Introduction

Isobutanol is a promising biofuel candidate because of its higher energy density and lower hygroscopicity than ethanol1,2. Many bacterial species such as Escherichia coli, Corynebacterium glutamicum, and Bacillus subtilis have been engineered to produce isobutanol by introducing Ehrlich pathway enzymes involved in fusel alcohol formation from 2-keto acids3. In these bacteria, pyruvate is converted to α-acetolactate by α-acetolactate synthase (ALS) and then converted to 2-ketoisovalerate (2-KIV) by sequential catalytic reactions of ketol-acid reductoisomerase (KARI) and dihydroxy-acid dehydratase (DHAD). Finally, isobutanol is produced from 2-KIV via heterologous Ehrlich pathway enzymes, 2-ketoacid decarboxylase (KDC) and alcohol dehydrogenase (ADH)4,5.

Saccharomyces cerevisiae, which can naturally produce isobutanol via valine catabolism (Fig. 1), is a promising host for the production of various alcohols due to its high alcohol tolerance6,7. There have been several efforts to increase isobutanol production in S. cerevisiae, but the production levels were still lower than those produced in the engineered bacteria8,9,10,11,12,13. Unlike bacteria, the isobutanol production pathway enzymes in S. cerevisiae are partitioned between the cytosol and the mitochondria (Fig. 1). In S. cerevisiae, cytosolic pyruvate is imported into the mitochondria via mitochondrial pyruvate carrier (MPC) complex14,15 and then pyruvate is converted to 2-KIV by sequential catalytic reactions of Ilv2 (ALS), Ilv5 (KARI), and Ilv3 (DHAD) in the mitochondrial matrix. The mitochondrial 2-KIV is then exported to the cytosol, and finally converted to isobutanol through Ehrlich pathway, involving endogenous KDCs and ADHs4. Therefore, subcellular compartmentalization of valine biosynthetic enzymes is one of the limiting factors for efficient isobutanol production in S. cerevisiae. To overcome this limitation, the whole isobutanol biosynthetic pathway has been relocated either to the cytosol by expressing the mitochondrial enzymes (Ilv2, Ilv3, and Ilv5) in the cytosol8,13 or to the mitochondria by expressing the cytosolic enzymes (KDCs and ADHs) in the mitochondria16. Previously, we tried to improve mitochondrial pathway by increasing mitochondrial pyruvate pool via overexpression of the MPC complex17. However, the majority of pyruvate still existed in the cytoplasm and converted to ethanol. Therefore, transport of cytosolic pyruvate to the mitochondria is a limiting factor for efficient production of isobutanol via mitochondrial pathway.

Figure 1
figure 1

Schematic illustration of an artificial cytosolic isobutanol biosynthetic pathway. Pyruvate is converted to 2-ketoisovalerate (2-KIV) by sequential catalytic reactions of α-acetolactate synthase (ALS), ketol-acid reductoisomerase (KARI), and dihydroxy-acid dehydratase (DHAD), and then 2-KIV is converted to isobutanol by 2-ketoacid decarboxylase (KDCs) and alcohol dehydrogenase (ADH). Deleted gene is indicated by double slash (//) and dashed line indicate multiple enzymatic actions.

Full size image

In this study, we developed S. cerevisiae strain with enhanced isobutanol production using artificial cytosolic isobutanol biosynthetic pathway. In a previous study, a cytosolic isobutanol production pathway was constructed by overexpressing codon-optimized Ilv2, Ilv5, and Ilv3 in the cytoplasm while deleting the innate ILV2 gene8. In this study, we found that B. subtilis (Bs) ALS (AlsS) is more active than the Ilv2 targeted to the cytosol, but overexpression of Bs alsS, and the subsequence accumulation of the pathway intermediate α-acetolactate led to a growth inhibition in S. cerevisiae. The growth inhibition could be alleviated when the downstream enzymes, Ilv5 and Ilv3, were expressed in the cytosol. To circumvent the growth inhibitory effect of α-acetolactate, we expressed Bs alsS under the control of copper-inducible CUP1 promoter. Furthermore, we developed methods selecting efficient isobutanol-producing strains with multi-copy integration of the Bs alsS and cytosol-targeted ILV5 and ILV3 genes, based on their opposite effects on cell growth when overexpressed.

