Stability-increasing Mutants of Glucose Dehydrogenase from Bacillus megaterium IWG3*

A glucose dehydrogenase gene was isolated from Ba- cillus megaterium IWG3, and its nucleotide sequence was identified. The amino acid sequence of the enzyme deduced from the nucleotide sequence is very similar to the protein sequence of the enzyme from B. mega- terium M1286 reported by Jany et ai. (Jany, K.-D., Ulmer, W., Froschle, M., and Pfleiderer, G. (1984) FEBS Lett. 166, 6-10). The isolated gene was muta- genized with hydrazine, formic acid, or sodium nitrite, and 12 clones (H35, H39, F18, F20, F191, F192, N1, N13, N14, N28, N71, and N72) containing mutant genes for thermostable glucose dehydrogenase were obtained. The nucleotide sequences of the 12 genes show that they include 8 kinds of mutants having the following amino acid substitutions: H35 and H39, Glu- 96 to Gly; F18 and F191, Glu-96 to Ala; F20, Gln-252 to Leu; F192, Gln-252 to Leu and Ala-258 to Gly; N1, Glu-96 to Lys and Val-183 to Ile; N13 and N14, Glu-96 to Lys, Val-112 to Ala, Glu-133 to Lys, and Tyr-217 to His; N28, Glu-96 to Lys, E. filter assay, the plasmids the two restriction for BanII, HindIII, PstI, and PuuII digestion, we of the and the plasmid pGDA1. Southern hybridization analysis (25) the 3.6-kb EcoRI-SaZI fragment of pGDAl hybridizes with the probe Subcloning Dehydrogenase Gene-pGDAl DNA nuclease, linker, PstI. dehydrogenase

The nucleotide sequencefs) reported in thispaper has been submitted 504805.
to the GenBankTM/EMBL Data Bank with accession number (s) $ To whom correspondence should be addressed.
stabilized by the addition of NaCl (4,5). It has been reported that the inactivation is due to the reversible dissociation of the tetramer into inactive protomers, and the mechanism of the reversible dissociation has been investigated by Pfleiderer (4, 6,7). Therefore, this enzyme is a good candidate to be redesigned to a more stable structure using the new mechanism of strengthening intersubunit association.
In this work, we isolated a glucose dehydrogenase gene from €3. megaterium JWG3. As the tertiary structure of this enzyme is not known, we treated the gene with chemical mutagens, and obtained 12 transformant colonies producing mutant enzymes with enhanced stability. These mutant enzymes provided us with information on the amino acid residues the alteration of which affects the stability of the enzyme by changing the tertiary structure of the protomer and/or by changing the strength of the intersubunit interaction.

EXPERIMENTAL PROCEDURES
Enzymes and Chemicals-Glucose dehydrogenase from B. megaterium IWG3 was a generous gift from Amano Pharmaceutical Co. Ltd. (Nagoya), and was used after purification by DEAE-Sephadex A-50 column chromatography and HPLC' (G3000SW and DEAE-SSW).

