1
Bridged Lactams as Model Systems for Amidic Distortion

Tyler J. Fulton, Yun E. Du, and Brian M. Stoltz

California Institute of Technology, Division of Chemistry and Chemical Engineering, Pasadena, California, 91125, United States

1.1 Introduction and Scope

Amide linkages are pervasive functional groups present in peptides, natural products, and bioactive materials [1]. The high degree of resonance contribution of the amide nitrogen nN to p*C=O (15-20 kcal/mol stabilization from resonance form 2, Scheme 1.1) imparts approximately 40% double bond character and renders typical amides planar and less reactive toward nucleophilic attack compared with other carbonyls and carboxylic acid derivatives (Scheme 1.1) [2]. Disruption of the nN to p*C=O resonance via distortion of the amide bond is a mode of amide activation with a rich history dating back to the l938 when Lukes first proposed that incorporation of an amide nitrogen into the bridgehead position of a bicyclic system would be "sterically impossible" and violate Bredt's rule [3]. Lukes further hypothesized that if amides bearing a bridgehead nitrogen were successfully synthesized, they would bear properties more akin to those of ketones. Bridged lactams have since served as a remarkable tool for the understanding of the amide bond and its properties. This chapter provides a review of bridged lactams up to 2020. A comprehensive discussion of bridged lactams is outside of the scope of this entry; therefore, the present chapter is focused on the reactivity and properties of bridged lactams as models for amidic distortion. The most recent comprehensive review of bridged lactams was published by Szostak and Aubé in 2013 [4], with an update of the field covering the period of 2014-2018 from Szostak [5]. For a full accounting of the synthesis of bridged lactams, we direct the reader to these recent reviews [4, 5].

The nomenclature and classifications of twisted amides were originally delineated by Yamada [6] and have become standard for discussion within the field [7]. Destabilization of amide resonance 2 can be affected by several general means: intramolecular steric repulsion, intramolecular steric restriction, intermolecular steric restriction, electronic delocalization (e.g. classical anomeric amides), and conformational effects (Figure 1.1) [4]. It is important to note that destabilized amide is a more general term, whereas nonplanar amide refers specifically to an amide that has been geometrically destabilized via steric and/or electronic factors, i.e. the contribution of resonance form 2 has been diminished. Qualitatively, nonplanar amides can be divided into Types A, B, and C as described by Yamada [6]. Type A amides contain a perpendicularly twisted N-C(O) bond with a nonpyramidalized trigonal geometry at nitrogen. Type B amides contain a significant pyramidalization of the nitrogen atom with a planar N-C(O) bond geometry. Type C amides are a combination of Type A and Type B wherein the N-C(O) bond is twisted with a pyramidalized nitrogen. Both Type A and Type C amides are known as twisted amides, while Type B amides are known as pyramidalized amides.

Scheme 1.1 Resonance structures of the amide bond.

Figure 1.1 Amidic distortion and types of nonplanar amides.

Figure 1.2 Types of bridged lactams.

Bridged lactams represent one specific class of nonplanar twisted amides wherein the amide nitrogen is embedded in the ring fusion of a bridged bicyclic scaffold, consequently resonance form 2 constitutes an anti-Bredt double bond in a bridged lactam. Steric restriction of the amide bond in such a way remains the most successful strategy by far for distortion of the amide bond. Bridged lactams therefore constitute privileged chemical motifs for understanding the structure and reactivity of distorted amide bonds in general [8]. In addition to the insights into the nature of the amide bond, amidic distortion is critical for understanding cis/trans-isomerization in peptides for protein folding [9]. Bridged lactams are divided into Type I scaffolds where the N-C(O) bond is on a bridge consisting of two or more carbon atoms or Type II where the N-C(O) bond is on a single carbon bridge (Figure 1.2). Type II bridged lactams are generally more strained than the corresponding Type I bridged lactams; however, they tend to be more resistant to hydrolysis due to medium ring scaffolding effects [10]. Bridged lactams wherein the 1-azabicyclo core contains more than 10 atoms will not be covered as these structures display physical properties and reactivity akin to that of planar amides.

