Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2014 Jul 11.
Published in final edited form as: J Med Chem. 2013 Jun 21;56(13):10.1021/jm400490e. doi: 10.1021/jm400490e

Development and Binding Mode Assessment of N-[4-[2-propyn-1-yl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-3H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-D-glutamic acid (BGC 945), a Novel Thymidylate Synthase Inhibitor that Targets Tumor Cells

Anna Tochowicz 1, Sean Dalziel 2, Oliv Eidam 3, Joseph D O’Connell 3rd 1, Sarah Griner 1, Janet S Finer-Moore 1, Robert M Stroud 1,*
PMCID: PMC3880649  NIHMSID: NIHMS497855  PMID: 23710599

Abstract

N-[4-[2-propyn-1-yl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-3H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-D-glutamic acid 1 (BGC 945, now known as ONX 0801), is a small molecule thymidylate synthase (TS) inhibitor discovered at the Institute of Cancer Research in London. It is licensed by Onyx Pharmaceuticals and is in Phase 1 clinical studies. It is a novel antifolate drug resembling TS inhibitors plevitrexed and raltitrexed that combines enzymatic inhibition of thymidylate synthase with α-folate receptor-mediated targeting of tumor cells. Thus, it has potential for efficacy with lower toxicity due to selective intracellular accumulation through α-folate receptor (α-FR) transport. The α-FR, a cell-surface receptor glycoprotein, which is over expressed mainly in ovarian and lung cancer tumors, has an affinity for 1 similar to that for its natural ligand, folic acid. This study describes a novel synthesis of 1, an X-ray crystal structure of its complex with Escherichia coli TS and 2’-deoxyuridine-5’-monophosphate, and a model for a similar complex with human TS.

Introduction

Thymidylate synthase (TS) (EC 2.1.1.45) has been recognized for decades as a key enzyme target for anti-cancer drugs because it plays a pivotal role in DNA replication1-3 TS directly methylates the 5C of uridine in 2’-deoxyuridine-5’monophosphate (dUMP), converting it to dTMP, in the sole de novo synthetic pathway to thymidine, which is required for DNA replication. Many inhibitors that compete with either the substrate (dUMP) or the cofactor 5,10-methylene-5,6,7,8,-tetrahydrofolate (mTHF) have been developed as drug leads. Also many X-ray structures of TS complexes with such inhibitors have elucidated their mechanisms of inhibition4-9 The structures encouraged rational design of analogs and different generations of structure and mechanism-based antifolate drugs, some of which are in clinical use.7, 10 However, treatment is often complicated by the problems of resistance and high toxicity.7, 11-13

CB371714, 15 (Fig. 1) is an early quinazoline-based folic acid analogue. Its clinical development was halted because of life-threatening renal and hepatic toxicity, and poor solubility.13 Raltitrexed (ZD1694) is a slightly modified analogue of CB3717 that is polyglutamatable by the folylpolyglutamate synthetase, which normally acts on the cofactor mTHF. Replacing the N-10 propargyl group and the benzene ring in CB3717 with a methyl group and thiophene ring, respectively, significantly improved solubility and potency and decreased the nephrotoxicity of the compound.13 Importantly, intracellular polyglutamylation of raltitrexed, which allows its cellular retention, does not decrease its activity. Raltitrexed is not approved by US Food and Drug Administration (FDA). Nevertheless, it became the first new drug for treatment of colorectal cancer since the mid 1990’s and it was licensed in Canada and many European countries for the treatment of metastatic colorectal cancer.16, 17 Subsequently, pemetrexed (LY231514), approved by the FDA in 2004, was licensed for the treatment of malignant pleural mesothelioma. Pemetrexed is another multi-targeted antifolate18, 19, which in 2008 was granted approval as a first-line treatment in combination with cisplatin for treatment of locally advanced and metastatic non-small cell lung cancer. Alone or in combination with other chemotherapeutics, pemetrexed also shows activity in a number of other tumors including head and neck, breast, bladder, cervical, gastric, pancreatic, ovarian and colorectal cancers.20-22 Pemetrexed is the first antifolate for which toxicity was reduced by a low-level folic acid and vitamin B-12 supplement.20, 23, 24. However, the optimization of folic acid supplementation to the level that decreases toxicity without compromising the antitumor effect of the drug still remains difficult and perhaps must be further explored.25 Plevitrexed (BGC 9331, ZD9331) is a non-polyglutamatable inhibitor that was developed as a result of the raltitrexed and pemetrexed adverse clinical effects. The effectiveness of Plevitrexed demonstrates that polyglutamylation is not required for potency of antifolates. It can be transported to tumor cells via both the α-folate receptor (α-FR) and the physiological reduced-folate carrier system (RFC).26 Clinical studies evaluating plevitrexed are still ongoing; the main interest in the drug is as an alternative treatment for gastric cancers for patients who are not able to tolerate platinum-based combination therapy.

Figure 1.

Figure 1

Chemical structures of the cofactor folic acid and thymidylate synthase inhibitors: CB3717, raltitrexed, pemetrexed, plevitrexed and 1.

1 (BGC 945, ONX 0801), Fig. 1, was designed to further reduce toxicity by more effectively targeting cancer cells that overexpress the α-FR.27 The α-FR is overexpressed in certain epithelial tumors, particularly ovarian cancer cells (more than 90% overexpress α-FR), and also lung, endometrial and mesothelioma tumors.28, 29 Importantly, the inhibitor is selectively transported via the α-FR and has reduced affinity for the widely expressed bidirectional RFC.27 The RFC is ubiquitously expressed and responsible for the uptake of conventional antifolates into normal tissues, and hence can cause TS-related toxicities in the bone marrow and gut. 1 emerged from a lead series of potent inhibitors that displayed low and high affinity for the RFC and the α-FR respectively (KD ~1mM and 1nM).30 The compound has an L-Glu-γ-D-Glu moiety that enhances binding to TS, and mimics the diglutamate metabolites of mTHF and certain conventional antifolates. The stereochemistry around the D-amino acid α-C prevents cleavage by peptide hydrolases. In the α-FR expressing human epidermoid (KB) or ovarian (IGROV-1) tumor cells its IC50 for inhibition of proliferation is ~1-10 nM and for cells that do not express the FR it is in low μM range.27 Thus some non-FR-mediated uptake into cells occurs at higher concentrations. However, short exposure to these higher concentrations, as observed in the plasma of KB or IGROV-1 tumor- bearing mice dosed with a single bolus injection of 1, leads to tumor selective TS inhibition because of rapid clearance from normal tissues. Compound 1 is licensed by Onyx Pharmaceuticals and is in Phase 1 clinical studies (ISRCTN 79302332).

