Solution Structure and Dynamics of Onconases with N-terminal and

Met23 Mutations


Vitaliy Y. Gorbatyuk#, Cheng-Kun Tsaix,†, Chi-Fon Chang# and Tai-huang Huang#,x,*

#Institute of Biomedical Sciences, Academia Sinica, Nankang, Taipei, Taiwan, Rep. of China

xDepartment of Physics, National Taiwan Normal University, Taipei, Taiwan, Rep. of China


Onconase (ONC)1, ranpirnase or P-30 protein, initially purified from the extracts of Rana pipiens oocytes and early embryos has been shown to exhibit anticancer activity both in vitro and in vivo and is in human phase III clinical trials for tumor therapy. We have determined the solution NMR structure of a recombinant onconase (M-1, Q1, M23L)rONC1. The 20 best solution structures have a backbone root-mean-square-deviation (r.m.s.d.) of 0.41±0.09 Å and heavy atoms r.m.s.d. of 0.80±0.06 Å for all residues with respect to the average structure. The energy minimized average NMR structure has a backbone r.m.s.d. of 0.72 Å from the X-ray crystallographic structure of native onconase. However, the orientations of the first N-terminal residues in the two structures are very different. Comparison of the 15N-HSQC spectra of (M-1, Q1, M23L)rONC and E1S-rONC shows the presence of substantial shifts of resonances assigned to residues at regions far beyond the vicinity of the mutation sites. The shifts are attributed to changes in dynamics rather than structure. Model-free analysis of the backbone amide 15N-T1, 15N-T2, and 15N-1H NOE relaxation data revealed that a1-helix and the b-sheets of (M-1, Q1, M23L)rONC display slow bending motion whilst E1S-rONC is much more rigid. The drastic differences in dynamics behavior between these two mutants offers an explanation for the change in stability and increased enzyme susceptibility. It also suggests the formation of hydrogen bonding network involving the N-terminal residue in E1S-rONC but not in (M-1, Q1, M23L)rONC.




A DFT/CDM Study of the Factors Governing the Protonation State of Zn-bound Histidine in Proteins


Yen-lin Lin1 and Carmay Lim*,1,2

Institute of Biomedical Science, Academic Sinica, Taipei 11529, Taiwan R.O.C., and

Department of Chemistry, National Tsing Hua University,, Hsinchu 300, Taiwan R.O.C.

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


We have performed systematic theoretical studies to elucidate the factors governing the His protonation/deprotonation state in Zn-binding sites, especially those containing the ubiquitous Zn-His-Asp/Glu triad. Specifically, we have addressed the following three questions: (1) How does the transfer of the Zn-bound His imidazole proton to the second-shell Asp/Glu carboxylate oxygen depend on the composition of the other first-shell ligands and the solvent accessibility of the metal-binding site? (2) Can any second-shell ligand with a proton acceptor group such as the backbone carbonyl oxygen also act as a proton acceptor? (3) What is the effect of the Asp/Glu in the Zn-His-Asp/Glu triad on the Zn-bound water protonation state? To address these questions, we used a combination of quantum mechanical and continuum dielectric methods to compute the free energies for deprotonating a Zn-bound imidazole/water in various Zn complexes. The calculations show that whether the Zn-bound His is protonated or deprotonated depends on (1) the solvent accessibility of the metal-binding site, and (2) the Lewis acid ability of Zn, which is indirectly determined by both the first- and second-shell Zn ligands. The calculations also show that the effect of the Zn-His-Asp/Glu interaction on the nucleophilicity of the Zn-bound water depends on the solvent accessibility of the catalytic Zn site. Furthermore, they show that the Asp/Glu side chain in the Zn-His-Asp/Glu triad can increase the negative charge of its partner, His, and create an anionic hole that may stabilize a cation in buried cavities, provided that the Zn complex is cationic/neutral. The findings of this work are in accord with available experimental data.




A method to predict ligand-binding sites from protein sequences

Tammy Man-Kuang Cheng*1,2 and Carmay Lim1, 2

1Department of Chemistry, National Tsing Hua University, Hsinchu 300

2Institute of Biomedical Science, Academia Sinica, Taipei 115


Loop regions often participate in forming ligand-binding sites and enzyme active sites. Here, we present a novel method to identify such "functional" loops from the protein sequence alone, without evolutionary or structural information. The method is based on the amino acid and backbone conformational preferences of functional loops, which are found to be distinct from those of alpha-helices, beta-strands, and non-functional loops. It can select at least 1 functional loop out of three candidates from a test set of 300 alpha/beta-, alpha-, and beta- proteins ~3-fold more successfully than random picking. It also successfully identified antibody-/ligand-binding site(s) in various virus proteins, Ig-alpha, and bird lysozymes.



