Crystal Structure of Octaprenyl Pyrophosphate Synthase from Hyperthermophilic Thermotoga maritima and Mechanism of Product

Chain Length Determination


Rey-Ting Guo1(郭瑞庭), Chih-Jung Kuo2(郭致榮), Tzu-Ping Ko2(柯子平),

Chia-Cheng Chou2(周家丞), Po-Huang Liang1,2*(梁博煌) and Andrew H.-J. Wang1,2*(王惠鈞)

1Taiwan International Graduate Program, Academia Sinica, Taipei 115, Taiwan

2Institute of Biological Chemistry, Academia Sinica, Taipei 115, Taiwan

No.128 Academia Road Section 2, Nan-Kang, Taipei, 11529, Taiwan


Octaprenyl pyrophosphate synthase (OPPs) catalyzes consecutive condensation reactions of farnesyl pyrophosphate (FPP) with isopentenyl pyrophosphate (IPP) to generate C40 octaprenyl pyrophosphate (OPP) which constitutes the side chain of bacterial ubiquinone or menaquinone. In this study, the first structure of long-chain C40-OPPs from Thermotoga maritima has been determined to 2.28 Å resolution. OPPs is composed entirely of a-helices joined by connecting loops and is arranged with 9 core helices around a large central cavity. An elongated hydrophobic tunnel between D and F a-helices contains two DDxxD motifs on the top for substrate binding and is occupied at the bottom with one large residue F132. The products of the mutant F132A OPPs are predominantly C50, longer than the C40 synthesized by the wild type OPPs, suggesting that F132 is the key residue for determining the product chain length. A76 and S77 located close to the FPP binding site and V73 positioned further down the tunnel were individually mutated to larger amino acids. A76Y and S77F mainly produce C20 indicating that the mutated large residues in the vicinity of the FPP site limit the substrate chain elongation. A76 is the 5th amino acid upstream from the first DDxxD motif on helix D of OPPs and its corresponding amino acid in FPPs is Tyr. In contrast, V73Y mutation led to additional accumulation of C30 intermediate. Based on the structure, we have sequentially mutated the large amino acids including F132, L128, I123, and D62 to small Ala underneath the tunnel. The products of the mutant F132A/L128A/I123A/D62A reach C95, beyond the largest chain length generated by all known trans-prenyltransferases. Further modifications of enzyme reaction conditions may allow the preparation of high molecular weight polyprenyl products resembling rubber molecule. The new structure of the trans-type OPPs significantly extends our understanding on the biosynthesis of long chain polyprenyl molecules.




Structural study of porcine β-Microseminoprotein

Iren Wang1, 2, Yuan-Chao Lou3, Wen-Chang Chang1, Shih-Hsiung Wu1, 2, Chinpan Chen3

1Institute of Biochemical Sciences, National Taiwan University, Taipei 106, Taiwan;

2Institute of Biological Chemistry, Academia Sinica, Taipei 115, Taiwan;

3Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan


A sperm motility inhibitor originally isolated from porcine seminal plasma has been confirmed to be identical to porcine β-microseminoprotein (MSP) based on peptide sequence, amino acid composition and mass spectral analysis. Porcine MSP contains 91 amino acid residues with a pyroglutamate at its N-terminus and is linked with five disulfide bridges. This protein inhibits competitively the activity of Na+, K+-ATPase purified from porcine cerebral cortex in a concentration-dependent manner. The half-maximal inhibition was achieved at an inhibitor concentration of 90 μM. To date, we have not found any known structure related with porcine MSP in Protein Data Bank. Therefore, knowing the structure/function relationship on MSP is an interesting and important task. For NMR structural study, we cloned, expressed, and purified the recombinant porcine MSP, which possesses 17 extra amino acids at the N-terminus compared to that of the native protein. The recombinant porcine MSP exhibits an inhibitory effect identical to the native protein. In addition, CD and NMR data showed that the secondary as well as tertiary structures are very similar between the native and recombinant proteins. On the basis of heteronuclear multidimensional NMR techniques, we have determined the three-dimensional structure of the recombinant porcine MSP which contains two domains formed with one four-stranded and three double-stranded antiparallel β-sheets. Except for the disulfide bridge of Cys34-Cys70, we suggested the driving force for the interaction of the two domains may be due to the electrostatic force. In this poster presentation, we will also discuss the backbone dynamic analysis of porcine MSP.




Crystal Structure Study of Inorganic Pyrophosphatase

from Helicobacter pylori


Ti-Chun Chao1* , Jia-Yin Tsai1, Po- Kai Lyu1,Chih-Tson Fong1,

     Chung-Yu Huang2, Hai-Mei Huang2, Yuh-Ju Sun1

     1Institute of Bioinformatics and Structural Biology, National Tsing Hua University,

Hsinchu 300, Taiwan, ROC
     2Institute of Biotechnology, National Tsing Hua University, Hsinchu 300, Taiwan, ROC

     No.101 Kuang Fu Road Section 2, Hsinchu300, Taiwan, ROC


Helicobacter pylori ( H. pylori ) is a gram-negative microaerophilic bacterium that is associated with peptic ulcers, gastric cancer and chronic gastritis. Inorganic pyrophosphatase [E.C.] (PPase) is an essential enzyme for energy metabolism in all cells. It catalyzes the hydrolysis of PPi that is synthesized during polymer (e.g. starch, sucrose, cellulose, protein, DNA, RNA) synthesis and then hydrolyzed to Pi, thus providing a thermodynamic pull favoring polymer synthesis. PPases are strongly dependent on divalent cations. The crystal structure of recombinant PPase from H. pylori (strain: 26695) was reported here.  PPase from H. pylori  is cytoplasmic and belongs to family I.

The subunit molecular weight of H. pylori PPase is about 20 kDa.  H. pylori PPase is recognized as a hexamer in solution according to the results of ultracentrifugation and gel filtration. The hexameric oligomerization was also seen in the structure. The PPase structure is composed of two extended α-helices and eight β-strands.  Crystal structures of PPase and PPi- PPase complex were solved here.





