Experimental and Molecular DockingStudies of Inhibitor Mechanisms forRibonucleotide reductase

Jinghui Luo

Degree project in applied biotechnology, Master of Science (2 years), 2009Examensarbete i tillämpad bioteknik 30 hp till masterexamen, 2009Biology Education Centre, Uppsala University, and Department of Biochemistry & Biophysics,Stockholm UniversitySupervisors: Astrid Gräslund and Ana Popovic-Bijelic

1

Summary Iron is an important factor for cell growth, and cell proliferation can be inhibited due to lack of iron. Ribonucleotide reductase (RNR), a bottleneck enzyme for synthesis of the precursor deoxyribonucleotides for DNA synthesis, requires iron for activity. Certain studies have shown that iron chelators can be used as potential agents against tumor growth. In this work, I have studied the effect of potential inhibitors of ribonucleotide reductase by molecular docking and UV spectroscopy. The binding sites and effect of iron depletion on mouse ribonucleotide reductase protein R2 were studied by application of para-alkoxyphenols and the iron chelator, 1,10-phenanthroline. The results from my UV spectroscopy indicate a synergistic relation between DL-dithiothreitol and 1,10-phenanthroline during chelation, and show that 1,10-phenanthroline can be considered as an inhibitor to RNR under reducing condition. Also, earlier studies have shown that another group of organic molecules, namely para-alkoxyphenols was effective in inhibiting mouse and tumor RNR. These molecules function as antitumor drugs and may exhaust deoxyribonucleotide pools in HIV-infected cells. By molecular docking, I present a model of the binding pocket in the RNR R2 protein for redox active para-alkoxyphenols. The results presented here suggest an efficient method to discover potential new enzyme inhibitors that interact with RNR protein R2 and interfere with the redox reactions that are involved in the enzyme activation as well as the enzyme reaction.

2

Abbreviations: Cis-diamminedichloridoplatinum(II) Cisplatin Dithiothreitol DTT Electron spin resonance EPR Hydroxyanisoles 4-HA Inhibition constant Ki Isopropyl-1-thio-β-D-galactopyranoside IPTG Ribonucleotide reductase RNR Small subunit of RNR R2 Para P 4-mercaptophenol 4MP 1,10-phenanthroline 1,10-phen Sodium dodecyl sulphate polyacrylamide gel electrophoresis SDS-page

3

Introduction

Ribonucleotide reductase (RNR) is the enzyme catalyzing the reduction of the four standard ribonucleotides to their corresponding deoxyribonucleotides, which provide the building blocks for DNA replication and repair in all living cells (Gräslund & Sahlin, 1996; Reichard & Ehrenberg, 1983; Gräslund & Ehrenberg, 2007).

There are at least three different classes of RNRs (Nordlund & Reichard, 2006; Jordan et al., 1998). In class I RNR, functioning in eukaryotes and some bacteria and viruses, the enzyme is comprised of two non-identical subunits, proteins R1 and R2, each of them a homodimer. The larger subunit R1 includes the redox-active cysteines, the substrate binding sites and the sites for the allosteric regulators of enzyme activity. The small subunit R2 may be considered an activator protein that contains a stable free radical on a tyrosyl residue and diferric site formed by a reaction between protein bound ferrous iron and molecular oxygen. The tyrosyl radical is formed in this reaction and its presence is required for RNR function. The stability of the radical depends on the integrity of the metal site (Shao et al., 2006; Stubbe, 1989; Högbom et al., 2004). The interaction between the R1 and R2 proteins is crucial for RNR enzyme activity. The proposed (Nordlund & Reichard, 2006) electron and proton transfer pathway from protein R2 to protein R1 in E. coli RNR is shown in Fig.1.

Figure 1. The electron transfer pathway with dotted line from the tyrosyl radical (Tyr365) in R2 to Cys-439 in R1 follows conserved and hydrogen-bonded amino acid residues. When Cys-439 in RNR R1 receives this electron and then becomes active, it reduces ribonucleotides to deoxyribonucleotides. The C terminus of R2 is important for forming the RNR holoenzyme complex, however, the detailed electron transfer pathway on C terminus of R1 with cartoon was not solved. Here, I modeled this C terminus and made figure by Pymol (http://pymol.sourceforge.net/). The pathway probably also includes Tyr-356 in R2, which is located at the RNR holoenzyme complex interface (Sjöberg, 1994; Fontecave, 1998; Lendzian, 2005). Therefore, the C terminus of protein R2 is important for forming the RNR holoenzyme complex (Uhlin, 1994).

4

Ribonucleotide reductase is a prime target for cytostatic drugs that can be used against tumors and bacteria (Liermann et al.,1990; Chaper et al., 1986; Tsimberidou et al., 2002; Nurbo et al., 2007; Cerqueira et al., 2005; Wnuk & Robin, 2006). Three main kinds of inhibitors of RNR have been explored. These include peptidomimetic inhibitors, iron chelators and catalytic inhibitors of R1 and R2 (Cerqueira et al., 2005; Harpstrite et al., 2007). Some of the RNR inhibitors show potent chemotherapeutic efficacy against serveral cancers (Cerqueira et al., 2005). A few of them, such as hydroxyurea (HU), Triapine and GTI-240 have been used in treatment of the disease (Shao et al., 2006). 4-HA (4-Hydroxyanisole) and cisplatin have been used clinically to treat maglinant melanoma and RNR was recognized as an alternative target for 2-, 3-, and 4-HA (hydroxyanisole) (Elford, 1984). The Lassmann group showed that 4-HA (4-hydroxyanisole) can inhibit the growth of tumor-RNR from Ehrlich ascites tumor cells (Lassmanm et al., 1991). Unfortunately the doses of 4-HA and cisplatin were too high for treatment of clinical melanoma therapy and the drug resistance of cisplatin increase the potential risk in the further treatment. However, p-alkyoxyphenols and iron chelators will provide efficacy against mammalian RNR from Ehrlich ascites tumor cells (Cerqueira et al., 2005; Heffeter et al.,2006) 1,10-phenanthroline

