The Holton Synthesis

The six reported syntheses are described in the following sections in order of their completion. The Holton and Nicolaou syntheses were published within days of each other, with the Nicolaou synthesis appearing first. The Holton synthesis is described first on the basis of the earlier submission date of the manuscript describing the work. 

Elaboration of the D-ring was achieved through the key intermediate 14.1.13, which was converted to an oxetane derivative by the steps of mesylation, osmylation, and treatment with base similar, although not identical pathways were used by all the other workers in this area. The synthesis was completed by conversion of the 1,2-carbonate protective group to the C-2 benzoate (another sequence used by most of the other investigators), and insertion of the C-9 oxygenation to give the protected baccatin III derivative 14.1.17. 

The overall yield of taxol from 14.1.1 in this synthesis is 4-5%. Compound 14.1.1 is readily available from the commercial intermediate (—)-patchino, but unfortunately it is the enantiomer of the structure required for taxol. Compound 14.1.1 with the correct absolute stereochemistry can be prepared from (—)-bomeol, but the additional steps required obviously detract from the elegance of the approach. 

---

Wender s construction of the A/B ring system [8a] in principle resembles the Holton synthesis, but makes use of the 4/6 precursor... 

The Holton synthesis of the structurally-analogous taxane ring system involves acid-catalyzed rearrangement upon epoxide opening of P-... 

All that remains before the final destination is reached is the introduction of the C-l3 oxygen and attachment of the side chain. A simple oxidation of compound 4 with pyridinium chlorochro-mate (PCC) provides the desired A-ring enone in 75 % yield via a regioselective allylic oxidation. Sodium borohydride reduction of the latter compound then leads to the desired 13a-hydroxy compound 2 (83% yield). Sequential treatment of 2 with sodium bis(trimethylsilyl)amide and /(-lactam 3 according to the Ojima-Holton method36 provides taxol bis(triethylsilyl ether) (86 % yield, based on 89% conversion) from which taxol (1) can be liberated, in 80 % yield, by exposure to HF pyridine in THF at room temperature. Thus the total synthesis of (-)-taxol (1) was accomplished. 

The first synthesis of Taxol was completed by Robert Holton and co-workers and is outlined in Scheme 13.53. One of the key steps occurs early in the synthesis in sequence A and effects fragmentation of 4 to 5. The intermediate epoxide 4 was prepared from a sesquiterpene alcohol called patchino. 35 The epoxide was then converted to 5 by a BF3-mediated rearrangement. 

In Holton s and Wender s work, the total synthesis was achieved by sequentially forming the AB ring through the fragmentation of epoxy alcohols derived from (—)-camphor and a-pinene. Nicolau s, Danishefsky s, and Kuwa-jima s total syntheses involved B ring closure connecting the A and C rings, whereas in Mukaiyama s synthesis, the aldol reaction was extensively applied to construct the polycyclic system. 

The total synthesis of taxol (1) by Holton and his associates [4] is based in the retrosynthetic analysis outlined in Scheme 13.6.7. 

Synthesis and coupling of the side chain to the tetracychc nucleus also proved to be a non-trivial synthetic challenge. Potter and Greene finally succeeded in a synthesis, but, with the low yields (approximately 50%), theirs was not a commercially viable approach [14]. Eventually, however, BMS scientists, in collaboration with the Holton team from FSU, developed a synthesis of the P-lactam for the side chain using the sequence shown in Scheme 7.2. 

At the time of the patent suit, it appears that BMS used a semisynthetic process that was licensed from Robert Holton s group at Florida State University. Holton s group not only discovered a highly efficient semisynthesis of paclitaxel from 10-DAB, but also were one of the first two groups (KC Nicolau s group was the other) to complete the total synthesis of paclitaxel, which means they assembled the molecule from commonly available precursors—essentially from scratch. The total synthesis, however, is not and probably never will be a commercially viable way to make this very complex organic molecule. 

Uses of FeCl3/Ac20 methodology in synthesis include the taxusin synthesis of Holton et al. (Sch. 2) [15], the Forsyth and Clardy synthesis of (+)-didemnenones A and B [16], and the synthesis of a nodulation factor (NodRf-III) by Fraser-Reid and coworkers [17]. 