Crystal Structure of Spermidine Synthase from Helicobacter pylori

Shiang-Yi Chien1*, Po-Kai Lyu1, Jia-Yin Tsai1, Chih-Tson Fong1,

Meng-Juan Lee2, Hai-Mei Huang2 and Yuh-Ju Sun1

1Institute of Bioinformatics and Structural Biology, National Tsing Hua University, Hsinchu, Taiwan

2Institute of Biotechnology, National Tsing Hua University, Hsinchu, Taiwan

101, Section 2 Kuang Fu Road, Hsinchu, Taiwan 300, ROC


Polyamines, such as putrescine, spermidine, and spermine, are polycationic mediators of cell proliferation and differentiation in most organisms. Spermidine synthase catalyzes the transfer of the aminopropyl group from decarboxylated S-adenosylmethionine to putrescine in the biosynthesis of spermidine. Here we report the crystal structure of spermidine synthase from Helicobacter pylori, which specialized in the colonization of the human stomach. Spermidine synthsae from H. pylori contains 262 amino acids with a molecular weight of 30.5 kDa. Its crystal structure has been determined to 2.5Å resolution by multiwavelength anomalous dispersion (MAD) from selenomethionine (Se-Met)-containing proteins and refined to a crystallographic R-factor of 22.1% and an R-free value of 29.3% so far. The enzyme shows the architecture of a dimer, and each monomer consists of a C-terminal domain with a Rossmann-like fold and an N-terminal beta-stranded domain. The larger C-terminal is a catalytic site and the N-terminal plays a role to maintain the dimer structure.




A Novel d(ApCpGpA)-Binding of a Bullfrog Ribonuclease Suggests the

Sialic Acid Binding


Jia-Yin Tsai1*, Yuh-Ju Sun1, Ching-Ju Tsai1, Jyung-Hurng Liu2, Chao-Tsen Chen3,

Charng-Sheng Tsai3, Liang-Yan Chen1, Pei-Tsung Cheng1, and You-Di Liao4

1Institute of Bioinformatics and Structural Biology, National Tsing Hua University, Hsinchu 300, Taiwan, ROC

2 Graduate Institute of Life Sciences, National Defense Medical Center Taipei 115, Taiwan, ROC

3 Department of Chemistry, National Taiwan University, Taipei 106, Taiwan, ROC

4 Institute of Biomedical Sciences, Academia Sinica, Taipei 115, Taiwan, ROC


Ribonucleases play a crucial role in gene expression by degrading RNA.  In addition to the intrinsic ribonucleolytic activity, bullfrog ribonucleases have been reported to exhibit cytotoxicity, lectin, and agglutinating activities toward tumor cells.  We report here two crystal structures of RC-RNase6, complexed with the dinucleotide (2’,5’ CpG) and tetranucleotide d(ApCpGpA), solved at 2.0 Å and 1.7 Å resolution, respectively.  The nucleotides’ binding modes are markedly different in the two complexes.  The retro binding (nonproductive) is found in the RC-RNase6×(2’,5’ CpG) complex.  Whereas a novel nucleotide binding is discovered in the RC-RNase6×d(ApCpGpA) complex and the d(ApCpGpA) binds to the surface of RC-RNase6.  Three significant residues (Thr14, Lys15, and Lys16) and two loop regions (residues 54-56 and 83-85) involved in the d(ApCpGpA) binding.  Among these residues, the positively charged residues on the surface of RC-RNase6 play important roles.  Residues 54-56 are in a similar location to the sialic acid binding site in cSBL (sialic acid binding lectin).  Meanwhile, the sialic acid binding of RC-RNase6 is detected by sialic acid-gold nanoparticle experiments, and in which multiple sialic acid binding sites are found at the surface of RC-RNase6.  We suggest that the novel nucleotide binding sites in the RC-RNase6×d(ApCpGpA) complex may represent the possible sialic acid binding sites for ribonuceleases cytotoxicity.





Studies of the Crystal Structures and the Substrate Specificities of

Escherichia coli Thioesterase I/Protease I /Lysophospholipase L1


Yu-Chih Lo1, Su-Chang Lin2, Jei-Fu Shaw3 and Yen-Chywan Liaw1

1Institute of Molecular Biology, Academia Sinica, Nan-Kang, Taipei 11529, Taiwan

2Graduate Institute of Life Sciences, National Defense Medical Center, Taipei, Taiwan

3Institute of Botany, Academia Sinica, Nan-Kang, Taipei 11529, Taiwan


Escherichia coli (E. coli) thioesterase I (TAP) is a multifunctional enzyme possessing activities of thioesterase, esterase, arylesterase, protease, and lysophospholipase. In particular, TAP has stereoselectivity for amino acid derivative substrates, hence it is useful for the kinetic resolution of racemic mixtures of industrial chemicals. The crystal structure of native TAP was determined at 1.9 Å, revealing a minimal SGNH-hydrolase fold. The structure of TAP in complex with a diethyl phosphono moiety (DEP) identified its catalytic triad, Ser10-Asp154-His157, and oxyanion hole, Ser10-Gly44-Asn73. The oxyanion hole of TAP consists of three residues each separated from the other by more than 3.5 Å, implying that all of them are highly polarized when substrate bound. The catalytic (His)Ce1-H···O=C hydrogen bond usually plays a role in the catalytic mechanisms of most serine hydrolases, however, there were none present in SGNH-hydrolases. We propose that the existence of the highly polarized tri-residue-constituted oxyanion hole compensates for the lack of a (His)Ce1-H···O=C hydrogen bond. This suggests that members of the SGNH-hydrolase family may employ a unique catalytic mechanism. In addition, most SGNH-hydrolases have low sequence identities and presently there is no clear criterion to define consensus sequence blocks. Through comparison of TAP and the three SGNH-hydrolase structures the currently known, we have identified a unique hydrogen bond network which stabilizes the catalytic center: a newly discovered structural feature of SGNH-hydrolases. On the other part, in order to investigate TAP’s versatile substrate specificities, we present the crystal structure of TAP in complex with octanoic acid (TAP-OCA; OCA, a free fatty acid with 8 carbon atoms). Structural comparison between the structures of native TAP and TAP-OCA shows a remarkable conformational rearrangement in loop75-80, called “switch loop movement”, upon OCA binding to the substrate-binding crevice of TAP. The switch loop movement is the acyl chain length dependent because it is absent in the TAP-DEP structure. OCA binding to the substrate binding crevice results in a continuous hydrophobic surface, which then triggers the switch loop movement. Together with previous substrate specificity assays, we suggest the effect of the switch loop movement is to stabilize the transition state intermediates of TAP during catalysis and is essential for TAP’s substrate preference for thioesters with long acyl chains. The finding of a sulfate ion binding site in the TAP structures and previous substrate specificity assays lead us to postulate that a putative CoA binding site is essential for efficient catalysis of thioesters.




