Nonenzymatic reactions between glucose and proteins (Maillard reactions) are an unavoidable feature of life. These reactions have mostly adverse effects on proteins by inhibiting enzymes, interfering with receptor function, and crosslinking structural proteins such as collagen. ^{1, 2} Traditionally, these reactions were thought to be totally nonenzymatic with no mechanisms to reverse or control them. This changed in the year 2000 with the discovery of Fructosamine-3-Kinase (FN3K) ^{3, 4}

This enzyme phosphorylates the first stable intermediates of the Maillard reaction, protein-bound fructosamines, to fructosamine-3-phosphates. Since these phosphoesters are intrinsically unstable, this reaction leads to decomposition of the glycation adducts, regeneration of an unmodified protein, and concomitant production of 3-deoxyglucosone. ⁴

When FN3K was isolated in our laboratory, it copurified with another closely related protein that has an identical MW but a slightly lower isoelectric point ( Fig. 1). ⁴ At the time, this second protein was assumed to be an isozyme of FN3K with the same substrate specificity, namely D-fructosamine. Subsequent work revealed, however, that this enzyme, renamed Frucose-3-Kinase-Related Protein (FN3KRP), is quite distinct from FN3K because it does not phosphorylate D-fructosamines. Extensive work by Van-Schaftingen et al. has shown that FN3KRP phosphorylates a number of other Amadori products on the C3-hydroxyl. These include D-erythrulosamine, D-ribulosamine and D-psicosamine. ⁵ While none of these ketosamines, nor their aldose precursors, are found in cells, Van Schaftingen et al. have suggested and documented ways in which they can be produced from erythrose-4-phosphate and ribose-5-phosphate by a combination of Maillard reactions of the phosphorylated metabolites and dephosphorylation of the resulting Amadori products by specific phosphatases. ^{6, 7}

Subsequent to Van Schaftingen’s work, using recombinant FN3KRP, I was able to show that, in addition to the compounds identified above, FN3KRP also phosphorylates glucitolamines such as N-methylglucamine (meglumine). ⁸ However, unlike the other substrates, this phosphorylation occurs on the C4-hydroxyl. This was a surprising result, the significance of which was not apparent until a recent closer re-examination of data on purified FN3K, as shown in Fig.1. As can be seen in figure, phosphorylating activities are evident with both FN3K and FN3KRP, even though the substrate used here is nominally just fructosamine. The activity of FN3K is easy to understand; however, the source of FN3KRP activity is not clear.

The most likely explanation for the phosphorylating activity of FN3KRP is that while the substrate used was nominally fructosamine, it was not completely homogeneous. As assessed at that time, it was only 95% pure but it probably also contained a compound (or compounds) that would be a substrate for FN3KRP. Given that the radioassay for FN3K used in these experiments contained saturating concentrations of fructosamine (1 mM vs. Km ~ 25 μM), it is likely that the FN3KRP substrate, even if present as only a small fraction in this preparation, would be phosphorylated sufficiently (by FN3KRP) to be detected.

Several candidates for the identity of this “contaminant” substrate are shown in Fig. 2, including the following: I)

A Schiff base of the amine with glucose