AU3513001A - Protein engineering of glucoamylase to increase pH optimum, ubstrate specificity and thermostability - Google Patents

Description

WO 98/03639 PCTYUS97/12983 -1- PROTEIN ENGINEERING OF GLUCOAMYLASE TO INCREASE pH OPTIMUM, SUBSTRATE SPECIFICITY AND THERMOSTABILITY FIELD OF THE INVENTION The field of the invention relates to mutations to produce a fungal glucoamylase enzyme that is more selective for the production of glucose rather than the a-1,6 linked disaccharide isomaltose, is more thermostable, and has increased pH optimum and produces increased amounts of glucose compared to wildtype enzymes.

BACKGROUND OF THE INVENTION Glucoamylase (EC 3.2.1.3) is a carbohydrase. Discovered in 1951, it is an exo-hydrolase that cleaves Dglucose from the nonreducing ends of maltooligosaccharides, attacking and at a much slower rate, a-(1,6)-glucosidic bonds. It is one of more than one hundred carbohydrases (EC 3.2.1) that cleave O-glycosidic bonds of either a- or 3configuration. The functional and structural relatedness of these enzymes is reflected in the presence of at least three discrete regions of sequence homology between glucoamylase and several a-amylases, aglucosidases, and transglucanosylases [Svensson, 1988], and a similar domain structure to carbohydrases that attack insoluble substrates [Knowles et al., 1987; Svensson et al., 1989)]. Aspergillus awamori glucoamylase (1,4-a-D-glucan glucohydrolase; EC 3.2.1.3) is one of the most important of the glucoamylases.

Glucoamylase is primarily used in industry for the production of high-fructose corn sweeteners in a process that involves 1) a-amylase to hydrolyze starch to maltooligosaccharides of moderate length (dextrin); 2) Glucoamylase to hydrolyze dextrin to glucose; and 3) WO 98/03639 PCT/US97/129 8 3 -2glucose isomerase to convert glucose to fructose. Corn sweeteners have captured over 50% of the U. S.

sweetener market, and the three enzymes used to make them are among the enzymes made in highest volume. In addition, glucose produced by glucoamylase can be crystallized or used in fermentation to produce organic products such as citric acid, ascorbic acid, lysine, glutamic acid or ethanol for beverages and fuel.

Approximately 12% of the country's corn production is processed with glucoamylase. Although glucoamylase has been successfully used for many years, it would be a more attractive product if it produced higher amounts of glucose instead of disaccharides, if it were more stable, and if it could be used in the same vessel with glucose isomerase.

Glucoamylase does not give 100% yield of glucose from dextrin because it makes various di- and trisaccharides, especially isomaltose and isomaltotriose, from glucose [Nikolov et al., 1989].

These products, formed at high substrate concentrations, result from the ability of glucoamylase to form a-(1,6)-glucosidic bonds. Glucoamylase is not as thermostable as either a-amylase or glucose isomerase. The optimum pH of GA (pH4-4.5) is lower than that of camylase (pH5.5-6.5) and glucose isomerase (pH7-8). Therefore glucoamylase hydrolysis must be done separately from the other enzymatic reactions in a different vessel and at lower temperatures, causing higher capital costs.

Glucoamylase from the filamentous fungus Aspergillus niger is the most widely used glucoamylase, and its biochemical properties have been extensively characterized. This enzyme is found mainly in two forms, GAI (616 amino acids; referred to as AA hereinafter) and GAII (512 AA), differing by the presence in GAI of a 104-AA C-terminal domain required for WO 98/03639 PCT/US97/12983 -3adsorption to native starch granules [Svensson et al., 1982; Svensson et al., 1989]. Both forms have a catalytic domain (AA1-440) followed by a Ser/Thr-rich, highly O-glycosylated region (AA441-512) [Gunnarsson et al., 1984]. The first thirty residues of this region are included in the three-dimensional structure of the enzyme [Aleshin et al., 1994; 1996; Stoffer et al., 1995]; they wrap around the catalytic domain like a belt. There is strong AA sequence homology among fungal glucoamylase's in four distinct regions of the catalytic domain that correspond to the loops that form the substrate binding site [Itoh et al., 1987]. In A.

niger glucoamylase these regions are AA35-59, AA104- 134, AA162-196, and AA300-320. The second and third regions partially or completely overlap the three regions of homology to c-amylases [Svensson, 1988]. In addition, the raw starch binding domain (AA512-616) has high homology to similar domains from several starchdegrading enzymes [Svensson et al., 1989].

Kinetic analysis showed that the substrate binding site is composed of up to seven subsites [Savel'ev et al., 1982] with hydrolysis occurring between subsites 1 and 2. The pKC's of hydrolysis, 2.75 and 5.55 [Savel'ev and Firsov, 1982], suggest that carboxylic acid residues at subsites 1 and 2 provide the catalytic acid and base for hydrolysis. Chemical modification experiments showed that three highly conserved residues, Aspl76, Glul79, and Glul80, are protected and are in the active site, suggesting that one or more of them are the possible catalytic residues [Svensson et al., 1990]. Chemical modification experiments also indicated that the highly conserved residue Trpl20 is essential, and is located in subsite 4 [Clarke and Svensson, 1984]. Trpl20 is homologous to Trp83 of Aspergillus oyzae a-amylase [Clarke and Svensson, 1984], which is also located in the active site of that WO 98/03639 PCT/US97/129 8 3 -4enzyme [Matsuura et al., 1984]. Site directed mutagenesis studies have indicated that Glu179 is the catalytic acid residue, while Glu400 is the catalytic base residue [Frandse et al, 1994; Harris et al, 1993; Sierks et al, 1990] Glucoamylases from A. niger [Svensson et al., 1983; Boel et al., 1984] and Aspergillus awamori [Nunberg et al., 1984] have been cloned and sequenced, and have identical primary structures. Innis et al.

[1985] and more recently Cole et al. [1988] have developed vectors (pGAC9 and pPM18, respectively) for glucoamylase expression in yeast, allowing convenient manipulation and testing of glucoamylase mutants.

SUMMARY OF THE INVENTION According to the present invention, a fungal glucoamylase (1,4-a-D-glucan glucohydrolase; EC 3.2.1) with decreased thermal inactivation (increased thermostability) and reduced isomaltose formation provided by the mutation Asn20Cys coupled with Ala27Cys forming a disulfide bond between the two is provided.

Cumulative thermostability is also provided for GA by including the mutation Asn20Cys coupled with Ala27Cys and at least one mutation from Table 13. An engineered GA including Ser30Pro, Glyl37Ala, and coupled with Ala27Cys provides even more thermostability. Cumulative thermostability is also provided for GA by including the mutation coupled with Ala27Cys and at least two mutations from Table 13.

The present invention also provides a fungal glucoamylase with reduced isomaltose formation including an Asn20Cys coupled with Ala27Cys mutation (S-S mutation) and at least one mutation selected from Table 14. In an embodiment Asn20Cys coupled with WO 98/03639 PCT/US97/12983 Ala27Cys mutation and a 311-314Loop (also referred to as 300Loop) mutation are included in an engineered GA.

In a further preferred embodiment the engineered glucoamylase with reduced isomaltose formation includes Asn20Cys coupled with Ala27Cys mutations Ser30Pro and Glyl37Ala.

The present invention also provides engineered fungal glucoamylase including a 311-314Loop mutation whereby reduced isomaltose formation is provided by the mutation. In a further embodiment fungal glucoamylase including a 311-314Loop mutation and at least one mutation from Table 14 are prepared whereby cumulative reduced isomaltose formation is provided by the additional mutation.

The present invention provides a fungal glucoamylase including a mutation Ser411Ala whereby increased pH optimum and reduced isomaltose formation is provided by the mutation. In an embodiment the Ser411Ala mutation is combined with at least one mutation from Table 15 whereby cumulative increased pH optimum is provided by the mutations. In an embodiment the Ser411Ala mutation is combined with at least one mutation from Table 14 whereby cumulative reduced isomaltose formation is provided by the mutations.

In a further embodiment an engineered fungal glucoamylase includes a mutation Ser411Ala and a mutation pair Asn20Cys coupled with Ala27Cys forming a disulfide bond between the two members of the pair whereby increased thermal stability, increased pH optimum and reduced isomaltose formation are provided by the mutations.

In a still further embodiment a fungal glucoamylase is engineered to include a Ser411Ala mutation and a mutation pair Asn20Cys coupled with Ala27Cys forming a disulfide bond between the two members of the pair and a 311-314Loop mutation whereby WO 98/03639 PCTUS97/12983 -6increased thermal stability, increased pH optimum and reduced isomaltose formation are provided by the mutations.

The present invention provides a method to obtain a fungal glucoamylase with reduced isomaltose formation by designing mutations to decrease the glucosidic linkage affinity of GA.

The present invention also provides a method to obtain a fungal glucoamylase with decreased thermal inactivation by designing mutations to decrease the enzyme's conformational entropy of unfolding and/or increase stability of a-helices, increase disulfide bonds, hydrogen bonding, electrostatic interactions, hydrophic interactions, Vanderwalls interactions and packing compactness.

The present invention also provides a fungal glucoamylase with increased pH optimum including changing the polarity, charge distribution and hydrogen bonding in the microenvironment of the catalytic base Glu400.

The present invention also provides a method of genetically engineering glucoamylase carrying at least two cumulatively additive mutations. Individual mutations are generated by site-directed mutagenesis.

These individual mutations are screened and those selected which show increased pH optimum and which show decreased irreversible thermal inactivation rates or reduced isomaltose formation. Site directed mutagenesis is then performed to produce enzymes carrying at least two of the isolated selected mutations. Finally the engineered enzymes are screened for cumulatively additive effects of the mutations on thermal stabilizing or reduced isomaltose formation by the produced enzymes carrying at least two of the isolated selected mutations. Alternatively, the engineered enzyme is screened for cumulatively additive WO 98/03639 PCT/US97/12983 -7effects of both of the mutations on pH optimum, thermostability and/or reduced isomaltose formation by the produced enzymes carrying at least two of the isolated selected mutations.

Vectors for each of the mutations and mutation combinations are also provided by the present invention as well as host cells transformed by the vectors.

DESCRIPTION OF THE DRAWINGS Other advantages of the present invention will be readily appreciated as the same becomes better understood by reference to the following detailed description when considered in connection with the accompanying drawings wherein: FIGURE 1 is a graph showing the relationship between temperature and kd for wild-type and proline substituted mutant GA's: S30P D345P E408P in Example 1.

FIGURE 2 is a graph showing effect of temperature on first-order thermoinactivation rate coefficients of wild-type A27C N20C A27C/N20C A471C/T72C A27C/N20C/G137A A27C/N20C/S436P AND G137A/S436P glucoamylases measured in pH 4.5 buffer.

FIGURE 3 is a graph showing initial reaction rates of wild-type A27C/N20C A471C/T72C and A29C/N20C/G137A glucoamylases with 4% maltose in 0.05 M sodium acetate (pH 4.5) as substrate at temperatures from 60°C to 76 0

C.

FIGURE 4 is a graph showing the effect of temperature on the activity of wildtype and mutant GA.

Error bars represent the standard deviation /rom three assays,. Wildtype S30P/G137A S-S/S30P/G137A FIGURE 5A-C are graphs showing the effect of WO 98/03639 PCT/US97/12 9 8 3 -8temperature on irreversible thermal inactivation rate coefficients of wildtype and mutant GA. Fig. Wildtype S30P G137A S30P/G137A Fig.

Wildtype S30P S-S (hexagon), (filed circle with empty center); Fig. 5C Wildtype (e) S30P/G137A S-S/S30P (filed circle with empty center), S-S/S30P/G137A FIGURE 6A-B are graphs showing saccharification of 28% Maltrin M100 by wildtype

S

3 0P/G137A (0) and S-S/S30P/G137A FIGURE 7 is a graph showing the 30% DE maltodextrin saccharification of wildtype and mutant glycoamylases: 300Loop S30P/G137A

S-S

S

3 0P/G137A/300Loop S-S/300Loop at 550C, enzyme concentration was 166.67 g/mL in each reaction. FIGURE 8 is a graph showing production of..

isomaltose by wildtype and mutant glucoamylases: Y116W Y175F R241K S411A S411G (hexagon), during glucose condensation at 55 0 C with D-glucose in 0.05M sodium acetate buffer at pH4.4 with 0.02% sodium azide for 12 days.

FIGURE 9 is a graph showing the production of glucose by wildtype and mutant glucoamylases: Y116W Y175F R241K S411A S411G (hexagon), during hydrolysis of DE 10 maltodextrin at 55 0 C with 28% maltodextrin in 0.05M sodium acetate buffer at pH4.4 with 0.02% sodium azide for 12 days.

FIGURE 10 is a graph showing the initial rates of glucose production by wildtype and S411A (0) glucoamylases during DE 10 maltodextrin hydrolysis at different pH values. Hydrolysis was performed at 360C with 28% maltodextrin in 25mM citrate-phosphate buffer at indicated pHs with 0.02% sodium azide for 4 days.

WO 98/03639 PCT/US97/12983 -9- DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT The present invention provides mutations for increased thermal stability, increased pH optimum and reduced isomaltose formation in the glucoamylase from fungal species which may provide increased glucose yields compared to wildtype glucoamylase. Predicted structure and known sequences of glucoamylase are conserved among the fungal species [Coutino et al, 19943. As an exemplar Aspergillus awamori glucoamylase (l,4-c-D-glucan glucohydrolase; EC 3.2.1.3; referred to as GA herein; SEQ ID No:l) is used, but any other fungal species including Aspergillus species glucoamylase can be used. The numbering of the glucoamylase amino acids herein is based on the sequence of the exemplar Aspergillus awamori.

Equivalent amino acid residue numbers are determined differently for different fungal species as is known in the art [Coutino et al., 1994].

The present invention provides a fungal glucoamylase with decreased thermal inactivation (increased thermostability) and decreased isomaltose formation provided by engineering the inclusion of a mutation pair Asn20Cys coupled with Ala27Cys which forms a disulfide bond between them (this mutation is abbreviated as Asn20Cys/Ala27Cys or Additional mutations providing decreased thermal inactivation are set forth in Summary Table 13.

Cumulative thermostability is also provided for GA by including at least two of the mutations in the enzyme as for example including mutations Ser30Pro and Glyl37Ala. Another example is to engineer S-S with Asn20Cys/Ala27Cys in the enzyme or to pair Glyl37Ala with S-S. Further, combinations of the individual mutations set forth in Table 13, particularly with S-S coupled with Ser30Pro also provide cumulative WO 98/03639 PCT/US97/1298 3 thermostability. In general two mutation combinations are made but triple mutations can also be constructed.

As for example, an engineered GA including the three mutations: Ser30Pro, Gly137Ala, and Asn20Cys/Ala27Cys provides even more thermostability.

By Asn20Cys coupled with Ala27Cys is meant a pair of mutations which is abbreviated as or Asn20Cys/Ala27Cys and between which is formed a disulfide bond as described herein in the Examples. In general, this is referred to as a single mutation since both are required to form the disulfide bond.

By cumulative is generally meant the additive (or nearly additive) effects of two or more mutations on the parameter of enzyme activity being measured.

The present invention also provides a fungal glucoamylase with reduced isomaltose formation and increased glucose yield including the Asn20Cys/Ala27Cys mutation (S-S mutation) and at least one mutation selected from Table 14. In an embodiment the Asn20Cys/Ala27Cys mutation and the 311-314Loop (300Loop) are included in GA. In a further preferred embodiment the engineered glucoamylase with reduced isomaltose formation includes Asn20Cys/Ala27Cys and with mutations Ser30Pro and Gly137Ala.

In an embodiment a glucoamylase with the 311-114 loop mutation is constructed to provide reduced isomaltose formation. By the 3 11-314Loop mutation is meant an insertional GA mutant with the sequence Tyr 3 11-Tyr312-Asn313-Gly314-Tyr311-Asn-Gly-Asn-Gly-Asn- Ser-Gln-Gly314 (311-314 Loop; SEQ ID No:2).

The present invention provides a fungal glucoamylase including a Ser411Ala mutation whereby increased pH optimum and reduced isomaltose formation is provided by the mutation. In an embodiment the Ser411Ala mutation is combined with at least one mutation from Table 15 whereby cumulative increased pH WO 98/03639 PCT/US97/12983 -11optimum is provided by the combined mutations. In a further embodiment the Ser411Ala mutation is combined with at least one mutation from Table 14 whereby cumulative reduced isomaltose formation is provided by the mutations.

In a further embodiment an engineered fungal glucoamylase includes a Ser411Ala mutation and the mutation pair Asn20Cys/Ala27Cys forming a disulfide bond between them whereby increased thermal stability, increased pH optimum and reduced isomaltose formation are provided by the mutations.

In a still further embodiment a fungal glucoamylase including a Ser411Ala mutation and a mutation pair Asn20Cys coupled with Ala27Cys forming a disulfide bond between the two members of the pair and a 311-314Loop mutation whereby increased thermal stability, increased pH optimum and reduced isomaltose formation are provided by the combination of mutations.

