Increase in Furfural Tolerance in Ethanologenic Escherichia coli LY180 by Plasmid-Based Expression of thyA

Abstract

Furfural is an inhibitory side product formed during the depolymerization of hemicellulose by mineral acids. Genomic libraries from three different bacteria (Bacillus subtilis YB886, Escherichia coli NC3, and Zymomonas mobilis CP4) were screened for genes that conferred furfural resistance on plates. Beneficial plasmids containing the thyA gene (coding for thymidylate synthase) were recovered from all three organisms. Expression of this key gene in the de novo pathway for dTMP biosynthesis improved furfural resistance on plates and during fermentation. A similar benefit was observed by supplementation with thymine, thymidine, or the combination of tetrahydrofolate and serine (precursors for 5,10-methylenetetrahydrofolate, the methyl donor for ThyA). Supplementation with deoxyuridine provided a small benefit, and deoxyribose was of no benefit for furfural tolerance. A combination of thymidine and plasmid expression of thyA was no more effective than either alone. Together, these results demonstrate that furfural tolerance is increased by approaches that increase the supply of pyrimidine deoxyribonucleotides. However, ThyA activity was not directly affected by the addition of furfural. Furfural has been previously shown to damage DNA in E. coli and to activate a cellular response to oxidative damage in yeast. The added burden of repairing furfural-damaged DNA in E. coli would be expected to increase the cellular requirement for dTMP. Increased expression of thyA (E. coli, B. subtilis, or Z. mobilis), supplementation of cultures with thymidine, and supplementation with precursors for 5,10-methylenetetrahydrofolate (methyl donor) are each proposed to increase furfural tolerance by increasing the availability of dTMP for DNA repair.

INTRODUCTION

Lignocellulose is an abundant renewable resource that can be converted into fuels and chemicals by biocatalysts. In contrast to starch, sugarcane, and sugar beet, the use of lignocellulosic residues and short rotation trees would not directly compete with food production (4, 9, 12). Lignocellulose is designed by nature to resist deconstruction. Dilute acid pretreatment of lignocellulose has been widely investigated to increase enzyme access to cellulose and hydrolyze hemicellulose (10, 35). During this acid pretreatment, small amounts of side products (furans, carboxylic acids, and aromatic species) are produced that retard microbial fermentation (27, 29).

Furfural and 5-hydroxymethylfurfural (HMF) are particularly important as inhibitors (3, 9, 15, 25). Both are formed by the dehydration of sugars (pentoses and hexoses, respectively) during pretreatment (10, 11). The toxicity of hydrolysates is correlated with the concentration of furfural and hydroxymethylfurfural (25, 36). Overliming with Ca(OH)₂ reduced the level of furfural and reduced toxicity (22, 23). Partial toxicity was restored by supplementing overlimed hydrolysate with furfural to the initial level (23).

Strategies have been developed to ameliorate the toxicity of dilute acid hydrolysates. These include reduction of furfural production by using phosphoric acid rather than sulfuric acid during pretreatment (6, 9, 10), the isolation of furfural-resistant mutants of yeast (Saccharomyces cerevisiae) (19) and bacteria (26), and directed genetic modifications (19, 26, 33). Native genes have been identified and overexpressed that catalyze furfural reduction with NADH or NADPH in yeast (7, 14, 17, 18) and bacterial biocatalysts (25, 33). Furfural and other aldehydes have many biological effects on eukaryotes and prokaryotes which may contribute to the inhibition of growth (36). Furfural has been proposed to induce the accumulation of reactive oxygen species (ROS) in yeast (2). Both furfural and ROS products can damage DNA, proteins, lipids, etc. (2, 3, 30). Furfural has been shown to cause DNA damage in Escherichia coli and Salmonella enterica serovar Typhimurium (16, 37) and to inhibit growth until furfural has been substantially metabolized to the less toxic alcohol (17, 33). To date, all genetic changes and genes associated with improvements in furfural tolerance have been limited to the native repertoire of the biocatalyst genome.

In this study, a chromosomal library of Bacillus subtilis YB886 was initially screened for plasmid clones that increased the furfural tolerance of E. coli LY180 (25), expecting to recover novel genes that are absent from or different in E. coli. Clones expressing B. subtilis thyA, a biosynthetic gene present in most organisms, were recovered. E. coli NC3 and Zymomonas mobilis CP4 libraries were subsequently tested with the same result. Clones containing thyA from B. subtilis, Z. mobilis, and E. coli increased furfural tolerance.

