Chirality, E/Z and atropisomerism

Two molecules can have the same atoms joined in the same order and still be different substances in the body, because a binding site is itself handed. This module teaches you to find the handedness, name it unambiguously, and say what it costs to get it wrong.

A stereocentre is an atom where swapping two attached groups gives you a different molecule. For carbon, the usual test is four different groups. Nitrogen can be one too, but it inverts rapidly, so in practice it only counts when it is locked in a ring or quaternised. Sulfur in a sulfoxide is a stereocentre, with its lone pair as the fourth group — that one is missed constantly.

With n stereocentres there are up to 2ⁿ stereoisomers. Four centres is sixteen compounds; six is sixty-four. That number is why stereochemistry is a synthesis and a regulatory problem as much as a chemical one.

"Left-handed" is not a name a database can store. The Cahn-Ingold-Prelog rules give an unambiguous one. The procedure is short:

• Rank the four groups by priority — higher atomic number wins.
• If two are tied at the first atom, go one bond further out and compare again. A double bond counts as two copies of the atom.
• Point the lowest priority group away from you.
• Trace 1 to 2 to 3. Clockwise is R; anticlockwise is S.

The trainer below does this on real structures. It outlines each group, asks you for the first atom's atomic number, and sorts what Rule 1 can sort — then makes you go one bond out where there is a tie.

Enantiomers are non-superimposable mirror images. Every stereocentre is inverted. They have identical melting points, solubilities, logP values and NMR spectra in an achiral environment. They differ only in the direction they rotate plane-polarised light, and in how they behave in a chiral environment — which the body entirely is.

Diastereomers are stereoisomers that are not mirror images: some centres inverted, not all. They are different compounds in every respect — different melting points, different solubility, different spectra, and they can be separated by ordinary chromatography. Enantiomers cannot.

The meso case is worth pausing on. Meso tartaric acid has two stereocentres and is nevertheless achiral, because it has an internal mirror plane: one half is the mirror image of the other. Counting stereocentres is not the same as counting chirality.

A racemate is a fifty-fifty mixture of two enantiomers. Historically most chiral drugs were sold this way, because separating enantiomers was hard and expensive. Regulators now expect a specific justification for doing so.

A racemic switch is redeveloping a marketed racemate as a single enantiomer. Esomeprazole from omeprazole, levocetirizine from cetirizine, escitalopram from citalopram. Sometimes the single enantiomer is genuinely better; sometimes the switch is mostly a patent strategy. Both happen, and the literature on each case is worth reading before forming a view.

The active enantiomer is the eutomer; the less active one is the distomer. The ratio of their potencies is the eudysmic ratio.

The mistake to avoid is treating the distomer as an inert diluent. It is a different compound, present at the same dose, and it has its own pharmacology. It may bind a different target and cause a side effect. It may be metabolised differently and inhibit an enzyme. It may be converted into the eutomer in vivo — (R)-ibuprofen is, by a specific epimerase, which is why the racemate works clinically. Or it may, as with thalidomide, interconvert with the eutomer in the body so that separating them achieves nothing at all.

Around a double bond there is no rotation, so substituents are locked. Z means the two higher-priority groups are on the same side; E means opposite. Priorities are assigned by the same CIP rules. The old cis and trans language is still used, and is unambiguous only when each carbon carries one hydrogen.

Atropisomerism is the case you met at the end of module 6. If rotation about a single bond is hindered enough, the two twisted forms become separable compounds — stereoisomers arising from a rotation barrier rather than from a stereocentre.

The regulatory treatment follows the interconversion rate directly, and the usual three classes are worth knowing:

• Class 1 — interconverts in seconds. Treat it as one compound with a conformational preference.
• Class 2 — interconverts over hours to days. The awkward middle: separable in principle, racemising on the shelf or in the patient.
• Class 3 — stable for years. Treat it exactly as you would a conventional enantiomer, and develop a single one.

A practical warning, because this will bite you. Stereochemistry is stored fragilely and is corrupted routinely.

• A SMILES without @ or @@ is not a claim that the molecule is achiral. It is usually a claim that nobody recorded the stereochemistry.
• InChI has several layers, and a comparison done at the wrong layer silently equates enantiomers.
• A 2D SDF with no wedge bonds has lost it entirely.
• Vendor catalogues frequently list a racemate under the name of one enantiomer.

So: undefined and achiral are not the same thing, and treating them as the same is how a docking run ends up with the wrong hand. Before you dock anything, check whether the stereochemistry in your input file is defined, and if it is not, find out what the compound actually was.