We present a synthetic route for 1 and subsequent structural biology for the 1.7 Å co-crystal structure of the ternary complex of E. coli TS (ecTS) with dUMP and 1. Further a model of human TS (hTS) in complex with dUMP and 1 is reported, highlighting the impact of chemical modifications on protein binding.

Results

Development of compound 1

The inhibitor 1 (Fig. 1 and Scheme 1A) belongs to a series of cyclopenta(g)quinazoline-based antifolates for which synthesis and development have been described previously.31, 32 These compounds were synthesized by stepwise addition of a p-aminobenzoate moiety, an N10-substituent, and a dipeptide moiety to a cyclopenta[g]quinazoline ring. Importantly, an L-Glu-γ-D-Glu dipeptide moiety replaces the glutamate moiety present in other antifolates.33, 34 Further intracellular polyglutamylation of TS antifolate inhibitors by folate polyglutamate synthase increases their affinity for TS.33,34 The L-Glu-γ-D-Glu dipeptide moiety of 1 serves the same purpose as the polyglutamate tail, increasing affinity for TS through electrostatic interactions with the positively charged polyglutamate-binding groove on the enzyme. The D-Glu protects 1 from peptide hydrolases in the cell. As mentioned above, several routes of synthesis of 1 are reported.31, 32 For preparation of the sample for X-ray crystallography as well as the initial Phase I clinical study drug substance, an innovative four-stage synthetic route was evolved and conducted under Good Manufacturing Practices at Regis Technologies, Inc. (Morton Grove, IL) (Scheme 1).35, 36 Amongst other improvements, this route did not require the cobalt catalyst system used in earlier routes.31, 32 A molar yield of 13.3 % with respect to the tricyclic acid starting material was achieved of very good quality. The purity results were: 99.5 % of area percent basis (a/a) by achiral HPLC, 99.6 % enantiomeric excess (ee) 6-S chiral purity, and 99.8 % L, D diastereomeric purity (see Methods). This process has subsequently been further improved for larger scale pharmaceutical production with increased robustness.

Scheme 1.

Scheme 1

A) Molecular Structure of 1 Trisodium Salt Drug Substance.

B) L-γ-glutamyl-D-glutamic acid tris(1,1-dimethylethyl) ester. C) N-[4-[2-propynyl-[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-D-glutamic acid tris(1,1-dimethylethyl) ester. D) N-[4-[2-propynyl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-D-glutamic acid. E) N-[4-[2-propynyl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl- D-glutamic acid sodium salt (1:3), ONX 0801 Trisodium Salt Amorphous Drug Substance.

Overall structure of the TS-dUMP-inhibitor complex

The structure of the TS ternary complex with the substrate dUMP and inhibitor 1 elucidates the binding mode of 1 and may be useful in further evolution of this class of drugs. We attempted to co-crystallize 1 with hTS and substrate, dUMP, but after extensive screening were only able to obtain small needles that did not diffract well. We were unable to optimize these crystal hits, so instead focused on the ecTS complex. The justification for this approach is that the active sites of hTS and ecTS are highly conserved and high affinity hTS inhibitors have been developed based on the high-resolution crystal structure of an ecTS complex.37

We determined the crystal structure of Escherichia coli TS (ecTS) in a complex with dUMP and 1 to a resolution of 1.7 Å. The crystals contained eight independent molecules (four obligate homodimers) in the asymmetric unit (Table 1). The second Glu (the D-Glu) of the di-Glu moiety of each inhibitor protrudes from the protein surface and contributes to the packing interface with an adjacent pseudo-two-fold related dimer. A similar interface cannot be formed in hTS because hTS has inserts with respect to ecTS in the vicinity of the ecTS interface. In the case of hTS, the di-Glu moiety may actually interfere with formation of well-ordered crystals, especially if it is conformationally disordered.

Table 1.

Data collection and refinement statistics.

PDB code 4ISK

Data Collection

Space group P21

N molecules per AU 8

Cell dimensions (Å)
a,b,c (Å) 95.9, 85.5, 134.3
β (deg) 109.4

Resolution (Å) 30.0 - 1.75

Total reflections 602132

Unique reflections 200975

Completeness (%) 99.8 (97.8)a

Rmerge (%) 7.0 (65.3)a

Redundancy 2.98 (3.04)a

I/σ(I) 8.79 (1.59)a

Wilson B-factor 20.41

Refinement

Resolution (Å) 29.8 - 1.75

No. of reflections (test set) 200689 (4014)

Rwork /Rfree (%) 0.18 / 0.23

No. atoms 37536
protein 35916
ligand (inhibitor 1) 368
ligand (UMP/UMC) 245
ions (Mg2+) 7
water 1000

Average B-factor (Å) 34.40

rms deviations
bond length (Å) ±0.013
bond angels (deg) ±1.44
a

Statistics for the highest resolution shell are shown in parentheses.

Both protomers contribute to each active site of a dimer. A molecule of 1 and dUMP are found in each active site. As expected, the ecTS enzyme is in an active conformation, also referred as a closed conformation. This is typical of TS ternary complexes with substrate and cofactor (or cofactor analogue) (Fig. 2A).8, 9, 38

Figure 2.

Figure 2

A) Cartoon representation of crystal structure of the ecTS homodimer (blue and grey) with substrate dUMP (green carbons) and inhibitor 1 (pink carbons) bound at the active site shown in CPK representation. B) A hydrated magnesium-cluster is found between enzyme and the second glutamate side chain of inhibitor 1. Final 2Fo-Fc electron density for the inhibitor and Mg2+ -water cluster (Mg ion and hydrogen bonds depicted in green) is shown in blue.

Substrate dUMP binding mode

The substrate dUMP and the inhibitor tricyclic ring system are bound in the same orientation and binding site in each of the subunits in the asymmetric unit. In five of the subunits there is a covalent bond between the Sγ of the catalytic Cys146 and 6C of the dUMP pyrimidine ring. Despite this result, in three of the four dimers the Cys146 adopts a different rotamer in one of the two active sites and is not covalently linked to dUMP 6C. This asymmetry in dUMP binding has been seen in other TS ternary complexes such as the ternary complex of hTS with raltitrexed and dUMP6 and is consistent with the half-of-the-sites reactivity of ecTS.39 A covalent bond between the catalytic cysteine and the substrate dUMP is not required for potent TS inhibition by cofactor analogs.

