Protecting-group-free synthesis of the bisindolylmaleimide GF109203X

The bisindolylmaleimide GF109203X is a highly selective inhibitor to Protein Kinase C (PKC) and has attracted much attention. However, its reported synthesis required protecting groups. In order to achieve a short and N-protecting-group-free synthesis of GF109203X, an investigation on N -alkylation of arcyriarubin A was carried out using different bases, solvents and equivalents of bromododecane. It is found that mono-N-alkylation and mono-N'-alkylation could be achieved respectively. Thus, a protecting-group-free synthesis of GF109203X was accomplished in 40% overall yield starting from indole. This work will lead to a short synthesis of mono - N' - alkylated arcyriarubins to accelerate drug discovery.


Introduction
Since the isolation of staurosporine (1)  1 in 1977, many indolocarbazoles including rebecamycin (2) 2 have been isolated from bacteria, fungi, and invertebrates. 3,4This family of compounds have attracted much attention because of their wide range of biological activities including inhibition of protein kinase C (PKC). 5 More importantly, a fruitful strategy has been developed to design more selective inhibitors by disrupting the planarity of the indolocarbazole ring system.Among them, bisindolylmaleimides (BIMs) 6,7 derived from arcyriarubin A (3) 8,9 are successful examples, which are of interest because of their PKC inhibitory activity.Mono-N'-substituted BIMs such as compounds 4-6 (Figure 1: for simplicity, the nitrogen in the maleimide ring is referred to as N, the other two nitrogens in indole rings are marked as N' and N'' respectively) represent a class of excellent PKC inhibitors.In particular, GF109203X ( 6) is a highly selective inhibitor to PKCs (IC 50 = 20 nM，17 nM, 16 nM and 20 nM for PKCα，βI，βII and γ respectively), 11 and it has been widely used in biological studies such as regulation of human platelets, 12 treating Alzheimer's disease, 13 prostate cancer 14 and so on.As a result, GF109203X has attracted attention from the synthetic community.In 1991, the chemical synthesis of GF109203X (6) was first reported by Toullec et al. 15 in 19% yield starting from indole, using an N-methyl protecting group on the maleimide NH.The bisindolylmaleimide core was built up via addition of indole-derived Grignard reagent to dibromide 10, 16,17 and the side-chain was introduced by mono-N'-alkylation of 9 (Scheme 1).Later on, Faul et al. 18 developed a more efficient synthesis of GF109203X in 51% yield starting from indole, using an O-trityl (Tr) protecting group (Scheme 1).The binsindolylmaleimide core was made via the coupling 19 between 12 and 13, and the side-chain was introduced by Nalkylation of indole-acetamide 14 at an early stage.Although protecting groups play an important role in organic synthesis, 20 their inclusion in a synthetic route increases the total number of steps by needing protection and deprotection and decreases atom-economy. 21][24] Considering mono-N'-substituted BIMs such as 4-6 are important lead compounds for PKC inhibitions and further chemical biology studies, we believe it makes sense to develop a general synthesis [25][26][27][28][29] of them all, based on arcyriarubin A with a protecting-group-free strategy, which requires the direct mono-N'-alkylation of arcyriarubin A. In this paper, we report the mono-N'alkylation of arcyriarubin A and its application to the protecting-group-free synthesis of GF109203X.

