could undergo splicing, although inefficiently, causing appearance of a shortened Sup35 isoform lacking its functional domain, which was also highly prionogenic. Our data suggest that truncated forms of amyloidogenic proteins, resulting from nonsense mutations or alternative splicing in rare somatic cells, might initiate spontaneous localized formation of amyloids, which can then spread, resulting in sporadic amyloid disease.

The [PSI⁺] nonsense-suppressor determinant of Saccharomyces cerevisiae is based on the formation of heritable amyloids of the Sup35 (eRF3) translation termination factor. [PSI⁺] amyloids have variants differing in amyloid structure and in the strength of the suppressor phenotype. The appearance of [PSI⁺], its propagation and manifestation depend primarily on chaperones. Besides chaperones, the Upf1/2/3, Siw14 and Arg82 proteins restrict [PSI⁺] formation, while Sla2 can prevent [PSI⁺] toxicity. Here, we identify two more non-chaperone proteins involved in [PSI⁺] detoxification. We show that simultaneous lack of the Pub1 and Upf1 proteins is lethal to cells harboring [PSI⁺] variants with a strong, but not with a weak, suppressor phenotype. This lethality is caused by excessive depletion of the Sup45 (eRF1) termination factor due to its sequestration into Sup35 polymers. We also show that Pub1 acts to restrict excessive Sup35 prion polymerization, while Upf1 interferes with Sup45 binding to Sup35 polymers. These data allow consideration of the Pub1 and Upf1 proteins as a novel [PSI⁺] detoxification system.

The [PSI⁺] nonsense-suppressor determinant of Saccharomyces cerevisiae is related to formation of heritable amyloids of the Sup35 (eRF3) translation termination factor. [PSI⁺] amyloids have variants in amyloid structure and in the strength of suppressor phenotype. Appearance of [PSI⁺], its propagation and manifestation depend primarily on chaperones. Besides chaperones, the Upf1/2/3, Siw14 and Arg82 proteins restrict [PSI⁺] formation, while Sla2 can prevent the [PSI⁺] toxicity. Here, we identify two more non-chaperone proteins involved in [PSI⁺] detoxification. We show that simultaneous lack of the Pub1 and Upf1 proteins causes lethality of [PSI⁺] cells with a strong, but not with weak suppressor phenotype. This lethality results from excessive depletion of the Sup45 (eRF1) termination factor due to its sequestration into Sup35 polymers. We also show that Pub1 acts to restrict excessive Sup35 prion polymerization, while Upf1 interferes with Sup45 binding to Sup35 polymers. These data allow considering the Pub1 and Upf1 proteins as a novel [PSI⁺] detoxification system.