Bacterial Strains, Phuges, and
Preparation of Plosmid and Phage DNAs-Plasmid DNAs were prepared by the alkaline-extraction method described by Birnboim and Doly (14) and further purified by CsCl/ethidium bromide density gradient centrifugation (15). M13 phage DNAs were prepared as described in Ref. 16.
Preparation of Probe DNA-The amino acid sequence of the Nterminal region of glucose dehydrogenase was identified by sequential Edman degradation using an automated protein sequencer (Applied Biosystem, model 470A); the sequence (1-29) was found to be the terminal region of the enzyme (5"GTAATAACAACAACTTTTC-same as that reported by Jany et al. (17). The probe for the N-CTTCCAGATCTTTATACAT-3') (38-mer) was chemically synthesized by an automated DNA synthesizer (Applied Biosystem, model 381A), and purified by reverse-phase HPLC (Cosmosil5C4-300). The purified oligonucleotide was labeled using [y3'P]ATP (5,000 Ci/ ' The abbreviations used are: HPLC, high-performance liquid chromatography; kb, kilobase pair. mmol, ICN) and polynucleotide kinase.
Filter Assay of Glucose Dehydrogenase Actiuity-Glucose dehydrogenase activity of transformants were detected as described in Refs. 20 and 21. Colonies on a plate were transferred to a paper filter, and the cells were lysed at room temperature. Then the filter was incubated at 60 "C for 20 min in 50 mM potassium phosphate, pH 6.5, containing 2 M NaCl and 50 mM EDTA (heat treatment). Positive clones were selected by incubating the filter at 30 "C for 10 min in 20 mM Tris/HCl, pH 8.0, 1 M NaCl, 0.1 M D-glucose, 0.5 mM 3-(4',5'dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, 0.5 mM 5ethylphenazinium ethylsulfate, and 50 p~ NAD.
Isolation of a Glucose Dehydrogenase Gene-Chromosomal DNA of B. megaterium IWG3 prepared by the method of Saito and Miura (22) was digested with EcoRI and BglII. After agarose gel (1%) electrophoresis (8) of the digested DNA, 3-4-kb fragments that hybridized with the synthetic DNA probe were obtained by extraction from the gel by the method of Vogelstein and Gillespie (23) using a Geneclean kit (BiolOl Inc.). The DNA fragments were ligated with pBR322 that had been digested with EcoRI and BamHI, and E. coli C600 was transformed with the ligated DNA. Ampicillin-resistant transformants (about 350 colonies) were analyzed by colony hybridization (24) with the probe DNA at 45 "C, and three positive colonies were obtained. These three clones were further tested for glucose dehydrogenase activity by the filter assay, and two of them were found to be positive. Since the plasmids from the two clones showed the same restriction pattern for BanII, HindIII, PstI, and PuuII digestion, we selected one of the clones and named the plasmid pGDA1. Southern hybridization analysis (25) showed that the 3.6-kb DNA.
EcoRI-SaZI fragment of pGDAl strongly hybridizes with the probe Subcloning of the Glucose Dehydrogenase Gene-pGDAl was digested with EcoRI and PuuII. A 1.5-kb DNA fragment was obtained by agarose gel electrophoresis as described above. The terminal singlestranded regions of the fragment were filled in by the Klenow fragment. The DNA fragment was then ligated with the PstI linker, and digested with BanII. The resulting terminal single-stranded regions of the fragment were removed by mung bean nuclease, the fragment was ligated with the EcoRI linker, and digested with EcoRI and PstI. The EcoRI-PstI fragment with 0.9 kb was inserted into pKK223-3, and E. coli JM105 was transformed with the hybrid plasmid. The glucose dehydrogenase activity of the transformants grown on an LB (19) plate containing ampicillin and isopropyl-6-D-thiogalactoside were detected by the filter assay. A plasmid harbored in one of the positive clones were selected and named pGDA2.
Chemical Mutagenesis-The EcoRI-PstI 0.9-kb DNA fragment containing the glucose dehydrogenase gene was inserted into M13mp18 and M13mp19. The single-stranded DNA of the hybrid plasmid was treated at 20 "C with hydrazine (for 5-20 min), formic acid (5-20 min), or sodium nitrite (1-3 h) by the method of Myers et al. (26). The mutagenized single-stranded DNA was annealed with the P I primer and made into a duplex form as described in Ref. 26. The duplex DNA was digested with EcoRI and PstI, the resulting 0.9kb fragment was inserted into pKK223-3, and E. coli JM103 was transformed with the hybrid plasmid.

RESULTS AND DISCUSSION
Glucose Dehydrogenase Gene from B. megaterium-We have isolated a glucose dehydrogenase gene from B. megaterium IWG3 as described under "Experimental Procedures," and cloned it in a plasmid, pGDA1. Fig. 1 shows the physical map of pGDA1. Southern hybridization experiments (25) using the synthetic DNA probe showed that the 0.5-kb BanII-Hind111 fragment encodes the N-terminal region of the enzyme. Therefore, we identified the nucleotide sequence of both strands of the 0.93-kb region flanked by EcoRI and Sau3AI ( Fig. 1) by the dideoxynucleotide chain-termination method (27) using a sequence kit (Toyobo Co. Ltd.); the sequence strategy is also shown in Fig. 1.