The ecTS ternary complex with dUMP and 1 has a novel magnesium cluster

The binding mode of 1 is well defined by the electron density and is schematized in Fig. 3 and shown in three-dimensions in Figs. 4 and 5A. There are at least three direct polar contacts between the ligand and the protein observed: the nitrogen at quinazoline position 3 forms a hydrogen bond with Asp169 (2.6 Å) and the carbonyl oxygen at position 4 makes a hydrogen bond with Gly173 (3.0 Å). The 2-hydroxy-methyl group is within hydrogen bonding distance of the backbone nitrogen of Ala263 (2.8 Å) and the carboxyl oxygen of Asp169 (2.9 Å). The nitrogen at position 1 hydrogen bonds with a conserved structural water molecule (Wat477). The cyclopenta moiety displaces Trp80 and the propargyl group is oriented towards Asn177 (Fig. 5B). The benzoate ring of the inhibitor makes a T-shaped edge-to-face π-π stacking interaction with Phe176.

Figure 3.

Figure 3

Schematic representation of 1 binding interactions with ecTS generated by the program LigPlot+.41 The covalent bonds of BGC 945 are depicted in purple, and orange for the protein. Hydrogen bonds are illustrated in green and hydrophobic interactions are represented as red eyelashes around the protein residue names and ligand atoms. Only a few waters are shown for clarity (cyan spheres), for example the water molecules of the Magnesium-cluster (Mg2+ in green) near the diglutamate moiety.

Figure 4.

Figure 4

Stick representation of the inhibitor-binding site in the ecTS complex focusing on the di-Glu environment. Water-mediated hydrogen bonds connecting the inhibitor to the protein are shown as yellow dashes. Waters involved in the network are shown as red spheres. The inhibitor is shown with pink carbons. The magnesium cluster is shown with cyan (for water) and green (for Mg+2) spheres. For comparison, polyglutamylated raltitrexed from an ecTS-dUMP-raltitrexed complex (PDB 2BBQ) is shown as thin green lines overlaid on 1.

Figure 5. Views of ecTS and hTS active sites bound to folate mimetic inhibitors.

Figure 5

A) Crystal structure of ecTS bound to 1 and dUMP. B) Crystal structure of ecTS bound to 1 (green carbon atoms) compared to the crystal structure of ecTS bound to raltitrexed (grey carbon atoms, PDB ID 2KCE). C) Model of hTS bound to 1. A crystallographic water molecule (Wat980) included in the modeling is depicted with hydrogens. D) Model of hTS bound to 1 (yellow carbon atoms) compared to the crystal structure of hTS bound to raltitrexed (cyan carbon atoms, PDB ID 1HVY). E) Superposition of crystal structure of ecTS bound to raltitrexed (grey carbon atoms) on crystal structure of hTS bound to raltitrexed (cyan carbon atoms). F) Superposition of the crystal structure of ecTs bound to 1 (green carbon atoms) on the model of hTS bound to 1 (yellow carbon atoms).

The L-Glu-γ-D-Glu moiety is a novel feature of 1 whose binding mode to TS has not been previously characterized. It binds to ecTS analogously to the polyglutamyl moiety of polyglutamylated raltitrexed40 except that the D-Glu is rotated 180° about its N-α-C bond. This rotation switches the orientations of D-Glu’s carboxyl group and side chain compared to the second Glu moiety (Glu-2) of the raltitrexed polyglutamate (Fig. 4). Thus the carboxyl group lies in the polyglutamate-binding groove of the protein, whose floor is formed by His51, Arg53 and Ser54. The major protein contact with Glu-2 in the raltitrexed complex is a van der Waals interaction with His51 in which the plane of the His51 imidazole ring is parallel to the plane of the Glu-2–Glu-3 amide linkage. His51 is in van der Waals interaction with the D-Glu carboxyl of 1, with the plane of its imidazole ring parallel to the plane of the D-Glu carboxyl (Fig. 4).

There are no direct hydrogen bonds between the L-Glu-γ-D-Glu moiety and the protein, but there are several water-mediated hydrogen bonds to the protein and two intramolecular polar hydrogen bonds (N-H – O-C=O) between the amides and the carboxyl groups. 1 also makes four hydrogen bonds with water molecules from a newly observed hexahydrated magnesium ion cluster. The Mg2+ cluster is sandwiched between the second glutamate side chain carboxylate of 1 and the side chain of protein residue Glu82, thereby forming ionic interactions with both the ligand and the protein (Fig. 2B, 3).

Structural differences between the ecTS complex with 1 and dUMP and its complex with raltitrexed and dUMP

1 differs from raltitrexed in five regions of the molecule: 1). the glutamate moiety of raltitrexed is replaced by a L-Glu-γ-D-Glu moiety; 2). the methyl at 2C of the quinazoline ring in raltitrexed is replaced by a hydroxymethyl; 3). in 1, a cyclopentan group is fused to the quinazoline ring system; 4). the 10N methyl is replaced by a propargyl group in 1; and 5). the thiophene ring is replaced by a benzene ring (Fig. 1). These molecular differences make 1 (MW 646) a larger molecule than raltitrexed (MW 458) and the active site has to expand to accommodate the larger ligand. Thus, superposition of the structures of the two ecTS ternary complexes reveals minor changes in the active site residue side chains (Fig. 5B), such as movements of Trp80, Asn177, Leu172 and His51. However, the backbone atoms of the two complexes closely overlap: the rmsd for the α-Cs of the complexes after they are superposed is 0.4Å. These mean deviations are similar to the rmsds between any two complexes in the crystal structure of the ecTS ternary complex with 1, which has four dimers per asymmetric unit.

Implication for binding of 1 to hTS (model)

There is no X-ray structure of 1 bound to hTS available. To understand how 1 could bind to hTS, the hTS-dUMP-1 complex was modeled (Fig. 5C) based on the structure of ecTS complex reported here. The suitability of this model is supported by the similarity of the crystal structures of hTS and ecTS ternary complexes with dUMP and raltitrexed (rmsd = 0.53Å over 1500 atoms) (Fig. 5E), and the high conservation of the residues within the folate binding site (12 out of 14 residues in a 5 Å radius around the ligand are identical, which corresponds to a sequence identity of 86 %).

