Plasmidic Expression of nemA and yafC* Increased Resistance of Ethanologenic Escherichia coli LY180 to Nonvolatile Side Products from Dilute Acid Treatment of Sugarcane Bagasse and Artificial Hydrolysate

Abstract

Hydrolysate-resistant Escherichia coli SL100 was previously isolated from ethanologenic LY180 after sequential transfers in AM1 medium containing a dilute acid hydrolysate of sugarcane bagasse and was used as a source of resistance genes. Many genes that affect tolerance to furfural, the most abundant inhibitor, have been described previously. To identify genes associated with inhibitors other than furfural, plasmid clones were selected in an artificial hydrolysate that had been treated with a vacuum to remove furfural. Two new resistance genes were discovered from Sau3A1 libraries of SL100 genomic DNA: nemA (N-ethylmaleimide reductase) and a putative regulatory gene containing a mutation in the coding region, yafC*. The presence of these mutations in SL100 was confirmed by sequencing. A single mutation was found in the upstream regulatory region of nemR (nemRA operon) in SL100. This mutation increased nemA activity 20-fold over that of the parent organism (LY180) in AM1 medium without hydrolysate and increased nemA mRNA levels >200-fold. Addition of hydrolysates induced nemA expression (mRNA and activity), in agreement with transcriptional control. NemA activity was stable in cell extracts (9 h, 37°C), eliminating a role for proteinase in regulation. LY180 with a plasmid expressing nemA or yafC* was more resistant to a vacuum-treated sugarcane bagasse hydrolysate and to a vacuum-treated artificial hydrolysate than LY180 with an empty-vector control. Neither gene affected furfural tolerance. The vacuum-treated hydrolysates inhibited the reduction of N-ethylmaleimide by NemA while also serving as substrates. Expression of the nemA or yafC* plasmid in LY180 doubled the rate of ethanol production from the vacuum-treated sugarcane bagasse hydrolysate.

INTRODUCTION

Sugars derived from lignocellulosic residues have the potential to serve as carbohydrate substrates for microbial fermentation into biobased products with minimal impact on food and feed (1–3). However, the deconstruction of lignocellulose and hydrolysis to sugar monomers requires harsh treatments, such as the use of dilute mineral acids at elevated temperatures (4, 5). Inhibitory side products, such as furfural, soluble fragments from lignin, and acetic acid, are formed during dilute acid pretreatment; these compounds retard growth and fermentation. The removal of inhibitors after dilute acid pretreatment typically involves additional process steps, such as fiber separation, countercurrent washing, and overliming (6, 7), all of which increase costs. Genetic engineering of resistance into biocatalysts represents a cost-effective approach for inhibitor mitigation.

Furfural is the dominant inhibitor in dilute acid hydrolysates of lignocellulose, a dehydration product of pentose sugars (primarily xylose). Many resistance genes associated with furfural tolerance have been identified for ethanologenic Escherichia coli LY180 and other strains of E. coli (8–12). Resistant derivatives of ethanologenic E. coli LY180, such as strains EMFR9 (13) and SL100 (20), have been isolated after serial transfers in AM1 mineral salts medium containing furfural (reagent) and in AM1 mineral salts medium containing toxic levels of sugarcane bagasse hydrolysate (SCBHz), respectively. Both selections resulted in mutants that are resistant to furfural, although SL100 is also resistant to other compounds. Despite the progress with furfural resistance, little progress has been reported with other inhibitors.

Artificial hydrolysates have been produced by heating xylose and mineral acids to provide a simpler mixture of inhibitors. As expected, furfural, a compound that increased the toxicity of other inhibitors in binary mixtures, was the most abundant side product and inhibitor (1). Additional reaction products in artificial hydrolysates included glycolaldehyde, formate, lactate, acetate, lactaldehyde, phenolics, and pseudo-lignin (15–18). Vacuum evaporation has been shown to remove furfural from hemicellulose hydrolysates and to reduce, but not eliminate, toxicity (19–21). Toxic nonvolatile compounds remained after furfural evaporation. Full toxicity was regained by the restoration of furfural (19).

In this study, we identified two genes that increase resistance to the nonvolatile compounds in dilute acid hydrolysates of sugarcane bagasse and in artificial hydrolysates. A vacuum-treated artificial hydrolysate (PX). PX was prepared by autoclaving 5% xylose in 1% phosphoric acid for 2 h at 140°C). Volatile constituents were removed under a vacuum to make PXV. PXV was used as a selection agent in broth to enrich for LY180 clones with plasmids containing resistance genes from SL100 (22). Two beneficial regions of the chromosome were identified: nemR′-nemA-gloA rnt ′lhr and dkgB′ yafC* yafD′. Clones expressing these genes were more resistant to a vacuum-treated sugarcane bagasse hemicellulose hydrolysate and an artificial hydrolysate than the parent strain (empty-vector control).

