1028
J. Comb. Chem. 2007, 9, 1028–1035
Skeletal Diversity in Small-Molecule Synthesis Using Ligand-Controlled Catalysis B. Lawrence Gray and Stuart L. Schreiber* Howard Hughes Medical Institute, Broad Institute of HarVard and MIT, Department of Chemistry and Chemical Biology, HarVard UniVersity, Cambridge, Massachusetts ReceiVed June 25, 2007 Two Pd-catalyzed reductive transformations of diynes tethered through a silyl ether linkage were developed, where the reaction outcomes were controlled solely by selection of phosphine ligand. We screened Pd precatalysts, ligands, and additives to optimize conditions selective either for reductive cyclization or hydrogenation of this substrate class. Sixteen silyl ether-tethered diynes were prepared and subjected to the best catalyst/ligand combinations for each pathway. Silacyclic dienes and silyl-tethered enyne products of these reactions were elaborated to densely substituted, stereochemically- and appendage-rich, bicyclic and tricyclic small molecules in 1–3 synthetic steps. These studies illustrate how small modifications to a transitionmetal catalyst can be used to access a diverse set of small molecules, in a fashion analogous to biosynthetic pathways such as terpene biosynthesis, where minor changes to enzyme structure direct skeletal differentiation. Introduction In biosynthetic pathways, for example in terpene biosynthesis,1 it is common to find a single biosynthetic intermediate converted to numerous small molecules having distinct skeletons. Here, minor differences in enzyme structure direct branching of common, polyunsaturated precursors to distinct skeletons (Figure 1, path a).2 Analogous pathways in diversity-oriented synthesis (DOS)3 would use minimal perturbations to a catalyst,4 rather than wholesale modification of reagent sets,5 to influence reaction outcome. Along these lines, we hypothesized that polyunsaturated compounds (II) might undergo different reductive transformations controlled by minor modifications of a transition-metal catalyst. Whereas metal-mediated transformations of alkyl-, ether-, and ester-tethered diynes have been used frequently in organic synthesis,6 similar reactions involving alkynes tethered through silyl ethers have been less well studied.7 For example, to date no reductive cyclization to produce silylcyclic dienes of type IV has been described. Herein, we report a catalyst-controlled branching pathway from diynes II that, coupled to complexity-generating steps, allows the rapid assembly of bicyclic (VII, VIII, and X) and tricyclic small molecules (IX), which are both suited for highthroughput, small-molecule screening8 (Figure 1, path b) and poised for post-screening optimization.5 Results and Discussion Synthesis of a Collection of Silyl Ether-Tethered Diynes. Silyl ether-tethered diynes II contain four discrete appending sites for incorporation of building-block diversity, and are constructed efficiently in three steps,9 rendering them well suited as branching intermediates for DOS pathways. * Corresponding author e-mail:
[email protected].
The synthesis of a collection of diynes commenced with alkynylsilane formation using alkynes 1–6 and either diisopropylchlorosilane 7 or diphenylchlorosilane 8 (Table 1). Treatment with N-bromosuccinimide resulted in the synthesis of the corresponding silyl bromides, which were converted to ethers of chiral propargyl alcohols 9–12 by reaction with triethylamine and catalytic DMAP. Terminal acetylenes bearing linear (entries 1–5 and 18–20), branched alkyl (entries 13–16), and aromatic groups (entries 6–12) were suitable substrates. Silyl etherification of chiral secondary alcohols bearing alkyl (entries 2, 3, 5, 8, 10, 11, 14, 15, 19), aryl (entries 1, 6, 13, 18), and mixed aryl ethers (entries 4, 9, 12, 16) also worked well. Development of a Ligand-Controlled Differentiation Process. In initial studies, it was found that treatment of 16 with Pd2dba3-CHCl3 adduct 33 (2.5 mol %), tri-o-tolylphosphine 38 (10 mol %), acetic acid, and excess triethylsilane in toluene at 23 °C10 gave an inseparable mixture of products 50 (via reductive cyclization) and 51 (via reduction of the terminal alkyne), along with cyclodimerization adduct 52 (>95% conv, 2.4:1 50:51, 38% 52).11 Reduction of the reaction temperature during formation of the catalyst increased the combined conversion and yield of 50 and 51 (10 °C: >95% conv, 1.9:1 50:51, 8% 52; –20 °C: >95% conv, 2.7:1 50:51, 9% 52). Conversely, the temperature at which substrate was introduced to the preformed catalyst was determined to have little effect on reaction efficacy (e.g., 23 °C: >95% conv, 2.7:1 50:51, 9% 52; 10 °C: >95% conv, 2.2:1 50:51, 8% 52; –20 °C: >95% conv, 1.6:1 50:51, 8% 52, when catalyst was formed at -20 °C). However, failure to increase the temperature to 23 °C after substrate addition resulted in incomplete reactions, despite improved ratios of diene to enyne (e.g., 0 °C: 14% conv, 3.2:1 50:51; -2 0 °C: 11% conv, 3.6:1 50:51; -7 8 °C: 18% conv, 3.0:1 50:51). Adopting a standard set of conditions from this early set of
10.1021/cc7001028 CCC: $37.00 2007 American Chemical Society Published on Web 10/26/2007
Skeletal Diversity in Small-Molecule Synthesis
Journal of Combinatorial Chemistry, 2007, Vol. 9, No. 06 1029
Figure 1. Examples of catalyst-based control of skeletal formation in cells and in the laboratory. (a) Terpene diversity generated by the actions of different enzyme cyclases on farnesyl pyrophosphate. (b) Small-molecule diversity initiated by different Pd/ligand combinations.
