Investigation of acyclic uridine amide and 5′-amido nucleoside analogues as potential inhibitors of the Plasmodium falciparum dUTPase

Graphical abstract


Introduction
Malaria is a major global problem, affecting millions of people each year. If not treated promptly; malaria can kill rapidly, especially children. In 2010, this accounted for an estimated 655,000 global malaria deaths (possibly ranging up to almost 1 million), 91% of which were in the African region. 1,2 Following the announcement of the goal of the global elimination of malaria; there is an urgent need for novel drug targets in order to overcome the current problems of resistance in the available antimalarial drugs 3,4 and to address new aspects of the disease.
Deoxyuridine 5 0 -triphosphate nucleotidohydrolase (dUTPase) is the enzyme that catalyses the hydrolysis of dUTP to dUMP. This provides dUMP, a precursor required for the biosynthesis of dTTP, whilst controlling the dUTP:dTTP concentration within the cell to levels that will prevent mis-incorporation of dUTP into DNA. 5 Due to the essentiality of dUTPases for cell viability in all organisms studied to date, including Escherichia coli and Saccharomyces cerevisiae, [6][7][8] it is therefore likely that dUTPases represent a novel target that yet remains to be explored in malaria.
We have previously reported the synthesis and biological evaluation of tritylated acyclic uracil analogues, [9][10][11] which were shown to be potent and selective inhibitors of the Plasmodium falciparum dUTPase (PfdUTPase). Interestingly these analogues, which were synthesised with varying chain lengths (Table 1), showed comparable activities to their preceding cyclic counterparts, 12 therefore are of equal interest in terms of inhibitory activities. More recently in an attempt to decrease lipophilicity and increase water solubility, we have shown when replacing the trityl group with a diphenyl moiety, in both acyclic and cyclic molecules, that it is possible under certain circumstances to retain PfdUTPase enzyme inhibition 13 (Table 1 and Table 2).
A clear advantage of the acyclic analogues is that they have reduced molecular weight in addition to decreased c log P values, therefore are possibly better candidates for the synthesis of an oral compound. 14 Additionally, these derivatives lack rigidity in their structure, thus could allow access to binding pockets not accessible with more rigid templates. The downsides of this strategy are entropic disadvantages and the possibility of multiple binding modes. One way in which to overcome this problem is to try and conformationally restrain the flexible chain by the insertion of one or more functional groups that are restricted in their rotation, thereby introducing a certain degree of rigidity. Appropriate choice of functional group may also give additional interactions with the active site, which may lead to an increase in potency and possibly also selectively. Additionally there is also potential for alteration and improvement of the pharmacokinetic properties of these compounds as anti-parasitic agents.
We have previously synthesised within our laboratories mono alkyl chain uracil acetamides with the amide bond insertion into the alkyl linker chain at the C-2,3 position (Fig. 1). 10 These were shown to exhibit weak inhibition of the PfdUTPase; therefore it did not seem that an amide linkage at this position was favourable.   Here we describe insertion of the amide bond at the C-4 position into the alkyl linker chain in order to probe the effects of a movement in the amide linkage in the tritylated derivatives. Diphenyl analogues were also included in this study, and only the 4C chain was synthesised as they were shown to have optimal activity and selectivity in the straight chain tritylated derivatives (Table 1). Additionally reversal of the amide linkage was also investigated (Fig. 2).
Finally insertion of the amide bond into the cyclic compounds gave an overall comparison of the effect of this increased rigidity in the restrained nucleoside upon enzyme inhibition (Fig. 3). The overall synthetic strategy (Scheme 1) in order to synthesise the first set of amides involved coupling of the relevant carboxylic acids to the amino alcohol linker chain, followed by attachment to the uracil base.
The initial amidation step was achieved starting from either diphenylacetic acid or triphenylacetic acid. HOBt and TBTU were used as coupling reagents with 3-aminopropanol to give intermediate compounds 3 and 4. Mitsunobu coupling using polymer   supported triphenylphosphine was carried out in order to attach the nucleobase, prior to deprotection of the benzoyl group yielding final compounds 7 and 8.
The synthesis was modified slightly in order to produce the corresponding compounds with the amide bond reversed (Scheme 2). Following ester hydrolysis of 9, 13 amidations of the resultant acid gave the desired compounds, 11 and 12.

