Structure-Based Design and Synthesis of Antiparasitic Pyrrolopyrimidines Targeting Pteridine Reductase 1

The treatment of Human African trypanosomiasis remains a major unmet health need in sub-Saharan Africa. Approaches involving new molecular targets are important; pteridine reductase 1 (PTR1), an enzyme that reduces dihydrobiopterin in Trypanosoma spp., has been identified as a candidate target, and it has been shown previously that substituted pyrrolo[2,3-d]pyrimidines are inhibitors of PTR1 from Trypanosoma brucei (J. Med. Chem.2010, 53, 221–229). In this study, 61 new pyrrolo[2,3-d]pyrimidines have been prepared, designed with input from new crystal structures of 23 of these compounds complexed with PTR1, and evaluated in screens for enzyme inhibitory activity against PTR1 and in vitro antitrypanosomal activity. Eight compounds were sufficiently active in both screens to take forward to in vivo evaluation. Thus, although evidence for trypanocidal activity in a stage I disease model in mice was obtained, the compounds were too toxic to mice for further development.


■ INTRODUCTION
Parasitic disease remains a major global health problem, especially in tropical and subtropical countries. The availability of curative medicines is extremely limited and, such as are available, are of limited effect. Listed among the most neglected tropical diseases, Human African trypanosomiasis (HAT) continues to threaten tens of millions of people across rural Africa. 1 However, concerted efforts in surveillance and intervention have brought the reported incidence to fewer than 7000 cases in 2011, with estimates of the total number of infected people being around 20 000. 2 The reduction in cases from an estimated 300 000 at the turn of the century 3 has stimulated a campaign to seek elimination of Trypanosoma brucei gambiense-related disease by 2030. The zoonotic nature of Trypanosoma brucei rhodesiense disease makes it more difficult to plan an elimination strategy.
The limitations of drugs currently used for HAT are several: there is a requirement for parenteral administration, severe toxicity is found in some cases, and increasing resistance is emerging in the parasites themselves. 3 This means that new drugs remain necessary if elimination is to be achieved and the threat of resurgence is to be removed. It is important to identify and to exploit new molecular approaches for the treatment of HAT and other infectious diseases. Pteridine reductase (PTR1) has been proposed to be a good target in African trypanosomes. The enzyme has been shown to be essential using genetic methods, 4 and it has been targeted by inhibitors of several compound classes. 5−8 Among these, pyrrolopyrimidines are interesting from the perspective of already possessing biological activity and providing templates for drug discovery; the pyrimidine ring and its substituents readily key into nucleobase and cofactor base binding sites in enzymes, and C5, C6, and N7 are suitable for introducing substituents to control selectivity and physicochemical properties. In the past 2 years alone, papers have appeared where such a scaffold has been exploited for protein kinase inhibition, 9−13 topoisomerase inhibition and antibacterial activity, 14−16 anti-inflammatory compounds, 17 antiparasitic compounds, 18 and dipeptidyl peptidase IV inhibitors. 19 In addition, pyrrolopyrimidines bring with them the advantage of carrying a pharmacophore with structural similarity to the recognition motif of the parasite's P2 aminopurine transporter, 20 a membrane protein capable of accumulating its substrates to internal levels that exceed external concentrations up to a thousand-fold. 21 Previously, we reported that a number of heterocyclic compounds including substituted pyrrolopyrimidines and furopyrimidines are inhibitors of PTR1 from Trypanosoma brucei and Leishmania major. 7,22 Here, we use structure-based design on the pyrrolopyrimidine template, taking into consideration appropriate synthetic strategies, to obtain compounds with antitrypanosomal activity in vivo.

