Skip to content


  • Research article
  • Open Access

Increasing stability of water-soluble PQQ glucose dehydrogenase by increasing hydrophobic interaction at dimeric interface

  • 1,
  • 1,
  • 1 and
  • 1Email author
BMC Biochemistry20056:1

  • Received: 13 August 2004
  • Accepted: 16 February 2005
  • Published:



Water-soluble quinoprotein glucose dehydrogenase (PQQGDH-B) from Acinetobacter calcoaceticus has a great potential for application as a glucose sensor constituent. Because this enzyme shows no activity in its monomeric form, correct quaternary structure is essential for the formation of active enzyme. We have previously reported on the increasing of the stability of PQQGDH-B by preventing the subunit dissociation. Previous studies were based on decreasing the entropy of quaternary structure dissociation but not on increasing the interaction between the two subunits. We therefore attempted to introduce a hydrophobic interaction in the dimeric interface to increase the stability of PQQGDH-B.


Amino acid residues Asn340 and Tyr418 face each other at the dimer interface of PQQGDH-B, however no interaction exists between their side chains. We simultaneously substituted Asn340 to Phe and Tyr418 to Phe or Ile, to create the two mutants Asn340Phe/Tyr418Phe and Asn340Phe/Tyr418Ile. Furthermore, residues Leu280, Val282 and Val342 form a hydrophobic region that faces, on the other subunit, residues Thr416 and Thr417, again without any specific interaction. We simultaneously substituted Thr416 and Thr417 to Val, to create the mutant Thr416Val/Thr417Val. The temperatures resulting in lose of half of the initial activity of the constructed mutants were increased by 3–4°C higher over wild type. All mutants showed 2-fold higher thermal stability at 55°C than the wild-type enzyme, without decreasing their catalytic activities. From the 3D models of all the mutant enzymes, the predicted binding energies were found to be significantly greater that in the wild-type enzyme, consistent with the increases in thermal stabilities.


We have achieved via site-directed mutagenesis the improvement of the thermal stability of PQQGDH-B by increasing the dimer interface interaction. Through rational design based on the quaternary structure of the enzyme, we selected residues located at the dimer interface that do not contribute to the intersubunit interaction. By substituting these residues to hydrophobic ones, the thermal stability of PQQGDH-B was increased without decreasing its catalytic activity.


  • Mutant Enzyme
  • Dime Interface
  • DCIP
  • Acinetobacter Calcoaceticus
  • CHARMM22 Force Field


Water-soluble quinoprotein glucose dehydrogenase (PQQGDH-B) from Acinetobacter calcoaceticus has great potential for application as a constituent of an electron mediator-type glucose sensor. The conventionally utilized enzyme for glucose measurement, glucose oxidase (GOD), inherently utilizes oxygen as its electron acceptor during the oxidation of glucose. In contrast, PQQGDH-B is completely independent of oxygen, resulting in improved accuracy and rapidity during glucose measurement. However, because the substrate specificity and stability of PQQGDH-B remain somewhat inferior to those of GOD, we have been engaged in the improvement of these enzymatic properties through protein engineering to further increase the application of PQQGDH-B in glucose monitoring systems [117].

The subunit structure of PQQGDH-B was determined to be homodimeric with no activity observed in its monomeric form [17]. Correct quaternary structure is essential for the formation of active enzyme and dissociation of the dimer conformation triggers inactivation of the enzyme. We have previously reported on the increasing of the stability of PQQGDH-B against the dissociation of quaternary structure by chemical cross-linking [12], by constructing tethered enzyme [13], and by introducing Cys residues at the dimer interface to form a novel intersubunit disulfide bond [14]. All of these attempts were based on decreasing the entropy by decreasing the possibility of dimer dissociation, but no attempts had been made to increase the interaction between the two subunits.

In this paper, we report on our rational designing of hydrophobic interaction in the dimer interface to increase the stability of PQQGDH-B. We identified protein regions at the dimer interface where potential novel hydrophobic interactions could be introduced by amino acid substitution, thereby increasing the stability.