Metal complexes have been reported as drugs against a variety of diseases (Thati & Noble, 2009; Harpstrite et al., 2007). Cisplatin (cis-diamminedichloridoplatinum(II)) is a well known metal-based anti-cancer drug and has been used in the treatment of a variety of human solid tumors, especially testicular cancer (Zhao & Lin 2005). However, drug resistance in tumor cells and cisplatin effects lead to limitations. Thus, efforts to develop new metal complexes would prove to be useful in overcoming the mentioned disadvantages. Phenanthroline is a well known heterocyclic organic compound (Figure 2a). As a bidentate ligand, 1,10-phenanthroline (1,10-phen) forms strong complexes with most metal ions such as Fe(II), Cu(II), Zn(II), Ni(II). Some studies show that chiral derivatives have important roles in enantioselective reactions, such as ketone reductions and oxidations (Plummer et al., 2008). KP722 is a new lanthanum compound containing three 1,10-phen molecules. Recent studies revealed that multidrug-resistant cancer cells were especially sensitive to KP722, which indicated that it was suitable for overcoming the disadvantages of cisplatin (Heffeter et al., 2006; Heffeter et al., 2007). Iron chelators such as triapine and hydroxyurea (HU) have a high potential for inhibiting ribonucleotide reductase (Shao et al., 2006). To my present knowledge, there are no reports on RNR inhibition by 1,10-phenanthroline. It has been shown that a similar compound bathophenanthroline (4,7-diphenyl-1,10-phenanthroline) can chelate iron from mouse R2 protein under anaerobic conditions and in the presence of hydroxyurea (Nyholm et al.,1993).

In the reaction of 1,10-phenanthroline with Fe(II), ferroin is formed (Fig. 2). The formation is a well known process and has been used for analytical determination of iron levels in blood plasma for several decades (Heffeter et al., 2006). The molar extinction coefficient of ferroin is 11000 M-1 cm-1 at 510 nm.

5

(a) (b)

Figure 2. The structure of (a) 1,10-phenanthroline and (b) ferroin,[Fe(o-phen)3]2+ which include three 1,10-phenanthroline and one Fe2+ with red color. The figure was made by Pymol (http://pymol.sourceforge.net/ ). Para-alkoxyphenols

Earlier studies have shown that p-allyloxyphenol, p-propoxyphenol and 4-hydroxyanisole are effective inhibitors of mouse RNR (Pötsch et al., 1994; Lassmann & Pötsch, 1995). They function as antitumor drugs and may exhaust deoxyribonucleotide pools in HIV-infected cells

(Pötsch et al., 1994; Lassmann & Pötsch, 1995; Pötsch et al., 1995; Wellnitz et al., 1997; Lassmann & Pötsch, 1995). These p-alkoxyphenol compounds interact with RNR protein R2 and are known to destroy the tyrosyl free radical (Högbom et al., 2004 & Pötsch et al., 1995). The rate of reduction of the tyrosyl radical increases in accordance with increasing length of the hydrophobic alkyl chain of the inhibitors. Docking

In the structure-based drug design, the interaction between drug and target protein can guide new drug discovery. Once binding amino acid groups around the inhibitor are identified, the pharmacophore of inhibitor could be built from their interaction (Graham, 2005). Experimentally, it takes much time to identify the bonding interactions that hold a ligand in the binding site, by e.g X ray or Nuclear Magnetic Resonance (NMR). Molecular docking can simulate the binding mode of different ligands in silico and can be finished in short time (Graham, 2005).

Docking simulation dock ligand conformers to targeted protein in the best ways. A conformer is a three dimensional conformation of a compound (Graham, 2005). Based on the certain rank solution and scoring fuction, the different conformers from simulation can be ranked (Graham, 2005). For docking assessment, the goal should be the reproduction of the bound conformation of a ligand into its target macromolecule. By using grid-based energy evaluation, such as combining with Monte Carlo optimization, searching conformational flexibility and heuristic free energy, Autodock program has become the most popular and accurate docking program in drug design (Morris, et al 1998). Aim

In this project I aimed to study the inhibition of RNR R2 protein by experiment and docking simulation. The project involved two kinds of inhibitors, 1,10-phenanthroline and p-alkoxyphenols. The detailed aims with respect to each inhibitor were to investigate the reaction between iron chelator and mouse RNR R2, and reveal the binding sites of p-alkoxyphenols on mouse RNR R2 and E. coli RNR R2, respectively.

6

Results Recombinant small subunit from mouse ribonucleotide reductase

To investigate the chelation of iron from mouse ribonucleotide reductase protein R2 by the inhibitor 1,10-phenanthroline, I first purified recombinant protein RNR R2 from E. coli. Fig. 3a shows that the purified R2 protein had a band at around 45 kDa. The protein was not very pure but it was enough for me to study the interaction between 1,10-phenanthroline and R2. The iron and tyrosyl radical in R2 were very important for this studies. Therefore, electron paramagnetic resonance (EPR) was implemented to detect the existence of radical. Electron paramagnetic resonance (EPR) is a technique which can detect unpaired electrons, such radical. The tyrosyl radical in mouse RNR R2 exhibited an electron spin resonance (EPR) doublet spectrum. The existence of the tyrosyl radical was also seen in the optical absorption spectrum and EPR spectrum of mouse R2 protein at 410 nm (Fig. 3b).

