ABSTRACT
Synthesis of chitin de novo from glucose involves a linear pathway in Saccharomyces cerevisiae. Several of the pathway genes, including GNA1, are essential. Genes for chitin catabolism are absent in S. cerevisiae. Therefore, S. cerevisiae cannot use chitin as a carbon source. Chitin is the second most abundant polysaccharide after cellulose and consists of N-acetylglucosamine (GlcNAc) moieties. Here, we have generated S. cerevisiae strains that are able to use GlcNAc as a carbon source by expressing four Candida albicans genes (NAG3 or its NAG4 paralog, NAG5, NAG2, and NAG1) encoding a GlcNAc permease, a GlcNAc kinase, a GlcNAc-6-phosphate deacetylase, and a glucosamine-6-phosphate deaminase, respectively. Expression of NAG3 and NAG5 or NAG4 and NAG5 in S. cerevisiae resulted in strains in which the otherwise-essential ScGNA1 could be deleted. These strains required the presence of GlcNAc in the medium, indicating that uptake of GlcNAc and its phosphorylation were achieved. Expression of all four NAG genes produced strains that could use GlcNAc as the sole carbon source for growth. Utilization of a GlcNAc catabolic pathway for bioethanol production using these strains was tested. However, fermentation was slow and yielded only minor amounts of ethanol (approximately 3.0 g/liter), suggesting that fructose-6-phosphate produced from GlcNAc under these conditions is largely consumed to maintain cellular functions and promote growth. Our results present the first step toward tapping a novel, renewable carbon source for biofuel production.
Saccharomyces cerevisiae ferments certain sugars very efficiently into ethanol, even under aerobic conditions. Glucose, fructose, and mannose are fermented via the Embden-Meyerhof pathway of glycolysis, and galactose requires the Leloir pathway. Certain yeast strains, particularly brewer's yeast strains, can ferment maltose, maltotriose, and melobiose as well. During recent years considerable efforts have been focused on generating S. cerevisiae strains that can also convert other sugars derived from plant biomass into ethanol (reviewed in reference 31).
Chitin is a renewable resource that is a copolymer of β-(1-4)-N-acetyl-d-glucosamine and N-glucosamine residues. When more than half of the residues are N-acetylglucosamine (GlcNAc) the polymer is termed chitin; otherwise, it is called chitosan (16). Chitin is the second most abundant polysaccharide in nature after cellulose, and the annual biotic production of chitin is in the range of several gigatons (12, 18, 29). Chitin is found in the exoskeleton of crustaceans, such as lobsters, shrimp, and crabs. The shells of these animals generate as much as 75% of the weight of the waste, and roughly half of that constitutes chitin. With an annual harvest of more than 5 megatons of shrimp (http://www.fao.org/ ), this represents a large resource.
Several industrial and medical applications rely on the use of chitin and its derivatives as a cationic polysaccharide (11). The absorption capacity of chitosan can be used to remove mercury from wastewater (20). Chitosan nanofiber scaffolds can be used to promote a hepatocyte-scaffold interaction in bioartificial livers (5). Furthermore, several new drug delivery systems are based on chitosan (22).
Chitin is also found in bacterial and fungal cell walls. In S. cerevisiae, chitin reinforces the bud site and marks sites of cell division (33).
Chitin extraction on an industrial scale uses a multistep procedure that includes grinding of the shells, deproteination and demineralization using harsh chemical treatments with NaOH and HCl, respectively, and subsequent bleaching with H2O2. This results in a crude chitin preparation that can be tailored using depolymerization enzymes, such as chitinases, chitodextrinases, or lysozyme (2, 9, 17, 34). To avoid chemical and enzymatic treatments for depolymerization and bleaching, ozone treatment has been suggested (26).
Several hyperthermophilic marine archaea and bacteria as well as various fungi can utilize GlcNAc as the sole carbon source, and chitinases produced by these organisms may have antifungal activity against plant pathogens (27).
