- Open Access
Combined inactivation of the Clostridium cellulolyticum lactate and malate dehydrogenase genes substantially increases ethanol yield from cellulose and switchgrass fermentations
Biotechnology for Biofuelsvolume 5, Article number: 2 (2012)
The model bacterium Clostridium cellulolyticum efficiently degrades crystalline cellulose and hemicellulose, using cellulosomes to degrade lignocellulosic biomass. Although it imports and ferments both pentose and hexose sugars to produce a mixture of ethanol, acetate, lactate, H2 and CO2, the proportion of ethanol is low, which impedes its use in consolidated bioprocessing for biofuels production. Therefore genetic engineering will likely be required to improve the ethanol yield. Plasmid transformation, random mutagenesis and heterologous expression systems have previously been developed for C. cellulolyticum, but targeted mutagenesis has not been reported for this organism, hindering genetic engineering.
The first targeted gene inactivation system was developed for C. cellulolyticum, based on a mobile group II intron originating from the Lactococcus lactis L1.LtrB intron. This markerless mutagenesis system was used to disrupt both the paralogous L-lactate dehydrogenase (Ccel_2485; ldh) and L-malate dehydrogenase (Ccel_0137; mdh) genes, distinguishing the overlapping substrate specificities of these enzymes. Both mutations were then combined in a single strain, resulting in a substantial shift in fermentation toward ethanol production. This double mutant produced 8.5-times more ethanol than wild-type cells growing on crystalline cellulose. Ethanol constituted 93% of the major fermentation products, corresponding to a molar ratio of ethanol to organic acids of 15, versus 0.18 in wild-type cells. During growth on acid-pretreated switchgrass, the double mutant also produced four times as much ethanol as wild-type cells. Detailed metabolomic analyses identified increased flux through the oxidative branch of the mutant's tricarboxylic acid pathway.
The efficient intron-based gene inactivation system produced the first non-random, targeted mutations in C. cellulolyticum. As a key component of the genetic toolbox for this bacterium, markerless targeted mutagenesis enables functional genomic research in C. cellulolyticum and rapid genetic engineering to significantly alter the mixture of fermentation products. The initial application of this system successfully engineered a strain with high ethanol productivity from cellobiose, cellulose and switchgrass.
Cellulose is the most abundant renewable feedstock on earth for biofuel production . However, the economic feasibility and sustainability of cellulosic biofuels are limited by the inefficient breakdown of recalcitrant cellulose fibers into sugars  and their fermentation into biofuels and other metabolites . While this biological conversion can be achieved in separate steps, a more economical solution will be to combine the processes into a single step, termed consolidated bioprocessing (CBP) . An efficient CBP scheme could either exploit a consortium of cellulolytic and ethanologenic microorganisms, or it could use a single microorganism with both activities. No natural microorganisms have been identified that possess all the necessary characteristics of an ideal CBP strain for industrial applications , therefore genetic engineering will likely be required for constructing an efficient CBP microorganism.
Clostridium cellulolyticum is a model mesophilic clostridial species for studying cellulose and hemicellulose degradation, and an excellent candidate for CBP strain development based on its ability to ferment its hydrolysis products to ethanol and organic acids [6, 7]. C. cellulolyticum has previously been transformed by electroporation , enabling random transposon-based mutagenesis , and laying the groundwork for more advanced genetic manipulation. Broadly, our goal is to build a genetic platform for the functional genomic analysis and genetic engineering of C. cellulolyticum to better understand the genetic and metabolic processes that would be required to enhance biofuel production.
Rational metabolic engineering includes three primary strategies to divert carbon and electron flow from byproducts to increase ethanol yield and purity: introduce heterologous genes, increase the expression of native genes, and disrupt genes required for competing pathways in mixed fermentations . Examples of the first strategy include the heterologous expression of the Zymomonas mobilis pyruvate decarboxylase and alcohol dehydrogenase genes to increase ethanol production in monosaccharide-fermenting Escherichia coli and later C. cellulolyticum. The heterologous expression of five genes in C. cellulolyticum converted pyruvate to isobutanol . Targeted mutagenesis experiments have also led to significant strain improvements. Deletions of the L-lactate dehydrogenase (ldh), phosphotransacetylase (pta) and acetate kinase (ack) genes of Thermoanaerobacterium saccharolyticum substantially increased the yield of ethanol from glucose and xylose fermentations, creating a homoethanologenic strain . However, a deletion of the pta gene of Clostridium thermocellum eliminated acetate production but had minimal effect on ethanol production from cellulose fermentation .
Targeted gene inactivation has never been reported for C. cellulolyticum, which significantly impedes genetic engineering efforts. Here, we describe a targeted gene knockout system for C. cellulolyticum based on a mobile group II intron originating from the Lactococcus lactis L1.LtrB intron, which has been successfully used in several clostridial species to disrupt genes [15–17]. This intron is a fragment of catalytic RNA that inserts into double-stranded DNA in a site-specific manner called retrohoming . The intron used in this study consists of a 915-bp L. lactis L1.LtrB-ΔORF intron flanked by short 5' and 3' exons and a downstream ltrA gene encoding a protein with endonuclease and reverse transcriptase activities . The target specificity is primarily determined by base pairing of intron RNA and target site DNA, and intron insertion does not require host-supplied factors; therefore, the intron can be easily modified to insert into virtually any DNA sequence in principle . However, a replicative plasmid and a strong promoter are required to drive the transcription of the intron and ltrA genes in each host strain .
As an initial step to increase ethanol production by C. cellulolyticum during cellulosic fermentation, we decided to disrupt the production of lactate, one of the cells' major fermentation products, by inactivating the lactate dehydrogenase (ldh) gene. C. cellulolyticum has two paralogs of the ldh gene that were both originally predicted to encode L-lactate dehydrogenase (LDH) enzymes: Ccel_0137 and Ccel_2485. An intron-based gene inactivation system was used to create markerless mutations in both genes to clarify their metabolic roles during growth on complex carbohydrates. The method applied here did not require the use of a retrotransposition-activated marker for integrant selection , allowing the rapid introduction of both Ccel_0137 and Ccel_2485 mutations in a single C. cellulolyticum strain.
