A Concise Silylamine Approach to 2-Amino-3-hydroxy-indoles with Potent in vivo Antimalaria Activity

Graphical Abstract

The development of a concise strategy to access 2-amino-3-hydroxy-indoles, which are disclosed as novel antimalarials with potent in vivo activity, is reported. Starting from isatins the target compounds are synthesized in 2 steps and in good yields via oxoindole intermediates by employing tert-butyldimethylsilyl amine (TBDMSNH2) as previously unexplored ammonia equivalent.

Malaria is the most deadly parasitic infectious disease with an estimated 300–500 million cases and a death toll of 0.8–1.2 million in 2008 alone.¹Of the four Plasmodium species that are relevant for humans, P. falciparum accounts for most of the fatalities.¹These already grim statistics are likely to become even grimmer as P. falciparum strains that are resistant to commonly used antimalaria chemotherapeutic agents emerge and spread throughout Africa and parts of Asia.

The most recent antimalarial drug class was introduced in 1996² In an effort to address the urgent need for new drugs, especially ones with new cellular targets or chemo types that could delay the emergence of resistance, we have screened small-molecule libraries for novel, drug-like compounds with whole-cell antimalarial activity and limited susceptibility to established mechanisms of drug resistance.³ Very recently, two similar efforts have also been reported.^4,5 Out of ~ 79,000 compounds,⁶ 104 inhibitors with nanomolar activity against drug-sensitive (3D7) and multidrug-resistant (Dd2, HB3) P. falciparum strains were identified.³ One of the hits, the 2-amino-5-chloro-3-hydroxy-3-phenylindole 1a (see Figure 1) was particularly attractive because of its unusual and compact 2-amino-3-hydroxy-indole core structure. In addition, compound 1a achieved excellent exposure in mouse pharmacokinetic studies, and more importantly, when tested in a 4-day suppressive P. berghei mouse model 1a demonstrated very good in vivo efficacy,⁷ causing a dose dependent decrease in parasitemia with undetectable levels of parasites at the highest dose tested and no signs of adverse side effects (Figure 1).

Figure 1.

In vivo antimalaria activity of 2-aminoindole 1a. Following inoculation of Swiss Albino mice with P. berghei parasites, 1a was administered i.p. once daily for 4 days, and parasitemia was determined on day 5.

Encouraged by this promising in vivo activity we began to explore and optimize this compound class. A thorough literature survey revealed that 2-amino-3-hydroxy-indoles are virtually unexplored for biological activity, which we suspect results at least in part from the limited synthetic methodology to access this class. To date, four distinct approaches have been reported.^8–11We found the methods reported by Bell et al. to prepare the 2-amino-3-hydroxy-indoles Via reaction of KCN with formylated or dichloroacylated 2-aminobenzophenones to be the most appropriate (Scheme 1).⁹

Scheme 1.

Traditional Synthetic Strategy to Access 2-Amino-3-hydroxy-indoles via Cyanohydrins

Unfortunately, this strategy, although readily scalable and attractive for process scale syntheses, is limited for early stage exploratory medicinal chemistry due to the lack of commercially available 2-aminobenzophenones and the inherent lack of enantioselectivity. As a result, we developed a general, short and efficient method that would (a) provide analogues in quantities that satisfy early stage drug discovery requirements, (b) tolerate a wide variety of functional groups, (c) begin with commercially available diversely functionalized building blocks, and (d) avoid the use of cyanide while (e) offering the potential to enantioselectively access 2-aminy-3-hydroxy-indoles.

We envisioned a three-step reaction sequence starting from isatins, which are widely available and inexpensive starting materials as shown in Scheme 2.

Scheme 2.

Retrosynthetic Analysis Starting from Isatin (2)

Grignard addition or Rh-catalyzed addition¹² of boronic acids to isatin 2 would give 3-hydroxy-3-aryl-oxindole 3. Oxindole 3 could then react with a protected ammonia equivalent (e.g., allylamine) to yield the corresponding protected 2-amino-3-hydroxy-indole 4, which on deprotection would result in the desired 2-amino-3-hydroxy-indole 1. Furthermore, boronic acids¹³ and electron-rich arenes¹⁴ can be added to isatins with high enantioselectivity. This method would also allow access to 2-amino-3-hydroxy-indoles enantioselectively.

