Highly Substituted Indole Library Synthesis by Palladium-Catalyzed Coupling Reactions in Solution and on a Solid Support

Abstract

3-Iodoindoles have been synthesized by the iodocyclization of N,N-dialkyl-o-(1-alkynyl)anilines, obtained by the Pd/Cu catalyzed coupling of terminal acetylenes with N,N-dialkyl-o-iodoanilines. These 3-iodoindoles undergo palladium-catalyzed Sonogashira and Suzuki coupling reactions to yield 1,2,3-trisubstituted indoles. These reactions have been applied to parallel library synthesis utilizing commercially available terminal acetylenes and boronic acids. The aforementioned chemistry has also been carried out on a chlorinated Wang resin as a solid support, affording 1,2,3,5-tetrasubstituted indoles after cleavage from the support. A diverse 42-member library of highly substituted indoles has been synthesized.

Introduction

Indoles are very important in medicinal chemistry and the indole moiety is prevalent in numerous naturally-occurring and synthetic biologically active compounds.¹ It is one of the most important nitrogen-containing pharmacophores,² and is present in various drugs.1b,1g Due to the importance of the indole nucleus, many synthetic approaches to this ring system have been developed in our research group and others and reported in the literature for the synthesis of substituted indoles.³ Biologically active natural products are a good indicator of lead structures that might possess biological activity. Due to the biological importance of compounds containing the indole nuclei, it is quite likely that libraries of low molecular weight indoles will display similar activity and thus serve as valuable tools for drug development. Several methods are known for the synthesis of indoles in solution phase⁴ and on a solid support⁵ by combinatorial methods, but 3-iodoindoles have not previously been examined as key intermediates for indole library synthesis.

Yamanaka et al. have reported the coupling of 3-iodoindoles with terminal acetylenes, but satisfactory results were obtained only when the N atom of the indole was protected with an electron-withdrawing 1-methanesulfonyl group.⁶ With an electron-donating group on the N atom of the 3-iodoindole, the C-I bond is electron-rich and this appears to limit further functionalization at the 3 position of the indole by palladium-catalyzed coupling reactions.

Previously, in our laboratory, we have synthesized N,N-dialkyl-o-(1-alkynyl)anilines (1) by coupling terminal acetylenes with N,N-dialkyl-o-iodoanilines in the presence of a Pd/Cu catalyst, which on iodocyclization yield 3-iodoindoles (2) in excellent yields (Scheme 1).⁷ We have previously reported individual examples of Sonogashira⁸ and Suzuki-Miyaura⁹ cross-coupling reactions, which provide the corresponding 1,2,3-trisubstituted indoles in good yields (Scheme 2).^6b We further optimized each of these processes in order to adapt them for library generation. We have previously also reported individual examples of these two coupling reactions on a solid support, followed by cleavage by base.¹⁰ The development of a solid phase version of this chemistry allows the multistep synthesis of highly substituted indoles and eliminates cumbersome purification steps. We herein report the successful synthesis of 1,2,3,5-tetrasubstituted indoles on a solid support by slight modifications of our earlier procedure and alternative cleavage reactions (Scheme 3).

Scheme 1.

Scheme 2.

Scheme 3.

Results and Discussion

Our previous work on 3-iodoindole synthesis reported good yields of single cyclization products from the corresponding N,N-dimethyl- or N-methyl, N-phenyl-o-(1-alkynyl)anilines (1). After the iodocyclization step in the former case, the N-atom of the 3-iodoindole is protected by a methyl group, and, in the latter case, by a phenyl group. Our desire for a low molecular weight indole library led us to choose methyl as the preferred N-protecting group. Therefore, our choice of R² was a methyl group in our solution phase library synthesis. 3-Iodoindole 2{1} was synthesized as our basic scaffold by using our previous cyclization method.⁷ The 3-iodoindoles 2{2} and 2{3} were similarly synthesized from the corresponding N, N-dimethyl-o-(1-alkynyl)anilines 1. Due to certain limitations in the types of R¹ and R² groups that can be employed in our iodocyclization methodology, we synthesized the 3-iodoindoles 2{4}¹¹ and 2{5}¹² by literature methods, while the 3-iodoindole 2{6} was obtained by treatment of 2{5} with NaBH₄. Thus, we choose a subset of 3-iodoindoles on the basis of the ease of synthesis from readily available starting materials and with different electron-donating and electron-withdrawing functionalities at the 2-position of the indoles (Figure 1).