MATERIALS AND METHODS

Strains and media.

Two strains of ethanologenic E. coli were used in this study: LY180 (23, 24) and SL100 (14, 22, 25). SL100 is a hydrolysate-resistant derivative of LY180, selected by serial transfers for more than a year in AM1 mineral salts medium (26) containing sugarcane bagasse hydrolysate (SCBHz). Except for plasmid constructions using Luria broth, strains were grown and maintained on either AM1 medium alone, AM1 medium mixed with an artificial hydrolysate, or AM1 medium mixed with sugarcane bagasse hydrolysate. Xylose was added as needed to provide 5% sugars. Media were adjusted to pH 6.5 prior to inoculation (incubation at 37°C).

SL100 chromosomal library.

Chromosomal DNA from SL100 was partially digested with Sau3A1 and was ligated (2- to 8-kbp fragments) into the dephosphorylated BamHI site of pUC19. TOP10F′ chemically competent cells were used as the host (100 mg ampicillin liter⁻¹). Colonies were pooled and were used to prepare a chromosomal library of plasmids.

Preparation of hydrolysates.

An artificial hydrolysate (PX) was prepared by autoclaving a mixture of xylose (50 g liter⁻¹) and phosphoric acid (10 g liter⁻¹) for 2 h at 140°C (Hirayama autoclave, model HA-305M; Amerex Instruments, Inc., Lafayette, CA). SCBHz was prepared using a Metso-Valmet continuous digester with a screw feeder (185°C, 7.5 min, 8 kg phosphoric acid per dry tonne sugarcane bagasse, 3 tonne h⁻¹) as described previously (19). Where indicated, volatiles such as furfural were removed from hydrolysates by evaporation (55°C) to half the original weight. Weight loss was replaced with distilled water, and the products were designated PXV and SCBHzV, respectively.

Construction of pLOI5883 derivatives for gene expression.

Plasmids, strains, and primers are listed in Table 1. The coding regions (ATG to TAA) of individual genes (nemA, gloA, rnt) were amplified using SL100 chromosomal DNA as a template unless specified otherwise. An artificial ribosomal binding site was supplied on the primers. Amplified genes were ligated into the pLOI5883 expression vector (EcoRI to HindIII [see Fig. S1 in the supplemental material]) between the Ptrc promoter and the rrnB terminator (11, 15) to construct pLOI5908 (nemA), pLOI5909 (gloA), and pLOI5910 (rnt). These three adjacent genes were also amplified together and ligated to construct pLOI5911 (nemA-gloA rnt). Similar plasmids were constructed for yafC (pLOI5913), yafC* (pLOI5914), and a combination of nemA and yafC* (pLOI5926). Sequences were confirmed by Sanger sequencing. An inducer (10 μM isopropyl-β-d-thiogalactopyranoside [IPTG]) was added where indicated.

TABLE 1.

