Insights into evolutionary interaction patterns of the 'Phosphorylation Activation Segment' in kinase

We are interested in studying the phosphorylation of the kinase activation loop, distinguishing the passage from the unphosphorylated to the phosphorylated form without allostery. We performed an interaction study to trace the change of interactions between the activation segment and the kinase catalytic core, before and after phosphorylation. Results show that the structural changes are mainly due to the attraction between the phosphate group and guanidine groups of the arginine side chains of RD-pocket, which are constituted mainly of guanidine groups of the catalytic loop, the β9, and the αC helix. This attraction causes propagation of structural variation of the activation segment, principally towards the N-terminal. The structural variations are not made on all the amino acids of the activation segment; they are conditioned by the existence of two beta sheets stabilizing the loop during phosphorylation. The first,β6-β9 sheet is usually present in most of the kinases; the second, β10-β11 is formed due to the interaction between the main chain amino acids of the activation loop and the αEF/αF loop.


667
©Biomedical Informatics (2019) on their catalytic site indirectly, by the displacement, to blocking key elements of the catalytic, or regulation domain [3]. The kinase can also be catalytically inactive due to the displacement of key elements participating in the catalytic domain, which is then allosteric regulation [3]. The activation of some kinases requires phosphorylation of their activation segment, particularly their activation loop which is a real phosphorylation regulatory site [3].The activation loop is involved in the kinase regulations which gives an essential role in the adoption of an active conformation. SLK phosphorylation at T183 and S189 levels is known [4].
Phosphorylation plays an important role in the activation and signalling processes of SLK [4]. Besides, the regulatory effect extends to the catalytic environment by the activation segment.
Researchers have recorded a deregulation of MAPK at low pH due to a structural rearrangement of the activation segment [5]. The latter may also have an auto-inhibitory action because of substrate blocking and stabilisation of an inactive conformation of the αC helix, such as the case of NDR1 in its non-phosphorylated state [6].
What is more, one of the most important mechanisms related to the activation segment is auto phosphorylation, where kinases dimerize and cause phosphorylation of the activation segment [7, 8,9]. So, the exchange of activation segment has proven to be a very important action in the auto phosphorylation mechanism [11], which can occur in one or both directions [12]. It is remarkable that autophosphorylation does not recognize the consensus sequences of substrates, which gives us a unique phosphorylation mechanism [10]. Concerning the physical aspect, the activation segment starts from the conserved motif DFG, which is the part of the activation loop, to the conserved motif APE (Figure 1) [3]. The N-terminal and C-terminal anchor points of the activation segment take place despite the folding of the activation loop under the effect of the phosphorylation of one of the kinase residues [3]. In fact, in many kinases, the interaction of the phosphorylated residue with the activation loop gives two states. One concerns the aspartate of the DFG motif pointing in the ATP binding site and coordinates two Mg 2+ ions (active state) DFGin, and the other concerns DFG pointing out of the ATP binding site (inactive state, DFGout) [13]. When the N-terminal or the C-terminal anchor points are disturbed, the kinase is in an inactive state. In general, it is commonly accepted that in the cell environment, kinases pass between the catalytically active conformation and inactive conformation [13].
There are also other secondary phosphorylation sites, both upstream and downstream of the primary site, improving the regulation of kinase activity . We report here the correlation of the "phosphorylation activation segment" P.A.S. and more specifically activation loop, with principal residues in kinase activity. We further explain the structural changes due to the P.A.S -kinase catalytic core interaction, benefiting from increasingly abundant databases of protein structures, and avoiding expensive quantum or dynamic calculations. Hence, a differential structural analysis between phosphorylated and non-phosphorylated P-kinases (kinase catalytic core) is completed.

Dataset analysis:
We have studied the conservation of secondary structures in the activation segment for selected P-kinases, before discussing the difference between compared 3D structures of the unphosphorylated structure considered as a reference, and phosphorylated structure for the extraction of the phosphorylation effect. For this, alignment of sequences, and a superposition of their secondary structures have been carried out.
We have performed an analysis of the cartesian deviations of the backbone chain by RMSDbb, side chain, by RMSDsc, and also the deviations of the dihedral angles σ (φ), σ (ψ) of each amino acid of our kinase to detect structural disparities between the phosphorylated and non-phosphorylated activation loop structures [30]. We have retained only variations which exceed 2 Å for RMSDbb and RMSDsc [31] and 20° for the dihedral angles [32]. Finally, to identify the impact of phosphorylation of the activation loop on the modes of its interaction with the remaining P-kinase motifs, we have attempted to enumerate the amino acids of activation loop which are essential for such interactions, using the PICI script [33].