Cyclic analogues
To prepare the cyclic amide analogues (trityl and diphenyl derivatives), 2 0 -deoxyuridine was protected at both the 5 0 and the 3 0 -positions as t-butyldimethylsilyl (TBDMS) ethers (Scheme 3). This was followed by selective monodeprotection of the 5 0 hydroxyl using pyridinium para-toluene sulphonate (PPTS). Oxidation of the alcohol was carried out in the presence of [bis(acetoxy)iodo]benzene (BAIB) and 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO) in water to give compound 15 in excellent yield. 15,16 The final amidations were once again carried out with tritylamine and diphenylmethylamine. The diphenyl product (16) was obtained in 69% yield; however no triphenyl amide was obtained. It is likely that the trityl group was too bulky to be able to react at the crowded centre of the activated ester intermediate. It is anticipated that had the trityl amide been formed, then it is likely that this would have been more potent than that of the diphenyl analogue as has been seen previously, although larger and more lipophilic. The diphenyl derivative on its own however is sufficient to serve as an indicator as to whether these compounds would show any affinity for the PfdUTPase. The final step was the removal of the protecting group from 16 in the presence of TBAF on silica.

Results and discussion
Biological activity was evaluated by testing all compounds against the recombinant PfdUTPase and human dUTPase (HsdUT-Pase) in order to determine inhibition constants and selectivity. Additionally compounds were screened in vitro against the chloroquine and pyrimethamine resistant, K1 strain of Plasmodium falciparum cultured in erythrocytes to evaluate antiplasmodial activity and the mammalian L6 cell line as a measure for cytotoxicity (Tables 3 and 4).

Acyclic analogues
The results for all the compounds tested are shown in Table 3 including the data for the straight chain acyclic analogues, 1f and 1k, for comparison. 9 A number of conclusions can be drawn from this data: Compounds 3 and 4 have been included and show the requirement of the uracil in inhibition of PfdUTPases (K i = >100 and 1000 lM). The N-benzoyl derivative 6 (K i >100 lM) is essentially inactive; removal of the benzoyl group from this trityl forward amide improves this value to 6.3 lM (compound 8).
The acid intermediate, 10, which lacks the trityl or diphenyl moiety clearly shows low affinity for PfdUTPase together with weak activity (25 lM) against P. falciparum.
Comparison of 'forward' amides 7 and 8 with 'reverse' amides 11 and 12 shows that the 'reverse' amides are more potent against the PfdUTPase (K i = 11 and 6.3 lM vs K i = 0.7 and 0.2 lM).
These results are noteworthy as a small change, such as reversing an amide bond has had a significant impact on the observed activities and illustrates the importance of correct orientation of any functionality. Additionally this is encouraging as it means that this extra rigidity in the chain is tolerated and the insertion of the amide group is not detrimental to activity if the correct orientation is attained.
The antiparasitic data remains constant for all the compounds ranging between 7 and 14 lM for 7, 8, 11 and 12, although the selectivity decreases compared to mammalian cells. There was no correlation between inhibition of the enzyme and inhibition of the parasite growth.

Cyclic analogues
The results of cyclic amide 17 alongside the original trityl (2b) and diphenyl (2c) derivatives are shown in Table 4. Compound 17 showed some inhibition of the PfdUTPase (K i = 8.3 lM), and selectivity compared to the human dUTPase was retained (K i = >100 lM).
This may suggest that the diphenyl group is constrained in a suboptimal conformation within the Plasmodium enzyme active site.
Activity against the parasite is also poor (EC 50 = >5 lM). In contrast to the acyclic series, this cyclic 'reverse amide' 17 was a weaker inhibitor of PfdUTPase than the corresponding amine 2c.

Conclusions
We have successfully prepared some conformationally restrained amide derivatives of the tritylated acyclic uridine derivatives.  potency of the compounds. Our 'reverse amides', 11 and 12, showed similar or greater potency to the alkyl analogues against PfdUTPase; whereas the 'normal amides' 7 and 8, showed reduced activity. This suggests that the normal amides constrain the trityl/ diphenyl group in a sub-optimal orientation and gives rise to unfavourable interactions in the active site. The 'reverse amides' retain activity against the enzyme, but show reduced selectivity in the cellular assay, indicating off-target effects of these compounds.