DNA Binding and Degradation by the HNH Protein ColE7

Kuo-Chiang Hsia1, Kin-Fu Chak2, Po-Huang Liang3, Yi-Sheng Cheng1, Wen-Yen Ku1, and Hanna S. Yuan1

1Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan, ROC.

2Institute of Biochemistry, National Yang-Ming University, Taipei, Taiwan, ROC.

3Institute of Biological Chemistry, Academia Sinica, Taipei, Taiwan, ROC.


The bacterial toxin ColE7 bears an HNH motif which has been identified in hundreds of prokaryotic and eukaryotic endonucleases, involved in DNA homing, restriction, repair or chromosome degradation.  The crystal structure of the nuclease domain of ColE7 in complex with a duplex DNA has been determined at a resolution of 2.5 Å.  The HNH motif is bound at the minor groove primarily to DNA phosphate groups at and beyond the 3’-side of the scissile phosphate, with little interaction with ribose groups and bases.  This result provides a structural basis for sugar and sequence independent DNA recognition and the inhibition mechanism by inhibitor Im7, which blocks the substrate binding site but not the active site.  Structural comparison shows that two families of endonucleases bind and bend DNA in a similar way to that of the HNH ColE7, indicating that endonucleases containing a “ba-metal” fold of active site possess a universal mode for protein-DNA interactions.





Backbone Dynamics and Thermodynamics of Free Escherichia coli Theosterase / Protease I and in Michaelis complex with Di-Ethyl-p-Nitrophenyl Phosphate

Ching-Yu Chou2*,1Sergiy I. Tyuhktenko1 and Tai-Huang Huang1,2 

1Institute of Biomedical Sciences, Academia Sinica, Nan-Kang, Taipei 11529, Taiwan and

2National Taiwan Normal University, Taiwan

No.128 Academia Road Section 2, Nan-Kang, Taipei, 11529, Taiwan


E. coli thioesterase/protease I (TEP-I) is a serine protease of the SGNH-hydrolase family. The residues involved in the catalytic process include the catalytic triad of Ser10, Asp154 and His157 and the oxyanion hole groups have been identified as the amide groups of Ser10 and Gly44 and the side chain of Asn73. We have applied 15N nuclear magnetic spin relaxation method to investigate the multiple time scale temperature dependent backbone motions in E. coli Theosterase / Protease I(TEP-I) and study the backbone motion of enzyme with inhibitor, Di-Ethyl-p-Nitrophenyl Phosphate (DENP).

Laboratory frame relaxation data at 278K, 285K, 295K, and 310K were analyzed by using modelfree and reduced spectral density approaches. Analysis of the relaxation measurements yields order parameters (S2) that reflect the degree of spatial restriction for backbone amide H-N vectors. From characters of thermodynamics analysis, we found that the TEP-I molecule as a whole is rather rigid. On the other hand, the active site pocket exhibits high degree of flexibilities with high entropy. The thermodynamic parameters extracted from the analysis will be presented.

The biological consequences and the changes in dynamics and thermodynamics at the intermediate stages of catalysis, i.e. the Michaelis complex and tetrahedral complex, are almost similar with free form. And the dynamic parameters extracted from the complex will be presented.





Structures and Electron Micrographs of Volvatoxin A2 Reveal pH-dependent Protein Assembly and Implications for Membrane Pore Formation

Su-Chang Lin*, Yu-Chih Lo , Jung-Yiaw Lin and Yen-Chywan Liaw

1Graduate Institute of Life Sciences, National Defense Medical Center, Taipei,

Taiwan 114, Republic of China

2Institute of Molecular Biology, Academia Sinica, Taipei, Taiwan 115, Republic of China

128 Academia Road Section 2, Nankang, Taipei 11529, Taiwan


Protein binding to and inserted into membranes are mechanisms essential to all the organisms, but, as yet remain largely unknown. Membrane pore-forming toxins (PFTs) are potential model systems for studying these mechanisms. We have determined the crystal structure of volvatoxin A2 (VVA2), a fungal PFT from Volvariella volvacea, using Br-MAD. Here we report three structures of VVA2 at different pHs, 5.5, 4.6 and 6.5, and one crystal structure of a VVA2 isoform. VVA2 consists of 199 amino-acid residues. The 1.42 Å and the 2.6 Å crystal structures of VVA2, obtained at pH 4.6 and pH 5.5, respectively, suggest that VVA2 molecules are organized as a weak oligomer in solution. These crystal structures and electron micrographs of VVA2 molecules in solution show that they can assemble into filament-like oligomers in a pH-dependent way. In addition, the 3.2 Å crystal structure of VVA2, obtained at pH 6.5, suggests that it can form a two-layered arc-shape oligomer or even a two-layered helical oligomer with 18 subunits per turn. This arc-shape oligomer may mimic the assembly of VVA2 on cell membranes. A unique ring model with 9-fold symmetry is postulated to elucidate the pore-formation mechanisms of VVA2.





The O-Glycosylated Linker Region Of Glucoamylase Affects Structure And Function Of The Starch-Binding Domain

Wei-Ting Liu*, Wei-I Chou, Shu-Chuan Lin and Margaret Dah-Tsyr Chang#

Institute of Molecular and Cellular Biology & Department of Life Science,

National Tsing Hua University, Hsinchu, Taiwan


The mature glucoamylase from Rhizopus oryzae consists of an N-terminal starch-binding domain (SBD, 106 aa), an o-glycosylated linker (L, 36 aa), and a C-terminal catalytic domain (CD, 437 aa).  The SBD processes the raw-starch binding ability and the CD can hydrolyze starch completely to glucose.  As for the linker region, the information is still quite limited.  We have constructed two recombinant clones, SBD and SBD-L.  Each protein was expressed in both E. coli and S. cerevisiae strains, directly purified by fast protein liquid chromatography.  Circular Dichroism analysis revealed that SBD possesses a β-barrel conformation and the existence of linker may influence the structure stability.  We found that SBD-L formed amyloid-like structures at high temperature employing Circular Dichroism analysis, Congo Red assays and transmission electron microscopy.  However, SBD along could only form insoluble aggregates under the same condition.  Further binding and amylolytic assays demonstrated that the dissimilar structures among these recombinant proteins resulted in the variations observed in the experiments.  Our results show that the linker region influences the folding of the starch-binding domain.





