Benzyl and Phenylthiomethyl Silanes: A New Class of Bifunctional Linchpins for Type II Anion Relay Chemistry (ARC)

Abstract

A new class of bifunctional linchpins bearing electrophilic sites β or γ to a silyl group have been designed, synthesized and demonstrated to be competent in tri-component unions exploiting Type II Anion Relay Chemistry (ARC). High diastereoselectivities were observed when a phenyl moiety (R₂) served as an anion stabilizing group (ASG) adjacent to a methyl susbtituent (R₁), while diastereomeric mixtures were obtained when a phenylthiol moiety served as the ASG, irrespective of α-substitution.

Efficient union of architecturally complex fragments constitutes a critical prerequisite in natural product total syntheses. One such tactic, Anion Relay Chemistry (ARC), a one-pot tri-component union tactic, has proven highly effective in our Laboratory.¹ First exploited in 1997, ² two structurally different epoxides were united in a single flask to furnish advanced fragments for the total syntheses of (+)-spongistatin 1 and 2.³ This tactic was subsequently employed in conjuction with a formal total synthesis of (+)-mycoticin⁴ and more recently in the synthesis of the aglycone skeleton of (+)-rimocidin.⁵ Now termed Type I ARC (Scheme 1A),¹ this strategy was based on the union of two epoxides as introduced by Matsuda^6a and Tietze.^6b Hale and coworkers subsequently took advantage of this protocol, employing the Tietze homocoupling version in their formal total synthesis of (+)-bryostatin 1.⁷

Scheme 1.

Type I and II Anion Relay Chemistry (ARC)

In 2004 we extended the Type I process to permit anion migration along the linchpin chain to generate a new reactive anion at a distal site (Scheme 1B).⁸ Termed Type II ARC,^1,9 initial nucleophilic attack at the electrophilic site of the linchpin (cf. aldehyde or epoxide), leads to an alkoxide, which upon a silyl C(sp³)→O Brook rearrangement,^8,9 triggered by addition of a polar solvent (HMPA) or by temperature increase, produces a new carbon anion. For silyl migration to occur, the newly generated anionic site must possess a viable anion stabilizing group (ASG). Capture of the derived anion is then possible with a variety of electrophiles to generate in a single flask diverse three-component adducts. The synthetic utility of the Type II ARC protocol was demonstrated in the construction of an advanced intermediate in our prospective syntheses of (+)- spirastrellolides A and B, architecturally complex sponge metabolites.¹⁰

Early in the development of the ARC technique, the dithiane moiety played the role of a highly effective ASG, furnishing after Brook rearrangement, a powerful nucleophile for reaction with a variety of electrophiles. One drawback however of the dithiane moeity as an ASG is the lack of reactivity of α-substituted linchpins such as 3 (Figure 1; R=CH₃).^10b We now recognize this lack of reactivity to be due to a preferred linchpin conformation that buries the terminus of the electrophilic epoxide in the dithiane ring. We therefore turned, in 2007, to possible alternative ASGs; preliminary studies with linchpins 4 and 5, bearing cyano or allyl groups proved promising.¹¹ A general search for other viable ASGs that would extend the scope and general utility of the ARC tactic was therefore initiated. Central to this venture was an ASG that would prove to be competent with α-substituted linchpins in order to access a variety polyketide natural products and/or natural product-like analogs for diversity oriented synthesis (DOS).¹² In this letter we report on one aspect of this program, the design, synthesis and evaluation of a new class of bifunctional Type II ARC linchpins that employ the phenyl or phenylthio¹³ group as the ASGs (Figure 1).

Figure 1.

