Genetically Engineered Saccharomyces Yeast Capable of Effective Cofermentation of Glucose and Xylose

Abstract

Xylose is one of the major fermentable sugars present in cellulosic biomass, second only to glucose. However, Saccharomyces spp., the best sugar-fermenting microorganisms, are not able to metabolize xylose. We developed recombinant plasmids that can transform Saccharomyces spp. into xylose-fermenting yeasts. These plasmids, designated pLNH31, -32, -33, and -34, are 2μm-based high-copy-number yeast-E. coli shuttle plasmids. In addition to the geneticin resistance and ampicillin resistance genes that serve as dominant selectable markers, these plasmids also contain three xylose-metabolizing genes, a xylose reductase gene, a xylitol dehydrogenase gene (both from Pichia stipitis), and a xylulokinase gene (from Saccharomyces cerevisiae). These xylose-metabolizing genes were also fused to signals controlling gene expression from S. cerevisiae glycolytic genes. Transformation of Saccharomyces sp. strain 1400 with each of these plasmids resulted in the conversion of strain 1400 from a non-xylose-metabolizing yeast to a xylose-metabolizing yeast that can effectively ferment xylose to ethanol and also effectively utilizes xylose for aerobic growth. Furthermore, the resulting recombinant yeasts also have additional extraordinary properties. For example, the synthesis of the xylose-metabolizing enzymes directed by the cloned genes in these recombinant yeasts does not require the presence of xylose for induction, nor is the synthesis repressed by the presence of glucose in the medium. These properties make the recombinant yeasts able to efficiently ferment xylose to ethanol and also able to efficiently coferment glucose and xylose present in the same medium to ethanol simultaneously.

Saccharomyces spp. are the safest and most effective microorganisms for fermenting sugars to ethanol and traditionally have been used in industry to ferment glucose (or hexose sugar)-based agricultural products to ethanol. Cellulosic biomass, which includes agriculture residues, paper wastes, wood chips, etc., is an ideal inexpensive, renewable, abundantly available source of sugars for fermentation to ethanol, particularly ethanol used as a liquid fuel for transportation. However, Saccharomyces spp. have been found to be unsuitable for fermenting sugars derived from cellulosic biomass. This is because most of the hydrolysates of cellulosic biomass contain two major fermentable sugars, glucose and xylose. Saccharomyces spp., including Saccharomyces cerevisiae, are not able to ferment xylose to ethanol or to use this pentose sugar for aerobic growth.

Even though Saccharomyces spp. are not able to metabolize xylose aerobically and anaerobically, there are other yeasts, such as Pichia stipitis and Candida shehatae, that are able to ferment xylose to ethanol and to use xylose for aerobic growth. However, these naturally occurring xylose-fermenting yeasts are not effective fermentative microorganisms, and they also have a relatively low ethanol tolerance (18). Thus, they are not suitable for large-scale commercial production of ethanol from cellulosic biomass.

The naturally occurring xylose-fermenting yeasts metabolize xylose by relying on xylose reductase (XR) to convert xylose to xylitol, on xylitol dehydrogenase (XDH) to convert xylitol to xylulose, and on xylulokinase (XK) to convert xylulose to xylulose 5-phosphate (17). The synthesis of these xylose-metabolizing enzymes in these yeasts, such as P. stipitis, requires the presence of xylose for induction and is also totally or at least partially repressed by the presence of glucose (5).

Although Saccharomyces spp. are not able to ferment xylose, they are able to ferment xylulose (14). Furthermore, they also are able to ferment xylose when xylose isomerase, a bacterial enzyme that catalyzes the conversion of xylose to xylulose, is present in the medium (10). Thus, previous studies have indicated that Saccharomyces spp. lack only the enzymes for conversion of xylose to xylulose. However, it has also been demonstrated that at least some Saccharomyces spp., if not all of them, contain very low levels of XK activity (12).

More than a decade ago, attempts were made to clone and express various bacterial xylose isomerase genes in S. cerevisiae, but these attempts failed to produce any recombinant Saccharomyces strains that were able to ferment xylose or utilize this sugar for growth. In this paper, we describe the development of a group of high-copy-number recombinant plasmids, collectively designated the pLNH plasmids (pLNH31, pLNH32, pLNH33, and pLNH34), that can transform some of the best glucose-fermenting Saccharomyces strains into effective xylose-fermenting yeasts which can also effectively utilize xylose for aerobic growth. These plasmids all contain the 2μm replication origin (2, 6), geneticin and ampicillin resistance genes (as the dominant selectable markers), and three xylose-metabolizing genes. The xylose-metabolizing genes cloned into the pLNH plasmids include the XR gene (XR), the XDH gene (XD), and the XK gene (XK) (collectively referred to below as the XYL genes). In addition, the 5′ control regions (both the transcriptional and the translational signals) of these XYL genes have been replaced by the effective 5′ control regions of the S. cerevisiae glycolytic genes. An ideal Saccharomyces strain for the fermentation of the sugars present in cellulosic biomass to ethanol not only should effectively ferment xylose to ethanol, but also should effectively coferment glucose and xylose simultaneously to ethanol. In this paper, we also demonstrate that each of the pLNH plasmids can transform a Saccharomyces strain to a recombinant yeast that can efficiently coferment glucose and xylose present in the same medium.