A Combined Experimental and Theoretical Study of Divalent Metal Ion Selectivity and Function in Proteins: Application to E. coli Ribonuclease H1


C. Satheesan Babu1, Todor Dudev1, R. Casareno2, J. A. Cowan2, and Carmay Lim1,3* 

 1Institute of Biomedical Sciences, Academia Sinica,Taipei 11529, Taiwan, R.O.C.

 2Department of Chemistry, The Ohio State University,100 West 18th Avenue, Columbus, OH 43210, USA.

 3Department of Chemistry, National Tsing Hua University, Hsinchu 300, Taiwan, R.O.C.


Structural and thermodynamic aspects of alkaline earth metal dication (Mg2+, Ca2+, Sr2+, Ba2+) binding to E. coli Ribonuclease H1 (RNase H1) have been investigated using both experimental and theoretical methods. The various metal-binding modes of the enzyme were explored using classical molecular dynamics simulations, and relative binding free energies were subsequently evaluated by free energy simulations. The trends in the free energies of model systems based on the simulation structures were also computed using a combination of density functional theory and continuum dielectric methods. The calculations provide a physical basis for the experimental results and suggest plausible role(s) for the metal cation and the catalytically important acidic residues in protein function. Magnesium ion indirectly activates water attack of the phosphorus atom by freeing one of the active site carboxylate residues, D70, to act as a general base through its four first-shell water molecules, which prevent D70 from binding directly to Mg2+. Calcium ion, on the other hand, inhibits enzyme activity by preventing D70 from acting as a general base through bidentate interactions with both carboxylate oxygen atoms of D70. These additional interactions to D70, in addition to the D10 and E48 monodentate interactions found for Mg2+, enable Ca2+ to bind tighter than the other divalent ions. However, a bare Mg2+ ion with two or less water molecules in the first shell could bind directly to the three active-site carboxylates, in particular D70, thus inhibiting enzymatic activity. The present analyses and results could be generalized to other members of the RNase H family that possess the same structural fold and show similar metal-binding site and Mg2+-dependent activity.




Monodentate vs. Bidentate Carboxylate Binding in Magnesium and Calcium Proteins: What are the Basic Principles?


Todor Dudev1 and Carmay Lim1,2

1Institute of Biomedical Sciences, Academia Sinica, Taipei 11529, Taiwan, R.O.C.

2Department of Chemistry, National Tsing Hua University, Hsinchu 300, Taiwan, R.O.C..


Aspartate and glutamate side chains are unique among the 20 amino acids in possessing carboxylate groups that can bind the metal cation either monodentately (via one of the carboxylate oxygens) or bidentately (with both carboxylate oxygens coordinated to the metal). Although statistical analyses of structures deposited in the Cambridge Structure Database (CSD) and the Protein Data Bank (PDB) have shown that the monodentate mode is generally more common than the bidentate mode in both small molecules and metalloproteins, very little is known about the physical factors that govern the mode of carboxylate binding. A number of questions remain: (i) Does the mode of carboxylate binding depend on the type of metal and its coordination number? (ii) What is the role of the immediate neighbors of the carboxylates, particularly the metal-bound water molecules, in determining carboxylate denticity? (iii) To what extent does the solvent accessibility of the metal-binding site (dielectric medium) affect mono/bidentate equilibrium?

The aforementioned questions are addressed here using a combined DFT/continuum dielectric approach to elucidate the physical principles determining the mode of carboxylate binding in metalloproteins. Our attention is focused mainly on magnesium and calcium proteins in view of the importance of carboxylate denticity for their function and selectivity. We systematically investigate (i) the role of the carboxylate immediate surrounding in determining its binding mode, (ii) the effect of the total charge of the metal complex on carboxylate denticity, (iii) the role of the metal and its coordination number, and (iv) the effect of the relative solvent exposure of the metal-binding site. Furthermore, the PDB is surveyed for Mg and Ca binding sites containing acidic amino acid residues. Based on the theoretical findings, we rationalize the differences observed in the PDB between the preferred mode of carboxylate binding in Mg and Ca proteins.   




Crystal Structure of the BAFF and BAFF-R Complex: Implications for the Receptor Activiation


Jie-Oh Lee

Department of Chemistry, Korea Institute of Science and Technology, Daejeon, Korea

373-1 Kusong-dong, Yusong-gu, Daejeon, Korea


B-cell activating factor (BAFF) is a key regulator of B-lymphocyte development. Its biological role is mediated by the specific receptors BCMA, TACI, and BAFF-R. We have determined the crystal structure of BAFF-R extracellular domain bound to BAFF at a resolution of 3.3 Å. The cysteine-rich domain (CRD) of the BAFF-R extracellular domain adopts a b-hairpin structure and binds to the virus-like BAFF cage in a 1:1 molar ratio. The conserved DxL motif of BAFF-R is located on the tip of the b-turn, and plays an indispensable role in the binding of BAFF. The crystal structure shows that a unique dimeric contact occurs between the BAFF-R monomers in the virus-like cage complex. Both of the CRDs of TACI contain the DxL motifs and simultaneously interact with the BAFF dimer in the virus-like cage.