Scheme 5.43 illustrates three applications of this methodology to total synthesis. The first exeunple is taken from Posner s synthesis of estrone and estradiol [211], the second from Posner s synthesis of methyl jasmonate [212], and the third from Holton s synthesis of aphidicolin [213]. The latter is particularly noteworthy in that two contiguous quaternary centers are created in the asymmetric addition with excellent selectivity. In the estrone synthesis, the chirality sense of the product is consistent with the nonchelate model, but the other two examples adhere to a chelate model. Note that the difference is the degree of substitution at the a-position of the enolate. 

The more convergent routes of Nicolaou et al. and Danishefsky et al. with B ring closure from a tethered A-C precursor are clearly the least efficient with respect to total yield (about 0.01 % starting from commercially available materials) if directly compared to the other two routes. Additionally the Nicolaou route needs enantiomer separation or a -presumably less favorable - asymmetric synthesis, whereas the Danishefsky route addresses absolute stereochemistry but involves more steps. Holton s synthesis shows quite good yields throughout with a total one of approximately 0,1 %, but has less room for improvements and is based on a not readily available starting material. The best synthesis in all respects to date is the one of Wender and coworkers, which is about 1-2 orders of magnitudes better in total yield (ca. 0.8%),... 

The /3-lactam synthon approach pioneered by Holton, Ojima, and Georg remains the most widely used method, and is the method currently used by Bristol-Myers Squibb for the commercial synthesis of taxol. 

The Wender synthesis was reported in two communcations in 1997 435, 436), but the full papers have yet to appear. The synthesis is linear of the form A- AB- ABC- ABCD, but it is very different from the formally similar Holton approach. 

The inefficiency of amide resonance in p-lactams and their spring-loaded reactivity associated with the release of angle and electronic strain played a defining role in the development of the key step in the Holton s synthesis of Taxol (Figure 11.58). ... 

Several different synthesis routes of paclitaxel have been published. The total synthesis has been first achieved independently by two teams in 1994, the Holton s and the Nicolaou s [41-43]. The third synthetic route was accomplished by Danishefsky s group in 1996 [44]. Total synthesis of Taxol is a complex task for chemists given the fact that the molecule consists of four complicated rings (A, B, C rings and the oxetane ring) and has 11 chiral centers. The synthesis processed more than 20 steps, and only 0.07% and 2.7% production rates for the Holton s and Nicolaou s routes were obtained, respectively [41, 42]. Mukaiyama et al. have proposed an improved method for the asymmetric total synthesis of Taxol by a different way [45]. An automated synthesizer with a 36-step synthesis sequences for intermediate of Taxol was developed by Doi et al. (2006) [46]. Due to the complexity of Taxol structure, the expensive chemical reagents, and the strict requirement for reaction, the process of the total synthesis of Taxol is multiple-step, costly, and commercially unfeasible for industrial application for Taxol manufacture. 

Polyepoxide Cyclization Cascade Based on the proposed biosynthesis of polyether toxins, Holton and coworkers used an electrophile-initiated epoxy alcohol cyclization cascade as a key step for the total synthesis of hemibrevetoxin B (37) (Scheme 4.7) [14], Treatment of 33 with N-phenylselenophthalimide (A -PSP) led to sequential cyclization to give 36 in 83% yield as a single stereoisomer. The cascade... 

Holton reported an influential application of epoxonium ions as intermediates in cascade reactions through the total synthesis of hemibrevetoxin B (11 Scheme 4.34) [69]. This process utilized A-(phenylseleno)phthalimide to form a sele-nonium ion from alkenyl epoxide 160. Exo-opening of the selenonium ion by the epoxide followed by m /o-opening of the resulting epoxonium ion provided 161 in an impressive... 

Scheme 14 Efficient formation of the seven-membered A ring from unprotected triol 75 in Holton s total synthesis of hemibrevetoxin B (77) [61]...

Holton, R.A., Somoza, C., Kim, H.B. etal. (1994) First Total Synthesis of Taxol. 1. Functionalization ofthe B Ring. Journal of the American Chemical Society, 116, 1597-1598. 

---