CHEMMEDCHEM FULL PAPERS DOI: 10.1002/cmdc.201300522 Fluorine-Containing 6,7-Dialkoxybiaryl-Based Inhibitors for Phosphodiesterase 10 A: Synthesis and in vitro Evaluation of Inhibitory Potency, Selectivity, and Metabolism Gregor Schwan,[a] Ghadir Barbar Asskar,[a] Norbert Hçfgen,[d] Lenka Kubicova,[a, e] Uta Funke,[b] Ute Egerland,[d] Michael Zahn,[c] Karen Nieber,[a] Matthias Scheunemann,[b] Norbert Strter,[c] Peter Brust,[b] and Detlef Briel*[a] Based on the potent phosphodiesterase 10 A (PDE10A) inhibitor PQ-10, we synthesized 32 derivatives to determine relationships between their molecular structure and binding properties. Their roles as potential positron emission tomography (PET) ligands were evaluated, as well as their inhibitory potency toward PDE10A and other PDEs, and their metabolic stability was determined in vitro. According to our findings, haloalkyl substituents at position 2 of the quinazoline moiety and/ or halo-alkyloxy substituents at positions 6 or 7 affect not only the compounds’ affinity, but also their selectivity toward PDE10A. As a result of substituting the methoxy group for a monofluoroethoxy or difluoroethoxy group at position 6 of the quinazoline ring, the selectivity for PDE10A over PDE3A increased. The same result was obtained by 6,7-difluoride substitution on the quinoxaline moiety. Finally, fluorinated compounds (R)-7-(fluoromethoxy)-6-methoxy-4-(3-(quinoxaline-2yloxy)pyrrolidine-1-yl)quinazoline (16 a), 19 a–d, (R)-tert-butyl-3(6-fluoroquinoxalin-2-yloxy)pyrrolidine-1-carboxylate (29), and 35 (IC50 PDE10A 11–65 nm) showed the highest inhibitory potential. Further, fluoroethoxy substitution at position 7 of the quinazoline ring improved metabolic stability over that of the lead structure PQ-10. Introduction The multigenic family of phosphodiesterases (PDEs) (EC 3.1.4.17) plays a critical role in control of the level, localization, and duration of intracellular 3’-5’-cyclic adenosine monophosphate (cAMP) and 3’-5’-cyclic guanosine monophosphate (cGMP) signals by specifically hydrolyzing these cyclic nucleotides. As the involvement of cyclic nucleotide second messengers in cell signaling and homeostasis is established, the regulation of these pathways in the brain by various PDE isoforms is an area of considerable interest, as they are involved in [a] Dr. G. Schwan, Dr. G. Barbar Asskar, Dr. L. Kubicova, Prof. Dr. K. Nieber, Prof. Dr. D. Briel Institut für Pharmazie, Universität Leipzig Brüderstr. 34, 04103 Leipzig (Germany) Fax: (+ 49) 341 9736889 E-mail: briel@uni-leipzig.de [b] Dr. U. Funke, Dr. M. Scheunemann, Prof. Dr. P. Brust Institut für Radiopharmazie, Forschungsstelle Leipzig Helmholtz Zentrum Dresden-Rossendorf Permoserstr. 15, 04318 Leipzig (Germany) [c] Dr. M. Zahn, Prof. Dr. N. Sträter Institut für Bioanalytische Chemie, Universität Leipzig Deutscher Platz 5, 04103 Leipzig (Germany) [d] Dr. N. Hçfgen, U. Egerland BioCrea GmbH, Meissner Str. 191, 01445 Radebeul (Germany) [e] Dr. L. Kubicova Department of Molecular Systems Biology, Faculty of Life Sciences University of Vienna, Althanstr. 14, 1090 Vienna (Austria) Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/cmdc.201300522.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim nearly all brain functions and in the etiology of neuropsychiatric diseases.[1, 2] The PDE10A isoform, isolated from different species and characterized regarding structure and function,[3–5] has received much attention in recent years, particularly in the context of schizophrenia[6–8] and Huntington’s disease,[9] which are both related to a role of PDE10A in the regulation of striatal dopaminergic neurotransmission.[10, 11] Outside of the CNS, the PDE10A expression pattern is low and variable across species, with high levels of transcripts in testis of humans and rats.[12] The enzyme is particularly sensitive to the opioid alkaloid papaverine[13] as well as the recently developed selective PDE10A inhibitor MP-10 (PF-02545920), and its trifluoroethylated analogue TP-10 (Figure 1), which both cause a robust increase in striatal tissue levels of cGMP and cAMP in vivo.[8] The search for novel PDE10A inhibitors is an increasingly competitive field, and the pharmaceutical industry has produced several PDE10A research programs.[14, 15] Accordingly, there is a great interest in non-invasively monitoring PDE10A expression and activity in normal brain, in neuropsychiatric diseases, and under potential therapy with selective PDE10A inhibitors. One approach to assess this aim is the development of PET radioligands, based on compounds that inhibit PDE10A (recently reviewed[14, 15]). Tu and co-workers[16] developed a selective PDE10A tracer using 11C-labeling of papaverine. However, [11C]papaverine seems unsuitable as a PDE10A radiotracer, due to its short retention time in the CNS and its low affinity and selectivity. ChemMedChem 0000, 00, 1 – 13 &1& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS www.chemmedchem.org Figure 1. Lead structures for PDE10A inhibitors. Further developments focused on MP-10 as a lead structure. Both the 18F-labeled derivative [18F]JNJ41510417[17] and the 11Clabeled parent compound[18] show PDE10A affinity in the low nanomolar range as well as high selectivity against other PDEs. Although in vivo evaluation in mice,[17] rat,[19] pig,[20] and nonhuman primate brain[19, 20] demonstrated specific binding, high plasma protein binding[17] and the presence of radiolabeled metabolites in the brain[18, 19] may limit further PET applications. In parallel to our own 18F-labeling work,[21] the same group isotopically 11C-labeled the quinoxaline derivative PQ-10,[22] which was developed as a drug candidate by Pfizer.[23] It was found to be unsuitable for further investigation, due to its low brain permeation in rats (0.06 % ID g 1 after 30 min), in addition to a lack of specific accumulation in the striatum as displayed by microPET imaging in rhesus monkeys.[22] Our 18F-labeled derivative displayed much higher brain uptake in mice striatum (2.3 % ID g 1) and excellent binding properties in vitro but failed to show specific binding in vivo.[21] An analysis of the interactions of PQ-10 with rat PDE10A indicates the suitability of positions 2, 6, and 7 for functionalization of the quinazoline skeleton. The nitrogen atom of the Gln 716 side chain forms two hydrogen bonds to the methoxy groups of PQ-10, and Phe 719 makes a p-stacking interaction with the quinazoline ring system (Figure 2). For the synthesized inhibitors, the binding mode was modeled based on the experimentally determined binding mode of PQ-10.[23] Figure 2 illustrates that both methoxy groups of PQ-10 have little solvent accessibility, so there is little space available for larger substitutions. Substituents at position 7 are oriented into a pocket of mixed hydrophobic and polar nature formed by the side chains of Gln 716, Val 668, Ile 682, Ala 679, Thr 675, and Ser 667. This cavity can probably accommodate substituents at C6 with a chain length of no more than three or four atoms. The methoxy group at position 6 also has a mixed polar and hydrophobic environment formed by the side chains of Met 703, Phe 719, Gln 716, and Tyr 683. There are also two water molecules nearby which are in contact with the bulk solvent at the protein surface. Longer linear chains appear to be possible at position 6, which would replace water molecule 154 and extend outside the binding cavity toward the solvent. Substitutions at position 2 point toward a larger cavity filled  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Figure 2. Scheme of the interactions of PQ-10 with rat PDE10A (PDB ID: 2OVY). Solvent accessibility of the inhibitor atoms is indicated by the size of the blue clouds. Blue shadows behind the residue labels indicate the strength of the interaction of these residues with the bound inhibitor. with water molecules bound tightly in a water network coordinated to the two metal ions and further polar groups. Based on these considerations, the aim of this study was to prepare fluorine-substituted derivatives as possible reference compounds, using PQ-10 as a lead structure, and examine these for their inhibitory effects and selectivity toward PDE10A with the objective of developing PET ligands for PDE10A imaging. Results and Discussion Chemistry The previously described PQ-10 (Figure 1) served as a lead structure for preparing new derivatives with structurally modified quinazoline and quinoxaline structures. These two aromatic systems are linked via an (R)-pyrrolidine-3-ol spacer, forming two isomers, of which the R enantiomer has been published as the biologically active form.