Identification of the first enantiopure Rac1–Tiam1 protein–protein interaction inhibitor and its optimized synthesis via phosphine free remote group directed hydroarylation

The remote substituent regiocontrol of phosphine free Heck hydroarylation has been exploited for the preparation of the first enantiopure inhibitor of Rac1–Tiam1 PPI.

Abstract

A phospine free hydroarylation reaction applied to norbornene derivatives is described for the first time and was exploited for the regioselective gram scale synthesis of AR-148, a known Rac1–Tiam1 PPI inhibitor. Umpolung conversion of the nitro group into free amine allowed the regiocontrol of the key arylation step via a long range effect. The effect of AR-148 in comparison with its enantiomers on Rac1 activation of has been evaluated and (–)AR-148 has been identified as the first enantiomerically pure inhibitor of Rac1–Tiam1 PPI.

Introduction

The small GTPase protein Rac1 has gained attention for its role in different pathologies, among which are cancer and cardiovascular diseases.1 Rac1 regulates events such as smooth muscle cell (SMC) migration2 and proliferation,3 and leukocyte-endothelial cell interaction.4 Rac1 activity depends on the equilibrium between the inactive GDP-bound and the active GTP-bound forms. This cycling is regulated by guanine nucleotide exchange factors (GEFs) which act as activators, and GTPase activating proteins (GAPs) and GDP dissociation inhibitors (GDIs) which act as negative regulators. The T-cell lymphoma invasion and metastasis 1 (Tiam1) protein, a specific GEF for Rac1, is crucial for cell–cell adhesion and cell migration. Small molecules interfering with the Rac1–Tiam1 protein–protein interaction (PPI) were identified and reported by us5,6 and by others7–11 (Fig. 1a). Afterwards we reported the identification of 2-amino-3-(phenylsulfanyl)norbornane-2-carboxylate as a privileged scaffold for the de novo design and synthesis of a structurally original family of Rac1 inhibitors12 (Fig. 1b). A G-LISA assay on SMCs demonstrated that among all the compounds tested, 1 (named AR-148) selectively and potently inhibits Rac1 without interfering with RhoA. A Boyden chamber chemotaxis assay and cell movement video microscopy analyses have shown how 1 (AR-148) affects cell migration in response to the chemotactic agent platelet derived growth factor BB (PDGF-BB).12

Fig. 1. Previous work: a) first generation of Rac–Tiam1 inhibitors; b) target compound AR-148 obtained in racemic form; c) AR-148 synthesis drawbacks: -compound 4, precursor of AR-148, is the minor reaction product, –2* used as the racemate of the two enantiomers.

In our previous work we designed and performed a divergent synthetic strategy from which a small library of aryl 2-amino-3-(phenylsulfanyl)norbornane-2-carboxylate was obtained.

The main difference between AR-148 and the other known Rac1–Tiam1 PPI inhibitors (Fig. 1a and b) consists in the presence of a three dimensional core heavily functionalized which position the aromatic groups involved in π–π interactions or π–cation interactions in very precise way (Fig. 2).

Fig. 2. AR-148 molecular interaction surface of Rac1-AR-148.12.

In modern drug discovery, compounds with three dimensional structures that escape the “flatland” of multiple arenes, become more central because of higher levels of insaturation and the presence of multiple stereocenters. The resulting structural complexity seems to provide better receptor ligand interaction with consequent improved potency, selectivity and decreased off-target effects.13

Focusing our attention on the efficient preparation of the most active compound AR-148, we identified two main drawbacks.

Firstly, the hydroarylation reaction was performed in a classical way in the presence of Ph₃P and an excess of iodoarene (3 equiv.) which gave the C-6 arylated isomer, precursor of AR-148, as the minor isomer (3 : 4 = 70 : 30)14 (Fig. 1c). Secondly, compound AR-148 was obtained in racemic form. Nowadays, defining the activity of both enantiomers is extremely important because of the great impact on reducing unexpected toxicity in further drug development. Moreover, once the eutomer is defined, its precise pharmacophore model could deeply improve the rational design of new derivatives. As a consequence, the preparation of the two enantiomers in pure form is crucial for evaluating their activity independently. In this work, we report an efficient protecting group free, gram scale selective synthesis of compound AR-148, the obtainment, via chromatographic separation, of both enantiomers and their pharmacological evaluation.

