Monday, March 11, 2013

Why not a semisynthesis?

Taken from a group meeting presentation on semisynthesis.

     One of the perks of being in the Baran lab is the weekly Saturday morning group meeting.  A free breakfast is accompanied by a short lit, long lit, and seemingly impossible mechanism problem to remind everyone of his or her incomplete (or inadequate) grasp of the literature.  New topics are not easy to come by owing to the fact that we have over ten years of archived presentations.  When my turn for long lit presentation arrived, Phil recommended the topic of semisynthesis.  Phil said, “Go for it, and take it any direction you want.”  Initially it seemed like a rather boring presentation, until a grad student recommended I take a look at a book by Tse-Lock Ho titled “Enantioselective Syntheses: Natural Products from Chiral Terpenes”.  If Ho would have been at Scripps it might have been called “Classics in Semisynthesis” and the pages illustrated with every color of the rainbow.  I decided to turn this into a game of “Find the Terpene”.   After a brief look at a natural product, the object is to identify the original terpene from which it came.  In conclusion, people have torn these cheap terpenes to pieces to arrive at nearly any sized ring or stereocenter your heart could desire.  Much can be accomplished if you are willing to go through the steps.  

      When it came time to create the intro slide I found it difficult to track down the first use of the term “semisynthesis.”  On top of that, any reasonable definition that I could find didn’t reflect the current use of the word in the literature.  Not to point the finger (since we are as guilty as anyone), but if the starting material is from the chiral pool, how can you call it a total synthesis?  Why the difference in the literature?  It seems that the word “semi-synthesis” has a negative connotation, is a dirty word, as if the authors took a shortcut or easy route.   But what if only one of ten stereocenters in a natural product is derived from the chiral pool?  Would we label this feat with such a disgraced title?  Where do we draw the line between semi and total? 

     This is where my presentation took a turn (maybe for the worse).  Since no adequate metric for comparison of semisyntheses existed, one needed to be created.  One must be created! (a little heavy on the sarcasm)  To keep it simple, I assumed the two variables consistent with all semi-syntheses were the incorporation of mass and complexity from the chiral pool.  Simple addition of these two variables would arrive at a given value, which I deemed naturality, or N.  The mass variable, M, consists of a percentage of mass retained from the starting material (m1) in relation to the mass of the target (m2).  The complexity variable, C, is a percentage of stereocenters remaining from the starting material (s1) over the total stereocenters in the target (s2).  After multiplying each term by 50 and adding them together, the resulting scale would range from 0-100.   A total synthesis would always give a value of 0, and the value would go up from there.  Unfortunately, as a synthetic organic chemist, this nears the extent of my mathematical prowess.

N = M + C = (m1/m2)50 + (s1/s2)50

     More variables could be added to include step counts, redox manipulations, functional groups, rings, etc.

(-)-limonene to (+)-carvone                                    (-)-limonene to (+)-bilobanone
m1 = C10H14 = 134                                                 m1 = C10H10 = 130
m2 = C10H14O = 150                                              m2 = C15H20O2 = 232
S1 = 1                                                                      S1 = 1
S2 = 1                                                                      S2 = 1
N = 45 + 50 = 95                                                    N = 28 + 50 = 78

     As expected, a degradative synthesis of jorumycin gave a very high value of N = 95.

     The smallest value I could find came from my previous lab, in which a single oxygen atom from lactic acid remains in rishirilide B, affording a value of N = 2.  

Next, I took various syntheses from the literature, assigned each a value, and put them in table. 




aeruginosin 298-A
rishirilide B
aeruginosin 205-B
N-methylwelwitindolinone C  isothiocyanate
lannotinidine B
aeruginosin 298-A
cortistatin A
epicoccin G
aeruginosin 298-A
aeruginosin 298-B

            Looking through the final table, we can see how they span the range from 5-95. Conclusions?  Anyone? Anything?  Would you use this approach to design a synthesis?  What would it be useful for?  My answers are no and I’ve yet to find one.  This could have been a waste of valuable time.  I may have created the first metric for comparison in which nothing useful can be derived from the results.  That may be a feat in itself.  The one realization that meandered out of this exercise would be that people misuse the term semisynthesis.  While this metric is most likely flawed, the one thing we might agree on is that if you can arrive at high value of naturality, you should hesitate to call it a total synthesis.  I’ll be first to admit that a simple label doesn’t encompass the innovation of a manuscript, and you shouldn’t get caught up in such trivial matters.  So maybe “semisynthesis” isn’t a foul word after all.  I could argue it is fitting for a number of total syntheses.  Or possibly, the term “total synthesis” is the one that is misunderstood or antiquated, as it seems to be a catch-all label for any synthesis of a natural product.


  1. Total synthesis is like rock climbing, an accomplishment to be proud of, indeed. But if there's a hiking trail that goes to the same place faster and easier, that's your semisynthesis. Outside of academic publications, you go the easy way every time, and there is elegance in the degree of efficiency. Taxol wouldn't be a drug today without a good semisynthesis.

  2. I've never understood the negative bias behind semisynthesis, either. Nature does it all the time (we'll ignore the iterative PKS mechanisms to make that point).

    Great post!

    1. P.S. - You may want to check the spelling on "Nicolaou" : )

    2. Fixed. Not my first misspelling of that name.

  3. An easy way to solve this problem is to just call everything a 'synthesis' and then let the reader decide. I agree with the sentiment in general though. Who cares what you started from as long as you make your target?

  4. I thinks it's not bad to try and come up with a way to measure a synthesis minus the artistic element of it. Step account and total yield is okay but not the most efficient. Kind of like a slugging number versus just batting and homeruns. Something that takes into account overall steps, yield, number of chiral center (weighed by difficulty), heteroatom, and probably others things too. That would be cool.

  5. Regarding Saturday morning seminars: Brian Stoltz explained to me once why he would not come to the lab on Saturdays. Even as a hard-working (and exceptionally productive) grad student he used to be, he already figured out that he needed to feel normal for at least a part of the wekend and that included Friday night partying + sleeping late and taking time off the next day. But he would always come to lab on Sunday morning at 9 am and have a full work day, so if his boss (and later the Chief) asked him on Monday about his project, he had a head start with few experiments running and other ones being already worked up and NMR-ed on Monday morning...

    The reason I mention this is that highly motivated bright young people need to learn how to maintain regular hours to stay productive, to rest and take some recreation time in an a completely purposeful, organized way so as to avoid burnout and depression when stuck in a difficult synthetic step. I think a healthy time management is not encouraged enough in synthetic chemistry programs.

    1. I couldn't agree more. I found there is nothing more counter-productive than working 24/7 and becoming a lab zombie.
      On the flip side, I think seminars on saturday morning are a good way to avoid putting a few hours hole in your lab schedule. I guess you could still take some free time at another point in the weekend.

  6. I generally say total synthesis if it is starting from commercially available compounds.
    What does it take to go from semi to total? Do you have to check that every moiety you add can be traced back to oil cracking?

    I see no problem in using the chiral pool and/or natural products to start a total synthesis.

    On the other side, if you use chiral pool-derived compounds as catalysts to set your chiral centers, it does not count as semisynthesis because it is not in the final structure, but you did use them.

    Same thing when you use enzymes to do a reaction with good ee... you kind of leave the hard work to Nature, it is a bit like using an elaborated compound nature does as a shortcut towards your target compound.

  7. This comment has been removed by the author.