Enzyme purification and inhibition assays
Both recombinant P. falciparum and human dUTPases were expressed in E. coli BL21 (DE3) cells which had been transformed with the pET11Pfdut and pET3Hudut (kindly provided by P.O. Nyman, Lund University, Sweden) expression vectors, respectively. For dUTPase purification, the same procedure was used for both  the human and the Plasmodium enzymes. Cell pellets from a 2.8 L IPTG-induced culture were resuspended in 70 mL of buffer A (20 mM MES, 50 mM NaCl, 1 mM DTT, pH 5.5) containing a protease inhibitor cocktail. The cells were lysed by sonication, and the cell extract was cleared by centrifugation at 14,000 rpm for 30 min. The supernatant was loaded onto a 40 mL phosphocellulose (Whatman P-11) column at 4°C and eluted with a 50 mM to 1 M NaCl gradient in buffer A. Protein was further purified by gel filtration chromatography on a Superdex 200 HA 10/30 column at 4°C. Pooled fractions were concentrated by centrifugation at 4°C and desalted using a PD-10 column. The enzyme was stored in 10 mM bicine and 5 mM MgCl 2, pH 8 at À80°C. Purified fractions contained dUTPase of P96% purity.
Nucleotide hydrolysis was monitored by mixing enzyme and substrate with a rapid kinetic accessory (Hi-Tech Scientific) attached to a spectrophotometer (Cary 50) and connected to a computer for data acquisition and storage as described previously. 17 Protons, released through the hydrolysis of nucleotides, were neutralized by a pH indicator in weak buffered medium with similar pK a and monitored spectrophotometrically at the absorbance peak of the basic form of the indicator. The ratio between the indicator and the buffer concentration  (Table 1) were obtained according to Eq. 1.
Activity against P. falciparum K1 strain and cytotoxicity assessment against L6 cells (rat skeletal myoblast cells) was determined as previously reported. 9

Chemistry
Solvents and reagents were purchased from commercial suppliers and used without further purification. Dry solvents were purchased in sure sealed bottles stored over molecular sieves. Reactions were performed in pre-dried apparatus under an atmosphere of argon unless otherwise stated. Normal phase TLC was carried out on pre-coated silica plates (Kieselgel 60 F 254 , BDH) with visualisation via either ninhydrin, PMA, or 254 nm UV light. Flash chromatography was performed using Combiflash Companion or Combiflash Rf and prepacked columns (silica gel) purchased from Redisep (Presearch), Silicycle (Anachem) or Grace Resolve. Preparative HPLC was performed using a Gilson (321-Pump, 153-UV-vis Detector) equipped with a Gilson liquid handler for injection and fraction collection and XBridge Prep C18, 5 lm, ODB, 19 Â 100 mm column (Waters) with 0.1% ammonia in water (solvent A) and acetonitrile (solvent B) as mobile phase. Melting points (mp) were measured on a Gallenkamp melting point apparatus and are uncorrected. 1 H NMR and 13 C NMR spectra were recorded on a Bruker Avance DPX500 spectrometer or on a Bruker Avance DPX300 using the applied solvent simultaneously as internal standard. Deuterated solvents were purchased from Goss. Chemical shifts (d) are given in ppm together with the multiplicity, relative frequency, coupling constants (J, Hz) and assignment. High resolution mass spectra were performed on a Bruker MicroTof mass spectrometer at University of Dundee. LC-MS analysis and chromatographic separation were conducted with a Bruker MicroTof mass spectrometer using an Agilent HPLC 1100 with a diode array detector in series. The column used was a Waters Xbridge C18, 3.5 lm particle size, 2.1 Â 50 mm column and the compounds were eluted with a gradient of 5-95% Acetonitrile/H 2 O + 0.1% ammonia.

General procedure A for the synthesis of amides 3-4, 11-12
The relevant carboxylic acid (1 equiv), HOBt (1.4 equiv) and TBTU (1.4 equiv) were dissolved in DMF. To this was added DIPEA (3 equiv) and stirred at room temperature for 1 h. To the mixture, the amine (3.2 equiv) was added and the reaction left to stir at room temperature overnight under an atmosphere of Ar. The mixture was diluted with CHCl 3 (25 mL) and extracted with H 2 O (5 Â 30 mL). The organic layer was dried with MgSO 4 and the solvent concentrated under reduced pressure. The product was purified by flash chromatography. Polymer supported triphenylphosphine (2.5 equiv; 3 mmol/g) was swelled in THF for 15 min. To this was added the corresponding alcohol (1 equiv) and the N-3 benzoyluracil (2 equiv) which were shaken at room temperature for a further 15 min before DIAD (2 equiv) in THF was added to the mixture. The reaction was shaken until consumption of the alcohol was seen by TLC. The resin was then filtered off and washed with THF and the solvent removed under reduced pressure. The crude product was then purified by flash chromatography (Hexane/EtOAc 40:60 and/or CHCl 3 / CH 3 OH 95:5).

General procedure C: hydrolysis of benzoyl esters for the synthesis of compounds 7-8
The benzoyl ester intermediate was dissolved in a solution of 0.2 M sodium methoxide in CH 3 OH and the reaction stirred at room temperature overnight until the disappearance of the starting esters was observed (TLC). The solution was neutralised with Dowex H + ion exchange resin, filtered and washed with methanol. The solution was concentrated in vacuo and the crude residue was purified by chromatography.