To accommodate 1 within the hTS active site, hTS residues within a radius of 5 Å of the ligand were minimized (see Methods). The binding mode of 1 in the model of the hTS complex is very similar to the binding mode of 1 in the crystal structure of the ecTS complex (Fig. 5F). The most profound changes during minimization were observed for residues Trp109 and Asn226, which rotated by 20 and 30 degrees respectively, to accommodate the cyclopentyl and propargyl groups (Fig. 5D). Also, the benzoate ring of 1 displaces L221 and F225. These rearrangements are similar to the structural differences observed between the ecTS ternary complex with 1 and ecTS-dUMP-raltitrexed (Fig. 5B). The hydrogen-bonding network of the quinazoline ring of 1 is highly conserved between human and E. coli TS. Subtle differences are observed in the interactions of the diglutamate moiety. One of two sequence differences in the binding site of 1 is that His51 in ecTS is Phe80 in hTS. In ecTS, the C-terminal glutamate carboxylate makes an anion π-π stacking interaction with His51. In theory, such an interaction is also possible with Phe80 in hTS, but in the hTS model the carboxylate shifts by 2 Å towards Arg49. Also, while the diglutamate moiety of 1 makes exclusively water-mediated hydrogen bonds with ecTS, Lys77 and Lys308 side chains interact directly with 1 in the model of hTS (Fig. 5C). These changes in protein-ligand interactions are likely observed because the energy minimization was carried out in the absence of water, but they could also reflect potential ionic interactions.

Discussion and Conclusions

Novel drug discovery strategies are pivotal for developing better cancer therapies. Effectiveness of antifolate drugs used in cancer treatment is limited by toxicity and mechanisms of drug resistance, including over-expression of hTS. The thymidylate synthase inhibitor 1 was developed to reduce toxicity toward non-tumor cells by specifically targeting the α-folate receptor, which is overexpressed in certain tumor cells. The IC50 for growth inhibition of α-FR-negative cells by 1 is ~7 μM compared to IC50s in the 1-300 nM range for cells over-expressing the α-FR receptors.27

1 was designed based on experience with a series of predecessors that bind to TS in a similar manner, primarily pemetrexed, a multitargeted drug licensed for lung cancer and mesothelioma treatment, and raltitrexed, which is specific for colorectal cancer. The novel features of 1 are the cyclopenta[g]quinazoline ring system and the side chain L-Glu-γ-D-Glu moiety. The purpose of this study was to elucidate how these novel structural features are accommodated in the active site of hTS to give a potent (Ki~1.2 nM) inhibitor. The crystal structure of 1 and dUMP bound to ecTS has been determined at a resolution of 1.7 Å and used to evaluate the structural determinants of binding affinity. The high resolution of the crystal structure allowed accurate description of the inhibitor binding mode and interactions with the protein.

Two crucial observations emerge from this study. Firstly, the novel diglutamate moiety binds in a solvent-mediated association with the surface of TS, analogously to the polyglutamate moieties of polyglutamylated antifolates. Crystal structures of ternary complexes of ecTS with polyglutamylated TS inhibitors have shown that the polyglutamylate tails bind to an electropositive binding groove on the enzyme surface, making mainly water-mediated hydrogen bonds with the protein.7, 40 Residues lining the groove are poorly conserved, which is consistent with the small number of direct hydrogen bonding interactions. Hence, there is considerable plasticity at the polyglutamate-protein interface; it is not surprising that the groove can bind a dipeptide with novel stereochemistry. The binding site for L-Glu-γ-D-Glu dipeptide, with the D-enantiomer stabilizing 1 against hydrolysis, overlaps the binding site for the first two glutamates in the polyglutamate tails, but with the carboxyl and side chain groups of the D-Glu in reversed positions. Thus, the carboxyl of the D-Glu is in the binding groove while the side chain is oriented away from the binding groove and towards the protein surface. The D-Glu makes water-mediated intermolecular hydrogen bonds with adjacent dimers in the crystal. hTS is not conserved in this region, and we speculate that if the D-Glu were oriented in a different conformation, or disordered, these stabilizing crystal packing interactions could not occur, which may explain our failure to co-crystallize hTS wth 1.

As co-crystallization of hTS with 1 was not successful, we used modeling to investigate how 1 may bind to this enzyme. Human and E.coli TS active sites are highly conserved (86 %), and comparison of crystal structures of ecTS- and hTS- with raltitrexed (Fig. 5E) showed that the folate inhibitor binding to both species was almost identical. Thus, similar changes to those observed in the crystal structure of ecTS-dUMP-1 were also predicted in a model of hTS-dUMP-1 (Fig. 5F). Analysis of this model revealed that hTS may accommodate the inhibitor by minor side chain movements of a few active site residues, mainly to harbor the cyclopentyl and propargyl groups. Furthermore, the model predicted two additional direct polar interactions between the C-terminal glutamate tail and the human protein (K77 and K308) (Fig 5C). The caveat with the modeling is that the tail could equally well make water mediated contacts with side chains in human TS, but water molecules were not included in the minimization.

The second key observation is a magnesium ion cluster linking a loop containing the invariant active site residue Trp80 with the end of the polyglutamate moiety. There is evidence in the literature that magnesium enhances catalysis by ecTS, but the basis for this effect was not known.42 Our structure is the first to show a hexa-hydrated Mg2+ cluster unambiguously bound to ecTS. It is consistent with the proposal that Mg2+ binding to a surface groove remote from the active site can enhance catalysis indirectly, by reducing the entropy of active site residues.43 The Mg2+ cluster also likely increases affinity of ecTS for the inhibitor by hydrogen bonding to the D-Glu. The loop residues that hydrogen bond to the cluster are not conserved in hTS, thus a Mg2+ cluster would not likely bind to the same site in hTS.

The potent and preferential transport of 1 by a receptor that is over-expressed in cancer cells strongly suggests that 1 might be the first highly-targeted antifolate for cancer treatment. Due to the well-characterized transport of 1 via α-FR, appropriate patient selection might benefit from use of diagnostic techniques to assess α-FR receptor level prior to initiation of treatment with this agent.30 This approach would facilitate a personalized identification of patients most suitable to the intrinsic cell targeting capability of 1. Our studies have developed a novel synthetic path to 1 and shown how this inhibitor binds to the validated cancer drug target TS. The novel di-Glu moiety on the inhibitor binds in the polyglutamate-binding groove and its mode of binding includes an ionic magnesium cluster coordinated by structural water, the inhibitor, and multiple residues on the TS enzyme.

Experimental Section

1. Cells and cell culture, protein expression and purification

ecTS was expressed in DH5α cells at 37 °C, induced at OD600 0.6-0.8 with 1 mM IPTG for 4 hours. Enzyme was purified by ion exchange chromatography (Q column) as described by Boles et al.44 and subsequently dialyzed into 20 mM KPO4 pH 7.4, 100 mM NaCl, and 2 mM DTT. The purification was then confirmed by SDS-PAGE gel.