Results and Discussion

Construction of a cytosolic isobutanol biosynthetic pathway in S. cerevisiae

The innate isobutanol biosynthetic pathway in S. cerevisiae is subdivided into the cytosolic and mitochondrial pathways (Fig. 1). Therefore, to increase isobutanol production, sufficient amount of pyruvate should be introduced into the mitochondria, but pyruvate is mainly fermented to ethanol in the cytoplasm under most culture conditions18. To overcome this compartmentalization problem, we relocated llv2, Ilv5, and Ilv3 to the cytosol by deleting the N-terminal mitochondrial targeting sequence as reported previously8. ILV2ΔN54, ILV5ΔN48, and ILV3ΔN19 were expressed under the control of a strong constitutive promoter, PTDH3, on CEN-based plasmids, in bat1Δald6Δ strain (JHY43), where the competing pathway genes, BAT1 and ALD6 involved in valine biosynthesis from 2-KIV and oxidation of isobutyraldehyde, respectively, were deleted17. JHY43 overexpressing ILV2ΔN54, ILV5ΔN48, and ILV3ΔN19 (JHY4301) produced 71.8 mg/L isobutanol (Fig. 2b) after 72 h culture. We next compared the activity of Ilv2ΔN54 with other ALSs from B. subtilis and Lactococcus lactis. The alsS genes from B. subtilis (Bs alsS) and L. lactis (Ll alsS) were cloned into p413GPD plasmid vector providing PTDH3-driven gene expression, resulting in p413GPD-alsS(B) and p413GPD-alsS(L), respectively. Each plasmid was introduced into JHY43 together with the expression vectors for ILV5ΔN48 and ILV3ΔN19. However, we could not obtain any transformants harboring p413GDP-alsS(B). Therefore, Bs alsS was expressed using a weaker promoter PADH1 by cloning the gene into p413ADH vector. JHY43 cells harboring p413ADH-alsS(B) (JHY4302) showed similar isobutanol production level compared to that of JHY4301 (Fig. 2b). However, JHY43 harboring p413GPD-alsS(L) (JHY4303), showed an increase in isobutanol production up to 96.6 mg/L (Fig. 2b). JHY4303 showed a growth defect compared with JHY4301 and 4302, suggesting that higher ALS activity might inhibit cell growth (Fig. 2a). The accumulation of α-acetolactate is toxic to S. cerevisiae (Supplementary Fig. S2). Considering the more severe effect of Bs alsS on growth inhibition when expressed using the TDH3 promoter, Bs AlsS might have the highest enzyme activity followed by Ll AlsS and Ilv2ΔN54 in S. cerevisiae. We confirmed the different enzymatic activities of the three ALSs by introducing α-acetolactate decarboxylase (AlsD) from B. subtilis, which convert α-acetolactate to acetoin19. Acetoin can be further converted to 2,3-butanediol by endogenous enzyme Bdh120. When the ILV2ΔN54, Ll alsS, or Bs alsS gene was expressed under the control of TDH3 promoter together with alsD, JHY43 strain overexpressing Bs alsS produced the highest levels of acetoin and 2,3-butanediol followed by strains overexpressing Ll alsS and ILV2ΔN54 (Supplementary Fig. S1). Ethanol production levels showed a negative correlation with the acetoin/2,3-butanediol levels, reflecting the competition between pyruvate decarboxylase (Pdc) and Als. (Supplementary Fig. S1).

Figure 2
figure 2

The effects of overexpressing various ALS genes and introducing Kozak sequence to ILV5VN48 and ILV3ΔN19. (a) The indicated cells were grown in SC-His, Trp, Ura medium containing 2% glucose for 72 h, and cell growth and glucose uptake rate were monitored. (b) Isobutanol production levels were detected by HPLC. Each value indicates the average ± SD of triplicate experiments.

