Abstract

Our project aimed to increase vitamin B2 production in the baker’s yeast Saccharomyces cerevisiae to produce vitamin B2-enriched food. We successfully overexpressed ADE4, RIB1, and RIB7 (all combinations) in S. cerevisiae S288C. The engineered strains can grow normally in YPD media and dough. The test results showed that the more genes got overexpressed, the more vitamin B2 was produced. Our best strain, co-overexpressing all three genes, increased the vitamin B2 content by 193% in liquid YPD media, and by 91% in steamed buns.

 

1. Design

1.1 Yeast selection: Saccharomyces cerevisiae S288C

S. cerevisiae, commonly known as baker's yeast, stands as one of the most frequently utilized food microorganisms worldwide. It has been used in baking and winemaking for thousands of years, therefore being our perfect choice. Another advantage of using S. cerevisiae for vitamin B2 production is that it has a well-studied natural riboflavin synthesis pathway (Gudipati et al., 2014). The S. cerevisiae S288C is a prototrophic strain capable of dough-leavening. S288C is fully sequenced and widely used in labs, which enables us to anticipate reliable outcomes during molecular manipulations.

 

1.2 Plasmid selection

S. cerevisiae S288C is susceptible to G418, which makes the G418 resistance gene (KanMX) a good selection marker (Vickers et al., 2013). And since we aimed to overexpress genes in S. cerevisiae, a plasmid containing a strong promoter was needed. pCEV-G4-Km is a classic S. cerevisiae-Escherichia coli shuttle cloning and expression vector containing KanMX and the TEF1 promoter. TEF1 promoter from the yeast Yarrowia lipolytica has been proven one of the strongest promoters of S. cerevisiae (Partow et al., 2010).

 

Figure 1. Plasmid map of pCEV-G4-Km.

 

1.3 Overexpression of ADE4, RIB1, and RIB7

According to the discovered riboflavin synthesis pathway, GTP is the precursor of riboflavin (Kowalski et al., 2008; Gudipati et al., 2014). So, our first try was to overexpress GTP. With all the ADE enzymes, we chose to overexpress the first enzyme, phosphoribosylpyrophosphate amidotransferase, encoded by ADE4. This was a tentative choice since no research has been done to determine the rate-limiting step in the pathway. Overexpression of ADE4 should lead to overexpression of IMP, which is the precursor of both GTP and ATP. Excess of GTP should lead to a higher amount of vitamin B2, and excess of ATP should facilitate cell growth as the energy carrier.

 

RIB genes are directly responsible for riboflavin synthesis. With RIB1, 7, 2, 3, 4, and 5 being discovered, we had a broad range of choices. Research in Ashbya gossypii and Pichia pastoris has shown that overexpression of any RIB gene leads to the overexpression of riboflavin, and the more RIB genes get overexpressed, the more vitamin B2 gets produced (Ledesma‐Amaro et al., 2015; Marx et al., 2008). This year, due to the time limit, we planned to only overexpress the first two enzymes, RIB1 (GTP cyclohydrolase II) and RIB7 (diaminohydroxyphoshoribosylaminopyrimidine deaminase). We plan to include more genes in the future.

 

Figure 2. GTP and riboflavin synthesis pathways of S. cerevisiae (Kowalski et al., 2008; Gudipati et al., 2014). Green circles indicate genes overexpressed in this project.

 

Then, we designed the overexpression vectors for ADE4, RIB1, and RIB7. BamHI and NheI would be used to linearize the pCEV-G4-Km. Each of the genes would be inserted between the two restriction sites, forming pCEV-G4-ADE4-Km, pCEV-G4-RIB1-Km, and pCEV-G4-RIB7-Km

 

Figure 3. A. Plasmid map of pCEV-G4-ADE4-Km; B. Plasmid map of pCEV-G4-RIB1-Km; C. Plasmid map of pCEV-G4-RIB7-Km.

 

1.4 Multi-gene co-overexpression

To determine which gene influences the production of vitamin B2 most and to determine how multiple genes work together, we planned to co-overexpress ADE4, RIB1, and RIB7 in all possible combinations. The combinations included RIB1+ADE4, RIB7+ADE4, RIB1+RIB7, and RIB1+RIB7+ADE4 .

 

The 2A self-cleaving peptides cleave a longer peptide into two shorter peptides. For the polycistronic gene expression system to work in our yeast, we planned to put multiple genes downstream of one P_TEF1 promoter and insert the 2A peptide sequences between genes. ERBV-1 and PTV have been proven to have the highest cleavage efficiency in S. cerevisiae (Souza-Moreira et al., 2018). Therefore, RIB1-PTV-ADE4, RIB7-PTV-ADE4, RIB1-ERBV-1-RIB7, and RIB1-ERBV-1-RIB7-PTV-ADE4 would be assembled, with the stop codon of the prior genes deleted. Two different 2A peptides should be used in RIB1-ERBV-1-RIB7-PTV-ADE4 to avoid unexpected recombination.