G A T C A T C A T A G C A G G A G T C A T G T T A G G G C T C G C A A A A G C T A A C T~T A T T A A C A S D
ATGTATAAAGATTTAGAAGGAAAAGTGGTCATAACATAACAGGTTCATCTACAGGTTTGGGA M e t T y r L y s A s p L e u G l u G l y L y r V a l V a l V~l V~l l l~T h~G l y S e~S e~T h~G l~L~~G l~ 20 * * * + *

A A A T C A A T G G C G A T T C G T T T T G C G A C A G A A A A A G C C A A A G T A G T T G T G A A C T A T C G T T C T L y s S e r M e t A l a I l e A r q P h e A l a T h r G l u L y s A l a L y r V~l V~l V~I A~~T y~A~g S~~ 40
* * x * * *

G C T G T C A A A G G T G A T G T A A C A G T T G A G T C T G A C G T T A T C A A T T T A G T T C A A T C T G C T A T T A l a V a l L y s G l y A s p V a l T h r V a l G l u S e r A s p V a l I l e A s n L e u V a l G l n S e r A l a I l e 80
I f t t t t

T C T C A T G A A A T G T C T T T A A G C G A T T G G A A T A A A G T C A T T G A T A C G A A C T T A A C G G G A G C T S e r H r s G l u M e t S e t L e u S e r~r p A~~L~~~~l e A s p T h r A s n t e u T h r G l y A l a
120 * X * x

T T T T T A G G C A G C C G T G A A G C G A T T A A A T A T T T T G T G G A A A A T G A T A T T A A G G G A A C A G T T
P h e L e u G l y S e r A r g G l u A l a l l e L y s T y r P h~V a l~~~A~p l l~L y~G l~T h r V~l

* A T T A A C A T G T C G A G T G T T C A C G A G A A A A T T C C T T G G C C A T T A T T T G T T C A T T A T G C A G C A
I l e A s n M e t S e r S e r V a l H~s G l u i y s i l e P r a T r p P r o L~~P h~V~l H~~T y~A l~A l~ 160 X + *

P r o G l u G l u l l e A l a A l a V a l A l a A l a T r p L e u A l a S e r S~~G l~A l~S~~T y~V~l T h~ 240
X I

G G A A T T A C G C T C T T T G C T G A C G G C G G T A T G A C A C A G T A C C C A T C A T T C C A A G C A G G A C G C
G l y I l e T h r L e u P h e A l a A s p G l y G l y M e t T h~~P~~S~~P h~G l~~l y A~g 2 6 0

