Deprotection, global

In the Fmoc protection approach, the acid-labile ferf-butyl groups are often used for side-chain protection. The base-labile Fmoc groups can be easily removed during a synthesis using piperidine (Fig. 4). The final global deprotection together with cleavage from the polymeric support is achieved with TFA. 

Elaboration of triol 88b to bryostatin 7 requires chemoselective hydrolysis of the Cl methyl ester in the presence of the C7 and C20 acetates, macrolide formation, installation of the C13 and C21 methyl enoates, and, finally, global deprotection. The sequencing of these transformations is critical, as attempts to introduce the C21 methyl enoate to form the fully functionalized C-ring pyran in advance of macrolide formation resulted in lactonization onto the C23 hydroxyl. In the event, trimethyltin hydroxide promoted hydrolysis [73] of the Cl carboxylate of triol 88b, and subsequent trie thy lsilylation of the C3 and C26 hydroxyls each occurs in a selective fashion, thus providing the seco-acid 89. Yamaguchi macrolacto-nization [39] proceeds uneventfully to provide the macrolide 67 in 66 % yield (Scheme 5.14). 

C-H activation at a primary benzylic site was the key step in very short syntheses of lig-nans 206 and 207 (Scheme 14.27) [138]. Even though both the substrate 203 and the vinyl-diazoacetate 204 contain very electron-rich aromatic rings, C-H activation to form 205 (43% yield and 91% ee) is still possible because the aromatic rings are sterically protected from electrophilic aromatic substitution by the carbenoid. Reduction of the ester in (S)-205 followed by global deprotection of the silyl ethers completes a highly efficient three-step asymmetric total synthesis of (-i-)-imperanene 206. Treatment of (R)-205 in a more elaborate synthetic sequence of a cascade Prins reaction/electrophilic substitution/lacto-nization results in the total synthesis of a related lignan, (-)-a-conidendrin 207. 

Exposure of 76 to HF in H20/MeCN for 20 hours was followed by a short microscale flash column. Analysis by H NMR spectroscopy (500 MHz, CD3CN, 16 hours) of the global deprotection product showed it to be a complex mixture, however there were some encouraging signals. Purification of this mixture was attempted by HPLC-MS (25 75... 

Our first disconnection revealed glycosyl monophosphate 7 and undecaprenyl monophosphate 8.11 This disconnection was chosen because previous work, directed toward the synthesis of the Park Nucleotide 3, had shown it possible to introduce an anomeric phosphate with anomeric selectivity in favor of the desired a-anomer. A second reason for choosing this disconnection was that undecaprenyl monophosphate was available for purchase from a commercial source. Thus, if we could identify a mild method for joining these two fragments, only a global deprotection step would be required to arrive at lipid I. 

At this point, we had achieved access to an orthogonally protected phosphomuramyl pentapeptide derivative poised to enable completion of our lipid I total synthesis. The technical hurdles that remained to be addressed were 1) identification of a mild method for installation of the lipid diphosphate linkage, 2) global deprotection, and 3) final purification of our synthetic lipid I. Our solutions to these problems are addressed in the following section. 

It now remained for us to apply the Coward protocol to our system and complete the synthesis of lipid I. Thus, phosphate 26 [Scheme 10], prepared by reductive cleavage of the phosphodiester protective groups of 9 (H2, Pd/C in MeOH, followed by pyridine, 91% yield), was converted to the corresponding phosphoroimidazolidate, whose formation was readily monitored via mass spectrometry. Excess carbonyldiimidazole was quenched via addition of methanol. The lipid phosphate salt was then added in portions via syringe until complete consumption of the phosphoroimidazolidate intermediate was observed. Mass spectrometry also allowed us to monitor the appearance of the desired lipid-linked diphosphate product. When the reaction was judged to be complete, the reaction solution was carefully concentrated and the crude product was treated with sodium hydroxide in aqueous dioxane in order to achieve global deprotection. The crude product was purified by reverse-phase... 

Following the lipid I analogy, we would introduce the lipid diphosphate at a late stage in the synthesis that would leave only a global deprotection as a final step.26 Disconnecting the diphosphate linkage provided the first retrosynthetic intermediates, disaccharyl pentapeptide... 

Further alkylation of the lithium (Z)-enolate of 25 with methyl iodide gave 26, introducing the C16 stereocentre (3 1 dr) and completing the carbon backbone. Oxidation at Cl and carbamate formation gave 27 which underwent a chelation-controlled reduction at C17 (30 1 dr). Finally, global deprotection completed the synthesis of discodermolide (1), with an overall yield of 4.3% achieved over 24 steps in the longest linear sequence. 

NaH to give the advanced enone 137 (Z E = 10 1). A further four steps were then required to complete the third-generation synthesis, beginning with C19 carbamate installation, followed by K-Selectride reduction to introduce the requisite (75)-stere-ocentre. Finally, global deprotection with concomitant S-lactonisation gave disco-dermolide (1) in 11.1% overall yield over 21 steps (longest linear sequence). 

With all four of the required protected acromelic acid analogues 106, 160,161,162 and 112 to 115 available in reasonable quantities as single diastereoisomers along with the C-4 epimers 116 to 119, it only remained to carry out global deprotection. 

---