data (see Supporting Information for full details), a systematic variation of Pd(0) and Pd(II) precatalysts, ligands, and hydride sources was undertaken in an effort to discover active catalysts selective for either reductive cyclization or enyne formation (Table 2).12 Several notable observations emerged from these screens. Both Pd(0) and Pd(II) complexes constitute competent precatalysts for this reaction; however, a change in metal precatalyst was not sufficient to induce a crossover in product selectivity to 51 (entries 1–5: 1.6:1–2.7:1 50:51).13 On the contrary, selection of the ligand was found to play a predominant role in tuning the reaction pathway,14 with trio-tolylphosphine 38 best for reductive cyclization (entry 2, 2.7:1 50:51), and monovalent, bulky biaryl phosphines such as X-PHOS 41, Cy2P-biphen 46, and t-Bu2P-biphen 48 were found best for selective hydrogenation (entries 10, 15, and 17, 1:3.7–1:6.8 50:51). Additionally, use of the polymeric PMHS as reductant resulted in very high selectivity for the enyne product (entry 6, 1:11.2 50:51); however, this additive proved difficult to separate from the product chromatographically and thus was not further pursued. (R)-MeO-MOP 42, a monodentate ligand bearing a phosphine and methoxy group, was also shown to be selective for 51 (entry 11).15 However, bisphosphines such as dppe 43 and dppb 44 abrogated catalytic activity (entries 12–13).16
Scope and Limitations of Pd2dba3-Catalyzed Reductive Transformations with Silyl Ether-Tethered Diynes. The generality of this ligand-controlled process was assayed next across a range of silyl ether-tethered diyne substrates (Table 3). The catalyst system preformed from Pd2dba3 34 and trio-tolylphosphine 38 and identified as the optimal system for reductive cyclization of 16 displayed good generality (10 of 16 cases tested). Sterics at the silylated alkyne proved important, with product ratios highest for compounds containing i-Pr (entries 11–14, 1.6–8.3:1 IV:III) and Ph(CH2)2 (entries 15–16; 4.5–7.1:1 IV:III). n-Hexyl substituted diynes underwent selective cycloreduction in four of five cases (entries 1, 2, 4, 5, 1.8–3.2:1 IV:III). The nature of substitution at silicon did not affect selectivity (compare entries 1 and 5; 1.8:1 vs. 1.9:1 IV:III). Similiarly, the effect of the secondary alcohol building block on selectivity was usually not strong enough to override catalyst preference (e.g., entries 11–14; 1.6–8.3:1 IV:III). Aromatic-substituted diynes, on the contrary, were not routinely diene selective, usually yielding enyne as the major adduct and often at incomplete conversions (entries 6–10; 1:1.5–4.4 IV:III).17 Additionally, with 15, discrimination between the two internal acetylenes was not achieved. In contrast to the above results, subjection to the complex derived from Pd2dba3 34 and X-PHOS 41 resulted in an enyne-selective reaction in 12 of 16 cases. Many of the trends
1030 Journal of Combinatorial Chemistry, 2007, Vol. 9, No. 06
Gray and Schreiber
Table 1. Synthesis of Silyl Ether-Tethered Diyne Substrates 13–32
entry
alkynea
silaneb
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1(R )C6H13) 1(R1)C6H13) 1(R1)C6H13) 1(R1)C6H13) 1(R1)C6H13) 2(R1)Ph) 2(R1)Ph) 2(R1)Ph) 2(R1)Ph) 2(R1)Ph) 3(R1)o-tol) 3(R1)o-tol) 4(R1)i-Pr) 4(R1)i-Pr) 4(R1)i-Pr) 4(R1)i-Pr) 5 6(R1)PhC2H4) 6(R1)PhC2H4) 6(R1)PhC2H4)
7(R )i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 8(R2)Ph) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 8(R2)Ph) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 7(R2)i-Pr) 8(R2)Ph)
1
2
yield (%)c 97 97 97 97 94 97 97 97 97 87 93 93 85 85 85 85 86 86 86 83
alcohold 9(R )Ph; R )H) 10(R3)C5H11; R4)H) 11(R3)C2H5; R4)CH3) 12(R4)H) 10(R3)C5H11; R4)H) 9(R3)Ph; R4)H) 10(R3)C5H11; R4)H) 11(R3)C2H5; R4)CH3) 12(R4)H) 10(R3)C5H11; R4)H) 10(R3)C5H11; R4)H) 12(R4)H) 9(R3)Ph; R4)H) 10(R3)C5H11; R4)H) 11(R3)C2H5; R4)CH3) 12(R4)H) 10(R3)C5H11; R4)H) 9(R3)Ph; R4)H) 10(R3)C5H11; R4)H) 10(R3)C5H11; R4)H) 3
4
yield (%)e
product
63 91 71 66 74 88 95 >95 95 >95 95 >95 70 >95 95 >95 54 >95
8 9 12