Full size image

Increase in isobutanol production by introducing a Kozak sequence into ketol-acid reductoisomerase and dihydroxy-acid dehydratase

The ALS is the enzyme competing with pyruvate decarboxylase (PDC) for pyruvate availability (Fig. 1). Therefore, strong ALS activity might be the key to redirect the pyruvate flux from ethanol production to isobutanol production. Bs alsS was successfully used to produce acetoin in S. cerevisiae when the gene was overexpressed together with B. subtilis alsD encoding α-acetolactate decarboxylase19. However, the toxic effect of Bs alsS overexpression together with the overexpression of ILV5ΔN48 and ILV3ΔN19, suggests that the following enzymes might not be active enough to convert the toxic intermediate generated by Bs AlsS. Since the transcription of ILV5ΔN48 and ILV3ΔN19 were already driven by the strong TDH3 promoter, we tried to increase the protein expression levels by improving their translational efficiency. To do this, Kozak sequence was introduced into each gene. In eukaryotes including S. cerevisiae, 5′ untranslated region (5′-UTR) plays an important role in translational initiation. Especially the Kozak sequence rich in adenine is localized from the start codon to the −6 position of the 5′-UTR in S. cerevisiae21,22. Therefore, we introduced five adenine sequences in front of the start codon (AAAAAATG) to increase the protein expression levels of the ILV5ΔN48 and ILV3ΔN19 genes (Supplementary Fig. S3). Expression vectors for ILV5ΔN48 and ILV3ΔN19 containing the Kozak sequence, (K)ILV5ΔN48 and (K)ILV3ΔN19, were introduced into JHY43 together with p413GPD-alsS(L). Compared with JHY4303 where ILV5ΔN48 and ILV3ΔN19 were expressed without Kozak sequence, JHY4306 expressing (K)ILV5ΔN48 and (K)ILV3ΔN19 produced about 1.4-fold higher level of isobutanol (133.9 mg/L), supporting the positive effect of increasing their protein expression levels in isobutanol production (Fig. 2b).

Enhancing isobutanol production by overexpressing alsS from B. subtilis using copper-inducible promoter, PCUP1

Although we could improve isobutanol production using Ll AlsS, the production level was still very low. Since Bs AlsS might have higher activity than Ll AlsS, but cannot be expressed using a strong constitutive promoter, we next considered using inducible promoter to overexpress Bs alsS at a specific time point. We chose the promoter of CUP1 gene encoding metallothionein, which is activated by Ace1 transcription factor in the presence of copper ions23,24,25. To express Bs alsS using PCUP1, PADH1 in p413ADH-alsS(B) plasmid was replaced with PCUP1, and the resulting plasmid p413CUP1-alsS(B) was introduced into JHY43 strain together with the expression plasmids for (K)ILV5ΔN48 and (K)ILV3ΔN19, generating JHY4307. JHY4306, where p413GPD-als(L) was introduced instead of p413CUP1-alsS(B), was used as a control. Cells were cultured for 7 h and then induced with 100 μM CuSO4. Copper induction in JHY4307 strain resulted in growth inhibition (Fig. 3a), but led to a 2.4-fold increase in isobutanol production from 84 mg/L to 205.2 mg/L (Fig. 3). In the control JHY4306 strain, the isobutanol production levels, which were not significantly affected by copper, were lower than those of copper-induced JHY4307 (Fig. 3). These results indicate that copper-inducible expression of Bs alsS using the PCUP1 is more effective in isobutanol production than constitutive expression of Ll alsS using the strong PTDH3 (Fig. 3).

Figure 3
figure 3

Expression of alsS from B. subtilis using copper-inducible promoter, PCUP1. JHY4306 and JHY4307 cells were grown in SC-His, Trp, Ura medium containing 2% glucose for 7 h and 100 μM of CuSO4 was added in the medium. Cells were further cultivated for 72 h, and the cell growth and glucose consumption (a) and Isobutanol production (b) were monitored. Each value indicates the average ± SD of duplicate experiments.