Results and Discussion
Construction of Ccel_0137 and Ccel_2485 mutations in C. cellulolyticum
Insertion mutations in the paralogous ldh genes, Ccel_0137 and Ccel_2485, were constructed using pLyc1217Er-based vectors that encoded targeted introns (targetrons) specific for each gene. C. cellulolyticum cells transformed with these vectors produced erythromycin-resistant colonies on agar plates. Colony PCR was performed using forward and reverse primers flanking the intron insertion sites of the targeted genes to screen transformants for the desired gene insertions (Figure 1). The proportion of transformants containing markerless chromosomal insertion mutations depends on the retargeting efficiency . Eight colonies transformed with pLyc1217Er0137 were screened by PCR for insertion in the Ccel_0137 gene: seven were wild-type, whereas one contained both an intron insertion allele (2,164-bp PCR product) and the wild-type allele (1,249-bp PCR product) (Additional file 1). This colony was streaked on an agar plate containing 15 μg/ml erythromycin, and colony PCR was performed until a single intron-inserted PCR band was identified in the tested colonies, demonstrating the homogeneity of the purified Ccel_0137 mutant. Similarly, the second ldh paralog was disrupted using vector pLyc1217Er2485, and eight colonies were screened for insertion in the Ccel_2485 gene: four showed an intron insertion allele (2,197-bp PCR product) and the wild-type allele (1,282-bp PCR product), whereas two contained a single intron insertion allele (2,197-bp PCR product) (Additional file 1). Colony 4 became the Ccel_2485 mutant used for further analysis. Colony PCR was performed to confirm the correct insertion sites and intron orientation of the isolated mutant strains (Figure 2A,B). PCR products from the mutants amplified by primers MdhF/MdhR and LdhF/LdhR were sequenced, verifying the correct intron insertions in the mutant strains (Additional file 2).
Because an intron inserted in the sense orientation could be spliced from the precursor mRNA by the plasmid-encoded protein LtrA  resulting in a wild-type mRNA, the plasmid was cured from all of the mutants to achieve an unconditional knockout before the fermentation and analytical experiments. The knockout plasmids pLyc1217Er0137 and pLyc1217Er2485 were cured from Ccel_0137 and Ccel_2485 mutants by serial passage in non-selective medium without erythromycin (Figure 2C). These plasmid-cured mutants were unable to grow in liquid medium supplemented with erythromycin, demonstrating efficient plasmid curing and markerless mutagenesis. The plasmid-cured Ccel_2485 mutant was transformed with plasmid pLyc1217Er0137, and a strain carrying an intron insertion in the Ccel_0137 gene was isolated (Figure 2D). Therefore, multiple gene inactivations in the same C. cellulolyticum strain could be introduced sequentially by pLyc1217Er-based plasmids targeting different genes.
Southern blots were performed on genomic DNA from the Ccel_0137, Ccel_2485 and double mutants after plasmid curing, using an intron-specific probe (Figure 2E). Both single mutants contained a single intron insertion in the chromosome with expected sizes of 4.1 or 4.9 kbp. The double mutant contained both bands. The knockout vector pLyc1217Er showed a band of 1.5 kbp, which was not present in the mutants. No band was detected in chromosomal DNA from wild-type cells. These hybridization experiments confirmed specific gene disruption, with no residual vector remaining in the mutants.
These Ccel_0137 and Ccel_2485 mutants are the first strains of C. cellulolyticum produced by targeted mutagenesis. The strong Clostridium pasteurianum ferredoxin promoter enabled intron and ltrA transcription in C. cellulolyticum. This mode of accurate and specific chromosomal insertions had a sufficiently high efficiency that mutants could be readily identified by screening a small number of colonies without chromosomal antibiotic counter selection. This high targeting efficiency is essential for C. cellulolyticum gene inactivation, due to the cells' low transformation efficiency. Another advantage of this intron-based knockout system is that the gene inactivation occurs in a one-step intron insertion process. The traditional gene disruption strategy based on homologous recombination usually takes two steps: one single crossover followed by a second crossover. That strategy requires at least two selectable markers, and a large number of colonies often need to be screened to identify a mutant . The intron-based gene knockout system used to construct these mutations was especially advantageous for the manipulation of these obligately anaerobic clostridia. The lack of a selectable marker in the intron and the ease of plasmid curing will also accelerate future efforts to produce strains with multiple mutations, introduced sequentially using pLyc1217Er. Because several nearly identical intron fragments would be present on the genome after multiple intron insertions, one possible problem is that homologous recombination might occur among these introns, causing genetic instability. Therefore, the long-term goal of genetic engineering of C. cellulolyticum using a multiple-intron insertion strategy may require the inactivation of the organism's recA gene.
Characterization of Ccel_0137 and Ccel_2485 single mutants
High-performance liquid chromatography (HPLC) analysis of fermentation products from Ccel_2485 mutant cells grown in defined VM medium with cellobiose showed that the lactate concentration decreased significantly to 0.09 ± 0.01 g/l, less than 10% of wild-type levels (Figure 3B). A small but significant (P < 0.0005) increase in acetate was observed for the mutant (0.75 ± 0.04 g/l versus 0.54 ± 0.06 g/l for wild-type cells). Ethanol concentrations were 30% higher in the mutant. In contrast, the Ccel_0137 mutant produced significantly more lactate than wild-type cells (1.4 ± 0.07 g/l versus 1.0 ± 0.22 g/l; P < 0.01) and 19% more ethanol than wild-type. Cultures of both mutants grew to similar maximum turbidities, with a similar exponential growth rate compared to wild-type cells (Figure 4). Only the inactivation of the Ccel_2485 gene produced the expected defect in lactate production, establishing that this locus encodes the primary lactate dehydrogenase.