S. cerevisiae produces chitin de novo from glucose (Fig. 1). The biosynthetic pathway of chitin is essential in S. cerevisiae, since a central metabolite, UDP-GlcNAc, is also used for providing GlcNAc moieties for glycosylation and the production of GPI anchors, which then attach to target proteins that in S. cerevisiae are a major component of the cell wall (13, 23). Therefore, several of the biosynthetic enzymes involved in the chitin synthesis pathway are encoded by essential genes in S. cerevisiae.
Chitin metabolism and GlcNAc catabolism in S. cerevisiae and C. albicans. The linear pathway for chitin synthesis in S. cerevisiae and C. albicans, starting from glucose, is shown. The involved proteins and cofactors are also shown. GFA1, GNA1, AGM1, and UAP1 are essential in S. cerevisiae. The yeast chitinase Cts1 is involved in mother-daughter cell separation but does not process chitin for uptake.
Candida albicans, a human fungal pathogen, is a dimorphic fungus that can switch between yeast and filamentous growth modes depending on environmental stimuli. In vitro specific conditions can be used to induce filament formation and mycelium development in C. albicans cells. This yeast-to-hypha transition requires a switch to 37°C and the presence of an external inducer, such as serum, amino acids, or GlcNAc (35). Interestingly, C. albicans harbors the genes required for a GlcNAc catabolic pathway in a gene cluster comprising six genes. This includes genes encoding proteins potentially required for GlcNAc uptake (NAG3 and NAG4) and GlcNAc phosphorylation (NAG5), deacetylation (NAG2), and deamination (NAG1), respectively. Transcription of some of these genes is induced upon GlcNAc availability, and C. albicans can utilize GlcNAc as a sole carbon source (reference 3 and references therein).
In this paper we describe the heterologous expression of C. albicans NAG genes in S. cerevisiae and study their effect on making a novel carbon source available for S. cerevisiae.
MATERIALS AND METHODS
Strains and media. C. albicans BWP17 and S. cerevisiae BY strains were used (Table 1). Cells were grown either in rich medium (yeast extract-peptone-dextrose [YPD]; 2% peptone, 2% glucose, 1% yeast extract) or in minimal medium at 30°C: complete synthetic medium (CSM); 2% glucose, 6.7 g/liter yeast nitrogen base [YNB] with ammonium sulfate and without amino acids, 0.79 g/liter CSM) or synthetic defined medium (2% glucose, 6.7 g/liter YNB, with ammonium sulfate and without amino acids). Minimal media were supplemented with the strain-specific requirements for amino acids.
Species and strains used in this study
Regulation of gene expression was done using the MET3 promoter. Activation of MET3p-controlled gene expression was achieved by growing cells in media lacking methionine and cysteine as described previously (7). Minimal media for GlcNAc utilization contained 6.7 g/liter YNB with ammonium sulfate without amino acids, 0.1% to 10% GlcNAc and 2 g/liter asparagine as nitrogen sources, and glutamate and other amino acids to complement auxotrophies. Addition of G418 (200 μg/ml) was used for plasmid selection in S. cerevisiae. For plasmid construction and amplification Escherichia coli DH5α served as a host.
Disruption of ScGNA1.Since ScGNA1 encodes an essential gene, the EUROSCARF collection does not contain a deletion strain for this gene. We used PCR-based gene targeting to disrupt GNA1 in a strain that harbors the NAG3 or NAG4 and the NAG5 genes. Transformants were generated using NATMX2 as a selectable marker gene, which confers resistance to the antibiotic nourseothricin (cloNAT; used at 200 μg/ml). Correct integration was verified by PCR. Primers were obtained from biomers.net and are listed in Table 2 .
Oligonucleotide primers used in this study
Molecular biology techniques.Standard molecular tools were used (24). Yeast transformation was done using the lithium acetate method (10). Bacterial transformation was done by electroporation. Plasmids were purified according to the manufacturers' instructions (Promega or Qiagen). DNA sequencing was performed by MWG-Biotech AG (Ebersberg, Germany). PCRs were performed using standard programs, e.g., 95°C for 5 min, 35 cycles of 95°C for 30 s, 52°C for 30 s, and 72°C for 2 min, and 72°C for 5 min. Taq polymerase was used. All enzymes were supplied by New England Biolabs.