2. Protein crystallization

An N-terminally His-tagged construct that was previously co-crystallized with inhibitors45 was used for screening crystallization conditions for the ternary complex of hTS with dUMP and 1. Several published protocols for co-crystallizing hTS with dUMP and antifolate inhibitors were tried, and small crystals were obtained using the protocol used by Almog et al.6 However, these diffracted to less than 5Å resolution or were twinned. Some crystal hits were also obtained from commercial crystallization screens, but these could not be optimized. Therefore, 1 was co-crystallized with ecTS and dUMP.

ecTS was concentrated to around 6 mg/mL, dUMP was added to a final concentration of 3.3 mM, and the inhibitor 1 was added to a final concentration of 3.3 mM. Additional inhibitor was added right before the crystal set-up. The optimized final crystallization condition was 20-21 % PEG 8000, 0.1 M TRIS pH 7.9, 0.2 M MgCl2, and 5 mM DTT at room temperature. To obtain better crystals conditions were optimized using additive screens. Finally, much better diffraction with MnCl2 and Ethanol (added separately) was observed and data were collected. Crystals were grown in hanging drops by the vapor diffusion method using a Mosquito nanoliter-scale robotic workstation (TTP Labtech). Crystals reached final size in ~15 days.

3. X-ray data collection and structure determination

X-ray diffraction data were collected from cryo-protected crystals (in 20 % ethylene glycol) at Beamline 8.3.1 of the Advanced Light Source (Lawrence Berkeley National Laboratories) on a 315r CCD detector. The monoclinic crystals diffracted to 1.75 Å resolution. Data were indexed, reduced, and scaled with XDS.46 The structure was determined by molecular replacement using Phaser47 and using liganded ecTS with the ligands removed (1AXW) as a search model. Restrained positional refinement and refinement of isotropic B-factors, and TLS refinement were carried out using Phenix.48 NCS restraints were not used because of small but significant conformational differences between the subunits in the asymmetric unit. Manual building was done using Coot.49 Data collection, processing statistics, R factors and statistics for the final structure are listed in Table 1.

4. Modeling of the hTS complex with dUMP and 1

Inhibitor 1 was modeled in the structure of hTS using the program PLOP (Protein Local Optimization Program).50 1 and dUMP were prepared using Schrodinger’s hetgrp_ffgen utility bundled with PRIME.51 The structure of hTS bound to raltitrexed (PDB ID: 1HVY)38 served as a template for the hTS model. One conserved water molecule (Wat980) from 1HVY was included in the minimization. 1 was placed in the hTS active site by superposing the ecTS-dUMP-1 complex (chain A) on hTS (1HVY, chain A). Protein residues within 5 Å radius of 1 were minimized to remove clashes of 1 with hTS. After the protein minimization, 1 was minimized in the enlarged binding site. dUMP was modeled without a covalent link to Cys195, as was observed in several subunits of the ecTS-dUMP-1 structure.

4. Synthetic chemistry and development of 1

Inhibitor 1 is D-Glutamic acid, N-[4-[2-propyn-1-yl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-3H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-, sodium salt (1:3); CAS registry number 1097638-00-0. The chemical structure of 1 is shown in Fig. 1 and Scheme 1A. The molecular formula of the trisodium salt is C32H30N5Na3O10 and has a molecular weight of 713.58 g/mol. The solid form is amorphous, hygroscopic, and with a glass transition temperature of approximately 120.6 °C.

Synthetic chemistry. General Information

A synthesis of 1 Drug Substance was established via a four-stage process referred to as the SR-1 process, and performed under contract by Regis Technologies (Morton Grove, IL). Chemicals and reagents were purchased from commercial suppliers and were used without further purification.

SR-1 Stage 1; L-γ-glutamyl-D-glutamic acid tris(1,1-dimethylethyl) ester (Scheme 1B)

N-[N-[(phenylmethoxy)carbonyl]-L-γ-glutamyl]-D-glutamic acid, tris(1,1-dimethylethyl) ester, 3.5 kg, was dissolved in ethanol and hydrogenated in the presence of 5% Palladium on Carbon under a 20 pound per square inch (gauge) hydrogen atmosphere. After the uptake of hydrogen ceased, the reaction was filtered through celite to remove the catalyst, and the catalyst cake was washed with methylene chloride. The resulting solution was stripped to dryness, then dissolved in ethylene dichloride, and stripped to dryness under reduced pressure to give 2.54 kg (approximately 94 % yield) Stage 1 product as a clear yellow oil. Ethylene dichloride was removed through subsequent processing in the remainder of the synthesis and verified by Gas Chromatography of the Drug Substance to be below < 5 ppm.

SR-1 Stage 2; N-[4-[2-propynyl-[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-D-glutamic acid tris(1,1-dimethylethyl) ester (Scheme 1C)

4-[2-propyn-1-yl-[6S-4,6,7,8,-tetrahydro-2-(hydroxylmethyl)-4-oxo-3H-cyclopenta[g]quinazolin-yl]amino]-benzoic acid, 1.52 kg, was charged to 32 kg dimethylformamide (DMF) solvent. To this mixture was added a solution of 2.53 kg L-γ-glutamyl]-D-glutamic acid tris(1,1-dimethylethyl) ester (Stage 1) dissolved in 10.1 kg of DMF. To this combined mixture was added 0.79 kg of 1-hydroxybenzotriazole hydrate, 0.87 kg of N-methylmopholine and 0.82 kg of dimethylaminopropyl ethylcarbodiimide hydrochloride. This mixture was stirred at room temperature for 12 hours, and then poured into water, and extracted into ethyl acetate. The product containing ethyl acetate portion was stripped to dryness to obtain 3.13 kg of oil. This material was purified using silica gel chromatography, eluting with methanol/methylene chloride. The eluent was stripped to dryness to give 1.22 kg (38 % yield) of the Stage 2 product.

SR-1 Stage 3; N-[4-[2-propynyl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl-D-glutamic acid (Scheme 1D)

N-[4-[2-propynyl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl- D-glutamic acid tris(1,1-dimethylethyl) ester (Stage 2), 407 g, was dissolved in 4.1 L of trifluoroacetic acid and stirred at 0 to 5 °C for about 2 hours. The triflouroacetic acid was removed by vacuum distillation, and replaced with 17 L of methylene chloride. The methylene chloride was removed by vacuum distillation, and a red oil resulted. Methyl tert-butylether (MTBE, 3.4 L) was charged to the oil and the mixture was stirred until a solid resulted. The solid was isolated by filtration, rinsed on the filter with MTBE, and slurried in acetonitrile, filtered and dried to give 292 g (77 % yield) of Stage 3 product.