Bacterial strains, plasmids, and primers

Strain, plasmid, or primer	Relevant characteristic(s) or sequence	Source or reference(s)
Strains
LY180	ΔfrdBC::(frgZm celYEc) ΔldhA::(frgZm casABKo) adhE::(frgZm estZPp FRT) ΔackA::FRT rrlE::(pdc adhA adhB FRT) ΔmgsA::FRT	23
SL100	LY180 improved for hydrolysate tolerance	22
TOP10F′	F′[lacIqTn10 (Tetr)] mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80lacZΔM15 ΔlacX74 recA1 araD139 Δ(ara-leu)7697 galU galK rpsL(Strr) endA1 nupG λ−	Invitrogen (Carlsbad, CA)
Plasmids
pUC19	bla lacZα ori	New England Biolabs
pLOI5883	lacIq Ptrc rrnB terminator bla RSF1010	11, 19
pLOI5905	Plasmid from library selection strain AQ23	This study
pLOI5906	Plasmid from library selection strain AQ35	This study
pLOI5907	Plasmid from library selection strain AQ49	This study
pLOI5908	nemA gene in pLOI5883	This study
pLOI5909	gloA gene in pLOI5883	This study
pLOI5910	rnt gene in pLOI5883	This study
pLOI5911	nemA-gloA-rnt genes in pLOI5883	This study
pLOI5912	Plasmid from library selection and in strain AQ78	This study
pLOI5913	yafC gene in pLOI5883	This study
pLOI5914	yafC mutant (yafC*) in pLOI5883	This study
pLOI5926	nemA-yafC* in pLOI5883	This study
Primers
nemA	Forward, AGTGAATTCAAGGAGATATACCATGTCATCTGAAAAACTGTATTCCCC	This study
	Reverse, AGTAAGCTTTTACAACGTCGGGTAATCGGTAT
gloA	Forward, AGTGAATTCAAGGAGATATACCATGCGTCTTCTTCATACCATGC	This study
	Reverse, AGTAAGCTTTTAGTTGCCCAGACCGCG
rnt	Forward, AGTGAATTCAAGGAGATATACCATGTCCGATAACGCTCAACTTAC	This study
	Reverse, AGTAAGCTTATTACACCTCTTCGGCGGC
nemA-gloA-rnt	Forward, AGTGAATTCAAGGAGATATACCATGTCATCTGAAAAACTGTATTCCCC	This study
	Reverse, AGTAAGCTTATTACACCTCTTCGGCGGC
nemRA upstream	Forward, CATTAACGGGTCTGGTCGGT	This study
	Reverse, GCAGAAATTTGCGTGGCTT
M13	Forward (−20), GTAAAACGACGGCCAG	This study
	Reverse, CAGGAAACAGCTATGAC
nemA sequencing	Forward, AATGTGGTGTCCGGCATCA	This study
	Reverse, CGCACGTCTGGTACTGGAA
yafC and yafC*	Forward, AGTGAATTCAAGGAGATATACCATGAAAGCCACGTCGGAAG	This study
	Reverse, AGTAAGCTTTTAAGCCTCTCTGACAGCTCC
99 A sequencing	Forward, GCAGGTCGTAAATCACTGC	This study
	Reverse, CTTCTCTCATCCGCCAAAAC
nemA-yafC*	Forward, GGATGGAAGCTTGGCTGTTTTGGCGG	This study

Enrichment for clones conferring resistance to PXV.

The SL100 chromosomal library was transformed into LY180 (parent), grown overnight in 100 ml AM1 mineral salts medium (250-ml flask) with 5% xylose and 100 μg ml⁻¹ ampicillin (37°C, 50 rpm), and inoculated (optical density at 550 nm [OD₅₅₀], 0.1) into 100 ml AM1 broth with 80% (vol/vol) PXV. Although little growth was observed after 48 h, cells were harvested by centrifugation (10 min, 2,000 × g) and were transferred to fresh AM1 broth containing 80% (vol/vol) PXV. Growth was abundant after 24 h. Plasmids were extracted, transformed into LY180, and tested for resistance to AM1 broth with 80% (vol/vol) PXV by measuring ethanol production (see Fig. S2 in the supplemental material).

Measuring toxicity.

The toxicity of hydrolysates was tested by measuring ethanol production in tube cultures (13- by 100-mm tubes; 4 ml broth) containing mixtures of water, hydrolysate, and the constituents of AM1 medium (26). Cultures were inoculated to an initial OD₅₅₀ of 0.1 and were incubated for 48 h at 37°C (13, 19). Due to color, ethanol production (determined by gas chromatography) and visual observation of cells were used to measure fermentation and confirm growth. In some experiments, cells were harvested by centrifugation, washed twice in AM1 medium, and resuspended in AM1 medium. This removed soluble color and allowed an estimation of cell growth by turbidity. In all cases, the shapes of the OD₅₅₀ curves were very similar to the ethanol measurements, indicating a close relationship (see Fig. S3 in the supplemental material).

Assay of NemA reductase activity.