Results and discussion

Chemistry

To optimize the synthesis of AR-148 in an eco-friendly way, we reasoned that a phosphine-free hydroarylation protocol could be of outstanding interest. Many phosphine ligands are expensive, toxic, unrecoverable and often water- and air-sensitive. In large-scale applications on industrial and semi-industrial scales, the phosphines might be a serious economic burden. Therefore we considered performing the reaction in the absence of phosphine.15 To our knowledge, phosphine-free Heck hydroarylation has never been realized on olefins and the only examples reported are limited to alkyne substrates.16,17 So, we planned to perform the hydroarylation using the same conditions described before, but in the absence of the phosphine ligand (Pd(OAc)₂/TEA/HCOOH in CH₃CN). The progressive addition to racemate 2 of 3-iodoaniline (1 equiv. × 3 every 4 hours) at a lower temperature (50 °C instead of 82 °C) gave the same mixture of regioisomers 3 and 4 in higher yield (80% instead of 75%) but with the same ratio (Scheme 1).

Scheme 1. Phosphine free methodology. Reagents and conditions: (a) Pd(OAc)₂ 5%, TEA 3.5 equiv., HCOOH 3.0 equiv., CH₃CN, 50 °C, (80%) *used as racemate of the 2 enantiomers.

The theoretical work done by M. B. Hall on the mechanism of ligand-free Heck type arylation18–20 and the works published by Cacchi on phosphine-free Heck hydroarylation on alkyne,16,17 shed light on the different possible scenarios regarding the mechanism of our reaction. Unfortunately, the presence of a slight excess of an organic base like TEA, the use of a coordinative solvent such as CH₃CN, the requirement of formic acid and the substrate itself, containing the norbornene core (ligand in the Catellani reaction21), make it very difficult to prove which catalytic cycle is really operating or if there are more than one operating at the same time. It is worthy to report that different solvents have been screened in the model reaction between simple norbornene SI5 and iodobenzene SI6 (DMF, DMSO, CH₃CN, DCM, MeOH, THF, DCE, Et₂O) and have always led, even if with variable yield, to the formation of the desired product SI7 (Fig. S3†). The good conversion obtained in DCE underlined how the reaction works, even in the absence of a coordinative solvent, is in accordance with the formation of the dianionic species isolated by Hartwig and co-workers for the ligand-free Heck arylation.19 Once a fairly satisfactory result was obtained which demonstrated the feasibility of a phosphine free hydroarylation protocol, we turned our attention to the selectivity of the Heck hydroarylation which is in fact the divergent point in the synthesis of 4, a precursor of the desired compound AR-148. In order to modulate the regioselectivity in favour of 4 instead of 3, several attempts were performed by using different reaction conditions but without satisfactory results22–24 (Table S1†).

Our previous studies14,25 disclosed that the regioselectivity is mainly driven by a long range effect exerted by the nitro group on C-2, (Path a, Fig. 3).26,27 We thus decided to exploit the umpolung conversion of the nitro- to the amine group, reasoning that an opposite long range effect should induce an inverse charge distribution of the double bond and consequently an opposite regiochemistry could be obtained (Path b, Fig. 3).

Fig. 3. Umpolung conversion of nitro- to amino group to modulate the long range effect on the hydroarylation reaction, (*data for regioselectivity in NBoc protected norbornene have been previously reported12,14).

By treating compound 2 with Zn/H₃PO₄ in THF from 0 °C to r.t., the amine precursor 5 has been prepared in very good yield (97%) (Scheme 2).

Scheme 2. Reagents and conditions: a) Zn, H₃PO₄ 1 M, THF, 0 °C-rt, 2 gram scale; b) phosphine free methodology m-iodoaniline 1.1 equiv., Pd(OAc)₂,TEA, HCOOH, CH₃CN, 45 °C, 1 gram scale; c) 4-iodonitrobenzene 1.1 equiv., Pd(tetrakis), K₂PO₄, CH₃CN, 70 °C, 1 gram scale; d) Zn, HCl 1 M, MeOH, 0 °C, 1 h, 1 gram scale; *used as the racemate of the 2 enantiomers.

After a brief optimization (Table S2†) to identify the adequate reaction conditions, compounds 7 and 8 have been obtained (yield 88%, 82% for the grams scale) by reacting 5 under the phosphine-free protocol previously optimized (Pd(OAc)₂/TEA/HCOOH in CH₃CN).

As expected thanks to an opposite long range effect, an inverse C5/C6 ratio was found to belong to the regioisomer 8, precursor of AR-148, the main product of the reaction (7 : 8 = 30 : 70).

Coordination of palladium by the free amine at the endo position of 5 cannot been excluded but is a non-productive pathway for the Heck hydroarylation reaction, in fact no endo-product has been detected on the norbornene core28–30 (see ESI† page S7 for NMR characterization and assignment of exo-substituted compounds 7 and 8).