FIG. 2.
Nucleotide sequence of glucose dehydrogenase gene. The nucleotide sequence was identified for both strands. The entire amino acid sequence deduced from the nucleotide sequence is shown below the coding region. The amino acid sequence underlined is the same as that identified by Edman degradation. SD represents a possible ribosome-binding site. Arrows indicate an inverted repeat structure, The amino acids that are not conserved in all the four known sequences (17,(28)(29)(30) are marked with asterisks. The amino acids that are changed in the sequences of the thermostable mutants (see Table I) are boxed. Fig. 2 shows the nucleotide sequence of the 0.93-kb fragment. There is an open reading frame of 783 base pairs available to encode a peptide of 261 amino acids (MI 28,085) which starts from ATG at position 60 and ends at TAA at position 843. This size agrees well with the MI of the subunit of the purified glucose dehydrogenase (Mr 30,000). The 29 amino acids of the N-terminal region deduced from the nucleotide sequence are the same as those identified by Edman degradation of the purified protein (the underlined amino acid sequence shown in Fig. 2). A possible ribosome-binding sequence (AGGAGG) is identified 9 bp upstream from the initiation codon. Downstream from the coding region, there is an inverted repeat sequence (16 base pairs each) followed by an AT-rich sequence that seems to be a transcriptional termination signal.
Chromosomal DNA of B. megaterium IWG3 was digested with HindIII or EcoRI + PstI, and analyzed by Southern hybridization (25) at 60 "C using the nick-translated 0.9-kb EcoRI-PstI fragment from pGDA2 as a probe DNA (Fig. 3). The amino acid sequence of glucose dehydrogenase from B. megaterium IWG3 can be deduced from the nucleotide sequence, and is similar to those from B. megaterium M1286 (17, 28, 29) and also to that from Bacillus subtilis (30); the amino acids that are not conserved in all four sequences are marked with asterisks in Fig. 2. The values of the sequence homology are about 83% between our enzyme and glucose dehydrogenase A from B. megaterium M1286 (28) and 82% between ours and the glucose dehydrogenase from B. subtilis (30). Our enzyme is almost the same as one of the glucose , the resulting DNA fragments were separated on a 1% agarose gel, transferred to a nylon filter, and hybridized at 60°C with a 32Plabeled 0.9-kb EcoRI-PstI fragment of pGDA2. The size of the fragments is given in kilobase pairs. dehydrogenases from B. megaterium M1286 whose protein sequence has been reported in Ref. 17; only 1 amino acid out of 261 is different: Leu-95 in our sequence is changed to Met-96 in the reported protein sequence (17) with some corrections (28,29). In spite of these high similarities, the reported protein sequence has two insertions (Glu-52 and Trp-150) compared with our sequence. This seems to be a very rare case, and must be a special example for enzyme evolution, provided these sequences are correct.
Isolation and Characterization of Genes for Thermostable Mutants-The EcoRI-PstI 0.9-kb DNA fragment was treated with hydrazine, formic acid, or sodium nitrite as described under "Experimental Procedures." The transformants (about 20,000 colonies) harboring the hybrid plasmid containing the mutagenized glucose dehydrogenase gene were analyzed for heat-resistant enzyme activity by the filter assay as described under "Experimental Procedures" with one alteration, that 2 M NaCl was omitted from the solution for the heat treatment at 60 "C. Under these conditions, native enzyme is inactivated and cannot be detected by the filter assay. Among the transformants, we obtained 12 positive clones; two of them were obtained from the hydrazine treatment (about 4,000 transformants), four from the formic acid treatment (about 7,500 transformants), and six from the sodium nitrite treatment (about 8,000 transformants).
We identified the nucleotide sequences of both strands of the EcoRI-PstI fragments (0.9 kb) from the 12 plasmids by the method of Henikoff (31). Table I summarizes the base substitutions caused by the chemical mutagenesis. Of the 12 clones, four pairs have identical genes: H35 and H39, F18 and F191, N13 and N14, and N71 and N72. There are two possibilities for getting these duplicate clones: the same pattern of base substitution occurred on two or more DNA fragments coding the glucose dehydrogenase gene, or transformants were duplicated by growth before the plate culture. The paired clones were obtained from one chemical mutagenesis experiment, and so we cannot select one of the possibilities.
We have thus obtained 8 kinds of mutant genes coding for thermostable glucose dehydrogenases. Among 26 base substitutions, 21 are transitions and 5 are transversions. All the transitions were made by the mutagenesis with hydrazine and sodium nitrite, and all the four types of transitions were observed (Table I). On the coding strand, however, there was no C to T transition. Four transversions were made by the mutagenesis with formic acid, one was from the mutagenesis with sodium nitrite, and a total of four types out of eight were observed (Table I). We mutagenized both coding and noncoding strands of the EcoRI-PstI DNA fragment using M13mp18 and M13mp19, respectively. This also increased the variety of mutant genes.
Not all of the base substitutions observed caused amino acid substitutions, and some of the base substitutions are silent when a mutant gene has several base substitutions. Table I also shows the amino acid substitutions observed in the mutant genes coding thermostable glucose dehydrogenases. The increased thermostability of the mutant enzymes were confirmed by measuring the remaining activity after heat treatment (at 60 "C for 20 min in 50 mM sodium phosphate buffer, pH 6.5) of crude enzyme solutions. After this heat treatment, the remaining activity of wild-type enzyme cannot be detected, but those of the mutant enzymes are 1-97% (Table I). These mutant enzymes seem to be classified into at least three groups based on the thermostability. Jany et al. (17) have reported that the modification of "yr-254 (in our enzyme Tyr-253) with tetranitromethane completely inactivates the enzyme without affecting the overall

