Fig. 4
Active ERK2 phosphorylates six (S/T)P-sites in hNHE1cdt in a distinct order and with different kinetics. a 15N,1H-HSQC spectra of unphosphorylated (blue) and ERK2 phosphorylated hNHE1cdt WT (red). Addition of catalytic amounts of aERK2 to 15N-labelled hNHE1cdt resulted in diagnostic changes in chemical shifts evident of specific S/T-phosphorylation events. b Zoom on S/T and SP/TP regions of a. Red labels/arrows indicate phosphorylation-induced peak shifts, and blue arrows/labels peak shifts of neighbouring residues sensing phosphorylation. c Time courses of hNHE1cdt phosphorylations and order of peak appearances. d Phosphorylation time courses. Zoom i–v on boxes in b. Peaks of unphosphorylated (left panel) and phosphorylated states (right panel) simultaneously disappear/appear with time, in a distinct order. Dashed lines encircle multiple peaks of the same residue reporting on distinct phosphorylation states. Stars indicate position of intermediates. e Relative positions of ERK2 phosphorylation sites (stars) to D-domains and the F-sites in hNHE1cdt f Changes of peak intensities with time reported on the phosphorylation rates at the individual sites. S693 and T779 were the first and most reactive aERK2 phosphorylation sites in hNHE1cdt, followed by S723 (I1), then S785 simultaneously with S726 (I2), followed by dually phosphorylated S723 and S726, and lastly S771. g Kinetics of T779 phosphorylation shown by peak disappearance of D776 and V777 concomitant with peak appearance for V777* and T779P. Apparent rate constants can be extracted by fitting the unphosphorylated disappearing peak or the appearing phosphorylated peak of either the phosphorylated residue or close neighbours, and should, for a two-state reaction, be the same irrespective of which peak is used for fitting. Two-state behaviour was observed for S693 and T779. h–j Phosphorylation of S723 and S726 showed more complex behaviour with interlinked rates. These residues are so close that a neighbouring phosphorylation event would influence their chemical shifts, leading to the observation of intermediates, i.e. phosphorylation of the neighbour in the self-unphosphorylated state, and vice versa (see also panel d). The observation of two intermediates I1 and I2 suggested parallel phosphorylation of S723 and S726. Both peaks of the dually unphosphorylated state disappeared with faster rates than the final peaks from the dually phosphorylated state appeared. Thus, the apparent phosphorylation rates of each site were highly dependent on the phosphorylation state of the neighbour, i.e. k 1 and k 4 as well as k 2 and k 3 are not identical. I1 appeared first with a rate similar to the fast decay of the unphosphorylated states (k 1, see also Table 1), whereas I2 appeared concomitantly with the dually phosphorylated state but much slower (k 2 and k 3) and with low intensities. Thus, although both orders of phosphorylation were observed, the main path was via intermediate I1, i.e. phosphorylation of S723 first, followed by phosphorylation of S726. k The weak peaks reporting on the unphosphorylated states of S771 and S785 initially gained intensity before starting to decrease due to phosphorylation (shown for S771). As peak intensities are strongly dependent on dynamics, this observed increase may result from altered dynamics caused by the nearby phosphorylation of T779. l Apparent rate constants for the individual phosphorylation sites and the effect of D-domain and F-site mutations. Docking site mutations do not change the order of phosphorylation events, yet modulate the individual rates in a distant dependent manner. Based on single measurements and standard deviations from the exponential fits, the apparent rates are significantly different except for S693 WT compared to D3-AXA, S771 WT compared to F2-AA, and S785 WT compared to D1D2-(AXA)2 (one-way ANOVA)