Chapter 2
Synthetic Methods for Macrocyclic Polyamines

Cyclization and polymerization reactions employ the same starting materials and compete with each other in most cases. Therefore, the major effort in the synthesis of macrocyclic compounds is to manipulate the orientation of the reactive sites to afford macrocyclic products rather than acyclic polymers. There are two general means to improve the ring-closure reaction: (i) performing the reaction in high-dilution conditions and (ii) using a suitable metal template to interact with the heteroatoms. Chemists have also developed a variety of other efficient strategies of accessing polyazamacrocycles without the use of high dilution or templates, such as the Richman-Atkins reaction, the crab-like cyclization, and the condensation of diacids with diamines. MPAs are not limited to macrocycles bearing amine functional groups but also include imine, amide, and other functional groups. In this chapter, we divide MPAs into five categories: saturated MPAs, aromatic subunit-containing MPAs, macrocyclic amides, macrocyclic imines, and polyaza cryptands. Although the synthesis of MPAs has been summarized well in one book [1] and multiple reviews [2], these treatments have discussed only some of the categories of MPAs. We will introduce some frequently used methods to prepare all five categories.

2.1 Ring-Closure Modes

2.1.1 Intramolecular Cyclization

A linear compound bearing reactive groups on each end can undergo intramolecular cyclization to afford a cyclic compound (Figure 2.1a). This intramolecular cyclization has been widely applied for the synthesis of cyclic peptides. It can also be used for the synthesis of MPAs. For example, nitrobenzyl-substituted cyclen was prepared by intramolecular nucleophilic substitution in DMF at 60 °C (Figure 2.1b) [3]. The most used 1 + 1 cyclization was not applicable to the synthesis of this product.

Figure 2.1 Intramolecular cyclization.

2.1.2 1 + 1 Cyclization

The most often used ring-closure mode for the synthesis of macrocycles is the 1 + 1 cyclization (Figure 2.2a). Each precursor has two reactive functional groups on the end. The 1 + 1 cyclization has been extensively used for the synthesis of polyazamacrocycles. The Richman-Atkins reaction is a typical 1 + 1 type cyclization (Figure 2.2b) [4].

Figure 2.2 [1 + 1] Cyclization.

2.1.3 2 + 2 Cyclization

In the synthesis of polyazamacrocycles, the 2 + 2 cyclization (Figure 2.3a) is often a side reaction of a 1 + 1 cyclization. The concentration of the substrates affects the type of cyclization. A higher concentration favors the 2 + 2 cyclization by favoring polymerization. For example, 36-membered MPAs were prepared by 2 + 2 cyclization (Figure 2.3b) [5]. When the reaction was conducted at 0.02 M, the 1 + 1 cyclization process was largely preponderant (95 : 5); by contrast, at 0.5 M, the 2 + 2 mode was greatly favored (10 : 90).

Figure 2.3 [2 + 2] Cyclization.

2.1.4 Other Cyclization Modes

There are many less used ring-closure modes for the synthesis of polyazamacrocycles, such as 3 + 3, 4 + 4, and 2 + 1. The 3 + 3 and 4 + 4 cyclizations occur as side reactions in the 1 + 1 or 2 + 2 cyclization mentioned earlier. The yields decrease as the size of the ring increases. These products can be isolated by careful chromatography.

2.2 The Synthesis of Saturated Macrocyclic Polyamines

2.2.1 Ring Closure Using Sulfonamides

The utilization of sulfonamides for ring closure was reported early. The introduction of a sulfonyl group to the nitrogen atom not only increases the acidity, thus facilitating the formation of salts with bases, but also forces the open-chain compounds into macrocycle-like conformations. The sulfonyl groups (nearly always tosyl moieties) of the resulting sulfonyl-substituted polyamines are cleaved to form saturated MPAs.

The 24-membered tetraazamacrocycle 2-1 was first prepared in 1954 via a 1 : 1 cyclization between disodium salts of sulfonamide with a dibromide compound in a high dilution in DMF [6]. The 10- to 12-membered triaza compounds were prepared in less than 30% yields using a similar method [7]. Cyclization between sulfonamide salts with a terminal dihalide is usually performed under high-dilution conditions and affords products in low yield. Remarkable progress was made in 1974, when Richman and Atkins utilized a terminal disulfonate rather than a terminal dihalide for this cyclization [4]. The reaction was conducted by reacting disodium salts of sulfonamide with a terminal ditosylate or dimesylate in DMF at elevated temperature without the use of high-dilution techniques or templates, and 9- to 21-membered rings containing three to seven heteroatoms were obtained in 40-90% yields.

