Design, Synthesis and Structural Analysis of D,L-Mixed Polypyrrolinones 2: Macrocyclic Hexapyrrolinones

Abstract

The design, synthesis and structural analysis of two macrocyclic D,L-alternating hexapyrrolinones has been achieved. These cyclic peptide mimics adopt a flat, hexagonal conformation, stabilized by intramolecular hydrogen bonding between adjacent pyrrolinone rings. Extensive NMR studies and X-ray analysis reveal respectively that the macrocyclic hexapyrrolinones aggregate in solution, and in the solid state form staggered stacked nanotube-like assemblies.

The evolution of the pyrrolinone scaffold towards a universal peptidomimetic possessing both conformational control and diversity was a significant theme of the Hirschmann-Smith collaboration for over 15 years. Through these efforts, we learned that the combined effects of α-stereogenicity of the pyrrolinone ring, intramolecular hydrogen bonding, and choice of side-chains determined the global minimum energy conformation of the polypyrrolinone chain. Homochiral polypyrrolinones (eg., all D, Figure 1),¹ that preferentially adopt an extended conformation, proved to be excellent β-strand/β-sheet mimics,² and as such, led to potent, orally bioavailable pyrrolinone-based enzyme inhibitors of aspartic acid proteases,³ as well as modest metalloprotease inhibitors,⁴ and peptide-pyrrolinone hybrid ligands for the class II MHC protein HLA–DR1.⁵ Alternatively, heterochiral polypyrrolinones (e.g., alternating D,L pyrrolinones; Figure 1), much like heterochiral polypeptides, adopt a turn structure,⁶ and as such have been employed to generate functional β-turn mimetics.⁷ Subsequent investigations of the extended heterochiral pyrrolinone motif led to the discovery that hexapyrrolinone (−)-1 adopts a flat G-shaped conformation that aggregates in solution and in the solid state self-assembles into a nanotube-like stucture.⁸

Figure 1.

(a) Homochiral (DDD) and Heterochiral Pyrrolinones (LDL); (b) Structure of D,L-Hexapyrrolinone (−)-1.

The nanotube-like architecture of (−)-1 in the solid-state, possessing termini in close proximity, readily suggested the design of macrocyclic hexapyrrolinones 2 and 3 (Figure 2a). Unencumbered with terminal substituents, we reasoned that such cyclic polypyrolinones might self-assemble into nanotubes.⁹ Pleasingly, Monte Carlo conformational searches¹⁰ for 2 predicted that the low energy conformations would possess a flat, hexagonal conformation (Figure 2b), in agreement with previous structural analysis of the acyclic heterochiral pyrrolinones such as (−)-1.

Figure 2.

(a) Prospective macrocyclic hexapyrrolinones 2 and 3; b) Stereoview of the lowest energy conformation of 2 derived via Monte Carlo conformational analysis.

Importantly, the predicted conformation presents hydrogen bonding acceptors and donors (cf. C=O and N-H, respectively) in an alternating pattern directed above and below the plane of the molecule, thus providing the potential for intermolecular hydrogen bonding in a nanotube-like array.

To access 2, we initially employed our iterative polypyrrolinone synthetic tactic in a linear fashion,2,6 beginning with the C terminus to generate the open-chain pentamer (−)-10. Although this approach to (+)-2 eventually proved successful (Supporting Information), we subsequently designed a more effective, convergent synthesis, beginning with (+)-4¹¹ and (−)-5 (Scheme 1).¹² Condensation to afford an intermediate imine, followed by treatment with KHMDS generated monopyrrolinone (+)-6, a common precursor for both (+)-7 and (−)-8. Hydrogenolysis furnished amine (+)-7, while treatment with LiBF₄ led to aldehyde (−)-8. Union of these two pyrrolinone building blocks was achieved in 82% yield by imine formation, followed by treatment with KHMDS. Acetal hydrolysis furnished trispyrrolinone (−)-9; a two-step sequence with pyrrolinone amine (+)-7 then delivered the pentapyrrolinone (−)-10. The critical final pyrrolinone ring construction, leading to macrocycle (+)-2 was achieved in a similar fashion, albeit in this case the yield was at best modest (ca. 12–13%). Notwithstanding the efficiency of the final cyclization, a sample (ca. 100 mg) of (+)-2 was prepared for structural analysis.

Scheme 1.

Convergent Synthesis of Macrocyclic Hexapyrrolinone (+)-2.

Assignment of structure (+)-2 was based principally on simplification of both the ¹H and ¹³C NMR spectra, in conjunction with HRMS identification of the parent ion. Pentapyrrolinone (−)-10 (an unsymmetrical molecule, Scheme 1) displays a distinct set of signals for the five chemically (and magnetically) different pyrrolinone units (e.g., vinyl and benzyl hydrogens, etc). Conversion to the cyclic C₃-symmetrical hexamer (+)-2 (Figure 3, Scheme 1) renders each benzyl and isobutyl pyrrolinone chemically and magnetically identical, resulting in isochronous NMR signals for the three monomeric units. Indeed, only two sets of signals are observed in both ¹H and ¹³C NMR spectra of (+)-2, corresponding to the two types of pyrrolinone rings.