Figure 3. (a): Sodium dodecyl sulphate polyacrylamide gel electrophoresis page of the purified small subunit of mouse ribonucleotide reductase. Lane 1 and 2 are large subunit R1 proteins of ribonucleotide reductase. Lane 3 and lane 4 are small subunit R2 proteins of mouse ribonucleotide reductase and lane 5 is calibration kit standards. The lane 3 and 4 show my purified mouse RNR R2 around 45 kDa (b): The signal of the tyrosyl radical from EPR. Axes: horizontal, magnetic induction, Gauss; vertical, signal strength, arbitrary units. EPR (Electron Spin Resonance) spectrum at 40 K of the tyrosyl radical in 30 μM small subunit of mouse ribonucleotide reductase. EPR conditions were 9.5 GHz microwave frequency, 3.2 mW power and 5 G modulation amplitude. The effect of dithiothreitol concentration on the chelation of iron from mouse small subunit protein

In order to study the role of the concentration of the reductant in the reaction system, I used different concentrations of dithiothreitol (DTT) as reductant, with the same concentration of 1,10- phenanthroline. Since 1,10-phenanthroline only forms complexes with Fe2+, DTT was used as

1 2 3 4 5

7

reductant for iron from mouse R2. In the first step, the ferroin absorbance was measured by UV spectroscopy. Fig. 4 shows that the maximum absorption of ferroin was at 510 nm. The formation of ferroin in the reaction between mouse R2 and 1,10-phenanthroline under different reducing conditions was then followed kinetically at 510 nm for 1500 sec (Fig. 5). It was seen that with increasing concentration of DTT the reaction was faster. The curve obtained for 0.9 mM DTT shows a maximum around 500 sec and was different from the rest of the curves. I suggest that in this case, when the concentration of DTT was very low, the formed ferroin was slowly oxidized to ferriin (which has an absorption maximum at 600 nm). Obviously, there was a saturation effect around 10 mM DTT and above this concentration there was no increase in the reaction rate.

00.050.1

0.150.2

0.250.3

0.350.4

0.450.5

300 400 500 600

λ (nm)

Abs

orba

nce

(AU

)

Figure 4. Light absorption spectrum in the range from 300-600 nm of ferroin formed in the reaction of 10 μM small subunit protein of ribonucleotide reductase, 90 μM 1,10-phenanthroline and 15 mM DTT after 40 minutes of incubation at room temperature.

Figure 5. The effect of the DTT concentration on the formation of ferroin in the reaction of 10 μM small subunit protein of ribonucleotide reductase with 90 μM 1,10-phenanthroline. The Y axis shows the concentration of ferroin complex forming by Fe2+ and 1,10-phenanthroline. The concentration of mouse R2 protein was 10 μM. The concentration of 1,10-phenanthroline was 90 μM. The concentrations of DTT were 0.90 mM, 2.56 mM, 10 mM, 16.7 mM, and 25 mM. The effect of 1,10-phenanthroline concentrations on the chelation of iron by mouse small subunit of ribonucleotide reductase

In order to detect chelation with the different concentrations of 1,10- phenanthroline, I varied the concentrations of 1,10-phenanthroline while keeping the concentration of DTT, constant at 10 mM. The effect of the concentration of 1,10-phenanthroline on iron chelation is shown in Figure 6. It was seen that with increased concentration of phenanthroline the initial reaction was faster. The curve for 20 μM Phe was somewhat distorted at early times, due to

8

problems with mixing of the sample or bubble formation.

Figure 6. The effect of 1,10-phenanthroline concentration on the formation of ferroin. Y axis show the concentration of ferroin complex forming by Fe2+ and 1,10-phenanthroline. The concentration of mouse R2 protein was 10 μM. The concentration of DTT was 10 mM. The concentrations of 1,10-phenanthroline were 90 μM, 180 μM, 270 μM, and 360 μM.

Summarizing, my data demonstrates that 1,10-phenanthroline could chelate iron (Fe2+) from the R2 protein in the presence of reductant, DTT. The concentrations of 1,10-phenanthroline and DTT play an important role in the chelation reaction. Removal of iron from the R2 protein leads to reduced enzyme activity (Shao et al., 2006). Therefore, improving the efficiency of iron chelation would improve the efficiency of anti-RNR drugs against tumor or bacteria. Binding sites of para-alkoxyphenols on small subunit of ribonucleotide reductase

Theoretically, my efforts in this study, based on computer simulations, focus on the inhibition mechanisms of RNR inhibitors based on p-alkoxyphenol compounds. P-propoxyphenol derivatives and 4-hydroxyanisole, compounds selected from literature that have different inhibition constants (Ki) for RNR were selected for docking which yield certain thermodynamic properties involved in the interaction between protein R2 and the ligands, such as Ki values and binding energies (ΔG). Molecular modeling and docking of R2 mutants involving Trp48 was implemented to understand better how p-propoxyphenol derivatives and 4-hydroxyanisole bind to protein R2. Furthermore, I investigated the interaction of para-quinones with protein R2 for determining preferred binding sites.

The destruction of the tyrosyl radical in protein R2 was known to be more efficient with increasing size of the hydrophobic alkyl substituent of the alkoxyphenol inhibitors (Pötsch et al., 1994). I therefore aimed at increasing the hydrophobicity of the alkoxy chain of the phenol compounds for binding RNR R2. I also included 4-mercaptophenol (4MP) in this series. 4MP is a redox active phenol derivative that consists of a hydrophobic and an ionizable part (Fig. 7). I also included some alternative compounds in the investigation. These were selected from quinones, which have nine redox states and were involved in the photosynthetic reactions (Rich & Bendall, 1979; Zhu & Gunner, 2005; Ishikita et al., 2003). The selected compounds (Fig. 7) were docked as ligands to RNR protein R2. In addition, mouse RNR protein R2 was docked with p-alkoxyphenol, in order to compare the differences in the binding pockets of mouse and E. coli protein R2.

