Roseophilin-inspired derivatives as transmembrane anion carriers

ABSTRACT Roseophilin is an alkaloid structurally related to prodiginines . The intriguing pharmacological properties of these derivatives have prompted us to prepare synthetic compounds 1–3 inspired by their structure and to explore their transmembrane anion transport activity. The methoxyfuran heterocycle impacts the anionophoric activity of the compounds as a result of the reduced hydrogen-bonding ability and electrostatic repulsions between the oxygen in the furan ring and the anions. The position of the furan was also found to be crucial for determining their anion transport activity. Overall, replacement of the characteristic methoxypyrrole moiety of prodiginines and tambjamines by the methoxyfuran found in roseophilin is detrimental to their ability as anion carriers, suggesting that the biological activity of roseophilin is likely not related to their potential activity as anion carriers. Compound 2, bearing a furan ring attached to a dipyrromethene moiety, was found to be the most active anion carrier. GRAPHICAL ABSTRACT


Introduction
In recent years, the development of small molecules capable of promoting the transmembrane transport of anions, anionophores, has attracted significant attention [1][2][3]. Using core supramolecular concepts, it is possible to apply rational design to produce this type of molecules [4,5]. Biological activities of these compounds and the prospects of developing future drugs based on this mechanism of action have fuelled the interest in this area [6][7][8]. The alteration of cellular homoeostasis could induce cellular toxicity and stress resulting in anticancer or antimicrobial activity [9][10][11][12][13]. Replacing the activity of faulty natural transport proteins could also be an interesting therapeutic approach to conditions such as cystic fibrosis, related to this problem [14][15][16]. In addition to purely synthetic anionophores, natural products represent an important inspiration for the design of these compounds. In particular, prodiginine and tambjamine alkaloids represent examples of biologically active anionophores ( Figure 1) [17,18]. These compounds are characterised by a methoxybipyrrole moiety and their intriguing pharmacological properties are linked to their activity as anionophores [19]. Roseophilin represents another group of natural alkaloids related to the prodiginine family (Figure 1) [20]. This product was first isolated from a culture broth of Streptomices. Its structure combines a strained macrocycle motif joined to an extended conjugated heterocycle core, responsible for the intense red colour of the compound. Although this product shares with its relatives the azafulvene structure, the methoxypyrrole ring of the prodiginines' backbone is replaced by a methoxyfuran moiety in roseophilin skeleton. The main biological application of roseophilin is related to its cytotoxicity against some cancer cell lines (human erythroid leukaemia and human epidermoid carcinoma) [21,22]. Its mechanism of action is still unknown, although it seems to be not associated to DNA damage under oxidative conditions [23]. Despite the efforts to find an effective total synthesis of this natural product and the different projects to study its biological activity, there are no previous reports on the synthesis of natural product-mimetics of roseophilin and their transmembrane anion transport properties have never been tested either. In this work, we decided to prepare derivatives inspired by the structure of roseophilin and to explore their anionophoric properties.

General experimental details
All reactions involving air-sensitive compounds were carried out under a nitrogen atmosphere. Oxygen was evacuated and purged with nitrogen, and deoxygenation processes were carried out by bubbling with a continuous nitrogen flow. Starting materials and solvents were obtained from commercial suppliers and used without further purification. TLC analysis was performed on aluminium-backed plates coated with silica gel 60 with F254 indicator or plastic TLC plates coated with aluminium oxide 60 neutral with F254 indicator; the plates were visualised under 254 or 366 nm light. Flash column chromatography was carried out in silica gel 60, 230 − 350 mesh ASTM; or aluminium oxide 90 activated, neutral, −60 mesh powder, Brockman grade I, 58 Å. NMR spectra were recorded on a Varian Mercury 300 MHz, Varian Unity Inova 400 MHz or Bruker Avance III HD 300 MHz. Chemical shifts for 1 H and 13 C NMR are reported in parts per million (ppm), using the residual solvent peak as reference, and 13 C NMR spectra were recorded using broad band proton decoupling. 1 H NMR coupling constants are reported in Hz and splitting pattern abbreviations are: s = singlet, t = triplet, dd = doublet of doublets, br = broad, m = multiplet. Multiplicities in 13 C NMR (CH 3 , CH 2 , CH and quaternary carbons as Cq) were determined by DEPT 135 experiments. High-resolution mass spectra (HRMS) were recorded in an Agilent 6545 Q-TOF spectrometer using +ESI and the results are reported as m/z. Melting points were measured using Gallenkampt Melting Point Apparatus combined with a calibrated thermometer (0-250 ºC, 1 ºC). For the measurements, not sealed capillaries were used and the values are not corrected.