MATERIALS AND METHODS

Strains, media, and growth conditions.

The strains, plasmids, and primers used in this study are listed in Table 1. LB medium containing xylose was used for constructions. AM1 minimal salts medium with xylose (20) was used for the maintenance and growth of ethanologenic strains. Solid medium contained 20 g liter⁻¹ xylose. Broth cultures and assays for MIC testing contained 50 g liter⁻¹ xylose. Batch fermentations contained 100 g liter⁻¹ xylose. Cultures were incubated at 37°C unless stated otherwise. Plates were incubated under argon.

Table 1.

Bacterial strains, plasmids, and primers used in this study

Strain, plasmid, or primer	Relevant characteristic(s) or sequencea	Source or reference
Strains
TOP10F′	F′ [lacIq Tn10(Tetr)] mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80lacZΔM15 ΔlacX74 deoR nupG recA1 araD139 Δ(ara-leu)7697 galU galK rpsL(Strr) endA1 λ−	Invitrogen
NC3	B/r derivative with mutation in hsr	34
LY180	ΔfrdBC::(Zmfrg celYEc) ldhA::(Zmfrg casABKo) adhE::(Zmfrg estZPp FRT) ΔackA::FRT rrlE::(pdc-adhA-adhB FRT) ΔmgsA::FRT	26
Plasmids
pCR2.1-TOPO	Plac bla kan	Invitrogen
pUC18	Plac bla	Sigma
pTrc99A	Ptrc bla lacIq	Pharmacia
thyA clones
pLOI5500	Plasmid from B. subtilis YB886 library	This study
pLOI5501	alrB of B. subtilis in pUC18	This study
pLOI5502	yncE and yncF of B. subtilis in pUC18	This study
pLOI5503	cotU and ynzJ of B. subtilis in pUC18	This study
pLOI5504	thyA with native RBS of B. subtilis in pUC18	This study
pLOI5505	alrB, yncE, and yncF of B. subtilis in pUC18	This study
pLOI5506	cotU, ynzJ, and thyA of B. subtilis in pUC18	This study
pLOI5509	thyA with native RBS of E. coli in pUC18	This study
pLOI5514	Plasmid from E. coli NC3 library containing thyA	This study
pLOI5515	Plasmid from E. coli NC3 library containing thyA	This study
pLOI5516	Plasmid from E .coli NC3 library containing thyA	This study
pLOI5531	Plasmid from Z. mobilis CP4 library containing thyA	This study
pLOI5532	Plasmid from Z. mobilis CP4 library containing thyA	This study
pLOI5534	thyA with native RBS of Z. mobilis in pTrc99A	This study
Primers
thyA cloning from B. subtilis Y886
alrBforEcoRI	ATCCGGAATTCCTCCTGAACAATCAGTTATTGAG	This study
alrBrevBamHI	TAGCGGGATCCTATGGAACAAGGAATCGACTGCC	This study
yncE+yncFforEcoRI	ATCCGGAATTCCGAATAATAGCAAGTACGGGATA	This study
yncE+yncFrev	TAGCGGGATCCGCATCCATGCTTTTATGGGGGGC	This study
cotU+ynzJforEcoRI	ATCCGGAATTCGGTTTACTTAGCTCGTGGATAAC	This study
cotU+ynzJrev	TAGCGGGATCCAGGATCTTATTCCACTCACTGTG	This study
thyAbsubforEcoRI	ATCCGGAATTCGGTATACACAAACACCTGACTAT	This study
thyAbsubrevBamHI	TAGCGGGATCCGGATATGATCTAGCTATTCATGG	This study
thyA cloning from E. coli NC3
thyAecolforEcoRI	ATCGCGAATTCCTTTCCATCCCGATGATTGTCGC	This study
thyAecolrevBamHI	TAGCGGGATCCTTCGGGAAGGCGTCTCGAAGAA	This study
thyA cloning from Z. mobilis CP4
thyAzmobForNcoI	CATGCCATGGCGTCGGAACATCAATATCTTCGCC	This study
thyAzmobRevXbaI	TAGTCTAGAACAAAGAAGGACTCATCCTTAA	This study

