Arginine vasopressin (AVP),a cyclic non-peptide hormone that is secreted mainly from the posterior pituitary gland,exerts multiple actions throughout the body by interacting with three Gprotein- coupled receptors,V1a,V1b and V2 ^[1]. The V2 receptors, which are localized in the renal collecting ducts,regulate water resorption and salt (NaCl) balance. Stimulation of V2 receptors by AVP reduces the urine excretion ^[2]. Thus,there is potential to develop a AVP V2 receptor antagonist for the treatment of disorders such as hyponatremia ^{[3, 4]},congestive heart failure ^{[5, 6]},renal disease ^[7],edema and syndrome of inappropriate antidiuretic hormone secretion (SIADH) ^[8].

A few AVP receptor antagonists have undergone sufficient clinical development to be on the market,such as conivaptan and tolvaptan for the treatment of hyponatremia in the USA. Lixivaptan, another promising AVP receptor antagonist,is still undergoing phase III clinical trials. All of them are based on the benzazepine scaffold,with conivaptan being a dual V1a/V2 receptor antagonist ^[9],whereas tolvaptan and lixivaptan being selective V2 receptor antagonists ^{[10, 11]}. Their structures were illustrated in Fig. 1.

Fig. 1

	Download: JPG larger image
Fig. 1. Chemical structures of conivaptan,tolvaptan and lixivaptan.

A number of benzazepine compounds as AVP receptor antagonists were reported ^{[12, 13, 14, 15, 16]}. Similar to conivaptan, tolvaptan and lixivaptan,most of their structures are comprised of a benzene-fused seven membered ring system and two aromatic rings linked through amide bonds. The benzazepine scaffold was probably metabolized by cytochrome P 450 (CYP) 3A4 and 2C19,in the similar way as diazepam,a benzodiazepine drug ^[17]. The AVP receptor antagonists with benzazepine scaffold might have drug- drug interactions with CYP 3A4 and/or CYP 2C19 inhibitors. It is reported that CYP3A4 is responsible for the metabolism of conivaptan ^[18] and lixivaptan ^[19]. Moreover,the FDA has included a black box warning that the coadministration with other drugs inhibiting the P450 system could result in an increase in conivaptan levels. Therefore,AVP receptor antagonists with other scaffolds should be introduced to avoid the drug-drug interactions. A number of compounds with non-benzazepine scaffolds were screened and the phenothiazine scaffold compounds were found to have excellent affinity to V2 receptors. Furthermore,our previous study indicated that the introduction of sulfonyl bond linkage may enhance the biological activity ^{[20, 21]}. Herein,the synthesis and biological activity of the potent AVP V2 receptor selective agonists based on the phenothiazine scaffold was presented here and the structure-activity relationship (SAR) was discussed. 2. Experimental 2.1. Synthesis and structure characterization

Two synthetic routes applied to prepare target compounds 1a-1y are illustrated in Scheme 1. Considered that 4 is a key intermediate for the synthesis of the target compounds and it is suitable for large scale preparation,procedure A is more efficient. Acylation of 2 with a p-nitrobenzoyl chloride,p-nitrobenzene sulfonyl chloride or m-nitrobenzene sulfonyl chloride provided 3, which was subsequently reduced using different methods to provide aniline 4 in high yields. However,when Pd/C was used to catalyze the reaction,the nitrobenzene could not be reduced completely due to catalyst poisoning probably caused by sulphur in the materials. Acylation of 4 with acyl chloride or substituted sulfonyl chloride provided target compounds 1a-1y illustrated in Table 1.

Scheme 1

	Download: JPG larger image
Scheme 1. Preparation of target compounds 1a-1y.

All the materials,reagents and solvents were obtained from commercial suppliers. Silica gel chromatography,if not mentioned, was conducted on Grace Reveleris Flash System X2 (Grace Davision Discovery Sciences,Columbia,MD,USA). HPLC data was obtained with an Agilent 1260 (Agilent Technologies,Inc.,Santa Clara,CA,USA) equipped with a Grace C18 column (5 μm, 250 mm × 4.6 mm,Lot No. 55/182).NMRspectra were recorded on a Bruker AV400 NMR (Bruker,Billerica,MA,USA) and HRMS were measured on a 7.0T FTMS System (IonSpec Corporation,Irvine,CA, USA). Melting points (uncorrected) were determined on YRT-3 Melting Point Tester (Precision Instrument of Tianjin University, Tianjin,China).