Synthesis of compounds
Compounds 7 and 9 were prepared according to previously reported synthetic procedures and showed identical spectroscopic properties to those reported therein [24][25][26].
2,4-dimethyl-pyrrole (0.11 mL, 1.00 mmol, 2.0 equiv.) was added to a deoxygenated solution of aldehyde 4 (0.10 g, 0.50 mmol, 1.0 equiv.) in methanol (5 mL) under nitrogen. Then, HCl in methanol (0.5 M, 1.97 mL, 1.0 mmol, 2.0 equiv.) was added dropwise to the mixture. The reaction was stirred under a nitrogen atmosphere at room temperature for 5 hours. After this time, the crude was concentrated to dryness. Then, the residue was purified by precipitation in a mixture of dichloromethane and diethyl ether. The resulting dark-green precipitate was filtered off in a plate to give the prodiginine-like product 1 (0.11 g, 70%).

Synthesis of Z
In a Schlenk flask, a solution of aldehyde 7 (0.30 g, 1.57 mmol, 1.0 equiv.) in degassed methanol (25 mL) was deoxygenated with several vacuum/nitrogen cycles. Then, 2,4-dimethylpyrrole (0.32 mL, 3.14 mmol, 2.0 equiv.) was added to the system with a syringe through a septum and, subsequently, HCl in methanol (0.5 M, 6.30 mL, 3.14 mmol, 2.0 equiv.) was added to the mixture dropwise. The resulting orange solution was stirred under nitrogen at room temperature for 16 hours. After this time, methanol was removed by rotary evaporation and the crude was recrystallised in a mixture of dichloromethane and hexane (1:1, v/v). The resulting crystalline purple solid was filtrated in vacuo and washed with hexane to afford the prodiginine-like product 2 as a dark purple solid (0.34 g, 70% In a round-bottom flask, cyclohexylamine (178 μL, 1.60 mmol, 2.0 equiv.) was added to a solution of aldehyde 4 (0.15 g, 0.80 mmol, 1.0 equiv.) in chloroform (15 mL). Then, acetic acid (100 µL) was added and the reaction mixture was stirred at 65°C until the starting material was consumed. Once the reaction had finished, the solvent was evaporated under reduced pressure and the crude was washed with 1 M HCl. The organic fractions were combined and dried over anhydrous Na 2 SO 4 . After evaporating to dryness, tambjamine 3 was isolated as a black solid (0.20 g, 82%).
Intermediate 5 (1.25 g, 4.20 mmol, 1.0 equiv.) was dissolved in tetrahydrofuran (6 mL) and a suspension of lithium hydroxide (1.03 g, 42.0 mmol, 10.0 equiv.) in methanol (6 mL) was added dropwise under nitrogen. The mixture was stirred at room temperature for one hour. After this time, the solvent was removed and the resulting solid was re-dissolved in chloroform and this solution washed with water. The organic phase was dried over anhydrous Na 2 SO 4 and concentrated to dryness. The residue was purified by column chromatography in silica gel using a mixture of hexane and ethyl acetate (1:1, v/v) as eluent to afford 4 as a dark green solid (0.65 g, 80%). 1 H NMR (300 MHz, CDCl 3 ) δ (ppm) 9.44 (1H, s), 6.93 (1H, br s), 6.69 (1H, br s), 6.33 (1H, s), 6.