[23] Without data on the inhibitory activity of the S form (compound 1) so far, this was synthesized first on the basis of (R)-pyrrolidine-3-ol, according to published methods.[23] We first varied position 2 of the quinazoline structure in two ways (Scheme 1): quinazolinones 3 a, b, d, and e were obtained from 2 a via conversion with the corresponding nitrile in dioxane/HCl in a high-pressure test tube, while 3 a and 3 c were prepared after converting the anthranilic acid with the appropriate acid anhydride at reflux, followed by the addition of ammonia or ammonium acetate. After chlorination with POCl3, chlorides 4 a–e, as well as commercially available 5, were coupled with (R)-2-(pyrrolidin-3-yloxy)quinoxaline 6[23] in dioxane/ water under alkaline conditions to obtain derivatives 7 a–f. ChemMedChem 0000, 00, 1 – 13 &2& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS www.chemmedchem.org Scheme 1. Preparation of 2-substituted derivatives 7 a–f. Reagents and conditions: a) R1CN, dioxane/HCl, 100 8C, 3 h, then 60 8C, 12 h for 3 a–b, 3 d–e; or b) (R1O)2O, MeCN, 0 8C!RT, 2 h, then NH4OAc or NH3, 0 8C!RT, 14 h, then HOAc, 120 8C, 3 h for 3 a, 3 c; c) POCl3, toluene, reflux, 2.5–4 h; d) compound 6, dioxane/H2O (5:1, v/v), K2CO3, 80 8C, 16 h. Phenol 15 was synthesized starting from vanillic acid 8 and building block 6, according to Scheme 2, to vary the substituents at position 7 on the quinazoline moiety. Compound 8 was converted in five steps into 7-(benzyloxy)4-chloro-6-methoxyquinazoline 13 using published methods.[24, 25] First, benzyl protecting groups were added to the benzoic acid and hydroxy group in 8 to obtain ester 9. This was followed by regioselective nitration[26] to afford 10. The product was then reduced using tin(II) chloride in ethyl acetate[24] to give aniline 11. The benzyl ester of 11 was cyclized to give lactam 12 at reflux using triethyl orthoformate as the C1 building block and ammonium acetate as the nitrogen source. After activation to afford 4-chlorine derivative 13 with phosphoryl chloride, we performed the coupling with 6 to afford intermediate 14 in the presence of potassium carbonate in aqueous dioxane at 80 8C. The coupling product 14 was treated with trifluoroacetic acid at reflux to remove the benzyl protecting group at position 7 on the quinazoline partial structure. The target ether derivatives, 16 a-j, were prepared by Oalkylation of the free phenolic functional group in 15. For the synthesis of 6-substituted derivatives, nitrobenzoate 17 was first converted into phenol 17 a (Scheme 3) via alkaline hydrolysis, followed by a six-step sequence in an analogous manner, as depicted in Scheme 2, to obtain 6-hydroxy quinazoline derivative 18. Subsequent O-alkylation resulted in derivatives 19 a–g. While the majority of compounds 16 and 19 were synthesized for in vitro analysis, we prepared derivatives 16 j and 19 g as one-step labeling precursors for 18F-labeling. The latter should lead to radiofluorinated ligands for imaging of PDE10A with PET.[21] The synthesis of fluoromethoxy derivatives 16 a and 19 a proved troublesome. As shown, for instance, in Scheme 4 for  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Scheme 2. Preparation of 7-substituted derivatives 16 a–j. Reagents and conditions: a) benzyl bromide (BnBr), K2CO3, DMF, 22 8C, 18 h; b) HNO3, H2SO4, CH2Cl2/acetic acid (HOAc), 0–5 8C, 4 h; c) SnCl2, EtOAc, 65 8C, 18 h; d) triethyl orthoformate [(EtO)3CH], NH4OAc, reflux, 4 h; e) POCl3, toluene, reflux, 2.5– 4 h; f) compound 6, dioxane/H2O (5:1, v/v), K2CO3, 80 8C, 16 h; g) TFA, reflux, 3 h; h) O-alkylation: R2-X (X = I, Br, OTs), K2CO3, (or Cs2CO3), DMF, 55 8C, 3– 48 h (or 70 8C, 3–5 h for 16 a and 16 j); ethylene carbonate (1,3-dioxolan-2one for 16 e), K2CO3, DMF, 150 8C, 5 h. 16 a, fluoromethyltosylate 23 was prepared as a reagent for synthesizing 16 a in three steps. First, p-toluenesulfonic acid was converted into silver salt 21 using silver(I) oxide. Reaction with diiodomethane produced the ditosylate 22 for conversion into fluoromethyltosylate 23. Tetra-n-butylammonium fluoride (TBAF), recrystallized from tBuOH, was used as the fluorinating agent.[27] Due to possible difluorination of 22 and instability of 23, a yield of only 8 % was observed after chromatographic purification. Fluoromethyltosylate 23 was then converted with phenol 15 using cesium carbonate to obtain a mixture of fluoromethoxy derivative 16 a and tosyl ester 24. As 16 a and 24 could not be separated chromatographically, the reaction mixture was treated with TBAF in dichloromethane to cleave tosyl ester 24, followed by final separation of 16 a from remaining phenol 15. Intermediate product 28 was synthesized according to Scheme 5,[28] based on commercially available 2,4-difluoro-1-nitrobenzene 25 for regioselective fluorine substitution at position 6 on the quinoxaline structure. First, the 2-fluorine atom in 25 was replaced by nucleophilic reaction with ethyl glycinate. The resulting ethyl N-arylglycinate 26 was reductively cyclized using SnCl2 or H2/Pd and then aromatized to quinoxalinole 28 using hydrogen peroxide. The 6,7-difluoroquinoxaline-2-ol 33 was synthesized from 4,5-difluoro-2-nitroaniline 31.[29, 30] After reduction to diamine 32 with SnCl2, chromatographic separaChemMedChem 0000, 00, 1 – 13 &3& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS Scheme 3. Preparation of 6-substituted derivatives 19 a–g. Reagents and conditions: a) 20 % NaOH, 95–100 8C, 12 h; b) benzyl bromide (BnBr), K2CO3, DMF, 22 8C, 18 h; c) SnCl2, EtOAc, 65 8C, 18 h, (!17 c); d) (EtO)3CH, NH4OAc, reflux, 4 h; e) POCl3, toluene, reflux, 2.5–4 h; f) compound 6, dioxane/H2O (5:1, v/v), K2CO3, 80 8C, 16 h, (!17 f); g) TFA, reflux, 3 h; h) O-alkylation: R2-X (X = I, Br, OTs, OSO2Ph), K2CO3, (or Cs2CO3), DMF, 55 8C, 3–48 h (or 70 8C, 3–5 h for 19 a and 19 g). www.chemmedchem.org Scheme 5. Synthesis of fluoroquinoxaline derivatives 30 and 35. Reagents and conditions: a) ethyl glycinate·HCl, K2CO3, DMF/H2O (15:1, v/v), reflux, 2 h; b) SnCl2/EtOH, reflux, 2–3 h or H2/Pd-C/MeOH, 22 8C, 24 h; c) 3 % H2O2, 1 m NaOH (aq), 85 8C, 2 h; d) SnCl2/EtOH, reflux, 1 h; e) glyoxylic acid monohydrate, EtOH, reflux, 1 h; f) (S)-N-Boc-3-pyrrolidinol, diethyl azodicarboxylate (DEAD), PPh3, THF, 22 8C, 18 h; g) TFA, CH2Cl2, 0–22 8C, 18 h (!29 a, 34 a); h) 6,7-dimethoxy-4-chloroquinazoline, dioxane/H2O (5:1, v/v), K2CO3, 80 8C, 14 h. Structure–activity relationships The inhibitory potency of novel quinazoline derivatives was determined via a scintillation proximity assay (SPA) using recombinantly produced PDEs. Human PDEs were used with the exception of bovine PDE6. The inhibitory potencies of the potential PET ligands 7 b, 16 c, and 19 c toward PDE1–10A were evaluated. Inhibitors with PDE IC50-values < 1000 nm were used to study the influence of substituents on selectivity. Variation on the quinazoline partial structure Scheme 4. Synthesis of fluoromethoxy derivative 16 a. Reagents and conditions: a) Ag2O, MeCN; b) CH2I2, MeCN, reflux 12 h, 60 8C, 12 h; c) TBAF·(tBuOH)4, MeCN, reflux, 2 h; d) compound 15, Cs2CO3, DMF, 70 8C, 3 h; e) TBAF·(tBuOH)4, CH2Cl2, 22 8C. tion under argon was required. Compound 32 was cyclized to quinoxalinole 33 with one equivalent of glyoxylic acid monohydrate at reflux. Similar to the synthesis of PQ-10,[23] quinoxalinoles 28 and 33 were converted using a Mitsunobu coupling with Boc-protected (S)-pyrrolidine-3-ol to intermediates 29 and 34, respectively. After removing the Boc protecting group in dichloromethane and trifluoroacetic acid, coupling with 6,7-dimethoxy-4chloroquinazoline was performed to obtain fluoroquinoxaline derivatives 30 and 35.