Mutations are indicated by the amino acid being replaced followed by the residue number followed by the replacing amino acid. Amino acids are abbreviated either with the three letter code or single letter code. Mutations are generated using site directed mutagenesis as is known in the art. The sequence and residue number are from the Wildtype (WT) or nonmutant enzyme. Biochemical characterization is performed as described herein below and in the Examples. The Examples provide exemplars of the analysis for an individual mutation to determine it's characteristics and provide exemplars of analysis for combinations of mutations to determine if the combination provides a cumulative effects.

By increased thermostability (or decreased thermal inactivation) is meant that at temperatures between 65 0 C and 77.5 0 C the mutants are irreversibly inactivated at a decreased rate compared to wildtype.

WO 98/03639 PCTIUS97/129 8 3 -12- The present invention provides a method to obtain fungal glucoamylases with decreased thermal inactivation by designing mutations to decrease the rate of irreversible thermal inactivation at temperatures between 65 0 C and 77.5 0 C compared to wildtype. This is accomplished by designing glucoamylases with decreased thermal inactivation by designing mutations to decrease the enzyme's conformational entropy of unfolding and/or increase stability of a-helices, increase disulfide bonds, hydrogen bonding, electrostatic interactions, hydrophic interactions, Vanderwalls interactions and packing compactness.

Basic mechanisms underlying protein thermostability and factors influencing reversible and irreversible thermal inactivation have been studied extensively [Argos et al., 1979; Klibanov, 1983; Wasserman, 1984; Ahern and Klibanov, 1985]. Factors involved in stabilizing proteins at high temperatures include 1) disulfide bonds, 2) noncovalent bonds such as salt bridges, hydrogen bonding, and hydrophobic interactions, and 3) conformational rigidity [Nosoh and Sekiguchi, 1988]. The causes of irreversible inactivation at high temperatures include 1) aggregation, 2) the formation of incorrect structures, 3) the destruction of disulfide bonds, 4) deamidation (especially of Asn at Asn-Gly sequences), and cleavage of Asp-X peptide linkages. It is apparent that replacement of even one residue can make a large difference in protein thermostability [Matsumura and Aiba, 1985], due to the small increases in free energy (20-30 kJ/mol) usually required to stabilize protein tertiary structures [Nosoh and Sekiguchi, 1988].

Genetic engineering to increase thermostability (or to decrease irreversible thermoinactivation) of enzymes has been successful in several cases [Perry and Wetzel, WO 98/03639 PCT/US97/12983 -13- 1984; Imanaka et al., 1986; Ahearn et al., 1987].

However, the mechanisms that govern thermostability are not fully understood, so that amino acid (AA) replacements that promote thermostability are not accurately predicted [Leatherbarrow and Fersht, 1986; Nosoh and Sekiguchi, 1988; Pakula and Sauer, 1989].

The method of the present invention allows for more accurate prediction.

By increased pH optimum is meant that the enzyme is functional at a higher pH, above that of wildtype.

The present invention also provides a method to design a fungal glucoamylase with increased pH optimum by changing the polarity, charge distribution and hydrogen bonding in the microenvironment of the catalytic base Glu400. For example, mutants S411G and S411A were designed to remove the hydrogen bond between Ser411 and Glu400 (see Example 8).

By increased selectivity is meant that there is decreased isomaltose formation due to decrease in the production of undesirable a-(1,6)-linked byproducts (reversion products) at high glucose concentrations [Lee et al., 1976]. As described above, GA hydrolyzes and synthesizes both and oe-(1,6) glucosidic bonds. Increasing selectivity indicates that the enzyme synthesizes al,6 linked products at a lower rate than wildtype as shown by reduced levels of isomaltose formation in condensation reactions with 30% glucose as a substrate compared to wildtype GA. Additionally, improved selectivity may result in increased glucose yields in saccharification reactions using 28% DE maltodextran as a substrate.

The present invention provides a method to obtain a fungal glucoamylase with reduced isomaltose formation by designing mutations to decrease the glucosidic linkage affinity of GA. That is mutations are designed in the active site to reduce isomaltose WO 98/03639 PCTYUS97/12 9 8 3 -14formation due to glucose condensation. The mutations are designed to have decreased ability to synthesize isomaltose while maintaining at least partial wildtype ability to digest al-4 linked substrates resulting in a lower ratio of isomaltose formation to glucose formation than wildtype. These mutations are made at positions that are not completely conserved based on homology analysis.

Kinetic studies have indicated that there are five to seven glucosyl binding subsites, and the catalytic site is located between subsites 1 and 2 [Hiromi et al., 1973, Hiromi et al., 1983, Meagher et al., 1989, Tanaka et al., 1983]. The solved three-dimensional structure of the catalytic domain of glucoamylase from Aspergillus awamori var X100, which has about homology with the corresponding regions of GAs from Aspergillus awamori and Aspergillus niger [Coutinho Reilly, 1994], contains thirteen alpha-helices, twelve of which are arranged in pairs forming an alpha/alpha barrel [Aleshin et al., 1992, Aleshin et al., 1994].

The active site is located in the cavity of the barrel center. In addition, homology analysis of thirteen amino acid sequences of glucoamylases showed that five conserved regions define the active site [Coutinho Reilly, 1994]. The mechanism of GA catalysis involves two carboxyl groups [Hiromi et al., 1966], Glul79 and Glu400 (in Aspergillus awamori or Aspergillus niger) [Frandsen et al., 1994, Harris et al., 1993, Sierks et al., 1990]. Glul79 protonates the oxygen in the glycosidic linkage, acting as general acid acatalyst, and Glu400 activates water (Wat500) for nucleophilic attack at carbon C-l, acting as a general base catalyst [Frandsen et al., 1994]. The crystal structures of glucoamylase complexed with the pseudotetrasaccharides (acarbose and D-gluco-dihydroacarbose), showed that there are two different binding conformers, pH 4-type WO 98/03639 PCT/US97/12983 and pH 6-type, for pseudotetrasaccharides at pH 4 [Aleshin et al., 1996, Stoffer et al., 1995]. Binding of the first two sugar residues of the pseudotetrasaccharides is the same, but there is an extraordinary variation in binding of the third and fourth sugar residues of the pseudotetrasaccharides [Stoffer et al., 1995].

The substrate specificity of an enzyme is determined by its ability to form a stable complex with a ligand in both the ground state and the transition state. The stability of the enzyme-ligand complex is affected by steric constraints, hydrogen bonding, van der Waal's and electrostatic forces, and hydrophobic contacts [see generally Fersht, 1985 Enzyme Structure and Mechanism, 2 nd edition, Freeman, San Francisco].

Site-directed mutagenesis was used to construct mutations at residues 119 and 121 to alter the hydrogen bonding between enzyme and substrate. Atom OG of Serll9 hydrogen bonds to the 3-OH of the fourth sugar residue of pseudo-tetrasaccharides only in the pH 6type conformer, whereas the amide nitrogen of Glyl21 hydrogen bonds to the 6-OH of the third sugar residue in both pH 4-type and pH 6-type conformers. These mutations are designed to change substrate specificity (decrease alpha-1,6 condensation reactions) while maintaining wild-type ability to hydrolyze alpha-1,4 linked substrates. Serll9 is not conserved and is replaced by Ala, Pro and Glu in other GAs. Mutant S119E was designed to strengthen the hydrogen bond between the enzyme and the fourth sugar residue of the substrate to stabilize the pH 6-type conformer, and to bring a negative charge near subsite 4 in order to increase electrostatic interactions in active site.

Mutant S119G was designed to remove the same hydrogen bond in order to destabilize the pH 6-type conformer.

Mutant S119W'was designed to remove the same hydrogen WO 98/03639 PCT/US97/1298 3 -16bond and to increase the hydrophobic interactions between the enzyme and the pH 6-type conformer. Glyl21 is highly conserved in all glucoamylase sequences except in Clostridium sp. G005 GA, which has high c-1,6 activity and in which Gly is replaced by Thr. Since the 0 and 0 angles of Gly121 would allow an alanine in this position without causing a conformation distortion, G121A was designed to introduce a Betacarbon at position 121 to displace the 6-OH group of the third sugar residue from its hydrogen bonding position. In addition, the double mutant G121A/S411G was designed to investigate additivity of the two substrate specificity mutations. S411G is shown herein to reduce the ratio of initial rates of isomaltose production (from glucose condensation reactions) to that of glucose production (from the hydrolysis of DE maltodextrin).

The following provide further examples of the strategies used for the design of mutations having increased selectivity.

300Loop mutation: According to the amino acid sequence homology study [Countinho and Reilly, 1994], it was found that GAs from Rhizopus and some other fungal families have a longer amino acid sequence and form a larger loop or cavity in the S4 conserved region compared to A. niger or A. awamori GAs. Since single mutation events alone are unlikely to bring about substantial increase in the specificity of bond hydrolysis or synthesis, an insertional mutant GA was designed, designated 300Loop or 311-314Loop (SEQ ID NO:2), and the inserted seven amino acids were adapted from Rhizopus GA because Rhizopus GA was the first enzyme to which the subsite theory was successfully applied [Himori et al., 1973]. The 300Loop mutation was designed to decrease the a-(1,6)-glucosidic affinity by introducing a larger loop into the S4 WO 98/03639 PCT/US97/12983 -17conserved region.

Tyrl75 is within the third conserved region.

The nearest distance between Tyrl75 and the fourth residue of inhibitor D-gluco-dihydroacarbose is 4.06 A [Stoffer et al., 1995]. Tyrl75 is replaced by Phe or Gin in several other glucoamylases. Changing Tyr175 to Phe was designed to increase the hydrophobic interaction between enzyme and substrate.

Glyl21Ala: Glyl21 is highly conserved in all glucoamylase sequences except in Clostridium sp. G005 GA, which has high a-1,6 activity and in which Gly is replaced by Thr. Since the 0 and V of Glyl21 would allow an alanine in this position without causing a conformation distortion, G121A was designed to introduce a 3-carbon at position 121 to displace the 6- OH group of the third sugar residue from its hydrogen bonding position.

Glyl21Ala with S411G (generally indicated as G121A/S411G): The double mutant was designed to investigate additivity (cumulative) of the two substrate specificity mutations. S411G reduces the ratio of initial rates of isomaltose production (from glucose condensation reactions, see Examples) to that of glucose production (from the hydrolysis of maltodextrin The present invention provides a method of engineering mutations for fungal glucoamylase and then preparing engineered enzymes carrying cumulatively additive mutations. The initial step is to generate individual mutations by site directed mutagenesis and screen the individual mutations as described in the Examples. Those individual mutations which show decreased irreversible thermal inactivation rates or reduced isomaltose formation or increased pH optimum are then selected for combinational analysis. In general mutations are selected which have at least WO 98/03639 PCT/US97/12 983 -18wildtype reaction rates.

Mutations are combined by site-directed mutagenesis to determine if their effects are additive as is discussed herein in the Examples. Site directed mutagenesis to produce enzymes carrying at least two of the isolated selected mutations is performed as is known in the art. These engineered enzymes are then screened for cumulatively additive effects on thermal stabilizing, pH optimum or reduced isomaltose formation. Alternatively the engineered enzymes carrying cumulative mutations are screened for cumulative effects on two or more of the parameters.

For biochemical characterization of the mutants, GA is purified from culture supernatants of 15-L batch fermentations by ultrafiltration, DEAE-Sephadex column chromatography, and column affinity chromatography using the potent inhibitor acarbose attached to a support [Sierks et al., 1989]. Purities of the resulting preparations are tested by standard techniques such as SDS-polyacrylamide gel electrophoresis and isoelectric focusing with narrowband ampholytes. Protein are measured by absorbance at 280 nm or by Bradford's method [1976]. GA activity is measured by a glucose oxidase/o-dianisidine assay (Sigma).

Selectivity is determined by any method known in the art but preferably by measuring the initial rate of isomaltose formation from 30% glucose condensation reactions at pH 4.4 and 55 0 C in 0.05M sodium acetate buffer and then by measuring the initial rage of glucose formation in 30% DE maltodextran hydrolysis reactions at pH 4.4 and 55 0

C

0.05M sodium acetate buffer. From the resulting initial rates, the ratio of isomaltose formation to glucose formation is calculated.

Thermostability is measured as is known in the art WO 98/03639 PCT/US97/12983 -19but preferably by incubating the enzyme at selected temperatures between 65 0 C and 77.5 0 C at 2.5 0 C intervals followed by activity analysis at 35 0 C using 4% maltose as substrate. When first-order decay is observed, as with WT GA, decay rate coefficients are determined.

Activation energies for decay are calculated from the rate coefficients at different temperatures.

pH optimum is measured as is known in the art but preferably at 45 0 C at 16 pH values, ranging for 2.2 to 7.0 using 0.025 M citrate-phosphate buffer with maltose or maltoheptaose as substrate.

Saccharification is measured as described in the Examples. Briefly, glucoamylase is incubated with DE maltodextran as substrate in 0.05M sodium acetate buffer at pH 4.4 at 55 0 C. Samples are taken at various times from 0.5 to 288 hours and the production of glucose determined.

The present invention provides vectors comprising an expression control sequence operatively linked to the nucleic acid sequence of the various mutant sequences disclosed herein, combinations of mutations and portions thereof. The present invention further provides host cells, selected from suitable eucaryotic and procaryotic cells, which are transformed with these vectors.

Vectors can be constructed containing the cDNA of the present invention by those skilled in the art and should contain all expression elements necessary to achieve the desired transcription of the sequences.

Other beneficial characteristics can also be contained within the vectors such as mechanisms for recovery of the nucleic acids in a different form. Examples are provided herein. Phagemids are a specific example of such beneficial vectors because they can be used either as plasmids or as bacteriophage vectors. Examples of other vectors include viruses such as bacteri6phages, WO 98/03639 PCT/US97/129 8 3 baculoviruses and retroviruses, DNA viruses, cosmids, plasmids, liposomes and other recombination vectors.

The vectors can also contain elements for use in either procaryotic or eucaryotic host systems. One of ordinary skill in the art will know which host systems are compatible with a particular vector.

The vectors can be introduced into cells or tissues by any one of a variety of known methods within the art (calcium phosphate transfection; electroporation; lipofection; protoplast fusion; polybrene transfection; ballistic DNA delivery; lithium acetate or CaCI transformation). The host cell can be any eucaryotic and procaryotic cells, which can be transformed with the vector and which will support the production of the enzyme.

The above discussion provides a factual basis for thermostable and selective mutants of fungal glucoamylase as well as methods of designing the mutations and screening for the cumulative effect of the mutations and vectors containing the mutations.

The methods used with and the utility of the present invention can be shown by the following non-limiting examples and accompanying figures.

EXAMPLES

General methods in molecular biology: Standard molecular biology techniques known in the art and not specifically described were generally followed as in Sambrook et al., Molecular Cloning: A Laboratory Manual, Cold Springs Harbor Laboratory, New York (1989), and in Ausubel et al., Current Protocols in Molecular Biology, John Wiley and Sons, Baltimore, Maryland (1989) and Rose, et al. Methods in Yeast Genetics: A Laboratory Course Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY (1990).

Polymerase chain reaction (PCR) was carried out WO 98/03639 PCT/US97/12983 -21generally as in PCR Protocols: A Guide To Methods And Applications, Academic Press, San Diego, CA (1990).

Oligonucleotides are synthesized as is known in the art. For example, an Applied Biosystems 380B DNA synthesizer can be used.

Materials: S. cerevisiae C468 (c leu2-3 leu 2-112 his 3-11 his 3-15 mal-) and the plasmid YEpPM18 were gifts from Cetus. Acarbose was a gift from Miles Laboratories. All restriction enzymes were purchased from Promega as well as T4 DNA ligase and pGEM-7Z(+), an E. coli phagemid vector, were from Promega. Maltose

(G

2 maltotriose

(G

3 maltotetraose

(G

4 maltopentaose maltohexaose

(G

6 maltoheptaose glucose oxidase, peroxidase, and c-naphthol were from Sigma. Isomaltose (iG 2 was purchased from TCI America.

DE 10 Maltodextrin with the average degree of polymerizations (DP) of 10, 6, and 4, respectively, were from Grain Processing Corporation. Highperformance thin-layer chromatographic (HPTLC) plates (LHPK silica gel 60 A, 20 x 10 cm) were from Whatman.

Site-directed mutagenesis: Site-directed mutagenesis was performed according to the Muta-Gene phagemid in vitro mutagenesis kit from Bio-Rad which is based on the method of Kunkel et al [1985]. A 1.7 kb XhoI-BamHI DNA fragment coding for the glucoamylase catalytic domain was cloned into a pBluescript II KS(+) vector from Stratagene. Oligonucleotides used as mutagenic primers are provided with the specific Example. The presence of the individual mutations was confirmed by sequencing and each mutated GA gene fragment was subcloned into YepPM18 [Cole, et al., 1988] and transformed into S. cerevisiae.