Transmembrane anion transport experiments
1-Palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) (Sigma-Aldrich) stock solutions (20 mg/ mL = 26.32 mM) were prepared in chloroform and kept in the freezer. To prepare POPC vesicles, 3 mL of the POPC stock solution were added to a roundbottom flask. The solvent was evaporated using a rotary evaporator (25°C), and lipids were dried overnight under high vacuum. On the next day, the lipid film was rehydrated with 1 mL of the NaCl internal solution corresponding to each assay. Then, the mixture was shaken in a vortex and subjected to seven freeze-thaw cycles (freezing by introducing the flask in a Dewar with liquid nitrogen, and then melting by introducing the flask into warm water). After this process, the lipid suspension was extruded twentynine times through a 200 nm polycarbonate nucleopore membrane using a LiposoFast Basic extruder (Avestin, Inc.). In order to remove the non-encapsulated chloride, the lipid suspension was introduced into a dialysis membrane, and the packet was dialysed against the required external solution (2 × 500 mL) for around 1.5 hours (45 min with each dialysis solution of 500 mL). Finally, the dialysed suspension of vesicles was placed into a 10-mL volumetric flask and made up with the corresponding external solution to obtain the vesicle stock suspension used in the assays (7.89 mM for 3 mL of POPC).
Chloride efflux experiments were performed using an Ion-Selective Electrode (ISE) (HACH console MM34O and a Chloride Selective Electrode 96 52 C HACH) [27]. To 5 mL of 0.5 mM POPC suspended in the corresponding external solution, an aliquot of the corresponding anionophore in DMSO was added to the experiment (the volume of DMSO was always less than 20 μL). The chloride efflux from the vesicles was monitored over time using a chlorideselective electrode. For NO 3 -/Clexchange experiments, the intravesicular solution contained 489 mM NaCl, 5 mM NaH 2 PO 4 , pH 7.2 (I. S. 500 mM), whereas the extravesicular solution was composed of 489 mM NaNO 3 , 5 mM NaH 2 PO 4 , pH 7.2 (I. S. 500 mM). In these assays, at t = 0 s the carrier was added to the experiment and its activity was recorded with a chloride-selective electrode. At the end of the experiment (t = 300 s), an aliquot of detergent was added to lyse the vesicles and release all the encapsulated chloride; this value was subsequently used to normalise the data. For 150 mM Na 2 SO 4 , 20 mM NaH 2 PO 4 , pH 7.2) was added, followed by the compound (t = 0 s). Finally, at t = 300 s an aliquot of a detergent was added to the experiment to lyse the vesicles and release all the entrapped chloride.

Synthesis of compounds
Prodiginines are characterised by a dipyrrolyldipyrromethene structure, whereas in roseophilin the azafulvene moiety is replaced by a methoxyfuran ring (Figure 1). Aiming to evaluate the impact of this major structural change in the anion transport properties of this compound prompted us to design compound 1. For comparison purposes, the furanylpyrrolo prodiginine 2 was also targeted. The synthesis of these compounds is presented in Scheme 1. The pyrrolyl-furan-2-carbaldehyde 4 precursor has been previously described by Challis [24]. We used a slightly modified procedure to prepare furan-2-carbaldehyde 6 using chloroform as solvent, in the presence of an excess of phosphorous oxybromide. A Suzuki-Miyaura coupling reaction with N-tert-butoxycarbonyl-2-pyrroleboronic acid, using Pd(PPh 3 ) 4 as catalyst and Na 2 CO 3 as base, yielded compound 5.
Suitable crystals for single-crystal X-ray diffraction analysis of compounds 1 and 2 as their hydrochloride salts were obtained, and the solid-state structures of these derivatives are shown in Figure 2.
These solid-state structures confirm that both compounds form hydrogen-bonding 1:1 complexes with the chloride anion. These derivatives show the expected planarity of the tris-heterocyclic system, reflecting their high conjugation. The tris-heterocycle skeleton of 1 displays a conformation where all heteroatoms from heterocycles are oriented towards the anion [28]. On the other hand, prodiginine 2 exhibits a conformation where the furan ring is flipped away to minimise the electrostatic repulsions of the furan oxygen atom with the anion, thus favouring the coordination of the latter. The NH groups of the pyrrole heterocycles interact, in both cases, with the chloride anion through hydrogen bonds. Moreover, in prodiginine 2 an additional hydrogen bond is formed between the CH group of the furan ring and the chloride anion.
Hydrogen bonds are quite directional, and distances and angles can be used as a measurement of their strength, being values close to 180° associated to stronger hydrogen bonds. In the case of 1, the values of the NH···Cl angles (167.9° and 153.6°) and the length of these non-covalent bonds (3.09 and 3.23 Å) are typical of moderately strong interactions [29,30]. The distance between the chloride anion and the oxygen atom of the furan ring is 3.60 Å. The high aromatic conjugation of 1 is also reflected on the C(sp 2 )-C(sp 2 ) distances. Thus, the length of the C-C bond connecting the pyrrole and furan is shorter than the typical values of the corresponding single bonds because it has a partial double bond character. NOESY NMR spectra in CDCl 3 also supported this disposition as the preferred conformation for 1 in solution (see Figure S22). For compound 2, the distances of both NH···Cl hydrogen bonds are similar (3.20 and 3.10 Å). Furthermore, the CH···Cl interaction displays a longer hydrogen-bonding length (3.51 Å), which is in agreement with the lower strength of this interaction. Regarding the angles, the NH···Cl hydrogen bonds display values of 174.6° and 166.6°, while that of the CH···Cl hydrogen bond is 151.9°. In compound 2 the C-C bonds around the methine group connecting the two pyrrole systems show similar distances (around 1.37 Å) and an angle of 134.7°. As well as in 1, the C(sp 2 )-C(sp 2 ) distances in compound 2 are intermediate between the typical values of single and double bonds, which highlights the high aromatic conjugation of the system. All crystallographic data of 2 evidence that its structure is similar to those of the click-prodiginines reported recently by us [31].
The availability of compound 4 prompted us to explore the preparation of compounds resembling the structure of tambjamines (Scheme 2). Tambjamines are remarkably stable Schiff bases (imines) synthesised by condensation of an amine and a carbonyl compound under mild acidic conditions. Thus, reaction of 4 with cyclohexylamine using acetic acid as catalyst allowed to obtain compound 3 as its hydrochloride salt after washing the crude with a hydrochloric acid solution. It should be noted that when aromatic amines were employed no reaction was observed, precluding the synthesis of aromatic-substituted imines. This result could be explained by the lower nucleophilicity of aromatic amines compared to aliphatic amines and the low electrophilicity of the carbonyl carbon of this derivative.