PRODEC : PROtein structural patterns Detector using Evolutionary Computing

Yin H. Chen(陳音璇) 1, Yuh J. Hu(胡毓志)* 1 and Ming J. Hwang(明經)* 2

1National Chiao Tung University Department of Computer & Information Science,

2Institute of Biomedical Science, Academia Sinica

No.1001Ta Hsueh Road, Hsinchu, Taiwan 300, ROC


Protein structures are well known highly related to their function. It will be much useful to using motifs to represent the most information. In this study we describe a new computational method, PRODEC (PROtein structural patterns Detector using Evolutionary Computing), to automatically discover sequence-structure patterns in proteins. PRODEC, based on genetic computation, begins with an initial population of random patterns. Through the evolutionary process, PRODEC iteratively improves the statistical significance of patterns by modifying their configurations. To evaluate each new pattern, a novel scoring function is developed to measure patterns that are both conserved in 1D sequence and 3D structure. A key feature of PRODEC is that each PRODEC pattern is a duo consisting of a sequence pattern and a structure pattern. Compared with conventional methods or their extensions, PRODEC has better performance to find flexible and long patterns. We expect the sequence-structure patterns to be useful in predicting protein structures from sequences, so as well in annotating protein functions.




Structural determination of the Mite Group 7 Allergen from

Dermatophagoides pteronyssinus

Tse-Hao Huang* and Shwu-Huey Liaw

Structural Biology Program, Institute of Biochemistry, Faculty of Life Science, National Yang-Ming University, Taipei, Taiwan

155, Li-Nong St., Sec.2, Peitou, Taipei, 11221, Taiwan


Allergic diseases such as atopic dermatitis, rhinitis and asthma afflict more than 20% of the world population. House dust mites have been regarded as one of the most important sources of indoor allergens. The two most important clinical mite species are Dermatophagoides pteronyssinus and Dermatophagoides farinae. The group 7 allergen from D. pteronyssinus (Der p 7) consists of a 17-residue leader peptide and a 198-residue mature protein. This allergen has been shown to react with about 50% of allergic sera. Our sequence analysis suggests that Der p 7 does not share significant homology to any other proteins, and may contain a novel protein fold. To gain structural insights into the allergenicity mechanism, we have obtained the first Der p 7 crystals and attempt to solve the phase problem using the selenomethionyl multiwavelength anomalous dispersion method.

The Der p 7 crystals have been grown at 15 oC in 20% polyethylene glycol 4000, 20 mM ammonium acetate and 100 mM sodium citrate (pH 5.6). The crystals belong to the P2221 space group with unit cell dimensions of a = 53.43 Å, b = 79.83 Å, c = 366.62 Å. The crystals diffract x-rays beyond 3.2 Å resolution. A self-rotation search revealed a six-fold point symmetry and thus suggested six monomers in each asymmetric unit. Furthermore, the circular dichroism spectra suggests that Der p 7 is a helical protein. Analytical ultracentrifugation analysis reveals that Der p 7 exists as in solution mainly as a monomer with trace amounts of trimer and hexamer.




Structural Analysis of the Cytidine Deaminase Superfamily: Conservation, Divergence, and Prediction

Cheng-Tsung Lai1* and Shwu-Huey Liaw2

1Bioinformatics Program, Institute of Genetics and 2Structural Biology Program, Faculty of Life Science, National Yang-Ming University, Taipei, Taiwan

155, Li-Nong St., Sec.2, Peitou, Taipei, 11221, Taiwan


One major challenge in the post-genomic era is to understand how nature has evolved one structural fold of various proteins for their different functions. Structural analysis of homologous proteins in a structural superfamily would provide an efficient path to understand the structural plasticity for the functional versatility. Our analysis has revealed that cytidine deaminases (CDAs), fungal cytosine deaminase (CD), B.subtilis, plant and some bacterial guanine deaminase (GD), dCMP deaminase (dCMPD), RNA editing enzymes and riboflavin biosynthesis protein RibG belong to the CDA superfamily. These deaminases catalyze the zinc-assisted conversion of the amino group of the cytosine, guanine or adenine moiety into a keto group.

To identify the putative substrate-recognition residues for each member, a comparative analysis of the available structures in the superfamily was first carried out. Then multiple sequence alignments were performed by the program ClustalW, followed by manual editing according to the structural information and secondary structure prediction using PSI-PRED. Finally, atomic models were built based on the known structures to reveal whether the conserved residues are potentially localized nearby the active-site cavity. The putative substrate-interacting residues for dCMPD and RibG have identified, while those for RNA editing enzymes are still under investigation. Obviously, these deaminases utilize different residues to interact with the common ribose and phosphate groups of their substrates. Our results also reveal that a large portion of the conserved residues are responsible for the superfamily's structural plasticity, whereas only a small portions are directly involved in its functional versatility. These prediction results would provide a structural basis for mutational analysis to elucidate the functional roles of the critical residues.




Crystal Structure of Bacillus subtilis Guanine Deaminase

Yu-Jui Chang1* and Shwu-Huey Liaw2

1Institute of Biotechnology in Medicine and 2Structural Biology Program, Faculty of Life Science,

 National Yang-Ming University, Taipei, Taiwan

155, Li-Nong St., Sec.2, Peitou, Taipei, 11221, Taiwan


Guanine deaminase (GD), a key enzyme in the nucleotide metabolism, catalyzes the hydrolytic deamination of guanine into xanthine. The first crystal of the 156-residue guanine deaminase from Bacillus subtilis was grown in polyethylene glycol 4000, ammonium acetate, sodium citrate. The crystals belong to space group C2221 with unit-cell parameters a = 84.91, b = 90.90, c = 80.19 Å, with one dimer per asymmetric unit. The structure has been solved at 1.15 Å resolution by the Se-MAD method. Unexpectedly, the C-terminal segment is swapped to form an inter-subunit active site and an intertwined dimer with an extensive interface of 3900 Å2 per monomer. The essential zinc ion is ligated by a water molecule together with His53, Cys83 and Cys86. The reaction intermediate was modeled into the active-site cavity based on the tightly bound imidazole and water molecules, showing the conserved deamination mechanism and the specific substrate recognition by Asp114 and Tyr156’. The closed conformation reveals that substrate binding seals the active-site entrance, which is controlled by the C-terminal tail. Therefore, the domain swapping in GD contributes not only for the oligomerization, the structural stability, but also for the substrate specificity. Finally, detailed structural comparison of the cytidine deaminase superfamily illustrates the functional versatility of the divergent active sites in guanine, cytosine and cytidine deaminases, and suggests putative specific substrate-interacting residues for other members such as the dCMP deaminase.




