Conformational Dynamics and Catalytic Backups in a Hyper-Thermostable Engineered Archaeal Protein Tyrosine Phosphatase

Abstract

Protein tyrosine phosphatases (PTPs) are a family of enzymes that play important roles in regulating cellular signaling pathways. The activity of these enzymes is regulated by the motion of a catalytic loop that places a critical conserved aspartic acid side chain into the active site for acid-base catalysis upon loop closure. These enzymes also have a conserved phosphate binding loop that is typically highly rigid and forms a well-defined anion binding nest. The intimate links between loop dynamics and chemistry in these enzymes make PTPs an excellent model system for understanding the role of loop dynamics in protein function and evolution. In this context, archaeal PTPs, which have evolved in extremophilic organisms, are highly understudied, despite their unusual biophysical properties. We present here an engineered chimeric PTP (ShufPTP) generated by shuffling the amino acid sequence of five extant hyperthermophilic archaeal PTPs. Despite ShufPTP's high sequence similarity to its natural counterparts, ShufPTP presents a suite of unique properties, including high flexibility of the phosphate binding P-loop, facile oxidation of the active site cysteine, mechanistic promiscuity, and most notably, hyperthermostability, with a denaturation temperature likely >130 °C (>8°C higher than the highest recorded growth temperature of any archaeal strain). Our combined structural, biochemical, biophysical and computational analysis provides insight both into how small steps in evolutionary space can radically modulate the biophysical properties of an enzyme, and showcase the tremendous potential of archaeal enzymes for biotechnology, to generate novel enzymes capable of operating under extreme conditions.

Keywords: archaea; catalytic backups; protein conformational dynamics; protein tyrosine phosphatase; sequence shuffling.

Figure 1.. Overview of the PTP catalytic mechanism.

Two-step mechanism catalyzed by PTPs, illustrating acidloop conformational changes, as well as the position of the catalytic aspartic acid. From Ref. , published by the American Chemical Society, under a CC-BY license. Copyright © 2024 The Authors.

Figure 2.. Key loops decorating PTP active sites.

(A) A comparison of the open and closed conformations of the acid-loop of the archetypal member of the protein tyrosine phosphatase superfamily, PTP1B, highlighting also the Q- and P-loops. (B) An overlay of the IPD- and P-loops of the archaeal PTPs, SsoPTP (PDB ID: 7MPD) and TkPTP (PDB IDs: 5Z5A and 5Z59), in both their unliganded and liganded forms. In both archaeal PTPs, the acid-loop is closed and in the same position in all structures, but the phosphate binding P-loop is flexible and takes on multiple conformations (note that only one conformation of SsoPTP’s P-loop could be captured structurally, the existence of the other was determined through NMR spectroscopy). Panel A was originally published in ref. under a CC-BY license. Published by the American Chemical Society. Copyright © 2021 The Authors.

Figure 3.. Sequence conservation demonstrates the archaeal nature of ShufPTP.

Multiple sequence alignment of characterized PTPs with the synthetic PTP ShufPTP highlights its similarities to thermophilic archaeal PTPs. Shown here is a portion of the sequence alignment among two archaeal PTPs from Thermococci, TgPTP and TkPTP, a third archaeal PTP SsoPTP, the human PTPs SHP-1 and PTP1B, and the bacterial PTP YopH, as well as the corresponding consensus sequence. Sequence alignment was performed using T-Coffee. The conserved IPD- or WPD-regions of the acid loop and the phosphate binding P-loop ((H/V)CX5R(S/T)) are highlighted in purple. Of the non-archaeal PTPs shown here, SHP-1 is an important anticancer drug target, and PTP1B and YopH are two of the most studied PTPs to date.^, ShufPTP shows 90% sequence identity to TgPTP, 85% to TkPTP, 39% to SsoPTP, 25% to SHP-1, 30% to PTP1B and 31% to YopH, respectively.

Figure 4.. ShufPTP adopts a canonical PTP structure.

(A) The active site of ShufPTP consists of four loop motifs, P-loop (Yellow), IPD-loop (blue), Q-loop (green), and E-loop (pink). All solved structures are largely superimposable with key differences seen in the P-loop motif. The vanadate ligand from PDB ID: 9E9U (this work) and key residues are shown in sticks. The dashed line refers to the P-loop region depicted in higher resolution in panels B-E. (B-E) Stick view of the different P-loop conformations and oxidation states. H-bonds are shown in teal, with differences in observed oxidation state, ligand, and conformation are listed to the right. (F) Chemical structure of the three observed oxidation states of C93. Atoms and bonds shown in red highlight the newly formed bonds caused by oxidation of the thiol.

Figure 5.. ShufPTP exhibits a bell-shaped pH-rate profile like extant PTPs.

The retention of the basic limb in the profile for the general acid variant D63N indicates the presence of an alternate general acid. Precipitation of the variant below pH 4.75 precluded acquisition of data at lower pH.

Figure 6.