DNA and proteins are simultaneously stabilized in thermophilic species

Hiroshi Nakashima1*, Kaoru Kido1 and Ken Nishikawa2

1School of Health Sciences, Faculty of Medicine, Kanazawa University,

5-11-80 Kodatsuno, Kanazawa 920-0942, Japan and 

2Center for Information Biology, National Institute of Genetics,

Yata 1111, Mishima, Shizuoka 411-8540, Japan


We report here how DNA and proteins from thermophiles are simultaneously adapted at high temperature. We compared the homologous DNA and protein sequences between thermophilic and mesophilic species of Methanococcus bacteria. DNA dinucleotide compositions are significantly different between two species, although remaining the G+C content unchanged. DNA sequences from thermophiles prefer dinucleotides such as AG/CT, GA/TC and so on. It is assumed that these dinucleotides increase the thermal stability of DNA through the change of the double-stranded DNA curvature. Our analysis showed that globular proteins are favorably coded on the strand where AG and GA dinucleotides are located, resulting in the increase of charged residues such as Glu, Lys and Arg in a protein sequence. It is known that protein thermostability is gained by an increase of charged residues on the protein surface. Taking all together, thermophiles might adapt at high temperature by changing DNA dinucleotide compositions so as to increase the DNA stability as well as to adjust the accompanied change of amino acid compositions towards the thermal stability of proteins.




AcrB Multidrug Transport Mechanisms in a POPC Bilayer Explored by Molecular Dynamics Simulations.


Hui-Hsuan Tu*, Jung-Hsin Lin

School of Pharmacy , National Taiwan University , Taipei , Taiwan

No.1 Jen-Ai Road Sec. 1 Taipei , 100 , Taiwan


AcrB is a crucial exporter in Escherichia coli, which is involved in the multidrug resistance phenomena in cancer chemotherapy and in treating bacterial infections. Recent structural determination of the AcrB transporter by X-ray crystallography has provided invaluable information for elucidating the promiscuity and specificity of drug transportation. The entire functional unit of the drug exportation machinery consists of three large subunits: AcrB, which belongs to the superfamily of resistance-nodulation-cell division (RND) transporter; AcrA, known as a membrane fusion protein (MFP); a mulitifunctional outer membrane channel, TolC. It has been clearly illustrated from the X-ray structure that AcrB is the major active part of the whole molecular machine.

Each monomer of the AcrB transporter consists of twelve transmembrane α-helices, within which two drug delivery pathways have been hypothesized. It is conceivable that the time for pumping out the drugs is highly correlated with the mean passage time of drugs along these pathways. However, molecular mechanism of AcrB drug recognition is still a subject of debate. On the other hand, the conformation of pore formed by three α-helices may be altered by the presence of substrates, which is still awaited for further clarification.

We have conducted molecular dynamics of AcrB in a POPC bilayer at full hydration to investigate the dynamic interaction of known drugs and their transporter. The pathways of the drug transportation from membrane exterior are constructed with biasing forces toward the central cavity of the AcrB trimeric complex. The simulated pathways are compared with previous hypothesis for drug exportation. Finally, the

substrate-induced opening of the pore is also investigated.




A Hierarchical Virtual Screening of Farnesyl Transferase Inhibitors Based

on Structure Analysis


Fang-Yu Su*, Jung-Hsin Lin

School of Pharmacy, National Taiwan University, Taipei, Taiwan

No.1 Jen-Ai Road Sec. 1 Taipei, 100, Taiwan


Farnesyltransferase (FTase) is a critical enzyme that post-translationally modifies Ras. FTase operates as a heterodimer, which consists of a 379-residue α-subunit and a 437-residue zinc-containing β-subunit. The zinc ion is embedded at the junction between a hydrophilic surface groove near the subunit interface and a deep cleft in the b subunit lined up with aromatic residues. This hydrophobic cleft, together with the location of a bound noncognate peptide, has been identified as isoprenoid and peptide substrate binding sites. Peptide conformation depends on zinc ion, which correctly orients the cysteine thiol (ate) for catalysis and stabilizes the peptide-substrate conformation.