■ COMPOUND DESIGN
Crystallographic studies of fused pyrimidines in the active site of the enzyme from T. brucei, TbPTR1, showed binding at the folate site, with the pyrimidine ring sandwiched between the cofactor nicotinamide and a phenylalanine, forming an array of hydrogen bonds with catalytic residues and the cofactor phosphate and ribose. 7 Two possible binding poses are observed. One orientation is termed the substrate-like pose and corresponds to how substrates such as oxidized pterins and  (PDB 3bmc) showing key residues that create the active site pocket. Potential hydrogen bonds are depicted as magenta dotted lines, and all atoms are colored according to atom type: O, red; N, blue; S, gold; P, orange; C, yellow (NADP+), cyan (TbPTR1), or black (folate). Phe97 is not labeled but is shown as thin cyan lines for clarity. (B) The active site with folate removed and key hydrogen-donor or -acceptor groups circled blue or red, respectively. Phe97 (thin cyan lines) is directly above the nicotinamide in this view. Scaffolds I and II are shown opposite, and possible hydrogendonor or -acceptor groups are designated D or A, respectively. Arrows on the schematic also indicate the intended direction of R1, R2, and R3 substitutions into hydrophobic parts of the active site. pteridines bind. The second orientation is that displayed by the archetypal antifolate methotrexate (MTX), in which the pteridine is rotated 180°about the N2−N5 axis in order to maximize hydrogen-bonding capacity. 23 In each case, the possibility emerged of introducing substituents at positions 4−7 of the pyrrolopyridine template to direct substituents into hydrophobic pockets near Cys168, Leu209, Pro210, Met213, and Trp221. The requirement for transport into trypanosomes and the possibility that the specific transporters found in trypanosomes that prefer 4-aminopyrimidine substrates might concentrate the inhibitors advantageously were also considered; 20,21 2,4-diaminopyrimidines were therefore considered to be important substructures at the outset of this work. Recognizing that physicochemical properties also play an important role in the biological activity of substituted pyrimidines, 4-alkoxy and 4-alkylamino substituents were investigated. To engage the hydrophobic pockets of PTR1 evident from crystallographic studies, 5-alkyl, 5-aryl, 6-alkyl, and 6-aryl pyrrolopyrimidines together with 5,6-disubstituted compounds were all studied ( Figure 1); the position and shape of the hydrophobic pockets influenced the selection of substituents in compounds for synthesis. In principle, N7 could also support a substituent, but, as will be shown below, the NH forms an important hydrogen bond with PTR1; consequently, N7 substituents were not investigated.
Polysubstituted pyrrolopyrimidines, in general, were found to be most active as inhibitors of PTR1 and of T. brucei in culture. One such compound (20) required a targeted synthesis (Scheme 3). 4-Chloropyrrolopyrimidine 7, protected by trifluoracetylation at N2 (16), was iodinated with Niodosuccinimide to give 17, which was deprotected to give intermediate 18. Sonogashira coupling followed by nucleophilic aromatic substitution provided 20, which was substantially active both in enzyme assays and against T. brucei in culture. Improved yields in the Songashira coupling with phenylacetylene were obtained when the 7-N-tosyl derivative (22) was used.
Recognizing the need for a flexible synthesis leading to pyrrolopyrimidines with 5-hydrophobic substituents, the Michael addition-based synthesis used previously 22 was extended to include a wide range of aryl and some aralkyl substitutents in both 4-oxo-and 4-aminopyrimidine series (Scheme 4). Although in principle triaminopyrimidine 24 might be expected to be more reactive than its 4-oxo relative (23), it was found that the opposite was the case; reaction rates were slower and yields were poorer in the preparation of the 2,4-diamino compounds than in the preparation of the 2amino-4-oxo compounds. It is likely that the mildly basic conditions used for the Michael addition step increased the reactivity of the 2-amino-4-oxo pyrimidine through formation of the anion. In this way, two groups of 5-arylpyrrolopyrimidines, 4-oxo 27a−d and 4-amino 29a−d, were obtained. This reaction was also applicable to the synthesis of 4alkylamino 31b and 4-alkoxypyrrolopyrimidines 31a and 31c. It is notable in passing that the nitroalkyl intermediates 26, 28, and 30 were found to be trypanocidal, although they are not active as inhibitors of PTR1 (Tables 1 and 2).
The developing SAR and crystal structure information showed that it was necessary to occupy both hydrophobic pockets at the active site of PTR1 ( Figure 1). For this, a new synthesis was necessary. Three types of pyrimidine substrate, 2,6-diamino-4-oxo 23, 2,4,6-triamino 24, and 2,6-amino-4alkylamino 32, were reacted with diaryl bromoketones 33 prepared by bromination of the corresponding arylmethyl ketones, which were either commercially available or prepared by Friedel−Crafts acylation of a suitably substituted benzene derivative. Condensation of the diarylbromoketones afforded three series of 5,6-diarylpyrrolopyrimidines, 4-oxo 34a−l, 4amino 35a−j, and 4-dimethylamino 36a−e, compounds in low to moderate yields (Scheme 5.) A single example of a 4cyclohexylamino analogue (37) and a morpholinoalkyl analogue (38), designed to provide improved solubility, were also prepared using the same methods.