Enzyme production and purification: Wild-type (WT) and mutant enzymes are produced by growing yeast at 30 0 C in 5.3 L SD His media for 72 hours at pH 4.5 in a 5.0 L fermentor. After 48 hours, i00g of dextrose WO 98/03639 PCT/US97/12983 -22and 22g of (NH 4 2 SO, in 300ml H 2 0 is added as a supplement. Following growth, the culture is centrifuged to remove yeast cells, the supernatant is concentrated by ultrafiltration, diafiltered against 0.5 M NaCl/0.1 M NaOAc, pH 4.5 and purified by acarbose-sepharose affinity chromatography. GA is eluted with 1.7 M Tris-Cl, pH 7.6, dialyzed against

H

2 0, further concentrated by ultrafiltration and diafiltered against 0.05 M NaOAc buffer, pH 4.5. The protein concentration is determined according to the Pierce bicinchoninic acid protein assay [Smith et al., 1985] using bovine serum albumin as a standard.

Enzyme activity assays: Enzyme activities were determined at 50 0 C using 4% maltose in 0.05 M NaOAc buffer pH 4.5 as substrate. One international unit (IU) of enzyme activity was defined as the amount of enzyme required to produce 1 ymol/min glucose at assay conditions. Following mixing enzyme with substrate, six 100 pil samples were removed at seven minute intervals over 42 minutes, the reaction stopped with pl of 4.0 M Tris-Cl, pH 7.0 and the glucose concentration was determined according to the Sigma peroxidase-glucose oxidase/o dianisidine glucose assay kit.

Irreversible thermal inactivation: Duplicate aliquotes of 40 pg/ml of purified wild-type and mutant enzymes were subjected to inactivation at six or more temperatures between 650 and 80 0 C at intervals of 0 C. Samples were removed at six different time points, immediately placed on ice and stored at 4°C for 24 hours. The residual activity of the inactivated samples along with a corresponding sample which had not been subjected to thermal inactivation, was determined as described above but at 35 0

C.

pH dependence of glucoamylase activity: pH depenoance of glucoamylase activity was measured at WO 98/03639 PCTYUS97/12983 -23- 0 C at 16 different pH values, ranging from 2.2 to using 0.025 M citrate-phosphate buffer [Mcllvane, 1921] with maltose or maltoheptaose as substrate. The ionic strength of the citrate-phosphate buffer was maintained at 0.1 by adding potassium chloride. The pK values of free enzyme and enzyme-substrate complex were measured at substrate concentrations smaller than 0.2 K, so that the initial rate was proportional to kcat/Km, and (ii) higher than 10 so that the initial rate was proportional to kca [Sierks Svensson, 1994, see also Whitaker (1994) Principle of enzymology for the food sciences, 2 nd edition, Marcel Dekker, NY].

The pK values of two catalytic groups of free enzyme and enzyme-substrate complex were calculated by fitting the initial rates as a function of pH values to the equation log Y log [C/(1+H/K 1 by using the software of Enzfitter. Y is the observed value of the parameter of interest kca/K, or kcat) measured at different pH values, C is the pH independent value of Y the maximal value of kca/Km or kca,), H is the concentration of hydrogen ion, K, and K 2 are dissociation constants of catalytic groups of enzyme.

When the values of apparent pK, and pK 2 were separated by less than 3 pH units, the pK values were adjusted by equations (1) 2 K, 4 and opt KK, 2 [Whitaker, 1994]. The concentration of hydrogen ion at the optimum pH, (H*)opt, was calculated from pHopt which is equal to the average of apparent pK, and pK 2

(IT)

1 and 2 (apparent K, and K 2 correspond to the concentrations of hydrogen ion when the pH values are equal to apparent pK 1 and pK 2 respectively.

The hydrolysis of DE 10 maltodextrin (Saccharification): Hydrolysis was performed at 35 0

C

and/or 55 0 C (as indicated in the text) with 28% (w/v) DE 10 maltodextrin as substrate in 0.05 M sodium acetate buffer at pH 4.4 with the addition of 0.02% WO 98/03639 PCT/US97/12983 -24sodium azide, used to inhibit microbial growth in the reaction mixtures. The enzyme concentration was 2.64 lM for both wild-type and mutant GAs. Samples were taken at various times (from 0.5 to 288 hours) and the reactions were stopped by adding samples to the same volume of 1 M Tris-HC1 buffer at pH 7.0, since Tris is a known inhibitor of glucoamylase [Clarke Svensson, 1984]. The production of glucose was determined by the glucose oxidase method [Rabbo Terkildsen, 1960].

Initial rates of glucose production were determined by fitting the experimental data to the equation c At/(l+Bt), where c is the product concentration, t is time, and A (the initial rate) and B are obtained from the nonlinear regression. At 55C, only the time points before 70 hours were used for the calculations, since the glucose production by that time had already declined for wild-type

GA.

Glucose condensations reactions: Glucose condensation reactions were performed at 35 0 C and 55 0

C

with 30% D-glucose as substrate in 0.05 M acetate buffer at pH 4.4 for 12 days with the addition of 0.02% sodium azide, used to inhibit microbial growth in the reaction mixtures. The enzyme concentration was 2.64 AM for both wild-type and mutant GAs. Samples were taken at various times and the reactions were stopped by adding samples to the same volume of 1 M Tris-HCl buffer at pH 7.0. High Performance Thin Layer Chromatography (HPTLC) and Imaging Densitometry were used to determine the production of isomaltose by a method modified from that described by Robyt et al..

[Robyt and Mukerjea, 1994]. One microliter of variously diluted samples and six different concentrations of standard (containing glucose, maltose and isomaltose) were applied to the HPTLC plates. The developing solvent system contained acetonitrile, ethyl acetate, 1-propanol and water in the volume proportions WO 98/03639 PCT/US97/12983 of 85:20:50:40. Only one ascent was used to develop the carbohydrate separation on HPTLC plates. After development, the plates were air-dried, dipped into an EtOH solution containing 0.3% alpha-naphthol and 5% H 2

SO

4 air-dried again, and incubated approximately 10 min at 120 0 C to visualize the carbohydrates. Densities of the isomaltose spots on HPTLC plates were quantified by Imaging Densitometry (Bio-Rad, Model GS-670), using Molecular Analyst software (Bio-Rad). The experimental data were fitted to the equation c At/(l+Bt), described above for the hydrolysis of DE 10 maltodextrin, to obtain the initial rates of isomaltose production.

EXAMPLE 1 STABILIZATION OF ASPERGILLUS AWAMORI GLUCOAMYLASE BY PROLINE SUBSTITUTION The following example is an exemplar of the methods and procedures that are used in the analysis of an individual mutation of a glucoamylase. To investigate the mechanisms governing Aspergillus awamori glucoamylase thermal stability, three proline substitution mutations were constructed. These mutations were predicted to increase GA stability by decreasing the enzyme's conformational entropy of unfolding.

Aspergillus awamori glucoamylase (a-1,4-D-glucan glucohydrolase, EC 3.2.1.3; GA) is an enzyme which catalyses the release of 3-glucose from the nonreducing ends of starch and related oligosaccharides.

GA is used in, and defines the rate limiting step of, the commercial conversion of starch to high glucose syrups which may be converted to fructose syrups by glucose isbmerase, or used in fermentations to produce WO 98/03639 PCT/US97/ 12 9 8 3 -26ethanol. GA is used industrially at 550-600C; at higher temperatures the enzyme is rapidly and irreversible inactivated. Therefore, a GA variant with increased thermostability would be advantageous industrially to decrease reaction times and/or to increase solids concentrations.

Previous work has shown that the natural stability of oligo 1, 6 -glucosidase [Suzuki et al., 1987] and pullulanase [Suzuki et al., 1991] can be positively correlated to the mole percent proline present in the protein, and a general rule for protein stability has been proposed [Suzuki, 1989]. This work has been extended to show that bacteriophage T4 lysozyme [Matthews et al., 1987] and Bacillus cereus ATCC 7064 oligo 1,6 glucosidase [Watanabe et al, 1994] can be stabilized by engineering proline into selected sites thereby decreasing the protein's conformational entropy of unfolding.

Three sites (Ser30, Asp345 and Glu408Pro) were selected for proline substitution based on structural and evolutionary considerations. Mutations at these sites were constructed using the cloned A. awamori gene [Innis et al, 1985] and the proteins were expressed in Saccharomyces cerevisiae [Cole, et al, 1988]. The stability of the mutant proteins was measured by their resistance to irreversible thermal inactivation at various temperatures. As shown herein, the mutation increased. However, unexpectedly the Glu408-.Pro mutation decreased and the Asp345-Pro mutation did not significantly alter GA stability.

Site-directed mutagenesis: Site-directed mutagenesis was performed as described herein above.

The following oligonucleotides were used as mutagenic primers: CAGAGTCCGCGCCCGGCACCCAAGCACCGTC (SEQ ID No:3),

AAGTCCAGCGACACAGGTGTGACCTCCAACGAC

(Asp345-Pro) (SEQ ID No:4) and CGAGCGGAAAGCTGC WO 98/03639 PCTUS97/12983 -27- GGGCCATCAGACTTGTC (Glu408-Pro) (SEQ ID Selection of sites for proline substitution: Based on the nearly identical catalytic domain of A.

awamori var X100 GA whose structure is known [Aleshin et al, 1992] three sites for substitution were chosen, which met the following criteria: 1) Ramachandran angles were within allowed values for proline [Ramachandran et al., 1963]. For this work the 0 and angles at the substituted site were restrained to the broad range 0 -900 to -400, V 1200 to 1800 or 0 -900 to -400, 0 -500 to 100. 2) Residues were highly solvent exposed, since mutation of residues in the core of the enzyme were thought to be more likely to decrease the enzyme's catalytic efficiency. 3) Residues didn't participate in hydrogen bonding with other amino acids. Additionally, based on sequence alignments with GA's from other organisms [Coutinho and Reilley, 1994b] only residues which met the above structural criteria and were not well conserved were selected for mutation. Ser30 could be aligned with proline in GAs from Hormoconis grisea var thermoidea and H. resiae GamP [Coutinho and Reilly, 1994b], which made it particularly attractive for proline substitution.

RESULTS

Specific activity None of the proline substitution mutations significantly altered enzyme specific activities of wild-type and mutant GA's at 50 0 C and pH 4.5. This suggests that these mutations did not significantly alter the enzyme's structure around the active site or alter its interaction with substrate.

Irreversible thermal stability Wild-type and mutant GA's were subjected to thermal inactivation at pH 4.5 as described in the WO 98/03639 PCTIUS97/ 12 9 8 3 -28experimental protocols. Semilogarithmic plotting of the percent residual activity against inactivation time yielded inactivation rate coefficients Figure 1 shows the relationship between temperature and kd for wild-type and mutant GA's. Based on these data, activation energies for thermal inactivation were calculated using transition state theory and melting temperatures the temperature at which the enzyme was 50% inactivated after 10 minutes were computed (Table As can be seen, the Glu408*Pro mutation greatly decreased, the Asp345Pro mutation did not significantly alter and the Ser30-Pro mutation increased GA stability.

It should be noted that although Table 1 shows that the Asp345-*Pro mutant GA demonstrated slightly increased AG' and Tm, these changes are generally not significant or that the Asp345-Pro mutant GA is more stable than wild-type since the kds for this mutant enzyme at two well separated temperatures (656 and 75 0 C) are essentially indistinguishable from wild-type (Figure 1).

The proline substitution mutations had different thermostabilities when measured by their resistance to irreversible thermal inactivation. When compared to wild-type GA, Glu408-Pro decreased, Asp345-Pro did not significantly alter and Ser30-Pro increased

GA

stability (Figure 1 and Table 1).

Glu408-Pro destabilized GA. As was first suggested by Schimmel and Flory [1968] and has been expanded by others [MacArthur and Thornton, 1991; Hurley et al, 1992] proline not only restricts the values for the site at which it exists, but also the 0,0 values of the preceding residue. These reports suggest that the values for the residue preceding proline should be restricted to approximately -1800 to -550 and 4= 550 to 1800 or 0 -1800 to -55 0 and WO 98/03639 PCT/US97/12983 -29- V= -300 to -700 for all residues in Xaa-Pro except for Xaa-Gly, for which the preceding still applies, but is extended to include 0=45 0 to 1800. In the published

A.

awamori var. X100 catalytic domain structure [Aleshin et al., 1992], Asp408 (0=-650, V=1460) which aligns with Glu408 in A. awamori GA, has 0,0 values within ranges acceptable for proline. However, the preceding residue Gly407 (0=800, =-5 0 has 0,V outside acceptable ranges for positions preceding proline. It is not surprising then, that the Glu408- Pro destabilized GA. Additionally, X-ray crystallography suggests that position 408, in the closely related A.

awamori var. X100 GA 2 lies within a 3-strand; a site not well suited for.proline substitution.

Asp345 (0 -650, V -260) and the preceding Thr344 (0 -1160, V =1780) have angle values lay well within allowed values for proline substitution at position 345. However, the Asp345-Pro mutant GA did not demonstrate stability significantly different from wild-type GA. This is particularly unexpected since position 345 lies at the N-terminus of an a-helix 2 a position previously shown to be particularly favorable for proline substitution [Watanabe et al, 1994].

-490, 1300) is preceded by Val29 -1270, 0 460) both of which have acceptable 0,0 angle values except Val29 P 460 which is slightly smaller than ideal for proline substitution at position In summary, when expressed in Saccharomyces cerevisiae, Glu408-Pro greatly decreased, Asp345- Pro, did not significantly alter and Ser30-Pro strongly stabilized the enzyme. The Ser30-Pro mutant GA showed a significantly decreased rate of irreversible thermal inactivation when analyzed between 650 and 77.5 0

C

without decreased enzyme activity. At 65 0 C a 1.7-fold decrease in thermal ihactivation rate coefficients was WO 98/03639 PCT/US97/129 8 3 seen and the activation energy for thermal inactivation was increased by 1.6 kJ/mol relative to wild-type

GA.

EXAMPLE 2 ENGINEERED DISULFIDE

BONDS

The following example is an exemplar of the methods and procedures that are used in the analysis of an individual mutation of a glucoamylase. The process of GA thermoinactivation is thought to be dominated by formation of enzymes with incorrect conformation [Munch and Tritsch, 1990]. Previous work supported this hypothesis. Site-directed mutagenesis has been used to eliminate sites of deamidation and peptide hydrolysis Chen et al., 1994 The corresponding mutations Asnl82->Ala and Asp257 Glu had reduced irreversible thermoinactivation rates at pH 4.5 below 70 0 C but increased rates above 70 0 C. Thus GA thermoinactivation is predominantly caused by "scrambled" structures rather than by deamidation and peptide hydrolysis.

Furthermore, mutations Glyl37-Ala, Glyl39-Ala and Glyl37/139-Ala/Ala, made to reduce helix flexibility, showed increased thermostability up to 75 0 C (Chen et al., 1996) apparently by slowing down the formation of incorrect structures.

To improve protein thermostability by preventing formation of incorrect structures, several strategies have been proposed including introducing covalent linkage such as disulfide bonds (Perry and Wetzel, 1984; Wetzel, 1987; Matsumura et al., 1989, Clarke and Fersht, 1993) There are a total of nine cysteine residues in A.

awamori GA, eight of which form disulfide-linked pairs, which are assumed to enhance the folding and stability of GA, residues 210 and 213, 262 and 270, 222 and 449 [Aleshin et al., 1992] and 509 and 604 [Williamson et al., 1992b] In this Example, additional disulfide WO 98/03639 PCT/US97/12983 -31bonds are introduced into GA to explore the effect on thermostability and catalytic activity. Two engineered disulfide bond mutants designated A27C/N20C (abbreviated S-S) and A471C/T72C were constructed. The new disulfide bond formed by A27C/N20C connects the Cterminus of helix 1 (Asn20) and a turn where residue Ala27 is located, while A471C/T72C bridges the Nterminus of helix 3 and the end of the highly O-glycosylated belt region together. The disulfide bonds are formed spontaneously after fermentation and have different effects on GA thermostability and catalytic activity.

Site-Directed Mutagenesis: Site-directed mutagenesis was performed as described herein above.

Oligonucleotide primers used are: 5'-CGT ACT GCC ATC CTG TGT AAC ATC GGG GCG GA-3' (N20C, AAT-TGT) (SEQ ID No:6), 5'-ATC GGG GCG GAC GGT TGT TGG GTG TCG GGC GCG- 3' (A27C, GCT-TGT) (SEQ ID No:7), 5'-CGA AAT GGA GAT TGC AGT CTC-3' (T72C, ACCTGC) (SEQ ID No:8), 5'-G AGT ATC GTG TGT ACT GGC GGC ACC-3' (A471C, GCTTGT) (SEQ ID No:9), with the underlined letters indicating the nucleotide mutations.