NemA reductase activity was measured as the N-ethylmaleimide (NEM)-dependent oxidation of NADPH (27, 28) unless stated otherwise. For enzyme assays, cells (50 ml) grown in AM1 xylose broth containing 100 mM morpholinepropanesulfonic acid (MOPS) (pH 7) were harvested by centrifugation at an OD₅₅₀ of approximately 1, washed twice with 10 ml of cold potassium phosphate buffer (50 mM; pH 7.0), and resuspended in phosphate buffer (3 ml). After disruption with 0.1-mm glass spheres for 20 s using a FastPrep-24 instrument (MP Biomedicals LLC, Santa Ana, CA, USA), cell debris was removed by centrifugation (20 min, 14,000 × g). The soluble protein fraction was used for assays of NEM-dependent activity at 22°C in 50 mM potassium phosphate (pH 7.0) with 0.2 mM NADPH, 0.1 mM NEM, cell extracts, and a vacuum-treated hydrolysate (PXV or SCBHzV) as indicated. Protein was measured with the bicinchoninic acid (BCA) reagent using bovine serum albumin as a standard. Vacuum-treated hydrolysates were tested as inhibitors of NemA activity using NEM as the electron acceptor. Vacuum-treated hydrolysates were also tested as inducers of NemA activity in LY180 and as sources of electron acceptors (without NEM) for NemA reductase using protein lysates of LY180(pLOI5908). One unit of activity is defined as the amount of enzyme that converts 1 μM NADPH to NADP⁺ min⁻¹.

Additional experiments were conducted to explore the potential proteolysis of NemA. Disrupted cell extracts were prepared from LY180 grown with 5% SCBHzV (induced) and AM1 medium alone (uninduced). Extracts were diluted 1:1 with phosphate buffer, and a mixture of equal amounts of induced and uninduced extracts was also made. All three samples were incubated for 9 h at 37°C. NemA activity was measured at time zero and at 3, 6, and 9 h. No loss of activity was observed with uninduced or induced LY180 extracts.

Expression of nemA as measured by real-time PCR.

The expression of nemA mRNA was determined by use of polA as a reference gene. RNA was isolated as described previously (23). Message abundance was compared for the parent strain, LY180, and the mutant strain, SL100, after growth in the presence and absence of vacuum-treated 5% SCBHz (inducer).

Bench-scale fermentations.

Hydrolysate resistance genes were expressed from pLOI5883 derivatives in LY180 during bench-scale fermentations (19) in AM1 medium (300 ml broth) containing 20% (vol/vol) SCBHzV. The results were compared to those for controls (empty vector) grown in the same medium and in AM1 medium without a hydrolysate. Sufficient xylose and glucose were added to provide 100 g total sugar liter⁻¹ (75 g glucose liter⁻¹ and 25 g xylose liter⁻¹), amounts similar to those for the hydrolysis of cellulose and hemicellulose. Fermentations were maintained at 37°C and pH 6.5 (by the automatic addition of 2 M potassium hydroxide). Mixtures were inoculated with a 24-h broth culture to provide an initial OD₅₅₀ of 0.10 (approximately 0.05 g [dry weight of cells] liter⁻¹).

Analyses.

Ethanol was measured as described previously (19) using an Agilent 6890N gas chromatograph equipped with flame ionization detectors and a 15-m HP-Plot Q Megabore column. Furfural and xylose were measured using an Agilent 1200 high-performance liquid chromatograph (HPLC) with an Aminex HPX-87P column. Experiments were conducted at least twice with three replicates each. Results are reported as averages with standard deviations. Significance was inferred (P ≤ 0.05) from a two-tailed Student t test by use of GraphPad Prism software for computations.

RESULTS AND DISCUSSION

Vacuum treatment removed furfural and decreased the toxicity of an artificial hydrolysate.

Previous studies have shown that xylose can be heated in dilute sulfuric acid to produce an artificial hydrolysate (17). However, the Florida process (25) for lignocellulose fermentation uses phosphoric acid pretreatment instead of sulfuric acid to avoid the need for exotic metallurgy. We confirmed that a similar brown, toxic artificial hydrolysate (PX) can be made from xylose and phosphoric acid (2 h at 140°C) in the absence of lignin and cellulose. The resulting PX contained 13.6 mM furfural (Table 2). AM1 broth containing 40% (vol/vol) PX was sufficiently toxic to completely inhibit the growth (as determined by visual examination) and fermentation of the parental strain, LY180 (Fig. 1A). SL100, a mutant of LY180 selected for resistance to sugarcane bagasse hydrolysate (SCBHz), was also more resistant to PX than LY180 and required 60% (vol/vol) PX for a similar degree of inhibition. Furfural removal by vacuum treatment (PXV) (Table 2) reduced the toxicity of hydrolysates to both LY180 and SL100 (Fig. 1B). Ethanol production by LY180 was partially inhibited with 60% (vol/vol) PXV and fully inhibited with 80% (vol/vol) PXV. Ethanol production by SL100 was inhibited by less than one-third with 80% (vol/vol) PXV. Prior selection for resistance to SCBHz with SL100 also coselected for resistance to volatile (primarily furfural) and nonvolatile inhibitors in the artificial hydrolysate (heated xylose in a phosphoric acid solution). These results confirmed that SL100 is a potential source of genes for resistance to residual toxins in PXV.

