Anthracene Bisureas as Powerful and Accessible Anion Carriers

Abstract Synthetic anion carriers (anionophores) have potential as biomedical research tools and as treatments for conditions arising from defective natural transport systems (notably cystic fibrosis). Highly active anionophores that are readily accessible and easily deliverable are especially valuable. Previous work has resulted in steroid and trans‐decalin based anionophores with exceptional activity for chloride/nitrate exchange in vesicles, but poor accessibility and deliverability. This work shows that anthracene 1,8‐bisureas can fulfil all three criteria. In particular, a bis‐nitrophenyl derivative is prepared in two steps from commercial starting materials, yet shows comparable transport activity to the best currently known. Moreover, unlike earlier highly active systems, it does not need to be preincorporated in test vesicles but can be introduced subsequent to vesicle formation. This transporter also shows the ability to transfer between vesicles, and is therefore uniquely effective for anion transport at low transporter loadings. The results suggest that anthracene bisureas are promising candidates for application in biological research and medicine.

Ak ey feature of transporters 1 and 2 is the 1,5-diaxial arrangemento f( thio)urea binding units. Parallel bonds between scaffold and (thio)urea positiont he latter so that all four NH can bind to Cl À simultaneously. Restricted rotation about these CÀNb onds ensures that intramolecular hydrogen bonding cannoto ccur.C onsidering alternative structures that might be easier to synthesize, we realized that the 1,8-disubstituted anthracenes 7 ( Figure 3) bear ac lose geometric resemblance to the 1,5-diaxial systems. Compounds 7 have previously been shown to function as receptorsf or anions [29][30][31][32] andn eutral guests, [33] but anionophore activity has not been investigated. Here, we report that anthracene bisureas 7 (X = O) can serve as outstandingly effective anionophores, competitive with the best of the 1,5-diaxial systemsa nd superior in some respects. Given its accessibility,t his system could be considered the methodofchoice for inducing rapid chloride transport in bilayer membranes.
[j] Measuredf or the ethyl ester analogue of 2a.T he ester side-chaini s not expectedt oa ffect affinities. [15] Chem (25 mm)a nd the influx of chloride monitored by the decay in lucigeninfluorescence. [39] Transport rates were quantified by fitting the inverse of the normalized fluorescence traces (F 0 /F)t os ingle exponential functiont oo btain approximate half-lives (t1 = 2 )a nd ad ouble exponentialf unction to obtain initial rates (I). Dividing I by the transporter/lipid ratio and averaging for ar ange of experiments at different loadings gives the specific initial rate, [I]. As described in previous work, [15] [I]i s independento ft he transporter to lipid loading and thus allows the performance of anionophores with widely different activities to be compared directly.
Results from these experiments are summarized in Ta ble 1 and Figure 5. All the anthracene bisureas were found to mediate chloride transport, with activities measurable at transporter to lipid loadings as low as 1:250k (1.6 nm overall in the aqueous suspension) ( Figure 5b). Consistentw ith previous observations, [15,16,20,25,28,34,35] the electron-deficient aryl termini N, F,a nd F2 promoted faster transport than unsubstituted P.T he most powerfulv ariant was the nitrophenyl bisurea 7ON,for which [I] was measured as 2100 s À1 .Ad ose-responses tudy employing six different loadings of 7ON (Figure 6a)r evealed that this compound was significantly active even at the lowest transporter to lipid ratio of 1:1000k. At this loading most LUVs contain either 1o r0transporter molecules, thust he activity observed corresponds to 7ON acting as as ingle molecule. [15] Comparing 7ON to earlier systems, the anthracene does not quite match decalin 2b,t he current record-holder at [I] = 3800 s À1 .H owever,t he examination of fluorescenced ecay traces showst hat in some respects the new system is more effective. We have previously found that powerful transporters such as 2b and 1b produce rapid initial drops in emission at the low loadings, but that traces tend to plateau at relatively high levels.T his is illustrated in Figure 7f or transporter:lipid =    :250k. We believe this is due to the absence of transporter molecules from many vesicles (see above), which therefore act as bystanders.I nc ontrast, the traces for 7ON tend towards lower emissionv alues at all loadings (Figures 6a and 7). Thus, as the experiment progresses,t he overall effect of 7ON becomes greater than any of the earlier systems. The reason for the difference is thought to be the ability of 7ON to transfer between vesicles,u nlike 2b,w hich is trapped in its originall ocation. Experimentss upporting this hypothesis are described in the next section.
Considering their accessibility,t he effectiveness of the anthracene bisureas is remarkable. In contrastt od ecalin 2b, which requires an ine-step synthesis, [15] the anthracenes are availablei nj ust two steps. Moreover, it is notable that 7ON and 7OF2 are ureas,w hereas 2b possesses the more favorable thiourea units. Decalin bisurea 2a is considerably less active than the anthraceneb isureas, suggesting that the anthracene scaffold is more effective than the decalin. It is more difficult to compare the transporters with systemsf rom other groups, but the dose-responsed ata for 7ON ( Figure 6) allows the estimation of an EC 50, 270 s , [46] am easure which is widely used by other laboratories. The EC 50, 270 s value calculated for 7ON is 0.0003 mol %, the lowest reported to date for chloride-nitrate exchange, and even lower than that of the natural anionophore prodigiosin. [28] In contrast to the ureas,a nd counter to trendso bserved previously, [15] the bisthiourea 7SF2 provedr elativelyi neffective (Figure 5a). While its modest chloride affinity may be af actor, it also showedl imited stability under the conditions of the transport experiment. [39] The anthracene bisthiourea design was therefore not pursued further.

