A gram-scale synthesis of a macrocyclic amidinourea with strong antifungal activity through a Fukuyama tri-protected polyamine intermediate

Systemic fungal infections are, nowadays, of crucial importance and, thus, in the last decade, we explored the great potential of natural and synthetic guanylated compounds, a great amount of work that led to the development of new non-azole antifungal compounds bearing a macrocycle, endowed with potent antifungal activity. We planned many biological assays to evaluate this class, implying always greater amount of compounds needed. This triggered us to setup a convenient strategy to prepare, in an easy and affordable way, grams of compound to be tested in excellent overall yield.


Introduction
][6][7] The highly polar nature of these compounds makes their handling difficult, time-consuming and sometimes expensive.Moreover, the selection of the proper protective groups for the nitrogen atoms is usually a pivotal moment in synthesis planning, above all if natural macromolecules preparation is involved.][15][16][17][18][19][20][21][22][23] In the last years, our research group focused its efforts on the preparation of a series of compounds bearing interesting antifungal properties, as we reported in previous works, [24][25][26][27][28][29][30][31][32] well represented by compound 1 (Figure 1).These compounds showed good activities against various Candida strains, most notably albicans, guillermondii, kefyr, glabrata.Their main point of strength, though, is their ability to improve the antifungal activity against strains presenting resistance to azoles, the most widely used antifungal drugs at the moment, as well as a good ADMET profile. 32Being these the most desirable characteristics novel antifungals should have, we decided to investigate them even further, so we faced the need to prepare amounts of compound 1 higher than what we usually managed.In the first described synthesis of 1, 27 the commercially available bis(8aminooctyl)amine was used.Two subsequent guanylations of the primary amino groups, by using N,N'-di-Boc-S-methylisothiourea and N,N'-di-Boc-N-[(E)-but-2-en]-S-methylisothiourea respectively, followed by a cyclization in refluxing THF in presence of TEA and a final Boc deprotection, led to the desired compound.However, since bis(8-aminooctyl)amine was no longer commercially available (with the exception of custom synthesis services), new synthetic paths had to be taken.Here follows a detailed analysis of these paths, with their obstacles and downsides, that finally led us to setup a convenient and versatile way to prepare our lead compound 1 in gram scale.