Anion transport experiments
Anion transport assays were carried out in 1-palmitoyl-2oleoyl-sn-glycero-3-phosphocholine (POPC) vesicles using a chloride-selective electrode. Full details on vesicles preparation and performance of the assays are described in the experimental section. The NO 3 -/Clexchange and the HCO 3 -/Clexchange assays were explored. Aliquots of DMSO solutions of the studied compounds were added to a suspension of the chloride-containing vesicles. Chloride efflux was monitored with a chloride-selective electrode and a detergent was added at the end of the experiments to lyse the vesicles and release all the encapsulated chloride, being this value used as 100% of chloride efflux. Performing the assays at various concentrations allows Hill analysis of the chloride efflux observed at 300 seconds and the calculation of the EC 50 and Hill parameter n. EC 50 represents the concentration needed to elicit 50% chloride release under the conditions explored. The obtained results are presented in Table 1. A representative example is shown in Figure 3.
Compound 2 was found to be the most active of these assays, displaying a submicromolar EC 50 value in the NO 3 -/Clexchange assay. This result is in agreement with the better stabilisation of the anion suggested by the solid-state structure of this compound (Figure 2), showing the involvement of the furan C-H in the hydrogen-bond cleft displayed by this derivative. Incorporation of the furan heterocycle as the B ring in roseophilininspired derivatives 1 and 3 is clearly deleterious for the transmembrane transport activity of these derivatives. When comparing the calculated EC 50 for 1 and 2 a 25fold increase is observed. A release of less than 50% chloride was observed for 3 even when explored at 1% mol carrier to lipid concentration under these conditions (see Figure S30). It should be noted that the parent prodiginine and tambjamine analogues of 1 and 3 are very active transmembrane transporters [32,33]. The marked drop in chloride efflux observed when the lipophilic nitrate is replaced by the more hydrophilic bicarbonate (and sulphate) is a result commonly observed in these experiments. It can be interpreted as the result of the differences in the ease of the anion being extracted into the membrane. These results also support an exchange mechanism as the main anion transport mechanism and rule out detergent effects exerted by these compounds.

Conclusions
Inspired by the structure of the natural product roseophilin, compounds 1-3 were prepared and characterised. These derivatives include a furan ring in their structure and are structurally related to other alkaloids, such as prodiginines and tambjamines. The anionophoric activity of these compounds was tested in model phospholipid vesicles using ISE-based assays. The results demonstrated that the substitution of a pyrrole ring by a furan heterocycle in these natural product mimetics is detrimental to the anionophoric activity of the compounds. The elimination of a hydrogen-bonding interaction with the anion, together with electrostatic repulsions between the oxygen of the furan ring of 1 and 3 and chloride, could explain these results. Moreover, a comparison between the anionophoric properties of prodiginines 1 and 2 showed that the position of the furan in these derivative compounds was crucial for the anion transport activity, being compound 2 in which the furan ring is attached to the dipyrromethene moiety, the most active derivative. These results suggest that, differently from prodiginines and tambjamines, the biological activity of roseophilin is likely not related to its potential activity as anion carrier.