Modeling novel dimer hydrophobic core

Among 19 amino acid residues located at the dimer interface, 8 residues were predicted not to be involved in the formation of hydrogen bonds, electrostatic interactions, or hydrophobic interactions (Fig. 1). We focused on the residues Asn340, Tyr418, Thr416, and Thr417, as they are not involved in the formation of the active site cavity and are therefore suitable candidates for amino acid substitution. Although residues Asn340 and Tyr418 face each other on the surface of the dimer interface (Fig. 2), no interaction exists between their side chains. We therefore simultaneously substituted Asn340 to Phe and Tyr418 to Phe or Ile to create the mutants Asn340Phe/Tyr418Phe and Asn340Phe/Tyr418Ile, respectively. Furthermore, the hydrophobic region composed of residues Leu280, Val282, and Val342 faces residues Thr416 and Thr417, again with no specific interaction. We therefore substituted Thr416 to Val and Thr417 to Val to create the double mutant Thr416Val/Thr417Val.
Figure 1
Figure 1

The amino acid residues located at the dimer interface. The two subunits are represented in red and blue, respectively, using the RasMol molecular visualization software 26. The 19 amino acid residues at the interface are shown in space filling format, of which 8 residues (orange and light blue) are predicted not be involve in hydrogen bond formation, electrostatic interaction, or hydrophobic interaction at the interface.

Figure 2
Figure 2

Hydrophobicity of PQQGDH-B dimer interface. Hydrophobic regions are shown in green and hydrophilic regions are shown in blue. Residues that have been substituted in this study are indicated. The structural images were generated using Molecular Operating Environment.

Characterization of mutant enzymes

The activity and stability of each of the three constructed double mutant enzymes were then analyzed. All three mutant enzymes showed slightly higher thermal stability than wild-type PQQGDH-B upon incubation for 10 min at various temperatures (Fig. 3). The temperatures resulting in lose of half of the initial activity were shifted by approximately 3°C higher in the mutants compared to the wild type (Table 1). As the time course of thermal inactivation at 55°C follows first-order kinetics (Fig. 4), half-lives were calculated using logarithmic regression of residual activity. The wild-type enzyme inactivates at 55°C with a half-life of 9.5 ± 1.4 min, while all three double mutants showed greater thermal stability, with half-lives of 5–6°C greater (Table 1).
Figure 3
Figure 3

Thermal stability of wild-type and mutant PQQGDH-Bs. Residual activity was measured at 25°C after 10-min incubations at different temperatures of the following protein samples (0.075 μg/mL): Wild-type , Asn340Phe/Tyr418Phe □, Asn340Phe/Tyr418Ile , and Thr416Val/Thr417Val .

Table 1

Kinetic parameters and thermal stability of PQQGDH-Bs.


V max(U/mg)

K m(mM)


t1/2 (min)b

Wild type



53.9 ± 0.9

9.5 ± 1.4




57.7 ± 0.4

14.9 ± 1.1




57.5 ± 0.3

15.5 ± 1.5




56.5 ± 0.9

14.8 ± 1.4

a Th represents the temperature at which half the initial activity is lost in 10 min.

b t1/2 represents the half-life at 55°C.

Figure 4
Figure 4

Time course of thermal inactivation of wild-type and mutant PQQGDH-Bs. Protein samples (0.075 μg/mL) were incubated at 55°C and aliquots were taken at different times to measure residual activity. The analyzed samples contained the following PQQGDH-Bs: Wild-type , Asn340Phe/Tyr418Phe □, Asn340Phe/Tyr418Ile , Thr416Val/Thr417Val .

Investigation of the kinetic parameters of the wild-type and mutant enzymes shows that the mutations did not significantly affect the overall kinetic properties of PQQGDH-B (Table 1). The specific activities of Asn340Phe/Tyr418Phe (3100 U/mg), Asn340Phe/Tyr418Ile (2500 U/mg), and Thr416Val/Thr417Val (2800 U/mg) were close to that of the wild-type enzyme (3030 U/mg). Except for Thr416Val/Thr417Val, which had a Km value of 16 mM, the mutants had Km values identical to 20 mM Km value of the wild-type enzyme.