1.2 General Properties of Bridged Lactams

1.2.1 Parameters of Amide Bond Distortion

The Winkler-Dunitz distortion parameters provide a quantitative assessment of amide bond distortion based on the twist angle (t), N-pyramidalization (?N), and C-pyramidalization (?C) (Figure 1.3) [11]. The twist angle t describes the magnitude of rotation about the N-C(O) bond. Quantitatively, t is 0° for a planar amide bond and 90° for a fully orthogonal nonplanar amide bond. Both ?N and ?C describe the tetrahedral character on the nitrogen and carbon, respectively, and range from 0° for a fully planar amide to 60° for fully pyramidalized amide bonds. In 2015, Szostak identified the additive Winkler-Dunitz distortion parameter (t + ?N) as a means to more accurately predict structural and energetic properties [12]. The additive parameter is an extremely powerful tool for predicting the properties of distorted amides, corresponding linearly to N-C(O) bond length, N- vs. O-protonation affinity, infrared frequencies, resonance energies, atomic charges, and frontier molecular orbital energies [12]. Nearly the entire Winkler-Dunitz scale is represented by Type II bridged lactams prepared via synthetic efforts [5, 8]. Type I bridged lactams tend to be more flexible and thus do not span the entire range of the Winkler-Dunitz scale, limiting their usefulness in comprehensive understanding of distorted amides (Figure 1.3).

Figure 1.3 Winkler-Dunitz parameters for quantitative assessment of amide bond distortion.

It should be noted that ?C values for bridged lactams are generally close to 0° (even for relatively undistorted bridged lactams) due to the heightened contribution of amino-ketone resonance form 1 (Scheme 1.1, vide supra). At this stage, it is important to note polarized resonance contributor 3 [13], which corresponds to a nitrogen-to-carbon p-electron transfer with little oxygen resonance overlap. Polarized resonance contributor 3 describes the considerably larger N-C(O) bond lengthening compared with a considerably smaller C=O bond lengthening observed upon amide bond distortion [1, 14, 15]. According to the Wiberg-Bader theory of atoms in molecules (AIM), amide resonance stabilization is attributed to the hybridization change of nitrogen from sp3 to sp2 in planarizing an orthogonal amide bond. The validity of this polarized resonance model 3 has been a topic of considerable ongoing debate, with recent extensive computational modeling of Type II bridged lactam amide bond rotation demonstrating structural and energetic changes consistent with classical resonance form 1 [12].

An important and underappreciated aspect of amidic distortion is the bending of the amide carbonyl oxygen toward nitrogen. Notably, C=O bending is indicative of impending N-C(O) bond cleavage [16, 17]. In 2016, Stoltz and coworkers proposed the bending angle parameter ? to describe the deviation from the imaginary CCN angle bisector (Figure 1.4) [16]. The ? value is calculated by the equation shown in Figure 1.4, with a positive ? corresponding to bending toward nitrogen and a negative ? corresponding to bending toward carbon. The amide carbonyl bending effect was first discussed by Bürgi and Schmidt in 1985 in the context of crystallographic and molecular orbital calculations of lactones and lactams [17]. In these studies, an anomeric effect was proposed to reduce a destabilizing mixing of an oxygen p-type lone pair orbital with the C-C(O) bonding orbital with simultaneous stabilizing mixing of the oxygen p-type orbital with the C-N antibonding orbital. Reported values of ? deviate significantly from computational models [12, 16, 17]; therefore only ? values derived from X-ray crystal structures will be discussed. Bridged lactam crystal structures published prior to the formalization of the ? parameter will be presented with ? values determined from crystal structures available via the Cambridge Crystallographic Data Centre when possible, thus some of the values...