Plasmid constructs.A divergent promoter construct harboring the Ashbya gossypii MET3 promoter and the Candida maltosa LEU2 promoter was constructed using pFA vector plasmids (25). To this end pFA-NATMX3-AgMET3p was cleaved with SmaI/BglII to eliminate NATMX3. pFA-CmLEU2 was cleaved with BamHI/PmeI to yield CmLEU2, which was inserted into the SmaI/BglII sites of the former plasmid.
The C. albicans NAG3 and NAG4 open reading frames (ORFs) were amplified with short terminator regions from genomic DNA and cloned as 1,767-bp and 1,906-bp fragments, respectively, into pRS-ScMET3prom or pRS-AgTEFprom (7) to generate plasmids 804 (pRS-ScMET3p-NAG3; 7,608 bp) and 805 (pRS-ScMET3p-NAG4; 7,747 bp) and plasmids C258 (pRS-AgTEFp-NAG3; 7,448 bp) and C259 (pRSAgTEFp-NAG4; 7,587 bp), respectively. Plasmid 805 contains only the XhoI site, and plasmid C259 contains an EcoRI site instead (see Fig. 2, below).
Plasmid constructs for expression of C. albicans NAG genes in S. cerevisiae. We used the pRS vector series as a plasmid backbone. Thus, all plasmids also harbored the CEN6/ARSH4 cassette for replication and plasmid distribution in S. cerevisiae. Expression of the NAG3 and NAG4 genes was achieved by using either the A. gossypii TEF or MET3 promoters as indicated. For construction details see Materials and Methods.
The C. albicans NAG5 ORF was amplified from genomic DNA and cloned via yeast in vivo recombination into pRS415-kanMX. In this way the kanR ORF was replaced with the CaNAG5 ORF, placing CaNAG5 under the control of the A. gossypii TEF promoter-generating plasmid, plasmid 363 (8,097 bp).
The C. albicans NAG1 and NAG2 ORFs with their terminators were amplified from genomic DNA and cloned as 916-bp BglII/XhoI and 1,392-bp BamHI/SacII fragments into pRS415. In vivo recombination via yeast was used to fuse the NAG genes with the divergent AgMET3prom-CmLEU2prom. This placed NAG1 under the control of the Candida maltosa LEU2 promoter and NAG2 under the control of the A. gossypii MET3 promoter. This cassette was then cloned next to CaNAG5 cleaved with SmaI/XhoI as a PmeI/XhoI fragment, generating plasmid C236 (pRS415-NAG1-NAG2-NAG5; 10,846 bp).
Fermentations.Fermentations were done in bench-scale tall-tube reactors containing 90 ml minimal medium 1.7 g/liter YNB without (NH4)2SO4, 2 g/liter asparagine, glutamate, histidine, and uracil, 0.5 g/liter MgSO4, 0.5 g/liter KH2PO4, 1 g/liter K2HPO4 containing 10% GlcNAc, with or without glucose (5 g/liter). G418 was added to ensure plasmid maintenance when necessary. An enzymatic ethanol assay (K-ETOH; Megazyme) was used to measure the ethanol content of the cultures. Ethanol was measured regularly over 2 weeks according to the manufacturer's protocol.
RESULTS
GlcNAc metabolism in S. cerevisiae and C. albicans.GlcNAc serves several purposes in fungal cells. It is used as UDP-N-acetyl-d-glucosamine in providing GlcNAc molecules for N-linked glycosylation, in GPI-anchored proteins and, of course, as a building block for chitin. In S. cerevisiae chitin is a component of the cell wall and is found enriched at the bud neck, forming the bud scars. UDP-GlcNAc is synthesized by way of four consecutive steps from fructose-6-phosphate. The four corresponding genes that take part in this process, GFA1, GNA1, PCM1/AGM1, and QRI1/UAP1, are all essential and conserved (Fig. 1). S. cerevisiae lacks genes for chitin catabolism. However, they can be found in various other organisms, including C. albicans. S. cerevisiae conceptually lacks four enzymatic activities for the degradation of chitin, which are a GlcNAc permease/transporter, a GlcNAc kinase, a deacetylase, and a deaminase (Fig. 1). Interestingly, genes involved in chitin catabolism are clustered within a 10-kb fragment in C. albicans. Although several of the genes in that cluster have been characterized in C. albicans (see reference 3), the functions of the putative NAG3 and NAG4 transporters have not been analyzed in detail. Nevertheless, we hypothesized that within this cluster all genes necessary for the utilization of GlcNAc are present, and we chose to express these genes in S. cerevisiae.