TABLE 2.

Concentrations of xylose and furfural in artificial and sugarcane bagasse hydrolysates

Dilute acid hydrolysate	Concn (mM)
	Xylose	Furfural
PX	296.7 ± 5.5	13.6 ± 1.4
PX + vacuum	296.0 ± 3.2	0.0 ± 0.0
SCBHz	275.4 ± 1.7	20.6 ± 1.3
SCBHz + vacuum	274.5 ± 1.9	0.0 ± 0.0

FIG 1.

Effect of vacuum evaporation on the toxicity of an artificial hydrolysate (PX) containing 50 g liter⁻¹ xylose and 10 g liter⁻¹ phosphoric acid (2 h at 140°C). (A) Toxicity to LY180 (parent) and SL100 (mutant selected for resistance to sugarcane bagasse hydrolysate) of PX without a vacuum treatment to remove volatiles. ETOH, ethanol. (B) Toxicity after vacuum treatment. (C) Example of results of screening for LY180 clones harboring pUC19 derivatives that express SL100 genes and confer resistance to an 80% (vol/vol) vacuum-treated artificial hydrolysate (PXV) in AM1 medium. Clones with the highest ethanol production contained SL100 nemA genes. Clones with intermediate ethanol production contained SL100 yafC* genes.

Cloning and sequencing of genes from SL100 that increase tolerance of vacuum-treated hydrolysates.

Differences in resistance to PXV between LY180 and SL100 provided a basis for the selection of genes conferring resistance (Fig. 1B; also see Fig. S2 in the supplemental material). Strain LY180 was transformed with a plasmid library of SL100 chromosomal fragments, and transformants were selected for growth in AM1 broth (5% xylose) containing sufficient PXV (80% [vol/vol]) to inhibit ethanol production and growth (by visual examination). Cells grew well after the second transfer and were harvested for plasmid purification. The resulting plasmids were back-transformed into LY180, diluted, and spread on AM1 solid medium (2% xylose) with 100 μg/ml ampicillin. A total of 475 clones were screened from three independent plasmid libraries of SL100. Each clone was tested for ethanol production in tube cultures containing 80% (vol/vol) PXV. Examples of the results of colony screening are shown in Fig. 1C. Positive clones produced 6-fold- to 10-fold-higher levels of ethanol than the control (empty pUC19 vector) and negative clones.

DNA fragments in 15 clones (those with the highest ethanol titers) were sequenced. Many were siblings with identical fragments of SL100 DNA, confirming the rigor of the selection. All clones fell into two groups (Fig. 2; see also Table S4 in the supplemental material). Twelve clones contained a large fragment with ′nemR-nemA-gloA rnt lhr′ (six unique fragments plus siblings), and three contained smaller fragments with ′dkgB yafC yafD′ (two unique fragments plus one sibling).

FIG 2.

Two chromosomal regions of SL100 with genes that increased resistance to a vacuum-treated artificial hydrolysate (PXV). The cloned fragments encompassing nemR′ to lhr′ and dkgB′ to yafD′ are shown between double forward slashes. Nucleotide mutations are shown as boxed capital letters. (A) Larger fragment with nemA. Although the cloned fragment did not contain any mutations, the adjacent repressor gene (nemR) contained a mutation in the upstream regulatory region. This mutation was present in strain SL100 and absent in strain LY180. The nemR mutation could affect the expression of nemA and other downstream genes. (B) Smaller fragment with yafC*. The smaller fragments contained a single nucleotide mutation (boxed capital letters) in the carboxy-terminal region of this predicted LysR-type regulator, designated yafC*. This mutation was present in SL100 but absent in LY180 and could affect the expression and function of many genes. Deletion of this gene has been shown to decrease survival after ionizing radiation (33).

Sequencing revealed that the nemR gene was incomplete in the plasmids with the large fragments and was unlikely to function. The nemA, gloA, and rnt genes were complete and did not contain mutations. The nemA and gloA genes are part of the cellular defense system for cytoplasmic detoxification (27). The rnt gene encodes RNase T, an exonuclease involved in trimming stable RNAs and in tRNA maturation (29–32). Although no mutations were found in the cloned ′nemRA-gloA, rnt, or lhr′ region, a chromosomal mutation in SL100 was found in the upstream regulatory region of nemR. This mutation was absent in LY180 and could affect nemRA-gloA expression (Fig. 2A). The yafC gene encodes a putative transcriptional regulator of unknown function (33). A single base mutation was found in the C terminus of the yafC coding region (D275G), designated yafC* (Fig. 2B) (present in SL100). This mutation (D275G) could alter the function of YafC and the expression of regulated genes.