 

Then, we designed the overexpression vectors for combinations of ADE4, RIB1, and RIB7. BamHI and NheI would be used to linearize the pCEV-G4-Km. Each of the gene combinations would be inserted between the two restriction sites, forming pCEV-G4-RIB1-PTV-ADE4-Km, pCEV-G4-RIB1-ERBV-1-RIB7-Km, and pCEV-G4-RIB7-PTV-ADE4-Km. Specifically, we planned to build the longest plasmid pCEV-G4-RIB1-ERBV-1-RIB7-PTV-ADE4-Km based on constructed CEV-G4-RIB7-PTV-ADE4-Km.

 

Figure 4. A. Map of RIB1-PTV-ADE4; B. Plasmid map of pCEV-G4-RIB1-PTV-ADE4-Km; C. Map of RIB1-PTV-RIB7; D. Plasmid map of pCEV-G4-RIB1-PTV-RIB7-Km; E.  Map of RIB7-ERBV-1-RIB4; F. Plasmid map of pCEV-G4-RIB7-ERBV-1-RIB4-Km; G. Map of RIB1-ERBV-1-RIB7-PTV-ADE4; H. Plasmid map of pCEV-G4-RIB1-ERBV-1-RIB7-PTV-ADE4-Km.

 

2. Build

2.1 Vector construction: single ADE4, RIB1, and RIB7

ADE4 (1533 bp), RIB1 (1035 bp), and RIB7 (732 bp) were acquired by PCR (Phusion High-Fidelity PCR Master Mix, Thermo Fisher, Waltham, MA, USA) using the genomic DNA of S. cerevisiae S288C as the template. BamHI and NheI (FastDigest, Thermo Fisher, Waltham, MA, USA) were used to linearize pCEV-G4-Km. Gibson Assembly (ClonExpress Ultra One Step Cloning Kit, Vazyme, Nanjing, Jiangsu, China) was performed to insert each gene between the restriction sites. The recombinant vectors pCEV-G4-ADE4-Km, pCEV-G4-RIB1-Km, and pCEV-G4-RIB7-Km were then transformed into S. cerevisiae S288C using the LiAc/SS Carrier DNA/PEG method (Gietz et al., 2006).

 

Figure 5. A. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pCEV-G4-ADE4-Km into E. coli DH5α; B. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pCEV-G4-RIB1-Km into E. coli DH5α; C. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pCEV-G4-RIB7-Km into E. coli DH5α.

 

2.2 Multi-gene co-overexpression

ADE4, RIB1, and RIB7 with different homologous overlaps were acquired by PCR (Phusion High-Fidelity PCR Master Mix, Thermo Fisher, Waltham, MA, USA) using the genomic DNA of S. cerevisiae S288C as the template. BamHI and NheI (FastDigest, Thermo Fisher, Waltham, MA, USA) were used to linearize pCEV-G4-Km. Gibson Assembly (ClonExpress Ultra One Step Cloning Kit, Vazyme, Nanjing, Jiangsu, China) was performed to insert each gene combination between the restriction sites.

 

Problem

We managed to construct pCEV-G4-RIB1-PTV-ADE4-Km, pCEV-G4-RIB7-PTV-ADE4-Km, and pCEV-G4-RIB1-ERBV-1-RIB7-Km. However, when we tried to use pCEV-G4-RIB7-PTV-ADE4-Km as the template to amplify RIB7-PTV-ADE4 by PCR, gel electrophoresis showed no bands. We had to adjust the PCR program to get the fragment we needed.

 

Problem solved

After several failed PCRs, we tried touchdown PCR from 68°C to 52°C. Although the gel electrophoresis results were still smeared, bands at around 2400 bp were identifiable. The fragments were successfully collected through proper gel extraction, enabling the construction of pCEV-G4-RIB1-ERBV-1-RIB7-PTV-ADE4-Km.

 

Figure 6. A. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pCEV-G4-RIB1-PTV-ADE4-Km (1-4), pCEV-G4-RIB7-PTV-ADE4-Km (5-8), and pCEV-G4-RIB1-ERBV-1-RIB7-Km (9-12) into E. coli DH5α; B. Gel electrophoresis of colony PCR products for verification of correct transformation of plasmid pCEV-G4-RIB1-ERBV-1-RIB7-PTV-ADE4-Km into E. coli DH5α.

