Lipomyces is one of the most effective starch-degrading yeasts known (Horn et al., 1988). This lipid-forming yeast hydrolyses starch completely due to the secretion of α-amylase, glucoamylase (Spencer-Martins and Van Uden, 1979), isoamylase (Spencer-Martins, 1982) and cyclodextrin glucanotransferase (Spencer-Martins, 1984) activities. Furthermore, of 81 starch-assimilating yeasts evaluated, the highest biomass production on starch medium was obtained with strains of Lipomyces kononenkoae and Lipomyces starkeyi (Spencer-Martins and Van Uden, 1977). When the efficiency of starch degradation was compared between Lipomyces strains and various Schwanniomyces strains, it was found that L. kononenkoae IGC4052B (CBS5608T) was the most efficient starch-degrading yeast among those screened (Horn et al., 1988). In this mutant strain, amylase production is not repressed by glucose (Van Uden et al., 1980). There is conflicting biochemical evidence for some of the properties of the purified α-amylase from L. kononenkoae IGC4052B, e.g. three independent groups have purified an α-amylase, reporting molecular weights of 76 kDa (Steyn and Pretorius, 1995; Prieto et al., 1995) and 38 kDa (Spencer-Martins and Van Uden, 1977). Differences in purification protocols and matrixes may partially explain some of the differences in molecular weight. However, the optimum temperature of 70 °C for the α-amylase purified by Prieto et al. (1995) is in stark contrast with the optimum temperature of 40 °C reported by Steyn and Pretorius (1995) and Spencer-Martins and Van Uden (1979). NH2-terminal sequencing of the α-amylase purified by Steyn and Pretorius (1995) and the translation of a cloned α-amylase gene (LKA1; GenBank Accession No. U30376) demonstrated that Lka1p is indeed the gene product of LKA1. The amino acid composition of the α-amylase purified by Prieto et al. (1995) differs significantly from that of Lka1p, ruling out the possibility that this α-amylase is Lka1p. Additional differences between α-amylases purified from L. kononenkoae IGC4052B are the optimum pH (Steyn et al., 1995, pH 4.0; Prieto et al., 1995, pH 4.5–5; Spencer-Martins and Van Uden, 1979, pH 5.5) and pI (Steyn and Pretorius, 1995, pI 4.17; Prieto et al., 1995, pI 3.5; Spencer-Martins and Van Uden, 1979, pI 7.1). Dissimilarities in amylase enzyme properties are not limited to the α-amylase produced by this yeast. For example, during purification of the α-amylase, Steyn and Pretorius (1995) observed two glucoamylase bands of Mr 100–200 kDa. Subsequently, Chun et al. (1995) reported the purification of two glycosylated glucoamylases, with a Mr of 150 kDa (GI) and 128 kDa (GII) respectively, from L. kononenkoae IGC4052B. These data are in contrast to the predicted Mr of 81.5 for the glucoamylase of L. kononenkoae IGC4052, reported by Spencer-Martins and Van Uden (1979). Taken together, these discrepancies highlight the need for a thorough molecular investigation of the properties of the purified amylases produced by Lipomyces spp.
Based on the conflicting data reviewed above, the purpose of this study was to screen a L. kononenkoae IGC4052B cDNA expression library in S. cerevisiae for other potential α-amylase-encoding genes. In this study, we describe the cloning and characterization of a cDNA and genomic copy of a second L. kononenkoae α-amylase-encoding gene (LKA2; GenBank Accession No. AF443872). The LKA2 gene was also expressed in S. cerevisiae. A comparison of the deduced primary sequences of amylase proteins is also presented.