C. albicans NAG3 or -4 and NAG5 can complement the loss of S. cerevisiae GNA1.The CaNAG3 and CaNAG4 open reading frames coding for putative GlcNAc transporters were placed under the control of the ScMET3 promoter. The CaNAG5 ORF, encoding a GlcNAc kinase, was put under the control of the Ashbya gossypii TEF promoter (Fig. 2). This was done to ensure the expression of the ORFs in S. cerevisiae, since the C. albicans genes are either constitutively expressed or induced upon growth in GlcNAc-containing medium. The corresponding Nag3p, Nag4p, and Nag5p should provide the S. cerevisiae cells with GlcNAc-6-phosphate, which is also the product of Gna1p (Fig. 1). Since we were unable to delete GNA1 in a haploid BY4742 background, as expected, we first transformed the NAG3- or 4- and the NAG5-containing plasmids into BY4742 and subsequently went on to delete ScGNA1 in this strain. We were able to delete ScGNA1 in strains carrying either NAG3 or NAG4 and NAG5. However, the gna1 deletion strain became dependent on the addition of GlcNAc to the growth medium and failed to grow on glucose-based media (Fig. 3). These results suggested that both Nag3p and Nag4p are functionally redundant as GlcNAc transporters and confirmed the function of Nag5p as a GlcNAc kinase. Upon nonselective growth in the presence of GlcNAc, both plasmids were maintained. This suggests that S. cerevisiae encodes/expresses neither GlcNAc transporter nor GlcNAc kinase activities. On the other hand, this result confirmed the successful expression of two of the four enzymatic activities required for chitin catabolism in S. cerevisiae. Furthermore, we noticed that strains expressing NAG3 or -4 and NAG5 became sensitive to elevated concentrations of GlcNAc and were unable to grow on solid or in liquid media containing >10 mM GlcNAc (data not shown, but see Fig. 4, below).
Expression of NAG3 or -4 and NAG5 in S. cerevisiae can functionally replace ScGNA1. Transformants of the respective strains were streaked onto CSM minimal medium containing glucose and relevant amino acids and G418 for plasmid selection. The plate in the upper panel was supplemented with 0.5 g/liter (2 mM) GlcNAc, while the plate in the lower panel was without GlcNAc. The plates were incubated for 4 days at 30°C prior to photography. The parental strain (BY4742) grew under both conditions.
Relief of GlcNAc sensitivity in S. cerevisiae by expression of the four NAG genes. Strains harboring the indicated genes on freely replicating plasmids were grown on CSM minimal medium supplemented with 12 mM GlcNAc. Note that strains harboring only a GlcNAc transporter (NAG3 or NAG4 when either MET3 promoter driven or TEF promoter driven) or a GlcNAc kinase (NAG5) were not sensitive to GlcNAc; strains expressing only a GlcNAc transporter and a GlcNAc kinase (NAG3 and NAG5 or NAG4 and NAG5) became GlcNAc sensitive at the concentration of GlcNAc used, a sensitivity that was repressed in a strain that encodes all four NAG genes. Plates were incubated for 4 days at 30°C and then photographed.