While our work was in progress, Kotter et al. (20), Walfridsson et al. (25), and Tantirungkij et al. (24) described the metabolic engineering of S. cerevisiae for xylose fermentation by cloning of only the XR and XD genes. The major differences between our recombinant Saccharomyces strains and the strains described by Kotter et al. (20), Walfridsson et al. (25), and Tantirungkij (24) are discussed below (see Discussion).

(Some of the results were presented at the Tenth International Symposium on Alcohol Fuel, Colorado Springs, Colo., 7 to 10 November 1993, and at the 16th Symposium on Biotechnology for Fuels and Chemicals, Gatlinburg, Tenn., 9 to 13 May 1994.)

MATERIALS AND METHODS

Strains and plasmids.

S. cerevisiae AH22, Saccharomyces sp. strain 1400 (a product of fusion between Saccharomyces diastaticus and Saccharomyces uvarum (11), and recombinant strain 1400(pLNH32) were the Saccharomyces strains used in this study. P. stipitis 5773 was used for comparison studies. Escherichia coli DH5α was used as the host during development and characterization of various recombinant E. coli plasmids or yeast-E. coli shuttle plasmids.

pUCKm10 (Fig. 1) is a high-copy-number yeast-E. coli plasmid which is identical to pUCKm8 (8) except that the two EcoRI restriction sites present in pUCKm8 are not present. pUCKm8, pUCKm10, pUC19, and pBluescript KS(−) (referred to below as pKS) were the plasmids used in this work.

FIG. 1.

Construction of high-copy-number yeast-E. coli shuttle plasmids pLNH31, pLNH32, pLNH33, and pLNH34. containing the XYL three-gene cassette KK-AR (or A*R)-KD. The XhoI DNA fragment containing KK-AR (or A*R)-KD was inserted into pUCKm10 at its SalI site.

Conversion of XR to AR or A*R, of XD to KD, and of XK to KK.

XR, the XR gene used in this study, was cloned from P. stipitis by PCR, and the 5′ control region of the XR gene was replaced by the 5′ control region of the S. cerevisiae alcohol dehydrogenase gene (ADC1) (1) as described by Chen and Ho (8). The resulting gene was designated AR or A*R. The 5′ control region of AR consisted of the original 5′ noncoding sequence of P. stipitis XR from positions −1 to −50, followed by the fragment cloned in pMA56 (1). Since it is not known what effect the nucleotide sequence (positions −1 to −50) from P. stipitis XR might have on expression of the XR gene in the Saccharomyces strains, the A*R gene was also constructed, in which the 5′ control region of XR was totally replaced by the intact 5′ control region of the ADC1 gene (positions −1 to −1500, noncoding sequence) (4).

The P. stipitis XDH gene, XD, was also amplified by PCR. The forward primer 5′-TCTAGACCACCCTAAGTCG-3′ and the reverse primer 5′-GGATCCACTATAGTCGAAG-3′ were used to amplify the intact XD gene (with its original 5′ and 3′ control sequences), and the forward primer 5′-CACACAATTAAAATGA-3′ and the reverse primer 5′-GGATCCACTATAGTCGAAG-3′ were used to amplify the promoterless XD gene (without the 5′ noncoding sequence) from the Pichia chromosomal DNA by using the previously published sequence of the Pichia XD gene (20). The amplified XD gene was first cloned and stored in pUC19. The original 5′ control region of XD was subsequently replaced by the intact 5′ control region (positions −1 to −911 of the noncoding sequence) of the S. cerevisiae pyruvate kinase gene (PYK) (7), and the resulting gene was designated KD. XD was fused to the 5′ control region of PYK by cloning both the fragment containing the 5′ noncoding sequence of PYK and the promoterless XD gene (with its original 3′ noncoding sequences) on pKS and then removing the extra nucleotides in the 5′ noncoding sequences by site-specific mutagenesis (21). The resulting plasmid was designated pKS-KD.

