Structure Activity Relationships: No one predicted Diacetyl Toxicity; What does Disacetyl Predict?

Bioorg Med Chem. 2003 Mar 20;11(6):853-62.
The role of dicarbonyl compounds in non-enzymatic crosslinking: a structure-activity study.

Meade SJ, Miller AG, Gerrard JA. Department of Plant and Microbial Sciences, University of Canterbury, Private Bag 4800, Christchurch, New Zealand.


The Maillard reaction is a complex network of reactions that has been shown to result in the non-enzymatic crosslinking of proteins. Recent attention has focussed on the role of alpha-dicarbonyl compounds as important in vivo contributors to protein crosslinking but, despite extensive research, the molecular mechanisms of the crosslinking reaction remain open to conjecture. In particular, no relationship between the structure of the carbonyl-containing compounds and their activity as crosslinking agents has been established. In an effort to elucidate a structure-reactivity relationship, a wide range of dicarbonyl compounds, including linear, cyclic, di-aldehyde and di-ketone compounds, were reacted with the model protein ribonuclease A and their crosslinking activity assessed. Methylglyoxal and glutaraldehyde were found to be the most efficient crosslinkers, whilst closely related molecules effected crosslinking at a much lower rate. Cyclopentan-1,2-dione was also shown to be a reactive crosslinking agent. The efficiency of methylglyoxal and glutaraldehyde at crosslinking is thought to be related to their ability to form stable heterocyclic compounds that are the basis of protein crosslinks. The reasons for the striking reactivity of these two compounds, compared to closely related structures is explained by subtle balances between competing pathways in a complex reaction network.


>>>>>>>>>>>>>>>>>>>>>>>>>>>>
BrooklynDodger(s) comments: Most toxicology textbooks and writings about risk assessment mention "structure activity relationships" as a method of hazard identification. There is some funding for computational toxicology. The Dodger(s) think(s) this is wasted space and money in occupational, environmental (including food) toxicology. It reflects the pharmacology roots of many toxicologists; pharmacologists who use this for suggesting "me too" drugs. Some of the key agents identified in recent years demonstrate the failure of this approach to spur research before the fact. Maybe the best known is n-hexane, methyl-n-butylketone (2-hexanone), 2,5-hexadione sequence causing peripheral neuropathy. The Dodger(s) haven't seen a convincing explanation of why n-heptane is not known to cause the peripheral neuropathy demonstrated in the laboratory after people got sick. Butadiene, chloroprene and isoprene (1,3 dienes) are carcinogenic by inhalation, especially butadiene with great potency.

Which brings us to popcorn lung. When diacetyl (2,3-butadiONE) was first proposed as a cause of popcorn lung, the Dodger(s) were skeptical. To a toxicologist, it didn't look like much, it was GRAS, and it was even in beer. So the issue is whether the toxicity of diacetyl could be anticipated from structure activity. The alpha, beta dione structure could suggest some kind of electrocyclic reaction, but it turns out that a much simpler effect was known, the Maillard Reaction, well known in the food chemistry circles. The Dodger(s) didn't find any special reference to diacetyl in the publication abstracted above.

The Dodger(s) would propose that chemicals containing the 1,3-diene or the 2,4-dione moiety be considered toxic until proven innocent. And that future computational toxicology systems be tested against these structure.