‘Click’ Reaction in Conjunction with Diazeniumdiolate Chemistry: Developing High-Load Nitric Oxide Donors

Abstract

The use of Cu(I)-catalyzed ‘click’ reactions of alkyne-substituted diazeniumdiolate prodrugs with bis- and tetrakisazido compounds is described. The ‘click’ reaction for the bis-azide using CuSO₄/Na-ascorbate predominantly gave the expected bis-triazole. However, CuI/diisopropylethylamine predominantly gave uncommon triazolo-triazole products as a result of oxidative coupling. Neither set of ‘click’ condition showed evidence of compromising the integrity of the diazeniumdiolate groups. The chemistry developed has applications in the synthesis of polyvalent and dendritic nitric oxide donors.

Diazeniumdiolate prodrugs¹ are efficient and reliable sources of nitric oxide (NO), a potent bioregulatory agent.^2,3 These prodrugs on suitable hydrolysis or enzymatic activation rapidly decompose to release NO at physiological pH. Targeted NO-releasing prodrugs have potential applications in treating cancer and other diseases. For example, JS-K (1, Figure 1) is an anticancer lead compound activated by gluthathione. JS-K (1) slowed tumor growth in several rodent models of cancer including leukemia, prostate cancer, multiple myeloma, and liver cancer.⁴ Another prodrug, V-PYRRO/NO (2, Figure 1) is activated by cytochrome P450 to release NO. It shows hepatoprotective properties against a variety of toxins in several animal models.⁵ Glycosidase activated N-acetyl glucosaminylated diazeniumdiolate GlcNAc-DEA/NO (3) significantly decreased Leishmania major parasite burden in infected macrophages.⁶ The 2,4-dinitrophenyl, vinyl and GlcNAc protecting groups are well studied for their trigger mechanisms to release NO from diazeniumdiolates.

Figure 1.

Structures of NO-releasing prodrugs JS-K (1), V-PYRRO/NO (2) and GlcNAc-DEA/NO (3).

Our laboratory is involved in NO-drug development mainly aimed to increase the potency of these prodrugs,⁷ and their effective site-directed delivery of NO. Some strategies we employ to achieve these ends are increasing the payload of NO per mole of prodrug (high-load NO-donors) and conjugating our NO-prodrugs with other biologically significant molecules. However, many of these prodrugs decompose under certain reaction conditions, thus limiting the scope of suitable reagents/reaction conditions. The Cu(I)-catalyzed azidealkyne cycloaddition (‘click’) reaction, independently discovered by the Sharpless⁸ and Meldal⁹ groups, involves mild reaction conditions. They are popular in medicinal chemistry because of their ease of execution, functional group tolerance, and potential to produce a library of compounds.¹⁰ Therefore the ‘click’ reaction was an attractive method to achieve the above-mentioned properties in NO-drug development.

The ‘click’ reaction between a suitably functionalized diazeniumdiolate prodrug having a terminal alkyne group and a polyazide can lead to multivalency and increased payload of NO. We planned to investigate the reaction of a bis-azide with the alkyne group attached to the diazeniumdiolate prodrug. Finn and co-workers reported mechanistic studies and reactivities of bis-azides in ligand-free Cu(I)-catalyzed ‘click’ reactions.¹¹ The report suggests that conformationally constrained 1,3-bis-azides 4 and 5 preferably form bis-triazole derivatives 6 and 7, respectively, over less hindered 8 and 9 (Figure 2), even if ten-fold excess of bis-azide over alkyne were used. In case of 4 and 5, formation of the first triazole ring catalyzes the subsequent cycloaddition to give the required bis-triazole predominantly.¹¹

Figure 2.

Preferential bis-triazole formation by hindered bisazides 4 and 5 [ref ¹¹].

We envisioned that addition of benzylidene protection to 2,2-di(azidomethyl)propane-1,3-diol 4 would further add to conformational strain in the system, and hence lead to preferential bis-triazole formation. Furthermore, suitably substituted benzylidenes have potential applications in synthesis of dendritic azides. Thus, the diol 4 was transformed into bis-azide 10 and tetrakis-azide 11 (Figure 3) (details in Supporting Information).

Figure 3.

Structures of bis-azide 10 and tetrakis-azide 11 synthesized for the ‘click’ reaction.

The diazeniumdiolate prodrugs 12-15 with terminal alkynes are shown in Figure 4 (details of their synthesis and characterization in Supporting Information). These alkynes represent a sample of O²-protected diazeniumdiolates with various modes of activation. Compound 14 is reported, and has shown anti-proliferative activity comparable to JS-K (1) against HL-60 and U-937 leukemia cell lines.¹²

Figure 4.

Diazeniumdiolate prodrugs 12-15 with terminal alkyne groups synthesized for use in the ‘click’ reaction.