9

Figure 7. Structures of selected known or potential ribonucleotide reductase R2 inhibitors which were made from www.bindingdb.org. Structure 1-5 are para-alkoxyophenol derivatives; 1, para-methoxy phenol; 2, para-propoxyphenol; 3, para-ethoxyphenol; 4, para-allyloxyphenol; 5, 4-mercaptophenol. Structure 6 is para-quinone and structure 7-10 are para-quinone derivatives.

1 6

2 7

3 8

4 9

5 10

10

Molecular docking in silico was used to detect the binding sites of R2 protein. The

investigated molecules (Fig. 7) were known or potential inhibitors to RNR. Considering a possible binding pocket on RNR R2, I used different centers for searching the site with the lowest binding energies. The two binding pockets with the lowest binding energy for the p-alkoxyphenol derivatives were localized in the interface between chain A and chain B (Fig. 8a). The pocket was localized close to Trp48 and Gln43 and was large enough to bind the p-alkoxyphenol derivatives (Fig. 8c).

For p-allyloxyphenol (compound 4, Fig. 7). I produced 100 independent conformers by Autodock 4 and distributed around the protein. There were 26 and 27 almost identical conformers in the pockets between chain A and chain B, respectively (Fig. 8a). Other conformers were distributed in more disorder and with higher binding energies, compared with the above mentioned preferred conformers. The binding energies to chain B were somewhat lower than to chain A, because building the crystal structure lead to somewhat non-symmetric protein structures in the dimer. The two binding pockets in each chain lay in the interface of the two chains. The binding sites of the compound were localized 3 Å from the surface of the R2 protein. The conserved residues Trp48, Glu115, His118, Asp237 and Glu238 lie around the pocket. Based on potential hydrogen bonds between Trp48 and p-allyloxyphenol, Trp48 plays an important role in binding the inhibitor (Fig. 8). Phe113 and Phe47 may be important for hydrophobic interaction (Fig. 8).

11

Figure 8. Model of p-allyloxyphenol bound at the active site of RNR R2 from E. coli. A, blue sticks represent conserved residues and lines represent p-allyloxyphenol. Two active site pockets lie at the interface between the two polypeptide chains A and B. B, schematic diagram of one binding pocket in RNR R2. The green sphere lies at the interface of the arrow and RNR R2, and shows the binding pocket for p-alloxyphenol. The gray arrow represents the coming direction of inhibitors from outside. C, Interaction between p-allyloxyphenol and residues Trp48, Phe47, Phe113 from chain B and another two Phe from chain A. Cyan, green and mauve colors represent conserved residues, interacting residues and p-allyloxyphenol, respectively.

Next, the p-alkyoxyphenol derivatives, p-methoxyphenol, p-propoxyphenol, and p-ethoxyphenol (compounds 1, 2 and 3, Fig. 7) were docked using the same procedure as before. Approximately 80, 40 and 70 conformers were distributed in two binding pockets for p-methoxyphenol, p-propoxyphenol and p-ethoxyphenol, respectively. Their preferred orientations were similar to that of p-allyoxyphenol. However, their hydrogen bonds were different from those of of p-allyoxyphenol. The double bond in the allyl group may affect the binding position. More details about the interaction between RNR R2 and p-allyoxyphenol derivatives are illustrated in Fig. 9.

Chain B

A

B C

12

Figure 9. A, Superimposition of p-propoxyphenol (red) and p-ethoxyphenol (green) on the two conserved chains. B, clusters of p-allyoxyphenol in the two binding pockets.

EPR and UV spectroscopy indicate that the tyrosyl radical in E. coli, mouse, herpes

simplex virus and mammalian RNR R2 can be reduced with different kinetics by the p-alkoxphenols and 4-hydroxyanisole (Lassmann & Pötsch, 1995). RNR R2 from tumor cells has a high reduction rate (Lassmanm et al., 1991). The differences may result from structural differences between the R2 proteins. The same docking protocol as for the E. coli R2 protein was run for mouse RNR R2 and p-allyoxyphenol. The results (Fig. 10) showed that the p-allyloxyphenol binding pocket in mouse R2 was located deeper inside the protein than in E. coli R2, and its distance was about 8 Å from the surface of the protein. The binding pocket has more hydrophobic groups around the hydrophobic alkyl substituent of p-allyoxyphenol. The pocket in one chain close to the interface between chains A and B may also be influenced by interaction with hydrophobic residues in the other chain. Phe102, Phe165, Phe62, Tyr95 and Met169 around the binding pocket could play important roles for interaction between the inhibitor and mouse R2. These results were consistent with previous experimental data showing that a specific hydrophobic site in the RNR R2 can interact with a hydrophobic p-substituted phenyl group (Gräslund & Voevodskaya, 2008).

13

Figure 10. A, Comparison of one binding pocket for p-allyloxyphenol from RNR R2 of mouse and E. coli. Cyan helix with pink stick spheres and green helix with yellow stick-spheres represent the ligand with mouse and E. coli R2, respectively. B, ensemble of p-allyoxyphenol and p-quinones in chain A of E. coli R2, showing Trp48. C, corresponding ensemble of p-allyoxyphenol and p-quinones around Trp48 in chain B.

The p-alkoxyphenols and p-quinones (Fig. 7) were docked to the E. coli R2 protein. The order of the estimated binding energies (ΔGb) and Ki values were approximately consistent with previous experimental data for each inhibitor-enzyme complex (Table 1). Among phenyl derivatives, 4MP could be a potential new inhibitor due to its very low binding energy represnting stable complex between the protein and inhibitor. Table 1. Comparison of inhibitor parameters between docking calculation and experimental RNR inhibition data.