Base changes by chemical mutagenesis and deduced amino acid substitutions in the mutant genes coding thermostable glucose dehydrogenases
Nucleotide changes are in the strand used for the chemical mutagenesis with the mutagens as described under "Experimental Procedures." M13mp18 and M13mp19 phages carry the coding and noncoding strands of the glucose dehydrogenase, respectively; here, "coding strand" is the DNA strand whose nucleotide sequence is the same as that of the mRNA for the enzyme. The numbers for the location of nucleotide change are taken from Fig. 2. The remaining activity was measured spectrophotometrically after the heat treatment at 60 "C for 20 min in 50 mM sodium phosphate, pH 6.5; the assay mixture contained 75 mM Tris/HCl, pH 8.0, 0.1 M glucose, and 2 mM NAD, and the reactions at 30 "C were recorded as the increase in absorbance at 340 nm. structure, and they concluded that Tyr-254 is involved either in catalysis or binding of substrates (17). The mutant N72, however, has the substitution of cysteine for Tyr-253 and has activity. These results indicate that Tyr-253 is not essential for the catalytic function. As the primary structures of these enzymes are almost the same, this discrepancy is not due to the difference in the structures of the two enzymes; the inactivation by the chemical modification of Tyr-254 may be due to a change in the conformation of the active site. The substitution of cysteine for Tyr-253, on the other hand, does not cause such a fatal change in the active site but increases the heat resistance of the enzyme, probably because cysteine is smaller than tyrosine. The role of Tyr-253 will be clarified by making mutant enzymes with different amino acids at this position. Fig. 2 shows that the amino acid substitutions cluster in a central region (96-133) and the C-terminal region (183-258), and no substitutions are observed in the N-terminal region  although there are two silent base substitutions in N28. In contrast, Fig. 2 also shows that 45% of the amino acids not conserved in all the four sequences are in the N-terminal region. This means that many amino acids in the N-terminal region can be replaced by others without large losses of activity, but none of the substitutions can increase the heat resistance of the enzyme. When a mutant enzyme has two or more amino acid substitutions, some of them may be neutral mutations that do not affect the activity and stability of the enzyme very much. Therefore, some mutants having additional neutral mutations in the N-terminal region could have been obtained, but they were not. This is an interesting phenomenon but we have to prepare more mutants to understand the role of the N-terminal region in the stability of the enzyme. It is also noteworthy that the mutated amino acid positions except Asp-108 and Gln-252 are all conserved in all the four sequences (Fig. 2). These amino acids are probably conserved because the stability of the enzyme must be kept at a desired level; that is, too stable a glucose dehydrogenase may not be good for the bacteria.

Mutant
Among 13 kinds of amino acid substitutions that increase the heat resistance of the enzyme, 8 substitutions decrease the number of negative charges and 4 of the 8 increase the positive charge of the enzyme at pH 6.5. These decreases in the total negative charge seem to contribute to the stabilization of the tertiary and quaternary structures of the enzyme by modulating charge-charge interactions, considering that this enzyme has an acidic isoelectric point (PI 4.7) (3) and is stabilized by a high ionic concentration (4,5). The mutants F20 and F192, on the other hand, have the same charge as the wild-type enzyme, and therefore, the stabilization mechanism of these mutants must be different from the others.
To know the effects of each amino acid substitution in the mutants, F192, N1, N13, and N28, we must prepare the mutants each having one of the substitutions. The effects of the combination of these mutations including those in H35, F18, F20, and N71 are also interesting. The mutation at Glu-96 is observed most frequently. To remove the negative charge, position 96 must be important for the increase in the stability. It is noteworthy that the mutant F18 has a higher stability than H35. This means that the methyl group of Ala-96 contributes to the increase in the stability; this stabilization mechanism is also interesting. These mutations may affect not only the stability but also the other properties of the enzyme. These properties of the mutant enzymes will be studied after purification of these enzymes.
In this work, we found that a single mutation of Glu-96 to glycine or alanine, Gln-252 to leucine, or Tyr-253 to cysteine increases the heat resistance of glucose dehydrogenase. These results indicate that any one of the four mutations can sta-bilize the tetrameric structure of the enzyme. Glu-96, Gln-252, and Tyr-253 are conserved in all the four glucose dehydrogenases except that Gln-252 is changed to lysine in the sequence of glucose dehydrogenase A from B. megaterium M1286 (Fig. 2). Therefore, these enzymes will also be stabilized by the above stability-increasing mutations.