SDS-PAGE and Thio-titration: SDS-PAGE was carried out using 0.75 mm thick 10% polyacrylamide gels following the method of Garfin [1990]. For thio-titration, GA at 2 mg/ml concentration was denatured by boiling in denaturing solution containing 2% SDS, 0.08 M sodium phosphate (pH 8.0) and 0.5 mg/ml EDTA [Habeeb, 1972] with or without 50 mM DTT [Pollitt and Zalkin, 1983] for 10 min. The denatured GA (reduced or non-reduced) was concentrated using Centricon 30 concentrators (Amicon, MA, USA) and the reduced GA was applied to Bio-spin 30 chromatography columns (Bio-Rad, CA, USA) pre-equilibrated with denaturing solution to remove DTT. The resulting solution as well as the non-reduced denatured GA sample were divided into two portions.

WO 98/03639 PCT/US97/ 12 9 8 3 -32- One portion was used for a protein concentration assay and the other portion was assayed for thio reduction by mixing with 4 mg/ml DTNB in denaturing solution with a 30:1 volume ratio, followed by incubation at room temperature for 15 minutes, and absorbance measurement at 412 nm with a molar absorptive value of 13,600 M-cm- [Habeeb, 1972].

GA Activity Assay: As described herein above, maltose was used as substrate in enzyme kinetics studies, with concentrations ranging from 0.2 Km to 4 K at 35°C and pH 4.5 as described previously [Chen et al., 1994b]. Kinetics parameters were analyzed by the program ENZFITTER. In residual enzyme activity assays, the conditions are the same as in the enzyme kinetics studies except that only one concentration of maltose is used as substrate. Specific activity assays were carried out with 4% maltose as substrate at 50 0

C

and pH 4.5. One unit (IU) was defined as the amount of enzyme required to produce limol glucose per min under the conditions of the assay. To compare the temperature optima of catalytic activities of wild-type and mutant GA, activities were assayed at pH 4.5 with 4% maltose as substrate at different temperatures.

Irreversible Thermoinactivation: As described herein above, purified wild-type or mutant GA proteins were incubated at five different temperatures from to 75 0 C at 2.5 0 C intervals at 40 pg/ml in 0.05 M NaOAC buffer (pH At six different time points, aliquots of the incubating enzyme were removed, quickly chilled on ice, stored at 4 0 C for 24 hours, and subjected to residual activity assay. The irreversible thermoinactivation of GA obeyed first-order kinetics [Chen et al., 1994b]. Thermoinactivation rate coefficients, kd were determined as described previously [Chen et al., 1994b].

Computer Modeling and Three-dimensional View of WO 98/03639 PCT/US97/12983 -33- Mutated Residues: The candidate residues of A.

awamori GA to form disulfide bond were modeled with the crystal structure of A. awamori var. X100 GA [Aleshin et al., 1992] (Igly in the Brookhaven Protein Data Bank) as reference by the SSBOND program (Hazes and Dijkastra, 1988) installed in a DEC 3100 workstation.

Selection of Mutation Site: Residues Asn20, Ala27 and Thr72, Ala471 were chosen to be replaced with cysteine. After the analysis of crystal structure of A. awamori var. X100 GA [Aleshin et al., 1992] by the program SSBOND, 132 pairs of residues were found that could potentially be sites for a disulfide bond. Pairs containing glycine were discarded on the assumption that glycine may be required for flexibility at that site. Also, the residues involved in hydrogen bonds and electrostatic interactions were eliminated.

Residues 20 paired with 27 as well as 72 paired with 471 were chosen as candidates for disulfide bond formation according to the geometrical analysis. Amino acid sequence alignment among related GAs showed that there is a disulfide bond between position 20 and 27 in Neurospora crassa [Coutinho and Reilly, 1994b], which suggested that introducing disulfide bond between position 20 and 27 would not cause unfavored interactions there in A. awamori GA. Furthermore, the 20/27 disulfide bond would link the C-terminus of helix 1 and the conserved S1 fragment of GA involved in substrate binding [Coutinho and Reilly, 1994a] to form a loop, near another loop very critical for catalysis containing Trp 120, a residue involved in substrate binding (Sierks et al., 1989]. Therefore, the proposed 20/27 disulfide bond was expected to stabilize GA by keeping the correct conformation for catalysis and substrate binding.

Another further candidate for a disulfide bond pair was between positions 471 and 72. This disulfide WO 98/03639 PCTIUS97/129 8 3 -34bond would link the N-terminus of helix 3 and the end of the 30-residue (440-470) highly O-glycosylated belt region to form a loop. This disulfide bond also would make an additional linkage between the catalytic domain and the O-glycosylated linker. This O-glycosylated linker has been proved to be important for GA thermostability by limiting the conformational space available to the GA unfolded peptide [Semimaru et al., 1995 and Williamson et al., 1992]. This disulfide bond could have a globe effect on the thermostability of GA because of this linkage. The side chain -OH group of Thr72 in A. awamori var. X100 GA [Aleshin et al., 1992] is hydrogen bonded to the main chain N atom of Asp73.

In A. awamori GA, however serine is found at residue 73 in place of Asp. It is possible that the hydrogen bond between residues 72 and 73 does not exist in A. awamori GA, and therefore replacing Thr72 with Cys will not disturb this interaction. This hydrogen bond is apparently not critical for GA since Thr72 is replaced by Ala, Lys or Val in other GAs [Coutinho and Reilly, 1994b] Engineered Disulfide Bonds Were Formed Spontaneously: After GA purification, the engineered disulfide bonds were found to be formed spontaneously o by the following two approaches. First, the mutant A471C/T72C has faster mobility than wild-type during SDS-PAGE under non-reducing conditions, suggesting that an additional disulfide bond forms a new loop retarding the migration. The possibility that a truncated enzyme was formed in this case was eliminated by DNA sequencing of the mutant cDNA and MALDI analysis. The MALDI data showed that the mutant GA had the same molecular weight as wildtype GA. Mutant A27C/N20C has the same migration as wild-type GA, which may be because the additional loop caused by the engineered disulfide bond is too small WO 98/03639 PCTUS97/12983 (seven residues) to affect migration.

Second, the new disulfide bonds were demonstrated by thio group titration. Comparing the numbers of free thio groups before and after the treatment of reducing reagent DTT, the total disulfide bonds in mutant and wild-type GA were deduced as reported in Table 2.

Wild-type, A27C/N20C, and A417C/T72C GA have in total 8.6, 10.9, and 10.4 free thio groups respectively according to the [SH]/molecule ratio in the presence of reducing reagent DTT (Table In the absence of DTT the numbers are 0.9, 0.9 and 1.3, respectively (Table This suggested that the number of disulfide bonds among wild-type, A27C/N20C and A471C/T72C are 4, 5 and respectively. Therefore, the introduced cysteine residues formed disulfide bonds instead of remaining free thiols.

Enzymatic Activity and Optima Temperature of Catalysis: The enzymatic properties of double mutations A27C/N20C and A471C/T72C were not changed compared with wild-type at 35 0 C and 50 0 C as shown in Table 3, while single mutations had significant reduced activity. Mutant A27C/N20C and A471C/T72C had specific activities at 50 0 C and kinetic parameters at 35 0 C very close to wild-type GA (Table The single mutant A27C had slightly increased Km but the same kca value as wild-type GA, and thus a reduced kc,/Km ratio of Mutant N20C had the same Km but both a decreased kca and kcac/K, ratio and a decreased specific activity at 50 0 C of more than Irreversible Thermoinactivation of GA: The irreversible thermoinactivation of wild-type and mutant GA was studied at 65 0 C, 67.5 0 C, 70 0 C, 72.5 0 C and 77.5°C with first-order irreversible theremoinactivation coefficients k d shown in Figure 2. Mutants A27C, A27C/N20C and A471C/T72C had smaller kd values than did wild-type GA within the measured temperature range, WO 98/03639 PCT/US97/12983 -36which means the activity decayed more slowly than wild type, while mutant N20C had greater kd value than wildtype at all temperatures except 75°C, which means that decayed faster than wild-type.

Table 4 shows the activation enthalpy entropy and free energy of unfolding at 65 0

C

and 75°C of wild-type and mutant GAs, calculated according to transition-state theory. The enthalpies of N20C and A27C/N20C decreased by 42 and 24 KJ/mol respectively, while no significant change occurs for A27C and A471C/T72C. Mutants N20C and A27C/N20C had decreased entropy of 115 kJ/mol and 75 kJ/mol respectively, while entropy of mutants A27C and A471C/T72C showed no significant change. Mutant A27C and A471C/T72C had a slightly higher AG' than wild-type GA at 65 0 C and 75°C kJ/mol), while the AG' of A27C/N20C was higher than that of wild-type by 1.5 and 2.2 kJ/mol at 65 0 C and 75 0 C respectively. Mutant had a decreased AG' by 3.0 and 1.8 kJ/mol at 65 0 C and 75 0 C, respectively, compared with wild-type

GA.

Therefore, the engineered disulfide bond mutant A27C/N20C significantly increased GA thermostability compared with wild-type GA while the single mutants produced either a slight increase (A27C) or a slight decrease (N20C) in thermostability. The other disulfide bond mutant had the thermostability identical to wild-type

GA.

EXAMPLE 3 MUTATION A27C/N20C IN COMBINATION WITH OTHER MUTATIONS In previous studies applicants have constructed the thermostable mutants G137A [Chen et al., 1996] and S436P (Li et al., 1996), which have the potential to be combined and improve thermostability additively. In this Example, these mutations are combined with each other and with A27C/N20C Example 2) to test their WO 98/03639 PCT/US97/12983 -37effects (cumulative/additive) on thermostability and GA activity.

Enzymatic Activity and Optima Temperature of Catalysis: The combined mutants A27C/N20C/G137A and A27C/N20C/S436P had increased specific activity while mutant G137A/S436P had similar specific activity to wild-type GA (Table The double mutants A27C/N20C and A471C/T72C as well as the combined mutant A27C/N20C/G137 had changed optimal temperatures for catalysis.

Relative activity assays at temperatures from to 74C (Figure 3) showed that wild-type, mutant A27C/N20C and A471C/T72C had the highest activity at 71 0 C, 720C and 72.5 0 C, respectively. From 600C to 67.50C, mutant and wild-type GA had very similar activities. However, when the temperature was above 700C, their relative activities differed substantially.

Mutants A27C/N20C and A27C/N20C/G137A had higher activity than wild-type consistently from 70°C to 760C with a peak at 72.5C, while mutant A471C/T72C had activity lower than wild-type from 70 0 C to 710C and 73°C to 74C but higher at 72°C which is its optimal temperature. Thus mutant. GAs A27C/N20C, A471C/T72C and the combined mutant A27C/N20C/G137A had increased temperature optima above wild-type GA by 1.5 0

C.

Irreversible Thermoinactivation of GA: The irreversible thermoinactivation of wild-type and mutant GA was studied at 65 0 C, 67.5 0 C, 700C, 72.5 0 C and 77.5 0

C

with first-order irreversible theremoinactivation coefficients kd shown in Figure 2. Mutants A27C, A27C/N20C and A471C/T72C, A27C/N20C/G137A, A27C/N20C/S436P and G137A/S436P had smaller kd values than did wild-type GA within the measured temperature range, which means the activity decayed more slowly than wild type, while mutant N20C had greater k, value than wild-type at all temperatures 'except 75C, which WO 98/03639 PCT/US97/129 83 -38means that N20C decayed faster than wild-type.

Table 4 shows the activation enthalpy entropy and free energy of unfolding at 65 0

C

and 75°C of wild-type and mutant GAs, calculated according to transition-state theory.

The helix flexibility mutant G137A showed additive thermostability when combined with either S436P or A27C/N20C. The combination S436P with A27C/N20C did not show additivity.

EXAMPLE 4 FURTHER STUDIES WITH COMBINED

MUTATIONS

To further investigate whether individual stabilizing mutations can cumulatively stabilize Aspergillus awamori glucoamylase mutant enzymes were constructed containing combinations of thermostabilizing mutations. Previous work has shown that the following mutations stabilize GA as demonstrated by decreased irreversible thermal inactivation rates when inactivated in the absence of carbohydrate: Ser30Pro (S30P; Example Glyl37-Ala (G137A), and Asn20-Cys/Ala27-Cys (which creates a disulfide bond between residues 20 and 27 and is therefore noted as S-S for convenience; Example To investigate whether individual stabilizing mutations can cumulatively stabilize GA, additional combined mutant enzymes were prepared utilizing the these three mutations.

Site-directed mutagenesis: The S-S/S30P/G137A combined mutant was constructed using the oligonucleotide listed above and a single stranded

DNA

template derived from a pBluescript II vector with a 1.7 kb XhoI-BamHI DNA fragment coding for the GA catalytic domain which already contained mutations conferring the S30P and G137A amino acid substitutions.

The presence of the individual mutations was confirmed WO 98/03639 PCT/US97/12983 -39by sequencing and each mutated GA gene fragment was subcloned into YEpPM18 [Cole et al., 1988] and transformed into S. cerevisiae.

Thiol analysis: 10 nmol of wild-type, S-S/S30P and S- S/S30P/G137A mutant GAs were incubated in 0.2 mM dithiobis(2-nitrobenzoic acid), 6 M GdnHC1, and Tris, pH 8 in duplicate [Fierobe et al., 1996]. The thiol concentration was calculated from a standard curve established using 0-30AM cysteine.

Irreversible thermal inactivation: Wild-type and mutant GAs were subjected to thermal inactivation at six or seven temperatures between 650 and 80 0 C at 2.5 0

C

intervals in duplicate. Following 24 hours at 4 0 C, the residual activities of the inactivated samples were analyzed at 35 0 C along with a corresponding sample which had not been inactivated [Chen et al, 1996].

Saccharification analysis: Saccharifications were performed in duplicate using stirring heating blocks and tightly sealed vials to prevent evaporation. Eight Ag/ml of wild-type and mutant GAs were assayed using 28% Maltrin DE 10 maltodextrin in 0.05 M NaOAc pH as substrate. At various times, sample was removed, diluted appropriately in 0.05 M NaOAc pH and the reaction stopped by adding 100pl of diluted sample to 40 Al of 4.0 M Tris-Cl, pH 7.0. The glucose concentration was determined by a glucose oxidase/o dianisidine assay [Banks and Greenwood, 1971].

RESULTS

Enzyme activities Table 5 shows the specific activities of the wildtype and mutant GAs at 50 0 C and pH 4.5 using maltose as substrate. None of the mutant GAs demonstrated reduced enzyme activity and the S30P/G137A and S-S/S30P/Gl37A mutants were somewhat more active than wild-type at 50 0 C. To further investigate this observation, the activities of these mutant enzymes were assayed at WO 98/03639 PCT/US97/12983 various temperatures between 350 and 680C (Figure 4).

The S30P/G137A and S-S/S30P/G137A mutant GAs were more active than wild-type at all temperatures examined.

Thiol analysis The formation of a disulfide bond between positions 20 and 27 in the Asn20-Cys/Ala27 Cys mutant GA has been confirmed (Example Table 6 shows the results of thiol analysis for the combined mutants Sand S-S/S30P/G137A. A. awamori GA has one free cysteine at position 320. The combined mutant GAs show slightly higher thiol content per molecule than wildtype which may reflect less than complete disulfide bond formation between positions 20 and 27. However, if the disulfide bond was completely unformed, the [SH]/protein would be expected to rise to approximately three with the addition of two free cysteine residues.

Therefore, we conclude that the disulfide bridge is formed at 70-80% the expected theoretical yield for complete formation.

Irreversible thermal inactivation Wild-type and mutant GAs were subjected to thermal inactivation at pH 4.5 between 650 and 80 0

C.

Semilogarithmic plotting of residual activity versus inactivation time yielded inactivation rate coefficients Figure 5 shows the effect of temperature on kd for wild-type and mutant GAs. As can be seen, the combined mutants are significantly more stable than the individual mutant enzymes.

Additionally, the temperature at which the enzymes were 50% inactivated after 10 minutes (Tm) was calculated by extrapolation from the thermal inactivation plots and transition state theory was used to calculate activation energies for thermal inactivation Table 7 shows the changes in AG' (AAG') and Tm for the combined mutant GAs relative to wild-type GA. These data clearly demonstrate that combining the individual WO 98/03639 PCTIUlS97/12983 -41stabilizing mutations can cumulatively stabilize the enzyme.

Saccharification analysis Figure 6 shows the results of saccharification analysis at 550 and 65 0 C for wild-type, S30P/G137A and S-S/S30P/G137A GAs using the industrial DE maltodextrin substrate Maltrin M100 (28% w/v) from Grain Processing Corporation. Complete conversion of 28% w/v DE 10 maltodextrin to glucose would result in a 1.71 M glucose syrup however, previous saccharification analyses in our laboratory have demonstrated that wildtype GA results in approximately 90% theoretical maximum glucose yield at 55 0 C (not shown). At 55°C no significant difference in glucose production was observed between the wild-type and mutant enzymes.