Crystal Structure of Hepatitis C Virus NS3 Helicase

Yun-Lin Wang* and Shwu-Huey Liaw

Structural Biology Program, Institute of Biochemistry and  Faculty of Life Science, National Yang-Ming University, Taipei, Taiwan

155, Li-Nong St., Sec.2, Peitou, Taipei, 11221, Taiwan


Hepatitis C virus (HCV), a positive-stranded RNA virus of the Flaviviridae family, is the major causative agent of transmitted non-A, non-B hepatitis and currently infects approximately 3% of the world’s population. The infection easily results chronic hepatitis, and then may lead to liver cirrhosis and hepatocellular carcinoma. Unfortunately, no effective treatments for hepatitis C virus are available. Nonstructural protein 3 (NS3) is a bifunctional enzyme with the N-terminal 180-residue serine protease domain and the C-terminal 450-residue helicase domain. The homologous NS3 proteins in all pesti-, poty-, and flaviviruses sequenced to date contain conserved helicase motifs, suggesting that NS3 helicase plays an important role in the virus life cycle and may be a potential target for developing antiviral drugs.

Helicases are enzymes which unwind duplex DNA or RNA in couple to NTP binding and hydrolysis. HCV NS3 helicase domain has been shown to unwind both RNA and DNA homo- and hetero-duplexes in a 3’ to 5’ directionality with strict requirement of a 3’ tail. Sequence comparisons have defined several helicase superfamilies and classified HCV NS3 helicase in superfamily II (SF2) with seven conserved motifs. The functional roles of the conserved motifs have been extensively studied by structural and mutational analyses. However, some motifs have not been clearly defined, even with conflicting results. Here we attempt to determine the crystal structure of HCV NS3 helicase in complex with the cofactor GTPγS and oligonucleotide. The crystals have been grown in 2.4 M NaCl, 0.1 M ADA (pH 6.5) and 2% glycerol. The crystals belong to the I4122 space group with unit-cell dimension of a = b = 196.24 Å, c = 127.55 Å, and with two molecules per asymmetric unit. The structure was determined by the Molecular Replacement method. Structural refinement and comparison are still under investigation.




Solution structure of Tc32, an inhibitor of K+ channel from venom of

Brazilian scorpion, Tityus cambridgei.

I-Che Huang1,2,3, Yuan-Chau Lou2, Iren Wang3, Shin-Hsiung Wu1,3, Chinpan Chen2

1Institute of Biological Chemistry, Academia Sinica, Taiwan

2.Institute of Biomedical sciences , Academia Sinica, Taiwan

3 Institute of Biomedical Sciences, National Taiwan University, Taipei 106, Taiwan

No.128 Academia Road Section 2, Nan-Kang, Taipei, 11529, Taiwan


Ion channels are involved in diverse biological processes and play essential roles in the physiology of all cells. An increasing number of human and animal diseases have been identified to relate with the defective function of ion channels. Scorpion venoms contain various polypeptides (channel blockers) that particularly affect the permeability of ion channels in cell membranes. Thus, understanding the structural basis of the specificity of scorpion toxins for these receptors could lead to the design of new ligands with controlled activity and potency for clinical applications. Tc32 is a 35 amino acids peptide from the venom of Brazilian scorpion, Tityus cambridgei, which exhibits blocking activity against Kv1.3 and Shaker B potassium channels with different Kd values and is classified as the first member of the new subfamily of  α-KTx(α-KTx18.1). Sequence homology analysis shows that the two key residues, Lys and Tyr at C-terminus previously confirmed essential for activity, are mutated in Tc32. We synthesized Tc32 with solid phase synthesis. After refolding and purification, two dimensional NMR and Circular Dichrorism were applied to determine the feature of secondary structure and three-dimensional solution structure of Tc32.  Based on the specific NOEs, 3 disulfide bonds were unambiguously identified (C7-C26, C12-C31, and C16-C33) that is the same as native Tc32 identified by enzyme digestion and mass analysis. Tc32 structure shows that it consists a flexible region at N-terminus, and a rigid secondary structural scaffold with anα-helix and a double-stranded antiparallel βsheet adopted by several other scorpion toxins. Although the rigid parts of the structure in Tc32 are similar with other scorpion toxins, the sequence alignment indicated that Tc32 may recognize the potassium channel using different residues. Comparing the structure of Tc32 and other scorpion toxins, structural and functional implications will be discussed in this poster.




Solution structure of the hypothetical protein HP0495 from

Helicobacter pylori

Chui-Lin Chiu, Ming-Tao Pai, C. C. Huang, Ya-Ping Tsao, Jya-Wei Cheng*
Institute of Biotechnology, National Tsing Hua University, Hsinchu 300, Taiwan, R.O.C.



Structural genomics attempts to provide three-dimensional structural information about a significant fraction of the proteins encoded by the genes sequenced in various genome projects. In this study, we use NMR to determine the solution structures of HP0495, a hypothetical protein from the human pathogenic bacteria Helicobacter pylori. HP0495 has 4 similar sequences referred to as 'hypothetical protein' by PSI-BLAST analysis against a non-redundant sequence database and is not associated with any known function based on its amino-acid sequence. Solution structure analysis revealed that the overall fold of this protein consists beta-strands and alpha-helices. We also investigate the conformational characteristics by hydrogen-deuterium exchange, and molecular dynamics. Furthermore, biochemical functions are analyzed to help the proteomic studies of the roles of Helicobacter pylori in human diseases.




