Cancer doesn't like sugar chemistry

I shuddered when I saw the sugary graphical abstract for Danishefsky's recent JOC Note, "A Practical Total Synthesis of Globo-H for Use in Anticancer Vaccines."  (doi: 10.1021/jo901682p) And then a moment later, I realized... anticancer vaccines?!  Where the hell have I been?  Apparently, this glycosphingolipid compound Globo-H is an antigen that was isolated in 1983 from a breast cancer cell line, and is now known to also be over-expressed on the surface of other types of cancer cells (prostate, ovary, lung, colon, and small cell lung cancers).  Administration with alpha-galactosylceramide induces antibodies against globo-H and SSEA3 (the pentasaccharide precursor of globo-H).  If I understand correctly, it's the equivalent of administering a dead or partial or similar virus in order to trigger the formation of antibodies.  The pentavalent vaccine also contains other antigens that are overexpressed on cancer cells (prostate and breast).  A search for "globo-h" on pubs.acs.org alone produces 118 hits, encompassing syntheses, development, and animal studies. 

There have been other syntheses, including that of Schmidt, Boons, Wong, Seeberger (2002 and 2007), and Huang referenced in Danishefsky's paper, and an earlier synthesis of his as well which was too low-yielding.  The ABC trisaccharide that will be incorporated later was synthesized according to the previously reported Cp₂Zr(OTf)₂-mediated coupling.  The starting material was synthesized via α-epoxidation followed by glycosylation to produce the DE disaccharide shown below; coupling of this molecule with the thiofucosyl donor (with fucose being a hexose deoxy sugar) selectively produced the α-trisaccharide. The conditions screened that produced the highest yield, 80%, were Cp₂Zr(OTf)₂ with 5:1 toluene:THF and 2,6-di-t-butylpyridine(hindered base) in the dark over 72 hours, but the conditions the authors moved forward with produced an impressive 78% yield in only 4 hours:

That's not so scary - silver triflate draws the chlorine away from TolSCl, which pulls ^-STol in after the alcohol displaces it. The next step involved a (to me) funky iodine reagent, I(coll)₂ClO₄ to promote iodosulfonimidation of the now DEF glycol.  The sulfonamide blocks any attack from the bottom face of the ring, forcing the nucleophile in the next step to attack from the top to form the β-isomer. Treatment of the crude intermediate with lithium ethanethiolate leads to the complete DEF donor in 75%, ready for coupling with the ABC acceptor.

The last coupling proceeds without fuss in 72% yield after treatment with methyl triflate. Deprotection of TIPS with TBAF and Bn with sodium reducing conditions followed by global peracetylation leads to the target hexasaccharide:

 

The last steps to append this molecule to the carrier protein KLH (keyhole limpet hemocyanin) were previously published by Danishefsky and co-workers, as was to covalently link the molecule to an amide (doi: 10.3987/COM-08-S(D)82). I'll add those as soon as Heterocycles loads on my work laptop tomorrow.