Compound (see Fig. 7)

ΔGb (kcal/mol) (calculated) 2

Ki (μM) (calculated) 3

IC50 (μM) (Pötsch et al. 1994) 4

1 -4.79 308.63 1100 2 -5.64 73.64 580 3 -5.31 128.78 650 4 -5.63 74.8 280 5 -5.59 79.33 —— 6 -5.10 183.95 —— 7 -5.68 68.11 —— 8 -5.66 70.87 —— 9 -4.31 698.18 —— 10 -4.56 452.61 ——

2ΔGb is change in binding free energy; 3Ki is the inhibition constant for a drug; 4IC50 is the concentration of a drug that results in 50% inhibition; ΔGb and Ki were calculated by Autodock 4.0.

Chain A

Chain B

A

C

B

14

In order to detect interaction between other redox-active compounds and RNR R2,

p-quinone derivatives were investigated and found to show low binding energies. For compound 7 (Table 1), most of the 100 conformers distributed in two binding pockets. 63 and 35 conformers were in the B and A chain, respectively. By superimposing conformers of p-allyoxyphenol and compound 7, I observed that they kept similar orientations and only a little rotation appeared due to the hydrophobic interaction between p-allyoxyphenol and the neigboring hydrophobic residues.

15

Discussion

There have been at least three proposed inhibition mechanisms that use the R2 protein as potential target: (1) access to and reduction of the radical or iron site directly, such as for the small molecule hydroxyurea (2) long range electron transfer from a binding site and (3) interference with another possibly more easily accessible free radical in RNR R2 (Gräslund & Voevodskaya, 2008; Cerez & Fontecave, 1992). The relatively large p-alkoxyphenol molecules cannot access the tyrosyl radical site directly for structural reasons (Nordlund & Eklund, 1993).

1,10-phenanthroline

My experimental studies showed that coincubation of mouse R2 protein with 1,10-phenanthroline in the presence of DTT led to generation of a ferroin spectrum. These data indicate that under reducing conditions, 1,10-phenanthroline can chelate the iron from mouse R2. The removal of iron from R2 results in lower tyrosyl radical content of the protein and therefore lowers enzyme activity. Thus, I demonstrated that 1,10-phenanthroline could be used as an RNR R2 inhibitor under reducing conditions. Recent studies have demonstrated that another kind of phenanthroline derivative, SmI2, a phenanthroline-epoxide compound, readily forms complexes with metal. Two SmI2 molecules were needed to form complexes with one Fe2+, which could improve chelation efficiency (Plummer et al., 2008). Para-alkoxyphenols

The studies concerning molecular docking with p-alkoxyphenol derivatives indicated that increasing size of the hydrophobic side chain in the substituted phenyl group improved the Ki value (Table 2), and therefore suggested that this type of RNR inhibitors act via a specific binding site on the protein. The suggested sites were shown in my docking experiments. Furthermore, preliminary attempts to dock 1,10-phenanthroline to the R2 protein showed a somewhat different binding site closed to C terminus surface of R2, while p-alkoxyphenols located inside the two chains.

My studies indicated that the binding sites of p-alkoxyphenols were localized 3 Å from the surface of E. coli ribonucleotide reductase R2 and that Trp48 had an important role in binding this group of inhibitors. For mouse R2, the binding pocket lay deeper than in E. coli R2 and more hydrophobic residues surrounded the binding pocket site. I hypothesize that p-alkoxyphenols and p-quinone also could play important roles in inhibiting electron transfer during catalysis. Trp48 was of fundamental importance in the pathway of electron transfer (Fontecave, 1998). These studies further suggest that p-quinone derivatives also could be efficiently bound to protein R2 and possibly interfere with the electron transfer during catalysis.

My results give examples of two types of mechanism that may be responsible for inhibition of RNR: iron chelation by 1,10-phenanthroline after reduction and interference with electron transfer during catalysis for the p-alkoxyphenols derivatives. The insights into binding sites, structure-activity relationships and iron chelation from my studies should be valuable for the future rational design of agents with potent chemotherapeutic efficacy against tumor cells.

16

Materials and Methods Expression of recombinant small subunit of ribonucleotide reductase

Native mouse RNR R2 protein was overexpressed from E. coli Rosetta 2 (DE3) p LysS pET-R2 plasmids (Lars Thelander, Umeå University). The bacteria were grown in 50 ml of LB medium (50.80 g Bacto Tryptone, 25.40 g Yeast extract, 50.80 g NaCl were added in a beaker and then added water up to 5 L with pH 7.5) containing carbenicillin (0.1 mg ml-1) and chloramphenicol (0.034 mg ml-1) overnight at 37 °C. Six flasks with 800 ml of LB medium +antibiotics were inoculated with 7 ml of the overnight culture and shaken vigorously at 37 °C until they reached an A595 of 0.8, when the temperature was decreased to 18 °C. The culture was induced with 500 µM isopropyl-1-thio-β-D-galactopyranoside (IPTG) and grown for an additional 20 h at 17 °C before harvesting. The flasks were taken from the incubator and kept in the cold room for at least 10 min. Bacteria were harvested by centrifugation for 20 minutes at 1300×g and 4 °C in a Beckman Coulter Avanti J-20 centrifuge. The cell pellets were resuspended in 100 ml of 50 mM Tris-HCl, pH 7.6 with 200 mM KCl and quickly frozen in liquid nitrogen. The cells were lysed by thawing and centrifuged in Beckman Ultracentrifuge LE 80 K for 30 minutes at 15000×g and 4 °C to pellet the cell debris. Nucleic acids were precipitated by mixing with 10% streptomycin sulfate and removed by centrifugation for 20 min at 4800×g, 4 °C in Beckman Avanti J-25 centrifuge. Mouse R2 protein was precipitated by ammonium sulfate (0.243 g/ml, yielding 40% saturation) (Mann et.al., 1991) and pellets were obtained by centrifugation for 20 minutes at 4800 g and 4 °C in Beckman Avanti J-25 centrifuge. The pellets were dissolved in 5 ml of 50 mM Tris-HCl, pH 7.6 and centrifuged for 10 minutes at 1600 g and 4 °C at Eppendorf desk-centrifuge to remove insoluble material. The supernatant was transferred to a new Sarstedt tube, quickly frozen in liquid nitrogen, and new Sarstedt tubes were stored at –20 °C. Purification of recombinant small subunit of ribonucleotide reductase