The Ccel_2485 gene is an ortholog of the Clostridium thermocellum ldh (Cthe_1053) gene that was characterized previously (Additional file 3) . The paralogous Ccel_0137 gene shares the same orientation as the adjacent Ccel_0138 gene on the C. cellulolyticum chromosome, forming a putative operon. The latter locus is predicted to encode a nicotinamide adenine dinucleotide phosphate (NAD(P))+-dependent malic enzyme, which catalyzes the oxidative decarboxylation of L-malate producing NAD(P)H and pyruvate. The cofactor binding site of the Ccel_0138 protein more closely resembles the sequence of NADP+-dependent malic enzymes than NAD+-dependent proteins , and the Ccel_0138 protein is probably an ortholog of the NADP+-dependent malic enzyme that was characterized previously from C. thermocellum. Both the sequence and organization of the Ccel_0137 and Ccel_0138 genes are conserved in C. thermocellum (Cthe_0345 and Cthe_0344). Therefore the Ccel_0137 protein was likely to function as a L-malate dehydrogenase (MDH). The MDH enzyme was proposed to act in concert with the malic enzyme to convert oxaloacetate to pyruvate and transfer electrons from NADH to form NADPH (Figure 5).
Malate dehydrogenase and malic enzyme activities were measured in cell-free extracts from wild-type, Ccel _2485 and Ccel _0137 mutants (Table 1). The Ccel_0137 mutant demonstrated 28% of the NADH-dependent malate dehydrogenase activity observed in wild-type cells, while activity in the Ccel_2485 mutant was not significantly different from wild-type. Although both mutants had significant malate dehydrogenase activity, the Ccel_0137 gene appears to encode the primary malate dehydrogenase. None of the extracts catalyzed the reduction of oxaloacetate using NADPH. All three strains had similar levels of malic enzyme activity (Table 1). This enzyme required NADP+ and L-malate for activity: reactions containing NAD+ or D-malate had 13% or 0.5% relative activity, respectively. These results indicate that the Ccel_0137 insertion has no polar effect on Ccel_0138 expression and support the model of oxaloacetate-dependent transhydrogenase activity illustrated in Figure 5. This pathway was predicted to be one of two primary modes of NADPH production for biosynthesis in C. thermocellum. Disrupting the MDH-malic enzyme transhydrogenase system would be expected to cause an increase in flux through alternative, ferredoxin-dependent transhydrogenase systems. Both the C. cellulolyticum and C. thermocellum genomes also encode homologs of the Clostridium kluyveri nfnAB genes that encode a ferredoxin-dependent transhydrogenase system, which provides another source of NADPH to these organisms . Ferredoxin is primarily reduced by the pyruvate:ferredoxin oxidoreductase complex. This NfnAB system catalyzes an electron bifurcation event, where the exergonic oxidation of reduced ferredoxin and NADH is coupled to the endergonic reduction of NADP+ (Figure 5). Unlike C. thermocellum, C. cellulolyticum lacks homologs of the cation-translocating ferredoxin:NAD oxidoreductase (Rnf complex) , which conserves energy (as ion-motive force) and transfers electrons from reduced ferredoxin.
The family of 2-hydroxyacid dehydrogenases includes both lactate and malate dehydrogenases, and a single discriminating amino acid substitution (Gln102Arg) converted the Geobacillus stearothermophilus lactate dehydrogenase into an efficient malate dehydrogenase . The corresponding amino acids are Gln86 in the Ccel_2485 protein and Arg81 in the Ccel_0137 protein. Only five amino substitutions in the E. coli MDH, including the Arg to Gln substitution, increased the enzyme's specificity for pyruvate by a factor of 4.6 × 109. Together with the biochemical data described above, these comparisons suggest that both enzymes catalyze lactate and malate dehydrogenase reactions, although the Ccel _2485 gene encodes the primary lactate dehydrogenase and the Ccel_0137 gene encodes the primary malate dehydrogenase enzyme. The clostridial LDH and MDH proteins share significant sequence similarity and have overlapping substrate specificity, therefore they could have diverged following the duplication of a single ancestral gene in the clostridial lineage. However, a phylogeny that includes more homologs does not support this hypothesis (Additional file 3). Instead, most LDH and MDH proteins appear to derive from separate linages, although bootstrap support is weak for deeply branching nodes in this tree. LDH activity levels were below the limits of detection for all C. cellulolyticum cell-free extracts (Table 1), potentially due to low levels of the enzyme at the tested growth stage or a requirement for an unrecognized allosteric activator. Future studies using purified enzymes will be required to measure specificity constants and elucidate the structural basis for substrate discrimination by these two enzymes.
Characterization of a Ccel_2485 (ldh) and Ccel_0137 (mdh) double mutant
Residual MDH and LDH activities in the single mutants suggested the two proteins have partially overlapping substrate specificities, which prevented the complete redirection of carbon flow in these strains. The markerless mutagenesis strategy developed here enabled the rapid construction of an ldh mdh double mutant. This mutant produced negligible lactate during fermentation (Figure 3), and had no detectable MDH activity although malic enzyme activity was significantly higher (Table 1). Most importantly, ethanol production doubled during growth on minimal VM medium with cellobiose (Figure 3). These results confirm that the paralogs are responsible for almost all lactate production in C. cellulolyticum.
Effects of mdh and ldh mutations on fermentation product mixtures from complex substrates
C. cellulolyticum wild-type and mutant strains were grown in defined VM and MTC media with different carbon sources to compare the metabolic products (Figure 3 and Additional file 4). Wild-type C. cellulolyticum grew poorly in defined glucose medium, fermenting only 26% of the sugar. Cultures of the ldh mdh double mutant fermented 22% of the glucose and produced only 7% of the lactate secreted by wild-type cells. This lactate concentration was 54% lower than values measured from ldh mutant cultures, which fermented 27% of the glucose. The residual lactate in the ldh mdh mutant could be D-lactate, formed by the putative D-2-hydroxyacid dehydrogenase Ccel_3425 . Acetate concentrations in the ldh mdh mutant cultures were 50% of wild-type levels, while ethanol production was slightly higher.
In contrast, all four strains grew vigorously in defined medium with cellobiose. Wild-type, mdh and ldh mutant cells grew with doubling times of approximately 3.5 h, while the ldh mdh mutant grew with a doubling time of 6.3 h and reached a lower maximum turbidity (Figure 4). Wild-type cells fermented 54% of the cellobiose to produce primary fermentation products including 12 mM ethanol. The ldh mdh mutant fermented the same proportion of substrate but produced significantly less acetate and more than twice the concentration of ethanol (Figure 3B). This ethanol yield is approximately 44% of the maximum theoretical conversion from 5 g/l cellobiose, and it represents 53% of the carbon and 79% of the electron equivalents in the primary fermentation products.