The use of the phosphine free protocol has been fundamental to obtain the desired compound as a single reaction product and to improve the atom economy of the process. Indeed, we hypothesized that the undesired Buchwald–Hartwig reaction of m-iodoaniline on itself or on the norbornene aliphatic amine of 5 does not work in absence of phosphine while the Heck arylation is not affected. Polymers of the m-iodoaniline and polyarylated products were in fact always detected by mass analysis of the crude using phosphine catalysed reaction conditions (Table S2†). Such undesired derivatives are completely suppressed using the phosphine free protocol and the amount of m-iodoaniline 6 is reduced from 3.0 to 1.1 equivalents.

Buchwald–Hartwig amination has been tested directly on the inseparable mixture of 7 and 8. Using Pd(tetrakis), K₃PO₄ and 1.4 equiv. 4-iodonitrobenzene in MeCN, compounds 9 and 10 were obtained in an almost quantitative yield (90%). K₃PO₄ is required instead of Cs₂CO₃ to avoid the arylation of the aminoester functionality on the norbornene core. The compounds 9 and 10 have been separated by flash chromatography and their structures were confirmed by NMR analysis (structural elucidation, ESI†). By performing the reduction of the nitro group of 10 with Zn/HCl 1 M in MeOH at 0 °C, compound (±)AR-148 was obtained in a very efficient way.

The whole protocol was scaled up and 1 gram of compound (±)AR-148 was obtained without the use of any protective group, with a higher atom economy, higher overall yield and in a regioselective way avoiding the use of phosphine in the Heck reaction (Scheme 2).

Finally, we faced the problem related to the availability of the two enantiomers of AR-148 for the pharmacological tests. We focused on the enantioselective synthesis of norbornene scaffold 2 obtained through the Diels–Alder cycloaddition reaction.12 We tried to perform the reaction by using chiral catalysts, but although we observed enhanced selectivity, very low yields always affected our results. The synthesis of chiral dienophiles (β-sulfanyl nitroacrylate esters of (–)menthol and (–)8-Ph-menthol) was very troublesome and the yields were not suitable as a first step of the synthesis. Therefore, we considered the exploitation of a removable chiral auxiliary to obtain the separation of the two corresponding diastereoisomers of AR-148. By reacting the free amine of AR-148 with (–)menthylchloroformate the two corresponding monomenthylcarbamate diastereoisomers were obtained and easily separated by HPLC, but unfortunately further hydrolysis of the chiral auxiliary always led to unreacted compounds or tarry materials.

Finally, through the use of chiral HPLC, we separated the two enantiomers of AR-148 with unsatisfactory results. However, excellent results were obtained in the chiral HPLC resolution of racemate (±)-10 (see the ESI†), (Scheme 2). The two enantiomers (–)-10 and (+)-10 were obtained with high enantiomeric purity (purity: >99% ee: >99).

After reduction of the single enantiomers (Zn/HCl in MeOH at 0 °C), the corresponding amines (–)AR-148 and (+)AR-148 were obtained. NMR analyses were performed confirming the structure and the purity of the final compounds.

Pharmacology

The effect of AR-148 and its enantiomers on Rac1 activation was investigated in human cultured smooth muscle cells (SMCs). Interestingly, the (–)AR-148 enantiomer reduced the intracellular levels of Rac1-GTP by 63.9 ± 15.0% at 10 μM, while (+)AR-148 was inactive. Accordingly, the racemate 10 μM AR-148 only partially inhibited the Rac1 activity (–17.2 ± 4.1%). These data clearly demonstrated an enantioselective effect of (–)AR-148 on Rac1 (Fig. 4, Experimental ESI†).

Fig. 4. Dose-dependent effect of AR148 and its enantiomers on Rac1-GTP levels. SMCs were seeded at a density of 2 × 10⁵ per 35 mm Petri dish and incubated with DMEM supplemented with 10% FCS. After 24 h, the compounds were added to the cultured medium at final concentrations of 1, 2.5, 5 and 10 μM, and after 4 h Rac1 GTP levels were evaluated by G-LISA assays.

Conclusions

In summary, we reported a gram scale, protecting group-free synthesis of compound AR-148, a known potent inhibitor of the Rac1–Tiam1 protein–protein interaction. The synthetic strategy applied includes the first example of a phosphine free Heck hydroarylation reaction on an olefin, which represents an improvement also in terms of eco-friendliness and atom economy of the process. We were able to exploit the remote directing group effect of the free amine for the selective obtainment of the desired regioisomer of the arylation reaction. Despite the impossibility of an enantiopure synthesis, we obtained the two enantiomers of AR-148, via HPLC separation of the precursors, and we demonstrated for the first time the enantioselectivity in the inhibition of Rac1–Tiam1 PPI by (–)-AR-148. Characterization of the absolute stereochemistry of (–)AR-148 and study of cocrystalization with the Rac1 protein are proceeding in order to define the pharmacophoric model for further drug design.

Conflicts of interest

The authors confirm that this article has no conflict of interest.