Synthesis of a Stereochemically Diverse Library of Medium-Sized Lactams and Sultams via S_NAr Cycloetherification

Abstract

We have implemented an aldol-based ‘build/couple/pair’ (B/C/P) strategy for the synthesis of stereochemically diverse 8-membered lactam and sultam scaffolds via S_NAr cycloetherification. Each scaffold contains two handles, an amine and aryl bromide, for solid-phase diversification via N-capping and Pd-mediated cross coupling. A sparse matrix design strategy that achieves the dual objective of controlling physicochemical properties and selecting diverse library members was implemented. The production of two 8000-membered libraries is discussed including a full analysis of library purity and property distribution. Library diversity was evaluated in comparison to the Molecular Library Small Molecule Repository (MLSMR) through the use of a multi-fusion similarity (MFS) map and principal component analysis (PCA).

Introduction

Diversity-oriented synthesis (DOS) is a commonly employed strategy for the facile assembly of structurally diverse molecules rivaling the complexity of natural products.¹ A primary goal of DOS is the generation of compounds with both skeletal and stereochemical diversity. Recently we described an aldol-based ‘build/couple/pair’ (B/C/P) strategy for the generation of a stereochemically diverse set of medium- and large-sized rings via a common linear template.² Here, we take advantage of this aldol-based strategy for the synthesis of fused pyridines (1) and fused sultams (2) (Figure 1). In the build phase, a series of asymmetric syn- and anti-aldol reactions were applied to produce four stereoisomers of a Boc protected β-hydroxy-γ-amino acid (3). Both stereoisomers of PMB-protected alaninol (4) were also obtained to complement the aldol-derived acids. In the couple step, all 8 stereoisomeric amides were synthesized from both the chiral acid and amine building blocks. The amide was subsequently reduced to generate a secondary amine (5). Finally, in the pair phase, we utilized an intramolecular S_NAr³ as the key cyclization step to access either the S_NAr-Pyr lactam (1)⁴ or the S_NAr-SO₂ sultam (2).⁵ This work is built off of our previous success with the S_NAr reaction for the synthesis of 8- and 9-membered lactams (6 and 7).² Two 8000-membered libraries were produced using solid-phase synthesis techniques. All 8 stereoisomers were prepared for each scaffold, providing not only structure-activity relationships (SAR) in primary screens, but also stereo/structure-activity relationships (SSAR). A sparse matrix library design strategy^6,7 was utilized to aid in the selection of diverse library members with built-in structural analogs and physicochemical properties suitable for high-throughput screening and downstream discovery.

Figure 1.

Synthesis of medium-sized ring scaffolds from a common linear intermediate.

Solution-phase synthesis of library scaffolds

The synthesis of S_NAr-Pyr scaffold 1 began with acylation of linear amine 5 using 5-bromo-2-chloronicotinoyl chloride 8, which afforded amide 9a–d in good yields (Table 1). The subsequent intramolecular S_NAr reaction showed strong stereochemical dependence. Amide 9a was converted directly to 10a in excellent yield (94%) upon treatment with TBAF in THF at 65 °C. However, application of these conditions to 9b provided lactam 10b in modest yield (71%). Fortunately, a two step protocol involving cyclization using NaH in THF once the TBS group had been deprotected with CsF proved effective. The syn-aldol derived substrates 9a and 9b were converted to 10a and 10b respectively in high yields (96–98%) without a need for chromatographic purification using this two step protocol. Meanwhile, S_NAr reaction of anti aldol-derived substrates 9c and 9d proved challenging, as the two step protocol led to formation of significant amount of oxazolidinone side-product (40–50%).⁸ Use of the one step deprotection/cyclization using TBAF in THF led to incomplete reaction even after 5 days and repeated addition of TBAF. Finally, choice of the solvent proved critical and use of TBAF in DMF led to complete conversion of the S_NAr reaction with minimal oxazolidinone formation (10–15%). Under these conditions, the anti-aldol derived substrates 9c and 9d gave 84% and 80% of 10c and 10d, respectively. As our plan called for loading onto solid support (SynPhase™Lanterns)⁹ using an acid-labile silicon linker, the Boc protecting group was exchanged for Fmoc, and subsequent DDQ mediated PMB removal led to the isolation of primary alcohol 1 (Scheme 1). S_NAr-pyr scaffolds 1a–d (and the corresponding enantiomers ent-1a–d) were prepared in 15–20 g quantities using this 5-step sequence starting from linear amine 5.