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim The inhibitory potency of PDE10A was found to decrease with the size of the R1 substituents at position 2 (Scheme 1). The 2substituted derivatives exhibited significantly lower PDE10A selectivity ratios toward PDE3A than PQ-10, as shown in Table 1. Among these, both the chloromethyl and methyl derivatives 7 d and 7 a bind more tightly to PDE3A than to PDE10A, distinguishing these from the other derivatives in this class. Additionally, derivatives 7 a and 7 d exhibited inhibitory effects toward PDE6A, whereas fluoromethyl derivative 7 b did not. Although there is enough space to accommodate substitutions at position 2 in the PQ-10 binding mode to rat PDE10A (Figure 2), all substitutions investigated in this study increased the IC50 value for PDE10A (Table 1). The may result from hydrophobic substituents introduced at position 2 being too near the three water molecules of this cavity and having to replace some of these waters. As these water molecules are involved in a larger hydrogen bonding network, including polar protein ChemMedChem 0000, 00, 1 – 13 &4& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS www.chemmedchem.org Table 1. PDE10A inhibition (IC50) and selectivity ratio toward PDE3A (human) and PDE6 (bovine) of R1-substituted derivatives at position 1 of the quinazoline. Compd R1 IC50 [nm][a] 7a 7b 7c 7d 7e 7f PQ-10 CH3 CH2F CF3 CH2Cl 4-F-C6H4 Cl H 104 141 407 169 952 122 16 PDE10A selectivity ratio[b] PDE3A PDE6 0.83 2.29 2.53 0.25 > 1.05 3.47 11.7 7.31 19.7 6.09 3.53 > 1.05 8.20 > 312 87 nm), respectively. The classical electroisosteric hydroxyethyl derivative 16 e (IC50 = 140 nm) had a slightly lower inhibitory activity than fluoroethyl derivative 16 c (IC50 = 106 nm). Fluoropropyl derivative 16 f showed IC50 values similar to that of chloroethyl derivative 16 g. Lipophilic and branched substituents decreased inhibitory potency further: 2,2,2-trifluoroethyl derivative 16 h showed an IC50 value of 253 nm, while 2,2-difluoride derivative 16 d showed higher inhibitory potency with an IC50 value of 132 nm. An isopropyl substituent (in 16 i) decreased inhibitory potency further to IC50 = 308 nm. Substituting the methyl group with a hydrogen atom (in 15) decreases activity 80-fold in the micromolar range. Similarly, the large lipophilic benzyl substituent causes a loss in inhibitory potency (compound 14). Figure 3 A shows the modeled binding mode of inhibitor 16 a. Replacement of the methoxy by a fluoromethoxy group does not disturb binding to PDE10. After energy minimization, [a] n  4 (n = number of measurements); sn < 15 % (sn = SEM, standard error of the mean). [b] Expressed as a ratio between the IC50 value for a given PDE and the IC50 value for inhibition of PDE10A (selectivity ratio of all compounds for PDE10A = 1). groups, the replacement of these waters is energetically unfavorable unless they are replaced with suitable polar substituents. Inhibitory potency decreased with increasing size of the substituent at position 7 on the quinazoline partial structure (Table 2); that is, by either exchanging a hydrogen atom for a fluorine atom (cf. PQ-10 vs. 16 a; IC50 = 16 nm vs. 24 nm) or extending the alkyl chain (cf. PQ-10 and 16 b, IC50 = 16 nm vs. Table 2. PDE10A inhibition (IC50) and selectivity ratio toward PDE3A and PDE4A of derivatives substituted with R2O at position 7 of the quinazoline. Compd R2 IC50 [nm][a] 14 15 16 a 16 b 16 c 16 d 16 e 16 f 16 g 16 h 16 i 16 j PQ-10 Bn H CH2F CH2CH3 CH2CH2F CH2CHF2 CH2CH2OH CH2CH2CH2F CH2CH2Cl CH2CF3 CH(CH3)2 CH2CH2OTs CH3 2130 1280 24 87 106 132 140 144 149 253 308 > 1000 (49 %) 16 PDE10A selectivity ratio[b] PDE3A PDE4A – – 1.38 3.16 0.84 1.55 2.35 0.82 1.19 – – – 11.7 – – > 41.7 8.52 5.30 > 7.58 4.16 23.4 > 6.71 – – – > 312 [a] n  4 (n = number of measurements); sn < 15 % (sn = SEM, standard error of the mean). [b] Expressed as a ratio between the IC50 value for a given PDE and the IC50 value for inhibition of PDE10A (selectivity ratio of all compounds for PDE10A = 1).  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Figure 3. Molecular interactions of the ligands with PDE10A. A) Modeled binding mode of 16 a to the active site of rat PDE10A. B) Modeled binding mode of 19 c to the active site of rat PDE10A. ChemMedChem 0000, 00, 1 – 13 &5& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS 16 a superimposed well with PQ-10, in agreement with the low IC50 value of 16 a. Longer substituents, as in inhibitors 16 c and 16 f, result in a significant decrease in inhibitory potency due to the restricted size of the cavity. Especially for inhibitor 16 j, the substituent is too large, which correlates with a very high IC50 value (> 1000 nm). Interestingly, replacement of the methoxy group by a phenolic hydroxy group results in a large increase in IC50 value (compound 15). This can be attributed to the loss of hydrophobic interactions of the methoxy group. In addition, the carbonyl oxygen of the Gln 716 side chain is 3.8  too far from the hydroxy proton of 15 to form a strong hydrogen bond, and no water molecule is appropriately positioned to act as a hydrogen bond acceptor for the hydroxy group. In addition to PDE10A, the 7-substituted derivatives particularly inhibit PDE3A; this substance class is known for its crossaffinity between PDE10A and PDE3A.[23] PDE10A selectivity ratios toward PDE3A are lowered with increasing chain length (PQ-10 vs. 16 c and 16 f), especially with the introduction of a fluorine atom. This effect may be related to the electrochemical characteristics of the fluorine atom, as this tendency also appears in the chloroethyl derivative (see 16 c and 16 g); the latter has a larger van der Waals volume and is capable of similar electrostatic interactions. In contrast, ethyl derivative 16 b and hydroxyethyl derivative 16 e favor PDE10A, albeit with a fourfold decrease in selectivity ratio relative to the lead structure. In addition, the ethyl substituent in 16 b decreases PDE10A selectivity over PDE4A compared with the methyl substituent (PQ-10) by more than 30-fold. Introduction of both a fluorine atom and the hydroxy group (16 c and 16 e) decrease the selectivity ratio with respect to PDE4A by a further 50% relative to compound 16 b. Inhibitory activity toward PDE10A decreases with increasing chain length in derivatives varied at position 6 but not as drastically as for those substituted at position 7. The IC50 values for fluoromethyl, fluoroethyl, and difluoroethyl derivatives 19 a, 19 c, and 19 d are 60, 64, and 51 nm, respectively (Table 3). Fluoroalkyl substitution is better tolerated at position 6 than at position 7. However, extending the chain to the fluoropropyl substituents involves stronger attenuation of inhibitory potency by 3.5-fold relative to the fluoroethyl derivative. In comparison with position 7, short chain fluorinated substituents are tolerated, with only moderate decreases in inhibitory potency. Monofluoroethyl and difluoroethyl derivatives 19 c and 19 d show a clear improvement in selectivity against PDE3A. PDE10A selectivity toward PDE4A, as shown by 19 a and 19 c, is greater than 65, in contrast to position 7 derivatives (Table 2, 16 a and 16 c). In the modeled binding mode of 19 c, the quinazoline moiety of the inhibitor is shifted to the right and cannot form any hydrogen bonding interactions with Gln 716 (Figure 3 B). This shift could also influence the p-stacking interaction. Replacement of the methoxy group of PQ-10 by an ethoxy or fluoroethoxy substituent is better accommodated at position 6 by the enzyme than the corresponding substitutions at position 7. Longer substituents also further increase the IC50 value, although position 7 is oriented close to the protein surface.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemmedchem.org Table 3. PDE10A inhibition (IC50) and selectivity ratio toward PDE3A and PDE4A of derivatives substituted with R3O at position 6 of the quinazoline. Compd R3 IC50 [nm][a] 17 f 18 19 a 19 b 19 c 19 d 19 e 19 f Bn H CH2F CH2CH3 CH2CH2F CH2CHF2 CH2CH2CH2F CH2CF3 > 1000 (41.1 %) 641 60.2 29.5 63.8 51 187 99.7 PDE10A selectivity ratio[b] PDE3A PDE4A – – 3.14 14.0 25.9 48.4 4.53 > 10.0 – – 69.4 – ~ 78.4 – – – [a] n  4 (n = number of measurements); sn < 15 % (sn = SEM, standard error of the mean). [b] Expressed as a ratio between the IC50 value for a given PDE and the IC50 value for inhibition of PDE10A (selectivity ratio of all compounds for PDE10A = 1). However, crystal structures with longer substituents are reported to retain the hydrogen bonds to Gln 716 (PDB ID: 3QPP and 3QPO).