Enhancement of furan aldehydes conversion in Zymomonas mobilis by elevating dehydrogenase activity and cofactor regeneration

Reduction evaluation of alcohol dehydrogenase and aldo-keto reductase gene expression

The results indicated that the expression of ZMO1771, ZMO1696, and ZMO1722 enhanced the furfural and HMF conversion, in which the expression of ZMO1771 encoding an iron-dependent alcohol dehydrogenase showed the most significant enhancement. 1.51 g/L of furfural or 1.89 g/L of HMF was converted at the mid-log phase of the fermentation after ZMO1771 was over-expressed, leading to the increase of furfural conversion by 33.63% and of HMF conversion by 35.00%, comparing to the control without ZMO1771 over-expression (Table 2). However, the cell growth, glucose consumption, and ethanol production of the three recombinants above did not show significant improvement in the synthetic medium with furfural or HMF addition. The rest of the seven genes, including two adh genes (ZMO0062 and ZMO1993) and five akr genes (ZMO0976, ZMO1344, ZMO1673, ZMO1773, and ZMO1984), did not show significant enhancement on furfural and HMF conversion. The in vitro enzymatic analysis in previous study [27] as well as in present study (Table 1) showed ZMO0976 encoding xylose reductase possessed NADPH-dependent furfural and HMF reduction activity, but no enhancement was observed in Z. mobilis ZM4 by expressing ZMO0976. The phenomenon suggests that the in vitro capacity of an enzyme may not go to the in vivo performance in Z. mobilis.

Then the inhibitors conversion of all the ten recombinants was further evaluated in corn stover hydrolysate containing 56.74 g/L of glucose, 24.80 g/L of xylose, 0.62 g/L of furfural, 0.34 g/L of HMF, 2.85 g/L acetic acid as well as 4-hydromethylbenzoaldehyde, vanillin, syringaldehyde, and other inhibitors. The results again showed that the expression of ZMO1771 significantly increased furfural and HMF conversion rate (Table 3; Fig. 1): furfural by 0.26 g/L at 12 h and 0.40 g/L at 24 h, HMF by 0.03 g/L at 12 h, and 0.07 g/L at 24 h, respectively, while other gene expression did not show the obvious effect. The significantly difference between the corn stover hydrolysate and synthetic medium is that the cell growth, glucose consumption, and ethanol production increased by 35.00, 78.79, 100% at 36 h, and 33.33, 74.14, 100% at 48 h, respectively (Fig. 1). The results suggest that the expression of ZMO1771 significantly increased the furfural and HMF conversion as well as ethanol productivity in the high inhibitors containing corn stover hydrolysate.

Recombinant strains	Furfural conversion (g/L)		HMF conversion (g/L)
	12 h	24 h	12 h	24 h
ZM4(pHW20a-gfp) (Control)	0.15 ± 0.01	0.33 ± 0.02	0.02 ± 0.01	0.05 ± 0.01
ZM4(pHW20a-ZMO0062)	0.15 ± 0.01	0.28 ± 0.02	0.01 ± 0.00	0.01 ± 0.00
ZM4(pHW20a-ZMO1696)	0.18 ± 0.00	0.33 ± 0.02	0.02 ± 0.00	0.05 ± 0.03
ZM4(pHW20a-ZMO1722)	0.18 ± 0.04	0.32 ± 0.05	0.02 ± 0.01	0.05 ± 0.01
ZM4(pHW20a-ZMO1771)	0.26 ± 0.02	0.40 ± 0.01	0.03 ± 0.01	0.07 ± 0.00
ZM4(pHW20a-ZMO1993)	0.16 ± 0.00	0.33 ± 0.01	0.01 ± 0.01	0.03 ± 0.00
ZM4(pHW20a-ZMO0976)	0.17 ± 0.02	0.31 ± 0.00	0.01 ± 0.00	0.03 ± 0.02
ZM4(pHW20a-ZMO1344)	0.14 ± 0.01	0.27 ± 0.01	0.01 ± 0.00	0.03 ± 0.01
ZM4(pHW20a-ZMO1673)	0.18 ± 0.00	0.29 ± 0.03	0.01 ± 0.00	0.06 ± 0.00
ZM4(pHW20a-ZMO1773)	0.13 ± 0.00	0.23 ± 0.01	0.02 ± 0.01	0.02 ± 0.01
ZM4(pHW20a-ZMO1984)	0.20 ± 0.02	0.31 ± 0.04	0.02 ± 0.00	0.04 ± 0.02

The corn stover hydrolysate contained 59.45 g/L of glucose, 24.03 g/L of xylose, 3.21 g/L of acetic acid, 0.38 g/L of HMF and 0.69 g/L of furfural at 30 °C. Standard deviations were derived from at least two independent determinations

Compared with the results in simple-defined RM medium, the improved conversion of furfural and HMF by expression of ZMO1771 may effectively alleviate the synergistic inhibition of furan aldehydes with other inhibitors in the hydrolysate, which then activate cell growth and ethanol production. Besides, the complicated hydrolysate components, such as the mineral element contents and other rich nutrients may also play a positive role to facilitate the reduction of furan aldehydes by ZMO1771, just as the previous studies on S. cerevisiae showed that enriched media, including complex media or high glucose concentrations, influence the inhibitor tolerance of cells [28, 29]. Thus, we recommend the use of a lignocellulosic hydrolysate for final fermentation evaluation of recombinant strains.

