Suggestion of suitable animal models for in vivo studies of protein tyrosine phosphatase 1b (PTP1B) inhibitors using computational approaches

Abstract

PTP1B is a prototypic enzyme of the superfamily protein tyrosine phosphatases (PTPs) which are critical regulators of tyrosine phosphorylation-dependent signaling events. It is a highly plausible candidate for designing therapeutic inhibitors of obesity and type 2 diabetes (T2D). In this study, a detailed comparative analysis to reveal the evolutionary relationship of human PTP1B among related vertebrates has been addressed. The phylogenetic trees were constructed with maximum likelihood algorithm by PhyML package on the basis of multiple sequence alignment (MSA) by ClustalΩ and T-coffee. Mutational variability of the sequences corresponding to the 3D structure (pdb: 2vev) was analyzed with Consurf software. The comparative analysis by inhibitor docking to different models was made to confirm the suitability of models. As a result, the PTP1B or PTP non-receptor type 1 homologies show high conservativity where about 70% positions on primary structures are conserved. Within PTP domain (3-277), the most variable positions are 12, 13, 19 and 24 which is a part of the second aryl binding site. Moreover, there are important evolutional mutations that can change the conformation of the proteins, for instance, hydrophilic N139 changed to hydrophobic Gly (mPTP1B); E132 to proline in the hydrophobic core structure or Y46 to cystein in pTyr recognition loop. These variations/differences should be taken into account for rational inhibitor design and in choosing suitable animal models for drug testing and evaluation. Moreover, our study suggests critically potential models which are Heterocephalus glaber, Tupaia chinensis, Sus scrofa, and Rattus norvegicus in addition to the best one Macaca fascicularis. Among these models, the H.glaber and R.norvegicus are preferable over M.musculus thanks to their similarity in binding affinity and binding modes to investigated PTP1B inhibitors.

Keywords: Animal model; Conservativity; Inhibitor docking; PTP1B; Phylogenetic study; Variation.

Figure 1

Multiple sequence alignment (part) of 24 vertebrate PTP1B amino acid sequences. The consensus sequence obtained with the parameters: identity 91.67%, significance 29.17%, gaps 50%. Residues numbered according to hPTP1B.

Figure 2

Unrooted phylogentic tree of 24 species’ PTP1B homologous sequences. Phylograms obtained by PhyML 3.0.

Figure 3

Variations in important binding sites of some sequences – (a) R24 second aryl binding site and pTyr recognition site; (b) R254 & G259 second aryl binding site and Q-loop motif. Conserved residues in these positions are shown in red. The yellow square indicates 6 species that have vigorous variations in these regions.

Figure 4

Mutational variability of 18 aligned PTPN1 sequences in corresponding to PTP1B structure [PDB: 2VEV]. Labeled residues indicate the most variable region(s). The figure was prepared by Chimera 1.8 with Consurf color codes.

Figure 5

Consurf color-coded multiple sequence alignment (part) with conservativity score of 18 PTP1B homologous sequences.

Figure 6

The correlation between the computational interaction energies and the observed binding energies ( ΔG _obs ) calculated from the experimental K _i values of investigated inhibitors. ΔG_obs = RT lnK_i with ΔG_obs: observed free energy change of binding; K_i : inhibition constant; R: gas constant (1.987 cal K ^-1  mol ^-1 ); T: room temperature (298.15 K).

Figure 7

Comparison in the binding site of hPTP1B (left) and of TuPTP1B (right) to the peptidomimetic compound 1. The binding pockets are visualizaed by LigPlot⁺ v.1.4. The ligands and protein side chains are shown in ball-and-stick representation, with the ligand bonds coloured in pink. Hydrogen bonds are shown as green dotted lines with H-bond lengths. Residues with direct/hydrophilic contacts are colored in green with brown backbone whereas ones with indirect/hydrophobic interactions are colored in black and indicated with the red spoked arcs.

Figure 8

Differences in binding sites of hPTP1B (left) and T.chinensis PTP1B (right) to compound 4. The analysis and illustration were made by using LigPlot⁺ v.1.4.

Figure 9

Comparison of the binding pocket of hPTP1B (left) and Tupaia PTP1B (right) for compound 5. The analysis and illustration were made by using LigPlot⁺ v.1.4.

Figure 10

Similarity of binding pocket of rat PTP1B model to hPTP1B leading to superiority of rat model over the mouse model – specific case with compound 4. The analysis and illustration were made by using LigPlot⁺ v.1.4.

Figure 11

Similarity of the binding pocket of rat PTP1B model to hPTP1B leading to superiority of rat model over the mouse model – specific case with compound 5. The analysis and illustration were made by using LigPlot⁺ v.1.4.