The ammonium-sulfate precipitated protein was desalted on a Sephadex G-25 column (40 ml) pre-equilibrated with 50 mM Tris-HCl, pH 7.6 and then purified on a DE52 column (5 ml) that was pre-equilibrated with 10 mM K-phosphate, pH 7.0, 30 mM KCl. The protein was loaded on the DE52 column and then the column was washed with 12 ml of 10 mM K-phosphate, pH 7.0, 30 mM KCl and eluted with 10 mM K-phosphate, pH 7.0, 70 mM KCl. The protein concentration was detemined by UV absorption at 280 nm. The concentration was calculated using the Lambert-Beer law. The molar extinction coefficient for monomeric mouse R2 protein was 62000 M-1 cm-1 (Mann et.al., 1991). The protein was concentrated by using six Millipore concentrators 50,000 MWCO (Eppendorf company) for 30 minutes at 1300×g and 4 °C in an Eppendorf table-top centrifuge. Sodium dodecyl sulphate polyacrylamide gel electrophoresis

The protein concentration was spectrophotometrically determined as above described. Samples were prepared by mixing 10 μl of protein with 10 μl of loading buffer (10 μl 1M Tris-HCl pH 8.0, 100 μl 0.1% BFB (bromphenol blue), 50 μl β-mercaptoethanol, 590 μl dH2O) in 1.5 ml eppendorf tubes. Phastgel was put on the running plane with a little water. Sodium dodecyl sulphate polyacrylamide gel electrophoresis was run on the PhastSystem (Amersham Pharmacia Biotech) using the protocols described in the PhastSystem manual. The proteins were identified according to their molecular weight with respect to Low molecular weight (LMW) calibration kit standards (Amersham Pharmacia Biotech).

17

Measurement of iron content

The iron content of all proteins was spectrophotometrically determined using Eagle Diagnostics Reagent Set (Eagle diagnostics, inc.). This method is based on the reaction of ferrous ions with ferrozine (3-2-pyridil-5,6-diphenyl-1,2,4-triazine), in which a stable magenta colored complex with absorption maximum at 560 nm is formed. The complex is very soluble in water and may be used for direct determination of iron in solution.

The standard was prepared from 1 ml of reagent A (hydroxylamine hydrochloride 220 mM in acetate buffer, pH 4.5 with surfactant), 200 μl of reagent B (ferrous chloride 500 g dl-1 in hydroxylamine hydrochloride) which is iron calibrator and 20 μl of reagent C (Ferrozine 16.7 mM in hydroxylamine hydrochloride).

For sample preparation, 50 μl of protein solution was first denatured by addition of 5 μl of perchloric acid (60 %). After addition of 145 μl of dH2O, all samples were centrifuged for 10 minutes at 4200×g and room temperature in a Biofuge desk-centrifuge. The supernatant was transferred to another eppendorf tube and 1 ml of reagent A and 20 μl of reagent C were added.

The standard and all samples were incubated at 50°C for 15 minutes and left at room temperature until the reaction was complete (i.e. there was no further increase in absorbance at 560 nm), which took approximately one hour. Absorbance of the standard and all samples was measured at 560 nm. The iron concentration was calculated using the following formula:

4179.0500 ×××=s

sampleFe A

AC

where CFe is the micromolar [μM] concentration of iron, Asample and As are absorptions at 560 nm of sample and standard, respectively; 500×0.179 is the concentration of the iron calibrator (0.179 μM); The factor 4 in the equation is the multiplication factor, which accounts for the difference in the volume of the protein and the iron calibrator (Narváez et al., 2006) Iron chelation by 1,10-phenanthroline

Optical absorption spectra were recorded at room temperature on a V-560 UV spectrophotometer. The iron chelation by 1,10-phenanthroline was followed kinetically as the formation of the product, ferroin, at 510 nm for 1500 seconds (Knízek & Musilová, 1968). Docking Procedure

All mol2 files of the ligands with hydrogens were constructed from http://rcmd-server.frm.uniroma1.it. The three-dimensional structure of E. coli RNR protein R2 (ID: 1av8) was downloaded from the Protein Data Bank (www.rcsb.org/pdb). Polar hydrogens were added on ribonucleotide reductase R2 by Accelrys Discovery Studio Visualizer (http://accelrys.com/products/discovery-studio). The two Fe metal ions have trivalent charges in the pdbqt file. Trp48 was altered to Phe48 in the R2 mutants by using Pymol. The orientation of Trp48 was in accordance with the crystal structure (ID: 1PJ0) (Voegtli et al., 2003). Autodock 4.1 was used to dock protein R2 and the alkoxyphenols and quinone derivatives. The Lamarckian Genetic Algorithm from MGLTools (http://mgltools.scripps.edu/) was used to search energetically supported binding modes (Morris, et al 1998). The running number was 100 and 250000 energy evaluations were applied for each run. AutoDockTools 1.4.6 were implemented to build the Autogrid box between the protein and ligands. The grid center was chosen on the residue Trp48 and the dimensions were 90×90×90 autogrid points (x,y,z directions) with 0.375 Å spacing. The atoms of the exocyclic group of the ligands were set as active torsions during docking in order that the confomation of inhibitor become flexible. The evaluation of energy functions and free energies was run with analysis in

18

tolerance of 2 Å rmsd. In order to get further refined information about the interaction between protein residues and ligands, energy minimization was performed by NAMD software for 500 steps using the Charmm force field. All calculations were run on a Linux x86 operating system.

