Synthesis of Structurally Diverse N-Substituted Quaternary-Carbon-Containing Small Molecules from α,α-Disubstituted Propargyl Amino Esters

Abstract

N-containing quaternary stereocenters represent important motifs in medicinal chemistry. However, due to their inherently sterically hindered nature, they remain underrepresented in small molecule screening collections. As such, the development of synthetic routes to generate small molecules that incorporate this particular feature are highly desirable. Herein, we describe the diversity-oriented synthesis (DOS) of a diverse collection of structurally distinct small molecules featuring this three-dimensional (3D) motif. The subsequent derivatisation and the stereoselective synthesis exemplified the versatility of this strategy for drug discovery and library enrichment. Chemoinformatic analysis revealed the enhanced sp³ character of the target library and demonstrated that it represents an attractive collection of biologically diverse small molecules with high scaffold diversity.

Keywords: diversity-oriented synthesis; drug discovery; medicinal chemistry; molecular diversity; quaternary stereocenters.

Figure 1

Examples of bioactive compounds containing a N‐substituted quaternary carbon.

Figure 2

Synthetic versatility of α,α‐disubstituted amino esters for the DOS of N‐substituted quaternary carbon containing small molecules.

Scheme 1

Synthesis of structurally diverse N‐substituted quaternary carbon containing small molecules from α,α‐disubstituted amino ester 1: a) amine derivatisation; b) pairing of the amino and ester moieties; c) ester derivatisation; d) pairing of the alkyne and ester moieties; e) alkyne derivatisations; f) pairing of the amino and alkyne moieties. For reaction conditions see Supporting Information.

Figure 3

(A) Exemplification of the synthesis of other derivatives via alkyne modifications. Reaction conditions: 19 a, 21 a, 22 a, 26 a: azidobenzene, CuSO₄⋅5 H₂O, tBuOH:H₂O, rt, 72–92 %; 21 b and 22 b: PdCl₂(PPh₃)₂, CuI, Et₃N, benzyl 2‐iodobenzoate, DMF, rt, 16 h, 60 % and 26 %, respectively; 21 c and 22 c: Pd/C, H₂, MeOH, rt, 2 h, 69 % and 87 %, respectively; 23 d: ethyl 2‐azidoacetate, Cp*RuCl(COD), PhMe, rt −60 °C, 21 h, 42 % (B) Synthesis of the phenyl‐containing derivative. Reaction conditions: i) Boc₂O, THF, 70 °C, o/n, 84 %; ii) ethyl 2‐azidoacetate, Cp*RuCl(COD), PhMe, 80 °C, 1 h, 87 %; iii) TFA, CH₂Cl_2, 2 h then NaHCO₃, (quantitative yield); iv) PhMe, 150 °C, o/n, 76 %. (C) Synthesis of an enantiopure component of the library. Reaction conditions: i) Boc₂O, THF, 70 °C, o/n, 90 %; ii) ethyl 2‐azidoacetate, Cp*RuCl(COD), PhMe, 50 °C, 1 h, 76 %; iii) TFA, CH₂Cl₂, 2 h then NaHCO₃, 93 %; iv) PhMe, 150 °C, o/n, 79 %.

Figure 4

Comparative PMI plot analysis of DOS library (blue circles), Maybridge core 1000‐member “Ro3” Fragment library (green triangles) and the DOS Ph virtual library (red squares). Compounds within “flatland” (represented by npr1+npr2 value <1.1, dashed line) could also be identified; the further from this area the molecules move the more they extend into three‐dimensional space.36.

Figure 5

Bioactive space coverage of the DOS library, compared to two targeted compound libraries and seven different bioactivity classes of common protein targets obtained from ChEMBL (based on Morgan fingerprints and Multi‐Dimensional Scaling, MDS; ellipses cover 90 % of the data in each class. It can be seen that the DOS library displays significantly wider coverage in bioactivity space than the two targeted libraries, covering all common classes of drug targets used for comparison.