Results and Discussion
Since the withdrawal from the market of the bis(8-aminooctyl)amine, two different solutions were explored (Figure 2).The first option was the synthesis of bis(8-aminooctyl)amine itself in four steps starting from 1,8-dibromooctane. 33In a second approach, the macrocyclic scaffold was obtained by reducing with DIBAL-H the amide obtained by coupling the mono-guanilated-1,8-diaminoctane and the Cbz-protected 8-aminooctanoic acid. 25oth the described approaches, though, are not suitable for obtaining grams of 1.The first approach, besides using the unsafe sodium azide, implies a quite heavy workload to obtain the bis(8-aminooctyl)amine.This is mainly ascribed to the purification of the 1-azido-8-bromooctane, which implies long isocratic hexane chromatography given the high similarity in terms of polarity the product and the starting material present.The difficulties that are faced in case one of the free amines needs to be purified is another aspect that involves overload of work.Besides, it is then used in a 3:1 ratio for the selective mono-alkylation, leading to a waste of unrecoverable starting material (alongside the wasting of time and expensive work).The second approach gets over this problem by anticipating the mono-guanylation in the synthetic process.DIBAL-H, though, is the only reducing agent that can be used for the amide reduction due to the presence of other carbonyl groups, which led to moderate to low yields, difficult purifications (including work-up with Rochelle salt leading to a viscous mass which is hard to handle), lack of a good reproducibility (since DIBAL-H quality tends to decay over time) and still not suitable for a scaling-up procedure.Over time, other alternatives were explored, but they failed to solve the abovementioned problems.To overcome the described limitations, we took inspiration from Fukuyama's work designing a novel synthetic strategy.This allowed us to prepare grams of 1 in a single batch, in the same timeframe of the old approaches.The keystone is to obtain intermediate bis (8-aminooctyl)amine in the tri-protected form 6, with trityl, Cbz and nosyl as protecting groups of choice.This intermediate offers many advantages, such as lowering of both time and costs of production (because protecting groups make it less polar and easier to handle and purify) and versatility, since the three different groups are orthogonal to each other, meaning they can be selectively removed in different conditions, allowing us to deprotect only the chosen amine; this avoids resorting to workarounds like the use of 3:1 ratio and waste of materials by means of simple one-pot-twosteps reactions.
The three subsequent reactions have the advantage of being one-pot-two-steps.Specifically, the first one (the deprotection of trityl group and guanylation of the primary salified amine) was carried out with TFA and Et 3 SiH in DCM and subsequent treatment with N,N'-Di-Boc-1H-pyrazole-1-carboxamidine and DIPEA to afford compound 7.In the deprotection step, the excess of Et 3 SiH was removed in vacuo as the remaining TFA and DCM, leaving the salified amine and the inert triphenylmethane ready for the guanylation step.7 is easily purified by flash chromatography given the lack of free amine groups, with an excellent overall yield of 99%.The following reaction, the nosyl deprotection, was carried out using thiophenol and K 2 CO 3 in dry DMF.The excess of thiophenol and K 2 CO 3 were eliminated during the work-up after washing with water and 5% NaHCO 3 , while the byproduct 1-nitro-4-(phenylsulfanyl)benzene is left in place, since it does not interfere with the next step.At this point, the solvent was removed and the crude product was heated at reflux with TEA in THF giving compound 8 (52% yield), again avoiding any possible problem during the purification of the intermediate.For this step, we exploited the tendency of di-Boc-protected guanidine to give amidinoureas when heated at reflux in THF in presence of an amine. 35,36This ability results from the conversion of one of the Boc protecting group of the guanidine into isocyanate through a t-butoxide elimination and subsequent nucleophilic attack by the amine.In our case, the reaction occurs through an intramolecular attack, which leads to the formation of a macrocycle.

Conclusions
The synthesis of compound 1 always presented drawbacks, ranging from a very low overall yield, challenging purifications and handling of highly polar compounds, wasting of starting materials due to low molar ratios, and use of hazardous chemicals.During the long time we have worked on these molecules, we always managed to prepare the lab-scale quantity required for the early development without too much hassle, but the more advanced steps of drug discovery and development required higher amount of compounds which were really hard to prepare following the old habits, despite the familiarity we had with them.Spending time and efforts in the setup of an alternative synthetic strategy appeared, then, a good investment for the future of the project.And indeed it was, since a gram-scale synthesis of 1 was successfully performed, following a convenient, versatile and efficient route, with a great overall yield, a low cost and adequate amount of time, despite the process still involved the use and preparation of polyamines.
Compound 1 obtained via this process proved to possess the same physico-chemical properties and pharmacological activities of our previous batches, making this synthetic method a viable and valuable route to overcome all the limitations and difficulties our previous strategies implied, easing the challenging purpose of finding always novel and more active antifungal compounds.To this aim, the synthesis goes through the obtainment of the key intermediate 6, whose synthetic flexibility could be very important in the planning of new derivatives.This process allowed us to perform advanced biological investigations on 1, e.g. in vivo assays, with very little requirements in terms of costs, manpower and time, pushing our understanding of its biological behavior even further, and allowing us to focus on the study of its mode of action.

Experimental Section
General.All reagents were used as purchased.CHCl 3 , CH 3 CN, DCM, TEA and DIPEA were dried by distillation from CaH 2 ; THF from Na/Benzophenon.Chromatographic purifications were carried out with flash columns packaged with silica gel 60, 230-400 mesh and reactions were monitored using TLC purchased from Merck with silica gel 60 F 254 . 1 H-NMR and 13 C-NMR data were acquired on Bruker Avance DPX 400.As internal standard the residual signal of deuterated solvents was used and the chemical shift are reported in ppm.Splitting patterns are described as s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet, dq = doublet of quartet, bs = broad signal.Mass spectra data were obtained using Agilent 1100 LC/MSD VL system (G1946C).

Figure 2 .
Figure 2. Summary of the synthetic strategies explored.