Inhibitors of FTase cause suppression of transforming activity of the Ras oncoproteins, and therefore farnesyltransferase inhibitors (FTIs) could be used as anticancer agents. Activated Ras is found in association with 30% of all human cancers. The enzyme protein farnesyltransferase (FTase) catalyzes the first step of the post-translational process by transferring the farnesyl group from farnesyl diphosphate (FPP) to the cysteine side chain in the CAAX sequence motif at the C-terminus of Ras. Once Ras proteins are activated after farnesylation, tumor cells will start proliferate.

In our study, a hierarchical virtual screening scheme is employed, which combines molecular weight filtering, pharmacophore filtering, and docking. ACD (Available Chemical Directory) is utilized as the basic chemical database and in the first filtering step only the chemical compounds with molecular weights less than 500 Da are chosen as candidates. A pharmacophore model constructed from known potent FTIs is then followed to perform the second stage filtering. Finally DOCK4.0 is used to simulate the molecular recognition process of FT and its inhibitors. The discovered novel lead can be further optimized and submitted to bioassays for investigating their ADME/T properties.




Computational study of amino acid-water interaction through Density Functional Theory method


Ajay Chaudhari, Prabhat K. Sahu, Shyi-Long Lee*

Dept. of Chemistry and Biochemistry, National Chung-Cheng University,

Ming-Hsiung, Chia-Yi- 621, Taiwan


Various configurations were investigated to find the most stable structures of glycine-(water)3 complex. Five different optimized conformers of glycine-(water)3 complex are obtained from density functional theory calculations using 6-311++G* basis set. Relaxation energy and many body interaction energies (two, three and four body) are also calculated for these conformers. Out of the five conformers, the most stable conformer has the total energy –513.8891537 Hartree and binding energy –27.28 Kcal/mol. It has been found that the relaxation energies, two body energies and three body energies have significant contribution to the total binding energy whereas four body energies are very small. The chemical hardness and chemical potential also confirmed the stability of the conformer having lowest total energy.




Applications of a 3-D Unpaired Electron Spin Density Equation for

Methyl group in Radiolysis and Biophysics


Shawn Shih

Department of Chemical Engineering

Yuan-Ze University, 135 Far East Road , Nei-Li ,Taiwan


A large class of stereochemical and related interactions in biochemistry are repulsive and others are attractive, but the relative orientation of two methyl groups and the amount of energy required to twist one relative to the other(the hindered rotation energy),or the alignment of such a group with respect to a conjugated ring to which it is attached(widely attributed to a mechanism called “hyperconjuation”. We used theories of both isotropic and anisotropic proton hyperfine interactions in the π-electron systems developed in the sixties. They are approximately by the magnetic dipole interactions between each proton and an electron spin magnetization that is didistributed in 2s and 2p Slater atomic orbitlas center on carbon atoms. We have extended these theories to the non-planar elefinic cation radicals, which are very important in both biocatalysis as well as in petroleum catalysis. A three dimensional unpaired electron spin density equation for the methyl group hinder rotation has been developed to handle some Jahn-Tellar vibronic molecular ions. It related the observed proton hyperfine splittings to the non-planar structures of the open chain alkene cation radicals generated by radiolysis and various oxidation methods. This valence bond theory and symmetry principles are in comparison with some MO calculations in the chemical literatures. Our localized V-B approaches is better in accordance with our experimental results. 




Using Discretization Techniques to Discover Gene Patterns for Diagnostics

Ru-Ting Yang 1*, Hsueh-Fen Juan2, and Ming-Syan Chen1

1Department of Electrical Engineering and Computer Science National Taiwan University

Department of Chemical Engineering and Graduate Institute of Biotechnology

National Taipei University of Technology,

1, Sec. 4, Roosevelt Road, Taipei, 106, Taiwan

1,Sec. 3, Chung-Hsiao E. Rd. Taipei, 106,Taiwan


After the draft of Human genome project was published and the high–throughput microarray technology is available, the discovery of discriminating gene patterns becomes important. In this work, we propose methods to analyze microarray data in both spatial and temporal directions. Entropy-based discretization is first performed with the purpose of not only getting rid of irrelevant genes, but also obtaining informative genes with specific cutting points that maximize the information gain and minimize the entropy. Candidate gene patterns, defined as the genes with the corresponding proper intervals, are generated. Gene patterns sharply distinguishing the different classes are discovered from these candidate ones. The discovery of discriminating gene patterns is based on the frequency change between different classes. Every class will have its own patterns which should not exist in the opposite class. These gene patterns are strong predictors for diseases and also indicators for the states of diseases helpful for clinical diagnostics and treatment suggestions. We implement these procedures spatially to the data sets of uterine leiomyoma. Meanwhile, discretization of the data sets of breast cancer cell lines is surveyed. From these discretized genes, we reveal trends under different treatments.