The first postcard out of the bag this week with the correct solution – Sarah Reisman – was from Tynchtyk. Congratulations.
As Reisman is new to the game as it were, I have already dished out all of the biographical details I could find in the clues so we will get straight into the chemistry.
The selection of last weeks Mega Chemist, the one and only Sarah Reisman of Caltech was in fact a little self serving. Obviously I really enjoy Reisman’s impressive total syntheses (otherwise selection for #MegaCC would obviously not have been so rapid) one of which is up for discussion today, but excitingly she has utilised my favourite reaction developed in the Procter group whilst I have been here; a SmI2 mediated reductive dialdehyde cyclisation cascade. Thus, here is my first opportunity to exalt the virtues of the sometimes derided SmI2 – the bread and butter of the Procter group.
The Reisman paper in question is the first total synthesis of (-)-maoecrystal Z published last year in JACS, and hot on the heels of the long sought after (-)-maoecrystal V which was finally pinned down by Yang the year before.
Now for me personally, natural products like maoecrystal Z bring me out in cold sweats when planning a synthesis; the general lack of functionality and the complex tetracyclic core with six contiguous chiral centers should prove a nightmare for most. Reisman though, appears not even a little afraid; a clever retrosynthesis and the impressive 12 step synthesis that realised maoecrystal Z show both her ambition and ability.
Diving straight into the key step of the retrosynthesis Reisman planned to utilise a SmI2 mediated dialdehyde cyclisation cascade (scheme 1) developed in the Procter group. You can see the development of this reaction here and here, though more excitingly you can see the first full application in an effective formal synthesis of pleuromutilin utilising SmI2 with tBuOH as a co-solvent.
Reisman proposed that the single electron reduction of the presumedly ‘more kinetically accessible aldehyde’, would generate a ketyl radical that would undergo intramolecular addition to the unsaturated lactone (pink line). A second electron transfer would subsequently generate a samarium enolate of the lactone which would undergo aldol cyclisation with the waiting aldehyde (blue line), generating the tetracylic core of maeocrystal. Reisman’s kinetic arguments for the selectivity of the reduction are solid, though it should also be considered that the single electron reduction of aldehydes may be reversible, and that the ability of the ‘less hindered’ aldehyde and hence its ketyl radical to undergo a facile 5-exo-trig cyclisation vs. a possible 4-exo-trig cyclisation for the alternate aldehyde may account in part, for the chemo-selectivity of the reaction.
Reisman’s awareness that the diastereocontrol may be difficult to predict led to an initial study on the 5-exo-trig radical cyclisation. The substrate was elegantly prepared from epoxidised gamma-cyclogeraniol using a modification of Gansauer’s modified conditions for intramolecular Ti (III) mediated couplings of epoxides and acrylic acids (scheme 2) generating spirocycle 1 as a single diastereoisomer in excellent yield.
Alkylation of 1 and a little more fiddling around gave 2 which allowed the radical cyclisation to be studied (scheme 3).
Initial cyclisation attempts using solely SmI2 in THF resulted in decomposition of the starting material but a screening of additives found that a SmI2-tBuOH-LiCl reagent system did the trick (we will come back to that later), giving the proposed product in respectable yield. Reisman cites the shown steric interactions for obtaining the correct stereochemistry (scheme 4).
With the first cyclisation looking good, they went for gold. Exchanging LiCl for LiBr they realised the target compound, generating both rings and four contiguous chiral centers in one pop, in an impressive in 54% overall yield (scheme 5).
With the key step nailed, as all of us who have worked in total synthesis know you are home and dr…….. or, back in the real world, your apparently simple protecting group chemistry to finish up proves formidable; and this was the case for Reisman. Happily though Reisman circumnavigated these issues and did get to maeocrystal Z in just four additional steps. An excellent total synthesis.
For me the most intriguing aspect of this synthesis (particularly as a Sm chemist) is the additives used in the cyclisation. It is well know that you can modulate the reactivity and often the selectivity of SmI2 using a host of additives including Lewis bases such as HMPA, DMPU; proton sources including but not exclusively H2O, MeOH, tBuOH; and metals and their salts, including LiCl and LiBr as demonstrated by Reisman.
In principle Reisman’s cyclisation simply requires the formation of a ketyl radical from an aldehyde which is easily achieved with SmI2 alone, though it was reported this led to degradation of the starting material – maybe due to reduction of the aldehyde (s) without cyclisation, or a selection of pinacol type products. She found, as was the case in the pleuromutilin synthesis that the addition of tBuOH facilitated reaction, though its role is somewhat unclear.
tBuOH is known to increase the reductive potential of SmI2 – pep it up a little if you like – though in principle this should not be required. The ‘other’ role tBuOH can play is as a proton source, though in my mind this can only be a bad thing; as we increase the number of protons floating around we risk protonating the enolate generated after the radical cyclisation, preventing the aldol reaction to form the second ring! This is an intriguing conundrum, with still no light to shine on it!
In addition to tBuOH Reisman also used LiCl or LiBr in her cyclisations, both of which are know to give SmI2 even more get up and go. The fact that these are required at all, never mind that they don’t lead to all sorts of side reactions is a mystery to me. Another conundrum still to be solved; though I do wonder what happened to the other 46% of starting material.
So despite SmI2 sometimes getting a bit of negative press, especially from industrial folk (and no, Sm is not radioactive) I think I have shown here its potential in total synthesis – 2 rings, 4 steroecenters in both maeocrystal Z and pleuromutilin in a SINGLE step.
The applications of SmI2 in functional group transformations and radical and anionic bond forming processes are massive, and the control and selectivity often achieved with the addition of additives is very impressive. You can find more information online – though I feel I have to mention this beautiful book to maintain favour with my boss.
And keep an eye out on Nature’s protocol exchange, an idiots guide to preparing SmI2 will be coming your way soon.