C-Terminal Truncation of Helicobacter pylori Fucosyltransferase The Region of Tandem Repeats Associated with Structure and Catalysision

Andrew H. Sheng-Wei Lin,1 Jen-Ru Li,1 Tsui-Min Yuan,1 Chun-Hung Lin1,2*

1Institute of Biochemical Sciences, National Taiwan University, Taipei, Taiwan

 2Institute of Biological Chemistry, Acdemia Sinica, Taiwan, No.128 Academia Road Section 2,

Nan-Kang, Taipei, 11529, Taiwan


Helicobacter pylori is an important pathogenic bacterium to cause gastric and duodenal ulcers in human. The infection is often associated with the occurrence of gastric cancer and lymphoma. To mimic the host’s cell-surface carbohydrates, this microorganism expresses fucose-containing Lewis x and Lewis y antigens as the specific epitopes on cell surface to prevent the immune surveillance. Herein we report the expression of the H. pylor a1,3-fucosyltransferase (FucT) in E. coli. that contains a tandem repeat of leucine zipper as well as a domain rich of basic residues in the C-terminal. Several mutants of different size were prepared by C-terminal truncation at different length in order to investigate the functional roles played by the two regions and improve the formation of soluble proteins. The results indicated that up to 80 amino acid residues, the more residues were deleted, the higher amount of the soluble form was obtained. All truncated FucTs catalyzed the transfer of fucose from GDP-fucose to N-acetyllactosamine at 5 mmole/min/mg. Further characterization indicated that the changes in quaternary structure (based on the study using analytical ultracentrigufation) and stability (by thermal denaturing) were observed in association with the C-terminal deletion.




NMR Structure of a Conserved Hypothetical Protein Xcc 975 from Xanthomonas Campestris pv. Campestris

Ko-Hsin Chin, Yu-Chen Hu, Jhe-Le Tu, Fu-Yang Lin, Shan-Ho Chou1,2

1Institute of Biochemistry, National Chung-Hsing University, Taichung, 40227, Taiwan

2Department of National Central University, Jung-Li, Taoyuan, 300, Taiwan


Xanthomonas campestris pv. campestris (Xcc) is a gram-negative phytopathogen that causes black rot in cruciferous plants. The solution structure of a conserved hypothetical protein, Xcc 975, from such a pathvar has been determined by using NMR technology. Xcc 975 belongs to the family PF01722 of BolA function in the Pfam databases for which no structural information is yet available. In the Pfam database, Xcc 975 of BolA-like function is equivalent to 0271 COG (Clusters of Orthologous) in the NCBI database, and is a stress-induced morphogen, which in E.coli causes round morphology when over-expressed. In the mean time, BolA may also participate in many other biological functions, including cell division, responses under unfavorable circumstances, and induction of the transcription of penicillin-binding proteins.

Xcc 975 is a single polypeptide chain of eighty amino acids with a molecular weight of 9 kDa. The solution structure of Xcc 975 was determined using NMR methodology and molecular dynamics techniques. The protein backbone HN, N, CA, and CB atoms were sequentially assigned using 2D 15N-HSQC, 3D heteronuclear triple-resonance HNCACB, CBCA(CO)NH, HNCO, and HNCACO spectra. The side chain atoms were assigned using 3D HCCH-COSY, HCCH-TOCSY, H(CCCO)NH, and (H)C(CCO)NH experiments. The HB2 and HB3 protons were stereospecifically assigned and the Chi-1 torsion angles determined using HNHB and HN(CO)HB spectra. The secondary structures were determined by CSI program and the backbone torsion angles f and q calculated by TALOS program. All these constraints, along with the NOEs determined from 3D-NOESY-15N-HSQC and NOESY-13C-HSQC spectra, were used as inputs for the CYANA program to calculate the final protein structures. Analysis of the results reveals a protein structure with a core comprising of a β-sheet, which includes paired adjacent anti-parallel β strands looped by a type-I turn, with the third β strand pairing in parallel with the second β strand. Threeαhelices were found connecting theβ strands with the topology of α-β-β-α-β-α. The structures and functions of the protein will be discussed in this report.




Structural and functional relationship of HATH domain of Human Hepatoma-derived Growth Factor

Shih-Che Sue1*, Wei-Tin Li1, Yu-Chieh Lin1, Shao-Chen Lee2, Wen-guey Wu2,

Jeou-Yuan Chen1and Tai-huang Huang1

1Institute of Biomedical Sciences, Academia Sinica, Nankang, Taipei, Taiwan, R.O.C.

2Institute of Bioinformatics and Structural Biology, College of Life Sciences,

National Tsing Hua University, Hsinchu, Taiwan, R.O.C.


Human hepatoma-derived growth factor (hHDGF) is a heparin-binding protein that was implicated in tumorigenesis, vascular development, cell proliferation and transcriptional activation. NMR solution structure reveals that hHDGF comprises an ordered N-terminal domain (HATH domain, ~ 100 a.a.) and a highly disordered C-terminal domain. The N-terminal HATH domain is capable to specifically interact with heparin. The disordered C-terminal region exhibits no heparin-binding ability but shows transformation activity to enhance cell saturation density and growth rate in H1299 cell. Meanwhile HATH domain is demonstrated to aid fluorescence protein, DsRED, to internalize into cytoplasm and nuclear. Based on the observation, HATH domain is proposed to serve as a carrier for protein internalization through binding with heparin near cell surface. Furthermore, two forms of HATH domain corresponding to monomer and dimer can be separated in physiologic condition by FPLC. The interconversion between monomer and dimer occurs slowly, however refolding of HATH domain from 8M urea-denatured state can convert monomer or dimer to the equilibrium proportion. The result indicates a large kinetic barrier that separates the two states and only decreasing kinetic barrier under denatured condition can speed up the equilibrium. A probable domain swapping is proposed. The dimeric HATH domain shows strong heparin-binding affinity of Kd ~ 10 nM that monomer lost such binding-behavior. Structural model is proposed to allow us to prospect the correlation between domain-swapping property and the heparin-binding mediated cell entry near cell surface.