The ‘click’ reaction of the alkynes 12-15 with bis-azide 10 was performed using CuSO₄/Na-ascorbate in THF:water (Table 1). The reaction proceeded quickly (15-45 min). However, we observed formation of two products. The major product in each case was the expected bis-triazole. The minor product was 5,5′ triazolo-triazole. No formation of monotriazole product was observed.

Table 1.

1,3-Dipolar cycloaddition using CuSO₄/Na-ascorbate for bis-azide 10.

azide	alkyne	% yield (bis- triazole)	% yield (triazolo- triazole)
10	12	60 (16)	12 (17)
10	13	67 (18)	14 (19)
10	14	75 (20)	10 (21)
10	15	63 (22)	trace (23)

CuSO₄·5H₂O (40 mol %)/Na-ascorbate (80 mol %), THF:H₂O (3:1), isolated product yields.

Formation of such triazolo-triazoles is reported in the literature, but often as undesired and uncharacterized impurities.¹³ They are a result of the oxidative coupling reaction of copper species after the triazole formation. To the best of our knowledge, there has only been one detailed report showing base dependence for the formation of triazolo-triazoles in an intermolecular cycloaddition reaction.¹⁴ In the cycloaddition reaction of azides 10 and 11, the benzylidene protection may bring the two newly formed triazole rings fairly close to enhance the formation of triazolo-triazole by oxidative coupling. These 5,5′-triazolo-triazoles remain under-explored heterocycles. Therefore, we attempted to synthesize triazolo-triazoles as the predominant product formed in case of bis-azide 10, as they may be good additions to heterocyclic linkers in medicinal chemistry. The reaction of alkynes 12-15 with 10 using CuI and diisopropylethylamine (DIPEA) as base predominantly gave the cycloaddition/oxidative coupling products (Table 2). Thus, steric congestion can be an additional factor for such oxidative dimerization reaction in Cu(I)-catalyzed cycloadditions.

Table 2.

1,3-Dipolar cycloaddition using CuI/DIPEA for bis-azide 10.

azide	alkyne	% yield (bis-triazole)	% yield (triazolo- triazole)
10	12	trace (16)	74 (17)
10	13	28 (18)	44 (19)
10	14	20 (20)	72 (21)
10	15	19 (22)	56 (23)

CuI (2 equiv)/DIPEA (2 equiv), isolated product yields.

In the case of the tetrakis-azide 11, an additional partially coupled product (triazolo-triazole bis-triazole adduct), along with the expected tetrakis-triazole and bis(triazolo-triazole) may occur. To minimize the formation of these by-products, and avoid difficulties of their chromatographic separations, the 1,3-dipolar cycloadditions of tetrakis-azide 11 with alkynes 12 and 13 were carried out in absence of oxygen using CuSO₄/Na-ascorbate (Table 3). Under nitrogen atmosphere, the cycloaddition reactions of 12 and 13 gave the desired tetrakis-triazoles 24 and 25, respectively, in acceptable yields. Efforts are underway to make tetrakis-triazoles of diazeniumdiolates with 2,4-dinitrophenyl and GlcNAc protecting groups.

Table 3.

1,3-Dipolar cycloaddition using CuSO₄/Na-ascorbate for tetrakis-azide 11.

azide	alkyne	% yield (tetrakis-triazole)
11	12	54 (24)
11	13	46 (25)

Under N₂ reaction, CuSO₄·5H₂O (80 mol %)/Na-ascorbate (160 mol %), THF:H₂O (3:1), isolated product yields.

The hydrolysis or enzymatic activation to release NO for these types of prodrugs is well established in the literature.^4-7,12 However, the NO release profile for the bis-triazole, triazolo-triazole and tetrakis-triazole linked prodrugs may have differences in their kinetics. In addition, the steric effect on enzymatic activation, their intracellular NO-release, biological evaluation, toxicity and physiological clearance of the linker^15,16 need to be investigated. Efforts are underway in these directions. If required, the phenyl ring of benzilidene protection in these compounds can be suitably substituted to improve their cell permeability and/ or solubility.

Thus the Cu-mediated ‘click’ reaction was utilized to tether two and four molecules of NO-donor prodrugs together, which would serve to increase the payload of NO. The reaction conditions did not compromise the integrity of the diazeniumdiolate functional group, which is sensitive to certain other reaction conditions. The reactions showed formation of unusual oxidative coupling products, which were well characterized. Use of CuI/DIPEA in acetonitrile to carry out the dipolar cycloaddition preferentially formed the oxidatively coupled triazole adducts, leading to unexplored heterocyclic linkers for biological applications. These results underline the need to further investigate the role of steric congestion for the formation of such Cu-catalyzed oxidative coupling products. A proper substitution in the phenyl ring of the synthesized substrates along with a proper coupling reaction may allow the synthesis of well defined dendritic NO-donors. These dendritic NO-donors can have potential applications in NO-releasing materials. The chemsitry established has potential applications in synthesizing hybrids of biologically important molecules and NO-donors. Efforts are underway for the synthesis of such NO-donor hybrids using the ‘click’ reaction.