^aLY180 is an ethanologenic derivative of E. coli W containing integrated genes from other organisms and various chromosomal mutations. Note that the celY gene (coding for endoglucanase) from Erwinia carotovora (celY_Ec) is expressed from a Z. mobilis surrogate promoter (Zmfrg) and integrated into the frd region. The Klebsiella oxytoca casAB operon (casAB_Ko) is expressed from a Z. mobilis surrogate promoter and integrated in the ldhA region. The Z. mobilis pdc-adhA-adhB artificial operon for ethanol production is integrated and expressed from the native rrlE promoter. The Pseudomonas putida estZ (estZ_Pp) gene (coding for acetyl esterase) is integrated into adhE together with a surrogate promoter from Z. mobilis. FRT, FLP recombination target; RBS, ribosome binding site. In the primer sequences, restriction enzyme sites used in construction are underlined.

Genetic methods and construction of genomic libraries.

Standard genetic methods were used for the isolation of DNA and plasmids, digestion with restriction enzymes, PCR amplification of DNA, and plasmid constructions (31). Enzymes were purchased from New England BioLabs (Ipswich, MA) and used as directed by the vendor. Plasmid constructions were confirmed by Sanger sequencing.

Genomic libraries were constructed using DNA purified from Bacillus subtilis YB886, E. coli NC3, and Zymomonas mobilis CP4. Genomic DNA was partially digested with Sau3AI. Fragments of 4 to 6 kb in length were purified from agarose gels and ligated into alkaline phosphatase-treated BamHI sites of either pUC18 (E. coli NC3 and Z. mobilis) or pCR2.1-TOPO (B. subtilis YB886). Ligation products were transformed into electroporation-competent E. coli TOP10F′ (Life Technologies, Grand Island, NY). More than 10,000 colonies were pooled for each organism and used to prepare master libraries of plasmid DNA. Subclones containing only the thyA gene (including the ribosomal binding site and terminator region) from B. subtilis and E. coli were constructed in pUC18. A subclone containing the Z. mobilis thyA gene was constructed in pTrc99A.

Screening for increased furfural tolerance.

Genomic libraries were transformed into chemically competent cells of LY180, spread on AM1 plates containing ampicillin (100 mg liter⁻¹), and incubated for 16 h at 30°C. This allowed expression of plasmid genes for an extended period. Selections were made on plates to provide a microenvironment in which individual clones can benefit from genes in ways that may be masked by competition in liquid culture. Approximately 20,000 of these small colonies were combined in AM1 medium and diluted to an optical density at 550 nm (OD₅₅₀) of 0.1. One drop of this suspension (10⁶ CFU ml⁻¹) was spread on each furfural plate (7.5 mM, 10 mM, or 12.5 mM furfural; 0.1 mM IPTG [isopropyl-β-d-thiogalactopyranoside]; 50 mg liter⁻¹ ampicillin). Large colonies appearing after 48 to 72 h of incubation at 37°C were retested for furfural resistance on plates. Furfural plates were stored under refrigeration and used within 48 h of preparation. Plasmid DNA was prepared from each confirmed positive clone, retransformed into LY180, and again tested on plates. Positive plasmids and subclones were sequenced.

Tube assay for furan tolerance (MIC).

Furfural and 5-hydroxymethylfurfural (HMF) toxicity was examined using culture tubes (13 mm by 100 mm) containing 4 ml of AM1 minimal salts medium (50 g liter⁻¹ xylose, 12.5 mg liter¹ ampicillin, 0.1 mM IPTG, furfural). Tube cultures were inoculated to an initial density of 43 mg (dry cell weight [dcw]) liter⁻¹. Cell mass was measured at 550 nm after incubation (37°C, 48 h) in a shaking water bath (26, 33).

Protein analysis and SDS-PAGE.

Cultures were grown overnight (37°C, 0.86 g dcw liter⁻¹) in tubes containing 4 ml AM1 (5 g liter⁻¹ xylose, 0.1 mM IPTG, 12.5 mg liter⁻¹ ampicillin), harvested by centrifugation (7,000 × g for 6 min at 4°C), washed twice with 10 ml of cold sodium phosphate buffer (50 mM, pH 7.0), and resuspended to a cell density of 4.3 g dcw liter⁻¹ in phosphate buffer (50 mM, pH 7.0) containing 1 mM dithiothreitol. Cells (1 ml) were disrupted using a Fastprep-24 homogenizer (MP Biomedicals, Solon, OH; 6 meters/second for 20 s). After centrifugation (14,500 × g, 20 min, 4°C), supernatant protein was measured using the Pierce bicinchoninic acid (BCA) protein reagent (Thermo, Waltham, PA). Each sample well was loaded with 20 μg protein (supernatant) and analyzed using 12% SDS-PAGE gels (Bio-Rad, Hercules, CA).