Cofactor preference of ZMO1771 and reduction modification on furfural and HMF

The reductase enzymes for furan aldehydes conversion were either NADH or NADPH dependent, and the regeneration of the NADH or NADPH cofactor is an important factor for furan aldehydes conversion. The enzyme activity assay showed that furan aldehydes reductase activity of Z. mobilis ZM4(pHW20a-ZMO1771) increased by 52.8% on furfural and 63.8% on HMF when NADPH was used as the cofactor, while the activity was almost same to the control when NADH was used (Table 1). The cofactor preference of NADPH was further proved using the purified protein from the expression of ZMO1771 in E. coli BL21: the reductase activity was 3.632 U/mg protein on furfural substrate and 1.999 U/mg protein on HMF substrate when NADPH was used as the cofactor, while no reductase activity was detected when NADH was used. These results confirm that the ZMO1771 encoding enzyme is NADPH dependent for furfural and HMF reduction, rather than NADH.

Preference of NADPH by the relevant reductase enzymes on the furfural and HMF conversion pathway may negatively interfere with the NADPH/NADP⁺ balance of Z. mobilis and then weaken its tolerance to furfural and HMF. Additional NADPH supply may further enhance the furfural and HMF conversion of Z. mobilis when ZMO1771 is over-expressed. Two transhydrogenase genes pntAB and udhA from E. coli involved in NADPH generation were selected and co-expressed with ZMO1771 to facilitate the availability of NADPH. The two genes encode the only two transhydrogenases documented in living organisms responsible for the interconversion between NADH and NADPH (NADH + NADP⁺ ↔ NAD⁺ + NADPH) [30]. Each gene was inserted into pHW20a together with ZMO1771, but in reversed direction under the regulation of Pgap promoter (Additional file 1: Figure S1). The PntAB and UdhA activities in the co-expression and the single expression of pntAB or udhA were essentially the same (0.037 U/mg protein for PntAB in both expressions, and 0.002 and 0.003 U/mg protein for UdhA activity, respectively), implying that each gene was expressed successfully in Z. mobilis.

The co-expression of the gene ZMO0367 encoding glucose-6-phosphate dehydrogenase (G6PDH) with ZMO1771 in Z. mobilis was also evaluated. The G6PDH activity of the co-expression of ZMO0367 was considerably high (3.011 U/mg protein), but the co-expression of ZMO0367 with ZMO1771 did not show any enhancement on furfural and HMF conversion in RM medium comparing to the single expression of ZMO1771 (Table 4).

Fermentation evaluation of the co-expression recombinants in corn stover hydrolysate

The co-expression recombinant of udhA with ZMO1771 in Z. mobilis was grown in the corn stover hydrolysate containing 59.45 g/L of glucose, 24.03 g/L of xylose, 3.21 g/L of acetic acid, 0.69 g/L of furfural, and 0.38 g/L of HMF as well as other inhibitors (Fig. 2). The co-expression exhibited the maximum conversion capacity, approximately 50% of the initial furfural was reduced at 12 h and completely reduced at 24 h, 12 h shorter than the time used for complete conversion of furfural by the recombinant with the ZMO1771 expression only. The new recombinant also completely converted HMF within 48 h, comparing with 12% of the residual HMF left even after 60 h’s culture by the recombinant with ZMO1771 expression only. The cell growth, glucose consumption, and ethanol production were also accelerated by 55.56, 60.00, and 46.67% at 36 h, respectively; at 48 h, the accelerations were 10.00, 12.12, and 25.00%, respectively. The maximum cell growth and ethanol titer of the co-expression recombinant were approximately equal to the control at 60 h.

The Ethanol fermentation performance in corn stover hydrolysate revealed that the co-expression of udhA with ZMO1771 promoted the furfural and HMF reduction rate by elevating the NADPH availability by delivering the proton from NADH which was not the cofactor of ZMO1771 for furfural or HMF conversion reaction. This regeneration of NADPH from NADH by UdhA, however, could compete for NADH with the biosynthesis of ethanol from acetaldehyde by ADHI or ADHII and inhibit the ethanol formation rate when furfural or HMF conversion reactions proceed. Figure 2c showed this tendency of slow ethanol formation by Z. mobilis during the period of furfural and HMF conversion. When furfural was completely converted and most of HMF was converted, the recombinant Z. mobilis of the co-expression of ZMO1771 with udhA started its faster rate of ethanol generation than the wild Z. mobilis strain in which furfural and HMF existed at relatively high levels. The fermentation results of the co-expression of ZMO1771 with udhA suggest that cofactor manipulation was an effective tool for increasing the conversion of furan aldehyde inhibitors in Z. mobilis.