■ COMPOUND EVALUATION IN VITRO
Target compounds and intermediates were first evaluated by a spectrophotometric enzyme assay that monitored the decrease in absorbance at 340 nm as NADPH is oxidized (Supporting Information). Stock solutions at concentration 100 mM in 100% (v/v) DMSO were prepared and screened in duplicate at 10 and 50 μM against 30 μg mL −1 TbPTR1 (0.96 μM). Several compounds that failed to dissolve adequately at the desired concentration were rejected, and this left 102 compounds that were assayed. Inhibition was calculated as a percentage compared to assessment in the absence of inhibitor with background NADPH oxidation subtracted from all measurements. Fifty-four compounds displaying at least 60−70% inhibition of TbPTR1 at 50 μM were further analyzed at concentrations from 0.025 to 100 μM, and K i app values were determined, assuming reversible competitive inhibition and 1:1 stoichiometry ( Table 1). The K i of MTX was measured routinely as a standard. 26 For consideration for progression to in vivo models of sleeping sickness in mice, targets optimally of less than 100 nM K i against PTR1 (but acceptably less than 400 nM) and an IC 50 of less than 500 nM in an in vitro assay of trypanocidal activity were set. Activity data are reported in Table 1. As noted above, some nitroalkylpyrimidine intermediates (26,28,30) also showed activity in the in vitro assay, and the results are reported in Table 2. Selected significant compounds were also assayed for toxicity in human HEK cells.

■ STRUCTURE−ACTIVITY RELATIONSHIPS
The first variation explored was the substitution of the 4-amino or 4-oxo group by 4-alkylamino and 4-alkoxy. This was done using 5-cyanopyrrolopyrimidines (3, 12−15), a class of compounds that had been found active in previous work, and two 5-unsubstituted compounds (8a,b), but the general lack of significant activity together with the crystallographic information (see below) implied that a hydrophobic substituent on the pyrrole ring was necessary for activity. The cyano vector in the crystal structures suggested that an appropriate variation was to examine hydrophobically substituted alkynes with a terminal hydrophobic substituent. The multistep synthesis leading to 20 was not convenient but did lead to a compound with good activity at the PTR1 and acceptable activity against T. brucei in vitro; 20 was taken forward for further evaluation as described below. The importance of a significantly sized hydrophobic substituent was emphasized by the low activity of 5-methylpyrrolopyrimidine 27e in the 4-oxo series and 5-cyanopyrrolopyrimidine 4 in the 4-amino series. In the 4-oxo series, 5-aryl substituents (27) improved the activity in all assays but not sufficiently to give compounds potent enough for progression. There was also the suggestion that a more flexible hydrophobic 5-substituent would not give the required activity, as shown by arylaminomethyl compounds 27d and 29d. The 4-amino series (29a−c), however, had several compounds with significant activity in the PTR1 assay. However, the activity in the cellular assay was disappointing, suggesting that the anticipated enhanced uptake into trypanosomes was not occurring.