PQQGDH-B is a 6-blade β-propeller protein, with each blade consisting of a 4-stranded anti-parallel β-sheet (W-motif) [18]. The strands in each W-motif are labeled A-D from the inside to the outside of the molecule [18, 19]. All strands are connected by loops, which are named according to the strands they connect. Based on PQQGDH-B structural information, these loop regions have been associated with a number of important functions, such as substrate binding, co-factor binding, and formation of the enzyme active site. As with other β-propeller proteins, the catalytic site and substrate-binding pocket of PQQGDH-B is made up of the cleft formed by loops DA and BC [20]. The enzyme surface composed of loops AB and CD is therefore located opposite the functional region. In the present study, we have introduced mutations at Asn340, Tyr418, Thr416, and Thr417, which are all located in the loop 5CD region. Considering that these residues are located far from the functional region and do not contribute in the structure of functional region, it is not surprising that their substitutions did not significantly alter the enzyme's kinetic parameters, particularly its catalytic activity.

Based on the predicted structure of the mutant enzymes shown in Figure 5, the binding energy of each subunit was calculated, using both the AMBER89 and CHARMM22 force fields, and compared with those of the wild-type enzyme (Table 2). As expected, the increases in hydrophobicity at the subunit interface resulted in increases in their binding energies calculated by both methods. Estimation of the number of hydrophobic interactions based on these same predicted models revealed 4 to 5 novel hydrophobic interactions at the interface of all the mutants, while none were found in the wild-type one. These results based on structural predictions are consistent with the observed improvements in thermal stability.
Table 2

Predicted binding energy of each mutant subunit calculated using AMBER89 and CHARMM22 force fields.


Binding energy (kcal/mol)




Wild type












Figure 5
Figure 5

Hydrophobicity of mutant PQQGDH-B dimer interface. Hydrophobic regions are shown in green while hydrophilic regions are shown in blue. The interfaces shown are those of the Asn340Phe/Tyr418Phe (left), Asn340Phe/Tyr418Ile (middle), and Thr416Val/Thr417Val (right) mutants of PQQGDH-B, with their respective mutation sites circled. The structural images were generated using Molecular Operating Environment.

The addition of hydrophobic interactions in other enzymes has been reported to result in thermostability increases of 2 to 10°C [2124], comparable to the results of the current study. Recent observations of enzymes from thermophilic organisms indicate that their extraordinary thermal stability is due to hydrophobic interactions. Oligomeric enzyme stability was also reported to be improved through engineering to increase the hydrophobic interaction at the oligomer interface [25]. However, these studies were based on homology analyses between mesophilic and thermophilic bacteria together with random mutagenesis library analyses [2124]. Although some sequences have been found to be homologous to PQQGDH-B, these are all putative ORFs with no functional information reported. Therefore, there exists no reliable template to help improve the stability of this enzyme by increasing the dimer interface interaction.


We improved PQQGDH-B's thermal stability by increasing the dimer interface interaction through rational design based on its quaternary structure. We demonstrated that this can be achieved by selecting residues located at the dimer interface that do not contribute to the intersubunit interaction and substituting them to hydrophobic ones via site-directed mutagenesis. In each case tested, the enzyme's thermal stability was increased without decreasing its catalytic activity. This rational design approach will provide relevant information for future designs by combining with other mutant PQQGDH-Bs with narrowed substrate specificity and improved catalytic efficiency.



Glucose, phenazine methosulfate (PMS), 2,6-dichlorophenolindophenol (DCIP), and magnesium chloride were obtained from Kanto Kagaku (Tokyo, Japan), 3-(N-morpholino) propane sulfonate (MOPS) from Dojin (Kumamoto, Japan), and pyrroloquinoline quinone from Mitsubishi Gas Chemical Company (Tokyo, Japan). All other regents were of analytical grade. Kpn I was obtained from TOYOBO (Osaka, Japan) and Hin dIII from New England BioLabs (Beverly, USA).