19

Acknowledgements I’m indebted to many people from DBB for let me enjoy past six months in Stockholm. Here I thank Astrid Graslund for giving me chance to join in your group, for encouragement, for wise supervising and insightful discussions; Ana Popovic-Bijelic & Nina Voevodskaya for teaching me in the wet lab; Torbjörn Astlind & Haidi Astlind for ensuring my project run smoothly; Worth mentioning was that lots of my experiment can not be run without the support of Torbjörn; Sebastian Wärmländer & Alex Peralvarez for introducing wonderful project on amyloid beta peptide to me; Pär Bjelkmar & Björn Wallner for patiently answering my questions and guiding me in the modelling. Peter Damberg & Jyri Jarvet for unconditional discussions on amyloid beta peptide project. Thanks to my friends in Uppsala for encouragement when I stayed in Stocholm at the beginning. Finally, many warm thanks to my parents and girlfriend for love and supports.

20

References Cerqueira, N.M., Pereira, S., Fernandes, P.A. and Ramos, M.J. 2005. Overview of

ribonucleotide reductase inhibitors: an appealing target in anti-tumour therapy. Curr Med Chem 12: 1283-94.

Chaston, T.B., Lovejoy, D.B., Watts, R.N. and Richardson, D.R. 2003. Examination of the Antiproliferative Activity of Iron Chelators. Clin Cancer Res 9: 402-14.

Elford, H.L. 1984. The effect of hydroxyanisoles on mammalian ribonucleotide reductase, Riley PA. (eds), Hydroxyanisole: recent advances in anti-melanoma therapy. IRL Press, Oxford.

Fontecave, M. 1998. Ribonucleotide reductases and radical reactions. Cell Mol Life Sci 54:684-9.

Ghisla, S. and Thorpe, C. 2004. Acyl-CoA dehydrogenases. A mechanistic overview. Eur J Biochem 271: 494-508.

Gräslund, A. and Voevodskaya, N. 2008. Ribonucleotide reductase in Chlamydia trachomatis: A new subclass Ic functioning without tyrosyl radical. In: Kristoffer A. (eds), Ribonucleotide Reductase, pp.278-280. Nova Science Publisher. New York.

Graham. P. 2005. Computer in medicinal chemistry. An introduction to medicinal chemistry. pp.326-379. Oxford University Press. Oxford.

Gräslund, A and Ehrenberg, A. 2007. Tyrosyl radical, diiron center and enzyme mechanism in ribonucleotide reductase. Appl Magn Reson 31: 447-455.

Gräslund, A. and Sahlin, M. 1996. Electron paramagnetic resonance and nuclear magnetic resonance studies of Class I ribonucleotide reductase. Annu Rev Biophys Biomol Struct 25:259-86.

Gerez, C. and Fontecave, M. 1992. Reduction of the small subunit of Escherichia coli ribonucleotide reductase by hydrazines and hydroxlamines. Biochemistry 31: 780-786.

Harpstrite, E.S., Prior, L.J., Rath, P.N. and Sharma, V. 2007. Metallophrobes: synthesis,characterisation,and potency of a novel gallium(III) complex in human epidermal carcinoma cells. J Inorg Biochem 101:1347-53.

Heffeter, P., Jakupec, M.A. and Körner, W. 2006. Anticancer activity of the lanthanum compound [tris(1,10-phenanthroline)lanthanum(III)]trithiocyanate (KP772; FFC24). Biochem Pharmacol 71:426-40.

Heffeter, P., Jakupec, M.A. and Körner, W. 2007. Multidrug-resistant cancer cells are preferential targets of the new antineoplastic lanthanum compound KP772 (FFC24). Biochem Pharmacol 73:1873-86.

Högbom, M., Stenmark, P., Voevodskaya, N., McClarty, G., Gräslund, A. and Nordlund P. 2004. The radical site in chlamydial ribonucleotide reductase defines a new R2 subclass. Science 305: 245-8.

Ishikita, H., Morra, G. and Knapp, E.W. 2003. Redox potential of quinones in photosynthetic reaction centers from Rhodobacter sphaeroides: dependence on protonation of GLU-L212 and Asp-L213. Biochemistry 42: 3882-92.

Jordan, A. and Reichard, P. 1998. Ribonucleotide reductase. Annu Rev Biochem 67: 71-98 Knízek M, Musilová M. 1968. Improvement of iron determination by use of ferroin-organic

dye ion-association complexes. Talanta, 15:479-81 Lendzian, F. 2005. Structure and interactions of amino acid radicals in class I ribonucleotide

reductase studied by ENDOR and high-field EPR spectroscopy. Biochim Biophys Acta 1707: 67-90.

Lassmann, G. and Pötsch, S. 1995. Structure of transient radicals from cytostatic-active p-alkoxyphenols by continuous-flow EPR. Free Radic Biol Med 19: 533-539.

Lassmanm, G., Liermann, B., Arnold, W. and Schwabe, K. 1991. Ribonucleotide reductase in

21

melanoma tissue. EPR detection in human amelanotic melanoma and quenching of the tyrosine radical by 4-hydroxyanisole. J Cancer Res Clin Oncol 117:91-95.

Liermann, B., Lassmann, G. and Langen, P. 1990. Quenching of tyrosine radicals of M2 subunit from ribonucleotide reductase in tumor cells by different antitumor agents: an EPR study. Free Radic Biol Med 9:1-4.

Mann, G. Graeslund, A. Ochiai, E.I. Ingemarson, R. Thelander, L. 1991. Purification and characterization of recombinant mouse and herpes simplex virus ribonucleotide reductase R2 subunit. Biochemistry 30: 1939–1947.