When 6-hydrophobic substituents were introduced, compounds with greatly improved PTR1 affinity were obtained. In 4-amino, 4-oxo, and 4-alkylamino series (6,15), compounds with good inhibitory activity were obtained; however, none of these compounds was sufficiently active in cellular assays to merit progression. Indeed, their activity was exceptionally low. It is possible that the 5-cyano group is sufficiently decreasing the basicity of these compounds so that they are not substrates for the transporters. Several compounds with more extended hydrophobic side chains, notably, phenylethyl, in both the 4amino and 4-oxo series also had good inhibitory activity against PTR1 (6c, 6d, 27d, 29d). One of these (6c) was active enough in the anti-trypanosomal assay in CMM medium to be considered for in vivo evaluation. More active compounds were found, however. Further investigations of substituent tolerance at C4 showed that alkoxy substitution afforded insoluble or weakly active compounds (31a, 31c) but that significant or good enzyme inhibitory activity was obtained with cyclohexylamino pyrrolopyrimidines (31b, 37). Once again, the cellular activity was lower than required for progression.   5-Methyl-6-phenyl pyrrolopyrimidine 34b was only modestly active, showing that more than a 6-aryl substituent was necessary for useful activity. A significant step forward came when two aryl substituents were introduced at both C5 and C6, as shown first by 5,6-diphenylpyrrolopyrimidines 34a and 35a This led to a clear increase in the anti-trypanosomal assay in CMM medium and, in the 2,4-diamino case (35a), to a compound that was at the margin as a further candidate for progression. A number of such compounds were made, and several of them displayed activity sufficient for further progression to in vivo evaluation. Notable, in terms of efficacy, are 5-phenyl-6-(4-fluorophenyl)pyrrolopyrimidine 35b and 5,6-  di(4-fluorophenyl)pyrrolopyrimidine 35c in the diamino series. The 4-dimethylamino compounds (36a−e), on the other hand, were insufficiently active. In the 4-oxo series especially, solubility prevented good enzyme assays from being obtained in some cases, but several compounds appropriate for progression were identified including 34g, 34l, 35f, 35g, and 35j. With the intention of filling the hydrophobic pockets as completely as possible, a branched alkyl substituent was introduced (34j, 35h), but this change did not improve activity. Similarly, the introduction of a strongly electronwithdrawing group (sulfone) gave only weakly active compounds with poor solubility (34k, 35i). An attempt to improve solubility with a flexible, polar ionic substituent (38) gave a compound with very low activity, and this strategy was not investigated further.

■ STRUCTURAL BASIS OF PTR1 INHIBITION
In previous work, we developed robust protocols to elucidate high-resolution structures of PTR1. 6−8 This allowed us to acquire new structural data that were combined with modeling approaches in an iterative process to guide compound design. Ultimately, this led us to determine isomorphous structures of 23 cocrystal complexes of the PTR1 tetramer (Figures 2 and 3 and Supporting Information). The high-resolution diffraction data and excellent quality of the electron density maps were sufficient to identify examples where multiple conformations and ligand orientations were present, for example, 8b, 29a, and 34i.
Superposition of the inhibitors as they are positioned in the TbPTR1 active site indicates that the core scaffold position and the hydrogen-bonding interactions of which it is capable are well-conserved. In addition, most of the inhibitors adopt the substrate-like pose described earlier. There are three exceptions: 2,4-diamino-5-phenylethylpyrrolopyrimidine 29d adopts the MTX-type orientation, 2,4-diamino-5-methyl analogue 29a displays both the substrate and MTX-type orientations, and 4-thiomorpholino-substituted compound 12d displays a distinct binding mode shifted away from the cofactor and catalytic residues. In all cases, we attribute the change in orientation to the presence of bulky hydrophobic substituents that would clash with polar components of the active site. 12d is unable to mimic interactions that stabilize substrate binding and therefore is only a modest inhibitor. The other two compounds (29a and 29d) participate in hydrogen bonding and van der Waals interactions similar to substrate and MTX binding poses and retain potency against PTR1.
The compound series that has been developed displayed good levels of PTR1 inhibition that are based on exploiting previously recognized molecular features, for example, by fulfilling the hydrogen-bonding capacity at the catalytic center. The widely dispersed appearance of the mainly hydrophobic substituents on the core scaffold ( Figure 2) demonstrates the conformational diversity available to interrogate the different regions of the active site, and this enhanced affinity for the target in some cases. In general, the key to improving potency, as exemplified by 3c, 20, and 35g, has been to incorporate a sufficiently bulky group capable of additional van der Waals interactions in the hydrophobic cavities adjacent to the substrate-binding site.
The 3-formylphenyl substituted compounds (6a and 15) are worth specific mention. They were designed to adopt the substrate-like pose but with the appropriate substituents that might form a covalent attachment to Cys168. This is indeed observed with stable thioester linkages being formed, presumably via oxidation of a first formed thioacetal adduct. The success here indicates that covalent modification might be exploited in future.

■ COMPOUND EVALUATION IN VITRO
Compounds were tested for activity, in vitro, against T. b. brucei using the Alamar blue assay, which demonstrates cell viability. Cells were tested in both rich HMI-9 medium and also a minimal medium, CMM, that does not contain added folate or biopterin. Results are shown in Tables 1 and 2. When inhibitors of dihydrofolate reductase (DHFR) are assessed in the latter medium, they are several hundred-fold more active than when tested in the rich medium, presumably due to the ability of the abundant folate to compete for binding sites on the target enzyme. The fact that for our compounds there was little difference in potency when using CMM or HMI-9 media would indicate that they are not exerting their activity through inhibition of DHFR, a fact consistent with structural predictions based on modeling within the DHFR-binding site that suggest regions of steric clash (data not shown).