Full size image

Multi-copy integration of Bs alsS at delta-sequences using a screening method based on the toxic effects of copper-dependent induction of Bs alsS

For more efficient production of isobutanol through regulating the expression levels of Bs alsS, we tried multi-copy integration of the PCUP1-controlled Bs alsS at delta-sequences, the retrotransposon Ty1 long terminal repeats (LTR), existing hundreds copies in the S. cerevisiae genome26,27. The delta-integration cassette containing PCUP1-Bs alsS-TCYC1 and G418-resistant marker (KanMX) was introduced into JHY43, and the transformants were first selected with high concentration of G418 (2 mg/ml), which might provide a selection power for the transformants with multiple KanMX marker genes (Fig. 4a). Next, a second screening of the strains with multiple integration of Bs alsS was performed based on the growth inhibitory effect of Bs alsS overexpression. The integrated copy number of the PCUP1-Bs alsS is expected to have a negative correlation with the cell growth rate upon inducing the expression of Bs alsS by copper. Therefore, the 26 transformants selected on G418 plate were cultured for 7 h and then 100 μM CuSO4 was added to the culture medium to induce the expression of Bs alsS. From this screening, JHY43D1, JHY43D2, and JHY43D3, showing significantly lower growth rates than that of JHY43 were selected (Supplementary Fig. S4).

Figure 4
figure 4

Multi-copy integration of Bs alsS into chromosome at delta-sequences (a) Experimental design to select yeast cells with multi-copy integration of Bs alsS. The delta-integration cassette containing Bs alsS and KanMX was introduced into JHY43, and selected in YPD medium containing 2 mg/ml G418. The selected transformants were treated with 100 μM of CuSO4, and the cells with reduced growth rate were selected. (b) The selected Bs alsS-integrated cells harboring plasmids p414GPD-(K)ILV5ΔN48, p416GPD-(K)ILV3ΔN19, and p413GPD (JHY43D1-1, JHY43D2-1, and JHY43D3-1), JHY43D2 harboring empty vectors (JHY43D2-C), and JHY4307 control expressing Bs alsS on p413CUP1-alsS(B) were inoculated to OD600 of 0.5 and were cultured in SC-His, Trp, Ura medium containing of 20 g/L glucose for 12 h, and 20 μM of CuSO4 was treated. Cells were further cultivated for 72 h. (c) Glucose consumption. (d) Isobutanol production. Each value indicates the average ± SD of triplicate experiments.

Full size image

Isobutanol production in the selected strains was investigated after introducing the expression plasmids for (K)ILV5ΔN48 and (K)ILV3ΔN19 into the strains JHY43D1, JHY43D2, and JHY3D3, resulting in JHY43D1-1, JHY43D2-1, and JHY43D3-1. Strain JHY4307 that expresses Bs alsS on p413CUP1-alsS(B) was used as a control. Cells were cultured for 12 h and the expression of Bs alsS was induced by adding 20 μM of CuSO4. JHY43D2-1 and JHY43D3-1 showed lower growth and glucose uptake rates compared with JHY4307 and JHY43D1-1 (Fig. 4b,c), but produced more isobutanol (Fig. 4d), reflecting a negative correlation between the cell growth and Bs alsS expression levels. JHY43D2-1 strain produced the highest level of isobutanol (290.4 mg/L), which is 1.7-fold higher than that produced in JHY4307 (172.7 mg/L), and even higher than that produced in JHY4307 after induction with 100 μM of CuSO4 (Fig. 3B). The best producer JHY43D2 contains about 4 copies of Bs alsS gene when the integrated gene copy number was determined by qPCR compared with JHY43DC, where one copy of Bs alsS gene was integrated into the URA3 locus (Supplementary Fig. S6a). Taken together, the strains with multiple integration of Bs alsS at delta-sequences were successfully isolated based on the growth inhibitory effect Bs alsS.

Assembly of the cytosolic isobutanol biosynthetic pathway by multi-copy integration of ILV5ΔN48 and ILV3ΔN19 genes at rDNA sites