In addition to the primary fermentation products, C. cellulolyticum secretes numerous minor products that together account for approximately 3% of the carbon and electron equivalents. To identify differences in the levels of these metabolites that could indicate changes in the carbon fluxes of mutants, we used gas chromatography-mass spectrometry (GC-MS) to analyze metabolites in the cell pellets and culture supernatants of cells grown on cellobiose (Table 2). Concentrations of extracellular pyruvate were highest in the ldh mutant and lowest in the mdh mutant; however, concentrations of pyruvate-derived amino acids were similar in the supernatants of all four strains. Malate concentrations were approximately 70% lower than wild-type in both mdh and ldh mdh mutants (0.65, 0.21 and 0.18 mM, respectively). The residual malate in the double mutant could be produced by the reductive carboxylation of pyruvate (catalyzed by the malic enzyme), or by fumarase (Ccel_2421-2422), which catalyzes the addition of water to fumarate produced by the argininosuccinate lyase enzyme (Ccel_1344). The accumulation of citrate and 2-hydroxyglutarate in the ldh mdh mutant suggests that the lack of MDH activity causes increased flux through citrate synthase, aconitase and the NADPH-producing isocitrate dehydrogenase enzymes. The 2-oxoglutarate product could be reduced using NADH to form 2-hydroxyglutarate , effectively reproducing the transhydrogenase activity of malate dehydrogenase and malic enzyme (Figure 5). Isotope incorporation experiments using Clostridium acetobutylicum suggest flux through both the oxidative and reductive branches of the incomplete citric acid cycle that cannot be explained by canonical genes in its genome [34, 35]; therefore future experiments will be required to determine the relative fluxes through both branches of the incomplete citric acid cycle in C. cellulolyticum.
Cells grown on crystalline cellulose (10 g/l) VM medium produced similar distributions of primary fermentation products compared to cellobiose cultures (Figure 3C). Wild-type C. cellulolyticum fermented 36% of the cellulose, compared to 26% in the ldh mutant and 30% in the mdh mutant. The ldh mdh mutant fermented a remarkable 50% of the cellulose to 2.7 g/l ethanol, about nine times more ethanol than wild-type cells produced. As a result of this metabolic shift, the molar ratio of ethanol to mixed acid production increased from 0.18 in wild-type cells to 15 in the ldh mdh mutant. As observed for cellobiose, this ethanol includes 79% of the electron equivalents in the primary fermentation products. Similar results were obtained from growth on MTC medium containing Avicel (Additional file 4).
In medium containing 5 g/l (33 mM) D-xylose, wild-type C. cellulolyticum fermented 52% of the carbohydrate, primarily forming lactate and acetate (Additional file 4). Cultures of both the ldh and ldh mdh mutants produced only trace amounts of lactate, but their acetate and ethanol levels were similar to wild-type cultures. Wild-type cultures grown on 5 g/l xylan fermented 40% of the carbohydrate to form 4.4 mM ethanol, 18.8 mM acetate and 1.9 mM lactate. Cultures of both the ldh and ldh mdh mutants fermented similar proportions of the xylan, but produced significantly more ethanol (8.0 and 12.3 mM, respectively, P < 0.01). Both the mdh and ldh mdh mutants produced significantly less acetate (8.9 and 10.7 mM, respectively, P < 0.05) compared to wild-type cells. Some of this acetate may have been produced by the hydrolysis of O-acetyl groups from the C-2 and C-3 positions of xylose residues in hardwood xylan .
Switchgrass contains substantial amounts of both cellulose and hemicellulose, which includes xylose, arabinose, galactose, mannose and hexuronic acids . Dilute acid pretreatment solubilizes and removes some of the hemicellulose sugars . Wild-type C. cellulolyticum cells grown on pretreated switchgrass consumed 37% of the switchgrass sugars, compared to 50% hydrolyzed by the ldh mdh mutant, measured by quantitative saccharification of the residual substrate (Figure 3F). Both strains preferentially hydrolyzed the monosaccharides from hemicellulose compared to cellulose. The product mixtures from the two cultures resembled those from xylan medium (Figure 3D, E): the wild-type cells produced significantly more acetate than the mutant (29 versus 18 mM, P < 0.0001) and the mutant produced more ethanol (28 versus 7.1 mM, P < 0.0005). The fourfold increase in ethanol production by the double mutant reflects a substantial diversion of both carbon and electrons from acetate and lactate to ethanol.
Construction of conditional mutations in Ccel_2137(pta) and Ccel_2136(ack)
Disruption of ldh and mdh eliminated lactate as a primary fermentation product and caused an unexpected decrease in acetate production. Nevertheless, acetate was still a significant fermentation product. To further shift fermentation towards ethanol, genes involved in acetate production were targeted for inactivation in the wild-type strain. However, using multiple approaches, no mutants were isolated with unconditional disruptions in either Ccel_2137 (pta) or Ccel_2136 (ack) genes. Retrotransposon vectors were constructed targeting pta and ack in the sense orientation and transformed in C. cellulolyticum (Additional file 2). From these transformations, strains were isolated that contain retrotransposon insertions into pta and ack (Additional file 5) in the presence of the plasmid-encoded LtrA, which catalyzes mRNA splicing  and enables translation to produce functional Pta and Ack proteins. These strains fermented cellobiose, primarily producing lactate and 30% less acetate than wild-type cells; however, they were unable to grow on VM medium with cellulose as the sole carbon source. Despite repeated attempts, the plasmid could not be cured from these mutants to eliminate LtrA production and mRNA splicing, suggesting that pta and ack may be essential under the conditions tested. The mRNA can only be spliced when the intron inserts in the sense orientation, so three additional plasmids were constructed to disrupt pta in the antisense orientation (Additional file 2). When these plasmids were transformed into C. cellulolyticum, hundreds of antibiotic-resistant colonies were screened and retrotransposition occurred, but no isolate showed PCR products consistent with a viable haploid or homozygous pta mutant (Additional file 5), further suggesting that acetate production is essential under the conditions tested . One possible explanation for the apparent essentiality of pta and ack in C. cellulolyticum is that H2 production could deplete the pool of reduced ferredoxin and NAD(P)H cofactors, forcing the cells to make fermentation products more oxidized than ethanol. Future experiments could determine the effects of the conditional pta and ack mutations on H2 production and electron-transfer cofactor pools, in order to infer why unconditional mutants could not be isolated using the present growth conditions.