New Bifunctional Linchpins for Type II ARC

Unlike earlier linchpins employing a disubstituted ASG (cf. dithiane), use of a monosubstituent ASG, such as phenyl or phenylthio moiety, leads to introduction of an asymmetric center at the silyl bearing carbon. The question thus arises as to the stereochemical outcome (i.e., retention, inversion or racemization) resulting from capture of the Brook derived anion. In the case of the 1,2-Brook rearrangement of α-silyl alcohols bearing a phenyl ASG, elegant studies by Brook^14b and Mosher^14c revealed that inversion occurs at carbon and retention at silicon.¹⁴ We therefore designed a series of enantiomerically pure linchpins (Figure 1, 6–10) possessing phenyl or phenylthio moieties as ASGs to explore both the viability of the Type II ARC process and the stereochemical outcome at carbon upon Brook rearrangement and anion capture.

Access to linchpins 6–10 called upon the protocol developed by Sato and coworkers¹⁵ that entails regioselective opening of disubstituted epoxide 11.^15a Reaction of 12 with 11 furnished (−)-13a and (+)-13b as a diastereomeric mixture (ca. 1:1). Removal of the trityl group with p-TsOH pleasingly furnished diastereomers (−)- 13a and (+)-13b that could be separated by flash chromatography. X-ray diffraction of the primary 4-bromobenzoate of diol (+)-13b, permitted assignment of the relative and absolute configurations.¹⁶ Oxidative cleavage of the diols led to aldehyde linchpins (−)-6a and (+)-6b in excellent yields. Epoxide linchpins (−)-7a and (+)-7b were prepared from (−)-13a and (+)-13b by monotosylation of the hydroxyl, followed by treatment with n-BuLi. Preparation of the non α-substituted linchpins, (−)-8a and (+)-8b, entailed a two step sequence employing the readily separable chloroalcohols (−)-15a and (+)-15b, derived from (S)-(+)-epichhlorohydrin. The relative and absolute configurations of (+)-15b were established by X-ray analysis of the corresponding 3,5-dinitrobenzoate.¹⁶

Construction of the phenylthio linchpins (−)-9a, (+)-9b, (−)-10a and (+)-10b proceeded in similar fashion (Scheme 3). For (−)-10a and (+)-10b, a three-step sequence: (1) acetylation, (2) mesylation and (3) epoxide formation upon deacetylation was employed. The relative and absolute configurations of (−)-9a and (−)-10a were again assigned by X-ray analysis. ¹⁶

Scheme 3.

Synthesis of Linchpins of 9a, 9b, 10a and 10b

With linchpins 6a–10b in hand, we evaluated both their viability as linchpins for the Type II ARC tactic and the stereochemial outcome at carbon upon Brook rearrangement followed by alkylation. Polar aprotic additives (cf. HMPA) were employed to trigger the Brook rearrangement⁹ of the initially generated litihium alkoxides. As illustrated in the Table 1 and Table 2, the phenyl and phenylthio moieties proved viable as ASGs for the ARC process. Initiating nucleophiles included n-BuLi, lithiated 2-methyl-1,3-dithiane and lithium di-n-butyl cuprate. Yields ranged from 50–81%. Of particular note is the stereochemical outcome. Addition of n-BuLi to linchpin (−)-6a furnished a diastereomeric mixture (95:5) with the all syn-product (−)-18 predominating (Table 1, Entry 1). The relative stereochemistry of (−)-18 was established by 2D NOESY analysis of the δ-lactone derived from (−)-18 upon oxidative cleavage and lactonization.¹⁶ The stereochemical outcome at the carbinol of (−)-18 was predicted via the Felkin-Anh model as observed by Sato upon addition of ethyl Grignard to vinylsilyl aldehydes;¹⁵ the stereochemical outcome of alkylation after Brook rearrangement only had potential precedent (i.e., inversion at carbon) based on the Brook^14b and Mosher^14c observations.

Table 1.