XK, the XK gene used in this study, was cloned from S. cerevisiae (15). The nucleotide sequence of the cloned XK gene containing 2,467 bp (bp 1 to 345 constitute the 5′ control region; bp 346 to 2118 constitute the coding sequence; and bp 2119 to 2467 constitute the 3′ noncoding sequence) has been analyzed and deposited in the EMBL data library (EMBL accession no. X61377). The promoterless XK gene can be amplified from S. cerevisiae chromosomal DNA by PCR with the following two primers: 5′-ATGTTGTGTTCAGTAAT-3′ and 5′-CAATAACCTAGCTCTTT-3′. The promoterless XK gene was also fused to the intact 5′ control region (positions −1 to −911 of the noncoding sequence) of PYK (7), and the resulting gene was designated KK. XK was fused to the 5′ control region of PYK by cloning both the fragment containing the intact 5′ noncoding sequence of PYK and the structural gene of XK with its native 3′ noncoding sequences on pKS and then removing the extra nucleotides in the 5′ noncoding sequences by site-specific mutagenesis (21). The resulting plasmid was designated pKS-KK.

Yeast transformation.

The plasmids used for transformation of the Saccharomyces strains described in this paper were all pUCKm8 or pUCKm10 derivatives which contain the geneticin resistance gene (the Km^r gene [also known as the kanamycin resistance gene]) and the ampicillin resistance gene (the Ap^r gene). The Km^r gene can be a dominant selectable marker for many yeasts, particularly Saccharomyces species (27). Transformation of the Saccharomyces strains by the pUCKm8 or pUCKm10 derivatives, such as pLNH31, pLNH32, pLNH33, or pLNH34, was carried out by electroporation largely by using the procedure described by Becker and Guarente (3), with some minor modifications. The yeast cells were grown to a density of 120 to 160 Klett units (KU) (optical density at 600 nm [OD₆₀₀], 12 to 16). A yeast cell suspension (60 μl containing 1 to 5 μl of plasmid DNA) in a 0.2-cm sterile Bio-Rad cuvette was electroporated with a Gene Pulser and a Pulse Controller (both obtained from Bio-Rad) set at 1.5 to 2.0 kV (the value varied from strain to strain), 25 μF, and 200 Ω. Authentic yeast transformants were selected as described below. The Km^r gene in a plasmid served as the primary selectable marker which rendered the host cells carrying the plasmid resistant to geneticin present in the medium. The concentration of geneticin used varied somewhat for different Saccharomyces yeasts. For example, fusion 1400 transformants were selected on YEPD (1% yeast extract, 2% peptone, 2% glucose) plates containing 40 μg of geneticin per ml, and S. cerevisiae AH22 transformants were selected on YEPD plates containing 50 μg of geneticin per ml. However, some untransformed yeast cells can be induced to become resistant to geneticin. Thus, geneticin is not ideal for direct selection of transformants, and not every geneticin-resistant colony is a true transformant. It has been reported that the Ap^r gene can be expressed in S. cerevisiae (19), and we found that it can also be expressed in fusion yeast strain 1400, as well as most other Saccharomyces strains. However, most Saccharomyces strains are resistant to ampicillin, and therefore the Ap^r gene cannot be used directly as a dominant selectable marker for selection of yeast transformants. Nevertheless, Saccharomyces strains exhibiting high levels of Ap^r gene expression produce penicillinase, and it is possible to identify yeast colonies containing the cloned Ap^r gene on special plates by the penicillinase test (9). The latter test provides a way to identify true transformants from geneticin-resistant colonies.

Preparation of seed cultures.

Yeast strains were grown on agar plates containing the proper media. For example, the untransformed Saccharomyces strains (such as strain 1400 or AH22) were grown on YEPD, and the transformed recombinant Saccharomyces strains [such as 1400(pLNH32)] were grown on YEPX (1% yeast extract, 2% peptone, 2% xylose). To prepare the seed culture for a new strain, first yeast cells from the agar plates were inoculated directly into test tubes, each containing 5 ml of medium [YEPD for strain 1400 or AH 22 and YEPX for 1400(pLNH32) and AH22(pLNH32)], and then the cells were incubated overnight in a shaker at 30°C and 200 rpm. Each of the 5-ml seed cultures was then transferred to a 300-ml Erlenmeyer flask containing 50 ml of the same sterile medium. Each flasks had a side arm (Bellco), which allowed us to directly monitor the growth of the yeast cultures with a Klett-Summerson photoelectric colorimeter. All of the seed cultures were incubated under the same conditions until they reached the late log phase (400 to 450 KU [OD₆₀₀, 80 to 90]) and then were stored at 4°C. The established seed cultures were transferred once a month by transferring 2 ml of each 1-month-old seed culture to another flask containing 50 ml of sterile medium and growing the cells to a density of 400 to 450 KU.