However, at 65 0 C the mutant GAs produced 8-10% more glucose than wild-type although none of the enzymes tested produced as much glucose as at 55 0 C probably due to thermal inactivation at the elevated reaction temperature.

In summary these data show that the S30P/G137A double mutant enzyme was more stable than either single mutant GA when analyzed for resistance to irreversible thermal inactivation between 65 0 C and 80 0 C. The S- S/S30P combined GA mutant was also more stable than either the S30P or the S-S mutant GAs. The S- S/S30P/G137A combined mutant was the most stable GA variant constructed, particularly at temperatures above 0 C when inactivated in a buffer system lacking monoor polysaccharides. Saccharification analysis showed that the mutant enzymes performed better at elevated temperatures than wild-type GA. Importantly, none of the combined mutant GAs showed decreased enzyme activity when analyzed at 50 0

C.

Discussion Sites of mutation WO 98/03639 PCT/US97/129 8 3 -42- As described in Example 2, the mutations and Ala27>Cys form a disulfide bond between the Cterminus of c-helix one and an extended loop between ahelices one and two. S30P and G137A were designed to stabilize the enzyme by reducing its conformational entropy of unfolding and are the most stabilizing in a series of proline substitution (Xaa-Pro) and Gly-Ala mutations respectively. Ser30 is located at the second position of a type II /-turn on an extended loop between a-helices one and two and Gly137 is located in the middle of the fourth a-helix.

It is of particular importance to note the positions of the S30P and the disulfide bond forming mutations. The disulfide bond is formed between positions 20 and 27; relatively close to position The fact that both the disulfide bond forming mutations and the S30P stabilize GA suggests that this region of the enzyme is critical for irreversible thermal inactivation and may represent a region of local unfolding important for thermal inactivation.

Additionally, previous investigators have suggested that a disulfide bond should not be engineered within four amino acids of a proline in primary sequence [Balaji et al, 1989]. This Example demonstrates that this rule is not absolute since thiol analysis showed that the disulfide bond was formed in the S-S/S30P and S-S/S30P/G137A combined mutants and thermal inactivation studies showed the stabilizing effects of the mutations were cumulative.

Cumulative stabilization Previous work by Applicants has shown that combining two stabilizing mutations does not necessarily stabilize GA [Chen et al, 1996]. The present study, however, demonstrates that combining stabilizing mutations, even mutations very close to each'other in the protein, can cumulatively stabilize WO 98/03639 PCT/US97/12983 -43- GA as measured by resistance to irreversible thermal inactivation.

The S30P/G137A mutant showed more than additive stabilization at low temperatures (650-70°C), but less than additive stabilization at high temperatures (77.5 0 -80 0 C) (Figure 5A and Table At 80 0 C the inactivation rate for the S30P/G137A combined mutant was nearly identical to the S30P individual mutant protein. This indicates that both regions are very important for low temperature thermal inactivation, but at high temperatures inactivation became governed by other processes.

It was somewhat surprising that combining the with the disulfide bond forming mutations resulted in cumulative stabilization. This is not only because the engineered disulfide bond is so close to the engineered proline as is discussed above, but also because both are targeting the same region of the protein (ie: the extended loop between a-helices one and two). It was expected that either the disulfide bond or stabilized this region maximally, and further stabilization at this site would not result in a functionally more stable enzyme. As can be seen in Figure 5B, this was not the case. Combining the mutations resulted in roughly additive stabilization at all temperatures examined between 650 and 80 0

C.

The S-S/S30P/G137A cbmbined mutant was no more stable than S30P/G137A GA at low temperatures 0 but was slightly more stable at higher temperatures (75 0 -80 0 C) (Figure 5C and Table 7).

Interestingly, the S-S/S30P GA is also more stable than S30P/G137A GA at high temperatures. Therefore, it appears that the introduced disulfide bond is particularly effective at stabilizing GA at high temperatures.

WO 98/03639 PCTUS97/1298 3 -44- EXAMPLE INDUSTRIAL

APPLICATION

To determine whether the thermal stabilizing mutations: S30P/G137A and S-S/S30P/G137A would enhance GA performance under industrial conditions wild-type and mutant enzymes were subjected to high temperature saccharifications (Figure Saccharification analysis showed that the mutant enzymes out-performed wild-type at 65 0 C but not at 55 0 C likely due to their increased stability.

Conclusion The S30P/G137A double mutant cumulatively stabilized GA as demonstrated by decreased irreversible thermal inactivation rates relative to either individual mutant enzyme when analyzed between 65 0 C and 0 C. Similarly, the S-S/S30P combined mutant also demonstrated cumulative stabilization. The S- S/S30P/G137A combined mutant was more stable than either of the "double" mutants, particularly at temperatures above 70 0 C. The S-S/S30P combined mutant had the same activity as wild-type and the S30P/G137A and S-S/S30P/G137A mutants increased enzyme activity by 10-20% when assayed between 350 and 68 0 C. The S30P/G137A and S-S/S30P/G137A mutant GAs decreased thermal inactivation rates approximately three fold relative to wild-type when inactivated in the presence of 1.71M glucose at 65oC. Additionally, at 55C no difference in glucose yield was observed between these mutant GAs and wild-type for the saccharification of the industrial substrate Maltrin M100, whereas at 65 0

C

the S30P/G137A and S-S/S30P/G137A GAs produced 8-10% more glucose than wild-type.

EXAMPLE 6 MUTATIONS WITH INCREASED

SELECTIVITY

Interactions between substrates and charged WO 98/03639 PCT/US97/12983 residues at subsites 1 and 2 of GA play a very important role in substrate specificity since the catalytic site is located between these sites.

Therefore mutations were designed and analyzed to determine residues within these regions where mutations would increase selectivity of the enzyme reaction. In addition, several mutations that were designed to have thermostability were also screened for selectivity, as well as mutations designed to increase the pH optimum.

Site-directed mutagenesis: Site-directed mutagenesis was performed as described herein above.

The following mutagenic oligonucleotide primers were synthesized at the Iowa State University Nucleic Acid Facility: 5'-GGT CTC GGT GAG CCC AGG TTC AAT GTC GAT-3' (Lysl08-Arg; SEQ ID No:10), 5'-GGT CTC GGT GAG CCC ATG TTC AAT GTC GAT-3' (Lysl08-Met; SEQ ID No:ll), GAC ACG TAC TGG AAC GGC AAC CCG-3' (Tyr312-Trp; SEQ ID No:12), and 5'-TAC CCT GAG GAC ACG TAC AAC GGC AAC GGC AAC TCG CAG GGC AAC CCG TGG TTC CTG TGC-3' (311-314 Loop; SEQ ID No:13), the underlined letters indicating the changed or added nucleotides.

Results Enzyme kinetics As shown in table 11, the kinetic parameters kca and KM for the hydrolysis of G 2 to G, as well as iG 2 in 0.05 M acetate buffer, pH 4.4, at 450C are given in Table 8. The 311-314 Loop mutant had kcat values 50-80% for all e-(1,4)-linked substrates and only 30% for iG,, KM values 50-75% for all substrates. The kcat values for Glyl37-Ala/Ser30Pro GA are 10-30% more, generally, than that of wild-type GA for all substrates. The KM values of Glyl37->Ala/Ser30>Pro GA are about half to twofold for all the a-(1,4)-linked substrates and essentially reached the wildtype level for iG 2 The ka values for the GA engineered to carry the triple mutation, S-S/Glyl37 Ala/Ser30Pro, ranged from 80 to WO 98/03639 PCT/US97/1 2 9 8 3 -46- 120% generally for all substrates, and the KM values are 30-80% for all substrates compared to wild-type

GA.

The kcat values for S-S GA are 85-110% for all substrates, and the S-S GA K, values are generally 110% for all substrates. However, the S-S GA values are 140% for G, and 190% for G 6 Values of kca/K, are 75-105%, 60-110%, 60-110%, and 60-120% for the Tyr312-Trp mutation, the combined Ser 3 0-Pro/Gly137-Ala double mutation, the combined S-S/Ser30>Pro/Gly137>Ala triple mutation, and the S-S engineered

GA,

respectively. The catalytic efficiencies for the 311- 314 Loop GA are 85-120% for all the a-(1,4)-linked substrates, and only 50% for iG 2 compared to wild-type

GA.

Table 8 shows the ratios of the catalytic efficiencies for G 2 to iG, for wild-type and mutant GAs.

GAs engineered with the 3 11-314Loop mutation and Lysl08-.Arg mutation have the highest (240%) and the lowest catalytic efficiencies for over a-(1,6)-linked substrates, respectively. The GAs engineered with the Tyr312-Trp and S-S mutations show and 20% increases for this ratio, respectively.

All other mutants had lower ratios, indicating poorer c-(l,4)-hydrolytic ability relative to c-(1,6)-hydrolytic ability than wild-type

GA.

Maltooligosaccharide hydrolysis GA engineered with the 3 11-314Loop mutation or with the S-S mutation had the highest average glucose yields (Figure The 3 11-314Loop GA had the lowest initial rates for glucose production 61%, and 82% compared to wild-type GA at 35, 45, and 55 0

C,

respectively) due to a specific activity only 60% that of wild-type GA (data not shown). Glucose concentrations decreased after reaching maximal values because of conversion to oligosaccharides.

Glucose condensation reactions WO 98/03639 PCT/US97/12983 -47-

IG

2 concentration profiles in 30% glucose condensation reactions at 35, 45, and 55"C were analyzed. GAs engineered with the LyslOBArg mutation had the highest and the 311-314Loop mutation as well as the S-S mutation the lowest equilibrium iG, concentrations at all three temperatures. Tyr312>Trp, Ser30-Pro/Glyl37-Ala, and S-S/Ser30-Pro/Glyl37-Ala GAs exhibited essentially the same equilibrium iG 2 concentrations as wild-type GA. For all the other engineered thermostable GAs tested, Ser436-Pro,

S-

S/Ser436-Pro, S-S/Glyl37->Ala, and Gly137- Ala/Ser436 Pro, all reached higher equilibrium iG 2 concentrations than did wild-type GA. Table 9 shows the initial rates of iG 2 formation in 30% glucose condensation reactions. -S-S and 311-314Loop mutant GAs have the lowest initial rates at all three reaction temperatures tested. Lysl08-Arg mutant GA showed the highest initial rates among all the mutant GAs tested at all three reaction temperatures. All the tested thermostable GAs except Ser30-Pro/Gly137-Ala and S- S/Ser30-Pro/Gly137-Ala had much higher initial rates than wild-type GA at 35 0 C, but they dropped to slightly higher or almost the same rate as wild-type GA at 55 0

C.

The specificity for a-(1,6)-linkage synthesis over a- (1,4)-linkage hydrolysis The ratio of the initial rate of iG 2 production in a 30% glucose condensation reaction to that of glucose formation in 30% DE 10 maltodextrin hydrolysis was calculated to estimate the selectivity for the synthesis of a-(1,6)-linked products over the hydrolysis of a-(1,6)-linked substrates. These iG 2 /glucose ratios and their relative ratios for wildtype and mutant GAs are given in Table 9. K108R and S- S mutants showed the highest and the lowest relative ratios among wildtype and all the mutant GAs at all reaction temperatures, respectively. Therefore, K108R WO 98/03639 PCT/US97/12983 -48had more specificity for a-(1,6)-linkages than linkages and S-S GA had more affinity for linkages than a-(1,6)-linkages. The 3 11-314Loop

GA

also showed very low relative ratios at these three temperatures.

EXAMPLE 7 ADDITIONAL SELECTIVITY MUTATION

ANALYSIS

Utilizing the methods as set forth herein above, additional mutations were screened for selectivity as shown in Table 10 and Figures 8 and 9.

Enzyme Kinetics: The kinetic parameters are seen in (kca, and Kj) for the hydrolysis of a-1,6-linked isomaltose and a-1,4-linked maltooligodextrins (DP2-7) at 450C and pH 4.4 are given in Table 10. Mutant Y175F was active. The kca and Km values were 83-141% and 106- 171%, respectively, that of wildtype for the different substrates tested and catalytic efficiencies were 69- 102% that of wildtype. Mutant R241K was also active.

Mutant S411G was highly active. The kca and K values were 93-129% and 83-203%, respectively, that of wildtype for the different substrates tested and catalytic efficiencies were 55-122% that of wildtype.

Mutant S411A had a similar catalytic efficiency ratio as wildtype. Mutants Y116W, R241K, and S411G had decreased catalytic efficiency ratios compared to that of wildtype

GA.

DE 10 Maltodextrin Hydrolysis: At 550C, the highest glucose yield was about 95% reached by engineered

GA

with mutant S411A at 216 hours compared to the wildtype yield of about 90% (Fig. All of the GAs, except S411A, reached their highest glucose yields rapidly.

The glucose yield of S411A slowly increased for an extended period of time. The initial rates of glucose production at 550C were 5 to 8 times higher that those at 356C.

WO 98/03639 PCT/US97/12983 -49- Glucose Condensation Reaction: Glucose condensation reactions were used to study the ability of wildtype and mutant GAs to synthesize isomaltose at high glucose concentrations (Figure The same concentrations of glucoamylases (2.64 pM) were used as in the hydrolysis of DE 10 Maltodextrin.

At 55 0 C, in spite of the different initial rates of isomaltose production for wildtype, R241K and Y175F, isomaltose production reached almost the same concentration at the last time point for these three mutant GAs (Figure indicating that the isomaltose production was close to equilibrium status. Isomaltose production for S411A and S411G was much lower than wildtype and almost linear as it was also at 35 0

C.

Unexpectedly, isomaltose production for Y116W had a different (lower) equilibrium status compared to wildtype. The initial rates of isomaltose production at 55 0 C were 5 to 7 times greater than those at 35 0

C.

R241K had a decreased initial rate of isomaltose production at 55 0 C compared to that of wildtype, and it also had a lower increase (about 5 times) in the initial rate of isomaltose production from 35 0 C to 0 C, compared to the wildtype increase (about 7 times). Y116W, Y175F, S411A and S41lG had increased initial rates of isomaltose production or about 7, 6, and 5 times, respectively from 35 0 C to 55 0

C.

Selectivity: The ratio of the initial rate of isomaltose production (from glucose condensation reactions) to that of glucose production (from hydrolysis of DE 10 maltodextrin) was calculated to evaluate selectivity for the synthesis of a-l,6-linked products versus the hydrolysis of a-1,4 linked substrates. This ratio represents the ability of a GA to synthesize isomaltose at a normalized level of DE maltodextrin hydrolytic activity.

Mutants Y175F, S411A and S411G had a decreased WO 98/03639 PCT/US97/129 8 3 ratio of the initial rate of isomaltose production to that of glucose production to that of glucose production by 12%, 35% and 56% at 35 0 C, respectively, and a decreased ratio by 24%, 60% and 62% at 55 0

C,

respectively, compared to wildtype. R241K had a very similar ratio to that of wildtype at both 35 0 C and 0

C.

EXAMPLE 8 MUTATIONS TO PROVIDE pH OPTIMIZATION Utilizing the methods as set forth herein above, additional mutations, S411G, S411A, S411C, S411H, S411D were screened for increased pH optimum as shown in Figure 10 and Tables 11 and 12.

Enzyme kinetics The kinetic parameters, and K, for the hydrolysis of a-l,4-linked maltose and maltoheptaose and a-l,6-linked isomaltose at 45 0 C and pH 4.4 are given in Table 11. Mutant S411G glucoamylase was highly active compared to wild-type, with an increased kca, and K, 13 30% and 11 59%, respectively, on the substrates tested. The catalytic efficiencies (kea/K) were 71 116% that of wild-type. Mutant S411A maintained 65 74% of wild-type catalytic efficiency with a slightly decreased kca r and a slightly increased Mutant S411C maintained 54 73% of wild-type catalytic efficiency with a decrease in both the kca

C

and Km values. Since mutant S411H and S411D had only about 6 12% of wild-type catalytic efficiency resulting from a seriously decreased kca and an increased the kinetic parameters for the hydrolysis of isomaltose were not determined. Only mutant S411H and S411D had large increases (5.5 to 7.5 kJ/mol) in the transition-state energy, A(AG), for the hydrolysis of maltose and maltoheptaose. The large increases of transition-state energy indicated that the introduction WO 98/03639 PCT/US97/12983 -51of histidine or aspartic acid into position 411 substantially destabilized the binding between GA and substrate in the transition-state.

pH dependence of GA activity The kinetic parameters, kcat/,K and kca,, of the hydrolysis of maltose by wild-type and mutant glucoamylases at different pH values were calculated from initial rates obtained at low (smaller than 0.2 K) and high (higher than 10 Kj) concentrations of maltose.