Anion Relay Chemistry of Linchpins 6a–8b

entry	NuLi	linchpin	products (yield%)a
1	n-BuLi
2	n-BuLi		(−)-18 (60%, d.r. >95:5)
3		(−)-6a
4		(+)-6b	(−)-19 (62%, d.r. >95:5)
5
6			(−)-20 (64%, d.r. >98:2)
7	(n-Bu)2CuLi	(−)-7a
8	(n-Bu)2CuLi	(+)-7b	(−)-21 (72%, d.r. >90:10)
9	n-BuLi
10
11			23 (76%, d.r. 1.8:1)
12	(n-Bu)2CuLi	(−)-8a
13	(n-Bu)2CuLi	(+)-8b	24 (73%, d.r. 1.1:1)

^(a)Isolated yields based on linchpins; diasteremer ratio was determined by ¹H NMR.

Table 2.

Anion Relay Chemistry of Linchpins 9a-10b

entry	NuLi	linchpin	products (yield%)a
1
2			25 (65%, d.r. ~1:1)
3	(n-Bu)2CuLi	(−)-9a
4	(n-Bu)2CuLi	(+)-9b	26 (72%, d.r. 1:4)
5
6			27 (55%, d.r. 1:1)
7	(n-Bu)2CuLi	(−)-10a
8	(n-Bu)2CuLi	(+)-10b	28 (60%, d.r. 1:1.9)

^(a)Isolated yields based on linchpins; diasteremer ratio was determined by ¹H NMR.

Interestingly, in the case of linchpin (+)-6b, addition of n-BuLi led to the same product [(−)-18] with the same diastereoselectivity as observed with (−)-6a (Entry 2). When employing lithiated 2-methyl-1,3-dithiane as the nucleophile both linchpins (−)-6a and (+)-6b again furnished the same product (−)-19, with similar high selectivities (Table 1, entries 3–4). Linchpins (−)-7a and (+)-7b also proved viable substrates for the ARC process; again single products (−)-20 or (−)-21 predominated when lithiated 2-methyl-1,3-dithiane or lithium di-n-butyl cuprate served as nucleophiles (Entries 5–8).¹⁶ In contrast, poor diastereoselectivities were obtained when linchpin (±)-6c¹⁷or linchpins (−)-8a and (+)-8b, lacking the methyl subtituent were employed (Table 1, Entries 9–13). Although there are few reports¹⁸ on diastereoselective alkylation of carbon anions generated via Brook rearrangement, our results are consistent with alkylation of configurationally labile benzyl anions.¹⁹

The stereochemical outcome employing linchpins 9a–10b bearing phenylthio moieties as the ASG proved different (Table 2). Although the ARC process proceeded in good yield (Table 2, Entries 1–8), diastereomeric mixtures resulted in all cases. That is, an α-chiral center adjacent to the reactive anion generated via Brook rearrangement [linchpins (−)-10a and (+)-10b] imposed little or no stereochemial bias towards the alkylation process (Table 2, Entries 5–8). This stereochemical outcome is not totally unexpected,²⁰ given that α-thio carbanions are known to be configurationally labile,²¹ even at −78 °C.

The divergence in stereochemical outcome observed with linchpins possessing the phenyl and phenylthio ASG with adjacent α-substituent can be understood in terms of A^1,3 strain interactions available only to the linchpins possessing the phenyl ASG (Scheme 4). A similar divergence was observed by Beak and coworkers in their pioneering work on directed amide alkylations.²²

Scheme 4.

A¹,³ Strain in diastereoselective alkylation

In summary, we have designed, synthesized and evaluated ten chiral non-racemic linchpins (6–10) for use in Anion Relay Chemistry (ARC). Both the phenyl and phenylthio moieties proved competent as anion stabilizing groups (ASG) for the ARC proccess. High diastereoselectivities were observed with linchpins possessing a methyl substituent α to the silyl group, when a phenyl moiety serves as the ASG, while poor diastereoselectivity is observed when a phenylthio moiety is employed as the ASG, irrespective of the presence of an α substituent. Studies directed towards both the design and synthesis of related linchpins, as well as analysis of the scope and limitations of Anion Relay Chemistry continue in our Laboratory.

Scheme 2.

Synthesis of Linchpins 6a, 6b, 7a and 7b