Oxidation At Methionine And Cysteine

The vulnerability is chemical. Divalent sulfur in the thioether of methionine and the thiol of cysteine has accessible lone pairs and a low oxidation potential, so it reacts with molecular oxygen, peroxides and hydroxyl radicals at rates that the carbon backbone does not approach. Tryptophan, tyrosine and histidine oxidise as well, more slowly and by more varied routes. Everything else in a typical sequence is comparatively inert.

Methionine oxidation is a two-step ladder. One oxygen atom gives methionine sulfoxide, a monoisotopic increase of 15.9949 Da. A second gives methionine sulfone, an increase of 31.9898 Da, and that step is slow enough under ordinary conditions that the sulfoxide is what is normally seen. The first step is partially reversible by chemical reduction; the second is not.

Cysteine is more branched. Two thiols can oxidise to a disulfide, losing two hydrogen atoms for a monoisotopic decrease of 2.0157 Da. A single thiol can go on to cysteine sulfinic acid at plus 31.9898 Da or cysteic acid at plus 47.9847 Da. It can also form a mixed disulfide with free cysteine or with glutathione present in the process stream, in which case the mass shift is that of the whole partner.

Mass shifts and where they show up

The retention-time column matters more than it looks. Adding oxygen to a thioether converts a nonpolar group into a strongly polar one, so a sulfoxide binds the stationary phase less tightly and comes off the column sooner. That is the reason an oxidation impurity generally appears as a shoulder or a satellite on the leading edge of the main peak rather than a well-resolved separate band.

What promotes it during and after synthesis

• Dissolved oxygen in any aqueous step. Air-saturated water at 25 degrees C holds roughly 8 milligrams of dissolved oxygen in every litre, measured by Clark-type electrode.
• Trace transition metals. Iron and copper at single-digit parts per billion catalyse Fenton chemistry that generates hydroxyl radicals; ICP-MS is the usual way these are quantified.
• Residual peroxides in excipients. Polysorbate and polyethylene glycol grades carry peroxide values that are measured in milliequivalents of active oxygen per kilogram and rise with storage time.
• Light. Ultraviolet and short-wavelength visible light generate reactive oxygen species indirectly through photosensitisers, which is why photostability is assessed separately under ICH Q1B.
• Cleavage and deprotection chemistry. Oxidising scavengers and strong acids used to remove side-chain protecting groups act on methionine as well.

Lyophilised solid is far less reactive than solution because the mobility required for the reaction is largely absent, but it is not inert. Residual moisture measured by Karl Fischer titration, headspace oxygen measured by laser-based headspace analysis, and the presence of an amorphous rather than crystalline matrix all bear on how much residual chemistry the solid can support.

The failure mode: a sulfoxide counted inside the main peak

The most consequential thing oxidation does to a record is not to be missed entirely. It is to be included. Purity by reverse-phase HPLC is calculated as the area of the main peak divided by the total integrated area, and integration boundaries are set at the points where the signal returns to baseline. A methionine sulfoxide that elutes 0.2 min ahead of the parent does not return to baseline in between. It forms an unresolved front shoulder, the integration software draws a single peak across both, and the reported purity figure counts the oxidised species as product.

The failure is quiet because every other part of the record looks correct. The chromatogram has one main peak. The mass spectrum, if acquired on the pooled material, shows the expected mass alongside a plus 16 Da signal that can be dismissed as in-source oxidation, which is a real artefact of electrospray and does occur. The way to separate the two possibilities is to change the gradient slope and see whether the shoulder resolves, and to check whether the plus 16 Da signal tracks the shoulder in time rather than appearing across the whole peak.

Disulfides are not always a defect

For a sequence with two cysteines, an oxidised disulfide may be the intended form rather than an impurity, and the reduced form the one to count against. The minus 2.0157 Da shift is the same chemistry read in the opposite direction. Where more than two cysteines are present, the number of possible pairings grows quickly and mass alone cannot distinguish between correctly and incorrectly paired isomers, because they are the same elemental composition.

Where our own position stops

Aurum publishes a reverse-phase HPLC purity figure with the detection wavelength stated and a mass confirmation. Neither of those is a statement about oxidation specifically. A methionine sulfoxide has essentially the same absorbance at 214 nm as the parent, because the peptide bond is the chromophore and oxidation does not change how many peptide bonds there are, so an unresolved sulfoxide contributes to the reported figure without altering the detector response in a way that would flag it. Forced oxidative degradation studies on each lot is not among the specifications Aurum publishes, separate oxidised-variant specification is not among the specifications Aurum publishes, and Headspace oxygen is not among the specifications Aurum publishes. Where a question turns on how much of a given lot is the sulfoxide rather than the parent, our published record does not answer it.