 

The recombinant vectors were then transformed into S. cerevisiae S288C.

 

3. Test

3.1 Growth rate tests

First of all, we wanted to make sure that our engineered yeasts can still grow normally. Each of the strains was inoculated into flasks containing liquid YPD media and cultured at 30 ℃, 180 rpm. On hours 0, 8, 24, 32, 48, 56, and 72, the OD600s were measured using a NanoDrop One spectrophotometer (Thermo Fisher, Waltham, MA, USA).

 

Figure 7. Growth curves of engineered S. cerevisiae strains in liquid YPD media for 72h.

 

All the strains had similar growth rates, which meant that the overexpression of ADE4, RIB1, and RIB7 did not dramatically influence cell metabolites. In general, strains overexpressing more than one gene showed a lower OD600 when reached the stationary stage. This might be due to the matter and energy used to reproduce being moved to the overproduction of vitamin B2.

 

3.2 Vitamin B2 production tests #1: in liquid YPD media

Then, we started to test the vitamin B2 production by each strain. The WT S. cerevisiae S288C and seven engineered strains were inoculated into YPD media, and cultured at 30 ℃, 180 rpm. On hours 0, 4, 8, 24, 28, 32, 48, 52, 56, and 72, the riboflavin concentrations in the supernatants were measured using high-performance liquid chromatography (HPLC) according to GB 5009.85—2016, National Standard of the People’s Republic of China.

 

 

Figure 8. Vitamin B2 contents in liquid YPD media containing engineered S. cerevisiae strains for 72h.

 

No obvious differences between the strains can be seen in the first 10 hours. This might be due to the low yeast density in the media before 10 hours, which was confirmed by Figure 7. After 10 hours, results confirmed that the WT S. cerevisiae S288C is capable of producing a small amount of vitamin B2. The overexpression of any single ADE4, RIB1, or RIB7 gene increased vitamin B2 production by about 40%. Compared to RIB7 (33% increase), ADE4 (44% increase) and RIB1(46% increase) were the better genes. S. cerevisiae showed massive vitamin B2 overproduction when two genes were co-overexpressed. RIB1-RIB7, RIB1-ADE4, and RIB7-ADE4 increased the vitamin B2 contents by 96%, 131%, and 111% respectively, all higher than that of the single gene overexpression strains. Amazingly, S. cerevisiae S288C overexpressing all three genes produced the most vitamin B2, resulting in a nearly three-fold increase (193% higher).

 

3.3 Vitamin B2 production tests #2: in self-made steamed buns

To determine whether our engineered yeasts can help to make vitamin-B2-enriched food, we made steamed buns using our yeasts in the lab. For each strain, 50 g of bread flour was mixed with 0.5 g of yeast and 50 mL of water. The dough was placed in plastic bags and incubated at 30 °C for until it rose.

 

In our lab, it took the dough about 4 hours to rise. However, most recipes say it should rise in less than 1.5 hours. This difference might be due to our original strain choice. We used a laboratory strain, S288C, as our chassis, but the recipes use commercial yeast strains. It turned out that S288C was not as efficient as the commercial strains in bread-making, which indicated that we should change our chassis to the commercial strains in the future.

 

Then, we put the dough in a food steamer for 15 minutes. The riboflavin concentrations in the buns were measured using high-performance liquid chromatography (HPLC).

 

 

Figure 9. A. Self-made teamed bun using S. cerevisiae S288C carrying pCEV-G4-RIB7-Km; B. Vitamin B2 contents in steamed buns leavened by engineered S. cerevisiae strains. *: p < 0.01; **: p < 0.001.

 

All the strains helped to make the vitamin-B2-enriched buns. Overexpression of RIB1, RIB7, ADE4, and RIB1-RIB7 contributed to a small amount of vitamin B2 increase (all < 25% increase, p > 0.01). RIB1-ADE4 and RIB7-ADE4 overexpression strains significantly (p < 0.01) increased the vitamin B2 amounts in the steamed buns by 67% and 64% respectively. The highest overproduction of vitamin B2 came from the strain stacking RIB1, RIB7, and ADE4, resulting in a 91% increase (p < 0.0001).

 

4. Learn

We proved that overexpression of ADE4, RIB1, and RIB7 in S. cerevisiae could lead to vitamin B2 overproduction without influencing cell growth rate. We successfully found the strain that produced the most vitamin B2 in bun-making, which is the strain co-overexpressing all three genes. In the future, we plan to introduce more ADE and RIB genes into S. cerevisiae to reach higher vitamin B2 production. We also plan to switch our chassis from the lab strain S. cerevisiae S288C to some commercial strains.

 

References