JHY43D2 strain harboring empty vectors (JHY43D2-C) showed a severe growth defect, which was relieved by overexpression of (K)ILV5ΔN48 and (K)ILV3ΔN19 (JHY43D2-1) (Fig. 4b). Overexpression of ILV5ΔN48 and ILV3ΔN19 might convert the toxic intermediate produced by Bs AlsS to less toxic isobutanol and other intermediates. We used this phenotype to generate stable isobutanol-production strain. To allow multiple integration of (K)ILV5ΔN48 and (K)ILV3ΔN19 genes into the JHY43D2 genome, rDNA (ribosomal DNA) repeats were used as the integration sites. The rDNA locus is composed of 150 tandem copies of a 9.1 kb repeat unit containing nontranscribed spacers (NTS1 and NTS2)28. The rDNA integration cassette containing PTDH3-(K)ILV5ΔN48-TCYC1-PTDH3-(K)ILV3ΔN19-TCYC1 and bleomycin selection marker (bleOR), flanked by NTS1 homology arms, was introduced into JHY43D2 (Fig. 5a). The transformants with multiple integration of (K)ILV5ΔN48 and (K)ILV3ΔN19 were selected on YPD plate containing 500 μg/mL of zeocin and 300 μM CuSO4 (Fig. 5a) (Supplementary Fig. S5). Copper-dependent induction of Bs alsS is toxic to the cells, thus cells with multiple integration of (K)ILV5ΔN48 and (K)ILV3ΔN19 could be selected on the medium containing high concentration of CuSO4 based on their faster growth. Five selected clones were grown in SC medium for 12 h and then treated with 20 μM CuSO4. All clones showed higher final cell densities (Fig. 5b) and higher glucose uptake levels (Fig. 5c) compared with the parental strain JHY43D2, reflecting the successful alleviation of the growth defect of JHY43D2. JHY43D24 strain showed the highest level of isobutanol production up to 227.2 mg/L. qPCR analysis revealed integration of 3 copies of (K)ILV5ΔN48 and (K)ILV3ΔN19 genes in the genome of JHY43D24 (Supplementary Fig. S6b). JHY43D24-53 strain having additional copy of (K)ILV5ΔN48 and (K)ILV3ΔN19 produced less isobutanol than JHY43D24, suggesting that our screening strategy was effective in isolating strains with optimal gene copy numbers (Supplementary Fig. S7). Furthermore, JHY43D24 strain showed consistent isobutanol production level after several subcultures, suggesting the stability of the integrated genes into the delta and NTS sites (Supplementary Fig. S8).

Figure 5
figure 5

Integration of (K)ILV5ΔN48 and (K)ILV3ΔN19 at NTS sites. (a) Experimental design for multi-copy integration of (K)ILV5ΔN48 and (K)ILV3ΔN19 at NTS sites. NTS-integration cassette containing (K)ILV5ΔN48, (K)ILV3ΔN19, and bleOR marker was introduced into JHY43D2 strain and selected on YPD medium containing 500 μg/ml zeocin and 300 μg/ml CuSO4. The selected transformants were further screened for improved cell growth upon induction of Bs alsS by adding 20 μM of CuSO4. (b) The JHY43D2 control and the selected strains JHY43D21, JHY43D22, JHY43D23, JHY43D24 and JHY43D25 were inoculated to OD600 of 0.5 and cultured in SC mix medium containing 20 g/L glucose for 12 h, and treated with 20 μM CuSO4. Cells were further cultivated for 72 h. (c) Glucose consumption. (d) Isobutanol production. Each value indicates the average ± SD of triplicate experiments.

Full size image

Enhancing isobutanol production in JHY43D24 by optimizing induction conditions of Bs alsS

The expression of Bs alsS is critical for isobutanol production, but high level expression exerts a negative effect though inhibiting cell growth. Therefore, isobutanol production could be optimized by regulating the expression time and expression level of Bs alsS. First, we determined optimal cell density to induce alsS expression. The final engineered strain JHY43D24 was cultured and Bs alsS expression was induced with 20 μM CuSO4 at various cell densities (Fig. 6). The earlier the induction, the stronger was the inhibition of growth and glucose uptake (Fig. 6a,b). Induction of alsS at OD600 of 5 led to the highest isobutanol production level (263.2 mg/L) (Fig. 6c). Induction of alsS at a later point of OD600 of 8 resulted in the best cell growth, but the isobutanol production level was the lowest among the tested induction conditions (Fig. 6).

Figure 6
figure 6

Optimization of copper induction time for isobutanol production. JHY43D24 cells were inoculated to OD600 of 0.5 and cultured in SC mix medium containing 20 g/L glucose. At the indicated cell densities, 20 μM of CuSO4 was added in the medium, and cell growth (a), glucose consumption (b), and isobutanol production (c) were monitored for 100 h. Each value indicates the average ± SD of triplicate experiments.