10-(4-Nitrobenzoyl)-10H-phenothiazine (3a). To a solution of 10H-phenothiazin (50.0 g,251 mmol) in CH₂Cl₂ (200 mL),Et3N (48.5 g,480 mmol) was added and the mixture was stirred at 0 °C. Then p-nitrobenzoyl chloride (59.8 g,322 mmol) dissolved in CH₂Cl₂ (150 mL) was added dropwise into the mixture and stirring was continued for another 2 h under room temperature. The reaction mixturewaswashedwith distilled water. Theorganic layer was dried over anhydrous magnesium sulfate and evaporated to give the crude product as a yellow solid,which was recrystallized from ethanol affording 3a (69.5 g) as white powder. Yield: 70%; purity: 98% (HPLC,methanol: water = 9:1,1 mL/min,the same below); mp: 224.7-225.9 °C; ¹H NMR (400 MHz,DMSO-d₆): d 7.25-7.30 (m,4H),7.49 (d,2H,J = 6.8 Hz),7.55 (d,2H, J = 8.8 Hz),7.60-7.62 (m,2H),8.14 (d,2H,J = 8.8 Hz); ESI-HRMS calcd. for C₁₉H₁₃N₂O₃S ([M+H]⁺) 349.0641,found 349.0643.

(4-Aminophenyl)(10H-phenothiazin-10-yl)ketone (4a). To a solution of 3a (41.8 g,120 mmol) in ethanol (250 mL),FeCl3 (1.0 g,60 mmol) and activated carbon (10.0 g) was added and stirred for 0.5 h. Hydrazine hydrate (80%,11.9 g,241 mmol) was added dropwise into the mixture. The mixture was heated to reflux and stirred for another 2 h. After filtration,most of the solvent was removed under reduced pressure and the remaining solution was cold to ambient temperature,and white crystals grew slowly. The intermediate was obtained by filtration as white solid (31.7 g, yield: 83%),mp: 181.9-183.0 °C. ¹H NMR (400 MHz,DMSO-d₆): d 5.62 (s,2H),6.33 (d,2H,J = 8.8 Hz),6.99 (dd,1H,J₁ = 1.8 Hz, J₂ = 7.0 Hz),7.21-7.25 (m,4H),7.45-7.47 (m,2H),7.52-7.54 (m, 2H). ¹³C NMR (400 MHz,DMSO-d₆): d 157.8,141.3,130.0,121.1, 120.9,117.6,117.0,116.7,116.2,110.6,102.3. ESI-HRMS calcd. for C₁₉H₁₅N₂OS ([M+H]⁺) 319.0900,found 319.0904.

N-(4-(10H-Phenothiazine-10-carbonyl)phenyl)butyramide (1a). To a solution of 4a (1.9 g,6 mmol) in CH₂Cl₂ (40 mL),Et3N (0.8 g,8 mmol) was added and the mixture was stirred at 0 °C. Then butyryl chloride (0.7 g,7 mmol) diluted in CH₂Cl₂ (10 mL) was added dropwise into the mixture and stirred for 2 h. The reaction mixture was washed with distilled water. The organic layer was dried over anhydrous magnesium sulfate and evaporated under reduced pressure. The resulting oil was purified by silica gel chromatography to give 4a (1.8 g,yield: 85%) as white powder. Purity: 98.6%; mp > 230 °C; ¹H NMR (400 MHz,DMSO-d₆): d 0.88 (t,3H,J = 7.4 Hz),1.57 (q,2H,J = 7.3 Hz),2.25 (t,2H,J = 7.4 Hz), 7.20-7.25 (m,6H),7.43-7.48 (m,4H),7.56-7.58 (m,2H),9.99 (s, 1H). ESI-HRMS calcd[7_TD$DIF]. for C₂₃H₂₁N₂O₂S ([M+H]⁺) 389.1318,found 389.1320.