On the basis of the well-precedented nature of the first step, we directed our efforts toward the conversion of 3-hydroxyoxindoles 3 to allyl-protected 2-amino-3-hydroxy-indoles 4. We were somewhat encouraged by variable conversions (~30–50%) obtained upon heating 5-chloro-3-hydroxyoxindole 3a with an excess of allylamine in the presence of 5 mol % of PTSA and 4 Å molecular sieves.¹⁵ Attempts to increase the yields by employing other drying agents such as Na₂SO₄, MgSO₄ or the use of Dean-Stark apparatus were not successful. We anticipated that better conversions could be achieved with the use of a Lewis acid, which not only would catalyze the transformation but would also efficiently entrap the water formed. Exploring a set of Lewis acids confirmed our hypothesis, with both SnCl₄ and Ti(OiPr)₄ efficiently and more importantly, reproducibly promoting the conversion of 3a to 4a (Scheme 3). Furthermore, while microwave conditions reduced the reaction times from 12–18 h to 40–60 min, the addition of NMM (N-methyl morpholine) as an acid scavenger allowed the amine coupling partner to be used in nominal amount.

Scheme 3.

Amidations with Allylamine and Triphenylsilylamine Conditions^a

a Conditions: (a) excess allylamine for 4a, (b) 2 equiv of Ph₃SiNH₂ and 4 equiv NMM for 5.

Disappointingly, the deprotection of the allyl group in 4a proved challenging despite the plethora of deallylation protocols for amines and amides.^16,17

Next, we turned our attention to identifying an ammonia equivalent that would allow for more facile deprotection. On the basis of literature precedents of employing a silylated ammonia equivalent in unrelated transformations,¹⁸ we decided to explore commercially available Ph₃SiNH₂. Unexpectedly, as shown in Scheme 3, the reaction of 3a with Ph₃SiNH₂ under the optimized conditions produced O-silylated 2-amino-3-hydroxy-indole 5, in 60% yield while only trace amounts of the desired desilylated product 1a were observed. Aminoindole 5 is probably formed by the intramolecular migration of the silyl group. Attempts to desilylate 5 using TBAF or other fluoride sources resulted in hydrolysis of 5 and in recovery of the original hydroxyoxindole starting material 3a.

Encouraged by the finding that a silylamine appears well-suited to install the desired amine functionality, we suspected that less sterically demanding analogues of Ph₃SiNH₂ such as tert-butyldimethylsilyl amine (TBDMSNH₂) might allow deprotection under condition that would prevent hydrolysis of the amine. Pleasingly, we were able to directly isolate the desired 2-aminoindole product 1a in 60% isolated yield. The structure of 1a was unambiguously established by X-ray analysis revealing that the newly installed substituent appears to favor the tautomer with an exoamine rather than an exoimine geometry (see Supporting Information).

Although TBDMSNH₂ has been utilized in the context of its ligand attributes for metal complexes,^19–21 this is, to our knowledge, the first report on the application of TBDMSNH₂ in organic synthesis.

The synthetic utility of this novel SnCl₄-promoted amidation reaction with TBDMSNH₂ as an ammonia surrogate was then explored using a series of substituted 3-hydroxy-oxindoles. A variety of functional groups in the indole ring, including halogens, nitro, and alkyl ether are well tolerated in the amidation process (Table 1). The amidation reaction worked equally well under Ti(OiPr)₄ promoted conditions.

Table 1.