Full size image

Next, we tried to find the optimal concentration of CuSO4. JHY43D24 cells were grown to OD600 of 5 and alsS was induced with various concentrations of CuSO4 (Fig. 7). The cell growth and glucose consumption rates decreased as increasing the concentrations of CuSO4 (Fig. 7a,b). Although 5 μM CuSO4 was enough to induce alsS expression and subsequent isobutanol production, the highest titer of isobutanol (263.1 mg/L) was produced in cells induced with 20 μM CuSO4, which is the same concentration we used in Fig. 6 (Fig. 7c). In most culture conditions, ethanol was the prominent byproduct and acetoin accumulated over time (Supplementary Fig. S9).

Figure 7
figure 7

Optimization of CuSO4 concentration for isobutanol production. JHY43D24 cells were inoculated to OD600 of 0.5 and cultured in SC mix medium containing 20 g/L glucose. At OD600 of 5, various concentrations of CuSO4 were added in the medium, and cell growth (a), glucose consumption (b), and isobutanol production (c) were monitored for 100 h. Each value indicates the average ± SD of triplicate experiments.

Full size image

Conclusions

In this study, we engineered S. cerevisiae strain to produce isobutanol using an artificial cytosolic isobutanol biosynthetic pathway consisting of AlsS from B. subtilis and endogenous Ilv5 and Ilv3 targeted to the cytosol by deleting the mitochondrial targeting sequences. AlsS is a key enzyme converting the cytosolic pyruvate to α-acetolactate, competing with the ethanol production pathway. However, we found that excessive production of α-acetolactate by overexpressing of Bs alsS is toxic to the cells, which can be alleviated by subsequent conversion of α-acetolactate to other chemicals by Ilv5ΔN48 and Ilv3ΔN19. Therefore, to reduce growth inhibition mediated by accumulation of the toxic intermediate α-acetolactate, Bs alsS was expressed under the control of a copper-inducible CUP1 promoter, and the translational efficiency of ILV5ΔN48 and ILV3ΔN19 was increased by adding Kozak sequence. Furthermore, the toxic effect of α-acetolactate was used to develop novel screening methods for multi-copy integration of the pathways genes. Multi-copy integration of Bs alsS into delta-sequences was screened based on the growth inhibition upon the induction of Bs alsS, whereas multi-copy integration of (K)ILV5ΔN48 and (K)ILV3ΔN19 with Kozak sequence into the rDNA sites of the alsS-integrated strain was screened based on the growth recovery effect. The strain JHY43D24 generated by this screening methods contains 4 copies of Bs alsS and 3 copies of (K)ILV5ΔN48 and (K)ILV3ΔN19. After optimizing the alsS induction conditions, the final engineered strain produced 263.2 mg/L isobutanol, showing about 3.3-fold increase compared to the episomal expression of the cytosol-targeted endogenous genes ILV2, ILV5, and ILV3 from strong constitutive promoters. However, the isobutanol production level is still very low compared to its production in other bacteria. Therefore, to further improve isobutanol production, it might be critical to improve the activities of Ilv5ΔN48 and Ilv3ΔN19, which require NADPH and iron-sulfur cluster, respectively.

Methods

Strains and culture conditions

All strains used in this study are described in Table 1. E. coli strain DH5α was used for construction of plasmids and cultured at 37 °C in Luria-Bertani (LB) medium containing 50 μg/mL of ampicillin. JHY43 strain, BAT1 and ALD6 deletion mutant derived from S. cerevisiae CEN. PK2-1C, was used as parental strain of all the engineered strains in this study29. Yeast cells were cultured in YPD medium or in synthetic complete (SC) medium (20 g/L glucose, 6.7 g/L yeast nitrogen base without amino acids, and 1.67 g/L amino acids mixture lacking histidine, tryptophan, leucine, and uracil) supplemented with auxotrophic nutrients as required.

