Skip to main content

Analytical Chemistry

Which Enantiomer, Anyway?

Assigning enantiomers (mirror-image isomers of a compound, for the non-organic-chemists in the crowd) can be a pain. By definition, no non-chiral technique can tell the difference between such things, and many of the chiral techniques will just tell you that they’re different, but not which one is which. Take, for example, chiral chromatography, in its various forms. You can often get a very nice separation and resolve a racemic mixture into two peaks, but you will have no idea what the absolute stereochemistry of each compound might be. In the drug labs, you’ll see these things labeled as “A” and “B” enantiomers, with the absolute assignment to be determined later. This language makes it all the way to the patents, with notations of “single unassigned enantiomer” and the like.

There are plenty of ways to, uh, resolve this situation, but none of them are completely general. If you can make a derivative of your compound with a chiral reagent, that’ll do the job if it crystallizes and you can get an X-ray structure. All the relative stereochemistries will become clear, and you can work back from the one that you know for sure in the group you’ve added on. “Mosher’s esters” are an example of this technique – those will often crystallize, but you can also read out the absolute chirality from the NMR spectra. The problem is that the group you’re derivatizing really needs to be pretty close to the enantiomeric center for this to work well – if you’re making Mosher esters of some alcohol way off on the other side of the molecule, you may well not see enough of a difference in the NMR to help you out much.

This new paper from Clemson suggests another technique: turn your alcohol into a sulfate (by reaction with sulfur trioxide/pyridine) and crystallize it with a guanidinium counterion. The sulfur atom is large enough for commonly available X-ray diffractometers to use anomalous dispersion to assign the chirality of the salt directly, and the sulfate/guanidinium combination has already been demonstated to be well-suited for growing crystals. It forms plenty of strong hydrogen bonds in all directions, which is just want you want for making the ordered solid phase thermodynamically appealing.

The paper demonstrates this technique on 17 test cases (16 alcohols and an amine, 1-phenethylamine). The chiral center doesn’t have to be right on the OH or NH2-bearing carbon, and the X-ray diffraction data seem robust. If you need to get your compound back, the sulfate hydrolyzes off readily. The paper does these on the tens-of-milligram scale for in-house X-ray equipment, but you could surely get away with less if you have synchrotron time. Overall it looks like a handy method, and we need all of those we can get for this problem.

Of course, if your molecule doesn’t have an OH or NH2 handle, you’re back to square one. The general solution to the which-enantiomer situation remains elusive!

22 comments on “Which Enantiomer, Anyway?”

  1. RM says:

    I wonder if some host-guest approach might work for some cases.

    That is, you have some derivatized chiral cyclodextran/cucurbituril-like molecules (probably a series) which provide a pocket for the molecule to bind into. Those host compounds are engineered such that they’re easy to crystallize/do NMR with. You get the benefit of the above mentioned derivitization approach, but you don’t necessarily need any covalent attachment points. (Instead, your issues are with finding a host molecule which actually binds your greaseball, and binds it in a specific enough fashion that the chiral centers of interest are well-resolved in the x-ray density/NMR spectrum.)

    1. Barry says:

      sure. In the context of a Drug Discovery effort, that’s called soaking a novel candidate into target crystal and solving the diffraction pattern of the complex. As long as complexation doesn’t require much movement on the host protein’s part, it often works.

    2. Anonymous says:

      Yes! You describe the “crystalline sponge” method, which Derek has written about himself. Very cool technology that Makoto Fujita, among others, are pushing. See below for Derek’s other posts on the subject (chronological order):

  2. Barry says:

    For the Drug discovery operation that’s tooled up for protein crystallography, it is routinely more satisfactory to co-crystallize the small molecule in question in a target active site and read the absolute chirality out (knowing that the amino acids in the protein are all (L)). Small-molecule crystallography requires quite a bit of different hardware and some different expertise.
    But it’s exciting that crystallography can be extended to species that are not themselves crystalline

  3. anon says:

    VCD anyone? FDA accepted.

  4. Naive reader says:

    Could some spectroscopic technique (VCD?) yield information that is different for each enantiomer? If so, computational techniques can calculate the spectrum for each possibility – is there is a qualitative or better match I would think this would be useful… Groups were doing accurate IR spectra prediction back in the 90’s and I assume it has developed from there. Especially has hardware is cheap.

    1. Conformer says:

      VCD surely can be used to assign enantiomers by comparison between simulation and measurement. The issue is that VCD is very sensitive to conformation, often a different conformation can lead to a change in sign for a given band. This can lead to confusion with the mirror image relationship between enantiomers.
      The spectra of enantiomers are mirror images but typically reflect a mixture of conformers. For analysis with sufficient certainty one would need to perform accurate conformational analysis, and this can turn out to be tougher than expected.
      Overall progress in VCD is fairly impressive over recent years. Still, hardware is not cheap though.
      Regular CD might be quicker to obtain, but calculating the spectra is far more difficult.