The so-called Richman-Atkins reaction is the cyclization of dimetal salts of sulfonamide with a terminal ditosylate or dimesylate in a dipolar aprotic solvent at elevated temperature without the use of high-dilution techniques or templates (Figure 2.4). However, the reaction of dimetal salts of sulfonamide with a terminal dihalide is also occasionally called the Richman-Atkins reaction. The Richman-Atkins reaction is probably the most widely used process for the synthesis of MPAs, especially for saturated MPAs.

Figure 2.4 A typical example of the Richman-Atkins reaction.

The tosylation of polyamines or glycol to prepare pertosylamides or terminal ditosylates is not complicated, but both precursors must be purified by crystallization or column chromatography before use. DMF, DMSO, and HMPA are good solvents for the cyclization, but DMF is more convenient. Purification of DMF is usually required. Dimethylamine, formaldehyde, and water impurities in DMF will decrease the yields of macrocyclic products.

The sulfonamide salts are prepared by adding sodium, sodium hydride, sodium methoxide or ethoxide, sodium hydroxide, sodium carbonate, potassium carbonate, potassium t-butoxide, or caesium carbonate to the solution of sulfonamide. The resultant salts can be isolated but are usually prepared in situ and used immediately because they are moisture sensitive. The nature of the bases affects the cyclization reaction. Caesium and potassium carbonate are the best bases for this reaction [8]. From an economic perspective, potassium is a good choice. The stronger bases accelerate the decomposition of the nucleophilic ditosylate precursor, thus decreasing the yield of the macrocyclic product. When the temperature is elevated, the decomposition occurs more quickly.

The leaving group X has a marked effect on the cyclization. As shown in Figure 2.5, the ditosylate and the dimesylate partner afforded 2-2 in good yields (80% and 66%, respectively), whereas dihalides gave lower yields of the product [9]. However, in another experiment, the leaving group had a different effect on the cyclization [8]. Bromides and mesylates gave the products in 70% and 75% yields, respectively, after 24 h at 30 °C. Chlorides and tosylates reacted more slowly under the same conditions; comparable yields were obtained at 50 °C. The requirement for different conditions may be due to the different effects of the leaving group. Usually, in a typical Richman-Atkins reaction, a terminal ditosylate or dimesylate gives better results than halides.

Figure 2.5 The effect of leaving groups.

Although most of the reactions were conducted at high temperature (100 °C), lower temperatures (room to ~50 °C) gave better results than high temperatures. Twelve macrocycles (17-35-membered rings, 4-8 nitrogen atoms) were prepared by condensation of sulfonamide salts with ditosylates in DMF at ambient temperature [10]. The yield was high, up to 99% (for [17]ane N4).

Usually, there is more than one possible synthetic pathway to produce the same final product. The choice of one preparative method over another depends mainly on the availability and ease of synthesis of the precursors. Iwata and Kuzuhara found that the coupling of the nucleophile containing the longest chain length with the electrophile containing the shortest chain length gave the best yields [10]. For example, products of macrocycles 2-3, 2-4, and 2-5 prepared through method (a) were higher than those through method (b). By employing this rule, a 52-membered MPA bearing 12 nitrogen atoms in the ring was obtained in 72% yield using 1,4-dibromobutane as the electrophile [11].

2.2.2 The Removal of Tosyl Protecting Groups

The tosyl groups can be removed by three general methods: (i) acid hydrolysis with concentrated sulfuric acid, (ii) cleavage with a HBr/acetic acid mixture, and (iii) reduction with other reducing agents, such as sodium in liquid ammonia and lithium aluminum hydride.

The most used detosylation method is acid hydrolysis with concentrated (90-97%) sulfuric acid. Impurities and water have obvious negative effects on the final yields. Deprotection is usually performed at elevated temperature (~100 °C). Removing more tosyl groups requires a longer reaction time. Fast detosylation was achieved by heating at 180 °C [12]. This method works very well with peraza macrocycles. However, acid hydrolysis does not always give good yields [13] and is less satisfactory for oxygen- and...