The effects of pH on the kcat/K, and kca, of maltose hydrolysis were used to determine the pK values (Table 12) of both the free enzymes and the enzyme-substrate complexes. Although wild-type GA had a higher catalytic efficiency (kcat/K) than all of the mutant glucoamylases at all of the pH values tested, mutants S411G and S411A had higher kca values than that of wild-type at some pH values. The uncomplexed and maltose-complexed S411H and S411D showed more narrow bell-shaped curves than that of wild-type.

The effects of pH on the hydrolysis of maltoheptaose by wild-type, S411G and S411A GAs were measured to further investigate the change of pK values and optimum pH of enzyme-substrate complexes using a long-length substrate. Surprisingly, not only S411G, but also S411A were highly active compared to wild-type at the optimum pH. Wild-type GA pKl values (ionization of the catalytic base) were 2.77, 2.11, and 2.6 for the free enzyme, the maltose-complexed form, and the maltoheptaose-complexed form, respectively. The pK 2 values (ionization of the catalytic acid) of wild-type were 5.80, 5.85, and 6.78 for the free enzyme, the maltose-complexed form, and the maltoheptaose-complexed form, respectively [Bakir et al., 1993, Hiromi et al., 1966, Sierks and Svensson, 1994]. Compared to wildtype, the S411G mutation increased the pKl of both the maltose-complexed form and the maltoheptaose-comolexed WO 98/03639 PCTUS97/12983 -52form by approximately 0'.6 units, whereas S411G had no effect on the pK2 of either enzyme-substrate complexes and only had a minor effect on the pKl and pK2 of the free enzyme. The combined effect of S411G on pKl and pK2 was an increased optimum pH of both the maltosecomplexed form and the maltoheptaose-complexed form by approximately 0.3 units.

The S411G mutation, however, had no effect on the optimum pH of the free enzyme. S411A and S411C had very similar effects on the pH dependence of maltose hydrolysis. S411A and S411C increased the pKj of the free enzyme and the maltose-complexed forms by 0.3 and 1.21 units, respectively. Surprisingly, S411A and S411C also increased the pK 2 of the maltosecomplexed form by approximately 0.5 units. In addition, S411A increased the pK, and pK 2 of the maltoheptaose-complexed form by 1.31 and 0.4 units, respectively. S411H increased the pKj of the free enzyme and maltose-complexed form by 0.33 and 1.47 units, respectively; however, it decreased the pK 2 of the free enzyme and the maltose-complexed form by 0.79 and 1.16 units, respectively. S411D increased the pKl of the free enzyme and the maltose-complexed form by 0.36 and 1.23 units, respectively. S411D also decreased the pK 2 of the maltose-complexed form by 0.32 units. For wild-type, S411G, and S411A GAs, the values of pK 1 pK 2 and pHop t for the maltoheptaose-complexed forms were higher than those of the corresponding maltose-complexed forms by approximately 0.5, 0.9 and 0.7 units, respectively. For S411G and S411A, the increases in pH optimum (compared to that of wild-type) obtained using the long-length substrate (maltoheptaose) were almost the same as that obtained using the short-length substrate (maltose).

All five mutants at position 411 showed a shift of 0.15 to 0.87 units in the optimum pH of the enzyme- WO 98/03639 PCTUS97/12983 -53substrate complex compared to wild-type (Table 12), mainly due to increased pK, values. Compared with other mutants, S411A was the best performing pH mutant.

S411A increased the optimum pH by 0.84 units while also maintaining a high level of both catalytic activity and catalytic efficiency (kcat/K,) The hydrolysis of maltodextrin The hydrolysis of 28% maltodextrin was used to study the pH dependence of GA activity at a high concentration of a long-length substrate. Maltodextrin is a mixture of maltodextrin with an average (and major) degree of polymerization of 10. The production of glucose by wild-type and S411A glucoamylases during the hydrolysis of maltodextrin 10 at 11 different pH values was determined, and used to calculate the initial rates of glucose production at different pH values (Figure 10). The production of glucose increased following a hyperbolic curve. S411A had higher initial rates of glucose production than wildtype when the pH values were above 6.6 (Figure Throughout this application, various publications are referenced by author and year and patents listed by number. Full citations for the publications are listed below. The disclosures of these publications and patents in their entireties are hereby incorporated by reference into this application in order to more fully describe the state of the art to which this invention pertains.

The invention has been described in an illustrative manner, and it is to be understood that the terminology which has been used is intended to be in the nature of words of description rather than of limitation.

Obviously, many modifications and variations of the present invention are possible in light of the above teachings. It is, therefore, to be understood WO 98/03639 PCT/US97/12983 -54that within the scope of the appended claims, the invention may be practiced otherwise than as specifically described.

Page(s)2> L 1 4 are claims pages They appear after the table listings WO 98/03639 PCTIUS97/12983 TABLE 1 Changes In AG1 and Trm for the mutant GAs relative to \vild-type.

AAG' ATm GA form (kJ/mol.)

(OC)

1.6 1.7 Asp345--*Pro 0.5 0.4 Glu4O8--*Pro -7.2 -6.7 TABLE 2 Summary of DTNI3-titratable sulfhydryl groups in wild-type and mutant

GA

with or without DTT reducing Enzymes [SHI/molecule No. of disulfide bonds* DFT±

DTT-

WT, 8.6 0.9 4 A27C/N20C 10.9 0.9 A471 C/T72C 10.4 1.3 No. of disulfide bonds= ([SH]/niolecule (DTT+)-[SH]/molecule (DTT-))/2 WO 98/03639 PCTIUS97/129 8 3 -56- TABLE 3 Catalytic properties of wild type and mutant GAs.

GA form W1.

A27C A27CIN20C A47 1C/T-72C 'A27C/N2OC /S4 36 A27CPN20C /GI37A G137A/S436P Specific Activity (lU/mg GA) 20.6±0.2 14.9±1.1 8.1±0.5 18,3±0.7 22 7± 1.5 22.5±1.8 2 4.2 8 25.0±0.9 0.7 2±0.03 0.86±0.11 0.70±0.05 0 .90±0.08 0.8 7±0.07 N/Dc T\I/r

N/D

mM,) cut 8.67±0.17 8 .02±0 .45 3 .97±0. 12 9.61±0.40 10.17±0.40

N/D

N/D

12.0 9.3 5.7 10.7 11.6

N/D

N/f) N/f) bProduced in shaking flasks standard error C not determined WO 98/03639 PCTfUS97/1 2983 -57- TABLE 4 Activation parameters for irreversible thcrmoinactivation of wild-type (WIT) and mutant GAs at pH GA form AW AS AG '(65 0 C) AG' (75 0

C)

(ki/mol) (J/rno[K) (ki/nol) (ki/mol) IT a 366± 1b769± 4 105.7 A27C 370±15 780±44 106.3 98.5 324±11 654±33 102.7 96.2 A27C/N2OC 342±16 694±46 107.2 100.2 A471C/T72C 365± 9 768±26 106.0 98.3 A27C/N200 352± 6 724±18 107.9 100.7 /S436P A27C/N20C 362± 1 75 1± 2 108.4 100.9 /G137A G137A1S436P 362±20 752±57 107.7 100.2 S436P C 351± 8 723±24 106.2 99.0 G I37Ad 330± 6 661±17 106.5 99.9 b produced by shaking flask standard error Li ef al., 1996 Chen el al., 1996 WO 98/03639 PCT/US97/129 8 3 -58- Table 5. Specific activities of wild-type and mutant GAs Specific activity' GA form (lU/mg) (IU/mg) Wild-type 21.1+0.1 S30P/Glyl37A 24.0+1.2 21.2+0.5 S-S/S30P/G137A 24.5 0.2 a Standard deviation resulting from three or more assays Table 6. Thiol analysis of wild-type and mutant GAs GA form [Protein] [SH] (IM)a [SH]/[Proteinj' Wild-type 10 8 0.8 10 11 1.1 S-S/S30P/G137A 10 13 1.3 Average of duplicate analysis Table 7. Changes in free energies for thermal inactivation (AAG4) and temperatures at which the enzyme is 50% inactivated after 10 minutes (ATm) relative to wild-type

GA

AAG*3 GA form GA form (kJ/mol) ATm b G137A 1.6 1.7 SSd1 0.8 1.2 A 1.2 1.4 S30P/G137A 3.5 3.2 S-S/S30P/G137A 4.4 3.9 Calculated at 65 0

C

b From Allen et. al.

8 SFrom Chen et. al.

6 d From Li et. al.

7 Kinetic parameters of wild-type and mutant GAs for hydrolysis of maltooligosaccharides DP 2-7 (G 2 -G7) at 45 0 C in 0.05 M acetate pH 4.4 Glucoamylase G2 G 3

G

4 G5 G6 G7 Wild- type kcat (s" 1 KM (mM) kcat/KM 1 m1) Lysl08Arg kcat (s" 1 co KM (mm) W kcat/KM (s mM' I (A G) b (kJ mol" 1 Tyr312Trp kcat (s" KM (mM) kcat/KM m41) (AG) (kJ mol"I) 300Loop kcat (s"1) KM (m) kcat/KM (s'mm A (AG) (kJ mol 1) 18.6 0 .4a 1.09 0.08 17.1 0.9 17.3 0.5 1.52 4 0.11 11.4 0.6 0.92 50.8 0.353 144 4 32.6 0.570 57.2 2.10 36.8 0. 343 107 6 0.67 17.2 0. 940 18.3 -0.16 14.7 0.738 20.0 -0.35 0.3 0.059 0.90 0.6 67.5 1.9 61.5 0.33 65.9 1.2 81.5 1.8 0.013 0.239 0.017 0.094 0.002 0.098 0.007 0.136 0.009 282 13 653 10 671 36 599 f 27 0.9 46.6 1.6 51.7 1.4 55.2 1.4 86.2 3.1 0.038 0.383 0.029 0.307 0.019 0.276 0.016 0.481 0.031 2.5 122 5 168 6 200 8 179 6 1.91 3.08 2.75 2.74 0.9 50.7 0.9 50.7 0.8 56.0 0.8 63.3 0.6 0.028 0.193 0.010 0.100 0.006 0.108 0.005 0.103 0.003 262 9 508 22 519 20 617 1 0.17 0.57 0.58 -0.07 0.6 34.1+ 0.8 43.0 0.6 41.4 0.8 41.9 0.7 0.019 0.114 0.008 0.072 0.004 0.064 0.005 0.083+0.005 300 17 598 28 642 47 506 -0.14 0.20 0.10 0.38 3.0 77.9 4 2.2 77.7 1.6 77.0 2.2 80.3 2.2 0.010 0.261 0.020 0.175 0.011 0.204 0.017 0.151 0.013 3.1 299 16 444 21 377 23 533 37 -0.15 1.02 1.52 0.31 0.3 25.9 .0.055 0.234 1.2 111 7 0.60 Ser30Pro/Gly137Ala kcat (s 1) 25.0 1.1 KM (nM) 1.62 0.11 kcat/KM mM 1 15.5 1.2 A(AG) (kJ mol 1) 0.27 50.2 0.596 84.2 1.42 SS/Ser3OPro/Glyl37Ala kcat 23.0 0.9 Km (mm) 1.66 0.07 kcat/KM (s-Imm- 1) 13.9 0.9 A (AG) (ki mol 0.55 kcat 20.7 0.6 KM.*(mm) 1.16 0.10 kcat/KM 1 mM 1) 17.8 1.1 42 .1 1.0 0.470 0.032 89.6 4.2 1 .26 72.0 2.1 72.2 1.0 79.5 1.7 81.5 1.4 0.236 0.019 0.172 0.007 0.157 0.011 0.198 0.010 305 17 420 13 505 26 410 -0.21 1.16 0.75 1.00 40.8 0.9 72.1 1.3 0.394 0.025 0.217 0.011 104 5 331 12 0.88 -0.42 76.5 0.8 76.4 2.1 71.8 0.6 0.132 0.005 0.184 0.015 0.114 0.003 579 16 414 26 632 0.32 1.28 -0.14 a Standard error bChange of transit ion- state energy A (AG) RTlfl[ (kcat/K.M) mut/ (kcat/KM) t] WO 98/03639 PCT/US97/12983 -61- TABLE 9 Initial rates of glucose and isomaltose productions in the hydrolysis of 30% maltodextrin M100'and 30% glucose condensations, respectively, and their relative ratios for wild-type and mutant glucoamylases at 35 C, 45 C, and 55 C.

Initial rates Ratios Enzymes Glucosea(G1) Isomaltoseb(iG 2 Ratios Relative 3 (jg/mL-h)xl0 3 (iG 2 /Gl)xl06 ratios 350C Wild-type 21.5 +0.6 c 289 5 13.5 1.00 LyslOBArg 22.0 +0.4 969 12 44.1 3.27 Tyr312Trp 17.9 +0.5 294 4 16.4 1.21 311-314Loop 13.8 +0.4 128 3 9.3 0.69 Ser30Pro/Glyl37Ala 27.7 +0.4 298 6 10.8 0.80 S-S/Ser30Pro/Glyl37Ala 30.1 +0.5 245 6 8.2 0.60 S-S 31.8 +0.6 135 3 4.2 0.31 Ser436Pro 29.9 +0.6 903 12 30.2 2.23 S-S/Ser436Pro 31.7 +0.5 824 12 26.0 1.92 S-S/Glyl37Ala 35.2 +0.6 982 15 27.9 2.07 Glyl37Ala/Ser436Pro 36.6 +0.7 776 10 21.2 1.57 450C Wild-type 66.2 +2.2 3880 60 58.7 1.00 LyslOBArg 50.2 ±2.0 6420 110 128 2.18 Tyr312Trp 52.6 ±2.1 3360 60 63.9 1.09 311-314Loop 40.4 ±1.8 1430 40 35.3 0.60 Ser30Pro/Glyl37Ala 76.3 ±2.7 3690 70 48.4 0.83 S-S/Ser30Pro/Glyl37Ala 84.3 ±3.0 3520 60 41.7 0.71 S-S 86.3 ±3.3 963 28 11.2 0.19 Wild-type 156 +3 4890 80 31.3 1.00 Lysl08Arg 101 +1 8200 120 81.0 2.59 Tyr312Trp 110 ±2 4440 70 40.5 1.29 WO 98/03639 PCTf1JS97/1 2 9 8 3 -62- TABLE 9 (continued) 3 l1-3l4Loop 128 +2 1890 50 14.8 0.47 Ser30Pro/Gly137Ala 157 +3 7200 110 45.8 1.47 S-S/Ser3OPro/Gly137Ala 167 +3 5690 100 34.1 1.09 S-S 164 +3 1230 40 7.5 0.24 Ser436Pro 218 +3 4710 80 21.6 0.69 S-S/Ser436Pro ND d 5130 100 ND

ND

SS/Gly137Ala 225 +3 5720 100 25 .4 0.81 Gly137Ala/Ser436Pro 208 +3 ND ND

ND

a Samples were taken from 30% M100 hydrolysis reactions in 0.05 M NaOAc buffer, pH- 4.4; glucose concentrations were determined by glucose oxidase method.

b Samples were taken from 30% glucose condensation reactions in 0.05

M.

NaOAc buffer, pH 4.4; isomaltose concentrations were determined by HPTLC.

CStandard error dNot determined Kinetic parameleles of wild-type and nutlant glucoainylase lor hydvolysis ol' isomaltose and mialtooligodextrins of DP 2 7 S ibstiale ,1 (G2) Enzyme Isoimallose Malose Mhalloriose Illeolruse MOlaheIause MaIIoIlexlosc Maltolcptaose (iG2) (iG2) (02) (G3) (G4) (05) (G6) (07) Wild-type 656 A.(S1) 0.72 ±O.01b 20.4 10.2 48.2 107 64.5 1 2.9 71.8 1 1 9 73 7 t 2 72 3 1 0.9 K, (mM) 23.5 0.6 101 0.03 025 001-1 0111 0.017 0.1101 0.010 0107 0010 0.083 1 0001 k~.,IKn 0.031 10.001 20.3 055 196 19 582 65 04- f 43 685 .17 870 t Y48F49W

ND

h. 0.23610016 1 99 0Ou Km (niM) ND' 9.9 1 1.8 ND NI) NI) ND 4.9 1 0.3 k,.

1 0 024 0.003 0408± 0.010 A(AG)' mol') 17.8 20 3 YII6W 498 0.69 ±0.02 11 7 1 0.2 19.1 ±0.3 5(9 t 19 5(1( 1 17 53 1 1 1.9 560 1.1 (mM) 28.8 2.5 0.98 10.06 0.20±0.01 0.2(1 002 0 132 01H 0143 0017 0.118 0008 (SnM') 0.024 0.001 12 0 1 1.60 98 6 256 1 17 378 30 372 32 475 1 A(AG) (KI mol*') 067 1.39 1.81 2.17 .45 1.62 1.60 Y175F 752 1.02 ±0.05 21.2 1 0.2 40.0 0.6 80 11 1 .8 796 19 765 IS 72.1 0.8 K. (mM) 40.1 ±4.3 1.13 ±0.04 02)±002 0187± 0012 01201 0010 0113± 0.008 0.095± 0.004 (S'mM) 0.025 0.002 18.8 ±0.5 136 6 429 19 666 ±42 677 37 761 27 A(AG) (KJ mol') 0.55 0.20 0.97 0.81 -0.05 0.03 0.35 'Determined at 45 0 C in 0.05 M sodium acelate buffer, pl 14.4.

bStandard error.