Mithramycin forms a stable dimeric complex by chelating with Iron(II): DNA-binding characteristic, subcellular distribution and biological

activity of this dimer complex


Ming-Hon Hou1,2 and Andrew H.-J. Wang 1,2*

1Institute of Biological Chemistry, Academia Sinica, Taipei, 115 Taiwan

2 Institute of Biochemical Sciences, National Taiwan University, Taipei, 106 Taiwan


The aureolic antibiotic mithramycin (Mith) is a DNA-binding anticancer drugs used clinically.  The Mith-d(TGGCCA)2 complex model reveals the coordination around the metal ion is in octahedral geometry by binding to the two oxygen atoms of each chromophore and two water molecules as the fifth and sixth ligands.  Herein, Mith formed a highly stable 2:1 drug-metal complex through the chelation with Fe2+ ion was studied by CD spectroscopy.  The chelating effect of Fe2+ is much greater than other divalent metal ions including Mg2+, Zn2+, Co2+, Ni2+ and Mn2+.  The [(Mith)2-Fe2+] complex binds to DNA and induces a conformational and thermodynamic change of DNA as shown by CD and UV thermal spectral analyses, respectively.  This dimer complex maintains structural integrity in cells based on intracellular Fe2+ assays and locates exclusively in the nucleus of K562 cells.  In DNA-break assay, the result showed that [(Mith)2-Fe2+] complex was able to promote the cleavage of plasmid DNA.  The [(Mith)2-Fe2+] complex exhibited higher cytotoxicity than the respective drugs alone in some cancer cell lines may due to its DNA cleavage activity.  Our results herein not only propose a possible anticancer mechanism for Mith but also suggest the Fe2+ derivatives of Mith as potential target to be explored in the future.




Structural Genomics of Xanthomonas campestris pv. campestris using High-Resolution NMR Techniques


Ko-Hsin Chin1, Jhe-Le, Tu1, Yu-Chen Hu2, Chao-Yu Yang1, Chiao-Li Chu1, Fu-Yang Lin2, Yen-Yu Wu2,

Kun-Chou Wei1, Yi-Ting Huang1, Wei-Ting Kuo1, Ji-Wei Lo1, Ping-Chiang Lyu3, Shih-Feng Tsai4, Shan-Ho Chou1,2

1Institute of Biochemistry, National Chung-Hsing University, Taichung, 40227, Taiwan

2Department of Life Science, National Central University, Jung-Li, Taoyuan, 300, Taiwan

3Institute of Biotechnology, National Tsing-Hwa University, Hsin-Chu, Taiwan

4Institute of Genetics, National Yang-Ming University, Taipei, Taiwan


The flood of sequence information available from the various genome projects coupled with the recent advances in molecular and structural biology has led to the concept of structural genomics on a genome-wide scale. Determination of protein three-dimensional structures is crucial for understanding their biological functions. In this respect, structural genomics is expected to pave way for understanding the intricate interactions among proteins in the whole organisms. Xanthomonas campestris is a gram-negative bacterium that is phytopathogenic to cruciferous plants and causes worldwide agricultural loss. However, it also produces exopolysaccharide (xanthan) that is of great industrial importance. Owing to the immense economic impact and the simpler genetic information (lacking introns), we have started a large scale program to determine and characterize the protein structures and functions in Xanthomonas campestris using high-resolution high-field NMR techniques. The final goal of this coordinated project is to obtain a structural database of Xanthomonas campestris containing all structural and function annotations.

This pathogen genome is first sequenced and its genetic content analyzed by bioinformatics method through our collaborators. The genes of interests are amplified by PCR and engineered to vectors containing GST, MBP, or THX gene segment for protein expression in high yields. After cleavage of fusion partner, the desired proteins are further purified for NMR studies.

The resulting 2D and 3D NMR spectra are processed by NMRPipe/NMRDraw programs, and analyzed by Sparky program. The protein backbone HN, N, CA, and CB atoms were sequentially assigned using 2D 15N-HSQC, 3D heteronuclear triple-resonance HNCACB, CBCA(CO)NH, HNCO, and HNCACO spectra. The side chain atoms were assigned using 3D HCCH-COSY, HCCH-TOCSY, H(CCCO)NH, and (H)C(CCO)NH spectra. The HB2 and HB3 protons were stereospecifically assigned and the c1 torsion angles determined using HNHB and HN(CO)HB spectra. The methyl groups were analyzed by constant-time (HM)CMC(CM)HM experiment, and the stereospecific assignments of the Pro-R and Pro-S methyl groups of Val and Leu made on the basis of the relative signs of cross-peaks observed in the 13C-1H CT-HSQC spectra of 15N, 10% 13C-enriched samples. The secondary structures were determined by CSI program and the backbone torsion angles f and q calculated by TALOS program. All these constraints, along with the NOEs determined from the 3D-NOESY-15N-HSQC, NOESY-13C-HSQC spectra, especially the critical methyl-methyl NOEs determined from the 3D 13C-13C NOESY (HSQC-NOESY-HSQC) or (HM)CMCB(CMHM)-NOESY spectrum, were served as inputs for CYANA program to calculate the final protein structures.

   Through the above-mentioned approaches, over 150 genes have been constructed in high-yield expression vectors, with more than twenty of the gene products exhibiting well resolved 2D 1H-15N HSQC spectra suitable for further structural studies. Several protein structures are being refined to yield final structures. The recent progress of this project will be reported.





Crystal Structure of CTX A3 and Hexasaccharide Heparin Complex from

Naja Atra

Hong-Hsiang Guan1,2*, Shao-Chen Lee 2,Chia-Hui Wang 2,Wei-Ning Huang 2 ,

Wen-Guey Wu 2 , Chung-Jung Chen 1

1Biology group,Synchrotron Radiation Research Center, Hsinchu, Taiwan

2 Institute of Bioinformatics and Structural Biology, National Tsing-Hua University, Hsinchu, Taiwan

Biology group,SynchrotronRadiation Research Center,Hsinchu 300,Taiwan


Cardiotoxin (CTX) is a major component of cobra toxin. Cobra cardiotoxins (CTXs) are basic proteins, composed of 60-62 amino acids, in which b-sheets form three finger-loop structures. When cobra bites animals and human, CTX can induce tissue inflammation. However, the CTX major target on cell membrane is still unclear now. Our previous studies have proved that heparan sulfate is the most potential target of CTX on cell membrane. We have determined the crystal structure of CTX A3 and hexasaccharide complex at 3.4 Å resolution using synchrotron radiation X-ray. Through the complex structure, we found that citrate anion plays an important role in prompting CTX A3 to form dimmer by interacting with Lys31 and Lys23.  The finding is important because anionic citrate is a major component (~50mM) of venom, but the specific role of citrate is still poorly understood.  Besides, Hexasaccharide heparin bind to CTX A3 by interacting with positively charged residue Lys12, Lys18 and Lys35. The binding site of hexasaccharide heparin from X-ray is similar to that of the disaccharide from NMR study.  It is also worthy to discuss that one of  CTX A3 dimmer is heparin bound form and the other is heparin non-bound form. This study suggests a novel role for venom citrate activity and identify specific sulfation pattern of heparan sulfate in binding to CTX A3.




