Pryor and Porter postulated several alternative mechanisms (2) that were inspired by the observation that the formation of HNE can be catalyzed by iron which "suggests the intermediacy of a hydroperoxide which is reduced to form an alkoxyl radical."  An unlikely possibility involves the generation of a hydroperoxy dihydropyran intermediate.  One difficulty with this mechanism is the geometric constraint of the C=C bonds in the alkoxyl radical intermediate which clearly would strongly disfavor cyclization.

A more reasonable scenario postulated by Pryor and Porter, involves an Epoxy Hydroperoxide Fragmentation (2).  The key cyclization of an allylic alkoxyl radical to a carbon-centered allylic radical had been detected earlier (3) and has subsequently been recognized as a "key step in hydroperoxide-amplified lipid peroxidation (4).  The 12,13-epoxy-9-hydroperoxide is a known product from the reaction of 13-HPODE and iron or hydrolysis of the methyl ester (5) which is generated during autoxidation of methyl linoleate (6).  Several alternative mechanisms, both homo and heterolytic, were suggested for the conversion of this hydroperoxide into 3,4-epoxynonanal that is known to rearrange readily to give HNE.

Yet another mechanism was proposed inspired by the detection of 10,13-dihydroxyoctadecadienoic acid (DHODA) as a product from the oxidation of LA.  This Hydroxyhydroperoxide Mechanism generates HNE by beta-scission of a key hydroxyalkoxyl radical that is produced from a hydroxy hydroperoxide (7).  Formation of the hydroxy hydroperoxide could involve hydrogen abstraction from an intermediate 13-HODE.  We note that an intramolecular H• transfer is geometrically favorable (vide infra), bolstering the likelihood of such a mechanism.  Besides HNE, the beta-scission also would generate a vinyl radical that delivers a vinyl peroxy radical by reaction with oxygen, and ultimately provides 9-oxononanoic acid (ONA) via ONA enol. 

Interestingly, an enzymatic route to these same products from LA was reported recently (8).  Thus, 9-HPODE, generated by the action of lipoxygenase on LA, is cleaved by the action of a hydroperoxide lyase to produce ONA and 3(Z)-nonenal (Enzymatic Pathway).  Alkenal oxygenase promoted oxygenation of this b,g-unsaturated aldehyde then delivers HNE. 

We postulated (9) a Peroxycyclization-Dioxetane Fragmentation Mechanism that predicts a competition between peroxycyclization, that leads to aldehyde fragmentation products, and H• transfer that produces the hydroperoxy ODAs 13-HPODE and 9-HPODE.  This scenario contrasts with that envisioned in the other mechanisms proposed for HNE generation that all consider one or the other of these hydroperoxides to be intermediates leading to HNE.  However, this difference between the mechanisms could easily be hidden if H• transfer is readily reversible.  Thus, hydroperoxides could serve as a reservoir of peroxy radicals if H• abstraction from the hydroperoxy oxygen occurs readily.  This is exemplified by the use of a hydroperoxide as the precursor for a peroxy radical in a Peroxycyclization Model Study presented below.

Peroxycyclization of an intermediate 5-peroxyeicosatetraenoyl radical followed by fragmentation of a dioxetane intermediate is a possible Peroxycyclization Route to OV-PC from AA-PC.  The hypothetical alternative pathways outlined above illustrate the potential competition for various isomeric pentadienyl and peroxyeicosatetraenoyl radical intermediates that may be crucial for a thorough understanding of factors influencing the generation and evolution of the end products of free radical-induced lipid oxidation.

As noted above, we recently postulated a possible mechanism for oxidative fragmentation of lipids: peroxycyclization to generate a dioxetane intermediate that fragments to generate two carbonyl groups (9).  Contemporaneously, this mechanism was recognized by others, and a model study was reported (see Peroxycyclization Model Study) in which one-electron oxidation of an allylic hydroperoxide produces acetone and chemiluminescence, presumably by cyclization of an alkylperoxyl radical to a dioxetane radical and subsequent dioxetane fragmentation that generates triplet acetone (10).  The generation of excited state carbonyl product is especially noteworthy as it is an expected consequence of orbital symmetry considerations.  This observation strongly supports the involvement of a dioxetane fragmentation because there is ample precedent for this unique phenomenon.

Although the model study supports the feasibility of a peroxycyclization dioxetane-fragmentation mechanism for the formation of HNE as well as HOOA and HODA derivatives, it remains to be demonstrated that lipid peroxy radicals actually undergo such a cyclization-fragmentation.  Furthermore, the simultaneous or exclusive operation of alternative mechanisms for the fragmentations which accompany free radical-induced lipid oxidations, have not been ruled out.