'clanges of transition-sate energy A(AG)= -RiT l I dNot determined.

cj, '0 0 0

(D(

H

Stibsirate k (G2) Enzymeu± Isulluose IalIose vlalluwt iosc MoIlolhrefluse M;ilropeiilose Maltohexaose Mahlolcplaose (iG2) (i02) (G2) (G3) (G6) (67) R241K 261 1.34 10.08t 201 jio3 168 j (1 73J 1 7.2 70 7 .t 21 75.8 t 27 80.6 1 1.6 K, (mM) 393 t 5.8 2.27 11 062 1 1.01 0115 0(19 0.19 1 0132 0.2101 0 0 0.20 1 0.01 (S mM') 0.034 0.003 8.9 0.3 76 1.3 164 1IX 368 21 373 25 411 16 A(AG)c (KJ mol') -0.28 2.19 2.51 336 1 52 1 61 1.98 S411 A 681 0.63 10.02 18.9 03 4-16 0.1 58.5 1.6 53.1 ±1.2 54.7 1.7 59. 0.6 Km (mM) 27.9 1 2.9 1.26 0.06 0.47 f 004 0.182 0.01.1 0.120t 0009 0.115 1 0.012 0.104 0.004 k,.Km(S'mM') 0.022 0.002 ISO 15 94.1 1 5.4 322 18 .113 1.27 476 41 570 ±17 A(AG) mol') 0.84 0.80 1 91 1.50 1 is 096 1.12 S41 IG 402 0 93 ±106 230 f±1.-I 55.1 t 1 6 59 7 1 1.8 75 1 1 2.1 75.9 f 4 3 84.0 K. (mM) 26.2 2.7 1.59 0 08 0.50 0 0.1 0092 0.010 0.09-1t 0010 0.125 1 0.024 0.132 0.012 kc/K. 0.036 0 004 14 5 1 06 108 16 6419 1 55 795 61 609 1 87 634 41 A(AG) (KJ mol') -0.39 0.89 156 -0.29 -0.52 0.31 0 84 SDetermined at 45 0 C in 0.05 M sodium acetate buffer, pTI14.A.

bStandard error.

'Changes of transition-state energy -RT Iu[ (kc/Ki)/(kd/KI)II.

dNot determined.

"i Kinetic parameters of wild-type anmd mmmlillt giicc tyllscs I'Mh Ihydrolysis of isorvu I ose, mIialtose and maltolieptaose Substrate Wild-type S4 IIG S41 IA S11 I IC S41II S41 I D Isonaltose (iG2) 0.72 1 0.011 0.93 j 0.06 0.63 0.02 0.22 j 0.01 (mnM) 23.5 0.6 26.2 i 2.7 279 2 9 12.3 i 0.9 NI NDd (S'rnM") 0.031 0.001 0036 1 0.004 0.022 0.U0 0.018 Li 0.001 4 A(AG) (KJ mo') -0.39 0.841 1.4 Mallose (G2) k. 20.4 02 23(1 0 04 189 ±i 03 7.78 1 0.07 5.31 0.15 1.36 0.05 K, (rnM) 1.01± 003 1 59 1 008 1.26 0.06 053 i 0012 367 025 3.58 0.11 20.3 1 0.6 14.5 t 06 15 0 05 148 06 1 15 i 006 1.22 1- 0.03 A(AG)C (KJ mol') 0.89 0.80 0.83 6.98 7.43 Malrolcptaose (G7) A. (S 72.3 0.9 8.1.0 2.5 59.4 1 0.6 33.0 0.5 32 4 0 9 15 8 i 0.3 0.083 0.004 0.132 0.012 0.1104 0.001 0.070 0.005 0336 0024 0.148 0.009 K (S* 1 mM') 870 j 35 634 .1 11 570 17 174 1 25 97 1 5 107 i A(AG)' (Ki moe') 0.84 1.12 1.60 5.81 5.54 'Determined at 45 0 C in 0.05 M sod(iul acelale butfer, p]1 4.4.

bStandard error.

'Changes of transition-state energy a(a) -RU l dNot determined.

0 %0 00 8

C

0 w0 00 pK values and optimum pl I of'wild-type and mnutant gincoinmylases for hydrolysis of maltose and rnaltoheptaose at

C

*Free enzymie Enzyme-subsirate complex En zyme-su bst rate complex (uneomplexed) .1 (nl lose-coil e)Ixed) (Jnailtohicptaose-compillexed) pKI pK 2 p pK I pK 2 ll pK I pK 2 PI-lupi H Wild-type 2.77 5.80 4.29 2.1 1 5.85 3 98 2 60 6 78 4.69 S41I10 3.01 5.57 4.2 9 1-68 58! 41 3. 2 2 6.73 4.98 S41 I A 3.11 5.86 4.49 3,32 6. 3 2 1 3.91 7.18 5.541 S4 II C 3.26 5.86 '4.56 3.32 6.38 4.85 ND" N D N D S4111-1 3.10 5.01 1.05 3.58 '1.69 4.13 ND N D N D S41 ID 3.13 5.72 4.42 3.341 5.53 4.44 ND) N D N D 'Not deter-mined.

,4 WO 98/03639 PCTIUS97/12983 -67- TABLE 13 Increases in free energies for thermal inactivation (AAG relative to wild-type GA calculated at

G"

GA form (kJ/m ol) S436P 1.6 G137A 0.8 s-s 1.2 S -S/S43 6P 2.2 G 13 7A1S436P S3OP/G137A S-S/S3OP S-S/G 137A 2.7 S-S/S3OP/G 137A 4.4 "AAG greater than zero indiactes increased thermnostability TABLE 14 Decrease in the relative ratio of initial rate of isomaltose formation from 30% glucose condensation reactions to that of glucose formation in 30% maltodextrin M100 hydrolysis reactions.

GA form Relative ratios, W ild-type 1.00 S-S 0.24 .0.77 G137A 0.54 Y175F 0.76 300Loop 0.47 S411A 0.40 S411G 0.38 S436P 0.70 S-S/G 137A 0.81 G121A/S411G 0.44 All the above reactions were carried out in 0.05M sodium acetate buffer, pH4.4, at 550 C.

'Ratio-, lower than 1.00 indicate increased specificity for over a-(1,6)-linked substrates.

WO 98/03639 PCT[US97/12983 -68- TABLE Increase in the optimum pH of the enzyme-substrate complex of mutant glucoamylases for hydrolysis of maltose at 450 C compared to that of wild-type.

GA form pH 0 P, Increase S41 1G S411A S411C S411H S411ID 0.26 0.84 0.86 0.15 0.46 'The pH optimum of the enzyme-substratc complex of wildtype glucoamylase for hydrolysis of maltose at 45" C was pH 3.98.

WO 98/03639 PCT/US97/12983 -69-

REFERENCES

Ahearn and Klibanov. The Mechanism of Irreversible Enzyme Inactivation at 100 0 C. Science, 228, 1280 (1985) Ahearn, et al. Control of Oligomeric Enzyme Thermostability by Protein Engineering. Proc. Nat.

Acad. Sci. U. S. 84, 675 (1987).

Aleshin, et al., 1992. Crystal structure of glucoamylase from Aspergillus awamori var. X100 to 2.2- A resolution. J. Biol. Chem. 267:19291-19298.

Aleshin, et al., 1994. Refined structure for the complex of acarbose with glucoamylase from Aspergillus awamori var. X100 to 2.4-A resolution. J. Biol. Chem.

269:15631-15639.

Aleshin et al., 1994, J. Mol. Biol, 238:575-591 Aleshin, et al., 1996. Crystallographic complexes of glucoamylase with maltooligosaccharide analogs: relationship of stereochemical distortions at the nonreducing end to the catalytic mechanism.

Biochemistry. 35: 8319-8328.

Argos, et al. Thermal Stability and Protein Structure.

Biochemistry, 18, 5698 (1979).

Bakir et al (1993) Protein Eng. 6:939-946 Balaji, et al 1989. Modification of protein stability by introduction of disulfide bridges and prolines; Geometric criteria for mutation sites. Biochem.

Biophys. Res. Commun. 160:109-114.

Banks, W. and Greenwood, C. J. 1971. The specific estimation of glucose Lsing glucose oxidase. Stgrke.

23:222-228.

Boel, et al. Glucoamylases G1 and G2 from Aspergillus niger Are Synthesized from Two Different but Closely Related mRNAs. EMBO 3, 1097 (1984).

Bradford. A Rapid and Sensitive Method for the Quantitation of Microgram Quantities of Protein Utilizing the Principle of Protein-Dye Binding. Anal.

Biochem., 72, 248 (1976).

Chen, et al (1996). Effect of replacing helical glycine residues with alanine on reversible and irreversible stability and production of-Aspergillus awamori glucoamylase. Protein Eng. 9:499-505.

WO 98/03639 PCT/US97/12 9 8 3 Chen, et al (1995) Protein Eng., 8, 575-582.

Chen, et al (1994a). Increased thermostability of Asn182-Ala mutant Aspergillus awamori glucoamylase Biotechnology and Bioengineering. 43:101-105.

Chen, et al (1994b) Biochem. 301, 275-281.

Clarke and Svensson. Identification of an Essential Tryptophanyl Residue in the Primary Structure of Glucoamylase G2 from Aspergillus niger. Carlsberg Res.

Commun., 49, 559 (1984).

Clarke and Fersht, (1993) Biochemistry, 32, 4322-4329.

Cole, et al. Stable Expression of Aspergillus awamoi Glucoamylase in Distiller's Yeast. Bio/ Technol., 6, 417 (1988) Coutinho and Reilly, (1994a) Protein Eng., 7, 393-400.

Coutinho and Reilly, 1994b. Structural similarities in glucoamylases by hydrophobic cluster analysis. Protein Eng. 7:749-760.

Fierobe et al., 1996. Mutational modulation of substrate bond-type specificity and thermostability of glucoamylase from Aspergillus awamori by replacement with short homologue active site sequences and thiol/disulfide engineering. Biochemistry. 35:8696- 8704.

Frandson et al., 1994. Biochemistry 33:13808-13816.

Garfin, (1990) In: Guide to protein purification. M. P.

Deutscher, Ed. Methods in Enzymology. vol. 182. Academic Press, San Diego, CA. pp. 425-441.

Gunnarsson, et al. Structural Studies on the O-Glycosidically Linked Carbohydrate Chains of Glucoamylase G1 from Aspergillus niger. Eur. J. Biochem., 145, 463 (1984) Habeeb, (1972). Methods in Enzymology. vol. Academic Press, San Diego, CA. pp. 457-464.

Harris et al., 1993. Refined structure for the complex of l-deoxynojirimycin with glucoamylase from Aspergillus awamori var. X100 to 2.4-A resolution.

Biochemistry. 32:1618-1626.

Hiromi et al., 1983, Mol. Cell. Biochem, 51:79-95 Hiromi et al. (1966) J. Biochem 59:469-475 WO 98/03639 PCT/US97/12983 -71- Hurley, et al., 1992. Flexible-geometry conformational energy maps for the amino acid residue preceding a proline. Biopolymers. 32:1443-1446.

Imanaka, et al. A New Way of Enhancing the Thermostability of Proteases. Nature, 324, 695 (1986).

Innis, et al. Expression, Glycosylation, and Secretion of an Aspergillus Glucoamylase by Saccharomyces cerevisiae. Science, 228, 21 (1985).

Itoh, et al. Nucleotide Sequence of the Glucoamylase Gene GLU1 from Saccharomycopsis fibuligera.

J.

Bacteriol., 169, 4171 (1987).

Klibanov. Stabilization of Enzymes Against Thermal Inactivation. Adv. Appl. Microbiol., 29, 1 (1983).

Knowles, et al. Cellulase Families and Their Genes.

Trends Biotechnol., 5, 255 (1987).

Kunkel. Rapid and Efficient Site-Specific Mutagenesis Without Phenotype Selection. Proc. Nat. Acad. Sci. U.

S. 82, 448 (1985).

Kunkel, et al 1987. Rapid and efficient site-specific mutagenesis without phenotypic selection. Methods Enzymol. 154:367-382.

Lee, et al. Pilot Plant Production of Glucose with Glucoamylase Immobilized to Porous Silica. Biotechnol.

Bioeng., 16, 1507 (1976).

Leatherbarrow and Fersht. Protein Engineering.

Protein Eng., 1, 7 (1986).

Matthews, et al., 1987. Enhanced protein thermostability from site-directed mutations that decrease the entropy of. unfolding. Proc. Natl. Acad.

Sci. U. S. A. 84:6663-6667.

MacArthur and Thornton, 1991. Influence of proline residues on protein conformation. J. Mol. Biol.

218:397-412.

Masumura, et al (1989) Nature, 342, 291-293.

Matsumura and Aiba. Screening for Thermostable Mutant of Kanamycin Nucleotidyltransferase by the Use of a Transformation System for a Thermophile, Bacillus stearothermophilus. J. Biol. Chem., 260, 15298 (1985).

Matsumura, et al.- A Cumulative Effect- of Intragenic Amino Acid Replacements on the Thermostability of a WO 98/03639 PCT/US97/1298 3 -72- Protein. Nature, 323, 356 (1986) Matsuura, et al. Structure and Possible Catalytic Residues of Taka-Amylase A. J. Biochem., 95, 697 (1984).

McIlvane (1921) Biochem. J. 49:183-6.

Meagher et al., 1989, Biotechnol Bioeng. 34:681-688 Munch and Tritsch. Irreversible Thermoinactivation of Glucoamylase from Aspergillus niger and Thermostabilization by Chemical Modification of Carboxyl Groups. Biochim. Biophys. Acta, 1041, 111 (1990) Nikolov, et al. Kinetics, Equilibria, and Modeling of the Formation of Oligosaccharides from D-Glucose by Aspergillus niger Glucoamylases I and II. Biotechnol Bioeng., 34, 694 (1989).

Nosoh and Sekiguchi. Protein Engineering for Thermostability. Biocatal., 1, 257 (1988).

Nunberg, et al. Molecular Cloning and Characterization of the Glucoamylase Gene of Aspergillus awamori. Mol.

Cell. Biol., 4, 2306 (1984).

Pakula and Sauer. Genetic Analysis of Protein Stability and Function. Ann. Rev. Genet., 23, 289 (1989).

Perry and Wetzel. Disulfide Bond Engineered into T4 Lysozyme: Stabilization. of the Protein Toward Thermal Inactivation. Science, 226, 555 (1984).

Pollitt and Zalkin, (1983). J. Bacteriology, 153, 27- 32.

Rabbo and Terkildsen (1960). Scandinav. J. Lab.

Investigation 12:402-407 Ramachandran, et al., 1963. Stereochemistry of polypeptide chain configurations. J. Mol. Biol. 7:95- 99.

Robyt Mukerjea (1994) Carbohydr. Res. 251:187-202 Savel'ev and Firsov. Carboxyl Groups at the Active Site of Glucoamylase from Aspergillus awamori.

Biochemistry S. S. 47, 1365 (1982).

Savel'ev., et al. Study of :.he Active Center of Glucoamylase from Aspergilius awamori. Biochemistry WO 98/03639 PCT/US97/12983 -73- S. S. 47, 330 (1982) Semimaru, et al (1995) Appl. Environ. Microbiol., 61, 2885-2990. Sierks Svensson (1994) Protein Eng. 7:1479-1480 Sierks, et al. Site-Directed Mutagenesis at the Active Site Trpl20 of Aspergillus awamori Glucoamylase.

Protein Eng., 2, 621 (1989).

Sierks, et al. Catalytic Mechanism of Fungal Glucoamy lase as Defined by Mutagenesis of Asp176, Glu179, and Glu 180 in the Enzyme from Aspergillus awamori.

Protein Eng., 3, 193 (1990).

Smith, et al., 1985. Measurement of protein using bicinchoninoc acid. Anal Biochem. 150:76-85.

Stoffer, et al., 1995. Refined structure for the complex of D-gluco-dihydroacarbose with glucoamylase from Aspergillus awamori var. X100 to 2.2 A resolution: dual conformations for the extended inhibitors bound to the active site of glucoamylase. FEBS Letters. 358:57- 61.

Suzuki, et al., 1987. A strong correlation between the increase in the number of proline residues and the rise in thermostability of five Bacillus oligo-1,6glucosidases. Appl. Microbiol. Biotechnol. 26:546- 551.

Suzuki, et al., 1991. A hyperthermostable pullulanase produced by an extreme thermophile, Bacillus flavocaldarius KP 1228, and evidence for the proline theory of increasing protein thermostability. Appl.

Microbiol. Biothnol. 34:707-714.