Table 1.

S_NAr cyclization to form lactams 10a–d

S_NAr reaction conditions: Method A a) CsF (5 equiv), DMF, 85 °C; (b) NaH (5 equiv), THF, 0 °C to rt; Method B TBAF (5 equiv), DMF, 65 °C.

Scheme 1.

Preparation of final S_NAr-Pyr scaffold 1

A similar approach as described above was utilized in the synthesis of fused bicyclic sultams.⁵ All stereoisomers of the common linear intermediate 5 were coupled with 4-bromo-2-fluorobenzene sulfonyl chloride 12 in good yield to give the S_NAr precursor 13 (Table 2). Similar to the pyridine fused systems, the S_NAr reaction involving sulfonamides 13 proceeded differently depending on the relative stereochemistry of the adjacent stereocenters. Sulfonamides 13a and 13b, both derived from the syn-aldol reaction, were easily converted in excellent yield to sultams 14a and 14b, respectively, by treatment with CsF in DMF at 85 °C. Sulfonamides 13c and 13d, both derived from the anti-aldol, were converted to sultams 14c and 14d utilizing the two-step approach that was employed for substrates 9a and 9b. First, treatment with CsF gave a mixture of uncyclized TBS deprotected material along with desired 14. Treatment of the mixture with NaH gave complete conversion to 14c and 14d, respectively, in good yield over the two steps. It is important to note, no matter the protocol used, the product could be isolated in sufficient purity without silica gel purification. This is in contrast to the S_NAr-Pyr substrates, where the use of TBAF required silica gel purification. Completing the synthetic sequence required the exchange of the Boc group for Fmoc followed by DDQ-mediated PMB removal to afford the desired S_NAr-SO₂ scaffolds 2a–d (and the corresponding enantiomers ent-2a–d) in good yield in 15–20 g quantities (Scheme 2).

Table 2.

S_NAr cyclization to form sultams 14a–d

S_NAr reaction conditions: Method A CsF (5 equiv), DMF, 85 °C; Method B a) CsF (5 equiv), DMF, 85 °C; (b) NaH (1 equiv), THF, 0 °C to rt.

Scheme 2.

Preparation of final S_NAr-SO₂ scaffold 2

Library Design

With the S_NAr-Pyr and S_NAr-SO₂ scaffolds in hand, a sparse matrix design strategy was implemented to select library members to be synthesized.⁶ A virtual library was constructed for each scaffold incorporating all possible building block combinations at R₁ (amine) and R₂ (aryl bromide) using a master list of reagents (R₁ = sulfonyl chlorides, isocyanates, acids and aldehydes; R₂ = boronic acids and alkynes). Physicochemical property filters were then applied to eliminate building block combinations that led to products with undesirable physicochemical properties. Property filters included the following: MW ≤625, ALogP -1 to 5, H-bond acceptors and donors ≤10, rotatable bonds ≤10 and TPSA ≤140. In order to increase the percentage of ‘Lipinski compliant’ products, a ‘75/25’ rule was also implemented where 75% of all library members had MW <500. A total of 1000 compounds per scaffold were selected from the remaining set using chemical similarity principles, maximizing diversity but retaining near neighbors for built-in SAR. The reagents selected for library production are shown below (Chart 1 and Chart 2). The same set of reagents was used for each stereoisomer thereby maintaining the ability to generate SSAR for each building block combination.

Chart 1.

Building blocks {1–72} used for amine capping at R₁.

Chart 2.

Building blocks {1–45} used for cross-coupling reactions at R₂.