Figure 1.

3-Iodoindole sublibrary.

The terminal alkyne sublibrary was chosen on the basis of commercially available acetylenes. Attempts were made to include heteroatoms in the acetylenes that could impart drug-like, hydrogen bond donor and/or acceptor properties to the indoles after Sonogashira coupling (Figure 2). For similar reasons, acetylenes 3{5} and 3{8} were chosen due to the increasing popularity of fluorine¹³ and sulfur¹⁴ atoms in drug molecules.

Figure 2.

Terminal acetylene sublibrary.

The boronic acids for the Suzuki-Miyaura reactions were also chosen on the basis of their commercial availability and their ability to provide the requisite diversity and drug-like properties to the indole scaffold after subsequent cross-coupling reactions (Figure 3). For instance, the methoxy-containing boronic acids 4{1} and 4{2} were chosen with a view towards increasing the polarity of the substituted indole. The N- heterocyclic boronic acids 4{3}, 4{4}, the O- heterocyclic boronic acids 4{5}, 4{6}, 4{7}, the indolylboronic acid 4{9}, and the benzamido boronic acid 4{11} were chosen to include heteroatoms and increase the drug-like nature of the corresponding indoles. The fluorine-containing acids 4{8} and 4{10} were desirable due to the importance of fluorine in medicinal chemistry.

Figure 3.

Boronic acid sublibrary.

Having chosen these sublibraries, we proceeded to prepare a diverse library of 1,2,3-trisubstituted indoles via solution phase chemistry as outlined in Scheme 2 and 1,2,3,5-tetrasubstituted indoles using a chlorinated Wang resin as the solid support as depicted in Scheme 3. The crude products have been analyzed by LC/MS, followed by purification by preparative HPLC or flash chromatography.

A summary of the results of the library synthesis is provided in Tables 1–3. Most of the crude products were subjected to preparative HPLC. Purities in the range of 70–100% have been achieved after purification. Most of the Sonogashira coupling reactions proceeded well, except for those run with the terminal alkynes 3{11} and 3{12}. Suzuki-Miyaura reactions with the boronic acids 4{10} and 4{11} with electron-withdrawing groups failed to give the desired coupling products. The boronic acids 4{3} and 4{8} gave decent yields of the coupling products 5{32} and 5{33} and excellent purities when reacted with 3-iodoindole 2{2}, but failed to give the corresponding trisubstituted indoles when coupled with 3-iodoindole 2{3}. On the solid support, the cleavage by MeMgBr was successful, but EtMgBr failed to give the anticipated products. Out of a total of 51 palladium-catalyzed processes attempted, around 80% were successful.

Table 1.

Library Data for Compounds 5{1–29}

compound	R	yielda (%)	purityb (%)
5{1}	4-MeOC6H4	12	97
5{2}	3,5-dimethoxyphenyl	38	94
5{3}	4-fluoro-3-methylphenyl	47	99
5{4}	4-H2NC6H4	27	95
5{5}	4-Me2NC6H4	30	79
5{6}	1-amino-1-cyclohexyl	-	-
5{7}	2-hydroxypropyl	-	-
5{8}	(1-hydroxy-1-methyl)ethyl	13	70
5{9}	C6H5	43	99
5{10}	3,5-dimethoxyphenyl	34	98
5{11}	4-Me2NC6H4	59	97
5{12}	3-thiophenyl	20	93
5{13}	(1-hydroxy-1-methyl)ethyl	33	90
5{14}	C6H5	45	98
5{15}	C6H5	89	100
5{16}	4-MeOC6H4	90	97
5{17}	3-MeOC6H4	79	95
5{18}	3,5-dimethoxyphenyl	94	91
5{19}	1-methyl-1H-imidazol-5-yl	82	90
5{20}	(1-hydroxy-1-methyl)ethyl	77	95
5{21}	C6H5	52	100
5{22}	3,5-dimethoxyphenyl	52	90
5{23}	1-methyl-1H-imidazol-5-yl	36	98
5{24}	3-thiophenyl	76	93
5{25}	(1-hydroxy-1-methyl)ethyl	43	96
5{26}	4-MeOC6H4	21	86
5{27}	3-MeOC6H4	36	92
5{28}	3,5-dimethoxyphenyl	3	100
5{29}	(1-hydroxy-1-methyl)ethyl	13	95

^aIsolated yield after preparative HPLC.