Testing subclones with single genes related to hydrolysate tolerance.

Each of the genes in the large nemA fragment (′nemR-nemA-gloA rnt), yafC (LY180), and yafC* (SL100) were cloned into expression vector pLOI5883 (see Fig. S1 in the supplemental material). An artificial ribosomal binding site was supplied by the primers used for amplification (nemA, gloA, rnt, and yafC*). SL100 served as a template for all the genes except yafC (wild type from LY180). Each gene was ligated into pLOI5883 (RSF1010-based expression vector) (see Fig. S1 in the supplemental material) to produce pLOI5908, pLOI5909, pLOI5910, pLOI5913, and pLOI5914, respectively (Table 1). Two combinations of genes were also constructed (nemA-gloA rnt with intergenic regions [pLOI5911] and nemA yafC* [pLOI5926]).

Expression plasmids were transformed into LY180, and the transformants were tested for ethanol production in AM1 medium containing 80% (vol/vol) PXV (Fig. 3A) and in AM1 medium containing 20% (vol/vol) SCBHzV (Fig. 3B). With AM1 containing 80% (vol/vol) PXV (without IPTG), ethanol production was significantly increased (P ≤ 0.05) over that with the empty vector (pLOI5883) by plasmids expressing nemA (pLOI5908), the 3-gene combination (pLOI5911), or the nemA-yafC* combination (pLOI5926). The addition of IPTG caused a small but significant (P ≤ 0.05) increase in ethanol production by LY180(pLOI5908), expressing nemA, and LY180(pLOI5911), expressing the 3-gene combination. When the transformants were tested in AM1 medium containing 20% (vol/vol) SCBHzV, ethanol production was significantly increased (P ≤ 0.05) by plasmids expressing nemA (pLOI5908), yafC* (pLOI5914), the 3-gene combination (pLOI5911), or the nemA-yafC* combination (pLOI5926) over that with the empty-vector control (pLOI5883) or yafC (wild type), with or without IPTG. The addition of the inducer provided a small benefit for nemA constructs but was detrimental for yafC* constructs (Fig. 3A and B). Plasmids expressing gloA, rnt, or yafC individually were of little benefit with 80% (vol/vol) PXV or 20% (vol/vol) SCBHzV. However, plasmids expressing the 3-gene combination (pLOI5911) or the nemA-yafC* combination (pLOI5926) were significantly better (P ≤ 0.05) at ethanol production than the plasmid expressing nemA alone when tested with 20% (vol/vol) SCBHzV (Fig. 3B). This plasmid-mediated increase in resistance was specific for nonvolatile components of hydrolysates and did not increase resistance to furfural (Fig. 3C). LY180 expressing nemA alone or in any combination exhibited an increase in resistance to vacuum-treated hydrolysates (PXV and SCHzV). None of the expression vectors increased furfural tolerance in LY180 (Fig. 3C).

FIG 3.

Plasmid expression of cloned SL100 genes increased the resistance of LY180 to hydrolysates. The pLOI5883 expression vector is included as a control (empty vector). An inducer (IPTG) was added as indicated. Shown is ethanol production by LY180 constructs with 80% (vol/vol) PXV (A), 20% (vol/vol) SCBHzV (B), or 10 mM furfural (C) in AM1 medium or by SL100 constructs with 35% (vol/vol) SCBHzV in AM1 medium (D).

Plasmids with various genes were also transformed into SL100, and the transformants were tested for resistance to 35% (vol/vol) SCBHzV (Fig. 3D). Although the effects were small, SL100 containing plasmids expressing nemA alone, gloA alone, or the 3-gene combination produced more ethanol (P ≤ 0.05) than the vector control or SL100 containing plasmids expressing rnt alone, yafC* alone, or the combination of nemA and yafC*. It was not surprising that none of the constructs with yafC* increased the resistance of SL100 to SCBHzV, since the yafC* mutation is already present on the SL100 chromosome. In contrast, the combination of nemA and yafC* was the most beneficial for ethanol production by LY180 in SCBHzV (Fig. 3B).

The addition of IPTG was generally beneficial for ethanol production with plasmid constructs lacking yafC or yafC*, a putative transcriptional regulator. The addition of IPTG decreased ethanol production in all constructs containing yafC or yafC*. Increasing the expression of this putative regulator appears to hinder cellular functions.