Coexpression of NAG1 and NAG2 suppresses GlcNAc sensitivity of S. cerevisiae strains carrying NAG3 or -4 and NAG5.To express CaNAG1 and CaNAG2 in S. cerevisiae we cloned these ORFs in the CaNAG5-containing plasmid. Here, CaNAG1 was placed under the control of the Candida maltosa LEU2 promoter, and CaNAG2 was placed under the control of the A. gossypii MET3 promoter (Fig. 2). This plasmid, carrying NAG1, NAG2, and NAG5, was then cotransformed into S. cerevisiae BY4742 together with a NAG4-containing plasmid. We observed that S. cerevisiae strains carrying the NAG3 or -4 and NAG5 genes were sensitive to >10 mM GlcNAc in the medium and failed to proliferate. The expression of NAG1 and NAG2 (together with NAG3 and NAG5) should provide the catabolic protein set that is needed to convert GlcNAc-6-phosphate into fructose-6-phosphate. We hypothesized that this should decrease the pool of GlcNAc-6-phosphate in the cells and suppress the GlcNAc sensitivity. All transformants were obtained on glucose media. Upon transfer to selective plates containing 12 mM GlcNAc (2.65 g/liter), strains carrying only NAG3, NAG4, or NAG5 were able to grow. Thus, the expression of only a GlcNAc transporter or a GlcNAc kinase did not result in GlcNAc sensitivity. This indicates that S. cerevisiae lacks GlcNAc transporter and GlcNAc kinase functions under these conditions. Upon expression of both a GlcNAc transporter and a GlcNAc kinase, cells became sensitive to elevated levels of GlcNAc. Notably, strains expressing all four NAG genes were able to grow in the presence of 12 mM GlcNAc, demonstrating the functional expression of these four genes in S. cerevisiae (Fig. 4).
Utilization of GlcNAc as the sole carbon source by S. cerevisiae.Evident functional expression of the four NAG genes in S. cerevisiae suggested that expression would be sufficient to support growth of S. cerevisiae on minimal media containing GlcNAc as the sole carbon source. To test this, S. cerevisiae was grown in both liquid and solid minimal media containing 2% GlcNAc but no glucose. In both cases we observed that expression of the four NAG genes was sufficient to sustain growth of S. cerevisiae on GlcNAc as the carbon source (Fig. 5). Growth, however, was slower than that of wild-type C. albicans on these media.
Utilization of GlcNAc as the sole carbon source in S. cerevisiae. The indicated strains were streaked onto minimal medium plates containing either 2% glucose or 2% (approximately 8 mM) GlcNAc as the sole carbon source or were grown in liquid minimal medium containing 2% GlcNAc as the sole carbon source. Plates were incubated for 4 days at 30°C and then photographed (A); liquid cultures were grown for 4 days at 30°C, centrifuged, and photographed (B).
Ethanol production based on GlcNAc.Uptake of GlcNAc and its catabolism generates fructose-6-phosphate, which can be used as a fermentable carbon source. Therefore, we sought to determine whether growth on GlcNAc also confers the ability to ferment GlcNAc and produce ethanol with this S. cerevisiae strain that expressed the four NAG genes. To this end we used S. cerevisiae BY4743 as a control strain and compared it to a BY4743 strain that was cotransformed with the NAG gene-containing plasmids. The strains harboring the NAG genes were able to grow in GlcNAc medium as described above and also fermented GlcNAc, as evidenced by the ethanol production. The ethanol yield obtained after 11 days of fermentation was low, at around 3 g/liter, and rather slow, as it continued for up to 2 weeks (Fig. 6).
Fermentation of GlcNAc by S. cerevisiae. Ethanol production by S. cerevisiae strains utilizing GlcNAc was monitored over 2 weeks and is plotted in grams/liter, including standard deviations. Two independent strains were tested in multiple experiments.
DISCUSSION
Chitin is a polymer of β-(1-4)-N-acetyl-d-glucosamine. The amino-sugar moieties can be partially deacetylated. Chitosan, the mostly deacetylated form, has several medical and cosmetic applications due to its nontoxic and nonallergenic properties, e.g., in tissue engineering (32). Chitin is the second most abundant polysaccharide in nature and as a renewable resource could be of interest for biofuel generation. In this study we engineered an S. cerevisiae strain that can utilize the hexose sugar GlcNAc as a carbon source. To achieve this, four genes were necessary to ensure the uptake and phosphorylation of GlcNAc as well as the deacetylation and deamination to convert GlcNAc-6-phosphate to fructose-6-phosphate, which can be fermented into ethanol.