^bUV purities determined at 214 nm after preparative HPLC.

Table 3.

Library Data for Compounds 5{46–51} Synthesized on a Solid Support

compound	R	yielda (%)	purityb (%)
5{46}	C6H5	69c	<90d
5{47}		-
5{48}	C6H5OMe-p	-
5{49}	C6H5	60c	<90d
5{50}		54	91
5{51}	C6H5OMe-p	64	95

^aIsolated yield after preparative HPLC.

^bUV purities determined at 214 nm after preparative HPLC.

^cIsolated yield after flash chromatography.

^dPurities determined by H¹ NMR spectroscopy after flash chromatography.

Our goal in synthesizing these low molecular weight heterocycles is for use in high-throughput screening projects. Therefore, we carried out an in silico evaluation of these library members to determine their agreement with Lipinski’s¹⁵ “rule of five” and Veber’s rules.¹⁶ The SYBYL¹⁷ program was used for the calculation of molecular weight, clog P, the number of hydrogen bond donors and acceptors, and the number of rotatable bonds for each library member (Table 4). According to these rules a potential drug molecule is more drug-like and more bioavailable if the clog P value is not more than 5, the molecular weight is less than 500, the hydrogen bond acceptors are not more than 10, the hydrogen bond donors are not more than 5, and the rotatable bonds in the molecule are not more than 12. One Lipinski violation is allowed for potential drug design. All of the indole library members are Lipinski compliant and no molecule has more than one Lipinski violation. The only violation that a molecule in the library had was clog P, which points towards potential solubility and delivery issues.

Table 4.

In silico parameters for gauging oral availability/drug-likeness

	Mean	St. Dev.	Range
Clog P	5.1	1.9	0.6 – 8.0
Mol. Weight	317	53	227–433
H-Bond Acceptors	2.0	0.9	0 – 4
H-Bond Donors	0.8	0.8	0 – 3
Rotatable Bonds	4.1	1.1	2 – 6

Conclusions

In conclusion, the synthesis of 4-iodoindoles and subsequent palladium-catalyzed Sonogashira and Suzuki-Miyaura cross-coupling reactions with various commercially available terminal alkynes and boronic acids have allowed the construction of a 42-member library of highly substituted indoles. The chemistry has been successfully transferred to a solid support and diversity has been achieved at the 5-position by different cleavage reactions. The average yield of the library was 46% and the average purity after purification was 94%.

Table 2.

Library Data for Compounds 5{30–45}

compound	Ar	yielda (%)	purityb (%)
5{30}	4-MeOC6H4	23	100
5{31}	3,4,5-trimethoxyphenyl	79	83
5{32}	3-fluoro-4-methoxyphenyl	42	100
5{33}	2-methoxy-5-pyridinyl	50	98
5{34}	benzo[1,3]dioxol-5-yl	39	99
5{35}	2-methoxy-5-pyrimidinyl	59	99
5{36}	4-H2NC(O)C6H4	-	-
5{37}	4-EtO2C-3-FC6H3	-	-
5{38}	3-fluoro-4-methoxyphenyl	-	-
5{39}	2-methoxy-5-pyridinyl	-	-
5{40}	2,3-dihydrobenzo[1,4]dioxin-6-yl	84	89
5{41}	2-methoxy-5-pyrimidinyl	2	100
5{42}	6-indolyl	25	94
5{43}	benzo[1,3]dioxol-5-yl	9	100
5{44}	6-indolyl	-	-
5{45}	4-(tetrahydropyran-2-yloxy)phenyl	9	91

^aIsolated yield after preparative HPLC.

^bUV purities determined at 214 nm after preparative HPLC.