Assay of thymidylate synthase.

ThyA activity was determined using a Beckman Coulter DU 800 spectrophotometer (8). A unit of activity is defined as the amount of enzyme that converts 1 μmol of dUMP to dTMP per minute at room temperature. This was determined by measuring the conversion of tetrahydrofolate to dihydrofolate (6,400-OD-unit increase in extinction coefficient at 340 nm). Assays were conducted in the presence and absence of 20 mM furfural.

Effect of B. subtilis thyA expression during batch fermentation.

Preseed cultures were grown in 16- by 150-mm tubes (16 h, 37°C) containing 10 ml of AM1 minimal salts medium (50 g liter⁻¹ xylose and 50 mg liter⁻¹ ampicillin). These were diluted into seed fermenters containing 300 ml AM1 minimal salts medium (100 g liter⁻¹ xylose, 0.1 mM IPTG, 12.5 mg liter⁻¹ ampicillin). Seed cultures were grown for 16 h and used to inoculate test fermenters (43 mg dcw liter⁻¹ inoculum) containing the same medium. Fermentation broth was maintained at pH 6.5 by the automatic addition of 2 M KOH (25, 33). Ethanol was measured using an Agilent 6890N gas chromatograph (Palo Alto, CA) equipped with flame ionization detectors and a 15-m HP-Plot Q Megabore column. Furfural was measured using a Beckman Coulter DU 800 spectrophotometer (21).

RESULTS

Isolation and characterization of furfural resistance genes from genomic libraries.

All previous searches for furfural resistance traits have been limited to investigations of native genomes. We expanded this approach for E. coli by investigating candidate genes from plasmid libraries of B. subtilis YB886. After transformation and selection on furfural plates, 15 large colonies were recovered from six plates. Plasmids from 13 of these clones were confirmed as positive for increased furfural resistance after back-transformation (Fig. 1A and B). All 13 plasmids contained the same B. subtilis DNA fragment based on length and sequence. This plasmid was designated pLOI5500. This DNA fragment was identified as a 5.7-kbp region of the B. subtilis chromosome (GenBank accession no. AL009126.3) containing seven genes. Subclones were constructed and tested for furfural tolerance on solid media (Fig. 2A). All activity was found to reside with the thyA gene encoding thymidylate synthase.

Fig 1.

Example of growth on AM1 plates lacking or containing furfural (48 h, 37°C). (A) No furfural added. (B) Furfural (7.5 mM) added. Strain LY180(pUC18) serves as the control with the unmodified vector. Strain LY180(pLOI5500) contains the thyA gene from B. subtilis YB886. Similar results were obtained with subclones containing functional thyA genes from E. coli LY81 and Z. mobilis CP4.

Fig 2.

Gene organization in original DNA fragments and subclones. Plasmids isolated directly from chromosomal libraries are shown without restriction enzymes. Subclone labels include restriction sites used during construction (see Table 1 for primers). Plasmids were isolated from three chromosomal libraries: B. subtilis YB886 (A), E. coli LY81 (B), and Z. mobilis CP4 (C). Constructs were scored for the ability to increase tolerance to furfural (10 mM) on plates.

This result with B. subtilis was unexpected, leading us to repeat the selection experiment with two additional chromosomal libraries: E. coli NC3 and Z. mobilis CP4. Ten large colonies were recovered from plates using the E. coli NC3 genomic library. Back-transformation of isolated plasmids confirmed that 8 of these plasmids carried a furfural resistance trait. These 8 clones contained three different DNA fragments, designated pLOI5514, pLOI5515, and pLOI5516. All three plasmids were transformed into LY180 and increased growth on plates containing 7.5 mM furfural (Fig. 2B). Sequencing revealed that all three DNA inserts contained the complete E. coli thyA gene.

Six large colonies were recovered from the Z. mobilis CP4 genomic library. Plasmids from only three of these clones conferred furfural resistance on plates after back-transformation. These three plasmids contained fragments of two different sizes, designated pLOI5531 and pLOI5532 (Fig. 2C). Both contained complete thyA genes.