Solid-phase Library Production

The construction of the S_NAr-Pyr and S_NAr-SO₂ libraries is outlined in Scheme 3. Scaffolds 1a–d and 2a–d (and the corresponding enantiomers) were loaded onto silicon-functionalized PS-SynPhase™ Lanterns (L-series) activated with TfOH in the presence of 2,6-lutidine (average loading level = 18 umol/Lantern).^9,10 The first diversity site, a secondary amine, was then revealed under standard conditions required for Fmoc removal (20% piperidine in DMF), and amines 16a–d and 17a–d were capped with the selected electrophiles (sulfonyl chlorides 1–11, isocyanates 12–26, carboxylic acids 27–50 and aldehydes 51–71) or skipped (72) to yield compounds 18a–d{1–72} and 19a–d{1–72}. Next, Sonogashira and Suzuki cross-coupling reactions were carried out to introduce appendage diversity at R₂. For the Sonogashira reaction, the Lantern-bound aryl bromides were heated at 60 °C in DMF overnight in the presence of DIEA, CuI, Pd(PPh₃)₂Cl₂, and the selected alkynes {1–24}. Meanwhile, for the Suzuki reaction, the Lanterns were heated at 60 °C in EtOH for 5 days in the presence of Et₃N, Pd(PPh₃)₂Cl₂ and boronic acids {25–44}. Removal of residual Pd and Cu was achieved by washing the Lanterns with 0.1 M NaCN. Finally, cleavage with HF-pyridine in THF afforded library members 20a–d{1–72,1–45} and 21a–d{1–72,1–45} with an average yield of 90 and 91% respectively.¹¹

Scheme 3.

Solid-phase synthesis of S_NAr-Pyr and S_NAr-SO₂ libraries on SynPhase™ Lanterns

All library products were analyzed by ultra-performance liquid chromatography, and compound purity was assessed by UV detection at 210 nm. An overview of compound purity for the S_NAr-Pyr and S_NAr-SO₂ libraries with respect to building blocks and stereochemistry is provided in Figures 2 and 3. The average purity of the S_NAr-Pyr library was 85%, with 89% of the library being >75% pure, while the average purity of the S_NAr-SO₂ library was 85%, with 91% of the library being >75% pure. (See Figures S5 and S6). In general, all building blocks performed well during the library production with the exception of certain reagent combinations. For example, compounds containing the dimethylisoxazole urea (18a–d{25} and 19a–d{25}) performed poorly in the subsequent Suzuki reaction, presumably due to reduction of the N-O bond, as an M+2 impurity was observed by LCMS for this reagent combination. Meanwhile products of reductive alkylation with 3-pyridyl benzaldehyde 18a–d{68} and 19a–d{68} performed poorly in the subsequent Sonogashira reaction. The success of the cross-coupling reaction in the presence of a free amine at R₁ to produce library members 20a–d{72,1–44} and 21a–d{72,1–44} was highly variable depending on the nature of the boronic acid and alkyne but in general was problematic. Surprisingly, the use of acetaldehyde for reductive alkylation at R₁ resulted in compounds (20a–d{2,1–72} and 21a–d{2,1–72}) of low purity for both libraries.

Figure 2.

Purity analysis (UV 210 nm) for S_NAr-Pyr Library. Library members are displayed as blocks of 8 stereoisomers (see legend) and reagents used for solid-phase diversification are shown on the x- and y-axes. (See Charts 1 and 2 for detailed list of reagents).¹²

Figure 3.

Purity analysis (UV 210 nm) for S_NAr-SO₂ Library. Library members are displayed as blocks of 8 stereoisomers (see legend) and reagents used for solid-phase diversification are shown on the x- and y-axes. (See Charts 1 and 2 for detailed list of reagents).¹²

Library Analysis

The S_NAr-Pyr and S_NAr-SO₂ libraries originate from the same linear intermediate (5, Figure 1) and vary in the pairing stage giving rise to different molecular architectures. The S_NAr-Pyr and S_NAr-SO₂ scaffolds have differences in their physicochemical properties such as molecular weight (372/407), ALogP (1.2/1.5) and TPSA, (75/87) that ultimately influence product selection (Table 3). As evident in Figures 2 and 3 the selected building block combinations vary significantly for these two scaffolds. For example, products are spread evenly in the S_NAr-Pyr library (Figure 2), as compared to the S_NAr-SO₂ library (Figure 3) for which small aliphatic building block combinations are favored (e.g., acids 27–34 and aldehydes 51–54). Analysis of the S_NAr-Pyr and S_NAr-SO₂ libraries reveals that the property profile for each library was within the intended range for the library design (MW ≤625, ALogP -1 to 5, H-bond acceptors and donors ≤10, rotatable bonds ≤10 and TPSA ≤140). Not surprisingly, S_NAr-SO₂ library members have higher mean values for MW and TPSA due to inherent differences between the initial scaffolds.