[31] Similar to the situation for position 7, substitution of the methoxy group at position 6 by a hydroxy group greatly diminishes inhibitory potency (compound 18). This is most likely due to loss of the hydrophobic contact of the methyl group and the inability of the hydroxy group of 18 to form hydrogen bonds to Gln 716 or a water molecule. Variation on the quinoxaline partial structure The 3’’-R-linked quinoxalinyl substituent is crucial to biological activity for lead structure PQ-10; S-form 1 (Table 4) only shows inhibitory activity in the micromolar range. On the other hand, fluoride substitution at position 6’ on the quinoxaline structure in compound 30 is not only tolerated but also improves inhibitory potency to PDE10A. Difluoride substitution in compound 35 is tolerated with regard to inhibitory activity and drastically improves the selectivity profile toward PDE3A. The derivatized positions are highly solvent-accessible (Figure 2), and there are no ordered water molecules nearby. Metabolic stability We performed in vitro investigations with S9 fractions from rat liver to estimate metabolic stability (see Experimental Section). We used C18 solid-phase extraction for metabolite isolation and concentration, and HPLC–MS analysis for structure elucidation. The structure of the primary metabolite was resolved using high-resolution mass spectrometry (HRMS) and characteristic ESI-MS–MS fragmentation. Lead structure PQ-10 and the 6-fluoroethoxyquinazoline derivative 19 c were subjected to a heavy phase 1 metabolism in vitro as already described for other 6,7-dimethoxyquinazoline ChemMedChem 0000, 00, 1 – 13 &6& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS www.chemmedchem.org Table 4. PDE10A inhibition (IC50) and selectivity ratio toward PDE3A of derivatives substituted with X1/X2 at positions 6’ and 7’ of the quinoxaline structure. Compd X1/X2 3’’-config. PQ-10 1 30 35 H/H H/H F/H F/F R S R R IC50 [nm][a] PDE10A/PDE3A sel. ratio[b] 16.1 1100 11.4 26.1 11.7 – 14.5 60.2 [a] n  4 (n = number of measurements); sn < 15 % (sn = SEM, standard error of the mean). [b] Expressed as a ratio between the IC50 value for PDE3A and the IC50 value for inhibition of PDE10A (selectivity ratio of all compounds for PDE10A = 1). derivatives by Uckun et al.[32] Position 7 was largely oxidatively demethylated, while the methoxy groups at position 6 of the 7-fluoroethoxy derivative 16 c were mostly metabolically stable (Figure 4). Figure 4. Comparison between 16 c, 19 c and lead structure PQ-10. Conclusions The following points can be made regarding structure–activity relationships for the lead structure and on the basis of thirtytwo derivatives (Figure 5): 1. Short-chain substituents at position 2 on the quinazoline structure decrease PDE10A inhibition and strengthen PDE3A inhibition. A fluoromethyl substituent (IC50 PDE3A/PDE10A = 2.3) showed the most favorable selectivity relationship relative to methyl (IC50 PDE3A/PDE10A = 0.83) and chloromethyl substituents (IC50 PDE3A/PDE10A = 0.25).  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Figure 5. Substituents and their influence on structure–activity relationships. 2. Fluoromethyl and fluoroethyl substituents at position 6 on the quinazoline structure decrease PDE10A inhibition potency three- to fourfold when compared with the lead structure (IC50 PDE10A = 16.1 nm vs. 51–63 nm). Monofluoroethyl and difluoroethyl derivatives 19 c and 19 d are two- and fourfold more selective toward PDE3A (IC50 PDE3A/PDE10A = 26 (19 c), 48 (19 d), and 12 (PQ-10), respectively, relative to PQ-10). 3. Alkoxy substituents at position 7 on the quinazoline structure decrease PDE10A inhibition potency with increasing size; this is probably due to spatial limitations in the active binding pocket. Monofluoride substitution reduces selectivity to PDE3A. The chemophysical interactivity of fluorine may explain the improved activity toward PDE3A, as the chlorine substituent has the same, albeit somewhat weaker, effect. 4. The R-form of the lead structure is the active form toward PDE10A compared with a 69-fold decrease in inhibitory potency for the S-form. 5. Fluorine substitution at position 6’ of the quinoxaline structure improves PDE10A inhibition relative to the lead structure (IC50 PDE10A = 11.4 nm vs. 16.1 nm) and improves the selectivity profile against PDE3A (IC50 PDE3A/PDE10A = 14.47 vs. IC50 PDE3A/PDE10A = 11.67). 6’,7’-Difluoride substitution moderately reduces PDE10A activity while improving selectivity against PDE3A 60-fold. 6. Structure modification at position 6 of the quinazoline ring plays an essential role in a selective PDE10A inhibitor, as shown in Table 3. Varying the substituents at positions 7 or 2 of the quinazoline ring (Table 1 and 2) in the reference compound decreases selectivity against PDE3A. We also found chain length to be crucial to selectivity; only one fluoroethyl substituent improved the selectivity profile while maintaining inhibitory potency. Variations in chain length at position 6 (such as 19 a and 19 e) reduce selectivity to PDE3A. Additionally, fluoride substitution on the side chain plays an essential role in selectivity toward PDE3A; only 6-ethyl derivative 19 b is similar to PQ-10 in selectivity. The following conclusions can be made with regard to use as a PET ligand: 2-fluoromethylquinazoline derivative 7 b is unsuitable as a PET ligand due to its low inhibitory potency and selectivity, while fluoroquinoxaline derivatives such as compounds 30 and 35 in this study may reveal good inhibitory poChemMedChem 0000, 00, 1 – 13 &7& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS tency and selectivity characteristics. Accessibility to radiosynthesis presents a challenge, however. Of the 6- and 7-fluoroethoxyquinazoline derivatives, 7-fluoroethoxy derivative 16 c is metabolically stable in vitro but did not selectively inhibit PDE10A against PDE3A and PDE4A in an in vitro inhibition assay. 6-Fluoroethoxy derivative 19 c is a highly selective PDE10 A inhibitor but not metabolically stable in vitro. Current in vivo PET studies will have to clarify whether the selectivity lacking in 16 c and the in vitro metabolic instability lacking in 19 c are relevant in opposing the use of 6-fluoroalkoxybiarylbased and 7-fluoroalkoxybiaryl-based, respectively, PDE10 A inhibitors as PET ligands. Experimental Section General experimental design: All melting points were obtained on a Melting Point B540 (Bchi) apparatus and are uncorrected. Mass spectra (MS) were taken in ESI mode on a Bruker Daltonics Apex II (FTICR-MS, skimmer voltage 100 V, drying gas 100 8C). Highresolution mass spectra were recorded using a 7 T Apex II FT-ICRspectrometer. (NMR) spectroscopy was performed using Varian Mercury 300BB (300 MHz) and Varian Mercury 400BB (400 MHz) spectrometers; the spectra were standardized on the solvent signal. KBr-IR spectra were taken on a Perkin-Elmer 16PC FT-IR spectrometer, and solid IR spectra were taken on a Perkin-Elmer SpectrumOne FT-IR spectrometer. Optical rotation was determined using a semi-automatic Perkin-Elmer 241 polarimeter. Merck silica gel 60-coated aluminum plates were used for thin-layer chromatography. Spot values were detected at 254 and 366 nm. Metabolites were detected using an Agilent G1515B DAD HPLC–MS UV detector (254 nm), an Agilent LC/MSD SL MS detector (fragmentor voltage 100 V), and a Nucleodur C18 end-capped column, 10 cm; solvent A was MeCN containing 0.1 % formic acid, solvent B was water and 0.1 % formic acid, using a time (min)/sol. A/sol. B gradient of 1:5:95, 1!7:5!99:95!1, 7!9:99:1; flow rate: 1 mL min 1. Merck K60 silica gel was used for preparative column chromatography. Purities of the target compounds (> 95 %) were determined by HPLC analysis. For more information on HPLC methods of analysis, see Supporting Information. Unless otherwise noted, all solvents and reagents were commercially available and used without further purification. Compounds 6,[23] 9,[24] 10,[24, 26] 11,[24] 12,[25] and 17 a[33] were prepared according to published methods. Intermediate compounds 17 b–f and 18 were analogously prepared from 17 a. Compounds 28[34, 28] and 33[29, 30] have been previously described, and their synthesis was carried out according to published protocols. Molecular modeling: The binding mode of the inhibitors to the active site of rat PDE10A was modeled based on the structure of PDB ID: 2OVY[23] with the MOE program (Schrçdinger LLC, New York, NY (USA), 2012). Hydrogen atoms were added, and the protein structure was energy-minimized with the MMFF94x force field. After replacing the bound inhibitor PQ-10 with the desired derivative, the structure was again energy-minimized. The final model was visualized with PyMOL (http://www.pymol.org). The ligand interaction scheme for inhibitor PQ-10 was prepared with MOE. General procedure 1 for the synthesis of 2-substituted 6,7-dimethoxyquinazoline-4(3H)-ones (3 a, b, d, and e): Compound 2 b (3 mmol, 1 equiv) and nitrile (7.5 mmol, 2.5 equiv) were dissolved in dioxane (6 mL) saturated with HCl in a high-pressure reaction container, and the mixture was subjected to ultrasound for 2.5 h. The reaction mixture was then heated at 100 8C for 3 h and heated  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemmedchem.org at 60 8C for another 12 h. The reaction mixture was then completely dissolved in water (15 mL). The pH was then increased to 6–7 with 1 n NaOH, and the resulting white precipitate was extracted by suction and washed with water. 