Materials and methods
Microbial strains, plasmids, media and enzyme assays
The microbial strains and plasmids used in this study are listed in Table 1. S. cerevisiae was cultured on synthetic dextrose (SD), synthetic complete (SC) and YPD media (Rose et al., 1990). YNBS medium (0.67% yeast nitrogen base without amino acids, 1.5% starch, 1% NaH2PO4H2O, 0.4% Na2HPO42H2O, 0.3% peptone, 0.1% yeast extract) was used to grow L. kononenkoae routinely. S. cerevisiae transformants were selected on PHSC−Ura or PHSS−Ura (PHSC refers to SC containing 0.4% Phadebas starch and lacking uracil, whereas PHSS contained 1% Phadebas starch instead of 2% glucose as a carbon source). Plates were incubated at 30 °C for 2–6 days and the digestion of starch was indicated by the appearance of blue haloes surrounding the colonies (Steyn and Pretorius, 1995). α-Amylase activity was quantitatively measured using the Phadebas amylase test (Pharmacia Diagnostics, Uppsala, Sweden) assay (Steyn and Pretorius, 1995). For the isolation of intracellular protein 50 ml YPD broth was inoculated to 0.1 OD600 from an overnight culture and grown at 30 °C for 48 h. Cells were harvested by centrifugation at 5000 rpm for 5 min and resuspended in 5 ml 50 mM Tris (pH 7.5; containing 10 mM NaCl2) buffer. 0.1 g 0.2 mm glass beads were added and the cells were vigorously vortexed for 3 min. After centrifugation at 6000 rpm for 2 min the supernatant, containing the intracellular protein extract, was carefully removed and used for the enzyme assays. The saccharolytic activity of α-amylase was assayed by the dinitrosalicylic acid (DNS) method (Miller et al., 1960). The standard reaction mixture, containing 700 µl substrate, 100 µl enzyme solution and 200 µl buffer, was incubated at 30 °C for 60 min. The reaction was stopped with 1.5 ml DNS reagent and boiled for 15 min. The contents of the assay tubes were diluted with 5 ml distilled water and absorbancy readings (A540) were taken at 540 nm in a Beckman DU-5 spectrophotometer. Various substrates were assayed: soluble starch (Lintner), dextrin, amylose, potato starch, amylopectin, glycogen (oyster) and pullulan. Polysaccharide substrates (0.5%) were heated on a magnetic stirrer prior to use in enzyme assays to facilitate their dissolution. For the determination of substrate specificity, the assay was carried out at optimum pH and temperature. The effect of pH on enzyme activity was determined by the DNS assay using various buffers. Citrate phosphate (100 mM) was used for the pH range 3.0–7.0, while Tris–HCl was used for the pH range 7.0–8.0. The DNS assay was also used to determine the effect of temperature on the enzyme activity. The reaction mixtures were incubated at temperatures of 20–80 °C. These incubations were buffered at the pH of the highest enzyme activity determined as described above. For the determination of the optimum pH and temperature, 0.5% of soluble starch (Lintner) was used as a substrate in the reaction mixture. Escherichia coli strains were routinely cultured on LB with 50 µg/ml ampicilin. Plasmid DNA was isolated using the Qiagen plasmid kit (Qiagen).
Table 1. Microbial strains and plasmids used in this study
Strains and plasmids
Eksteen JM, Steyn AJC, Van Rensburg P, Cordero Otero RR, Pretorius IS. 2002. The evaluation and comparison of recombinant Saccharomyces cerevisiae strains expressing α-amylase and glucoamylase genes from Lipomyces kononenkoae and Saccharomyces fibuligera (submitted for publication).
Bi-directional sequencing was performed using the Ready Reaction DyeDeoxy Terminator Cycle Sequencing kit (Applied Biosystems, Foster City, CA, USA) and analysed on an Applied Biosystems Model 373A DNA sequencing system. Oligodeoxyribonucleotides (20–30 nucleotides) were synthesized by the phosphoramidate method on an Applied Biosystems Model 394 DNA/RNA apparatus. DNA was amplified using Expand™ High Fidelity mix (Boehringer Mannheim, Indianapolis, IN) according to the manufacturer's instructions. PCR products were purified by phenol/chloroform extraction and ethanol precipitation. Restriction enzyme digestion of the PCR products was performed at 37 °C for 14 h. The following oligodeoxyribonucleotides were used as primers: LKA2-F 5′-GCGAATTCATGCGGTTAAATCTCAAAC-3′ and LKA2-R 5′-GCCTCGAGTTAAGAAC-AAAATTTCCCAG-3′, for PCR amplification of the genomic copy of LKA2 (LKA2g). Both primers contained restriction endonuclease sites for EcoRI (italicized) and XhoI (underlined), respectively. S. cerevisiae was transformed using the method described by Gietz et al. (1992). E. coli DH5α was prepared according to Dower et al. (1988) and electrotransformed using a Bio-Rad Gene Pulsar Apparatus (Bio-Rad, Richmond, CA). The isolation of plasmid DNA from the yeast transformants was performed according to Adam and Polaina (1991). The nucleotide sequence of the LKA2 gene was deposited in Gene Bank under Accession No. AF443872.
Construction of library and expression vectors
A genomic library of L. kononenkoae IGC4052B was constructed in cosmid pYC1 (Hohn and Collins, 1980), as described before (Steyn et al., 1996). High molecular weight DNA was isolated from L. kononenkoae IGC4052B according to Steyn and Pretorius (1995). Plasmid pJUL3 was constructed by cloning the EcoRI–XhoI fragment of pJUL1 into EcoRI–XhoI-digested pJC1, thereby positioning the LKA2 cDNA fragment between the promoter and terminator elements of the PGK1 gene. Oligonucleotides LKA2-F and LKA2-R were used to amplify LKA2g from genomic cosmid pools (Steyn and Pretorius, 1995). The digestion of the genomic PCR products with EcoRI and XhoI and the subsequent cloning into plasmid pJC1 resulted in the plasmid pJUL2.
