Insights into the importance of WPD-loop sequence for activity and structure in protein tyrosine phosphatases

Abstract

Protein tyrosine phosphatases (PTPs) possess a conserved mobile catalytic loop, the WPD-loop, which brings an aspartic acid into the active site where it acts as an acid/base catalyst. Prior experimental and computational studies, focused on the human enzyme PTP1B and the PTP from Yersinia pestis, YopH, suggested that loop conformational dynamics are important in regulating both catalysis and evolvability. We have generated a chimeric protein in which the WPD-loop of YopH is transposed into PTP1B, and eight chimeras that systematically restored the loop sequence back to native PTP1B. Of these, four chimeras were soluble and were subjected to detailed biochemical and structural characterization, and a computational analysis of their WPD-loop dynamics. The chimeras maintain backbone structural integrity, with somewhat slower rates than either wild-type parent, and show differences in the pH dependency of catalysis, and changes in the effect of Mg²⁺. The chimeric proteins' WPD-loops differ significantly in their relative stability and rigidity. The time required for interconversion, coupled with electrostatic effects revealed by simulations, likely accounts for the activity differences between chimeras, and relative to the native enzymes. Our results further the understanding of connections between enzyme activity and the dynamics of catalytically important groups, particularly the effects of non-catalytic residues on key conformational equilibria.

Fig. 1. (Top) PTPs harbor a number of conformationally flexible loops decorating the active site, including the phosphate-binding P-loop (shown in green) and the WPD-loop (shown in orange). The WPD-loop exhibits two distinct conformations: an open, non-catalytic conformation, and a closed, catalytically active one, shown here complexed with the transition state analog (TSA) vanadate. The vanadate ion is shown in spheres and key catalytic side chains are shown in sticks. (Bottom) during the PTPase-catalyzed reaction, which involves a two-step reaction mechanism, the WPD-loop closes toward the P-loop. This brings a conserved aspartic acid into position to protonate the leaving group, followed by a subsequent rate-determining step where the same residue acts as a general base to activate a water molecule in the hydrolysis of the phosphocysteine intermediate. This figure is adapted from ref. . Copyright 2021 American Chemical Society.

Scheme 1. Kinetic scheme for the PTP catalytic cycle including WPD-loop equilibria. The catalytically functional WPD-loop closed conformational states are shown in blue.

Fig. 2. Active site hydrogen bonding differences between (A) ligand-free (PDB ID 2CM2 (ref. 78)) and (B) ligand-bound (PDB ID 3I80 (ref. 23)) WT PTP1B. The WPD-loop is shown in orange, and the phosphate-binding P-loop is shown in green. The sidechain conformation of E186 changes between the open and closed forms of the WPD-loop.

Fig. 3. Structural differences in chimeras compared to parent enzymes. WT PTP1B and YopH structures are shown in panels (A), (C) and (E), while Chimera 3, 4, and 7 structures are shown in panels (B), (D) and (F). WT PTP1B is colored in orange, WT YopH in dark green, Chimera 3 in blue, Chimera 4 in tan, and Chimera 7 in light green. (A) PTP1B and YopH have very similar active site secondary structures, except for the loop structure containing K116 in PTP1B that is absent in YopH. (B) Chimeras 3, 4, and 7 have retained structural integrity in the active site relative to the parents; however, unlike the parent enzymes, the ligand-free form of Chimera 3, but not 4, indicates a shift to a more thermodynamically favorable closed conformation of the WPD loop. The loop of Chimera 7 is also closed, but, as discussed in the text, this likely results from the presence of a molecule of HEPES buffer in the active site. (C and D) Structures of WT PTP1B and YopH, as well as Chimeras 3, 4 and 7, in complex with VO₄ revealed the formation of transition state analogs (TSAs) for the second hydrolysis step (Fig. 1). These TSAs possess a trigonal bipyramidal geometry, in which the active site cysteine is one of the apical ligands. The similarity of these transition state analogs shows the chimeras are able to assume analogous conformations to those of the parents to stabilize the associated transition states. (E and F) A comparison of WO₄-bound structures of (E) WT PTP1B and YopH and (F) Chimeras 3 and 4 show a different sidechain conformation of Asp181 than the corresponding residue in native YopH. The sidechain of K116 and the presence of a magnesium ion is suspected to direct the Asp carboxylate away from the active site in the chimera structures.

Fig. 4. The pH rate profiles for WT PTP1B, WT YopH, and Chimeras 3, 4 and 7. Retained bell-shaped profiles for the chimeras imply maintained general acid catalysis and therefore functional WPD-loop closure. Note the logarithmic scale on the y-axis. Kinetic data for the PTP1B chimeras and WT enzyme were all obtained from thawed aliquots of protein that had been frozen and stored under the conditions described in Materials and methods.

Fig. 5. (A) Experimental and calculated reaction free energies for WT PTP1B and YopH and the four chimeras studied in this manuscript. Experimental values are obtained from the k_cat values whilst calculated values are obtained from our EVB simulations of the hydrolysis reaction (see the Materials and methods). Errors for the calculated values are the standard errors of the mean from 30 EVB replicas per system. (B) The largest per residue differences in electrostatic contributions to transition state (TS) stabilization (difference between the residue contribution in each chimera relative to WT PTP1B) as determined by our EVB calculations. The absolute values of all significant (>0.1 kcal mol⁻¹) side chain contributions to TS stabilization can be found in Fig. S12 and Table S8.

Fig. 6. (A–E) Free energy landscapes for each enzyme projected from the first two time-lagged independent components (TICs), which describe the slowest motions of the WPD-loop and E-loop residues (all enzymes are projected on the same TICs). The centroid of the clusters determined from our Markov state models (MSMs) are indicated on each plot with C_x and O_x referring to a closed and open WPD-loop conformation respectively, and Int_x for an intermediate between the closed and open states. (F) Representative structures of the 7 clusters identified from the MSM generated for WT PTP1B, with structures colored according to whether they belong to the closed, open or an intermediate WPD-loop conformational state. The P-loop is colored in yellow for reference.

Fig. 7. (Left panels) color mapping of the differences (WT–chimera) in the calculated per residue C_α RMSF (ΔRMSF) for the simulations of the closed WPD-loop. Color mapping was performed from blue (positive ΔRMSF) through to white (0 ΔRMSF) to red (negative ΔRMSF). In practice, a blue residue would mean increased rigidity for the given chimera over WT PTP1B and vice versa (see Fig. S16 for this data in graphical form). Note, however, that the differences observed in simulations initiated from the closed conformation, shown here, are far more subtle than those observed in simulations initiated from the open conformation of the WPD-loop, shown in Fig. S14 and S15. (Right panels) differences in the hydrogen bonding network between WT PTP1B and each chimera for simulations of the closed WPD-loop conformation. Hydrogen bonds with a higher occupancy in the WT are shown as black cylinders between the donor and acceptor atoms, with red cylinders used to indicate H-bonding interactions which have a higher occupancy in a given chimera. The width of the dash indicates the magnitude of the difference in the occupancy of the hydrogen bond between the two enzymes. The P-, Q- and E-loops are colored green, magenta, and orange respectively, with the WT PTP1B WPD-loop colored cyan and the chimera WPD-loops colored yellow.