2-(Fluoromethyl)-6,7-dimethoxyquinazoline-4(3H)-one (3 b): Procedure 1 using 2 a (630 mg, 3.00 mmol, 1 equiv), fluoroacetonitrile (0.45 mL, 7.50 mmol, 2.5 equiv), and dioxane/HCl (6 mL) produced 3 b (C11H11FN2O3 ; M = 238.22 g mol 1) as a colorless solid (yield: 714 mg, 3.00 mmol, 99 %): mp: 275–276 8C (under decomposition); 1 H NMR (300 MHz, [D6]DMSO): d = 3.85 (s, 3 H, CH3), 3.87 (s, 3 H, CH3), 5.37 (d, JH,F = 46.2 Hz, CH2F), 7.24 (s, 1 H, CH), 7.44 (s, 1 H, CH), 11.12 ppm (bs, 1 H, NH); 13C NMR (75 MHz, [D6]DMSO): d = 56.5 (OCH3), 56.8 (O-CH3), 81.1 (d, JC,F = 170.7 Hz, 1 C, CH2F), 105.9 (CH), 107.1 (CH), 114.9 (C), 142.2 (C), 149.8 (C-OCH3), 151.7 (d, JC,F = 18.9 Hz, 1 C, C-CH2F) 155.5 (C-OCH3), 160.7 ppm (C=O); 19F NMR (282 MHz, [D6]DMSO): d = 222.6 ppm (t, JF,H = 46.4 Hz, 1F); IR (KBr): ~ n = 2828 (CH3), 3159, 3123, 1720, (-CONH-) 1655, 1483 cm 1 (arom. C=C); LRMS-ESI( ): m/z calcd C11H12FN2O3 + [M H] + : 239.0, detected: 239.1. 6,7-Dimethoxy-2-(trifluoromethyl)-quinazoline-4(3H)-one (3 c): 2Amino-4,5-dimethoxybenzoic acid (2 b, 197 mg, 1.00 mmol, 1 equiv) was dissolved in CH3CN (5 mL). Pyridine (0.4 mL, 5.00 mmol, 5 equiv) was added, and trifluoroacetic anhydride (0.42 mL) was slowly add dropwise to the ice-cooled mixture. The mixture was stirred for 2 h while slowly warming to ambient temperature. Next, the solution was slowly added dropwise to an icecooled solution of ammonium carbonate (384 mg, 4.0 mmol, 4 equiv) in CH3CN (5 mL) and stirred overnight while the solution slowly warmed to room temperature. After removing the solvent under reduced pressure, the oily, highly viscous precipitate was heated at 120 8C in acetic acid (5 mL) for 2.5–3 h. The reaction mixture was then concentrated by low-pressure solvent evaporation for drying. An EtOH/water (5 mL, 1:1, v/v) solution was added, and the colorless solid precipitate was washed in ice-cooled EtOH/ water solution and dried to give 3 c (yield: 139 mg, 0.51 mmol, 50 %): mp: 284 8C (aq. EtOH); 1H NMR (300 MHz, [D6]DMSO): d = 3.93 (s, 3 H, O-CH3), 3.91 (s, 3 H, O-CH3), 7.32 (s, 1 H, CH), 7.50 (s, 1 H, CH), 13.40 ppm (s, 1 H, NH); 13C NMR (75 MHz, [D6]DMSO): d = 55.9 (O-CH3), 56.2 (O-CH3), 104.9 (CH), 108.8 (CH), 115.4 (CH), 118.0 (q, JC,F = 274 Hz, CF3), 142.5 (C), 150.2 (C-OCH3), 155.0 (C-OCH3), 160.7 ppm (C=O), one quarternery C signal was not detectable; 19 F NMR (282 MHz, [D6]DMSO): d = 69.4 ppm (s, CF3); LRMS-ESI(+): m/z calcd C11H11F3N2O4 [M + H2O]: 292.1, detected: 291.8; IR (KBr): ~ n = 2899 (CH3), 3039, 1664, 1492 (-CONH-), 1608, 1518 cm 1 (arom. C=C). General procedure 2 for the synthesis of 4-chloro-quinazolines 4 a–e, 13, and 17 e: The respective quinazoline-4H-one derivative (1 mmol) was extracted three times from the remaining water component by co-distillation with toluene (10 mL). The reaction mixture was then heated in toluene (10 mL) with phosphorylchloride (3 mL) for 2.5–4 h at reflux until a clear solution formed. After cooling, the excess phosphorylchloride was removed under reduced pressure by co-distillation with toluene. The solid was obtained in almost quantitative yield and was further converted as a raw product. As an alternative to distillation of the excess phosphorylchloride, the reaction could be cooled and set on ice, the mixture extracted in several passes with CH2Cl2, the organic phases concentrated and dried over Na2SO4, and the solvent removed under reduced pressure. The resulting solid was further converted as a raw product. ChemMedChem 0000, 00, 1 – 13 &8& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS 4-Chloro-6,7-dimethoxy-2-methylquinazoline (4 a): General procedure 2 was performed using 3 a to afford 4 a as a red solid: 1 H NMR (300 MHz, CDCl3): d = 2.94 (s, 3 H, C-CH3), 4.08 (s, 3 H, OCH3), 4.12 (s, 3 H, O-CH3), 7.40 (s, 1 H, CH), 7.69 ppm (s, 1 H, CH); IR (KBr): ~ n = 2964 (CH3), 1504 cm 1 (N=C). 4-Chloro-2-(fluoromethyl)-6,7-dimethoxyquinazoline (4 b): General procedure 2 was performed using 3 b to afford 4 b as an orange solid: 1H NMR (300 MHz, CDCl3): d = 3.85 (s, 3 H, O-CH3), 3.87 (s, 3 H, O-CH3), 5.57 (d, JH,F = 46.7 Hz, CH2F), 7.37 (s, 1 H, CH), 7.37 ppm (s, 1 H, CH). 13C NMR (75 MHz, CDCl3): d = 56.6 (O-CH3), 56.8 (O-CH3), 83.4 (d, JH,F = 175.0 Hz, 1 C, CH2F), 102.9 (CH), 107.1 (CH), 118.5 (C), 149.3 (C), 151.7 (C-OCH3), 157.3 (C-OCH3), 158.7 (d, JH,F = 18.2 Hz, 1 C, C-CH2F), 159.6 ppm (C-Cl); 19F NMR (282 MHz, CDCl3): d = 222.5 ppm (t, JH,F = 46.7 Hz, 1F); LRMS-ESI(+):m/z calcd: 257.0 [M + H] + : detected: 256.9; m/z calcd: 257.0 [M + Na] + : 279.0, detected: 278.9. General procedure 3 for the coupling of 4-chloroquinazolines with 2-(pyrrolidin-3-yloxy)quinoxalines to quinoxaline-2-yloxy)pyrrolidine-1-yl)-quinazolines (7 a–f, 14, 17 f): 4-Chloroquinazoline 4 a–e in excess was added to a solution of 6 (215 mg, 0.69 mmol, 1 equiv) in dioxane/water (12 mL, 5:1 v/v) (min 1.1 equiv with regard to the quinazoline-4-3(H)-one derivative) and K2CO3 (5 equiv). The suspension was stirred at 80 8C for 16 h. After removing the solvent under reduced pressure, the residue was re-dissolved in CH2Cl2 (20 mL) and the organic layer was washed in water, 5 % citric acid, and brine. After drying the organic layer over Na2SO4, the solvent was removed under reduced pressure. (R)-6,7-Dimethoxy-2-methyl-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1-yl)-quinazoline (7 a): General procedure 3 was performed using 4 a to afford a crude product, which was purified by recrystallization from MeOH to produce 7 a as colorless crystals (yield: 180 mg, 0.43 mmol, 62 %): mp: 118–120 8C; [a]D23 = + 1928 (0.43, CHCl3); 1H NMR (300 MHz, CDCl3): d = 2.57–2.30 (m, 2 H), 2.58 (s, 3 H), 3.95 (s, 3 H), 3.98 (s, 3 H), 4.44–4.07 (m, 4 H), 5.97–5.89 (m, 1 H), 7.45 (s, 1 H), 7.18 (s, 1 H), 7.58 (ddd, JH,H = 8.4, 7.0, 1.5 Hz, 1 H), 7.69 (ddd, JH,H = 8.4, 7.0, 1.5 Hz, 1 H), 7.84 (dd, JH,H = 8.3, 1.0 Hz, 1 H), 8.01 (dd, JH,H = 8.2, 1.4 Hz, 1 H), 8.47 ppm (s, 1 H); 13C NMR (75 MHz, CDCl3): d = 26.1, 31.7, 48.8, 55.8, 56.2, 56.2, 74.6, 104.6, 106.9, 108.1, 127.0, 127.4, 129.1, 130.4, 139.1, 139.8, 140.2, 146.9, 149.3, 154.0, 156.5, 159.5, 162.1 ppm; IR (KBr): ~ n = 2957 CH, 1571 (C=N), 1511 cm 1 (C=N); HRMS-ESI(+) m/z calcd: [M + H] + : 418.18737, detected: 418.18690, m/z calcd: [M + Na] + : 440.16931, detected: 440.16920., m/z calcd [2 M + Na] + : 857.34940, detected: 857.34843. (R)-2-Chloro-6,7-dimethoxy-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1-yl)-quinazoline (7 f): A suspension of 2,4-dichloro-6,7-dimethoxyquinazoline (1.03 g, 3.8 mmol, 1 equiv), potassium carbonate (2.7 g, 19.5 mmol, 5.14 equiv) and 6 (1.19 g, 3.8 mmol, 1 equiv) in dioxane/water (30 mL, 5:1, v/v) was stirred for 24 h at 80 8C, yielding a colorless solid precipitate. After removing the solvent under reduced pressure, water (40 mL) was poured into the reaction mixture. The precipitate was filtered and washed in water and MeOH to produce 7 f as a colorless solid (yield: 1.50 g, 3.42 mmol, 88 %: mp: 213–214 8C (MeOH); [a]D23 = + 2138 (0.81, CHCl3); 1H NMR (300 MHz, CDCl3): d = 2.61–2.32 (m, 2 H), 3.96 (s, 2 H), 3.97 (s, 2 H), 4.44–4.13 (m, 4 H), 6.00–5.87 (m, 1 H), 7.15 (s, 1 H), 7.45 (s, 1 H), 7.59 (ddd, JH,H = 8.4, 7.0, 1.5 Hz, 1 H), 7.69 (ddd, JH,H = 8.4, 7.0, 1.5 Hz, 1 H), 7.84 (dd, JH,H = 8.3, 1.1 Hz, 1 H), 8.02 (dd, JH,H = 8.2, 1.4 Hz, 1 H), 8.47 ppm (s, 1 H); 13C NMR (75 MHz, CDCl3): d = 31.6, 49.1, 56.2, 56.3, 56.3, 74.3, 107.2, 127.1, 129.2, 140.1, 108.5, 139.2, 104.6, 127.4, 130.5, 139.6, 147.7, 154.6, 155.3, 156.3, 160.4 ppm, one quaternary C signal was not detectable; IR (KBr): ~ n = 2877 (CH), 1575 (C=N),  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemmedchem.org 1499 cm 1 (C=N); HRMS-ESI(+): m/z calcd [M + H] + : 438.13274, detected: 438.13234. m/z calcd [MNa] + : 460.11469, detected: 460.11478. m/z calcd [2 M + Na] + : 897.24016, detected: 897.23952. 