Table 3.

Property analysis for S_NAr-Pyr and S_NAr-SO₂ libraries

Property	SNAr-Pyr Scaffolda (n = 1)	SNAr-SO2 Scaffolda (n = 1)	SNAr-Pyr Libraryb (n = 7045)	SNAr-SO2 Libraryb (n = 6690)
MW	372	407	484	504
ALogP	1.2	1.5	2.7	2.9
TPSA	75	87	95	105
Rotatable Bonds	4	4	7.3	7.2
HBA	5	5	6.1	6.0
HBD	2	2	1.3	1.3

^aProperty analysis of bare scaffolds, where R₁ and R₂ = H.

^bProperty analysis (mean value) of all registered library samples passing QC requirements (purity >75%).

The structural diversity of library members resulting from S_NAr-Pyr and S_NAr-SO₂ pathways was analyzed in comparison to the NIH Molecular Library Small Molecule Repository (MLSMR) as we intended to submit a subset of these compounds to the collection at the time of the analysis. We employed multi-fusion similarity (MFS) maps for the comparison of each collection using extended connectivity fingerprints (ECFP_4) for molecular representation and Tanimoto coefficient as the similarity measure.¹³ In this method, each molecule in the test set (S_NAr-Pyr and S_NAr-SO₂ library members) is compared to every molecule in the reference set (MLSMR) and the largest similarity score and the mean similarity score to the reference set is obtained. The resulting mean similarity (X-axis) and maximum similarity (Y-axis) values are plotted in two dimensions as a scatter plot facilitating the visual characterization and comparison. Figure 4 shows the MFS map comparing S_NAr-Pyr and S_NAr-SO₂ libraries to the MLSMR. Each data point in the map depicts a compound from the test set and its location was influenced by the reference set. (The reference compounds themselves do not appear in the plot.) The maximum mean similarity of each library is 0.15 indicative of their overall structural diversity with respect to the MLSMR reference set. There are no compounds with maximum similarity equal to or greater than 0.45 in the MLSMR, which clearly illustrates the regions of chemical space unexplored by the MLSMR.

Figure 4.

Multi-fusion similarity map comparing the S_NAr-Pyr (red) and S_NAr-SO₂ (blue) libraries to the MLSMR. The reference set (MLSMR) is not shown on the map (see text for details).

We also carried out a principal components analysis (PCA)¹⁴ using 16 structural and physicochemical descriptors (including MW, ALogP, rotatable bonds and TPSA) for the MLSMR and S_NAr-Pyr and S_NAr-SO₂ libraries. The PCA plot is shown in Figure 5. The DOS libraries property space is embedded in the MLSMR property space with ~75% of compounds being Lipinski compliant. This observation is particularly important because it dispels the common notion that DOS may not render compounds with good properties.¹⁵ While occupying the desirable property space we have covered new chemical space of structural diversity.

Figure 5.

Principal component analysis of S_NAr-Pyr and S_NAr-SO₂ libraries compared to MLSMR. The 16 molecular descriptors used for PCA included molecular weight, ALogP, LogD, number of rotatable bonds, TPSA, number of nitrogen, oxygen, halogen, and sulfur atoms, hydrogen bond donors and acceptors, aromatic rings, ring fusion degree, normalized number of ring systems, chiral centers, and fraction of sp³.

Conclusion

In summary we have implemented an aldol-based ‘build/couple/pair’ (B/C/P) strategy for synthesis of stereochemically diverse 8-membered lactam and sultam scaffolds via S_NAr cycloetherification. A sparse matrix design strategy was implemented to select library members for synthesis to achieve a balance between diversity and built-in structural analogs. Analysis of final library members illustrates that the sparse matrix design achieved the intended outcome of structural diversity and favorable physicochemical properties. Screening of these compounds is currently underway in multiple biochemical and cell-based assays.