Enzyme assays.

To analyze the XR, XDH, XK activities, strain AH22 or fusion 1400 transformants carrying pLNH32, pUCKm10-AR, pUCKm10-A*R, pUCKm10-KD, or pUCKm10-KK were cultured in YEPD containing the proper concentration of geneticin until the density was 200 KU (20 OD₆₀₀, 20). Cells (10 ml) were harvested by centrifugation, washed with 10 ml of sterile water, resuspended in 400 μl of buffer (0.1 M sodium phosphate, pH 7.2), and chilled in a freezer for at least 30 min until they were frozen. The cells were then thawed and transferred to a 1.5-ml Eppendorf tube. About 0.3 g of acid-treated glass beads and 150 μl of 0.1 M 2-mercaptoethanol were added to the cells, which were then disrupted by vortexing. The resulting cell extracts were assayed for XR and XDH activities as described by Bolen and Detroy (5) and for XK activity as described by Shamanna and Sanderson (23). For comparison, 10 ml of P. stipitis cells cultured in YEPX until the density was 200 KU processed in the same way, and cell extracts were used to determine the XR, XDH, and XK activities by the same procedures.

Cofermentation of glucose and xylose or fermentation of xylose by yeasts.

For cofermentation studies, 2 ml of a seed culture of strain 1400 or 1400(pLNH32) was inoculated into 100 ml of YEPD nonselective medium in a 300-ml Erlenmeyer flask equipped with a side arm. The culture was incubated aerobically at 30°C in a floor shaker at 200 rpm until the density was 400 to 450 KU, and there was no more than a 20-KU difference between the cultures used in the comparison studies. Twenty milliliters of 50% glucose and 10 ml of 50% xylose were then added to each culture, making the final glucose and xylose concentrations approximately 8 and 4%, respectively. After thorough mixing, 1 ml of the supernatant broth (containing some cells) was removed from each flask to serve as the zero time sample. The flasks were then sealed with two layers of Saran Wrap to allow fermentation to occur anaerobically and were incubated in a shaker under identical conditions. Samples (1 ml) of the fermentation broth were removed at intervals and used for analyses of substrates and fermentation products, including glucose, xylose, ethanol, xylitol, and glycerol. To analyze fermentation of xylose by 1400(pLNH32), the cells were cultured in YEPX.

Analysis of fermentation products.

A high-performance liquid chromatograph (Hitachi Ltd., Tokyo, Japan) equipped with an RI detector was used to analyze the concentrations of glucose, xylose, xylitol, ethanol, and glycerol. A Bio-Rad type HPX-87H ion-exclusion column was used. The mobile phase was 0.005 M H₂SO₄ at a flow rate of 0.4 ml/min.

RESULTS

Expression of AR, A*R, KD, or KK in S. cerevisiae transformants cultured in medium containing glucose.

Previously, we described the cloning of a BamHI fragment containing AR or A*R into the yeast-E. coli shuttle plasmid pUCKm8. The resulting plasmids were designated pLXR10 and pLXR13, respectively (8). These plasmids were used to transform S. cerevisiae AH22. The resulting transformants carrying pLXR10 or pLXR13, cultured in rich medium containing glucose as the sole carbon source (YEPD containing 50 μg of geneticin per ml), were found to be able to synthesize high levels of XR, and the enzyme activity was similar to that present in wild-type P. stipitis cultured in YEPX (Table 1). In control experiments, untransformed parent S. cerevisiae cells and P. stipitis cells cultured in YEPD were found to contain practically no detectable XR activity.

TABLE 1.

In vitro XR, XDH, and XK enzyme activities

Strain	Sp act (nmol/mg of protein−1 min−1)
	XR (NADPH)a	XDH (NAD)b	XKc
1400(pLNH32)	411	839	100
1400	54	22	NDd
P. stipitis 5773	417	410	35.2

^aThe specific activity of XR is expressed as the amount of NADPH oxidized by the enzyme. 

^bThe specific activity of XDH is expressed as the amount of NAD reduced by the enzyme. 

^cThe specific activity of XK is expressed as the amount of xylulose phosphorylated by the enzyme. 

^dND, not detectable. 