The intron-based targeted mutagenesis method proved to be an efficient genetic tool to interrogate fermentative pathways in C. cellulolyticum. The markerless retrotransposition strategy enabled the rapid disruption of Ccel_2485 and Ccel_0137 genes in C. cellulolyticum, and their combination resulted in a strain that produced remarkably high yields of ethanol during polysaccharide and switchgrass fermentations. These experiments distinguished the primary LDH and MDH paralogs, while illustrating their partially overlapping substrate specificity in vivo. The synergy between ldh and mdh mutations all but eliminated lactate production, as expected, and significantly decreased acetate production. These cells produced substantially more ethanol than C. cellulolyticum strains that expressed the heterologous Z. mobilis pyruvate decarboxylase and alcohol dehydrogenase proteins . The C. cellulolyticum strains also secreted a number of minor fermentation products, identified by metabolomic analyses that improve our understanding of carbon flux and applicable pathways, while providing constraints for future metabolic models. Establishing complete carbon and electron balances will be essential to modeling and manipulating CBP fermentations using these organisms.
Medium and culture conditions
E. coli TOP10 cells (Invitrogen, Grand Island, NY) were used for cloning and were grown at 37°C in LB medium supplemented with 50 μg/ml kanamycin or 15 μg/ml chloramphenicol as appropriate. C. cellulolyticum H10 was cultured at 34°C anaerobically in modified VM medium  or MTC medium , with various carbon sources. For complex medium, the modified VM medium was supplemented with 2.0 g/l yeast extract. For agar plates, 1.0% (weight/volume) of Bacto agar (BD, Franklin Lakes, NJ) was added to the medium. The modified VM medium was prepared anaerobically and was supplemented with 15 μg/ml erythromycin as appropriate. A list of all C. cellulolyticum strains is presented in Additional file 6.
The E. coli-C. cellulolyticum shuttle vector used in this study was pWH199, which was modified from plasmid pAT187 as described previously . To insert the intron into pWH199 downstream of a Clostridium pasteurianum ferredoxin (Fd) promoter , a DNA fragment containing the intron and ltrA was amplified from pJIR750ai (Sigma-Aldrich, St. Louis, MO) by PCR using primers pJIR750aiXmaIF and pJIR750aiXhoIR (Additional file 7). The PCR product was digested with Xma I and Xho I enzymes and then ligated into pWH199 treated with the same enzymes, resulting in plasmid pLyc1217Er. The correct construction of pLyc1217Er was verified by sequencing. The intron integration sites were chosen by calculating all possible sites for insertions into Ccel_0137 and Ccel_2485, using an online intron design tool at http://www.clostron.com. The program predicted multiple intron insertion sites across the genes. Based on the consideration of both optimal gene inactivation and efficient insertion, a sense integration site 216 bp downstream of the start codon was chosen for Ccel_0137, and a sense integration site 516 bp downstream of the start codon was chosen for Ccel_2485. Four PCR primers for each integration design, IBS, EBS1d, EBS2 and EBSu were created by the online intron design tool. To insert the intron to the targeted genes, a 340-bp Xma I-Bsr GI intron fragment was amplified by a one-step crossover PCR, using external primers IBS and EBS1d and internal primers EBS2 and EBSu. The PCR template was a 613-bp DNA fragment amplified with primers pWH199F2 and pintronR1 using pLyc1217Er as the template. The 340-bp intron fragments were ligated into pLyc1217Er treated by Xma I and Bsr GI to form pLyc1217Er0137 and pLyc1217Er2485. All correct constructs targeting different genes were verified by sequencing using pintronR1 as the primer. Insertion vectors targeting the Ccel_2137 (pta) and Ccel_2136 (ack) genes were similarly constructed; these are described in Additional file 2.
The plasmids were transformed into wild-type C. cellulolyticum by electroporation as previously described  with modifications. Cells were grown in complex modified VM medium containing 5 g/l cellobiose and 2 g/l yeast extract for 17 to 24 hours to reach early to middle log phase (optical density at 600 nm of 0.3 to 0.5). The transformation was performed at 4°C under anaerobic conditions. The cells were harvested and washed twice with ice-cold anoxic electroporation buffer containing 270 mM sucrose, 1 mM MgCl2 and 5 mM sodium phosphate buffer, pH 7.4. The washed cells were resuspended in electroporation buffer and stored on ice. The plasmid DNA was treated with MspI methyltransferase (New England Biolabs, Ipswich, MA) for 3 h, followed by purification with the DNA Clean and Concentrator Kit (Zymo Research, Irvine, CA). For each transformation, a 50-μl cell suspension was mixed with 2.0 μg of methylated plasmid DNA. The cells were electroporated in 2-mm gap electroporation cuvettes (BTX, Holliston, MA) with a MicroPulser electroporator (Bio-Rad, Hercules, CA), inside an anaerobic chamber. The voltage was 1.5 kV, and the time constant was 5 ms. The electroporated cells were transferred to 10 ml of complex modified VM medium, and recovered for 24 h at 34°C. The cells were collected by centrifugation, and the cell pellet was spread on complex modified VM agar plates supplemented with 15 μg/ml erythromycin. The plates were incubated at 34°C anaerobically in BD GasPak plastic bags (BD, Franklin Lakes, NJ) for 5 to 7 days until single colonies appeared. To create the ldh mdh mutant, the Ccel _2485 (ldh) mutant was transformed using pLyc1217Er0137, and a strain containing an intron insertion in Ccel_0137 was isolated.
The donor plasmid pLyc1217Er in mutants was cured by culturing the cells in modified VM liquid medium without erythromycin for 4 to 5 days with daily subculture. Cells from the final subculture were spread on modified VM agar plates without erythromycin and incubated anaerobically in BD GasPak bags at 34°C. Isolated colonies were screened by colony PCR using plasmid-specific primers pWH199F2 and pintronR1. Colonies lacking PCR products were picked for further verification. The donor plasmids could not be cured from ack or pta mutants using these methods.