S. cerevisiae readily ferments glucose, fructose, and mannose. Maltose fermentation is limited to certain strains, e.g., brewer's yeast strains harboring one or several MAL loci encoding a maltose permease, a maltase, and a transcriptional activator that regulates the maltose-inducible expression of the former genes (19, 21). Interestingly, the MAL genes in S. cerevisiae are clustered and telomere associated, as are the NAG genes in C. albicans (4; http://www.candidagenome.org/ ).
The utilization of pentose sugars has recently been achieved in a series of experiments and enables the fermentation of lignocellulosic biomass (14, 15). However, the possibility of using chitin and GlcNAc from animal biomass or bacterial cell wall fractions for fermentation by S. cerevisiae has not been examined. This may in part be due to the high costs necessary for enzymatic depolymerization of chitin. Recently, however, it was shown that nonhydrolytic accessory Cbp21-like proteins may increase the efficiency of chitinases (8). Cbp21-like proteins, first identified in Serratia marcescens, a very efficient chitin degrader, have now been identified in a variety of chitinolytic microorganisms (28, 30).
We chose the C. albicans NAG genes for expression in S. cerevisiae since they occur in a clustered fashion and a previous characterization of some of the genes demonstrated several enzymatic activities required for chitin catabolism. The NAG3 and NAG4 paralogs encode putative transporters of the major facilitator superfamily (MFS), but their GlcNAc transporter activity had not been demonstrated previously (3). Another gene with similarity to the MFS, encoded by CaNGT1, was recently suggested to act as a GlcNAc transporter. NGT1 was found to be required for efficient GlcNAc uptake and for the yeast-to-hypha transition at low GlcNAc concentrations. Furthermore, expression of NGT1 in S. cerevisiae promoted GlcNAc uptake (1). Our studies show that NAG3 and NAG4 both promote GlcNAc uptake in S. cerevisiae. Together with NAG5, NAG3 or NAG4 can replace the essential S. cerevisiae GNA1 gene in producing GlcNAc-6-phosphate. This demonstrates the role of the NAG3 or -4 and NAG5 genes in GlcNAc uptake and phosphorylation, respectively.
UDP-GlcNAc plays a central role in yeast metabolism by providing GlcNAc moieties added during N-linked glycosylation of proteins and the synthesis of GPI anchors and for the polymerization of chitin (6). We found that yeast strains expressing NAG3 or -4 and NAG5 are sensitive to elevated GlcNAc concentrations in the medium. This sensitivity was alleviated by the coexpression of NAG1 and NAG2. On the other hand, overexpression of NAG1 and NAG2, e.g., by using the divergent S. cerevisiae histone H2A-H2B promoter, in otherwise-wild-type strains resulted in inviable strains, presumably due to depletion of the UDP-GlcNAc pool (unpublished results). This was corroborated by expressing a truncated and nonfunctional NAG1 gene together with the other NAG genes. In this case GlcNAc sensitivity was not suppressed and growth on chitin as a sole carbon source did not occur.
The S. cerevisiae strains expressing the NAG genes tolerated high concentrations of GlcNAc (10%), and the utilization of GlcNAc was not subject to glucose repression, based on the promoters used for the expression of the NAG genes. This study provides promising initial results for GlcNAc utilization from a shellfish waste stream on an industrial scale by S. cerevisiae. The physiological consequences for yeast growth and proliferation need to be analyzed in more detail. Chitin levels at septal sites, for example, seemed to be increased based on fluorescence microscopy observations. Fermentation of GlcNAc, however, was slow and has to be improved substantially. Several issues should be addressed to progress from proof-of-principle studies to use under industrial conditions in the future: (i) adaptation of the Candida NAG genes to S. cerevisiae codon usage; (ii) analysis of the conversion of fructose-6-phosphate into glucose-6-phosphate; (iii) chromosomal integration of the NAG genes; (iv) use of different fermentation regimens regarding, for example, pH and the addition of other cofactors.
ACKNOWLEDGMENTS
We thank Michael Thorsen for running the fermentation tests.
This research was supported by the Friedrich Schiller University and Leibniz Institute for Natural Products Research and Infection Biology, Hans Knöll Institute, Jena, Germany.
FOOTNOTES
- Received 9 January 2009.
- Accepted 20 July 2009.
- Copyright © 2009 American Society for Microbiology