Table 1 Strains used in this study.
Full size table

To produce istobutanol, yeast cells harboring proper plasmids were pre-cultured in 4 mL selective SC medium containing 20 g/L glucose, inoculated to OD600 of 0.2 or 0.5 in 6.5 mL of the same medium in 50 mL conical tube, and then cultured at 30 °C with constant shaking at 170 rpm. For copper induction, overnight culture cells were diluted to OD600 of 0.2 or 0.5, incubated for 7 h to 12 h in SC medium containing 20 g/L glucose, and then induced with appropriate concentrations of CuSO4.

Construction of plasmids

Plasmids and primers used in this study are described in Table S1 and Table S2, respectively. The N-terminally truncated ILV genes (ILV2, ILV5, and ILV3) were obtained by PCR amplification from S. cerevisiae genomic DNA using specific primer pairs (ORF N F and ORF R), generating ILV2ΔN54, ILV5ΔN48, and ILV3ΔN19. These DNA fragments were cloned between BamHI and XhoI sites of p413GPD, p414GPD, and p416GPD30, respectively, resulting in p413GPD-ILV2ΔN54, p414GPD-ILV5ΔN48, and p416GPD-ILV3ΔN19 plasmids. The ORF DNA fragments of alsS(B) from B. subtilis, and alsS(L) from L. lactis were amplified by PCR from each genomic DNA using target-specific primer pairs (ORF F and ORF R). To generate single-gene-expression plasmids, PCR-amplified products were cloned into p413GPD or p413ADH plasmids, resulting in p413GPD-alsS(L) and p413ADH-alsS(B). In p414GPD-ILV5ΔN48 and p416GPD-ILV3ΔN19 plasmid, ILV5ΔN48 and ILV3ΔN19 were replaced with the Kozak sequence-ILV5ΔN48 and Kozak sequence-ILV3ΔN19 fragments obtained by PCR amplification with primer pairs (ORF_K F and ORF R), generating p414GPD-(K)ILV5ΔN48 and p416GPD-(K)ILV3ΔN19 plasmids. Also, in p413ADH-alsS(B), PADH1 part was replaced with PCUP1, generated by PCR with CUP1p F and CUP1p R primers, generating p413CUP1-alsS(B).

To construct a vector for delta-integration, Delta6M-alsS plasmid was constructed as previously described with minor modifications31. Two half fragments of YARCdelta4, YARCdelta4-1 (167 bp, fragment1) and YARCdelta4-2 (170 bp, fragment2), were obtained by PCR amplified from S. cerevisiae genomic DNA. AmpR-expression cassette (fragment3) was obtained by PCR from p413GPD. These 3 DNA fragments were combined by using overlapping PCR, and the resulting DNA fragment containing YARCdelta4-1, AmpR, and YARCdelta4-2 was cloned between the NheI and NotI sites of pUG6MCS32, resulting in Delta6M. Gene-expression cassette, PCUP1-alsS (B)-TCYC1, PCR-amplified from p413CUP1-alsS(B) was cloned between NheI and NotI sites of Delta6M, resulting in Delta6M-alsS.

To construct NTS66M-53 plasmid for NTS site-integration, two half DNA fragments of NTS1-2, NTS1-2a (400 bp) and NTS1-2b (400 bp), were amplified from S. cerevisiae genomic DNA. AmpR cassette and bleOR cassette were obtained from pUG66 vector by PCR. These 4 PCR products were assembled by overlapping PCR, and the resulting DNA fragment, NTS1-2a-bleOR-AmpR-NTS1-2b, was ligated with ILV5ΔN48-expression cassette (PTDH3-Kozak sequence-ILV5ΔN48-TCYC1), resulting in NTS66M-5 plasmid. ILV3ΔN19-expression cassette (PTDH3-Kozak sequence-ILV3ΔN19-TCYC1) flanked by MauBI and and NotI sites were cloned into AscI and NotI site of the NTS66M-5, resulting in NTS66M-35 plasmid.

Analytical methods

Cell growth was determined by the measurement of an optical density at 600 (OD600) with spectrophotometer (Varian Cary® 50 UV-Vis). To determine isobutanol, glucose, and ethanol concentrations, 500 μL of culture supernatants filtered through a 0.22 μm syringe filter were analyzed by High performance liquid chromatography (HPLC) following a previously described procedure17.