C. cellulolyticum genomic DNA was extracted using a Wizard Genomic DNA Purification Kit (Promega, Madison, WI). A total of 10 μg of genomic DNA was digested with Eco RI, which does not cut the inserted intron fragment, and was separated by agarose gel electrophoresis. The DNA was transferred to Hybond-N nucleic acid transfer membrane (GE Healthcare, Piscataway, NJ) and autocrosslinked using a Stratagene UV Stratalinker 1800 (Agilent, Wilmington, DE), followed by baking at 80°C for 30 minutes. The probe was amplified by PCR using dNTPs mixed with Biotin-11-dUTP (Thermo Scientific Fermentas, Glen Burnie, MD) and intron-specific primers pintronF1 and pintronR1. Hybridization and detection were performed using a North2South Chemiluminescent Hybridization and Detection kit (Thermo Scientific, Rockford, IL), following the manufacturer's instructions.
Metabolic product analysis
Wild-type and mutant strains were grown in 10 ml of modified VM cellobiose medium or 50 ml of MTC medium. Carbon substrates in these defined media were 5.0 g/l D-glucose, 5.0 g/l D-cellobiose, 5.0 or 10 g/l Avicel PH-105 crystalline cellulose (FMC BioPolymer, Philadelphia, PA), 5.0 g/l D-xylose, 5.0 g/l xylan from birch wood or 10 g/l switchgrass (dry mass; pretreated with diluted sulfuric acid). For fermentation product analyses cultures were sampled after incubation for 5 to 14 days, when fermentation was complete. The samples were filtered through 0.2 μm filters, acidified and analyzed for primary fermentation products (lactate, acetate and ethanol) using HPLC . These data from four strains were compared using the KaleidaGraph program (v. 4.1.2, Synergy Software, Reading, PA) to perform one-way analysis of variance (ANOVA) with Dunnett's multiple comparison test (0.05 significance level).
The extent of substrate conversion to primary fermentation products was calculated from the molar carbon atom ratio of primary fermentation products to substrates, assuming a stoichiometric ratio of 2 lactate, 2 acetate + 2 CO2 or 2 ethanol + 2 CO2 per glucose equivalent. For switchgrass analysis, conversion efficiency and sugar composition was determined using quantitative saccharification .
More detailed metabolic profiles were obtained by GC-MS. Supernatant and cell pellet samples for metabolomic analysis were collected from duplicate stationary phase cultures grown in defined VM medium with 5.0 g/l cellobiose. Aliquots containing 250 μL of supernatant or cell lysate and 10 μL of sorbitol (0.1% w/v) were transferred by pipette to a vial and stored at -20°C overnight. The samples were thawed and concentrated to dryness under a stream of N2. The internal sorbitol standard was added to correct for subsequent differences in derivatization efficiency and changes in sample volume during heating. Trimethylsilyl derivatives were prepared from each sample for analysis by GC-MS . The GC-MS data indicated the presence of a number of 2-hydroxyacids. To confirm whether these were induced from 2-oxoacids with reactive carbonyl groups, a test sample was additionally prepared using a double derivatization protocol that preferentially protects carbonyl groups . Briefly, 200 μL of methoxamine reagent was added to the test sample, which was heated at 30°C with stirring for 90 min. Then, 800 μL of N-methyl-N-(trimethylsilyl)trifluoroacetamide + 1% trichloromethylsilane was added and the sample was heated at 37°C for 30 minutes. The sample was analyzed by GC-MS after 2 h storage at room temperature and again after 1 day.
Enzymatic activity assays
Wild-type and mutant cells were grown in 100 ml of modified VM cellobiose medium to late log phase (optical density at 600 nm of 0.65 to 0.85). The cells were harvested, washed in buffer containing 50 mM N-[Tris(hydroxymethyl)methyl]-2-aminoethanesulfonic acid (TES)-NaOH and 10 mM MgCl2 (pH 7.2), and resuspended in 10 ml of buffer. The cells were lysed using a French pressure cell and press (Thermo Electron, Needham Heights, MA) at 10,000 psi. The lysate was centrifuged at 18,000 g for 10 min to produce cell-free extract. Total protein measurements were made with the BCA Protein Assay Kit (Thermo Pierce, Rockford, IL), using bovine serum albumin as a standard. Malate dehydrogenase and malic enzyme activities were measured at room temperature using a continuous spectrophotometric assay . Standard malate dehydrogenase reaction mixtures (1 ml) contained buffer (described above), 200 mM KCl, 125 μM NADH, 2 mM sodium oxaloacetate (prepared fresh), and cell-free extract (10 to 14 μg protein). Control reactions without oxaloacetate were used to determine the rate of non-specific NADH oxidation in these samples. Standard malic enzyme reaction mixtures (1 ml) contained buffer, 200 mM KCl, 0.2 mM NADP+, 2 mM L-malate, and cell-free extract (14 to 21 μg protein). Control reactions without malate were used to determine the rate of non-specific NADP+ reduction. Lactate dehydrogenase reaction mixtures (1 ml) contained buffer, 200 mM KCl, 1 mM pyruvate, 1 mM fructose-1,6-diphosphate, 0.2 mM NADH, and cell free extract (18 μg protein). Specific activities measured from triplicate technical replicates data from four strains were compared using one-way ANOVA with Dunnett's multiple comparison test (0.05 significance level).