      1. MDB says:

        Another post in support of VCD — the overwhelming majority of these systems would be slam dunks by this methodology (which could be done in a fee-for-service fashion or in collaboration with an academic lab possessing an instrument if bringing the hardware in-house is prohibitively expensive).

        See this nice recent review, published in a location where organic chemists would hopefully see it (only the cover article, in case you missed it…)

        DOI: 10.1021/acs.joc.9b00466

        I am not the author of this work.

  5. dasf says:

    Just two things: I think you meant guanIdinium, not guanAdinium throughout the paper.
    Second: Am I blind, or did you forget to link the paper?

  6. Andrew says:

    Is there a link to the paper?

    1. Derek Lowe says:

      There would be if I had remembered to include it (!) Fixed. . .

  7. Andrew says:

    Do you mean this paper by Kolis and Whitehead? (Not Clemson as you quote in the article)

    1. Steven K says:

      Kolis and Whitehead are located at Clemson

  8. Chris Gripton says:

    Those are some nice alkylating agents they’re making there…

  9. luysii says:

    If you really want to understand where optical rotation comes from (and how mirrors and diffraction gratings work) read the following chapters of Vol. I of the Feynman lectures on physics, something I never had time to do until I retired. Quantum mechanics is not involved, it’s all very classical. Most chapers are under 10 pages.

    ▶Chapter 27.Geometrical Optics
    ▶Chapter 28.Electromagnetic Radiation
    ▶Chapter 29.Interference
    ▶Chapter 30.Diffraction
    ▶Chapter 31.The Origin of the Refractive Index
    ▶Chapter 32.Radiation Damping. Light Scattering
    ▶Chapter 33.Polarization
    ▶Chapter 34.Relativistic Effects in Radiation

    1. KazooChemist says:

      Yes, the Feynman Lectures on Physics are a great read. Now available free online:

  10. KN says:


  11. xray mistery says:

    According to what I was told, it is not required anymore to have a “heavy atom” (such as bromine) in the molecule in order to determine absolute stereochemistry by XRay. This is opposite to what most of us learnt in the past. I guess modern XRay diffraction just got better? Can someone confirm?

    1. Ivan Bushmarinov says:

      Indeed the heavy atom is not exactly necessary anymore, if you have good crystals, some luck and either enough oxygens in your molecule (or just about any crystals and a microfocus copper X-ray source). Parson’s quotients (strictly speaking, a correction of a 50-year-old error in the method X-rya-based structure determination ) were implemented in the main crystallographic packages a few years ago.

  12. Osmium_fan says:

    Along similar lines:
    If your compound in question bears a double bond to react with OsO4, this might be helpful—

  13. Rina Dukor says:

    Thank you to everyone who has so elegantly described VCD. Thank you MDB for placing a link to the most recent review.

    My name is Rina Dukor and I am one of the co-founders of a company that commercialized VCD. I have not written here before (although do read this blog often) and thought this would be a good first time for an update on VCD:

    1. VCD is now used by all major pharmaceutical companies for determination of absolute configuration through their in-house equipment. Many 2nd and 3rd tier companies use lab services offered by BioTools or through collaborations with many academics who have the equipment.
    2. In our rough estimate, there has been more than 7000 compounds for which AC has been determined using VCD!!
    3. There are two chapters in USP on VCD – & ;; there are also over a hundred patents for NDA’s where AC is proven with VCD!
    4. I think over 1000 papers have been published with use of VCD for some stereochemical assignments. There are many academic groups specializing in VCD research and many more using it as a tool.

    Someone mentioned that you need careful conformational assignment by saying “The issue is that VCD is very sensitive to conformation, often a different conformation can lead to a change in sign for a given band. This can lead to confusion with the mirror image relationship between enantiomers.”

    BUT in fact it is not an issue – it is an advantage of VCD (and others spectroscopic techniques) that conformers and their populations can be determined by VCD! Yes, one needs to be careful with conformational search to make sure that none of the important conformers are missed but there are now many excellent programs that do just that. And yes, some bands might switch signs but the advantage of VCD vs CD, for example, is that there are many bands in the fingerprint region and all of the bands will not switch on conformers thus making absolute configuration assignment possible.

    Other advantages of VCD (if not yet obvious) is the measurement is done in solution so no crystal growth, no heavy atoms, or any molecule modification required.

    Derek – you conclude your blog by stating: “The general solution to the which-enantiomer situation remains elusive!”. ……..I know I am very biased (before anyone jumps on me for being commercial here), but thousands of compounds, over a thousand papers and everything I listed above – I would argue that VCD does indeed provide a GENERAL SOLUTION to the ‘which enantiomer’ question.

Comments are closed.