The mechanism of a one-substrate transketolase reaction

Abstract Transketolase catalyzes the transfer of a glycolaldehyde residue from ketose (the donor substrate) to aldose (the acceptor substrate). In the absence of aldose, transketolase catalyzes a one-substrate reaction that involves only ketose. The mechanism of this reaction is unknown. Here, we show that hydroxypyruvate serves as a substrate for the one-substrate reaction and, as well as with the xylulose-5-phosphate, the reaction product is erythrulose rather than glycolaldehyde. The amount of erythrulose released into the medium is equimolar to a double amount of the transformed substrate. This could only be the case if the glycol aldehyde formed by conversion of the first ketose molecule (the product of the first half reaction) remains bound to the enzyme, waiting for condensation with the second molecule of glycol aldehyde. Using mass spectrometry of catalytic intermediates and their subsequent fragmentation, we show here that interaction of the holotransketolase with hydroxypyruvate results in the equiprobable binding of the active glycolaldehyde to the thiazole ring of thiamine diphosphate and to the amino group of its aminopyrimidine ring. We also show that these two loci can accommodate simultaneously two glycolaldehyde molecules. It explains well their condensation without release into the medium, which we have shown earlier.

shows the wide range scale (480-550Da) of a typical ESI-MS spectrum of ThDP and its intermediates which are formed in the reaction of holoTK with HPA in the absence of NaCNBH3. To make this spectrum more readable we deleted minor peaks with amplitude below 20% of the 485.065 peak amplitude, even considering the risk to lose some important peaks which sometimes go as minor (for example, the peaks 483.050 and 467.056). However, despite this sacrifice, there are still enormous amounts of foreign peaks here because of the high content of polluting traces, coming from reagents, decaying protein, column and products of their interaction. Mass spectra in the presence of NaCNBH3 are similarly unreadable (not shown). Without refinement of these spectra we could only investigate known and expected peaks 483.050, 485.065, 487.081, which were reproducibly found indeed in most of initial experimental spectra after extending the scale ( Fig. 1 of main text).
However none of multiple spectra had expected masses of tetrose derivatives of ThDP (545.087, 543.072 and 527.077). We had to clarify, whether they just disappeared or were converted into compounds with different masses. To answer this question we excluded foreign and insignificant peaks via the following refinement procedure. First, we deleted from the mass spectrum given on Fig. 1SA all masses which were also present in mass spectra of buffer and holoTK control. However, the resulted differential spectrum still had too many peaks (Fig.1SB), most of which we could not identify.
On the next step of refinement we used the criteria of reproducibility: the essential peaks must be present in all experiments conducted under the same conditions while at different dates with different batches of reagents, columns, protein etc. Fig. 1SC shows the synthetic spectrum obtained by uniting (intercrossing) two independent spectra obtained from two independent experiments of the same settings (without NaCNBH3) but with several months interval between them. Only peaks whose masses coincide at 3 digits after the dot are presented in the synthetic double spectrum (Fig.1SC). As we see, the peak quantity dropped significantly. The same refinement procedure was performed for cyanborohydride positive spectra with the same excellent result (Fig.1SD). Now we had 9 refined peaks in total, and three of them (483, 485 and 487) we identified and discussed in the main text of the manuscript. From the remaining 6 peaks three, 501.060, 527.058 and 543.053, were fully reproducible both in non-reducing and reducing conditions as it is seen on tetra-spectrum, obtained via crossing two double spectra Fig.1SC +Fig.1SD. 4 (Fig.1SE).
The first of these three peaks, mass 501.0603 can be identified with high precision of 0.0005 as hydrated 483.050 mass: 483.0503+18,0105= 501,0608.

Supplementary Fig. 1S
Mass spectra of intermediates formed in the one-substrate reaction.
A -Typical ESI-MS spectrum of ThDP and its intermediates formed in reaction of holoTK with HPA in the absence of NaCNBH3 (experiment 1). Conditions are same as for the Fig.1  B -Same as A, but from initial spectrum are deleted all masses which were also present in mass spectra of buffer and of holoTK control.
C -The intercrossing of two B-spectra obtained in two independent experiments (experiment 1, black lines + experiment 2, blue lines) of the same setting without NaCNBH3. Only peaks whose masses coincide at 3 digits after the dot are presented in the synthetic double spectrum. D -Same as C but in presence of NaCNBH3, the intercrossing of experiments 3 (green line) and 4 (orange line).
E -Coincident peaks from all four experiments are presented by tetra-spectrum, the intercrossing of C and D. Experiments 1-4, respectively: black, blue, green and orange lines.