The BamHI-XhoI fragment containing the cloned KD gene was isolated from plasmid pKS-KD (Fig. 2) and subcloned into pUCKm10 (Fig. 1) at its BamHI and SalI sites. The resulting plasmid, pUCKm10-KD, was used to transform S. cerevisiae AH22. When the transformants were cultured in YEPD supplemented with 50 μg of geneticin per ml, they were able to synthesize high levels of XDH, and the enzyme activity was comparable to the activity present in 1400(pLNH 32) cultured in YEPD (Table 1). Untransformed AH22 did not have any detectable XDH activity.

FIG. 2.

Construction of E. coli plasmids containing the XYL gene cassette KK-AR (or A*R)-KD. KK, S. cerevisiae XK gene fused to the promoter of the S. cerevisiae pyruvate kinase gene; AR (or A*R), P. stipitis XR gene fused to the promoter of the S. cerevisiae alcohol dehydrogenase gene; KD, P. stipitis XDH gene fused to the promoter of the S. cerevisiae pyruvate kinase gene. The double dagger (‡) indicates that the XhoI site was regenerated after ligation, and the asterisks (∗) indicate an intact ADC1 promoter.

The BamHI-XhoI fragment containing the cloned KK gene was also isolated from pKS-KK (Fig. 2) and subcloned into pUCKm10. The resulting plasmid, pUCKm10-KK, was used to transform S. cerevisiae AH22. The resulting transformants were able to synthesize high levels of XK, like 1400(pLNH32) (Table 1), when they were cultured in YEPD supplemented with the proper levels of geneticin. Untransformed AH22 did not have any detectable XK activity when it was cultured in YEPD.

S. cerevisiae AH22 was used as the host to test the expression of these cloned genes in a Saccharomyces strain because AH22 can be very effectively transformed by electroporation under the conditions described in this paper. For example, more than 70% of geneticin-resistant AH22 colonies were true transformants.

Construction of E. coli plasmids containing the XYL gene cassette, KK-AR (or A*R)-KD.

Our strategy for developing the desired recombinant yeast-E. coli shuttle plasmid containing the XYL genes was first to subclone AR (or A*R), KD, and KK into pKS in such a way that the three genes formed a cassette which could be excised from the pKS plasmid together as a fragment by a single restriction endonuclease digestion. As a result, the three-gene cassette could be easily inserted (by a single ligation and transformation) into various yeast-E. coli shuttle plasmids for simultaneous expression of all three genes in yeast cells. Figure 2 is a schematic diagram showing the construction of four such recombinant plasmids, pKS-KK-A*R-KD-1 and -2 and pKS-KK-AR-KD-3 and -4 containing the XYL gene cassette (referred to below as pKRD-1, -2, -3, and -4, respectively). The XYL gene cassette can be removed from these plasmids by digestion with restriction endonuclease XhoI. The four plasmids differed only in AR gene orientation and in the promoter of the AR gene, as described in Materials and Methods.

Construction of high-copy-number yeast-E. coli shuttle plasmid containing the three-gene cassette KK-AR-KD or KK-A*R-KD.

As shown in Fig. 1, pUCKm10 is a 2μm replicon-based yeast-E. coli shuttle plasmid and contains the Km^r and Ap^r genes as selectable markers. The pUCKm10 plasmid could be introduced into most of the Saccharomyces strains that we tested by using Km^r and Ap^r for selection of the putative transformants, as described in Materials and Methods. The XhoI fragments containing the three-gene cassette were isolated from pKRD1, -2, -3, and -4 and inserted into the SalI site of pUCKm10; this resulted in eight possible plasmids, which were collectively designated the pLNH plasmids. Four such plasmids, pLNH31, -32, -33, and -34 (containing the XYL gene cassettes from pKRD1, -2, -3, and -4, respectively) were characterized and have the general structures shown in Fig. 1.

Xylose as the ideal selective pressure in rich medium for Saccharomyces transformants carrying the pLNH plasmids.