7-(Benzyloxy)-4-chloro-6-methoxyquinazoline (13): General procedure 2 was performed using 12[25] to yield compound 13 as pale-yellow crystals (yield: 199 mg, 0.66 mmol, 94 %): mp: > 140 8C (under decomposition); 1H NMR (300 MHz, CDCl3): d = 4.12 (s, 3 H), 5.47 (s, 2 H), 7.35–7.45 (m, 6 H), 8.11 (s, 1 H), 9.02 ppm (s, 1 H). (R)-7-(Benzyloxy)-6-methoxy-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1-yl)-quinazoline (14): General procedure 3 was performed using 6 (270 mg, 1.25 mmol, 1 equiv) and 13 (358 mg, 1.19 mmol, 0.95 equiv) in 12 mL dioxane/water (5:1, v/v) and potassium carbonate (866 mg, 6.27 mmol, 5.0 equiv). The foamy yellow crude product was recrystallized from 20 mL MeOH/EtOAc (95:5) to yield 14 as colorless crystals (yield: 305 mg, 0.64 mmol, 57 %): Rf = 0.56 (EtOAc/MeOH, 9:1, v/v), 0.59 (CHCl3/MeOH, 95:5, v/v); mp: 86–90 8C (MeOH/EtOAc); [a]D23 = + 157.98 (c 1.06, CHCl3); 1H NMR (300 MHz, CDCl3): d = 2.45 (m, 2 H, 4-CH2), 3.97 (s, 3 H, 6-OCH3), 4.17 (m, 2 H, 5CH2), 4.29 (m, 1 H, 2-CH2), 4.41 (dd, JH,H = 1.32 Hz, JH,H = 8.20 Hz, 1 H, 2-CH2), 5.27 (s, 2 H, CH2, benzyl), 5.94 (m, 1 H, 3-CH), 7.28 (s, 1 H, 8CH, quinazoline), 7.30–7.40 (m, 3 H, 3,4,5-CH, benzyl), 7.47 (m, 2 H, 2,5-CH, benzyl), 7.49 (s, 1 H, 5-CH, quinazoline), 7.58 (ddd, JH,H = 1.5 Hz, 7.1 Hz, 8.4 Hz, 1 H, 6-CH, quinoxaline), 7.68 (ddd, JH,H = 1.5 Hz, 7.1 Hz, 8.4 Hz, 1 H, 7-CH, quinoxaline), 7.84 (dd, JH,H = 1.1 Hz, 8.3 Hz, 1 H, 8-CH, quinoxaline), 8.02 (dd, JH,H = 1.3 Hz, 8.2 Hz, 1 H, 5CH, quinoxaline), 8.48 (s, 1 H, 3-CH, quinoxaline), 8.50 ppm (s, 1 H, 2-CH, quinazoline); 13C NMR (75 MHz, CDCl3): d = 31.7 (4-CH2), 49.0 (5-CH2), 56.0 (2-CH2), 56.5 (6-OCH3), 70.9 (CH2, benzyl), 74.7 (3-CH), 105.1 (5-CH, quinazoline), 109.0 (8-CH, quinazoline), 110.5 (4a-C, quinazoline), 127.1 (6-CH, quinoxaline), 127.5 (8-CH, quinoxaline), 127.7 (2,6-CH, benzyl), 128.4 (4-CH, benzyl), 128.9 (3,5-CH, benzyl), 129.3 (5-CH, quinoxaline), 130.6 (7-CH, quinoxaline), 136.1 (1-CH, benzyl), 139.3 (4a-C, quinoxaline), 139.8 (3-CH, quinoxaline), 140.3 (8a-C, quinoxaline), 147.9 (8a-C, quinazoline), 148.1 (6-C, quinazoline), 153.2 (2-CH, quinazoline), 153.3 (7-C, quinazoline), 156.5 (2-C, quinoxaline), 159.4 ppm (4-C, quinazoline); IR (KBr): ~ n = 2914 (-CH2-, -CH-), 1664, 1643, 1500 (conj. -C=N-), 1570, 1501 cm 1 (arom. C=C); UV/VIS (CH3CN): l (loge) = 220 (4.63), 246 (4.66), 324 nm (4.21); MS (EI): m/z (%) = 479 (M + , 4), 333 (M + C8H6N2O, 99), 266 (M + C12H12N3O, 13), 214 (M + C16H13N2O2, 14); HRMS-ESI(+): m/z calcd: C28H26N5O3 + [M + H] + : 480.20302, detected: 480.20276. (R)-6-Methoxy-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1-yl)-quinazoline-7-ol (15): Compound 14 (441 mg, 0.92 mmol, 1 equiv) was held at reflux in TFA (5 mL) for 3 h. After co-distillation of the excess TFA with toluene, the residue was suspended in a solution of Na2SO4 (20 mL, 100 g L 1) for 12 h. Compound 15 was collected by filtration and washed in water (20 mL) and MeOH (20 mL) to yield a colorless solid (yield: 350 mg, 0.90 mmol, 98 %): mp: 260– 265 8C; 1H NMR (300 MHz, [D6]DMSO): d = 2.68 (m, 2 H, 4-CH2), 4.16 (s, 3 H, 6-OCH3), 4.51 (m, 4 H, 2-CH2, 5-CH2), 6.12 (m, 1 H, 3-CH), 7.31 (s, 1 H, 8-CH, quinazoline), 7.46 (s, 1 H, 5-CH quinazoline), 7.64 (ddd, JH,H = 1.5 Hz, 7.0 Hz, 8.3 Hz, 1 H, 6-CH quinoxaline), 8.0 (ddd, JH,H = 1.5 Hz, 7.0 Hz, 8.4 Hz, 1 H, 7-CH, quinoxaline), 8.07 (dd, JH,H = 1.4 Hz, 8.3 Hz, 1 H, 8-CH, quinoxaline), 8.21 (dd, JH,H = 1.3 Hz, 8.2 Hz, 1 H, 5CH, quinoxaline), 8.52 (s, 1 H, 3-CH, quinoxaline), 8.74 ppm (s, 1 H, 2-CH, quinazoline); 13C NMR (75 MHz, [D6]DMSO): d = 30.8 (4-CH2), 48.1 (5-CH2), 55.3 (2-CH2), 55.4 (6-OCH3), 74.9 (3-CH), 104.5 (5-CH, quinazoline), 109.9 (8-CH, quinazoline), 107.4 (4a-C, quinazoline) 126.9 (6-CH, quinoxaline), 127.0 (8-CH, quinoxaline), 128.7 (5-CH, quinoxaline), 130.5 (7-CH, quinoxaline), 138.4 (4a-C, quinoxaline), 139.58 (3-CH, quinoxaline), 140.1 (8a-C, quinoxaline), 148.0 (6-C, quinazoline), 148.9 (8a-C, quinazoline), 152.2 (2-CH, quinazoline), ChemMedChem 0000, 00, 1 – 13 &9& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS 153.3 (7-C, quinazoline), 156.4 (2-C, quinoxaline), 158.4 ppm (4-C, quinazoline); IR (KBr): ~ n = 3000(br) (OH), 2946 (-CH2-, -CH-), 1570, 1492 (conj. -C=N-), 1570 cm 1 (arom. C=C); UV (MeOH): l (loge) = 220 (4.80), 246 (4.82), 323 nm (4.42); HPLC-LRMS-ESI(+): m/z (tr) calcd C21H20N5O3 + [M + H] + : 390.2, detected: 390.2 (4.27 min, method A); ESI-MS–MS (+): m/z = 244 ([M + H + ] C8H6N2O); HRMSESI(+): m/z calcd: C21H20N5O3 + [M + H] + : 390.15607, detected: 390.15565; MS (EI): m/z (%) = 389 (M + , 5), 243 (M + C8H6N2O, 100), 216 (M + C9H6N2O2, 13), 176 (M + C12H12N3O, 49). General procedure 4 for the O-alkylation of 15 and 18 to compounds 16 a–j and 19 a–g: The corresponding base was suspended under argon atmosphere for 15 min in a solution of 15 or 18, as applicable. After addition of the alkylation reagent, the solution was stirred at 55 8C for 3–48 h until 15 or 18, respectively, were completely consumed (TLC control). The residue was dissolved in CH2Cl2 and water after evaporating the solvent under reduced pressure. After washing the organic layer in 1 n NaOH, water, and brine and drying over Na2SO4, the solvent was evaporated under reduced pressure. (R)-7-Ethoxy-6-methoxy-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1yl)-quinazoline (16 b): General procedure 4 was performed using 15 (100 mg, 0.23 mmol, 1 equiv), K2CO3 (54 mg, 0.35 mmol, 1.5 equiv), ethyl bromide (80 mL, 0.92 mmol, 4 equiv), and DMF (2 mL). The crude product was purified by column chromatography (CHCl3/MeOH, 98:2!97:3, v/v) to produce 16 b as a colorless solid (yield: 22.6 mg, 0.05 mmol, 24 %): Rf = 0.59 (CHCl3/MeOH, 9:1, v/v); mp: 105–106 8C; [a]D23 = + 179.28 (c 0.25, CHCl3); 1H NMR (300 MHz, CDCl3): d = 1.54 (t, JH,H = 7.0 Hz, 3 H, CH3-ethoxy), 2.62– 2.35 (m, 2 H, 4-CH2), 3.97 (s, 3 H, O-CH3), 4.15–4.55 (m, 4 H, 2-CH2 5CH2), 4.43 (q, JH,H = 7.0 Hz, 2 H, CH2-ethoxy), 5.96 (m, 1 H, 3-CH), 7.29 (s, 1 H, 8-CH, quinazoline), 7.50 (s, 1 H, 5-CH, quinazoline), 7.60 (ddd, JH,H = 8.4 Hz, 7.0 Hz, 1.5 Hz, 1 H, 6-CH, quinoxaline), 7.70 (ddd, JH,H = 8.4 Hz, 7.0 Hz, 1.5 Hz, 1 H, 7-CH, quinoxaline), 7.85 (dd, JH,H = 8.3 Hz, 1.0 Hz, 1 H, 8-CH, quinoxaline), 8.03 (dd, JH,H = 8.2 Hz, 1.5 Hz, 1 H, 5-CH, quinoxaline), 8.49 (s, 1 H, 3-CH, quinoxaline), 8.52 ppm (s, 1 H, 2-CH, quinazoline); 13C NMR (75 MHz, CDCl3): d = 14.6 (2-CH3, ethoxy), 31.7 (4-CH2), 49.1 (5-CH2), 56.2 (6-OCH3), 56.4 (2-CH2), 64.9 (1-CH2, ethoxy), 74.5 (3-CH), 104.7 (5-CH, quinazoline), 107.1 (8-CH, quinazoline), 109.7 (4a-C, quinazoline), 127.1 (6-CH, quinoxaline), 127.4 (8-CH, quinoxaline), 129.2 (5-CH, quinoxaline), 130.5 (7-CH, quinoxaline), 139.2 (4a-C, quinoxaline), 139.7 (8a-C, quinoxaline), 140.2 (3-CH, quinoxaline), 145.2 (4-C, quinazoline), 148.1 (6-C, quinazoline), 152.2 (2-CH, quinazoline), 153.9 (7-C, quinazoline), 156.4 (2-C, quinoxaline), 159.3 ppm (8a-C, quinazoline); IR (KBr): ~ n = 2930 (CH3, -CH2-, -CH-), 1726, 1572, 1508 (conj. -C=N-), 1219 (-Ar-O-CH2), 1135 (O-CH3), 837 (isolated H, quinazoline), 772 cm 1 (4H quinoxaline); UV/VIS (MeOH): l (loge) = 225(4.21), 250 (4.23), 325 nm (3.99); HRMS-ESI(+): m/z calcd: C23H23FN5O3 + [M + H] + : 418.18737, detected: 418.18728; MS (EI): m/z (%) = 417 (M + , 4), 271 (M + C8H6N2O, 100), 244 (M + C9H6N2O2, 13). (R)-7-(Fluoromethoxy)-6-methoxy-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1-yl)-quinazoline (16 a): Compound 15 (140 mg, 0.36 mmol, 2 equiv) and Cs2CO3 (290 mg, 0.90 mmol, 5 equiv) were suspended in DMF (2 mL), and a solution of 23 (36 mg, 0.17 mmol, 1 equiv) in DMF (2 mL) was added dropwise. The reaction mixture was stirred for 3 h at 70 8C under argon atmosphere. After evaporating the solvent under reduced pressure, the residue was re-dissolved in CH2Cl2 (10 mL), and TBAF·(tBuOH)4 (500 mg, 0.90 mmol, 5 equiv) or an equivalent solution of TBAF in THF was added. The solution was stirred at room temperature until side product 24 decomposed to 15 (TLC control, CHCl3/MeOH, 95:5). After washing the organic layer with water and brine and drying over Na2SO4, the  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemmedchem.org solvent was evaporated under reduced pressure. The residue was purified