Compounds were also tested against additional T. b. brucei strains that are defective in transporter proteins that might be expected to contribute to their uptake. 20,21 The TbAT1 knockout (KO) line has lost the P2 aminopurine transporter, which is also known to carry melaminophenylarsenicals and a Blank cells imply not tested in relevant assay. K i app is the apparent dissociation constant for the enzyme inhibitor complex, before correction for the inhibition modality-specific influence of substrate concentration relative to K m . As the inhibitors compete for binding with the pterin substrate, K i can be calculated according to the equation K i app = K i /(1 + S × K m −1 ), where S and K m refer to the pterin substrate. K i app and IC 50 values refer to the data presented in Tables 1 and 2. Cli is the clearance in vivo. fu is the fraction unbound by plasma. t 1/2 is the half-life of the compound. clogP and ligand efficiency (LE) were obtained using Pipeline Pilot (Accelrys Inc). diamidines, 27 whereas the B48 line has additionally lost another diamidine/melaminophenylarsenical transporter, the HAPT1 carrier (now known to be an aquaglyceroporin designated AQP2). 28,29 Trypanocidal activity on these two transporter mutants showed hardly any changes as compared with that of wild-type trypanosomes (data not shown), indicating that our compounds enter via routes other than, or in addition to, the P2 transporter and HAPT1. PTR1 inhibitors were also tested against related leishmania parasites (Leishmania mexicana M379, data not shown), where activity was moderate and weaker, indicating differences in its PTR1 from that of trypanosomes and in its biological context.

■ PHARMACOKINETIC AND IN VIVO EVALUATION
Progression of compounds into in vivo evaluation was based on a target property profile that included the following data: (1) affinity of compound for TbPTR1 preferably <100 nM but acceptably lower than 400 nM. Several compounds were found with K i in the range 120−140 nM, but selection of compounds for in vivo evaluation also took strongly into account the activity against T. brucei in culture. The use of a high substrate concentration to ensure a robust assay implies equally a high value for K i in the enzyme assays. (2) Selectivity of compound with respect to human dihydrofolate reductase, which was inferred from the relative activity in HDMI and CMM media.
(3) Time required to kill T. brucei in culture desirably within 6 h but acceptably within 48 h; many compounds were found that killed within 48 h, but none was as rapid as 6 h (kill curves and the methods used are described in the Supporting Information). (4) Physicochemical parameters (Table 3): logP range desirably of 2−3.5 and pK a between 4.5 and 9.2. For the most active compounds, clogP was in the range 3−3.6 and calculated pK a was typically 4−5 for 4-amino compounds and 9−9.6 for 4-oxo compounds (Pipeline Pilot, Accelrys Inc., was used for both logP and pK a estimation, see the Supporting Information). (5) Metabolic stability in microsomal assay: desirably t 1/2 > 6 h but acceptable >2 h; t 1/2 for the most active compound (35b) was >8 h. (6) In vitro trypanocidal activity desirably IC 50 < 100 nM but acceptably <800 nM; several compounds were obtained with IC 50 < 500 nM (Table 1).
Compounds for which PK data were obtained are shown in Table 3. These showed selectivity for killing parasites over a human cell line (HEK) in culture, again indicative of on-target activity and at an appropriate level of discrimination to justify further progression. Ligand efficiencies with respect to both PTR1 inhibition and in vitro anti-trypanosomal activity were all good, and the PK properties were acceptable for in vivo evaluation. LC−MS measurements showed significant accumulation of 34a and 35a into trypanosomes. The PK data showed good stability and relatively slow clearance, particularly for 35b, one of the most potent compounds. Four compounds in the 4amino series (35b, 35e, 35g, 35j) have been evaluated for efficacy in vivo. On observation of toxicity in the first round of experiments, MTD studies with 35b, 35e, and 35g showed severe toxicity and fatality at 100 mg/kg, some transient clinical signs at 30 mg/kg, but no clinical signs at 10 mg/kg; therefore, 30 mg/kg was selected as the dose for in vivo efficacy studies. Infected mice were found to be more sensitive to the toxic effects of the drugs, but once daily dosing only up to 4 days was found to be tolerated. In the experiments with 35b, 35e, and 35g, one mouse in each group survived long enough to demonstrate reduction of parasitaemia from ∼10 8 to below detectable limits. The surviving mice receiving 35e and 35g showed a relapse of parasitaemia on day 10. The 4-oxo series was also evaluated with the hope that toxicity would be reduced. 34h and 34i did not show acute toxicity when administered at a single dose of 30 mg/kg. However, 34i was not curative, and toxicity was still observed with 34h, although parasitaemia was greatly reduced with this compound.