Although true pLNH transformants of yeast strain 1400 can be obtained by selecting colonies that not only grow on plates containing geneticin but also are positive for the penicillinase test (see Materials and Methods), geneticin is not an ideal selective pressure for culturing and maintaining strain 1400 transformants carrying the pLNH plasmids for at least two reasons. One reason is that the transformed yeast cells can all be gradually induced to become resistant to geneticin without relying on the action of the Km^r gene product. This allows plasmid-free yeast cells to grow in the presence of geneticin. Therefore, geneticin can lose its effectiveness as the agent which exerts selection pressure to maintain the pLNH plasmids in the strain 1400 transformants. The second reason is that geneticin is too costly and too great an environmental hazard to be used to maintain cultures for large-scale ethanol production. It is known that in minimal medium all microorganisms require a carbon source, such as glucose or xylose, for growth. Thus, xylose could be used as the selection pressure to maintain the transformants in minimal medium supplemented with xylose as the sole carbon source. However, yeast-based industrial ethanol fermentation is usually carried out in an inexpensive rich medium, such as corn steep liquor, and most microorganisms should be able to grow in rich medium without a carbon source. Nevertheless, we discovered that most Saccharomyces strains (more than 10 Saccharomyces strains were tested, and there were no exceptions) do require a carbon source for growth even in rich medium, such as YEP (1% yeast extract, 2% peptone), as shown in Fig. 3. This seems to be a special characteristic of a few genera of yeasts but not all yeasts. For example, it was found that P. stipitis and C. shehatae do not require a carbon source to grow in rich medium, such as YEP. The latter discovery made it possible to use xylose as the ideal selection pressure to culture and maintain the pLNH transformants in both rich and minimal media. Furthermore, it also provided an alternative mechanism for selection of true transformants carrying the pLNH plasmids in case certain yeasts could not use the Km^r and Ap^r genes as the selection markers.

FIG. 3.

Growth of S. cerevisiae in rich medium with or without a carbon source (sugar). (A) S. cerevisiae AH22 cultured in YEP or YEPD. (B) S. cerevisiae AH22 cultured in YEP or YEPXylu (1% yeast extract, 2% peptone, 2% xylulose). Symbols: ○, YEP; •, YEPD; ▴, YEPXylu.

Transformation of Saccharomyces sp. strain 1400 with the pLNH plasmids and the ability of the resulting transformants to utilize xylose for growth.

Saccharomyces sp. strain 1400 (11), which can tolerate high temperatures and high ethanol concentrations, should be an ideal strain for producing fuel ethanol. Thus, this yeast strain was used as the host to receive the pLNH plasmids. Transformation of Saccharomyces sp. strain 1400 with the pLNH plasmids and selection of pLNH transformants of strain 1400 were carried out as described in Materials and Methods. Each pLNH plasmid was able to transform strain 1400 to a xylose-utilizing strain. The 1400 recombinant strains were initially identified by their ability to grow on YEPD plates containing geneticin (40 μg per ml) and by their ability to form halos on ampicillin test plates (9). Each one was verified by its ability to transfer the specific plasmid from the transformant back into E. coli (26). Most importantly, these transformants were able to grow in YEPX. Table 2 clearly shows that a pLNH transformant of strain 1400, such as 1400(pLNH32), can grow in rich medium containing xylose as the sole carbon source (YEPX) but the untransformed parent strain 1400 cannot. These results further demonstrate the usefulness of xylose or a carbon source as an effective selective agent for selection and maintenance of recombinant Saccharomyces strains.

TABLE 2.

Growth of Saccharomyces recombinant and parent yeast strains in media containing glucose and xylose

Strain	Medium	Density (KU) ata:
		Zero time	2 h	4 h	6 h	8 h	11 h	15 h
1400(pLNH32)	YEPD	21	22	44	120	289	455	472
	YEPX	18	21	37	68	122	450	550
1400	YEPD	17	17	39	120	290	433	460
	YEPX	54	66	72	75	75	79	79

^aDensity was determined with a Klett-Summerson photoelectric colorimeter; 1 KU was equivalent to 0.1 OD₆₀₀ unit at densities between 20 and 200 KU and to 0.15 to 0.2 OD₆₀₀ unit at densities of more than 300 KU. 

Comparison of the abilities of 1400(pLNH32) and parent strain 1400 to cofermenting a mixture of glucose and xylose.

Because pLNH32 and related plasmids are 2μm-based high-copy-number plasmids, the transformants were relatively stable and could be cultured for four or five generations in rich, nonselective medium, such as YEPD, without losing much of their xylose-fermenting ability. This allows the use of these recombinant yeast strains to ferment xylose or to coferment glucose and xylose in YEPD. To compare the abilities of 1400(pLNH32) and related transformants containing other pLNH plasmids to coferment glucose and xylose with the ability of the parent strain 1400 to coferment glucose and xylose, these strains were cultured in YEPD (50 ml) until the density was 400 to 450 KU (initial density, 20 KU [OD₆₀₀, 2]), and then 10 ml of 50% glucose and 5 ml of 50% xylose were added to each culture to start the fermentation. The results obtained with 1400(pLNH32) and parent strain 1400 are shown in Fig. 4. The results obtained with strain 1400 transformants containing other pLNH plasmids were similar to the results obtained with 1400(pLNH32). These results demonstrated that the genetically engineered strain 1400(pLNH32) and related transformants could effectively coferment most of the 8% glucose and 4% xylose to ethanol in 48 h. In contrast, parent strain 1400 could ferment glucose to ethanol but could not ferment xylose.