Literature searches and amino acid alignments revealed no special biochemical or genetic features associated with the thyA gene from B. subtilis. Comparison of amino acid sequences revealed 30 to 44% identity among the three thyA genes, with 43 to 55% similarity. Plasmids containing the thyA coding regions from all three organisms caused similar increases in furfural tolerance on plates (Fig. 2).

The thyA genes from B. subtilis and E. coli were subcloned into pUC18. The thyA gene from Z. mobilis was subcloned into pTrc99A. All three subclones were tested in tube assays with furfural (Fig. 3A) and provided an equal benefit. The effectiveness of the B. subtilis thyA was limited to furfural and did not improve tolerance to 5-hydroxymethylfurfural (HMF) (Fig. 3B). Other clones were not tested with HMF.

Fig 3.

Increased expression of thyA increased the MIC for furfural. (A) MIC for furfural. (B) MIC for 5-hydroxymethylfurfural. (C) SDS-PAGE of cytoplasmic extracts from strains harboring pUC18 (lane 1) and pUC18 derivatives containing complete thyA genes. Plasmids pLOI5500 (lane 2; original clone) and pLOI5504 (lane 3; subclone) contain the thyA gene from B. subtilis. Plasmid pLOI5509 (lane 4; subclone) contains the thyA gene from E. coli. Arrows indicate putative ThyA protein from plasmid expression. Protein sizes are in agreement with previous reported (5, 13) predicted values of 32,500 Da for E. coli and 30,500 Da for B. subtilis. M, molecular mass marker lane.

SDS-PAGE revealed prominent protein bands corresponding to the predicted sizes of ThyA from E. coli (32,500 Da) and B. subtilis (30,500 Da) in protein extracts from the subclones (Fig. 3C). No band was evident in the longer, original clone of B. subtilis thyA (pLOI5500), indicating that high levels of expression are not required for furfural tolerance. The B. subtilis thyA subclone was expressed at higher activities than the E. coli subclone (Table 2) and also may be advantageous for future chromosomal integration. Further studies used only the subcloned thyA gene from B. subtilis (pLOI5504).

Table 2.

Effect of 20 mM furfural on ThyA activity

Construct	dUMP concn (mM)	Tetrahydrofolate concn (mM)	Activity (U mg−1 protein) witha:
			0 mM furfural	20 mM furfuralb
LY180(pUC18)	0.042	0.281	0.009 ± 0.001	NDc
LY180(pLOI5509)	0.042	0.281	0.117 ± 0.007	ND
LY180(pLOI5504)	0.042	0.281	0.972 ± 0.030	0.919 ± 0.011
LY180(pLOI5504)	0.042	0.187	0.403 ± 0.049	0.432 ± 0.071
LY180(pLOI5504)	0.011	0.281	0.419 ± 0.018	0.381 ± 0.023

^aCytoplasmic extracts were assayed for ThyA activity in the presence and absence of furfural. No significant inhibition was observed (P > 0.05, Student's t test).

^bActivities with and without furfural were not statistically different (P > 0.05).

^cND, not determined.

ThyA increased furfural tolerance during batch fermentation.

Strain LY180(pUC18) was unable to grow and ferment 10% xylose in the presence of 10 mM furfural (Fig. 4), although furfural was partially metabolized during the 96 h of incubation. Replacement of this vector with pLOI5504 containing the thyA gene from B. subtilis improved furfural tolerance. Growth and fermentation were restored after an initial lag of 48 h. During this lag, growth remained slow until furfural was completely metabolized to the less toxic furfuryl alcohol (Fig. 4) (33). After 96 h, ethanol production by LY180(pLOI5504) in 10 mM furfural medium was equivalent to that of the control lacking furfural after 48 h.

Fig 4.

Batch fermentation of recombinant strain LY180(pLOI5504) in the presence of 10 mM furfural. Strain LY180(pUC18) containing the vector was included as a control. Both strains are grown in the presence and absence of furfural. (A) Cell mass. (B) Furfural. (C) Ethanol.

Supplementation with thymine or thymidine increased furfural tolerance.