SR-1 Stage 4; N-[4-[2-propynyl[(6S)-4,6,7,8-tetrahydro-2-(hydroxymethyl)-4-oxo-1H-cyclopenta[g]quinazolin-6-yl]amino]benzoyl]-L-γ-glutamyl- D-glutamic acid sodium salt (1:3), ONX 0801 Trisodium Salt Amorphous Drug Substance synthesis (Scheme 1E)

The Stage 3 product, 557 g, was slowly added to a solution of 278 g sodium bicarbonate and 5.5 L water, resulting in a clear solution. This solution was added to a column containing 32 kg of Sepabeads SP-207SS (Resindion, Mitsubishi Chemical Corporation) and eluted with water, followed by dilute acetonitrile in water. The pure product containing fractions were combined, filtered through a 2 micron cartridge, and lyophilized. The lyophilized solid was redissolved in 20 L water and ultra-filtered using a 5,000 MW cut-off cartridge. The filtrate was lyophilized to give 370 g (67 % yield) of a light-beige powder, ONX 0801 trisodium salt amorphous Drug Substance with a purity by HPLC of 99.5 % a/a, by achiral HPLC 99.6 % ee 6-S chiral purity, and 99.8 % L, D diastereomeric purity.

Acknowledgments

Research was supported by NIH grant GM 051232. We thank Dr. James Holton and Dr. George Meigs at the Advanced Light Source in Berkeley, at beamline 8.3.1. for their assistance during data collection.

Abbreviations

TS

thymidylate synthase

hTS

human thymidylate synthase

ecTS

Escherichia coli thymidylate synthase

dUMP

2’-deoxyuridine-5’-monophosphate

TMP

thymidine-5’-monophosphate

mTHF

N5,N10-methylene-5,6,7,8-tetrahydrofolate

α-FR

α-folate receptor

RFC

reduced-folate carrier system

DDT

dichlorodiphenyltrichloroethane

PMSF

phenylmethylsulphonyl fluoride

a/a

area percent basis

ee

enantiomeric excess

Footnotes

Associated Content

The atomic coordinates 4ISK have been deposited in the Protein Data Bank.