Site-directed mutagenesis

The structural gene for wild-type PQQGDH-B was previously amplified by polymerase chain reaction (PCR) and inserted into the expression vector pTrc99A (Pharmacia) to create pGB [15]. A 1.2-kbp Kpn I-Hin dIII fragment containing the PQQGDH-B gene was transferred from pGB to pKF18k and mutagenesis was carried out with the Mutan-Express Km kit (Takara) according to the manufacturer's instructions with the oligonucleotides Asn340Phe (5'-GGTGGGACAAAGAA TTTACCAGTCC-3'), Tyr418Phe (5'-CGGTACAGCGTCATCAAA AGTAGTGC-3'), Tyr418Ile (5'-CGGTACAGCGTCATCAAT AGTAGTGC-3'), and Thr416Val/Thr417Val (5'-CAGCGTCATCATAAACAAC GCTATAAGTTGGATC-3'). The mutations (underlined) were confirmed by automated DNA sequencing (ABI PRISM Genetic analyzer 310, Applied Biosystems). The mutated genes were digested with Kpn I and Hin dIII and were replaced into pGB to construct expression vectors containing mutated PQQGDH-B. Numbering of the amino acid positions starts from the first residue of the signal peptide (24 residues).

Enzyme preparation and assay

The PQQGDH-B genes were expressed in Escherichia coli and the enzymes purified as previously reported [15, 16]. Following a 30-min preincubation in 10 mM MOPS-NaOH (pH 7.0) containing 1 μM PQQ and 1 mM CaCl2 at room temperature (25°C) to produce the holoenzyme, GDH activity was measured by using 0.6 mM PMS and 0.06 mM DCIP. The enzyme activity was determined by measuring the decrease in absorbance of DCIP at 600 nm.

Analysis of PQQGDH-B stability

The thermal stability of wild-type and mutant PQQGDH-B was determined with 0.075 μg/mL protein, as previously reported [15]. Thermal inactivation experiments were carried out by incubating each holoenzyme in 200 μL of 10 mM MOPS-NaOH, pH 7.0, at 55°C. Aliquots were taken every 5 min and kept at 4°C for 2 min, followed by incubation at room temperature for 30 min. The residual enzyme activity was determined as described above. Since the initial time course for thermal inactivation at 55°C followed first-order kinetics, the thermal stability of each mutant enzyme was expressed as a half-life. The thermal stability of Asn340Phe/Tyr418Phe, Asn340Phe/Tyr418Ile and Thr416Val/Thr417Val were also determined by incubating purified enzyme at various temperatures for 10 min. The residual activities were determined as described above, and were compared with the initial activities.

Prediction of three-dimensional structure and quaternary-dimensional structures

Three-dimensional and quaternary structures were predicted using Molecular Operating Environment (MOE) (Chemical Computing Group Inc., Quebec, Canada). By using the available PDB data of the wild-type PQQGDH-B, 1QBI [18], we made the appropriate substitutions with all possible side-chain orientations to predict the structures of the Asn340Phe/Tyr418Phe, Asn340Phe/Tyr418Ile and Thr416Val/Thr417Val mutants. After addition of hydrogen atoms to the PQQGDH-B structure and optimization of orientation of some hydrogen atoms by MOE, the structures were subjected to energy minimization using the AMBER89 or CHARMM22 force field within the MOE program until the final energy gradient was < 0.01 kcal/mol·Å.



We are grateful to Rie Yamoto for her assistance in preparing the manuscript.