Results and discussion
Cloning of a cDNA copy of LKA2
The production of α-amylase, glucoamylase and isoamylase by L. kononenkoae IGC4052 is subject to carbon catabolite repression (Van Uden et al., 1980). Currently, our research efforts are focused on the glucose derepressed mutant strain IGC4052B, which produces large quantities of α-amylase in media containing glucose or starch as the carbon source (Van Uden et al., 1980). Previously, our laboratory reported the purification and characterization of an α-amylase from L. kononenkoae IGC4052B (Steyn and Pretorius, 1995), identified the α-amylase-encoding gene LKA1 (Steyn et al., 1995) and studied the expression of a cDNA and genomic copy of LKA1 in S. cerevisiae and Sz. pombe (Steyn et al., 1996). To investigate the presence of a second putative amylase, we transformed plasmid cDNA library pools into S. cerevisiae YPH259 and screened them on PHSC−Ura plates for halo-producing colonies. The use of Phadebas tablets in minimal medium enabled us to easily distinguish between glucoamylase- and α-amylase-secreting cells (Steyn and Pretorius, 1995). We obtained a single S. cerevisiae clone (LKA2) showing a distinctly small blue halo on PHSC−Ura medium, which is in contrast to the other large, halo-producing clones. Free plasmid DNA was isolated from these transformants and transformed into E. coli DH5α. Plasmids isolated from the bacterial clones were reintroduced into strain YPH259 to validate the secretion of α-amylase on PHSC−Ura and starch plates. Plasmids isolated from halo-producing yeast clones were digested with XhoI and EcoRI and the inserts were cloned into pBluescript KS (Stratagene) for sequence analysis. The sequence results showed that the translated cDNA insert is not identical to LKA1 and that it showed homology to a large number of α-amylases present in bacteria, yeast and fungi (see below). Sequence analysis of the inserts obtained from two large halo-producing clones identified cDNA fragments containing the LKA1 open reading frames (ORFs).
We conclude that we identified a second novel L. kononenkoae α-amylase-encoding gene, LKA2, which can be expressed and produced in S. cerevisiae in a biologically active form.
Nucleotide sequence of LKA2
Restriction analysis of a plasmid pUS1 obtained from an α-amylase-producing S. cerevisiae transformant showed the presence of a cDNA fragment with an approximate size of 1.7 kb. The nucleotide sequence of the 1663 bp cDNA contained a single ORF of 1497 nucleotides and the translated amino acid sequence showed homology to α-amylases from a variety of organisms. We designated this α-amylase-encoding gene LKA2. The nucleotide sequence of the 1.66 kb EcoRI–XhoI fragment is presented in Figure 1. The LKA2 ORF starts with an ATG start codon at position 1 and ends with a TAA stop codon at position 500, followed by a poly(A) region. We assumed that translation initiation occurs at the first in-frame AUG at the 5′ end of the mRNA encoded by the cDNA fragment, as proposed by Kozak (2001). The nucleotide sequence context of the LKA2 translational start signal, GCTCTGATGCGG, does not resemble the yeast consensus sequence, WAMAMAAUGUCY, for highly-expressed genes (Hamilton et al., 1987). However, it does resemble the eukaryote consensus sequence, TCATCATGCG (Kozak, 2001). The TAA termination signal at position 1497 of the LKA2 ORF is followed by several stop codons that are both in and out of frame. The LKA2 3′ non-coding region contains sequences similar to the TAG and TA(T)GT sequences that are believed to play a role in the termination and polyadenylation of yeast genes (Zaret and Sherman, 1982). These include the TAG, TAGT and TACG sequences that are present 9–55 nucleotides downstream from the stop codon (Figure 1).
The coding region of LKA2 has a G + C content of 47.47%, which is similar to that of the total DNA (48.3%) from Lipomyces spp. (Kurtzman, 1984). The codon usage of the LKA2 gene indicates that the most abundant amino acids are Leu (8.82%), Val (8.02%), Ala (8.02%), and Ser (7.41%). The amino acid composition of the α-amylase purified from L. kononenkoae by Prieto et al. (1995), namely 25.3% (w/w) Thr, 14.3% (w/w) Ser, 0.2% (w/w) Lys and 0.8% (w/w) Tyr, differs significantly from that of LKA2, which strongly suggests that these α-amylases are dissimilar.