It is evident that there is significant host toxicity in both the 4-oxo and 2,4-diamino series of 5,6-diarylpyrrolopyrimidines that could not be removed by aryl substituent modification. This series of compounds is nevertheless trypanotoxic and inhibits PTR1 in in vitro assays. The detailed information obtained from the crystallographic studies provides a strong basis for the design of further series of compounds based on interactions other than those in the hydrophobic pocket, for example, the covalent binding described above. Future work will address whether they actually act through inhibiting this target in vivo. The compounds demonstrate mammalian toxicity, possibly through interaction with a CNS receptor, that must also be addressed in future chemical modifications.

■ EXPERIMENTAL SECTION
Cell culture. Trypanosomes. Bloodstream form T. b. brucei (strain Lister 427) was cultured at 37°C in a humidified 5% CO 2 environment in either HMI-9 (Gibco) or Creek's minimal medium (CMM) 30 supplemented with 10% FBS. Cells were routinely diluted when reaching the mid-log-phase concentration of 2 × 10 6 cells/mL and maintained in culture for no more than 30 passages.
Alamar Blue Assay. Against Trypanosomes. Compounds' potency against trypanosomes was assessed by Alamar blue assay. 31 Cells were seeded at a final concentration of 2 × 10 4 cells/mL into serial dilutions of the test compound. After 48 h incubation at 37°C and 5% CO 2 , 20 μL of resazurin dye (Sigma-Aldrich) solution at 0.49 mM was added, and cells were incubated for a further 24 h. The reduction of resazurin was measured with a fluorimeter (FLUOstar Optima, BMG Labtech) using 544 nm excitation and 590 nm emission wavelengths. Output was plotted using the IC 50 determination algorithm of the GraphPad Prism 5.0 software. All experiments were carried out in duplicate and repeated on at least three independent occasions.
Against HEK Cells. To test compounds' activity against HEK cells, 3 × 10 5 cells/mL were seeded in a 96-well plate and allowed to adhere for 3 h at 37°C and 5% CO 2 . An equal volume of doubling serial dilutions of test compound was then added to the wells, and cells were incubated for a further 16 h before addition of resazurin (20 μL of a 0.49 mM solution). After 24 h, fluorescence was measured and data analyzed as described above.
In Vivo Experimental Protocol. Experimental groups of three ICR mice were infected with 1 × 10 5 bloodstream T. b. brucei strain 427 obtained from a donor mouse. Twenty-four hours postinfection mice were injected intraperitoneally with 30 mg/kg daily of test compound (repeated for four consecutive days), prepared in a 10% v/v DMSO, 40% v/v PEG, and 50% v/v water solution. Parasitaemia was monitored daily by tail venipuncture. Infected, untreated mice were used as control. All procedures were carried out by licensed animal workers and complied with the UK Animals (Scientific Procedures) Act, 1986, and with the national and University of Glasgow maintenance and care guidelines.
Compound Synthesis. General. 1 H and 13 C NMR spectra were measured on a Bruker DPX 400 MHz spectrometer with chemical shifts given in ppm (δ values) relative to proton and carbon traces in solvent. Coupling constants are reported in Hz. IR spectra were recorded on Shimadzu IRAffinity-1 spectrometer. Elemental analysis was carried out on a PerkinElmer 2400 analyzer series 2. Accurate mass was measured using Thermo Exactive MS and a Thermo U3000 HPLC system. Ionization was carried out by an ESI source (not heated). Anhydrous solvents were obtained from a Puresolv purification system, from Innovative Technologies, or purchased as such from Aldrich. Melting points were recorded on a Reichert hotstage microscope and are uncorrected. Chromatography was carried out using 200−400 mesh silica gels, or using reverse-phase HPLC on a Waters system using a C18 Luna column.