Results and discussion:
The structures retained by the filters used during the study of the structural differentiation between P-kinase, before and after phosphorylation of the activation loop, are listed in  (2019) bonds with R50 (αC helix), R126 (catalytic loop), R150 (β9) and Y180 (loop αEF/αF). T160 is exhibiting deviations of the backbone chain (RMSFbb) and side chain (RMSFsc) exceeding 3Å and a deviation of the dihedral angle σ (ψ) of about 80°; the same interactions were obtained by L. N. Johnson [39]. We can prove the flexibility of the activation loop amino-acids compared to all explored amino acids (from F4 to F286), due to their ability to move easily to other amino acids according to specific interactions. The deviation will rise to the disappearance of interactions between the amino acids of activation loop (V156, R157, T158) and those (G176, C177, K178, Y180) of the loop αEF/αF, which will leads to the total demolition of the secondary structure β10-β11. GSK3β kinase catalytic domain mainly shows modifications of the dihedral angles σ (φ) and σ (ψ) at the level of the amino acids: R209 to Y216 of the activation loop. Following these structural modifications, the propagation of the phosphorylation affects only a few amino acids of the activation loop, from the phosphorylated residue to the N-terminal R209, and does not affect any amino acid of the C-terminal of the activation segment after phosphorylation. We also note that there is an insufficient variation in interactions, whereas there is a conservation of secondary structures (β6-β9) and (β10-β11). Kinase reactive subdomains show no significant interaction variation.  I146, I147, H148, R149). The dual phosphorylation causes a broad propagation along the activation loop, but also some amino acids Mg 2+ loop (A172, R173, and H174) since the two β sheets (β6-β9 and β10-β11) will not support the activation loop of the unphosphorylated structure. This propagation gives rise to a release of the active site, following dilation of the activation loop. The kinase emphasizes substantial cartesian modifications at the level of the activation segment, ranging from 3.1 to 12.1 Å, which caused an increase in the number of interactions, that is rational with its importance in the stability of the structure.

3-Phosphoinositide-dependent kinase 1 (PDK1) is a kinase of the AGC group of the PKB family which contains a PH domain [48].
Since the pleckstrin homology (PH) domain promotes binding PDK1 to the plasma membrane [49]. It is involved in a large variety of processes, including cell proliferation, differentiation, and apoptosis [50]. Auto phosphorylation of the activation segment at S241 is necessary for PDK1 activity [50]. Phosphorylation of PDK1 at S241 deviates its RMSD in the order of 3.6 Å at the main chain and 4.8 Å in the side chain, and of the order of 144.8° at the dihedral angle σ (ψ). This deviation induces an attraction between the phosphate group with the guanidine groups of the two R129 arginines of the Cα and R204 of the catalytic loop. This last result coincides with the results published by Komander, but which adds interaction with K128 of β9 and T126 of Cα helix [51]. The phosphorylation of PDK1 kinase does not give rise to a significant structural change, following conservation of secondary structures, but it gives rise to a change of interactions between Q236, A237, R238 of the activation loop and C260, S258, A259 of the loop αEF/αF. This phosphorylation is followed by some changes in the interactions of several amino acids between the activation loop and the catalytic body of the kinase, marking the disappearance of hydrogen interaction of R204 of the HRD motif, with A239 and a change at R238 by hydrogen interactions with S258 and A259. Additionally, there is an appearance of hydrogen bonds at the amino acid level F242 with R204.
Spleen tyrosine kinase (Syk) is a cytoplasmic tyrosine kinase SYK family In case of AURORA kinase monophosphorylation at T287, the activation loop remains stabilized with the beginning of the formation of the β10-β11 sheet and the β6-β9 sheet. This stability is reflected in the small fluctuation of the entire segment. All these fluctuations do not prevent some structural changes scattered between the two lobes and marked by the conservation of multiple interactions between Q177 and W277, and the disappearance of a hydrogen interaction with G276, which belongs to the DFG motif. As regards monophosphorylation at T288, in an earlier work by Bayliss, we have found a difference in the analyzed structure,, since it is complexed with TPX2. We note that we do not have the same interactions with T288, whereas there is the appearance of interactions that resemble the case of T287 phosphorylation. This finding allows us to say in the case of the work of Bayliss, that these interactions are not due to the phosphorylation but to the complexation of TPX2 [1]. For the doubly phosphorylated structure, it is noted that R285, of the phosphorylated structure T288, binds with R180 and R255, the same goes with those where the phosphorylated structure binds to T287 as if there exists some competition between the T287and T288 phosphorylated structures.

Conclusion:
Results show that the structural adaptation of the activation loop after phosphorylation is mainly due to hydrogen bonds formed between the phosphate group with amino groups (R-NH2) of lysines or with guanidine groups (R-CH5N3) of arginines. The multiplicity of these phosphate groups interactions represent anchors, which stabilize the activation loop. The strongest anchor belonging to the N-terminal concerns arginines or lysines of the Cα and β9 and also arginines of the HRD motif, as well. The second remarkable anchor, which belongs to C terminal, concerns arginines or lysines of the αEF and leads to a rearrangement of the loop P+1 amino acids. Moreover, these anchors allow interactions propagation towards the N-terminal lobe of the activation segment. Further, it should be noticed that the activation loop stability is conditioned by the existence of interactions between the main chains of the activation loop and the αEF/αF loop. Besides, the structures containing β6-β9 sheets in the N-terminal or β10-β11 in the middle of the activation loop strengthen its stability during phosphorylation. Finally, we find out that interactions' variations are acting on the most essential regions at the level of kinases with the hydrogen bonds, affecting the most conserved motifs in kinases: DFG for the SYK kinase, APE for the PAK1 kinase and HRD for CDK2, MAPK14, PDK1, PAK1, AurA as well as other ones which react on the important loops, like GSK3β on Mg 2+ loop and MAPK14 on P+1 loop.

Acknowledgments:
We want to thank Mr. R. Boulguid for support.

Author contributions:
Adil Ahiri did data analysis and interpretation, article writing. Aziz Aboulmouhajir did data analysis and interpretation, article reviewing, article writing; Crtomir Podlipnik did data analysis and interpretation, article reviewing and Hocine Garmes did article reviewing. The open access charges for this article is fully sponsored by Biomedical Informatics (P) Ltd, India.

Conflicts of Interest:
The authors declare that they have no conflict of interest.