Discovery at the VICB
Fast Efficient Approach to Macrocycles
By: Carol A. Rouzer, VICB Communications
Published: January 3, 2017
An approach to the synthesis of a class of large cyclic peptides offers flexibility and efficiency in a traditionally challenging field.
A growing interest among organic chemists is to develop the ability to synthesize small molecules that contain large rings. Such macrocyclic compounds show promise for the development of sensors, new materials, and novel therapeutics. In addition, many natural products that exhibit interesting biological activities contain macrocycles, and the ability to synthesize these molecules enables chemists to incorporate structural variations that might enhance desired and reduce unwanted properties. The synthesis of large rings is challenging, however, as the reactions intended to unite the two ends of a linear precursor frequently result, instead, in joining the precursor molecules to each other. Furthermore, most past syntheses of macrocyclic compounds have focused on generating a single target molecule, providing little opportunity to create libraries of similar molecules for testing. These challenges led Vanderbilt Institute of Chemical Biology investigator Jeffrey Johnston and his doctoral student Suzanne Batiste to develop a new approach to the synthesis of a class of molecules called cyclodepsipeptides (Figure 1). Their method enables control of multiple structural features of the compounds produced in a relatively quick and efficient synthetic scheme. [S. M. Batiste & J. N. Johnston, Proc. Natl. Acad. Sci. U.S.A., (2016) published online December 14, DOI: 10.1073/pnas.1616462114].
FIGURE 1. (A) Structures of amino acids typical of those found in peptides and proteins. The structure includes a carboxylic acid group, an amino group on the carbon next to the carboxylic acid (the α carbon), and one of up to 20 different substituents designated here as R or R´. The orientation of the substituents on the α carbon is important, and in proteins, it is always as shown here. In bacteria and some other lower organisms, alternative substituents, structures, and geometric orientations may be found. The linking of the carboxylic acid group of one amino acid and the amino group of a second one to form an amide bond (also known as a peptide bond, shown in red) forms a dipeptide. Addition of more amino acids yields tripeptides, tetrapeptides, etc. (B) Depsipeptides are structurally similar to peptides with the exception that some of their subunits are hydroxy-acids rather than amino acids. As a result the linkages between them are esters (blue) rather than amides.
Peptides are polymers made up of amino acids, compounds that contain a carboxylic acid group and an amino group (Figure 1). Out of the hundreds of known amino acids, there are 20 that are commonly used to make proteins in biological systems. Joining the amino group of one amino acid to the carboxyl group of a second one to form an amide bond yields a dipeptide. This process can be continued to make long chains of amino acids. Depsipeptides are similar with the exception that some monomers used in their construction are hydroxy-acids rather than amino acids, and the linkage joining them together is an ester bond. Cyclodepsipeptides result from the head-to-tail joining of a depsipeptide chain to form a ring.
As a target compound for their synthetic scheme, Batiste and Johnston chose verticilide (Figure 2), a 24-membered cyclodepsipeptide natural product that has garnered interest due to its ability to block the ryanodine receptor in insects. A previously reported total synthesis of this compound required 14 steps. Verticilide is a polymer of a six atom monomer that contains an internal amide bond and terminal carboxyl and hydroxyl groups. Monomer 1, (Figure 3) is identical to the verticilide monomer with the exception that it lacks a methyl group on the amide nitrogen atom. The researchers synthesized Monomer 1 using methods developed in their laboratory that are specially designed to preserve the desired three dimensional geometry of the product molecule. Once the monomer was made, they subjected it to Mitsunobu reaction conditions, in order to make ester linkages between the carboxylic acid and hydroxyl groups of the monomer. They hypothesized that two things would occur - monomers would be linked together, and at least some of the resulting polymers would cyclize to form macrocyclic compounds.
FIGURE 2. Chemical structures of verticilide (top left) and bassianolide (top right). General structure of a cyclodepsipeptide showing structural features that can be readily modified using the synthetic scheme developed by Batiste & Johnston. Figure reproduced with permission from S. M. Batiste & J. N. Johnston, Proc. Natl. Acad. Sci. U.S.A., (2016) published online December 14, DOI: 10.1073/pnas.1616462114. Copyright 2016, Batiste & Johnston.
Their approach was a success. The reaction produced four major macrocyclic products, containing 18, 24, 30, and 36 atoms in the ring at yields of 9, 15, 16, and 24%, respectively. The 24-membered product was structurally the same as verticilide except that it lacked the methyl groups on the amide nitrogen atoms (highlighted in orange in Figure 2). The chemists added those groups in one final step, yielding verticilide in 9% yield following a six-step synthetic scheme.
Although their efforts had yielded a shorter synthetic scheme for verticilide, the investigators hypothesized that they could improve the yield of the desired product by adding salts to the reaction mixture during the cyclization step. Prior studies had shown that complexation of ions from a salt to the precursor molecules during cyclization reactions could affect their conformation, thereby altering the size of the macrocycle ultimately obtained. This approach had not been previously applied to Mitsunobu reaction conditions, or to cyclodepsipeptide synthesis, however. To test their hypothesis, the investigators attempted monomer polymerization in the presence of a variety of salts. They discovered that salt addition effectively eliminated production of the 18-atom ring, and different salts favored rings of other sizes. Addition of NaBF4 maximized production of the 24-atom macrocycle, increasing its yield to 25%. However, the 36-atom ring remained the most highly favored product under all conditions, with its yield maximized to 44% in the presence of KBF4.
These findings demonstrated that the reaction conditions could control, at least to some degree, the size of the product macrocycle. The investigators hypothesized that further control might be gained by altering the structure of the monomer. To test this hypothesis, they synthesized Monomer 2 (Figure 3), a 12-atom long compound comprising two units of Monomer 1 joined by an ester linkage. Subjecting Monomer 2 to Mitsunobu conditions yielded 24-, 36-, and 60-atom ring products in yields of 63, 10, and 5%, respectively. Inclusion of NaBF4 increased the yield of the 24-atom ring to 89%, practically eliminating formation of the other two products. Other salts increased formation of the larger macrocycles, but the 24-atom ring product predominated under all conditions. The improved yield obtained in the presence of NaBF4, enabled Batiste and Johnston to complete a total synthesis of verticilide in 8 steps with an overall 36% yield.
FIGURE 3. Monomers used for the synthesis of verticilide (monomers 1 and 2), and bassianolide (monomer 3).
To see if their approach could be generalized to other cyclodepsipeptides, the chemists next turned to the total synthesis of a more complex natural product, bassianolide (Figure 2). This macrocycle first required the synthesis of the tetradepsipeptide Monomer 3 (Figure 3). Exposing Monomer 3 to Mitsunobu reaction conditions produced the desired ring at a yield of 11% in the absence of added salt. Addition of NaBF4 increased the yield to 31%. Methylation of the amide nitrogen atoms completed the synthetic scheme. The entire synthesis of bassianolide comprised 8 steps and provided an overall yield of 9%, a considerable improvement over the previously reported 2.8% yield achieved using more conventional approaches.
These results represent a major step forward in the synthesis of cyclodepsipeptides. Not only did the investigators achieve improved product yields in fewer steps, their method also provided opportunities to easily alter various aspects of the final product's structure, including size of the ring, substituents present on the ring, and the geometry of those substituents (Figure 2, bottom). This added versatility will enable the exploration of a large number of novel compounds, greatly expanding the structural diversity of molecules that may be used for the development of new materials and therapeutics in the future.