by column chromatography (CHCl3/MeOH, 100:0!100:0.5) to produce 16 a as a pale-yellow solid (yield: 63.5 mg, 0.15 mmol, 41 %); mp: 90–93 8C; [a]D23 = + 154.58 (c 0.53, CHCl3); 1H NMR (300 MHz, CDCl3): d = 2.41–2.64 (m, 2 H, 4-CH2), 4.00 (s, 3 H, 6OCH3), 4.21–4.53 (m, 4 H, 5-CH2, 2-CH2), 5.89 (ddd, JH,F = 52.9, 5.1, 2.8 Hz, 2 H, CH2F fluoromethoxy), 5.95–5.99 (m, 1 H, 3-CH), 7.58 (s, 1 H, 8-CH, quinazoline), 7.59 (ddd, JH,H = 8.3, 7.0, 1.6 Hz, 1 H, 6-CH, quinoxaline), 7.65 (m, 1 H, 5-CH, quinazoline), 7.70 (ddd, JH,H = 8.4, 7.1, 1.5 Hz, 1 H, 7-CH, quinoxaline), 7.84 (dd, JH,H = 8.3, 1.3 Hz, 1 H, 8CH, quinoxaline), 8.02 (dd, JH,H = 8.2, 1.3 Hz, 1 H, 5-CH, quinoxaline), 8.48 (s, 1 H, 3-CH, quinoxaline), 8.49 ppm (s, 1 H, 2-CH, quinazoline); 13 C NMR (75 MHz, CDCl3): d = 1.6 (4-CH2), 49.3 (5-CH2), 56.5 (2-CH2), 56.5 (6-OCH3), 74.3 (3-CH), 99.9 (d, JC,F = 223.9 Hz, CH2F, fluoromethoxy), 106.4 (5-CH, quinazoline), 106.6 (br, 8-CH, quinazoline), 110.2 (4a-C, quinazoline), 127.3 (8-CH, quinoxaline), 127.4 (6-CH, quinoxaline), 129.2 (5-CH, quinoxaline), 130.7 (7-CH, quinoxaline), 139.3 (4a-C, quinoxaline), 139.6 (3-CH, quinoxaline), 140.0 (8a-C, quinoxaline), 148.4 (br, 2-CH, quinazoline), 148.9 (6-C, quinazoline), 151.7 (d, JC,F = 2.2 Hz, 7-C, quinazoline), 156.1 (2-C, quinoxaline), 158.6 (4-C, quinazoline), 158.7 ppm (8a-C, quinazoline); 19F NMR (282 MHz, CDCl3): d = 152.8 ppm (t, J = 52.9 Hz); IR (KBr): ~n = 2930 (-CH2-, -CH-), 1571, 1501 cm 1 (conj. -C=N); HRMS-ESI(+): m/z calcd: C22H21FN5O3 + [M + H] + : 422.16229, detected: 422.16215 + C22H20FN5O3Na [M + Na] + : 444.14424, detected: 444.14432. C44H40FN10O6Na + [2 M + Na] + : 865.29926, detected: 865.29964; C50H45FN10O8SNa + [M + M26 + H] + : 965.31993, detected: 965.32215; MS (EI): m/z (%) = 421 (M + , 6), 275 (M + C8H6N2O, 100), 208 (M + C12H12N3O, 54). (R)-6-(2-Fluoroethoxy)-7-methoxy-4-(3-(quinoxaline-2-yloxy)-pyrrolidine-1-yl)-quinazoline (19 c): General procedure 4 was performed using 18 (43 mg, 0.11 mmol, 1 equiv), K2CO3 (30 mg, 0.21 mmol, 2 equiv), 2-brom-1-fluoroethane (19 mg, 0.15 mmol, 1.5 equiv) and DMF (2 mL). The product was purified by column chromatography (CHCl3/MeOH 100:0!98:2 v/v) to produce 19 c (C23H22FN5O3 ; M = 435.46 g mol 1) as a colorless solid: Yield: 43 mg (0.10 mmol, 90 %); mp: 113–115 8C; [a]D23 = + 132.468 (c 0.92, CH3OH); 1H NMR (300 MHz, CDCl3): d = 2.42 (m, 1 H, 4-CH2), 2.50 (m, 1 H, 4-CH2), 3.97 (s, 3 H, 7-OCH3), 4.16–4.42 (m, 6 H, 5-CH2, 2-CH2, 1CH2 fluoroethoxy), 4.72 (m, 1 H, 2-CH, fluoroethoxy), 4.88 (m, 1 H, 2CH, fluoroethoxy), 5.93 (m, 1 H, 3-CH), 7.28 (s, 1 H, 8-CH, quinazoline), 7.62 (s, 1 H, 5-CH, quinazoline), 7.56 (ddd, JH,H = 1.5 Hz, 9 Hz, 9 Hz, 1 H, 6-CH, quinoxaline), 7.66 (ddd, JH,H = 1.5 Hz, 9 Hz, 9 Hz, 1 H, 7-CH, quinoxaline), 7.83 (dd, JH,H = 1.1 Hz, 9 Hz, 1 H, 8-CH, quinoxaline), 8.00 (dd, JH,H = 1.3 Hz, 9 Hz, 1 H, 5-CH, quinoxaline), 8.45 (s, 1 H, 3-CH, quinoxaline), 8.47 ppm (s, 1 H, 2-CH, quinazoline); 13 C NMR (100 MHz, CDCl3): d = 31.6 (4-CH2), 49.2 (5-CH2), 56.3 (2CH2), 56.4 (6-OCH3), 69,5 (d, JC,F = 20,4 Hz, 1-CH2, fluoroethoxy); 74.5 (3-CH), 82.3(d, JC,F=171.3 Hz, 2-CH2, fluoroethoxy), 106.5 (5-CH, quinazoline), 108.7 (8-CH, quinazoline), 109.4 (4a-C, quinazoline), 127.2 (6-CH, quinoxaline), 127.4 (8-CH, quinoxaline), 129.2 (5-CH, quinoxaline), 130.5 (7-CH, quinoxaline), 139.2 (4a-C, quinoxaline), 139.7 (3-CH, quinoxaline), 140.2 (8a-C, quinoxaline), 146.9 (8a-C, quinazoline), 147.0 (4-C, quinazoline), 152.2 (6-C, quinazoline), 155.2 (2-CH, quinazoline), 156.4 (7-C, quinoxaline), 159.2 (4-C, quinazoline)., 162,7 ppm (2-C, quinazoline); 18F NMR (282 MHz, CDCl3): d = 223 ppm (tt, JH,F = 47.37, 28.20 Hz,1F); IR (KBr): ~ n = 2929, 1490 (-CH2-, -CH-), 1616, 1570, 1506 (conj. -C=N-), 1471 (arom. C=C), 1304 (C-F), 1222 (-Ar-O-CH2), 1136 (O-CH3), 882 (isolated H, quinazoline), 761 cm 1 (4H quinoxaline); UV/VIS (MeOH): l (loge) = 218 (4,60), 246 (4.65), 324 nm (4.24); HRMS-ESI(+): m/z calcd: C23H23FN5O3 + [M + H] + : 436.17794, detected: 436.17745; MS (EI): m/z (%) = 435 (M + , 19), 289 (M + C8H6N2O, 100), 273 (M + ChemMedChem 0000, 00, 1 – 13 &10& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS C8H6N2O2, 70), 222 (M + C12H12N3O, 31), 175 (M + C10H10N2O, 17), 146 (M + C15H17FN3O2, 14), 90 (C17H18FN4O3, 11), 68 (C4H4O, 26). (R)-tBu-3-(6-fluoroquinoxalin-2-yloxy)-pyrrolidine-1-carboxylate (29): Compound 28 (492 mg, 3 mmol, 1.5 equiv), PPh3 (786 mg, 3 mmol, 1.5 equiv), DEAD (1.5 mL, 1.5 equiv, 40 % in toluene) was added to a solution of (S)-tBu-3-hydroxypyrrolidine-1-carboxylate (374 mg, 2 mmol) in THF (15 mL). The reaction mixture was stirred overnight at room temperature. After evaporating the solvent, the residue was re-dissolved in CH2Cl2 (55 mL), washed in 1 n NaOH (10 mL) and 5 % citric acid (10 mL), and dried over Na2SO4. After evaporating the solvent under reduced pressure, the crude product was purified by column chromatography (n-hexane/EtOAc, 3:1, v/v) to produce 29 as a waxy solid (yield: 640 mg, 1.92 mmol, 96 %): [a]D23 = + 74.248 (c 0.59, CHCl3); 1H NMR (300 MHz, CDCl3): d = 1.46 (s, 9 H), 2.24 (m, 2 H), 3.57 (m, 3 H), 3.75 (dd, 1 H, JH,H = 12.6 Hz, 4.7 Hz), 5.70 (m, 1 H), 7.44 (m, 1 H), 7.65 (m, 1 H), 7.79 (m, 1 H), 8.45 ppm (s, 1 H); 13C NMR (75 MHz, CDCl3) d = 28.7 (3 C), 44.2, 51.8, 79.9, 113.2 (d, JC,F = 10.1 Hz, 1 C), 120.0 (d, JC,F = 12.4 Hz, 1 C), 129.0 (d, JC,F = 4.7 Hz, 1 C), 137.2, 139.5 (d, JC,F = 6.0 Hz, 1 C), 140.8, 154.7, 156.3, 159.3, 162.6 ppm (d, JC,F = 278 Hz, 1 C); 19F NMR (282 MHz, [D6]DMSO): d = 113.9 ppm (m, 1 F); IR (solid): ~ n = 2983 (CH), 1573 (C=N), 1478 cm 1 (C=N); LRMS-ESI(+): m/z calcd C17H21FN3O3 [M + H] + : 334.1, detected: 334.1. (R)-4-(3-((6-Fluoroquinoxaline-2-yl)-oxy)-pyrrolidine-1-yl)-6,7-dimethoxyquinazoline (30): A solution of 29 (667 mg, 2.00 mmol) in CH2Cl2 (10 mL) and TFA (0.78 mL, 10 mmol) was stirred at 0 8C. The solution was stirred overnight and allowed to temper at room temperature. After evaporating the solvent and the excess of TFA under reduced pressure, the oily brown residue was dissolved in dioxane/water (5:1 v/v), 4-chloro-6,7-dimethoxyquinazoline (500 mg 2.20 mmol, 1.1 equiv) and K2CO3 (1.40 g, 10.1 mmol, 5 equiv) was added. The reaction mixture was stirred for 14 h at 80 8C. After removing the solvent under reduced pressure, the residue was re-dissolved in water and CH2Cl2. The organic layer was washed in NaOH (aq, 1 m) and brine and dried over Na2SO4. After evaporating the solvent, the crude product was purified by column chromatography (n-hexane/iPrOH mixtures), and recrystallized from MeOH to produce 30 as a colorless solid (yield: 440 mg, 1.00 mmol, 52 %): mp: 190–191 8C (MeOH); [a]D23 = + 154,58 (c 1,02, CHCl3); 1H NMR (300 MHz, CDCl3): d = 2.48 (m, 1 H). 2.50 (m, 1 H), 3.99 (s, 3 H), 4.00 (s, 3 H), 4.15 (m, 2 H), 4.25–4.35 (m, 1 H), 4.42 (m, 1 H), 5.92 (m, 1 H), 7.26 (s, 1 H), 7.34 (s, 1 H), 7.44–7.50 (m, 1 H, 7CH-quinoxaline), 7.66 (dd, JH,H = 9.0 Hz, JH,F = 2,8 Hz, 1 H), 7.82 (dd, JH,H = 9.0 Hz, JH,F = 5.6 Hz, 1 H), 8.48 (s, 1 H), 8.50 ppm (s, 1 H); 13 C NMR (75 MHz, CDCl3): d = 31.8, 49.2, 56.2, 56.4, 74.6, 104.6, 106.5, 109.8, 113.3 (d, JC,F = 21.8 Hz, 1 C), 120.2 (d, JC,F = 25.1 Hz, 1 C), 129.0 (d, JC,F = 9.3 Hz, 1 C), 137.1, 139.6 (d, JC,F = 12.1 Hz, 1 C), 140.6, 148.1, 150.6, 152.1, 154.6, 155.1, 159.3, 161.1 ppm (d, JC,F = 246 Hz, 1 C); 19F NMR (282 MHz, [D6]DMSO): d = 113.5 ppm (m, 1 F); IR (solid): ~ n = 2935 (CH), 1574 (C=N), 1510 cm 1 (C=N); HRMS-ESI(+): m/z calcd C22H19FN5O3 + [M + H] + : 422.16229, detected: 422.16187. In vitro inhibition potency PDE testing procedure: We used human PDE10A1 (AB 020 593) expressed in SF21 cells using a baculovirus vector. The reaction mixture of 100 mL consisted Tris·HCl/ 5 mm MgCl2 buffer (pH 7.4), 0.1 mm [3H]cAMP and a defined enzyme quantity. The reaction was triggered by addition of substrate solution and was allowed to continue at 37 8C for 30 min. The enzymatic reaction was then terminated by adding 10 mL of YSi SPA beads. The reaction mixture was measured after 60 min using a scintillation counter (Microbeta Trilux). The Hill plot two-parameter model was used to determine IC50 values. At least four measurements were taken to determine the IC50 values (SEM   2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemmedchem.org 15 %).