Results and Discussion
We commenced our study by making the starting material arcyriarubin A according to Faul's procedure 18,30 (Scheme 2).Indole reacted with oxalic chloride followed by quenching the reaction with methanol at 0 o C to give methyl indole-3-glyoxylate 12 in 69% yield, which was previously synthesized by using NaOMe in MeOH at -60 o C. 18 This was reacted with indole-3-acetamide 14 in the presence of t-BuOK, the intermediates was converted into arcyriarubin A in 96% yield by acidic aqueous work-up.Investigation of the reaction of arcyriarubin A (3) with n-C 12 H 25 Br was carried out to see if any mono-N'-substituted product can be obtained.Firstly, we examined the N-alkylation of arcyriarubin A 31,32 in acetone using K 2 CO 3 as base (Table 1, entry 1).The red color of arcyriarubin A in acetone turned deeper when it was treated with K 2 CO 3 , which was reacted with 1.2 eq. of n-C 12 H 25 Br for seven hours to give mono-N-alkylation product 17 in an isolated yield of 18%.When the reaction was run at 55 o C for 16 hours using three equivalents of K 2 CO 3 and three equivlanets of bromide, the mono-N-alkylation product (17) was isolated in 73% yield, accompanied by the mono-N'-alkylated product (16) and N,N'-dialkylated product (18) in yields of 7% and 15% respectively.The 1 H-NMR spectrum of 16 recorded in CDCl 3 showed singlets at 8.65 and 7.86 ppm, which were assigned to the NH of indole and maleimide respectively.Two sets of indole aromatic protons (10 H) could be seen in the range of 8.00-7.00ppm, which confirmed the asymmetric structure of 16.Similarly, singlets at 8.56 ppm (2H) and 7.53 ppm (1H) were seen in 17 and 18, and one set of aromatic protons for 17, while two sets for 18.The IR spectra of all the three compounds show weak peak at around 3300 cm -1 and a strong absorption at around 1700 cm -1 , which are characteristic of NH and C=O.Finally, all three compounds gave satisfactory HR-MS (EI) results (see supplementary materials).a Yield based on the recovered starting material (BORSM) of 3.
The mono-N-alkylation of arcyriarubin A other than mono-N'-alkylation resulted as the major product under the aforementioned conditions because the NH of the maleimide unit is more acidic than the NH of indole.Thus we speculated that the use of stronger base would make a difference to the N-alkylation reaction of arcyriarubin A. Indeed, when the reaction was carried out at ambient temperature in THF for 48 hours using LiHMDS, t-BuOK or NaH, the color of the reaction mixture turned from red to purple upon the addition of the base to arcyriaribin A. Also, the purple color changed back to red gradually when the reaction mixture was treated with two equivalents of bromododecane.In all cases, the starting material 3 was not fully converted, the mono-N'-alkylated product 16 was isolated as major product, the di-N',N''-alkylated product 19 as minor product, and the mono-N-alkylated products were not isolated (See Table 1, entries 3-5).The isolated yields of mono-N'-alkylated product 16 were 30%, 30% and 40% respectively in the case of LiHMDS, t-BuOK and NaH respectively, and 50%, 60% and 70% based on recovered starting material (BORSM).Besides 16, the di-N',N''-alkylated product 19 was isolated in a yield of 21%, 10% and 15% correspondingly, and 35%, 20% and 27% BORSM.We also tried the reaction at 55 o C for four hours using six equivalents of bromododecane, and full conversion of starting material 3 was observed (See Table 1, entry 6-8), the mono-N'-alkylated product 16 was isolated in a yield of 30% and 26% for LiHMDS and t-BuOK, and the di-N',N''alkylated product 19 a yield of 60% and 62%, respectively.In the case of NaH as base, the di-N',N''-alkylated product 19 was obtained in 91% isolated yield, and the mono-N'-alkylated product 16 was not observed.The characterization of 19 was carried out by 1 H-NMR. 13C-NMR, HR-MS, and IR measurements (see supplementary materials).Thus, it is crucial to run the reaction at ambient temperature for a long period using NaH in order to obtain mono-N'alkylated product.
A plausible mechanism for mono-N-substitution and mono-N'-substitution is shown in Scheme 3. The pK a of maleimide NH is around 10, 33 hence it can be deprotonated easily in the presence of K 2 CO 3 to give the corresponding anion 20, which reacted with alkyl bromide to give N-substituted product 21.However, the pK a of an indole NH is about 21, 33 so for deprotonation it requires bases with a pK a higher than 21.When arcyriarubin A was treated with strong bases such as NaH, all three NHs were completely deprotonated to afford the trianion 22 presumably (Scheme 3), 34,35  Encouraged by the above results, we turned our attention to the synthesis of GF109203X itself (Scheme 2).Some factors were considered in order to obtain optimized conditions: (i) the electrophile 15 is in the form of its HCl salt, which will consume base as well, so that more equivalents of NaH are required; and (ii) lowering the amounts of electrophile 15 will decrease the proportion of di-N',N''-alkylated product.Under the optimized conditions, arcyriarubin A (3)  was dissolved in THF and treated with six equivalents of NaH, followed by addition of 1.3 equivalents of bromide 15 in an ice bath.The reaction was continued at ambient temperature for 48 h, standard work-up was followed by flash column chromatography and thereafter crystallization from acetone to give GF109203X (6) as a red solid, in an isolated yield of 61%.The starting materials arcyriarubin A was recovered in 23% yield, i.e. 84% yield BORSM.It is worth mentioning that the purification of 6 from starting material 3 by chromatography was quite tricky owing to the presence of the polar dimethylamino group.Fortunately, after careful screening of developing solvents for thin-layer chromatography, we found a special combination of triethylamine and acetone (around 1/12 by volume) to be very helpful for the purification of 6 by flash column chromatography.The peak at 412.1906 [M + ] in high-resolution mass spectrum revealed the existence of the desired product 6 (reqired 412.1899).The IR data of 6 showed the presence of NH (3396 cm -1 ) and imide carbonyl (1748, 1703 cm -1 ).
1 H-NMR spectrum of 6 recorded in DMSO-d 6 (Figure 2) confirmed that 6 had been obtained: two broad singlets at 11.68 ppm (1H) and 10.90 ppm (1H) correspond to NH and N''H, 10 protons in the region of 7.78-5.53result from two different kinds of indole rings, and the triplets at 4.24 ppm (2H) and 1.82 ppm (2H) as well as a broad singlet at 2.03 ppm (8H) are from the N,N-dimethyl-3-aminopropanyl side-chain (Figure 2).The 1 H-NMR data of our synthetic sample matched well with the ones reported by Toullec 15 and Faul 18 (see supplementary materials).