Differential scanning calorimetry (DSC) of ShufPTP. (A) Profiles of heat capacity versus temperature for solutions of ShufPTP and thioredoxin at 1 mg/mL and pH 7. The profiles have been shifted in the y-axis for the sake of clarity The four uppermost profiles correspond to ShufPTP solutions and differ in the upper temperature of the DSC scan, as it is visually apparent. The lowermost profile corresponds to a thioredoxin solution and, unlike the profiles for ShufPTP, shows a prominent denaturation transition (heat capacity peak). (B) Amount of non-aggregated (soluble) protein in solutions of ShufPTP extracted from the calorimetric cell after cooling from a DSC scan. The amount of soluble protein is given as function of the highest temperature reached in the DSC scan.

Figure 7.. Characterizing prospective catalytic backups in ShufPTP.

Potential backups identified by tracking the distances between the phosphate group at the phosphoenzyme intermediate and prospective catalytic carboxylate side chains, extracted from our MD simulations of wild-type and D63N ShufPTP. (A) Kernal density estimates (KDE) of the distances between the phosphate group and prospective catalytic residues, defined as the distance between the P atom of the phosphate group and the closest oxygen atom of the carboxylate side chain of each key residue. (B) The % simulation time the side chains of each of the three most likely candidates (D63, E41 and E132) spent within 6Å of the phosphate group phosphorus atom (shown in blue). Further, to be catalytically viable, it is necessary for a water molecule to bridge the respective side chain and the phosphate group, in order to act as a nucleophile (Figure 1). For each prospective catalytic side chain, we calculated the % of simulation time a water molecule is positioned within both 3.5Å of the carboxylic acid of the respective side chain, and of the phosphorus atom of the phosphate group (red, measured based on distances to the nucleophilic oxygen atom). (C) The dramatic increase in proximity of E132 to the C93 side chain with increased temperature can be attributed to the increased flexibility of the Q-loop. (D) Illustration of the catalytically active conformations of D63 (IPD-loop) and E132 (Q-loop), with a catalytic water molecule bridging the side chain and the phosphate group.

Figure 8.. Empirical Valence Bond evaluation of reaction barrier for the dominant and backup mechanism of ShufPTP.

(A) Representative snapshots of phosphoenzyme intermediate and product state structures extracted from empirical valence bond (EVB) simulations of the hydrolysis reaction catalyzed by wild-type ShufPTP. Shown here are stationary points for the D63-as-base and E132-as-base mechanisms. (B) Calculated ΔG^‡ values (kcal mol⁻¹) for the D63-as-base and E132-as-base mechanisms in wild-type (WT) and D63N ShufPTP were compared to the experimental values obtained from the kinetic data (k_cat, Table 1) for both variants. The red dashed line indicates the experimental activation free energy for the reaction catalyzed by wild-type (WT) ShufPTP, and the blue dashed line indicates the experimental activation free energy for the reaction catalyzed by the D63N ShufPTP mutant. The error bars represent the standard error of the mean on the calculated activation free energies over 20 individual EVB trajectories for each system. The raw data for this figure are shown in Table S2.

Figure 9.. Conformational dynamics of the IPD- and P-loops in the unliganded forms of ShufPTP and TkPTP.

Simulations were initiated from the IPD-loop-closed and low P-loop state of each enzyme. (A) The ensemble of sampled conformations during simulations of ShufPTP and TkPTP at the IPD-loop-closed unliganded state was visualized and colored based on the root mean square fluctuations (RMSF, Å) of loop C_α-atoms. (B) A stabilizing hydrophobic network at the base of the IPD-loop of was identified and compared to the interactions present in TkPTP. This network was obtained by calculations using Key Interaction Networks (KIN). (C) Assessment of the P-loop and IPD-loop conformational ensembles based on kernel density estimation (KDE) analysis of the root mean square deviations (RMSD, Å) of the backbone atoms of the IPD- and P-loops of ShufPTP and TkPTP. These data indicate a less conformationally diverse ensemble in TkPTP than in ShufPTP.

Figure 10.. Comparison of P-loop conformational transitions between ShufPTP and TkPTP.

(A) 2D histograms of the P-loop conformation sampled in our MD simulations of each enzyme, defined based on a root mean square deviation (RMSD)-like metric (Eq. 1 of the main text). The deviation between active (low) and inactive (high) P-loop state X-ray crystal structures (CS) of ShufPTP and TkPTP corresponds to φ and ψ angles of 116.1 degrees and 104.7 degrees respectively. The angle RMSD of two ShufPTP intermediates to active and inactive CS are: 9E9L = (78.6, 95.2) degrees, and 9E9M = (78.6, 71.6) degrees. Simulations were initiated from inactive (high) P-loop states of each enzyme, with the IPD-loop in its loop-closed state. Corresponding analysis initiated from the intermediates and active (low) P-loop state is shown in Figures S8 and S9. The corresponding crystallographic positions of the P-loop are denoted by stars, as annotated on the figure. (B) Illustration of structures from the local maxima of the three states observed on the histogram of ShufPTP, in order to illustrate the key structural differences between the MD intermediate and the inactive (high) state highlighted. (C) Comparison of active (low), intermediate and inactive (high) P-loop conformations of TkPTP and ShufPTP, based on crystallographic data (PDB IDs: 5Z5A, 5Z59, 9E9N, 9E9L, 9E9M and 9E9U).