To identify a putative signal peptide, we used the programs SIGNALP (web) and PSORT (web). Both programs identified cleavage of a presumed signal peptide between A23 and K24 of Lka2p. The calculated Mr of the unmodified Lka2p (499 amino acids) precursor is 55.3 kDa. Processing of the first 23 amino acids of Lka2p during secretion gave rise to a 52.7 kDa (476 amino acids) mature protein with a theoretical pI of 4.71. In contrast, the calculated Mr of mature Lka1p is 65.7 kDa (596 amino acids), with an experimentally determined pI of 4.17 (Steyn et al., 1995; Steyn and Pretorius 1995). It is interesting to note that Lka2p is approximately 100 amino acids shorter than Lka1p, which might suggest a different biological role. Lka2p contains seven putative glycosylation sites (N46–L–T, N92–T–T, N162–T–S, N177–Q–S, N254–Q–S, N378–T–S, N473–G–T), whereas Lka1p contains only two, which are similar to that of the Schwanniomyces occidentalis amylases SWA2 (Claros et al., 1993) and AMY1 (Strasser et al., 1989). The α-amylase from Saccharomycopsis fibuligera (ALP) contains only one potential glycosylation site (Itoh et al., 1987). Intriguingly, Lka2p contains six C residues, which potentially could form disulphide bonds when secreted or when translocated to the cell wall (Figure 1). A sulphated tyrosine signature is present at position 291 (Figure 1).
We conclude that the translated amino acid sequence of Lka2p differs significantly from that of Lka1p and agrees with previously conflicting data (Chun et al., 1995; Steyn and Pretorius, 1995), thus providing clear evidence for the presence of a second α-amylase in L. kononenkoae.
Cloning of a genomic copy of LKA2
The transformation of S. cerevisiae YPH259 with pJUL2 containing the genomic copy of LKA2 (LKA2g) did not result in halo formation on PHSC−Ura plates, suggesting that biologically active α-amylase was not synthesized in S. cerevisiae. We therefore concluded that LKA2 might contain an intron, as does LKA1 (Steyn and Pretorius, 1995). PCR amplification using primers designed from the cDNA sequence of LKA2 (LKA2-F, -R), and cosmid DNA prepared from genomic DNA (Steyn and Pretorius, 1995), revealed the presence of the expected size band of ca. 1.7 kb in several cosmid pools (results not shown). Subsequent cloning and sequencing of LKA2g confirmed that the amplified PCR product was indeed the genomic copy of LKA2 and revealed the presence of a 55 bp intron between nucleotide positions 304 and 305 of LKA2: 5′-GTTAGTATTGATACATATTCATTTGACTACCG-TCTG GCTGACAAAAACTTAATAG-3′. The position of the intron and the donor, acceptor and splicing sites of LKA2 differ from those of LKA1. The LKA2 intron is 8 bp shorter than the LKA1 intron of 63 bp and contains a donor consensus sequence (italicized) that is similar to that of Sz. pombe (GTAA/TGT) (Prabhala et al., 1992). However, the branch site (underlined) is identical to the LKA1 sequence and correlates with the common branch motif shared by mammals and S. cerevisiae (T/AAYTRAY) (Zhang and Marr, 1994). The 3′-splicing site of the LKA2 intron (bold) resembles the Sz. pombe splicing sequence (T/ATAG).
According to our knowledge, this is the second report of a yeast α-amylase containing an intron and, more specifically, its occurrence in the Lipomycetaceae family could contribute significantly to our understanding of the gene organization in this family.