FIG. 4.

Comparison of the abilities of recombinant Saccharomyces sp. strain 1400(pLNH32) and parent strain 1400 to coferment glucose and xylose. (A) Strain 1400(pLNH32). (B) Strain 1400. Utilization of glucose and xylose and production of ethanol, xylitol, and glycerol were determined. Symbols: ▪, glucose; •, xylose; ▴, ethanol; □, xylitol; ▵, glycerol.

XR, XDH, and XK activities in Saccharomyces transformant 1400(pLNH32).

To further establish that the ability of the strain 1400 transformants containing the pLNH plasmids to effectively metabolize xylose was due to the presence of the cloned XYL genes, the strain 1400 transformant containing plasmid pLNH32 was cultured in YEPD to a density of 200 KU (OD₆₀₀, 20), and the resulting cells were used to determine XR, XDH, and XK activities, which were compared with the activities present in parent strain 1400 and P. stipitis, as shown in Table 1.

DISCUSSION

This paper describes the strategies used to develop recombinant Saccharomyces strains that can effectively ferment xylose and, in particular, can coferment glucose and xylose present in the same medium. Using Saccharomyces sp. strain 1400 as an example, we successfully demonstrated that by cloning XR, XD, and XK, each of which was fused to an efficient glycolytic promoter, a superior glucose-fermenting Saccharomyces strain can be transformed to a recombinant yeast strain that is able to effectively metabolize xylose aerobically for growth and anaerobically for the production of ethanol as the overwhelmingly major product. Furthermore, the resulting recombinant xylose-fermenting Saccharomyces strains can also effectively coferment glucose and xylose present in the same medium. This unique property should make the recombinant Saccharomyces strains particularly effective in large-scale production of ethanol from fermenting mixed sugars present in cellulosic biomass hydrolysates.

Although only the results obtained with strain 1400(pLNH32) are presented here, the transformants containing pLNH31, pLNH33, or pLNH34 do not differ significantly in their ability to coferment glucose and xylose or to utilize xylose for growth. The presence of either AR or A*R in the transformants also does not significantly affect their ability to ferment xylose or utilize it for growth.

Our recombinant Saccharomyces strains can effectively ferment xylose in the presence of glucose, as shown in Fig. 4. However, they do not depend on the presence of glucose to ferment xylose. They can effectively ferment xylose in the absence of glucose when they are cultured in xylose medium, as shown in Fig. 5.

FIG. 5.

Fermentation of xylose by recombinant Saccharomyces sp. strain 1400(pLNH32) cultured in the xylose-containing medium YEPX. Utilization of xylose and production of ethanol, xylitol, and glycerol were determined. Symbols: •, xylose; ▴, ethanol; □, xylitol; ▵, glycerol.

As mentioned above, while our work was in progress, Kotter et al. (20), Walfridsson et al. (25), and Tantirungkij et al. (24) described the development of their recombinant S. cerevisiae strains, which was accomplished by cloning only the P. stipitis XR gene and XDH gene. However the recombinant yeast strains of these groups ferment xylose extremely slowly and produce little ethanol. Furthermore, the major fermentation product produced by these recombinant yeast strains is not ethanol but xylitol. Most importantly, there are fundamental differences between our strategy and the strategies used by the other three groups to develop recombinant yeast strains. For example, as mentioned above, our recombinant Saccharomyces strains not only contain the cloned XR gene and the XDH gene but also contain the cloned XK gene. As a result, our recombinant yeast strains can effectively ferment high concentrations of xylose almost completely to ethanol, and very little xylitol is produced as a by-product (Fig. 4A and 5).

Our strategy to fuse the cloned XR, XD, and XK genes to highly effective glycolytic promoters (or expression signals) made our recombinant yeast strains better than the recombinant yeast strains developed by others, as well as naturally occurring microorganisms in fermenting substrates containing both glucose and xylose. The other groups did not include this strategy in their designs. As a result, the expression of the cloned XR, XD, and XK genes in our recombinant yeast strains does not require the presence of xylose for induction, and also the expression of the cloned genes is not repressed by glucose in the cultural medium. This allowed our recombinant Saccharomyces strains to effectively coferment glucose and xylose without a lag period, as shown in Fig. 4A. However, for the conversion of cellulosic biomass to ethanol, the fact that our yeast strains do not require xylose for induction and the fact that they are able to maintain expression of the xylose-metabolizing genes in the presence of glucose provide a much greater benefit than metabolizing xylose without a lag period. The naturally occurring xylose-fermenting yeasts, such as P. stipitis and C. shehatae, which ferment xylose as effectively as our recombinant xylose-fermenting Saccharomyces strains but do not have the properties mentioned above, are not able to ferment xylose at all when glucose is present in the medium, even though they effectively ferment xylose in the absence of glucose. Data comparing the ability of our recombinant yeast strains to ferment xylose and to coferment glucose and xylose with the ability of the naturally occurring xylose-fermenting yeasts P. stipitis and C. shehatae to ferment xylose and to coferment glucose and xylose will be published elsewhere.