Zhang YHP, Himmel ME, Mielenz JR: Outlook for cellulase improvement: screening and selection strategies. Biotechnol Adv 2006, 24: 452-481. 10.1016/j.biotechadv.2006.03.003
Li Y, Irwin DC, Wilson DB: Increased crystalline cellulose activity via combinations of amino acid changes in the family 9 catalytic domain and family 3c cellulose binding module of Thermobifida fusca Cel9A. Appl Environ Microbiol 2010, 76: 2582-2588. 10.1128/AEM.02735-09
Lynd LR, Weimer PJ, van Zyl WH, Pretorius IS: Microbial cellulose utilization: fundamentals and biotechnology. Microbiol Mol Biol Rev 2002, 66: 506-577. 10.1128/MMBR.66.3.506-577.2002
Lynd LR, Zyl WHv, McBride JE, Laser M: Consolidated bioprocessing of cellulosic biomass: an update. Curr Opin Biotechnol 2005, 16: 577-583. 10.1016/j.copbio.2005.08.009
Shaw AJ, Podkaminer KK, Desai SG, Bardsley JS, Rogers SR, Thorne PG, Hogsett DA, Lynd LR: Metabolic engineering of a thermophilic bacterium to produce ethanol at high yield. Proc Natl Acad Sci USA 2008, 105: 13769-13774. 10.1073/pnas.0801266105
Petitdemange E, Caillet F, Giallo J, Gaudin C: Clostridium cellulolyticum sp. nov., a cellulolytic, mesophilic species from decayed grass. Int J Syst Bacteriol 1984, 34: 155-159. 10.1099/00207713-34-2-155
Mohand-Oussaid O, Payot S, Guedon E, Gelhaye E, Youyou A, Petitdemange H: The extracellular xylan degradative system in Clostridium cellulolyticum cultivated on xylan: evidence for cell-free cellulosome production. J Bacteriol 1999, 181: 4035-4040.
Jennert KCB, Tardif C, Young DI, Young M: Gene transfer to Clostridium cellulolyticum ATCC 35319. Microbiology 2000, 146: 3071-3080.
Blouzard J-C, Valette O, Tardif C, de Philip P: Random mutagenesis of Clostridium cellulolyticum by using a Tn 1545 derivative. Appl Environ Microbiol 2010, 76: 4546-4549. 10.1128/AEM.02417-09
Nevoigt E: Progress in metabolic engineering of Saccharomyces cerevisiae . Microbiol Mol Biol Rev 2008, 72: 379-412. 10.1128/MMBR.00025-07
Ingram LO, Conway T, Sewell GW: Genetic engineering of ethanol production in Escherichia coli . Appl Environ Microbiol 1987, 53: 2420-2425.
Guedon E, Desvaux M, Petitdemange H: Improvement of cellulolytic properties of Clostridium cellulolyticum by metabolic engineering. Appl Environ Microbiol 2002, 68: 53-58. 10.1128/AEM.68.1.53-58.2002
Higashide W, Li Y, Yang Y, Liao JC: Metabolic engineering of Clostridium cellulolyticum for production of isobutanol from cellulose. Appl Environ Microbiol 2011, 77: 2727-2733. 10.1128/AEM.02454-10
Tripathi SA, Olson DG, Argyros DA, Miller BB, Barrett TF, Murphy DM, McCool JD, Warner AK, Rajgarhia VB, Lynd LR, Hogsett DA, Caiazza NC: Development of pyrF -based genetic system for targeted gene deletion in Clostridium thermocellum and creation of a pta mutant. Appl Environ Microbiol 2010, 76: 6591-6599. 10.1128/AEM.01484-10
Chen Y, McClane BA, Fisher DJ, Rood JI, Gupta P: Construction of an alpha toxin gene knockout mutant of Clostridium perfringens type A by use of a mobile group II intron. Appl Environ Microbiol 2005, 71: 7542-7547. 10.1128/AEM.71.11.7542-7547.2005
Heap JT, Kuehne SA, Ehsaan M, Cartman ST, Cooksley CM, Scott JC, Minton NP: The ClosTron: mutagenesis in Clostridium refined and streamlined. J Microbiol Methods 2010, 80: 49-55. 10.1016/j.mimet.2009.10.018
Tolonen AC, Chilaka AC, Church GM: Targeted gene inactivation in Clostridium phytofermentans shows that cellulose degradation requires the family 9 hydrolase Cphy3367. Mol Microbiol 2009, 74: 1300-1313. 10.1111/j.1365-2958.2009.06890.x
Lambowitz AM, Zimmerly S: Mobile group II introns. Annu Rev Genet 2004, 38: 1-35. 10.1146/annurev.genet.38.072902.091600
Frazier CL, San Filippo J, Lambowitz AM, Mills DA: Genetic manipulation of Lactococcus lactis by using targeted group II introns: generation of stable insertions without selection. Appl Environ Microbiol 2003, 69: 1121-1128. 10.1128/AEM.69.2.1121-1128.2003
Perutka J, Wang W, Goerlitz D, Lambowitz AM: Use of computer-designed group II introns to disrupt Escherichia coli DExH/D-box protein and DNA helicase genes. J Mol Biol 2004, 336: 421-439. 10.1016/j.jmb.2003.12.009
Shao L, Hu S, Yang Y, Gu Y, Chen J, Yang Y, Jiang W, Yang S: Targeted gene disruption by use of a group II intron (targetron) vector in Clostridium acetobutylicum . Cell Res 2007, 17: 963-965. 10.1038/cr.2007.91
Heap JT, Pennington OJ, Cartman ST, Carter GP, Minton NP: The ClosTron: a universal gene knock-out system for the genus Clostridium . J Microbiol Methods 2007, 70: 452-464. 10.1016/j.mimet.2007.05.021
Huang I-H, Waters M, Grau RR, Sarker MR: Disruption of the gene ( spo0A ) encoding sporulation transcription factor blocks endospore formation and enterotoxin production in enterotoxigenic Clostridium perfringens type A. FEMS Microbiol Lett 2004, 233: 233-240. 10.1111/j.1574-6968.2004.tb09487.x
Özkan M, Yılmaz Eİ, Lynd LR, Özcengiz G: Cloning and expression of the Clostridium thermocellum l-lactate dehydrogenase gene in Escherichia coli and enzyme characterization. Can J Microbiol 2004, 50: 845-851. 10.1139/w04-071
Hsieh J-Y, Hung H-C: Engineering of the cofactor specificities and isoform-specific inhibition of malic enzyme. J Biol Chem 2009, 284: 4536-4544.
Lamed R, Zeikus JG: Thermostable, ammonium-activated malic enzyme of Clostridium thermocellum . Biochim Biophys Acta 1981, 660: 251-255.
Lamed R, Zeikus JG: Ethanol production by thermophilic bacteria: relationship between fermentation product yields of and catabolic enzyme activities in Clostridium thermocellum and Thermoanaerobium brockii . J Bacteriol 1980, 144: 569-578.