Compounds 1b-1y were synthesized following the method of 1a. Their structures were characterized by HRMS,¹H NMR and ¹³C NMR partly (see [8_TD$DIF]Supporting information). 2.2. Biological activity evaluation

The human recombinant vasopressin V1a (Cat. ES-361-C) and V2 (ES-363-C) receptors in 1321N1 host cell were purchased from Perkin Elmer Inc. (Waltham,MA,USA). The Sprague-Dawley rats were obtained from Shanchuanhong Experimental Animals Co.,Ltd (Tianjin,China). The in vitro evaluation was done by the same method we reported previously ^{[20, 21]}. The in vitro radioligand binding assay was performed to determine the binding affinity of the candidates to human V2 and V1a receptors. We investigated some potent derivatives for in vivo diuretic activity in conscious hydrated male Sprague-Dawley rats at 8 weeks of age (body weight: 260 ± 20 g). Urine volume was measured during the 20 h after oral administration of the test compounds as well as the reference compound tolvaptan and lixivaptan. 3. Results and discussion

The structures of the target compounds 1a-1y and evaluation of the biological features were summarized in Table 1. The binding affinity was determined by the radioligand binding assay on V1a and V2 over-expressing cells. These compounds had specific affinity to human AVP receptors. Furthermore,they showed high selectivity to V2 receptors. Generally,as shown in Table 1,when R₂ was an alkyl group or alogen-substituted alkyl group with straight chain (1a-1e),the biological activities were obvious worse than that of substituted benzene group (1f-1y). When R₂ was the substituted benzene ring,the different substituted positions of halogen only slightly affected the binding affinity to V2 receptors of the compounds (e.g. 1i and 1j). In contrast,the different substituted positions ofmethyl (e.g. 1g and 1h) and nitro (e.g. 1m and 1n) group may change the binding affinity significantly.

Table 1

Table 1 Structures of potent AVP V2 receptor agonists based on the phenothiazine scaffold.

Table 1 Structures of potent AVP V2 receptor agonists based on the phenothiazine scaffold.

Moreover,CF₃-substituted phenothiazine compounds exhibited relatively higher binding affinity to V2 receptors than nonsubstituted compounds (e.g. 1h vs. 1w,1j vs. 1x,1n vs. 1y),which indicated that the electron-withdrawing groups attached to the phenothiazine scaffold may increase the affinity to V2 receptors. Strong electron-withdrawing substituent groups of R₂ also increased the binding affinity significantly (e.g. 1g and 1o). It is obvious that the sulfonyl bond linkage enhanced the binding affinity (e.g. 1c vs. 1e),and compounds with dual sulfonyl bond linkage exhibited higher V2 receptor affinity activity than that with only single sulfonyl bond linkage (e.g. 1o vs. 11v).

Several compounds presented encouraging binding affinity, with both remarkable binding affinity and selectivity for the V2 receptor. The compounds with satisfactory binding affinity were selected to conduct the in vivo diuretic assay,with tolvaptan and lixivaptan as the reference compounds. As shown in Table 1,it is evident that the selected compounds have exhibited significant diuretic activity,especially compounds 1n,1r,1t and 1v,as they strongly increased urine volume compared with the control group. Compounds 1r and 1v showed excellent diuretic activities which were equivalent to tolvaptan and lixivaptan. The diuretic activity and binding affinity of the compounds did not match well probably due to specific differences in the vasopressin receptors. Therefore it may be difficult to draw a direct comparison between the diuretic assay in rats and the binding assay in cells expressing the human receptors. 4. Conclusion

Twenty-five derivatives of phenothiazine designed as AVP V2 receptor antagonists were synthesized and characterized by 1H NMR,HRMS and HPLC. Their biological activity was evaluated by in vitro radioligand binding assay and in vivo diuretic assay. Several compounds exhibited both high affinity and promising selectivity for V2 receptors and selected for in vivo test. Meanwhile,the selected compounds showed promising diuretic results in rats, especially compounds 1r and 1v,which produced total urine volumes equivalent to tolvaptan and lixivaptan during the experimental period. Through the present studies,compounds 1r and 1v,which exhibited promising efficacy both in vitro and in vivo,could be novel AVP V2 receptor antagonist candidates. Further preclinical studies are however still required. Acknowledgment

This project was supported by National Major Scientific and Technological Special Project for ‘‘Significant New Drugs Development’’ (Nos. 2011ZX09401-009 and 2013ZX09102014).

Supplementary data associated with this article can be found,in the online version,at http://dx.doi.org/10.1016/j.cclet.2015.01.022.