Authors’ Affiliations

Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-13 Naka-machi, Koganei, Tokyo 184-8588, Japan


  1. D'Costa EJ, Higgins IJ, Turner APF: Quinoprotein glucose dehydrogenase and its application in an amperometric glucose sensor. Biosensors. 1986, 2: 71-87. 10.1016/0265-928X(86)80011-6.View ArticlePubMedGoogle Scholar
  2. Yokoyama K, Sode K, Tamiya E, Karube I: Integrated biosensor for glucose and galactose. Anal Chim Acta. 1989, 218: 137-142. 10.1016/S0003-2670(00)80291-3.View ArticleGoogle Scholar
  3. Smolander M, Livio H-L, Rasanen L: Mediated amperometric determination of xylose and glucose with an immobilized aldose dehydrogenase electrode. Biosensors & Bioelectronics. 1992, 7: 637-643. 10.1016/0956-5663(92)85021-2.View ArticleGoogle Scholar
  4. Sode K, Nakasono S, Tanaka M, Matsunaga T: Subzero temperature operating biosensor utilizing an organic solvent and quinoprotein glucose dehydrogenase. Biotechnol Bioeng. 1993, 42: 251-254. 10.1002/bit.260420214.View ArticlePubMedGoogle Scholar
  5. Ye L, Hammerle M, Olsthoorn AJJ, Schuhmann W, Schmidt H-L, Duine JA, Heller A: High current density "wired" quinoprotein glucose dehydrogenase electrode. Anal Chem. 1993, 65: 238-241.View ArticleGoogle Scholar
  6. Katz E, Schlereth DD, Schmidt H-L: Reconstitution of the quinoprotein glucose dehydrogenase from its apoenzyme on a gold electrode surface modified with a monolayer of pyrroloquinoline quinone. Electroanal Chem. 1994, 368: 165-171. 10.1016/0022-0728(93)03094-6.View ArticleGoogle Scholar
  7. Kost GJ, Vu H-T, Lee JH, Bourgeois P, Kiechle FL, Martin C, Miller SS, Okorodudu AO, Podczasy JJ, Webster R, Whitlow KJ: Multi-center study of oxygen-insensitive handheld glucose point-of-care testing in critical care/hospital/ambulatory patients in the United States and Canada. Crit Care Med. 1998, 26: 581-590. 10.1097/00003246-199803000-00036.View ArticlePubMedGoogle Scholar
  8. Laurinavicius V, Kurtinaitiene B, Liauksminas V, Jankauskaite A, Simkus R, Meskys R, Boguslavsky L, Skotheim T, Tanenbaum S: Reagentless biosensor based on PQQ-dependent glucose dehydrogenase and partially hydrolyzed polyarbutin. Talanta. 2000, 52: 485-493. 10.1016/S0039-9140(00)00396-9.View ArticlePubMedGoogle Scholar
  9. Razumiene J, Meskys R, Gureviciene V, Laurinavicius V, Reshetova MD, Ryabov AD: 4-Ferrocenyllphenol as an electron transfer mediator in PQQ-dependent alcohol and glucose dehydrogenase-catalyzed reactions. Electrochemistry Commun. 2000, 2: 307-311. 10.1016/S1388-2481(00)00024-2.View ArticleGoogle Scholar
  10. Schmidt B: Oxygen-independent oxidases A new class of enzymes for application in diagnostics. Clinica Chim Acta. 1997, 26: 33-37. 10.1016/S0009-8981(97)00164-2.View ArticleGoogle Scholar
  11. Mullen WH, Churchhouse J, Vadgama P: Enzyme electrode for glucose based on the quinoprotein glucose dehydrogenase. Analyst. 1985, 110: 925-928. 10.1039/an9851000925.View ArticlePubMedGoogle Scholar
  12. Takahashi Y, Igarashi S, Nakazawa Y, Tsugawa W, Sode K: Construction and characterization of glucose enzyme sensor employing engineered water soluble PQQ glucose dehydrogenase with improved thermal stability. Electrochemistry. 2000, 68: 907-911.Google Scholar
  13. Sode K, Shirahane M, Yoshida H: Construction and characterization of a linked-dimeric pyrroloquinoline quinone glucose dehydrogenase. Biotechnol Lett. 1999, 21: 707-710. 10.1023/A:1005518610946.View ArticleGoogle Scholar