[35] PDE3 and PDE4 were also tested with recombinant enzymes respectively; PDE6 was isolated from bovine retina. The enzymatic activity of these isoforms was tested in the presence of 0.1 mm cAMP, 0.5 mm cAMP, and 2 mm cGMP, respectively. For a detailed list of all examined IC50 values, see Supporting Information. In vitro metabolism: S9 fraction: BD 452593 male Fischer rat liver S9; buffer solution: 80 mmol Tris·HCl pH 7.4, 20 mmol MgCl2 ; cosubstrate: 20 mmol NADPH/H*4 well + (16.6 mg) in 1 mL water (1 mL corresponds to 20 nmol NADPH/H); master solution: 5.5 mmol of substance in 1 mL DMSO (1 mL corresponds to 5 nmol of substance); 898 mL buffer solution, 50 mL co-substrate, and 2 mL master solution were added to a reaction tube and kept at a temperature of 37 8C for ~ 5 min. The solution was then added to the enzyme solution and the tube was inverted twice; 200 mL (t0) was extracted and 20 mL CH3CN was added to this fraction, with the product stored on ice while the remaining solution was left in the water bath. Samples (200 mL each) were taken every 7.5, 15, and 60 min; 80 mL CH3CN was then added to each sample, and they were centrifuged at 13 000 rpm for 10 min. The t0 sample was treated in the same way. The supernatant was transferred to an Eppendorf tube containing 50 mL of a suspension corresponding to ~ 10 mg of C18 silica gel in water/CH3CN (95:5, v/v). After 30 min agitation at 500– 700 rpm, the mixture was centrifuged at 8 000 rpm for 3 min, and the aqueous supernatant was removed by pipet. CH3CN (200 mL) was now added to the C18 phase, agitated for 30 min at 500– 700 rpm, and then centrifuged at 13,000 rpm for 10 min. Supernatant (150 mL) was then carefully extracted without disturbing the solid phase and transferred to a sample vial. The samples were examined using HPLC–MS. Acknowledgements Financial support from the European Regional Development Fund (ERDF) is gratefully acknowledged. Keywords: 3D QSAR · drug design phosphodiesterase 10 A · quinazolines · fluorine · [1] E. Bollen, J. Prickaerts, IUBMB Life 2012, 64, 965 – 970. [2] G. W. Reierson, S. Guo, C. Mastronardi, J. Licinio, M. L. Wong, Curr. Neuropharmacol. 2011, 9, 715 – 727. [3] S. H. Soderling, S. J. Bayuga, J. A. Beavo, Proc. Natl. Acad. Sci. USA 1999, 96, 7071 – 7076. [4] K. Loughney, P. B. Snyder, L. Uher, G. J. Rosman, K. Ferguson, V. A. Florio, Gene 1999, 234, 109 – 117. [5] K. Fujishige, J. Kotera, H. Michibata, K. Yuasa, S.-i. Takebayashi, K. Okumura, K. Omori, J. Biol. Chem. 1999, 274, 18438 – 18445. [6] N. J. Bray, F. M. Leweke, S. Kapur, A. Meyer-Lindenberg, Curr. Opin. Neurobiol. 2010, 20, 810 – 815. [7] J. Kehler, J. Nielsen, Curr. Pharm. Des. 2011, 17, 137 – 150. [8] C. J. Schmidt, D. S. Chapin, J. Cianfrogna, M. L. Corman, M. Hajos, J. F. Harms, W. E. Hoffman, L. A. Lebel, S. A. McCarthy, F. R. Nelson, C. ProulxLaFrance, M. J. Majchrzak, A. D. Ramirez, K. Schmidt, P. A. Seymour, J. A. Siuciak, F. D. Tingley, R. D. Williams, P. R. Verhoest, F. S. Menniti, J. Pharmacol. Exp. Ther. 2008, 325, 681 – 690. [9] R. J. Kleiman, L. H. Kimmel, S. E. Bove, T. A. Lanz, J. F. Harms, A. Romegialli, K. S. Miller, A. Willis, S. des Etage, M. Kuhn, C. J. Schmidt, J. Pharmacol. Exp. Ther. 2011, 336, 64 – 76. [10] T. F. Seeger, B. Bartlett, T. M. Coskran, J. S. Culp, L. C. James, D. L. Krull, J. Lanfear, A. M. Ryan, C. J. Schmidt, C. A. Strick, A. H. Varghese, R. D. Williams, P. G. Wylie, F. S. Menniti, Brain Res. 2003, 985, 113 – 126. [11] V. Lakics, E. H. Karran, F. G. Boess, Neuropharmacology 2010, 59, 367 – 374. [12] K. Fujishige, J. Kotera, K. Omori, Eur. J. Biochem. 1999, 266, 1118 – 1127. ChemMedChem 0000, 00, 1 – 13 &11& These are not the final page numbers! ÞÞ CHEMMEDCHEM FULL PAPERS [13] J. A. Siuciak, D. S. Chapin, J. F. Harms, L. A. Lebel, S. A. McCarthy, L. Chambers, A. Shrikhande, S. Wong, F. S. Menniti, C. J. Schmidt, Neuropharmacology 2006, 51, 386 – 396. [14] J. Kehler, J. P. Kilburn, Expert Opin. Ther. Pat. 2009, 19, 1715 – 1725. [15] J. Kehler, Expert Opin. Ther. Pat. 2013, 23, 31 – 45. [16] Z. Tu, J. Xu, L. A. Jones, S. Li, R. H. Mach, Nucl. Med. Biol. 2010, 37, 509 – 516. [17] S. Celen, M. Koole, M. De Angelis, I. Sannen, S. K. Chitneni, J. Alcazar, S. Dedeurwaerdere, D. Moechars, M. Schmidt, A. Verbruggen, X. Langlois, K. Van Laere, J. I. Andrs, G. Bormans, J. Nucl. Med. 2010, 51, 1584 – 1591. [18] Z. Tu, J. Fan, S. Li, L. A. Jones, J. Cui, P. K. Padakanti, J. Xu, D. Zeng, K. I. Shoghi, J. S. Perlmutter, R. H. Mach, Bioorg. Med. Chem. 2011, 19, 1666 – 1673. [19] S. Celen, M. Koole, M. Ooms, M. De Angelis, I. Sannen, J. Cornelis, J. Alcazar, M. Schmidt, A. Verbruggen, X. Langlois, K. Van Laere, J. I. Andrs, G. Bormans, NeuroImage 2013, 82, 13 – 22. [20] C. Plisson, C. Salinas, D. Weinzimmer, D. Labaree, S. F. Lin, Y. S. Ding, S. Jakobsen, P. W. Smith, K. Eiji, R. E. Carson, R. N. Gunn, E. A. Rabiner, Nucl. Med. Biol. 2011, 38, 875 – 884. [21] U. Funke, W. Deuther-Conrad, G. Schwan, A. Maisonial, M. Scheunemann, S. Fischer, A. Hiller, D. Briel, P. Brust, Pharmaceuticals 2012, 5, 169 – 188. [22] Z. Tu, S. Li, J. Xu, L. Jones, R. Mach, J. Nucl. Med. 2009, 50, 618. [23] T. A. Chappie, J. M. Humphrey, M. P. Allen, K. G. Estep, C. B. Fox, L. A. Lebel, S. Liras, E. S. Marr, F. S. Menniti, J. Pandit, C. J. Schmidt, M. Tu, R. D. Williams, F. V. Yang, J. Med. Chem. 2007, 50, 182 – 185. [24] A. Pandey, D. L. Volkots, J. M. Seroogy, J. W. Rose, J. Yu, J. L. Lambing, A. Hutchaleelaha, S. L. Hollenbach, K. Abe, N. A. Giese, R. M. Scarborough, J. Med. Chem. 2002, 45, 3772 – 3793. [25] K. Matsuno, J. Ushiki, T. Seishi, M. Ichimura, N. A. Giese, J.-C. Yu, S. Takahashi, S. Oda, Y. Nomoto, J. Med. Chem. 2003, 46, 4910 – 4925.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.chemmedchem.org [26] E. A. Arnott, J. Crosby, M. C. Evans, J. G. Ford, M. F. Jones, K. W. Leslie, I. M. McFarlane, G. J. Sependa, WO/2008/053221, 2008. [27] D. W. Kim, H.-J. Jeong, S. T. Lim, M.-H. Sohn, Angew. Chem. Int. Ed. 2008, 47, 8404 – 8406; Angew. Chem. 2008, 120, 8532 – 8534. [28] a) S. T. Hazeldine, L. Polin, J. Kushner, J. Paluch, K. White, M. Edelstein, E. Palomino, T. H. Corbett, J. P. Horwitz, J. Med. Chem. 2001, 44, 1758 – 1776; b) W. C. Lumma, R. D. Hartman, W. S. Saari, E. L. Engelhardt, V. J. Lotti, C. A. Stone, J. Med. Chem. 1981, 24, 93 – 101. [29] K. L. Bhat, D. M. Flanagan, M. M. Joulli, Synth. Commun. 1985, 15, 587 – 598. [30] S. Piras, M. Loriga, G. Paglietti, Farmaco 2004, 59, 185 – 194. [31] C. J. Helal, Z. Kang, X. Hou, J. Pandit, T. A. Chappie, J. M. Humphrey, E. S. Marr, K. F. Fennell, L. K. Chenard, C. Fox, C. J. Schmidt, R. D. Williams, D. S. Chapin, J. Siuciak, L. Lebel, F. Menniti, J. Cianfrogna, K. R. Fonseca, F. R. Nelson, R. O’Connor, M. MacDougall, L. McDowell, S. Liras, J. Med. Chem. 2011, 54, 4536 – 4547. [32] F. M. Uckun, J. Thoen, H. Chen, E. Sudbeck, C. Mao, R. Malaviya, X.-P. Liu, C.-L. Chen, Drug Metab. Dispos. 2002, 30, 74 – 85. [33] a) M. Greenwood, R. Robinson, J. Chem. Soc. 1932, 1370 – 1376; b) C. Wber, A. Bielik, . Demeter, I. Borza, G. I. Szendrei, G. M. Keserű, I. Greiner, Tetrahedron 2005, 61, 9375 – 9380. [34] K. Aoki, T. Obata, Y. Yamazaki, Y. Mori, H. Hirokawa, J.-i. Koseki, T. Hattori, K. Niitsu, S. Takeda, M. Aburada, K.-i. Miyamoto, Chem. Pharm. Bull. 2007, 55, 255 – 267. [35] N. Hçfgen, H. Stange, R. Schindler, H.-J. Lankau, C. Grunwald, B. Langen, U. Egerland, P. Tremmel, M. N. Pangalos, K. L. Marquis, T. Hage, B. L. Harrison, M. S. Malamas, N. J. Brandon, T. Kronbach, J. Med. Chem. 2010, 53, 4399 – 4411. Received: December 12, 2013 Revised: February 25, 2014 Published online on && &&, 0000 ChemMedChem 0000, 00, 1 – 13 &12& These are not the final page numbers! ÞÞ FULL PAPERS Fluor your health: Phosphodiesterase 10 A (PDE10A) has emerged as an attractive target for the development of 18 F-labelled brain imaging agents for positron emission tomography. A series of fluorinated dialkoxybiaryl compounds were synthesized and evaluated as PDE10A inhibitors, assisted by QSAR docking studies. The 7-fluoromethoxy derivative appears to be a promising candidate for further development.  2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim G. Schwan, G. Barbar Asskar, N. Hçfgen, L. Kubicova, U. Funke, U. Egerland, M. Zahn, K. Nieber, M. Scheunemann, N. Sträter, P. Brust, D. Briel* && – && Fluorine-Containing 6,7Dialkoxybiaryl-Based Inhibitors for Phosphodiesterase 10 A: Synthesis and in vitro Evaluation of Inhibitory Potency, Selectivity, and Metabolism ChemMedChem 0000, 00, 1 – 13 &13& These are not the final page numbers! ÞÞ