ThyA is a key enzyme in the de novo biosynthesis of dTMP, the primary source of pyrimidine deoxyribonucleotides for DNA synthesis (Fig. 5) and DNA repair (28). This deoxynucleotide can also be supplied by the catabolic salvage pathway. Four intermediates in these pathways were tested for their effects on furfural tolerance (Fig. 6A and B). A small benefit was observed from the addition of deoxyuridine. No benefit was observed from the addition of deoxyribose. However, thymine and thymidine each caused similar dose-dependent increases in furfural tolerance, with an optimum at 0.2 mM. At this concentration, the MIC for furfural was increased from 10 mM to 15 mM, equivalent to the increase provided by plasmid expression of thyA (Fig. 3A). The combination of expression of B. subtilis thyA from pLOI5504 and supplementation with thymidine was no more effective than either alone (data not shown).

Fig 5.

Pathways for the synthesis of dTMP. (A) De novo biosynthesis pathway. (B) Salvage pathway. Genes are shown in italics. Dotted arrows show enzymatic links between these two pathways. This diagram is based on charts from EcoCyc (28).

Fig 6.

Effect of biosynthetic intermediates for TMP on the furfural tolerance of LY180 (10 mM, 48 h). (A) Dose-response curve. (B) MIC. (C) Summary bar graph. THF, tetrahydrofolate.

Supplementation with serine and tetrahydrofolate increased furfural tolerance.

Metabolites involved in the biosynthesis of 5,10-methylenetetrahydrofolate (serine and tetrahydrofolate) were also investigated (Fig. 6C). This pathway is associated with trimethylpyrimidine inhibition and with thymineless death (1, 28). Supplementation with serine or tetrahydrofolate was beneficial (0.1 mM and 0.5 mM) for furfural tolerance. Neither alone was as effective as 0.1 mM thymine. However, the combination of 0.1 mM serine and 0.5 mM tetrahydrofolate was equivalent to 0.1 mM thymidine. Together, these results with plasmids and supplements demonstrate that furfural tolerance can be increased by approaches that could increase the supply of pyrimidine deoxyribonucleotides.

ThyA activity is not inhibited by furfural. ThyA activity was increased up to 100-fold by expression from plasmids (Table 2). Although this increase was shown to increase furfural tolerance, addition of furfural in vitro did not inhibit thymidylate synthase activity. To examine the possibility that furfural is a competitive inhibitor masked by excess substrates, activity measurements were repeated at subsaturating levels of each substrate. No significant furfural inhibition of ThyA was observed under any of these conditions (P > 0.05). Although increased expression of thyA is clearly beneficial for furfural tolerance, this enzyme does not appear to be the immediate target site for furfural.

DISCUSSION

The thyA gene was recovered in plasmids from three different genomic libraries during selection for furfural tolerance and was shown to increase furfural resistance. However, furfural did not directly inhibit thymidylate synthase. Furfural is known to have multiple sites of inhibitory action (2, 3, 25). One of these is the depletion of NADPH by the action of YqhD, a furfural oxidoreductase that converts furfural to the alcohol form (26, 32). This enzyme has a low K_m for NADPH and competes with biosynthetic enzymes (26). Furfural tolerance was improved by deleting the yqhD gene, by expressing low levels of NADH/NADPH transhydrogenase (pntAB) (25), and by expressing an NADH-dependent furfural reductase (fucO) (33). The durations of the furfural-induced lags in growth and fermentation were shortened by each of these genetic changes, similar to the effects of ThyA expression and thymidine addition on the furfural-induced lag.

NADPH is required for the de novo synthesis of deoxyribonucleotides during the reduction of ribose and for the regeneration of the methyl donor (5,10-methylenetetrahydrofolate) for ThyA (24). It is possible that this creates competition for available NADPH. However, complete blockage of dTMP synthesis by furfural is unlikely since thymineless death rapidly follows TMP starvation by folate inhibitors such as trimethylpyrimidine (1). With furfural, cells continue to grow slowly and metabolize furfural (19, 26, 33). Furfural is known to damage DNA in E. coli (16) and to induce oxidative stress in yeast (2, 3). Repair of DNA damage is essential for growth (1). An added requirement for dTMP during DNA repair (1, 24) could partially deplete the TMP and NADPH pools and slow growth. Alternatively, furfural could partially inhibit reactions in the folate cycle, limiting the methyl donor for dTMP biosynthesis by ThyA, analogous to aminopterin and trimethylpyrimidine and thymineless death (1). Both mechanisms are consistent with our data that demonstrate that supplementation with thymine or thymidine (but not deoxyuridine or deoxyribose), supplementation with precursors of 5,10-methylenetetrahydrofolate (serine plus tetrahydrofolate), and increased expression of ThyA are beneficial for furfural tolerance.