The fact that we developed a set of broad-host-range pLNH plasmids by using the Km^r and Ap^r genes as the primary selection markers also contributed greatly to the development of effective recombinant xylose-fermenting Saccharomyces strains, such as 1400(pLNH32). The development of such plasmids was intended to provide a means by which wild-type yeast strains, particularly diploid or polyploid industrial strains (such as strain 1400) that may have superior properties desirable for ethanol fermentation, can easily be transformed by these vectors. Furthermore, since there are other factors besides the enzymes encoded by the cloned XYL genes that may seriously affect the effectiveness of yeast xylose fermentation, we do not expect that all Saccharomyces strains that receive the same cloned XYL genes will ferment xylose with the same efficiency. Thus, the pLNH plasmids can also be excellent tools for screening new Saccharomyces strains that may be superior for cofermenting glucose and xylose. Nevertheless, since we did not have to screen many yeast strains to obtain strain 1400, which is an effective host for the expression of the XYL genes, there may be other Saccharomyces strains that are as effective as or more effective than strain 1400 as hosts for expression of the XYL genes for efficient fermentation of xylose to ethanol.

It is desirable to design a broad-host-range selective mechanism that requires only inexpensive and nontoxic chemicals as the selective pressure to select and maintain Saccharomyces transformants, and our discovery that Saccharomyces strains require a carbon source to grow even in rich medium provided an answer. This discovery provided an ideal and effective mechanism that could be used to screen and maintain true transformants containing the cloned XYL genes. Furthermore, it also provided a rapid and effective method for differentiating xylose-fermenting Saccharomyces transformants from other xylose-metabolizing microorganisms, including the naturally occurring xylose-fermenting yeasts, such as P. stipitis and C. shehatae. This is because there are very few microorganisms that require the presence of a carbon source in rich medium to grow.

The fact that xylose can serve as a selective agent to maintain the pLNH plasmids even in rich medium, such as YEPX, coupled with the fact that the pLNH plasmids are stable, high-copy-number plasmids in Saccharomyces strains, allows the pLNH transformants of Saccharomyces strains to effectively ferment xylose to ethanol in the absence of selection for at least four or five generations, as shown in Fig. 4. This makes it possible for industry to use these recombinant yeast strains, such as 1400(pLNH32), to carry out large-scale fermentation of the glucose and xylose present in cellulosic biomass to ethanol in the absence of selection. To achieve this, the only requirements are that the process is a batch fermentation process and that the yeast cells are propagated in the early stages in medium containing xylose as the major or sole carbon source.

Several groups in the United States and Canada are using our xylose-fermenting recombinant strain 1400 yeasts to perform various studies, and they have all verified what we report here. One group that used 1400(pLNH32) to ferment corn fiber sugars to ethanol has published their results (22).

Nevertheless, additional improved xylose-fermenting Saccharomyces strains can still be developed. For example, even though the Saccharomyces transformants carrying the pLNH plasmids are very stable and can be maintained by an ideal selection mechanism (in which xylose is used as the selection pressure), an ideal transformant should be completely stable and not require the use of selection pressure at any stage of growth or fermentation.

Recently, we have been very successful in developing stable derivatives of xylose-fermenting recombinant yeast strain 1400, designated 1400(LNH-ST) strains (or LNH-ST strains). The latter Saccharomyces strains are completely stable and do not require any selection pressure to maintain their ability to ferment xylose to ethanol or their ability to utilize this sugar for growth. Furthermore, these stable xylose-fermenting recombinant yeast strains also ferment xylose to ethanol slightly more efficiently than 1400(pLNH32) or any other pLNH transformant of strain 1400. The stable strains were developed by integrating multiple copies of the XYL genes into the yeast chromosome. Brief descriptions of the ability of these stable xylose-fermenting Saccharomyces strains to ferment mixed sugars in the absence of any selection pressure have been presented previously (16). The details concerning the development and characterization of such stable xylose-fermenting Saccharomyces strains will be described elsewhere.