Conclusions
The mono-N-substitution and mono-N'-substitution of arcyriarubin A (3) with bromododecane was accomplished by using K 2 CO 3 in acetone and NaH in THF.A plausible mechanism for mono-N-substitution versus mono-N'-substitution of arcyriarubin A was proposed.Based on this result, a protecting-group-free synthesis of GF109203X was achieved in three steps and 40% yield from indole.This work will lead to a general synthesis of mono-N'-alkylated BIMs, and accelerate chemical biological study of this class of important molecules.

Experimental Section
General.All reagents were used as received from commercial sources without further purification or prepared as described in the literature.Tetrahydrofuran was distilled from sodium and benzophenone immediately before use.Reactions were stirred using Teflon-coated magnetic stir bars.Analytical TLC was performed with 0.20 mm silica gel 60F plates.Chromatographic purification of products was carried out by flash chromatography on silica gel (230-400 mesh).Melting points were determined using an electrothermal melting point apparatus or DSC.Infrared spectra were recorded on a Nicolet 8700 Fourier transform spectrometer.NMR spectra were measured in CDCl 3 (with TMS as internal standard) or DMSO-d 6 on a Bruker AV400 or Varian INOVA-400M ( 1 H at 400 MHz. 13 C at 100 MHz) magnetic resonance spectrometer.Chemical shifts (δ) are reported in ppm, and coupling constants (J) are in Hz.The following abbreviations are used to designate the multiplicities: s = singlet, d = doublet, t = triplet, q = quartet, m = multiplet.High-resolution EI mass spectra (HR-EI-MS) were recorded on an GCT CA127 Micromass UK mass spectrometer.

Scheme 2 .
Scheme 2. Two-step synthesis of arcyriarubin A from indole.

Scheme 3 .
Scheme 3. Proposed mechanism for the N-substitution versus N'-substitution of arcyriarubin A.

Scheme 4 .
Scheme 4. The synthesis of GF109203X via N'-substitution of arcyriarubin A.