Vacuum-treated hydrolysates contain substrates for NemA.

NemA is a versatile NADPH-dependent flavoprotein reductase (old yellow enzyme) capable of reducing a broad range of organic compounds, including electrophiles (quinones, glyoxals, trinitrotoluene) (27) and even inorganic substrates, such as nitrates and chromates (28, 34). Considering the diversity of compounds formed by acid treatment of xylose (17, 35), it is not surprising that some components of PXV (Fig. 4, left) and SCBHzV (Fig. 4, center) can serve as electron acceptors for NemA/NADPH (Table 3). Although activity in LY180 was low with a hydrolysate as the sole source for electron acceptors, the activity in LY180 harboring the nemA expression vector (pLOI5908) was twice that with the respective vector controls, in agreement with measurements of nemA-encoded activity. With vacuum-treated hydrolysates as potential substrates, activity plateaued or declined with increases in the concentrations of PXV and SCBHzV Fig. 4, left and center). These unexpected kinetics can be attributed in part to the dual action of hydrolysate components as both substrates for NemA and inhibitors of NemA activity with NEM as the substrate.

FIG 4.

Effects of vacuum-treated hydrolysates on NemA activity. (Left and center) A vacuum-treated artificial hydrolysate (PXV) or a vacuum-treated sugarcane bagasse hydrolysate (SCBHzV) can serve as a substrate for NADPH-dependent reduction by NemA (open symbols with broken lines, right axis). Vacuum-treated hydrolysates also inhibit N-ethylmaleimide reduction by NemA (filled symbols with solid lines, left axes). (Right) Induction of NemA activity (with NEM as the electron acceptor). Shown is the activity induced in LY180 and SL100 during growth in AM1 medium alone (control), with vacuum-treated PX, and with vacuum-treated SCBHz (filled bars, left axis). IPTG-induced NemA activity in LY180 harboring pLOI5908 (open bar, right axis) is also included for purposes of comparison.

TABLE 3.

Hydrolysate as a substrate and an inhibitor of NemA activity in protein extracts of LY180

Hydrolysate and concn (%)	Sp act (U mg−1) of NemA in LY180 harboring the following plasmid with the indicated substrate:
	nemA expression vector		Empty vector
	Hydrolysate	NEM (0.1 mM)	Hydrolysate	NEM (0.1 mM)
None	0.0127 ± 0.0004	1.0940 ± 0.0430	0.0100 ± 0.0004	−0.0020 ± 0.0004
PXV
6.25	0.034 ± 0.006	0.438 ± 0.025
12.5	0.034 ± 0.008	0.238 ± 0.014	0.015 ± 0.001	−0.004 ± 0.001
25	0.039 ± 0.0048	0.108 ± 0.014	0.016 ± 0.001	−0.002 ± 0.001
SCBHzV
0.125	0.018 ± 0.004	0.416 ± 0.036
0.25	0.018 ± 0.003	0.281 ± 0.0165	0.011 ± 0.001	−0.002 ± 0.001
0.625	0.020 ± 0.003	0.128 ± 0.002	0.008 ± 0.001	−0.002 ± 0.001
1.25	0.007 ± 0.002	0.014 ± 0.004

Vacuum-treated hydrolysates contain inhibitors of N-ethylmaleimide reduction by NemA.

NemA activity is typically measured with NEM as the electron acceptor, although the physiological substrate for this enzyme is unknown (36, 37). The NEM-dependent activity of this enzyme was inhibited by the addition of the vacuum-treated hydrolysate PXV (Table 3; Fig. 4, left) or SCBHzV (Table 3; Fig. 4, center). SCBHzV was a more potent inhibitor. The addition of 1.25% (vol/vol) SCBHzV to the reaction mixture was sufficient to inhibit 99% of the NemA activity with NEM as the substrate (1.1 U mg⁻¹ protein without an inhibitor). In contrast, the addition of 25% (vol/vol) PXV to the reaction mixture inhibited only 90% of NemA activity. Both hydrolysates appear to contain a combination of substrates and inhibitors that affect NemA activity. The dose-dependent inhibition by hydrolysates may be responsible for the unusual kinetics observed when hydrolysates were tested as sources of electron acceptors for NemA activity (Fig. 4, left and center).

Vacuum-treated hydrolysates as inducers of nemA activity.