References

  • 1.Heidelberger C, Danenberg PV, Moran RG. Fluorinated pyrimidines and their nucleosides. Adv Enzymol Relat Areas Mol Biol. 1983;54:58–119. [PubMed] [Google Scholar]
  • 2.Santi DV, Danenberg PV. Folates in Pyrimidine Nucleotide Biosynthesis. In: Blakley RL, Benkovik SJ, editors. Folates and Pterins. Vol. 1. John Wiley and Sons; New York: 1984. pp. 345–398. [Google Scholar]
  • 3.Jackman AL, Newell DR, Gibson W, Jodrell DI, Taylor GA, Bishop JA, Hughes LR, Calvert AH. The biochemical pharmacology of the thymidylate synthase inhibitor, 2-desamino-2-methyl-N10-propargyl-5,8-dideazafolic acid (ICI 198583) Biochem Pharmacol. 1991;42:1885–1895. doi: 10.1016/0006-2952(91)90586-t. [DOI] [PubMed] [Google Scholar]
  • 4.Stout TJ, Stroud RM. The complex of the anti-cancer therapeutic, BW1843U89, with thymidylate synthase at 2.0 A resolution: implications for a new mode of inhibition. Structure. 1996;4:67–77. doi: 10.1016/s0969-2126(96)00010-x. [DOI] [PubMed] [Google Scholar]
  • 5.Stout TJ, Tondi D, Rinaldi M, Barlocco D, Pecorari P, Santi DV, Kuntz ID, Stroud RM, Shoichet BK, Costi MP. Structure-based design of inhibitors specific for bacterial thymidylate synthase. Biochemistry. 1999;38:1607–1617. doi: 10.1021/bi9815896. [DOI] [PubMed] [Google Scholar]
  • 6.Almog R, Waddling CA, Maley F, Maley GF, Van Roey P. Crystal structure of a deletion mutant of human thymidylate synthase Delta (7-29) and its ternary complex with Tomudex and dUMP. Protein Sci. 2001;10:988–996. doi: 10.1110/ps.47601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Sayre PH, Finer-Moore JS, Fritz TA, Biermann D, Gates SB, MacKellar WC, Patel VF, Stroud RM. Multi-targeted antifolates aimed at avoiding drug resistance form covalent closed inhibitory complexes with human and Escherichia coli thymidylate synthases. J Mol Biol. 2001;313:813–829. doi: 10.1006/jmbi.2001.5074. [DOI] [PubMed] [Google Scholar]
  • 8.Montfort WR, Perry KM, Fauman EB, Finer-Moore JS, Maley GF, Hardy L, Maley F, Stroud RM. Structure, multiple site binding, and segmental accommodation in thymidylate synthase on binding dUMP and an anti-folate. Biochemistry. 1990;29:6964–6977. doi: 10.1021/bi00482a004. [DOI] [PubMed] [Google Scholar]
  • 9.Matthews DA, Appelt K, Oatley SJ, Xuong NH. Crystal structure of Escherichia coli thymidylate synthase containing bound 5-fluoro-2’-deoxyuridylate and 10-propargyl-5,8-dideazafolate. J Mol Biol. 1990;214:923–936. doi: 10.1016/0022-2836(90)90346-N. [DOI] [PubMed] [Google Scholar]
  • 10.Jackman AL, Taylor GA, Gibson W, Kimbell R, Brown M, Calvert AH, Judson IR, Hughes LR. ICI D1694, a quinazoline antifolate thymidylate synthase inhibitor that is a potent inhibitor of L1210 tumor cell growth in vitro and in vivo: a new agent for clinical study. Cancer Res. 1991;51:5579–5586. [PubMed] [Google Scholar]
  • 11.Farrugia DC, Ford HE, Cunningham D, Danenberg KD, Danenberg PV, Brabender J, McVicar AD, Aherne GW, Hardcastle A, McCarthy K, Jackman AL. Thymidylate synthase expression in advanced colorectal cancer predicts for response to raltitrexed. Clin Cancer Res. 2003;9:792–801. [PubMed] [Google Scholar]
  • 12.Jarmula A. Antifolate inhibitors of thymidylate synthase as anticancer drugs. Mini Rev Med Chem. 2010;10:1211–1222. doi: 10.2174/13895575110091211. [DOI] [PubMed] [Google Scholar]
  • 13.Jackman AL, Forster M, Ng M. Targeting thymidylate synthase by antifolate drugs for the treatment of cancer. In: Neidle S, editor. Cancer Drug Design and Discovery. Elsevier; New York: 2007. pp. 198–226. [Google Scholar]
  • 14.Jones TR, Calvert AH, Jackman AL, Brown SJ, Jones M, Harrap KR. A potent antitumour quinazoline inhibitor of thymidylate synthetase: synthesis, biological properties and therapeutic results in mice. Eur J Cancer. 1981;17:11–19. doi: 10.1016/0014-2964(81)90206-1. [DOI] [PubMed] [Google Scholar]
  • 15.Clarke SJ, Jackman AL, Judson IR. The history of the development and clinical use of CB 3717 and ICI D1694. Adv Exp Med Biol. 1993;339:277–287. doi: 10.1007/978-1-4615-2488-5_27. discussion 289-290. [DOI] [PubMed] [Google Scholar]
  • 16.Wilson KS, Fitzgerald CA, Barnett JB, Gill S, Khoo KE. Adjuvant therapy with raltitrexed in patients with colorectal cancer intolerant of 5-fluorouracil: British Columbia Cancer Agency experience. Cancer Invest. 2007;25:711–714. doi: 10.1080/07357900701518388. [DOI] [PubMed] [Google Scholar]
  • 17.Wilson KS, Malfair Taylor SC. Raltitrexed: optimism and reality. Expert Opin Drug Metab Toxicol. 2009;5:1447–1454. doi: 10.1517/17425250903307455. [DOI] [PubMed] [Google Scholar]
  • 18.Walling J. From methotrexate to pemetrexed and beyond. A review of the pharmacodynamic and clinical properties of antifolates. Invest New Drugs. 2006;24:37–77. doi: 10.1007/s10637-005-4541-1. [DOI] [PubMed] [Google Scholar]
  • 19.Goldman ID, Zhao R. Molecular, biochemical, and cellular pharmacology of pemetrexed. Semin Oncol. 2002;29:3–17. doi: 10.1053/sonc.2002.37461. [DOI] [PubMed] [Google Scholar]
  • 20.Adjei AA. Pemetrexed (ALIMTA), a novel multitargeted antineoplastic agent. Clin Cancer Res. 2004;10:4276s–4280s. doi: 10.1158/1078-0432.CCR-040010. [DOI] [PubMed] [Google Scholar]
  • 21.Adjei AA. Clinical studies of pemetrexed and gemcitabine combinations. Ann Oncol. 2006;17(Suppl 5):v29–32. doi: 10.1093/annonc/mdj946. [DOI] [PubMed] [Google Scholar]
  • 22.Tomao F, Panici PB, Frati L, Tomao S. Emerging role of pemetrexed in ovarian cancer. Expert Rev Anticancer Ther. 2009;9:1727–1735. doi: 10.1586/era.09.141. [DOI] [PubMed] [Google Scholar]
  • 23.Scagliotti GV, Shin DM, Kindler HL, Vasconcelles MJ, Keppler U, Manegold C, Burris H, Gatzemeier U, Blatter J, Symanowski JT, Rusthoven JJ. Phase II study of pemetrexed with and without folic acid and vitamin B12 as front-line therapy in malignant pleural mesothelioma. J Clin Oncol. 2003;21:1556–1561. doi: 10.1200/JCO.2003.06.122. [DOI] [PubMed] [Google Scholar]
  • 24.Vogelzang NJ, Rusthoven JJ, Symanowski J, Denham C, Kaukel E, Ruffie P, Gatzemeier U, Boyer M, Emri S, Manegold C, Niyikiza C, Paoletti P. Phase III study of pemetrexed in combination with cisplatin versus cisplatin alone in patients with malignant pleural mesothelioma. J Clin Oncol. 2003;21:2636–2644. doi: 10.1200/JCO.2003.11.136. [DOI] [PubMed] [Google Scholar]
  • 25.Chattopadhyay S, Moran RG, Goldman ID. Pemetrexed: biochemical and cellular pharmacology, mechanisms, and clinical applications. Mol Cancer Ther. 2007;6:404–417. doi: 10.1158/1535-7163.MCT-06-0343. [DOI] [PubMed] [Google Scholar]