Declaration of Purity. All final compounds were equal to or more than 95% pure by HPLC and 1 H NMR (400 MHz).
Leading to 34c, 35b, and 36c. 1-(4-Fluorophenyl)-2-phenylethanone. 33 Aluminum chloride (16.0 g, 0.119 mmol, 1.2 mol equiv) was added to fluorobenzene (50 mL, 51.2 g, 0.533 mmol, 5.0 mol equiv) with stirring and cooling with ice water under nitrogen. Phenacyl chloride (13.8 mL, 16.13 g, 0.104 mmol, 1.05 mol equiv) was added dropwise while keeping the temperature below 20°C. The reaction mixture was stirred for a further 15 min, and then it was heated at 50°C for 5 h, after which time the reaction mixture was left stirring at room temperature for 9 h. Hydrolysis was carried out by diluting with dichloromethane, pouring the reaction mixture onto crushed ice (50 g), and extracting the resulting suspension with HCl (2M, 30 mL). The organic phase was then cautiously washed with a saturated aqueous solution of sodium hydrogen carbonate and brine. The organic layer was dried (Na 2 SO 4 ) and filtered, and the solvent was removed under reduced pressure to give solid material, which was washed with n-hexane. The desired material (21.00 g, 94%) was obtained as a pale yellow solid, mp 83−85°C (Lit. mp 82°C), 33 R f = mmol) were dissolved in DMF (4 mL, dry). The reaction mixture was heated at 60°C for 4 days with stirring under nitrogen. DMF was removed in vacuo, and the residue was applied to a silica gel column chromatography and eluted with methanol/ethyl acetate (1:9; R f = 0.4). The product was obtained as yellow solid ( 35c−j were similarly prepared. N 4 ,N 4 -Dimethyl-5,6-diphenyl-7H-pyrrolo[2,3-d]pyrimidine-2,4-diamine (36d). N 4 ,N 4 -Dimethyl-2,4,6-pyrimidinetriamine (0.500 g, 3.26 mmol) and 2-bromo-1,2-diphenylethanone [desyl bromide] (0.898 g, 3.26 mmol) were dissolved in DMF (4 mL, dry), to which were added potassium iodide (0.541 g, 3.26 mmol) and cesium carbonate (1.062 g, 3.26 mmol). The reaction mixture was heated at 60°C for 24 h. DMF was removed in vacuo, and the residue was dissolved in ethyl acetate and methanol, to which was added silica gel, and the solvents were removed under reduced pressure. The residue was applied to a silica gel column chromatography and eluted with ethyl acetate (R f = 0.3). The required product was obtained as a yellow solid (0.310 g, 29%), mp > 230°C. 1  was dissolved in toluene (15 mL, dry), to which was added morpholine (5.20 g, 0.06 mol) with stirring at room temperature. The reaction mixture was heated under reflux overnight, and then the solvent and excess morpholine were removed under reduced pressure. The crude material obtained was dissolved in ether and extracted with 30% NaOH (aqueous). The organic layer was extracted with water, dried (Na 2 SO 4 ), and filtered, and the solvent was removed under reduced pressure to give the required product as a thick brown oil. The crude product was applied to a silica gel column chromatography and eluted with ethyl acetate to give pale yellow thick oil (6.325 g, 98%). 1  Hydrobromic acid in acetic acid (33%, 1 mL) was added at room temperature with stirring. Bromine (0.5 mL) in chloroform (10 mL) was added to the reaction mixture at room temperature with stirring. The dropwise addition continued until slight bromine coloration remained. The stirring was continued for further 30 min. Aqueous sodium sulfite (10%) solution was added, and the reaction mixture was then extracted. The organic layer was collected and washed with a saturated solution of sodium hydrogen carbonate followed by brine, the organic layer was dried (Na 2 SO 4 ) and filtered, and the solvent was removed under reduced pressure to give the required product (1.020 g, 91%) as pale yellow oil. This material was used in the next experiment without further purification.

Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript.

■ NOTE ADDED AFTER ASAP PUBLICATION
The version of this paper that was published ASAP on July 29, 2014, contained incorrect versions of both Scheme 4 and Scheme 5. The corrected version was reposted July 30, 2014.