Suzuki, 1989. A general principle of increasing protein thermostability. Proc. Jpn. Acad. Ser B.

65:146-148.

Suzuki, et al. Site-Directed Mutagenesis Reveals Functional Contribution of Thr218, Lys220, and Asp304 in Chymosin. Protein Eng., 4, 69 (1990).

Svensson. Regional Distant Sequence Homology Between Amylases, a-Glucosidases and Transglycosylases.

FEBS

Lett., 230, 72 (1988) Svensson, et al. Characterization of Two Forms of Glucoamylase from Aspergillus niger. Carlsberg Res.

.Commun., 47, 55 (1982).

WO 98/03639 PCT/US97/12983 -74- Svensson, et al. The Complete Amino Acid Sequence of the Glycoprotein, Glucoamylase Gl, from Aspergillus niger. Carlsberg Res. Commun., 48, 529 (1983).

Svensson, et al. Sequence Homology Between Putative Raw-Starch Binding Domains from Different Starch- Degrading Enzymes. Biochem. J. Lett., 264, 309 (1989).

Svensson, et al. Identification of Carboxylic Acid Residues in Glucoamylase G2 from Aspergillus niger That Participate in Catalysis and Substrate Binding. Eur J. Biochem., 188, 29 (1990).

Tanaka et al., 1983, J. Biochem., 93:1037-1043 Watanabe, et al., 1994. Multiple proline substitutions cumulatively thermostabilize Bacillus cereus ATCC7064 oligo-1,6-glucosidase. Eur. J. Biochem. 226:277-283.

Wasserman. Thermostable Enzyme Production. Food Technol., 38, 78 (1984).

Wetzel, R. (1987) Trends Biochem. Sci., 12, 478-482.

Williamson, et al (1992) Biochem. 282, 423-428.

WO 98/03639 PCT/US97/12983 SEQUENCE LISTING GENERAL INFORMATION: APPLICANT: Allen, Martin Fang, Tsuei-Yun Li, Yuxing Liu, Hsuan-Liang Chen, Hsui-Mei Coutinho, Pedro Hanzatko, Richard Ford, Clark (ii) TITLE OF INVENTION: PROTEIN ENGINEERING OF GLUCOAMYLASE

TO

INCREASE pH OPTIMUM, SUBSTRATE SPECIFICITY

AND

THERMOSTABILITY

(iii) NUMBER OF SEQUENCES: 12 (iv) CORRESPONDENCE

ADDRESS:

ADDRESSEE: Kohn Associates STREET: 30500 Northwestern Hwy.

CITY: Farmington Hills STATE: Michigan COUNTRY: US ZIP: 48334 COMPUTER READABLE FORM: MEDIUM TYPE: Floppy disk COMPUTER: IBM PC compatible OPERATING SYSTEM: PC-DOS/MS-DOS SOFTWARE: PatentIn Release Version #1.30 (vi) CURRENT APPLICATION

DATA:

APPLICATION

NUMBER:

FILING DATE:

CLASSIFICATION:

(viii) ATTORNEY/AGENT

INFORMATION:

NAME: Kohn, Kenneth I.

REGISTRATION NUMBER: 30,955 REFERENCE/DOCKET NUMBER: 0812.00001 (ix) TELECOMMUNICATION

INFORMATION:

TELEPHONE: (248) 539-5050 TELEFAX: (248) 539-5055 INFORMATION FOR SEQ ID NO:1: SEQUENCE CHARACTERISTICS: LENGTH: 616 amino acids TYPE: amino acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: protein (iii) HYPOTHETICAL:

NO

(vi) ORIGINAL SOURCE: ORGANISM: Aspergillus WO 98/03639 PCT/US97/1298 3 -76- (xi) SEQUENCE DESCRIPTION: SEQ ID NO:1: Ala Thr Leu Asp Ser Trp Leu Ser Asn Giu Ala Thr Val Ala Ar Thr 1 5 10 Ala Asp Phe Asp Tyr Asp Glu Ala Asn 145 Asn Leu His Ser Gin 225 Arg Asp Arg Tyr Arg 305 Ile Ser Tyr Leu Ile Leu Thr Leu 130 Gly Asp Trp Arg Cys 210 Ser Se r Pro Ala Thr 290 Tyr Leu Asn Gly Ile Thr Trp Phe Arg Ser Ala Ser Ser 100 Ala Tyr 115 Arg Ala Tyr Thr Leu Ser Glu Glu 180 Ala Leu 195 Ser Trp Phe Trp Gly Lys Giu Ala 260 Leu Ala 2*75 Leu Asn Pro Glu Asn Val Thr Asn Gin Gly Thr Thr Ser Tyr 165 Val1 Val1 Cys Thr Asp 245 Ala Asn Asp Asp Ala 325 Ile Val1 Arg Gly 70 Al a Al a Gly Al a Thr 150 Val1 As n Giu Asp Gly 230 Al a Cys His Gly Thr 310 Gly Ala Asp 55 Asp Ile Gly Ser Met 135 Al a Ala Gly Gly Ser 215 Ser As n Asp Lys Leu 295 Tyr Ala Ser 40 Ser Thr Val1 Leu Trp 120 Ile Thr Gin Se r Ser 200 Gin Phe Thr Asp Glu 280 Ser Tyr Asp 25 Pro Gly Ser Gin Gly 105 Gly Gly Asp Tyr Ser 185 Ala Ala Ile Leu Ser 265 Val1 Asp Asn Gly Ala Trp Ser Thr Asp Leu Val Leu Leu Leu Ser 75 Gly Ilie Ser 90 Giu Pro Lys Arg Pro Gin Phe Gly Gin 140 Ile Val Trp 155 Trp Asn Gin 170 Phe Phe Thr Phe Ala Thr Pro Glu Ile 220 Leu Ala Asn 235 Leu Gly Ser 250 Thr Phe Gin Val Asp Ser Ser Glu Ala 300 Gly Asn Pro 315 Val Ser Asn Pro Lys Thr Thr Ile Asn Pro Phe Asn 110 Arg Asp 125 Trp Leu Pro Leu Thr Gly Ile Ala 190 Ala Val 205 Leu Cys Phe Asp Ile His Pro Cys 270 Phe Arg 285 Val Ala Trp Phe Gly Ala Asp Tyr Leu Val Giu Asn Ser Gly Val Asp Gly Pro Leu Asp Val Arg 160 Tyr Asp 175 Val Gin Gly Ser Tyr Leu Ser Ser 240 Thr Phe 255 Ser Pro Ser Ile Val Gly Leu Cys Thr.Leu Ala Ala Glu Gin Leu Tyr Asp Ala Leu Tyr Gln Trp Asp WO 98/03639 PCT7US97/12983 -77- Lys Gin Gly Ser Leu Glu Val Thr Val Ser Leu Asp Phe Phe Lys 350 Ala Thr Val 385 Gin Trp Val Cys Ser 465 Pro Ala Thr Glu Thr 545 Pro Tyr Asp Leu Tyr 370 Ser Tyr Ser Pro Ala 450 Trp Thr Ser Ala Asn 530 Ser Leu Lys Pro Asp Ile Glu Ser 405 Ala Trp Ser Ile Gly 485 Ser Val Leu Ile Val 565 Arg Glu Ala Val Thr 390 Asp Leu Gly Ala Val 470 Ser Thr Thr Val Ala 550 Thr Ile Tyr Ala Asp 375 His Gly Leu Glu Ile 455 Ala Val Ser Phe Gly 535 Leu Val Glu Thr Trp 615 Thr 360 Ala Ala Glu Thr Thr 440 Gly Thr Thr Thr Asp 520 Ser Ser Thr Ser Val 600 Arg Thr Lys Ser Leu 410 Asn Ala Tyr Gly Thr 490 Ser Thr Ser Asp Pro 570 Asp Gin Tyr Thr Asn 395 Ser Asn Ser Ser Thr 475 Ser Thr Ala Gin Lys 555 Ala Ser Ala Ser Phe 380 Gly Ala Arg Ser Ser 460 Thr Lys Ser Thr Leu 540 Tyr Gly Val Cys Ser 365 Ala Ser Arg Arg Val 445 Val Thr Thr Cys Thr 525 Gly Thr Glu Glu Gly 605 Ser Asp Met Asp Asn 430 Pro Thr Thr Thr Thr 510 Thr Asp Ser Ser Trp 590 Thr Ser Gly Ser Leu 415 Ser Gly Val Ala Ala 495 Thr Tyr Trp Ser Phe 575 Glu Ser Ser Phe Glu 400 Thr Val Thr Thr Thr 480 Thr Pro Gly Glu Asp 560 Glu Ser Thr Ala Thr Val Thr Asp Thr 610 INFORMATION FOR SEQ ID NO:2: SEQUENCE CHARACTERISTICS: LENGTH: 7 amino acids TYPE: amino acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: peptide WO 98/03639 PCTFUS97/12983 -78- (xi) SEQUENCE DESCRIPTION: SEQ ID NO:2: Asn Gly Asn Gly Asn Ser Gin 1 INFORMATION FOR SEQ ID NO:3: SEQUENCE CHARACTERISTICS: LENGTH: 31 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:3: CAGAGTCCGC GCCCGGCACC CAAGCACCGT C 31 INFORMATION FOR SEQ ID NO:4: SEQUENCE CHARACTERISTICS: LENGTH: 33 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4: AAGTCCAGCG ACACAGGTGT GACCTCCAAC GAC 33 INFORMATION FOR SEQ ID SEQUENCE CHARACTERISTICS: LENGTH: 32 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID CGAGCGGAAA GCTGCGGGCC ATCAGACTTG TC 32 INFORMATION FOR SEQ ID NO:6: SEQUENCE CHARACTERISTICS: LENGTH: 32 base pairs TYPE: nucleic acid WO 98/03639 PCT/US97/12983 -79- STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:6: CGTACTGCCA TCCTGTGTAA CATCGGGGCG GA 32 INFORMATION FOR SEQ ID NO:7: SEQUENCE CHARACTERISTICS: LENGTH: 33 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:7: ATCGGGGCGG ACGGTTGTTG GGTGTCGGGC GCG 33 INFORMATION FOR SEQ ID NO:8: SEQUENCE CHARACTERISTICS: LENGTH: 25 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:8: GAGTATCGTG TGTACTGGCG GCACC INFORMATION FOR SEQ ID NO:9: SEQUENCE CHARACTERISTICS: LENGTH: 30 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:9: WO 98/03639 PCT/US97/12 9 83 GGTCTCGGTG AGCCCAGGTT CAATGTCGAT INFORMATION FOR SEQ ID SEQUENCE

CHARACTERISTICS:

LENGTH: 30 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc

"PRIMER"

(xi) SEQUENCE DESCRIPTION: SEQ ID GGTCTCGGTG AGCCCATGTT CAATGTCGAT INFORMATION FOR SEQ ID NO:11: SEQUENCE

CHARACTERISTICS:

LENGTH: 27 base pairs TYPE: nucleic acid STRANDEDNESS: single TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc

"PRIMER"

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:11: GAGGACACGT ACTGGAACGG CAACCCG 27 INFORMATION FOR SEQ ID NO:12: SEQUENCE

CHARACTERISTICS:

LENGTH: 60 base pairs TYPE: nucleic acid STRANDEDNESS: single: TOPOLOGY: linear (ii) MOLECULE TYPE: other nucleic acid DESCRIPTION: /desc "PRIMER" (xi) SEQUENCE DESCRIPTION: SEQ ID NO:12: TACCCTGAGG ACACGTACAA CGGCAACGGC AACTCGCAGG GCAACCCGTG GTTCCTGTGC

Claims (3)

1. 3. The fingal glucoamylase according to claim 1 or claim 2, comprising at least one mutation selected from Table 13 wherein cumulative thermal stability is provided by the additional mutation.
2. 4. The fungal glucoamylase according to any one of claims 1 to 3, comprising mutations Ser30Pro, Glyl27Ala wherein cumulative thermal stability is provided by the additional mutations. 5 The fungal glucoamylase according to any one of claims 1-4, comprising at least one mutation from Table 14 wherein cumulative reduced isomaltose formation is provided by the additional mutation. 6 The fungal glucoamylase according to any one of claims 1-5, comprising a
3. 331-314 Loop mutation wherein cumulative reduced isomaltose formation is provided by the mutation. 7. The glucoamylase according to any one of claims 1-6, wherein reduced isomaltose formation is provided by the 311-314 Loop mutation. 8 The glucoamylase according to any one of claims 1-7, wherein increased pH optimum and reduced isomaltose formation is provided by the mutation of Ser41 lAla 9. The fungal glucoamylase according to any one of claims 1-8, comprising at least one mutation from Table 15 wherein cumulative increased pH optimum is provided by the mutation. -82- A fungal glucoamylase including a Ser411Ala mutation and a mutation pair coupled with Ala27Cys forming a disulfide bond between the two members of the pair. II. The glucoamylase according to any one of claims 1-10, wherein increased thermal stability, increased pH optimum and reduced isomaltose formation are provided by the Ser41 lAla mutations and the mutation pair Asn2-Cys coupled with Ala27Cys forming a disulphide bond between the two members of the pair. 12 A fungal glucoamylase including a Ser411Ala mutation and a mutation pair coupled with Ala27Cys forming a disulfide bond between the two members of the pair and a 311-314 Loop mutation. 13. The glucoamylase according to any one of claims 1-12, wherein increased thermal stability, increased pH optimum and reduced isomaltose formation are provided by the Ser411Ala mutation and a mutation pair Asn20Cys coupled with Ala27Cys forming a disulphide bond between the two members of the pair and the 311-314 Loop mutation. 14. A vector containing the cDNA for an engineered glucoamylase according to any one of claims 1-13. A host cell transformed with the vector of claim 14. 16. A fungal glucoamylase according to any one of claims 1-15, wherein the glucoamylase is an Aspergillus glucoamylase. 17. The glucoamylase according to claim 16, wherein the glucoamylase is Aspergilhls awamori glucoamylase. 18. A method to obtain a fungal glucoamylase with decreased thermal inactivation by introducing mutations, selected from the group consisting of a pair coupled with Ala27Cys forming a disulphide bond between the two members of the -83- pair, a 311-314 Loop mutation, and a mutation of Ser411Ala, having decreased conformational entropy of unfolding, or increased stability of ca-helices, or increased disulfide bonds or hydrogen bonding and electrostatic interactions and hydrophic interactions and Vanderwalls interactions and packing compactness. 19. A method to obtain a fingal glucoamylase with reduced isomaltose formation by introducing mutations selected from the group consisting of a pair coupled with Ala27Cys forming a disulphide bond between the two members of the pair, a 311-314 Loop mutation, and a mutation of Ser411Ala, to decrease the ct- (1,6) -glucosidic linkage affinity. 20. A method to obtain a ftingal glucoamylase with increased pH optimum by changing the polarity, charge distribution and hydrogen bonding in the microenvironment of the catalytic base Glu400. 21. A method of selecting mutations for fungal glucoamylase to be used for constructing glucoamylases with cumulative mutations by designing and generating individual mutations by site directed mutagenesis; screening the individual mutations and selecting those which show at least increased pH optimum, decreased irreversible thermal inactivation rates or reduced isomaltose formation; performing site directed mutagenesis to produce enzymes carrying at least two .of the isolated selected mutations for either increased pH optimum, decreased irreversible thermal inactivation rates or reduced isomaltose formation, and screening for cumulatively additive effects of the mutations on pH optimum, thermal stabilizing or reduced isomaltose formation by the produced enzymes carrying at least two of the isolated selected mutations. -84- 22 A method to obtain a fungal glucoamylose with reduced isomaltose formation by introducing at least one mutation from Table 13 or Table 14 to decrease the ca- (1,6)-glucosidic linkage affinity. 23 A fungal glucoamylase substantially as herinbefore described with reference to any one of Examples 1 to 8. 24 A fungal glucoamylase substantially as hereinbefore described with reference to any one of Figures 1 to A method to obtain a fungal glucoamylase with decreased thermal inactivation substantially as hereinbefore described with reference to any one of Examples 1 to 6. 26. A method to obtain a fungal glucoamylase with decreased thermal inactivation substantially as hereinbefore described with reference to any one of Figures 1 to 7. 27. A method to obtain a fungal glucoamylase with reduced isomaltose formation substantially as hereinbefore described with reference to any one of Examples 1 to 8. 28. A method to obtain a fungal glucoamylase with reduced isomaltose formation substantially as hereinbefore described with reference to any one of Figures 1 to 29. A method to obtain a fungal glucoamylase with increased pH optimum substantially as hereinbefore described with reference to Example 8. A method to obtain a fungal glucoamylase with increased pH optimum substantially as hereinbefore described with reference to Figure DATED THIS 6 TH DAY OF APRIL 2001. IOWA STATE UNIVERSITY RESEARCH FOUNDATION INC. By their Patent Attorneys LORD COMPANY PERTH, WESTERN AUSTRALIA