  • 26.Theti DS, Jackman AL. The role of alpha-folate receptor-mediated transport in the antitumor activity of antifolate drugs. Clin Cancer Res. 2004;10:1080–1089. doi: 10.1158/1078-0432.ccr-03-0157. [DOI] [PubMed] [Google Scholar]
  • 27.Gibbs DD, Theti DS, Wood N, Green M, Raynaud F, Valenti M, Forster MD, Mitchell F, Bavetsias V, Henderson E, Jackman AL. BGC 945, a novel tumor-selective thymidylate synthase inhibitor targeted to alpha-folate receptor-overexpressing tumors. Cancer Res. 2005;65:11721–11728. doi: 10.1158/0008-5472.CAN-05-2034. [DOI] [PubMed] [Google Scholar]
  • 28.Jackman AL, Theti DS, Gibbs DD. Antifolates targeted specifically to the folate receptor. Adv Drug Deliv Rev. 2004;56:1111–1125. doi: 10.1016/j.addr.2004.01.003. [DOI] [PubMed] [Google Scholar]
  • 29.Elnakat H, Ratnam M. Distribution, functionality and gene regulation of folate receptor isoforms: implications in targeted therapy. Adv Drug Deliv Rev. 2004;56:1067–1084. doi: 10.1016/j.addr.2004.01.001. [DOI] [PubMed] [Google Scholar]
  • 30.Pillai RG, Forster M, Perumal M, Mitchell F, Leyton J, Aibgirhio FI, Golovko O, Jackman AL, Aboagye EO. Imaging pharmacodynamics of the alpha-folate receptor-targeted thymidylate synthase inhibitor BGC 945. Cancer Res. 2008;68:3827–3834. doi: 10.1158/0008-5472.CAN-08-0135. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.Henderson EA, Bavetsias V, Theti DS, Wilson SC, Clauss R, Jackman AL. Targeting the alpha-folate receptor with cyclopenta[g]quinazoline-based inhibitors of thymidylate synthase. Bioorg Med Chem. 2006;14:5020–5042. doi: 10.1016/j.bmc.2006.03.001. [DOI] [PubMed] [Google Scholar]
  • 32.Bavetsias V, Marriott JH, Melin C, Kimbell R, Matusiak ZS, Boyle FT, Jackman AL. Design and synthesis of Cyclopenta[g]quinazoline-based antifolates as inhibitors of thymidylate synthase and potential antitumor agents. J Med Chem. 2000;43:1910–1926. doi: 10.1021/jm991119p. [DOI] [PubMed] [Google Scholar]
  • 33.Bavetsias V, Jackman AL, Kimbell R, Gibson W, Boyle FT, Bisset GM. Quinazoline antifolate thymidylate synthase inhibitors: gamma-linked L-D, D-D, and D-L dipeptide analogues of 2-desamino-2-methyl-N10-propargyl-5,8-dideazafolic acid (ICI 198583) J Med Chem. 1996;39:73–85. doi: 10.1021/jm950471+. [DOI] [PubMed] [Google Scholar]
  • 34.Jodrell DI, Gibson W, Bisset GM, Boyle FT, Judson IR, Jackman AL. The in vivo metabolic stability of dipeptide analogues of the quinazoline antifolate, ICI 198583, in mice. Biochem Pharmacol. 1993;46:2229–2234. doi: 10.1016/0006-2952(93)90613-2. [DOI] [PubMed] [Google Scholar]
  • 35.Onyx Pharmaceuticals, I. Efficient peptide couplings and their use in the synthesis and isolation of a cyclopenta[g]quinazoline trisodium salt. 2012
  • 36.Onyx Pharmaceuticals, I. Synthesis of cyclopentaquinazolines. 2012
  • 37.Reich SH, Fuhry MA, Nguyen D, Pino MJ, Welsh KM, Webber S, Janson CA, Jordan SR, Matthews DA, Smith WW, et al. Design and synthesis of novel 6,7-imidazotetrahydroquinoline inhibitors of thymidylate synthase using iterative protein crystal structure analysis. J Med Chem. 1992;35:847–858. doi: 10.1021/jm00083a007. [DOI] [PubMed] [Google Scholar]
  • 38.Phan J, Koli S, Minor W, Dunlap RB, Berger SH, Lebioda L. Human thymidylate synthase is in the closed conformation when complexed with dUMP and raltitrexed, an antifolate drug. Biochemistry. 2001;40:1897–1902. doi: 10.1021/bi002413i. [DOI] [PubMed] [Google Scholar]
  • 39.Spencer HT, Villafranca JE, Appleman JR. Kinetic scheme for thymidylate synthase from Escherichia coli: determination from measurements of ligand binding, primary and secondary isotope effects, and pre-steady-state catalysis. Biochemistry. 1997;36:4212–4222. doi: 10.1021/bi961794q. [DOI] [PubMed] [Google Scholar]
  • 40.Kamb A, Finer-Moore J, Calvert AH, Stroud RM. Structural basis for recognition of polyglutamyl folates by thymidylate synthase. Biochemistry. 1992;31:9883–9890. doi: 10.1021/bi00156a005. [DOI] [PubMed] [Google Scholar]
  • 41.Laskowski RA, Swindells MB. LigPlot+: multiple ligand-protein interaction diagrams for drug discovery. J Chem Inf Model. 2011;51:2778–2786. doi: 10.1021/ci200227u. [DOI] [PubMed] [Google Scholar]
  • 42.Maley GF, Maley F, Baugh CM. Differential inhibition of host and viral thymidylate synthetases by folylpolyglutamates. J Biol Chem. 1979;254:7485–7487. [PubMed] [Google Scholar]
  • 43.Wang Z, Sapienza PJ, Abeysinghe T, Luzum C, Lee AL, Finer-Moore JS, Stroud RM, Kohen A. Mg2+ binds to the surface of thymidylate synthase and affects hydride transfer at the interior active site. J Am Chem Soc. 2013 doi: 10.1021/ja400761x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Boles JO, Cisneros RJ, Weir MS, Odom JD, Villafranca JE, Dunlap RB. Purification and characterization of selenomethionyl thymidylate synthase from Escherichia coli: comparison with the wild-type enzyme. Biochemistry. 1991;30:11073–11080. doi: 10.1021/bi00110a009. [DOI] [PubMed] [Google Scholar]
  • 45.Carosati E, Tochowicz A, Marverti G, Guaitoli G, Benedetti P, Ferrari S, Stroud RM, Finer-Moore J, Luciani R, Farina D, Cruciani G, Costi MP. Inhibitor of ovarian cancer cells growth by virtual screening: a new thiazole derivative targeting human thymidylate synthase. J Med Chem. 2012;55:10272–10276. doi: 10.1021/jm300850v. [DOI] [PubMed] [Google Scholar]
  • 46.Kabsch W. Integration, scaling, space-group assignment and post-refinement. Acta Crystallogr D Biol Crystallogr. 2010;66:133–144. doi: 10.1107/S0907444909047374. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47.Storoni LC, McCoy AJ, Read RJ. Likelihood-enhanced fast rotation functions. Acta Crystallogr D Biol Crystallogr. 2004;60:432–438. doi: 10.1107/S0907444903028956. [DOI] [PubMed] [Google Scholar]
  • 48.Adams PD, Grosse-Kunstleve RW, Hung LW, Ioerger TR, McCoy AJ, Moriarty NW, Read RJ, Sacchettini JC, Sauter NK, Terwilliger TC. PHENIX: building new software for automated crystallographic structure determination. Acta Crystallogr D Biol Crystallogr. 2002;58:1948–1954. doi: 10.1107/s0907444902016657. [DOI] [PubMed] [Google Scholar]
  • 49.Emsley P, Cowtan K. Coot: model-building tools for molecular graphics. Acta Crystallogr D Biol Crystallogr. 2004;60:2126–2132. doi: 10.1107/S0907444904019158. [DOI] [PubMed] [Google Scholar]
  • 50.Kalyanaraman C, Bernacki K, Jacobson MP. Virtual screening against highly charged active sites: identifying substrates of alpha-beta barrel enzymes. Biochemistry. 2005;44:2059–2071. doi: 10.1021/bi0481186. [DOI] [PubMed] [Google Scholar]
  • 51.Sherman W, Day T, Jacobson MP, Friesner RA, Farid R. Novel procedure for modeling ligand/receptor induced fit effects. J Med Chem. 2006;49:534–553. doi: 10.1021/jm050540c. [DOI] [PubMed] [Google Scholar]

RESOURCES