  14. Igarashi I, Sode K: Stabilization of quaternary structure of water-soluble quinoprotein glucose dehydrogenase. Mol Biotech. 2003, 24: 97-103. 10.1385/MB:24:2:97.View ArticleGoogle Scholar
  15. Sode K, Ohtera T, Shirahane M, Witarto AB, Igarashi S, Yoshida H: Increasing the thermal stability of the water-soluble pyrroloquinoline quinone glucose dehydrogenase by single amino acid replacement. Enz Microbial Technol. 2000, 26: 491-496. 10.1016/S0141-0229(99)00196-9.View ArticleGoogle Scholar
  16. Igarashi S, Ohtera T, Yoshida H, Witarto AB, Sode K: Construction and characterization of mutant water-soluble PQQ glucose dehydrogenases with altered Km value – site-directed mutagenesis studies on the putative active site. Biochem Biophys Res Commun. 1999, 264: 820-824. 10.1006/bbrc.1999.1157.View ArticlePubMedGoogle Scholar
  17. Igarashi S, Okuda J, Ikebukuro K, Sode K: Molecular engineering of PQQGDH and its applications. Arch Biochem Biophys. 2004, 428: 52-63. 10.1016/ ArticlePubMedGoogle Scholar
  18. Oubrie A, Rozeboom HJ, Kalk KH, Duine JA, Dijkstra BW: The 1.7Å crystal structure of the apo-form of the soluble quinoprotein glucose dehydrogenase from Acinetobacter calcoaceticus reveals a novel internal conserved sequence repeat. J Mol Biol. 1999, 289: 319-333. 10.1006/jmbi.1999.2766.View ArticlePubMedGoogle Scholar
  19. Faber HR, Groom CR, Baker HM, Morgan WT, Smith A, Baker EN: 1.8Å crystal structure of the C-terminal domain of rabbit serum haemopexin. Structure. 1995, 3: 551-559. 10.1016/S0969-2126(01)00189-7.View ArticlePubMedGoogle Scholar
  20. Oubrie A, Rozeboom HJ, Kalk KH, Olsthoorn AJJ, Duine JA, Dijkstra BW: Structure and mechanism of soluble quinoprotein glucose dehydrogenase. EMBO J. 1999, 18: 5187-5159. 10.1093/emboj/18.19.5187.PubMed CentralView ArticlePubMedGoogle Scholar
  21. Kallwass HKW, Surewicz W, Parris W, Macfarlane ELA, Luyten MA, Kay CM, Gold M, Jones JB: Single amino acid substitution can further increase the stability of a thermophilic L-lactate dehydrogenase. Protein Eng. 1992, 5: 769-774.View ArticlePubMedGoogle Scholar
  22. Kirino H, Aoki M, Aoshima M, Hayashi Y, Ohba M, Yamagishi A, Wakagi T, Oshima T: Hydrophobic interaction at the subunit interface contributes to the thermostability of 3-isopropylmalate dehydrogenase from an extreme thermophile, Thermus thermophilus. Eur J Biochem. 1994, 220: 275-281.View ArticlePubMedGoogle Scholar
  23. Akanuma S, Yamagishi A, Tanaka N, Oshima T: Serial increase in the thermal stability of 3-isopropylmalate dehydrogenase from Bacillus subtilis by experimental evolution. Protein Sci. 1998, 7: 698-705.PubMed CentralView ArticlePubMedGoogle Scholar
  24. Ohkuri T, Yamagishi A: Increased thermal stability against irreversible inactivation of 3-isopropylmalate dehydrogenase induced by decreased van der Waals volume at the subunit interface. Protein Eng. 2003, 16: 615-621. 10.1093/protein/gzg071.View ArticlePubMedGoogle Scholar
  25. Glaser F, Steinberg DM, Vakser IA, Ben-Tal N: Residue frequencies and pairing preferences at protein-protein interfaces. Proteins. 2001, 43: 89-102.View ArticlePubMedGoogle Scholar
  26. Sayle S, Milner-White EJ: RasMol: Biomolecular graphics for all Trends in Biochemical Sciences. TIBS. 1995, 20: 374-PubMedGoogle Scholar


© Tanaka et al; licensee BioMed Central Ltd. 2005

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.