Direct Conversion of 3 to 1 with TBDMSNH₂

entry	R1	R2	yield (%)a	EC50 (nM)
				3D7	Dd2
a	5-CI	Ph	60 (56)b	216	507
b	5-I	Ph	64	289	338
c	5-CI	o-OMePh	43	24	57
d	5-CI		45	1160	1127
e	5-CI		58	104	465
f	5-F	p-OMePh	52	85	359
g	5-CI	iso-butyl	50	232	270
h	5-CI	m-CO2MePh	38	783	1578
i	5-CI	m-AcPh	30b	529	1525
j	5-CI	CH2-(o-tolyl)	53	63	139
k	4-Br	CH2Ph	51	>5000	>5000
l	5-OMe	Ph	49 (48)b	24	385
m	H	o-tolyl	53	112	152
n	5-CI	1-naphthyl	67	30	37
o	5-NO2	Ph	30b	>5000	>5000
p	5-CI	3’-(5’-OMe-indolyl)	40b	>5000	>5000
q	5-CI	p-SO2NH2Ph	45b,c	>5000	>5000
rd	H	Ph	39	>5000	>5000
	Chloroquine			16	216

^aIsolated yields.

^bTi(OiPr)₄ was used instead of SnCl₄ and NMM.

^cTBS-protected aminoindole was formed, which on treatment with HF/pyridine gave 1q; yield reported is over two steps.

^dThe indole nitrogen is methylated (see Supporting Information).

The substrate scope with respect to the substituent at the 3-position of oxindoles was also found to be broad. Thus, the aryl group at the 3-position possessing functional groups such as ester (entry h) and alkyne (entry d) could be employed in this process. In the case of an aryl group substituted with an enolizable ketone, the use of Ti(OiPr)₄ in lieu of SnCl₄ was critical for the success of the reaction (entry i, Table 1), as is the indole with free nitrogen as R² (entry p).

Interestingly, an aryl group substituted with the sulfonamide group yielded the aminoindole with TBDMS group on sulfonamide nitrogen with this method. However, deprotection with HF/pyridine gave the aminoindole 1q (entry q). It is worth noting that several of these 2-amino-3-hydroxy-indoles would be difficult to access or would require functional group manipulations via a cyanohydrin route significantly lengthening the synthesis. In addition, our methodology allowed for direct access of N(1)-alkyl iminoindole using the corresponding oxindole as starting material (entry r).

To ascertain whether an enantioenriched 3-hydroxyoxin-dole would transform into 2-amino-3-hydroxy-indole under the reaction conditions without any loss in enantiomeric excess, we prepared chiral 3-hydroxyoxindole (S)-3a²²(>95% ee) and subjected it to SnCl₄ promoted amidation reaction with TBDMSNH₂. Gratifyingly, the corresponding 2-amino-3-hydroxy-indole (S)-1a was obtained in good yield without any measurable loss of enantiomeric excess (>95%ee) indicating no racemization occurred during the reaction (Scheme 4).

Scheme 4.

Synthesis of Enantioenriched 2-Amino-3-hydroxy-indole 1a

The initial exploratory SAR against drug-sensitive (3D7) and drug-resistant (Dd2) parasite strains indicates that 2-amino-3-hydroxy-indoles are in general more active when R² is either an ortho-substituted electron-rich aromatic ring (entries c and m, Table 1) or a naphthyl moiety (entry n, Table 1), which is possibly a result of the increased dihedral angle of the biaryl system and thereby generating a more favorable binding conformation for its target(s). 2-Amino-3-hydroxy-indoles 1 with benzylic and alkyl groups at R² also displayed potent antimalarial activity for R¹ as 5-Cl. Furthermore, protection of the 3-hydroxyl (Scheme 3 compound 5; EC₅₀ >5 μM) or 2-amino groups (Scheme 3 compound 4a; EC₅₀ > 5 μM) was found to be detrimental. The 3-hydroxyl-oxindoles (3) which potentially can be regenerated by hydrolysis of the 2-amino group did not exhibit any significant antimalarial activity.

A 2-fold difference in in Vitro activity between the two enantiomers of 1a was noted and interestingly, racemate was found to be more active than either of the two enantiomers (Scheme 4).

In summary, we have identified 2-amino-3-hydroxy-indoles as a novel chemical class with potent in Vitro and in vivo antimalaria activity. We have developed a concise synthetic strategy to efficiently synthesize analogues in quantities sufficient for medicinal chemistry exploration. This method establishes the unprecedented use of TBDMSNH₂ as an ammonia surrogate and allows for the first enantioselective synthesis of 2-amino-3-hydroxy-indoles. It is likely that TBDMSNH₂ will find use in other transformations requiring protected ammonia equivalents.