Mechanistic studies
Although the anthracene bisureas wered esigned as anion carriers, it is also possible that they could act through formation of self-assembled channels. The linear relationship between initial transportr ates and transporter:lipid ratios for 7ON (Figure 6b)p rovides one line of evidencef or the carrier mechanism;i fm ore than one transporter molecule is required to form the active complex, one would normally expect reduced effectiveness at lower concentrations. [47] To support this conclusion,t ransport was also studied in vesicles composed of 1,2-dipalmitoyl-sn-glycero-3-phosphocholine (DPPC),w hich undergoes at ransitionb etween gel and liquid phases at 41 8C. [48] The transition is expectedt oa ffect transport rates for mobile carriers,b ut much lesss of or channels. As anticipated, 7ON provedi nactive at 25 8C( gel phase), but active at 45 8C( liquid phase), consistent with the carrier mechanism. [39,49] Studies werea lso undertaken to test the ability of the transporters to move between vesicles,a si mplied for 7ON by the results discussed earlier." Delivery vesicles" containing 7ON but not lucigenin were mixed with "receiver vesicles" containing lucigeninb ut not 7ON,b efore addition of chloride. Fluorescence decay traces implied that the transporter was transferred rapidly to the receiver vesicles,e quilibrationo ccurring in 5min. [39] In contrast, the same experiments with the more lipophilic 7OF2 showed negligible transfer on thes ame timescale. Despite the ability of 7ON to exchange between vesicles, experimentsd esigned to detect leachingf rom the membranes gave negative results, implyingt hat the equilibrium concentration in water is very low. [39] Deliverability An important requirement for practical applications of anionophores is that they must be readily deliverable to target membranes. The most activec holapod and trans-decalin carriers do not fulfil this criterion well. Decalin 2b,i np articular,i sa lmost inactivew hen added to preformed LUVs and is only effective when incorporated in the vesicles as they are prepared. [15,45] To provide aquantitative estimate of deliverability,wehave developed av ariant of the lucigenin assay in which the vesicles are formed without transporter,a nd the latter is then added using as tandardized procedure, before the introduction of chloride. [17] The decay of F 0 /F is followed, and the initial rate I is measured. Deliverability (D)i sq uantified by dividing I for this experiment by that observed when the anionophore was preincorporated. Fluorescence decay traces for both types of experiment, applied to the four anthraceneb isureas, are shown in Figure 8. Values of D for the anthracene bisureas, as well as 1a, 1b,a nd 2b,a re listed in Ta ble 1. The results show that deliverability for 7OF and 7OF2 is only moderate, but that 7OP and 7ON are transferred quite efficientlyt ot he vesicles. In particular,t he deliverability of 7ON,a tD = 0.82, contrasts starkly with that of 2b (D = 0.03). The good deliverability of 7ON probablyr elatest oi ts moderate lipophilicity (c log P = 6.9, see Ta ble 1). We suspect that highly lipophilic agents such as 2b (c log P = 11.6) form intractable aggregates after addition to the aqueous phase,and these interact poorly with the membranes. Though 7ON presumably also aggregates, the individualm olecules are less lipophilica nd this could lead to improved availability.

Conclusion
In conclusion, we have shown that anthracene1 ,8-bisureas are exceptionally effective and practical anion transporters. The most powerful promote chloride/nitrate exchange at rates comparable to the highest previously observed, while being far more accessible than the earlier systems. Dinitrov ariant 7ON combinesh igh activity with good deliverability in a manner unmatchedb yp revious systems. Ta king into account its ability to transfer between vesicles,i ti sa rguably the most effective agent currently available for transporting chloride across vesicle membranes at low dosages. The anthracene scaffold has potential for further modificationt oc ontrol binding affinities, lipophilicities etc. We believe the design has promise for application in tools for biomedical research, and perhaps in the treatment of channelopathies such as cystic fibrosis.