Figure 3.

Concentration dependence of ¹H NMR chemical shifts in CDCl₃ of N-H(1) (□) and NH(2) (◆) protons of (+)-2.

The propensity of macrocycle (+)-2 to self-assemble in solution was demonstrated via a series of ¹HNMR studies in CDCl₃ similarly employed for in the study of (−)-1.^8,13 The cyclic structure of (+)-2 permits each N-H of the individual macrocycles to be involved in intramolecular hydrogen bonding, thereby lessening their solvent exposure, and thus a relatively small effect of concentration was observed on the chemical shifts.⁶ Nonetheless, a concentration dependence of the N-H protons was observed (Figure 3) despite the intramolecular interactions. The measurable down-field shift for the N-H signals with increasing concentration is consistent with aggregation mediated by intermolecular hydrogen bonding.

Single crystal X-ray analysis of (+)-2 would provide the strongest evidence possible for nanotube assembly; unfortunately, crystals of (+)-2 suitable for X-ray analysis have not, as yet, been forthcoming. The failure to obtain suitable crystals of (+)-2 prompted the synthesis of hexapyrrolinone 3, wherein the isobutyl groups were substituted for isopropyl groups, with the expectation that the reduced flexibility of the isopropyl side-chains might facilitate crystal growth.

The synthesis of (+)-3 was completed in a convergent fashion, similar to that employed for (+)-2 (see Supporting Information). Not surprisingly, the macrocyclization step proceeded in even lower yield given the steric encumberance of the isopropyl side-chains (ca. 2% yield). Concentration dependent NMR analysis of (+)-3 revealed nearly identical solution-state aggregation as observed for (+)-2 (see Supporting Information). Gratifyingly, compared to (+)-2, superior crystals of (+)-3 could be obtained by slow evaporation in 1:1 CHCl₃:trifluoro- ethanol, and the solid-state structure analysis using synchrotron X-ray diffraction was achieved.¹⁴ Interestingly, unlike the open chain hexapyrrolinone (−)-1, (+)-3 assembles into a staggered nanotube-like array (Figure 4). Comparison of this structure, with that of (−)-1,8 provides both interesting similarities and differences. The monomers of (+)-3 assemble in an antiparallel fashion, as observed for (−)-1. Alternatively, the staggered array adopted by (+)-3, possesses only four intermolecular pyrrolinone-pyrrolinone hydrogen bonds between each pair of monomers, compared to six for (−)-1. Additionally, the staggered nanotube structure of (+)-3 forms an infinite array in the crystal lattice (four molecules/unit cell as illustrated in Figure 4b–d). Given the pyrrolinone scaffold comprises a designed peptidomimetic, our work has obvious parallels to the cyclic peptides pioneered by Karle,¹⁵ Lorenzi,¹⁶ and Ghadiri.^17,9 Of particular interest, the 1987 report by G. P. Lorenzi et al. discloses a series of D,L-alternating cyclic hexapeptides.¹⁶ In contrast to the hexapeptides, which were reported to be completely insoluble in all common organic solvents, (+)-2 and (+)-3 are moderately soluble in most polar organic solvents (i.e., EtOAc, EtOH), and display excellent solubility in CHCl₃. Importantly the structure of cyclic hexapyrrolinone (+)-3 overlays remarkably well with the Lorenzi cyclic hexapeptides (Figure 5), illustrating the close correspondence between the pyrrolinone and peptide units, and suggesting the possibility that the two macrocycles would form heterogenous aggregates.

Figure 4.

The X-ray structure of (+)-3: (a) a single molecule; (b) representative stereoview of the infinite staggered nanotube array, viewed from the side. The staggered nanotube structure is more clearly visible with the side-chains removed, from the top (c), and from the side (d).

Figure 5.

Overlay of the X-ray structure of cyclic hexapyrrolinone (+)-3 (blue) with the X-ray structure of cyclo(-D-Leu-L-MeLeu-D-Leu-L-MeLeu-D-Leu-L-MeLeu-) (red).¹⁶

In summary, the design, syntheses and structural analysis of macrocyclic D,L-hexapyrrolinones (+)-2 and (+)-3 have been achieved. Studies by NMR suggest that the macrocyclic hexapyrrolinones aggregate in solution, while the solid-state structure of (+)-3, determined via X-ray crystallography, revealed an extended, staggered nanotube-like array, stabilized by four intermolecular hydrogen bonds between pairs of pyrrolinone rings. Importantly, the solid state structure of (+)-3 demonstrates the ability of macrocyclic pyrrolinones to assemble into nanotube-like structures. Studies to both improve the final macrocyclization as well as to exploit the structural properties of such cyclic heterochiral polypyrrolinones continue in our laboratory.