Anaerobic and Aerobic Structures of Ferredoxin ll from Desulfovibrio Gigas Reveal Electron Transfer Mechanism

Yin-Cheng Hsieh,1,2* Ming-Yih Liu,4 and JeanLeGall4,Chun-Jung Chen,1, 3

1Biology Group, National Synchrotron Radiation Research Center, Hsinchu 30077, Taiwan;

2Institute of Bioinformatics and Structural Biology; 3Department of Physics,

National Tsing-Hua University, Hsinchu 30077; 4Department of Biochemistry

and Molecular Biology, University of Georgia, Athens, GA 30602, U. S. A.


Ferredoxin II (Fd II) is a small electron transfer protein, isolated from the strict anaerobic sulfate-reducing bacterium, Desulfovibrio gigas.  The protein contains 58 amino acids and an iron-sulfur cluster.  The cluster [3Fe-4S] spontaneously undergoes conversion to [4Fe-4S] when it is used as an electron mediator in the phosphoroclastic reaction.  This two-form interconversion appears to have physiological significance.  We have recently obtained both aerobic and anaerobic Fd II crystals of the high-resolution quality.  Both structures are independently determined by the iron single-wavelength anomalous dispersion (Fe-SAD) method using synchrotron radiation X-ray. 

The structure of aerobic Fd II has been refined to 0.9 Å ultra-high resolution in the space group P21212.  Its [3Fe-4S] cluster is bound with Cys8, Cys14, and Cys50, whereas Cys11 extends away from cluster.  Cys18 and Cys42 form a disulfide bridge to maintain the protein folding.  Five isolated Zn2+ ions around the protein are located and bound with Glu, Asn and Asp, respectively, which indicates the transition metals, other than iron, could be incorporated into [3Fe-4S] center.  On the other hand, the anaerobic Fd II structure from the crystals grown under anaerobic condition has also been refined to 1.4 Å resolution.  The anaerobic structure shows the different iron-sulfur cluster, disulfide bridge, Cysteine resiudue conformations as well as crystal packing.  Here we present the structure comparison between aerobic and anaerobic Fd II at ultra-high resolution which reveals the unique iron-storage function and electron transfer mechanism of ferredoxin II from Desulfovibrio gigas.




Self-assembly of archaeal RadA proteins into long and fine helical filaments

Ming-Hui Lee*1,2, Chih-Hsiang Leng2, Yuan-Chih Chang3, Yi-Kai Chen1, Chia-Seng Chang3, Fu-Fei Hsu2,

 Andrew H.-J. Wang1,2and Ting-Fang Wang1,2

1Institute of Biological Chemistry, Academia Sinica,, Taipei 115, Taiwan

2National Core Facilities of High Throughput Protein Production, Taipei 115, Taiwan and

3Institute of Physics, Academia Sinica, , Taipei 115, Taiwan

No.128 Academia Road Section 2, Nan-Kang, Taipei, 115, Taiwan


Archaeal RadA protein, like bacterial RecA and eukaryotic Rad51 proteins, is able to promote strand exchange between DNA strands with homologous sequences. We applied the parallel cloning and expression system to screen for soluble RadA fusion protein in E. coli. Using a carbon nanotube(CNT) tip for high-resolution atomic force microscopy (AFM) image, we report for first time that RadA proteins can form long (up to 1mm in length) and fine helical filaments (10nm pitch) in the absence of DNA.




Molecular Visualization of Yeast Dmc1-ssDNA complexes by Atomic Force Mcroscopy Using Carbon Nanotube Tips

Yuan-Chih Chang2, Yu-Hui Lo1*, Chih-Hsiang Leng1,3, Ming-Hui Lee1, Douglas K. Bishop4,

Chia-Seng Chang2, and Ting-Fang Wng1

1Institute of Biological Chemistry, Academia Sinica, Nan-Kang, Taipei 11529, Taiwan,

2Institute of Physics, Academia Sinica, Nan-Kang, Taipei 11529, Taiwan,

3the Vaccine Research and Developmental Center, National Health Research Institute. Taipei, Taiwan, and

4the Department of Radiation and Cellular Oncology, University of Chicago, Illinois 60637


Saccharomyces cerecisiae DMC1 gene encodes the meiosis-specific RecA homologue protein that catalyzes singl-end invision (SEI) reaction during meiotic DNA recombination. Here we report the high-resolution images of yeast Dmc1 proteins obtained by atomic force microscopy (AFM) using carbon nanotube (CNT) tips. We show that yeast Dmc1 protein from ring structures in the absence of DNA. In contrast, no single or stacked yeast Dmc1 protein rings were observed on single stranded DNA substrates. Thus, yeast Dmc1 proteins from nucleoprotein filaments that are resemble to those of the bacterial RecA and yeast Rad51 proteins, and subsequently promote homologous DNA recombination.




Purification, Crystallization and preliminary X-ray crystallographic analysis

of Cystein-rich Secretory Protein(CRISP) from Naja atra

Yu-Ling Wang1,2*, King-Xiang Goh 2, Wen-Guey Wu 2 and Chung-Jung Chen 1

1Biology group,Synchrotron Radiation Research Center, Hsinchu, Taiwan

2 Institute of Bioinformatics and Structural Biology, National Tsing-Hua University, Hsinchu, Taiwan

Biology group,SynchrotronRadiation Research Center,Hsinchu 300,Taiwan


Cysteine-rich secretory proteins (CRISPs) may play a role in the innate immune system and are transcriptionally regulated by androgens in several tissues. They are mostly found in epididymis and granules of mammals. A number of snake venoms, from habu snake, erabu sea snake, Conus textile, etc., contain CRISP family proteins. Natrin has a cysteine-rich C-terminal tail and belongs to a family of CRISPs. We have purified the natrin protein from Naja atra (Taiwan cobra) venom using a three-step chromatography procedure. The protein was crystallized using the hanging-drop method. Data were collected to 1.7 Ǻ resolution using synchrotron radiation and the crystals belonged to space group C2221 , with unit-cell parameters a = 59.172, b = 65.038, c = 243.156 , α = β = γ=90ú . There are two protein molecules in the asymmetric unit. Matthews coefficient is estimated to be 2.32 Å3  Da-1 and a solvent content of 45.13%.