Crucial role of 4-deoxy-L-erythro-5-hexoseulose uronate reductase for alginate utilization revealed by adaptive evolution in engineered Saccharomyces cerevisiae

In brown macroalgae, alginate and D-mannitol are promising carbohydrates for biorefinery. Saccharomyces cerevisiae is widely used as a microbial cell factory, but this budding yeast is unable to utilize either alginate or D-mannitol. Alginate can be depolymerized by both endo-type and exo-type alginate lyases, yielding a monouronate, 4-deoxy-L-erythro-5-hexoseulose uronate (DEH), a key intermediate in the metabolism of alginate. Here, we constructed engineered two S. cerevisiae strains that are able to utilize both DEH and D-mannitol on two different strain backgrounds, and we also improved their aerobic growth in a DEH liquid medium through adaptive evolution. In both evolved strains, one of the causal mutations was surprisingly identical, a c.50A > G mutation in the codon-optimized NAD(P)H-dependent DEH reductase gene, one of the 4 genes introduced to confer the capacity to utilize DEH. This mutation resulted in an E17G substitution at a loop structure near the coenzyme-binding site of this reductase, and enhanced the reductase activity and aerobic growth in both evolved strains. Thus, the crucial role for this reductase reaction in the metabolism of DEH in the engineered S. cerevisiae is demonstrated, and this finding provides significant information for synthetic construction of a S. cerevisiae strain as a platform for alginate utilization.


Supplementary methods
Plasmids. All plasmids used in this study are listed in Supplementary Table S1. Genomic PCR was conducted using S. cerevisiae BY4742 genomic DNA as a template with either KOD-Plus Neo (Toyobo, Japan) for cloning or KOD FX Neo (Toyobo) for verification. The amplified sequence was confirmed after cloning. DSF1 and HXT17 were amplified by genomic PCR with primers (7 and 8 for DFS1, 5 and 6 for HXT17) and inserted into FseI and SalI of pAT426, respectively, yielding pMK5503. pMK5503 was digested by XhoI/SacI, and the resultant XhoI/SacI fragment carrying DSF1 and HXT17 was inserted with In-fusion into EcoRV of pFA6a-kanMX6, a PCR fragment amplified with primers 1 and 2 (a template; pFA6a-kanMX6), resulting in pMK5547. The pseudogene sequence (YIL174w/175w/176c, 902-b, corresponding to the sequence of 8,794-9,696 nt of chromosome IX; SGD) was amplified with genomic PCR (primers 3 and 4) and inserted into HpaI of pMK5547 with disruption of the original HpaI, yielding pMK5550 in which only one HpaI exists at 500 nt in YIL174w/175w/176c. pMK5076, which contained the cyc8Δ1139-1164 allele from MK4416 (YCplac33-cyc8 (4456) 1 ), was digested with EcoRI/NaeI, blunted, and ligated, resulting in pMK5482 lacking CEN4/ARS1. Only one EcoNI exists in 1,822 nt in the cyc8Δ1139-1164 allele.

(Supplementary
The yopt_A1-R', which is flanked by 5'-TEF1t and 3'-CYC1t, was inserted into SalI of pRS423, producing pMK5090. The pMK5090 was digested with SnaBI/XcmI, blunted, and then self-ligated to remove only the 2 µ region, yielding pMK5090-1. A probable pseudogene, YOL153C, was amplified with genomic PCR with primers 11 and 12 and inserted into BamHI of pMK5090-1 with disruption of this original BamHI, producing pMK5474 in which only one NcoI exists at 903 nt in YOL153C (1,746 b). The yopt_DHT1 was inserted into SalI of pAT422, resulting in pMK5531. The yopt_DHT1 flanked by 5'-TDH3p and 3'-TDH3t was amplified by PCR with primers 13 and 14 and with pMK5531 as a template, and was inserted into SacI of pMK5474 with a disruption of the original SacI, yielding pMK5552 in which only one NcoI exists at 903 nt in YOL153C (1,746 b). A c.50A>G mutation was introduced in A1-R' with inverse PCR using primers 42 and 43 and with pMK5552 as a template, yielding pMK5552-1. The A1-R'(c.50A>G), that was again amplified with PCR using primers 44 and 45 and with the pMK5552-1 as a template, was inserted into EcoRV/XhoI of pMK5552, creating pMK5859 that is the pMK5552 in which only c.50A>G was introduced into A1-R'.
The yopt_A1-I was amplified by PCR with primers 26 and 27, inserted into NdeI/BamHI of pET-14b (Novagen), yielding pMK5631. The c.50A>G was also introduced into A1-R' in pMK4699 (A1-R' in pET-21b) 4 with PCR using primers 46 and 47 and with pMK4699 as a template, resulting in pMK5922.

Strains.
All strains used in this study are listed in Supplementary Table S2.
Transformation of S. cerevisiae was conducted with the lithium acetate/single-stranded carrier DNA/polyethylene glycol method 8 . ADE2 and TRP1 were deleted from BY4742 as described previously using 5-fluoroorotic acid 9 , yielding MK5315. In order to replace CYC8 with the cyc8Δ1139-1164 allele in the genomic DNA of the D452-2 strain, pMK5482 linearized by the digestion of EcoNI, was introduced into the D452-2 strain, resulting in MK5502, as described using 5-fluoroorotic acid 9 .
To introduce yopt_kdgK and yopt_eda, pMK5471 linearized by the digestion of PstI was introduced into the D452-2 and MK5315 strains, producing MK5517 and MK5524, respectively, in which the targeting sites were confirmed by genomic PCR with primers (21 and 22; 18 and 23) followed by sequencing. To further introduce both yopt_A1-R' and yopt_DHT1, pMK5552 linearized by the digestion of NcoI was introduced into the MK5517 and MK5524 strains, yielding MK5590 and MK5591, respectively, in which the targeting sites were confirmed by genomic PCR with primers (21 and 24; 23 and 25). Likewise, to integrate both yopt_A1-R'(c.50A>G) and yopt_DHT1, pMK5859 was linearized by the digestion of NcoI and was again introduced into the MK5517 and MK5524 strains, yielding MK5906 (D_DEH+_E17G) and MK5909 (BY_DEH+_E17G), respectively. Again, the targeting sites were confirmed by genomic PCR with primers (21 and 24; 23 and 25).
To introduce DSF1/HXT17, pMK5550 was linearized by the digestion of HpaI and was introduced into the MK5590 and MK5591 strains, yielding MK5609 (D_DEH+) and MK5622 (BY_DEH+), and also into the BY4742 and D452-2 strains, yielding MK5580 and MK5583, coli transformant showing kanamycin-resistance due to KanMX from pMK5550, and the sequence of the plasmid was determined. The ρ 0 mutants were created with 25 µg/mL ethidium bromide as described 10 .