Morris, G. M., Goodsell, D. S., Halliday, R.S., Huey, R., Hart, W. E., Belew, R. K. and Olson, A. J. 1998. Automated Docking Using a Lamarckian Genetic Algorithm and and Empirical Binding Free Energy Function. J. Comp Chem 19: 1639-1662.

Narváez, A.J., Voevodskaya, N., Thelander, L., and Gräslund, A. 2006. The Involvement of Arg265 of Mouse Ribonucleotide Reductase R2 Protein in Proton Transfer and Catalysis. J Biol Chem. 275:25365-71.

Nordlund, P. and Eklund, H. 1993. Structure and function of the E.coli ribonucleotide reductase protein R2. J Mol Biol 232: 123-164.

Nordlund, P. and Reichard, P. 2006. Ribonucleotide reduactase. Annu Rev Biochem 75: 681-706.

Nurbo, J., Roos, A.K., Muthas, D., Wahlström, E., Ericsson, D.J., Lundstedt, T., Unge, T. and Karlén A. 2007. Design, synthesis and evaluation of peptide inhibitors of mycobacterium tuberculosis ribonucleotide reductase. J Pept Sci 13: 822-32.

Nyholm, S., Thelander, L., Gräslund, A. 1993. Reduction and loss of the iron center in the reaction of the small subunit of mouse ribonucleotide reductase with hydroxyurea. Biochemistry 32:11569-74.

Plummer, J.M., Weitgenant, J.A., Noll, B.C., Lauher, J.W., Wiest, O. and Helquist, P. 2008. Synthesis, structure, and metal complexation behavior of a new type of functionalized chiral phenanthroline derivative. J Org Chem 73:3911-4.

Pötsch, S., Sahlin, M., Langelier, Y., Gräslund, A., Lassmann, G. 1995. Reduction of the tyrosyl radical and the iron center in protein R2 of ribonucleotide reductase from mouse, herpes simplex virus and E. coli by p-alkoxyphenols. FEBS Lett 374: 95-99.

Pötsch, S., Drechsler, H., Liermann, B., Graslund, A. and Lassmann, G. 1994. p-Alkoxyphenols, a new class of inhibtors of mammalian R2 ribonucleotide reductase: possible candidates for antimelanotic drugs. Mol Pharmacol 45: 792-6.

Rich, P.R. and Bendall, D.S. 1979. Mechanism for reduction of cytochromes by quinols in solution and its relevance to electron transfer reactions. FEBS Lett 105: 189-194.

Reichard, P. and Ehrenberg, A. 1983. Ribonucleotide reductase—a radical enzyme. Science 221: 514-519.

Schaper, K.J., Seydel, J.K., Rosenfeld, M. and Kazda, J. 1986. Development of inhibitors of mycobacterial ribonucleotide reductase. Lepr Rev 57: 254-64.

Shao, J., Zhou, B., Chu, B. and Yen, Y. 2006. Ribonucleotide reductase inhibitors and future drug design. Curr Cancer Drug Targets 6: 409-31.

Silva, K.E., Elgren, T.E., Que, L. and Stankovich, M.T. 1995. Electron transfer properties of the R2 protein of ribonucleotide reductase from E.coli. Biochemistry 34: 14093-14103.

Sjöberg, B.M. 1994. The ribonucleotide reductase jigsaw puzzle: a large piece falls into place. Structure 15: 793-6.

Stubbe, J.A. 1989. Protein radical involvement in biological catalysis? Annu Rev Biochem 58: 257-85.

Sturgeon, B.E., Burdi, D., Chen, S.X., Huynh, B.H., Edmondson, D.E. and Stubbe, J. 1996. Reconsideration of X, the diiron intermediate formed during cofactor assembly in E.coli ribonucleotide reductase. J Am Chem Soc 118: 7551–7557.

22

Thati, B., Noble, A. and Creaven, BS. 2009. Role of cell cycle events and apoptosis in mediating the anti-cancer activity of a silver(I) complex of 4-hydroxy-3-nitro-coumarin-bis(phenanthroline) in human malignant cancer cells. Eur J Pharmacol 602:203-14.

Tsimberidou, A.M., Alvarado, Y. and Giles, F.J. 2002. Evolving role of ribonucleoside reductase inhibitors in hematologic malignancies. Expert Rev Anticancer Ther 2:437-48

Uhlin, U. and Eklund, H. 1994. Structure of ribonucleotide reductase protein R1. Nature 370: 533-9.

Voegtli, W.C., Sommerhalter, M., Saleh, L., Baldwin, J., Bollinger, J.M. and Rosenzweig, A.C. 2002. Variable coordination geometries at the diiron(II) active site of ribonucleotide reductase R2. J Am Chem Soc 125: 15822-15830.

Wellnitz, U., Pötsch, S., Garbe, C. and Lassmann, G. 1997. Inactivation of ribonucleotide reductase in tumor cells and inhibition of tumor cell growth by p-alkoxyphenols. Melanoma Res 7: 288-298.

Willems, J.P., Lee, H.I., Burdi, D, Doan, P.E., Stubbe, J. and Hoffman, B.M. 1997. Identification of the protonated oxygenic ligands of ribonucleotide reductase intermediate X by Q-band H-1, H-2 CW and pulsed ENDOR. J Am Chem Soc 119: 9816-9824.

Wnuk, S.F. and Robins, M.J. 2006. Ribonucleotide reductase inhibitors as anti-herpes agents. Antiviral Res 71: 122-6.

Zhu, Z.Y. and Gunner, M.R. 2005. Energetics of quinone-dependent electron and proton transfer in Rhodobacter sphaeroides photosynthetic reaction. Biochemistry 44: 82-96.

Zhao, G. and Lin, H. 2005. Metal complexes with aromatic N-containing ligands as potential agents in cancer treatment. Curr Med Chem Anticancer Agents 5:137-47.