Wang S, Huang H, Moll J, Thauer RK: NADP+reduction with reduced ferredoxin and NADP+reduction with NADH are coupled via an electron-bifurcating enzyme complex in Clostridium kluyveri . J Bacteriol 2010, 192: 5115-5123. 10.1128/JB.00612-10
Biegel E, Schmidt S, González JM, Müller V: Biochemistry, evolution and physiological function of the Rnf complex, a novel ion-motive electron transport complex in prokaryotes. Cell Mol Life Sci 2011, 68: 613-634. 10.1007/s00018-010-0555-8
Wilks HM, Hart KW, Feeney R, Dunn CR, Muirhead H, Chia WN, Barstow DA, Atkinson T, Clarke AR, Holbrook JJ: A specific, highly active malate dehydrogenase by redesign of a lactate dehydrogenase framework. Science 1988, 242: 1541-1544. 10.1126/science.3201242
Yin Y, Kirsch JF: Identification of functional paralog shift mutations: Conversion of Escherichia coli malate dehydrogenase to a lactate dehydrogenase. Proc Natl Acad Sci USA 2007, 104: 17353-17357. 10.1073/pnas.0708265104
Chistoserdova LV, Lidstrom ME: Purification and characterization of hydroxypyruvate reductase from the facultative methylotroph Methylobacterium extorquens AM1. J Bacteriol 1991, 173: 7228-7232.
Buckel W, Barker HA: Two pathways of glutamate fermentation by anaerobic bacteria. J Bacteriol 1974, 117: 1248-1260.
Amador-Noguez D, Feng X-J, Fan J, Roquet N, Rabitz H, Rabinowitz JD: Systems-level metabolic flux profiling elucidates a complete, bifurcated tricarboxylic acid cycle in Clostridium acetobutylicum . J Bacteriol 2010, 192: 4452-4461. 10.1128/JB.00490-10
Crown SB, Indurthi DC, Ahn WS, Choi J, Papoutsakis ET, Antoniewicz MR: Resolving the TCA cycle and pentose-phosphate pathway of Clostridium acetobutylicum ATCC 824: Isotopomer analysis, in vitro activities and expression analysis. Biotechnol J 2011, 6: 300-305. 10.1002/biot.201000282
Lindberg B, Rosell K-G, Svensson S: Positions of the O -acetyl groups in birch xylan. Sven Papperstidn 1973, 76: 30-32.
Dien BS, Jung H-JG, Vogel KP, Casler MD, Lamb JFS, Iten L, Mitchell RB, Sarath G: Chemical composition and response to dilute-acid pretreatment and enzymatic saccharification of alfalfa, reed canarygrass, and switchgrass. Biomass Bioenerg 2006, 30: 880-891. 10.1016/j.biombioe.2006.02.004
Schell DJ, Farmer J, Newman M, McMillan JD: Dilute-sulfuric acid pretreatment of corn stover in pilot-scale reactor. Appl Biochem Biotechnol 2003, 105-108: 69-85.
Yao J, Zhong J, Fang Y, Geisinger E, Novick RP, Lambowitz AM: Use of targetrons to disrupt essential and nonessential genes in Staphylococcus aureus reveals temperature sensitivity of Ll.LtrB group II intron splicing. RNA 2006, 12: 1271-1281. 10.1261/rna.68706
Raman B, Pan C, Hurst GB, Rodriguez M, McKeown CK, Lankford PK, Samatova NF, Mielenz JR: Impact of pretreated switchgrass and biomass carbohydrates on Clostridium thermocellum ATCC 27405 cellulosome composition: a quantitative proteomic analysis. PLoS ONE 2009, 4: e5271. 10.1371/journal.pone.0005271
Graves MC, Rabinowitz JC: In vivo and in vitro transcription of the Clostridium pasteurianum ferredoxin gene. Evidence for "extended" promoter elements in gram-positive organisms. J Biol Chem 1986, 261: 11409-11415.
Yang S, Tschaplinski T, Engle N, Carroll S, Martin S, Davison B, Palumbo A, Rodriguez M, Brown S: Transcriptomic and metabolomic profiling of Zymomonas mobilis during aerobic and anaerobic fermentations. BMC Genomics 2009, 10: 34. 10.1186/1471-2164-10-34
Sluiter A, Hames B, Ruiz R, Scarlata C, Sluiter J, Templeton D, Crocker D: Determination of Structural Carbohydrates and Lignin in Biomass. Laboratory Analytical Procedure (LAP) NREL/TP-510-42618. Golden, CO: National Renewable Energy Laboratory; 2008.
Welthagen W, Shellie R, Spranger J, Ristow M, Zimmermann R, Fiehn O: Comprehensive two-dimensional gas chromatography-time-of-flight mass spectrometry (GC × GC-TOF) for high resolution metabolomics: biomarker discovery on spleen tissue extracts of obese NZO compared to lean C57BL/6 mice. Metabolomics 2005, 1: 65-73. 10.1007/s11306-005-1108-2
Drevland RM, Waheed A, Graham DE: Enzymology and evolution of the pyruvate pathway to 2-oxobutyrate in Methanocaldococcus jannaschii . J Bacteriol 2007, 189: 4391-4400. 10.1128/JB.00166-07
We thank Dr Wendy Higashide for providing pWH199 plasmid, Cassie Bruno for GC-MS analysis, and Dr Jonathan Mielenz for helpful discussions. This work is supported by the BioEnergy Science Center at Oak Ridge National laboratory. The BioEnergy Science Center is a US DOE Bioenergy Research Center supported by the Office of Biological and Environmental Research in the DOE Office of Science. Oak Ridge National Laboratory is managed by UT-Battelle, LLC, for the US DOE under contract DE-AC05-00OR22725.
This manuscript has been authored by UT-Battelle, LLC, under Contract No. DE-AC05-00OR22725 with the US Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, worldwide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes.
The authors declare that they have no competing interests.
YL planned and carried out genetic experiments, analyzed all results and jointly drafted the manuscript. TT and NE planned and executed metabolomic experiments and analyzed the results. CH and MR performed metabolic product analyses and analyzed the results. DG performed enzyme activity analyses. JL, CS, AG, YY and DG planned experiments, analyzed results and helped draft the manuscript. All authors reviewed the final manuscript draft.