Sequence analysis of LKA2 and Lka2p and comparison with sequences other amylases
Sequence analyses were performed using the GCG program package (version 8) of the Genetics Computer Group, Madison, WI, USA (Devereux et al., 1984). Using GAP and BESTFIT to compare LKA1 with LKA2, we observed an identity and similarity of 35 and 39% respectively. A Gap-BLAST search revealed that mature LKA2 shows homology to fungal (Aspergillus nidulans: 36% identity), yeast (Debaryomyces occidentalis, S. fibuligera, Sz. pombe, Cryptococcus sp. S–2: 34%, 34%, 33% and 38% identity, respectively) and bacterial (Thermoactinomyces vulgaris: 31% identity) α-amylases. In addition, we found significant similarities to bacterial cyclodextrin glycosyltransferases (Thermoanaerobacterium thermosulfurigenes and Bacillus licheniformis: 30% and 33% identity, respectively), which hydrolyse non-reducing cyclic oligodextrins (Figure 2). With homologous amylases, residues and conserved regions involved in substrate and Ca2+ binding are indicated (Buisson et al., 1987; Itoh et al., 1987; Claros et al., 1993). Janse et al. (1993) and Chiba (1997) provided an extensive list of four conserved regions, A, B, B′ and C that are present in the catalytic domain of α-amylases, cyclodextrin glucanotransferases, maltases, pullulanases, isoamylases, oligo-1,6-α-glucosidases, and branching enzymes. Lka2p contains the conserved residues 137DVVINH142, 223GIRLDTARH231, 251EALN254 and 312FLDNQD317 in regions A, B, B′ and C, respectively. However, two residues in these regions of Lka2p differ from the conserved residues present in a wide spectrum of α-amylases (Figure 2). The most significant differences are the presence of L226 (region B), A252 (region B′), D314 and Q316 (region C), which potentially could influence catalytic activity and substrate binding and/or specificity. Holm et al. (1990) reported that mutating K237 (corresponding to R230 of LKA2) to R in region B of the B. stearothermophilus α-amylase had no effect on enzyme activity. However, this conclusion was based solely on estimations of halos on starch plates and did not directly address the catalytic effect or substrate affinity as a result of the changes. Matsui et al. (1992) altered K210 (corresponding to R230 in region B of LKA2) to R or N in the active site of the S. fibuligera α-amylase and reported a three-fold enhancement of short substrate-specific (maltose) catalytic activity of the mutant enzyme (K210→R). Based upon this observation and the homology (38%) between Lka2p and the S. fibuligera Alp1p, it is tempting to speculate that Lka2p might have a high affinity for short substrates. Holm et al. (1990) suggested that W266 (corresponding to L253 in region B′ of Lka2p) is in contact with the docked substrate and corresponds to Trp206 in the barley α-amylase, which has been shown to be essential for enzymatic function (Svensson et al., 1987). The uncharged residue Q is present in region C of both Lka1p (Q443) and Lka2p (Q318) in the place of the highly conserved, charged H, which is involved in substrate binding (Kuriki and Okada, 1995). Søgaard et al. (1993) have shown that the corresponding residue, H290, in barley α-amylase I is involved in the stabilization of the transition state.
We conclude that the amino acid alterations in the conserved regions of Lka2p most probably affect the catalytic activity and/or substrate binding of Lka2p and needs to be investigated in greater detail.
Properties of α-amylase expressed by S. cerevisiae transformants
We tested the ability of the non-amylolytic yeast S. cerevisiae Σ1278b to secrete active α-amylase when transformed with pJUL3. Σ1278b transformants were able to form blue halos on PHSC−Ura plates within 96 h, indicating the stable expression of the L. kononenkoae IGC 4052B LKA2 gene under control of the PGK1 promoter, and effective secretion of the encoded gene product by S. cerevisiae Σ1278b. The amylase activity of the transformed S. cerevisiae strain was primarily cell wall associated. Intracellular extracts of Σ1278b–pJUL3 were used in the determination of the Lka2p activity at different temperatures. LKA2 activity using the DNS assay was measured between 20 °C and 80 °C (Figure 3A). The optimum temperature for Lka2p was 60 °C, indicating a specific activity (specific activity refers to 1.0 U/mg). At temperatures below 55 °C, Lka2p activity decreased gradually, whereas incubation at temperatures over 65 °C drastically decreased Lka2p activity. At 30 °C and 40 °C, the activity was 45% and 50% of the maximum, respectively. Figure 3B shows that the optimum pH is 3.5. At pH 6, the activity was reduced by almost 20%, while the activity at pH 4 was reduced to 85% of the maximum. The substrate specificity of the amylolytic enzyme was determined using a number of different glucose polymers, containing either α-1,4-glucosidic linkages or a mixture of α-1,4- and α-1,6-glucosidic linkages. The enzyme showed high reactivity towards soluble starch (Lintner), dextrin and amylose (Table 2), but only small amounts of reducing sugars were liberated from amylopectin, glycogen and pullulan. The reaction specificity of Lka2p is very similar to that of an α-amylase with associated dextrinase activity and several bacteria and filamentous fungi do produce such enzymes (Hamilton et al., 2000).
Table 2. Substrate specificity of the L. kononenkoae LKA2-encoded α-amylase secreted by the S. cerevisiae transformant
Relative activity (%)
We conclude that the amylolytic capability of L. kononenkoae depends on the reaction specificity of Lka1p and Lka2p. Lka2p, with poor side pullulanase and glucoamylase activity, have great dextrinase activity, thus complementing Lka1p which have better side pullulanase and glucoamylase activity.
The National Research Foundation is gratefully acknowledged for supporting this work financially.