The nemR operon contains three genes (nemR-nemA-gloA) that are repressed by nemR in the absence of an inducer (27). In AM1 medium without an inducer, NemA activity (with NEM as the electron acceptor) in SL100 was >20-fold higher (0.09 U/mg protein) than in lysates from LY180 (parent strain), a finding consistent with a role for NemA in hydrolysate resistance (Fig. 4, right). NemA activity in LY180 harboring the nemA expression plasmid pLOI5908 (1.1 U mg⁻¹) (Fig. 4, right panel, right axis) was 13-fold higher than that in SL100 (Fig. 4, right panel, left axis) and >250-fold higher than that in LY180 alone. NemA activity in LY180 was induced 2.5-fold and 8-fold (0.035 U) by growth in the presence of 50% (vol/vol) PXV and 5% (vol/vol) SCBHzV, respectively. In contrast, a high level of NemA activity was produced in SL100 without any hydrolysate (0.09 U), increasing less than 2-fold when 5% (vol/vol) SCBHzV was included during growth (Fig. 4, right). This partial derepression of nemA in SL100 is presumed to result from the base mutation in the upstream regulatory region of nemR in SL100, a transcriptional regulator.

Transcriptional regulation of nemA was confirmed by measuring message levels (by real-time PCR) with polA as the reference gene. Induction with 5% (vol/vol) SCBHzV increased mRNA levels in LY180 5-fold over those in uninduced LY180. In the resistant mutant SL100, nemA mRNA was >200-fold more abundant than in the parent strain.

Potential regulation by proteinases was also investigated. Disrupted cell extracts were prepared from LY180 grown with 5% SCBHzV (induced) or in AM1 medium alone (uninduced). Extracts were diluted 1:1 with phosphate buffer, and a mixture of equal amounts of induced and uninduced extracts was also made. All three samples were incubated for 9 h at 37°C. NemA activity was measured at time zero and at 3, 6, and 9 h. No loss of NemA activity was observed with uninduced or induced LY180 extracts, or with the mixture.

SL100 was more resistant to PXV and SCBHzV than the parent strain and had higher uninduced levels of NemA activity. Expression of nemA from plasmids in LY180 (the parent strain) increased NemA activity and nemA mRNA expression and increased tolerance to vacuum-treated hydrolysates. This effect was specific for vacuum-treated hydrolysates and did not increase tolerance to furfural. SL100 exhibited a further increase in resistance to vacuum-treated hydrolysates when nemA and gloA genes were coexpressed from a single plasmid, pLOI5911 (Fig. 3D).

Plasmid expression of resistance genes improved fermentation performance.

Plasmids containing nemA or yafC* were expressed individually in LY180 during batch fermentations of 20% (vol/vol) SCBHzV in AM1 medium (supplemented with glucose and xylose to make 100 g total sugar liter⁻¹). Under these conditions, the maximum rate of ethanol production with LY180(pLOI5883) (empty-vector control) was approximately half that observed without any hydrolysate added (Fig. 5, top right). Production of 30 g ethanol liter⁻¹ required 8 days with a hydrolysate and the empty vector but only 2 days in AM1 medium without a hydrolysate. Expression of yafC* or nemA (individually) substantially improved ethanol production, reducing the time required to produce 30 g ethanol liter⁻¹ to 3 days.

FIG 5.

Plasmid expression of nemA and yafC* increases resistance to 20% (vol/vol) SCBHzV in AM1 medium during pH-controlled fermentation. Glucose and xylose were added to adjust the sugar concentration to 100 g liter⁻¹. An empty vector (pLOI5883) in AM1 medium with or without a hydrolysate (labeled AM1) is included as a control. Fermentations were sampled for ethanol at 24-h intervals. (Top left) Total sugars; (top right) ethanol; (bottom left) glucose; (bottom right) xylose.

These two genes may be useful for engineering hydrolysate resistance into future biocatalysts for renewable products. Plasmid expression of yafC has been shown to increase survival of ionizing radiation (32). Both genes improved the rates of growth (observations of turbidity) and sugar utilization (Fig. 5, top left and bottom). With either nemA or yafC*, fermentation of glucose was complete after 3 days, but with the empty vector control, it required more than 8 days. The fermentation of xylose was partially inhibited by the addition of a hydrolysate. In a hydrolysate medium, xylose was used concurrently with glucose but at a much lower rate. After 2 days, the rate of xylose utilization was increased by plasmids expressing nemA or yafC*. However, fermentations failed to completely utilize xylose even after 8 days. Without a hydrolysate, xylose fermentation was completed after 4 days. Near-theoretical yields were obtained from both sugars.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.