How Intrinsic Molecular Dynamics Control Intramolecular Communication in the Signal Transducers and Activators of Transcription Factor STAT5

Florent Langenfeld1, Yann Guarracino1, Michel Arock1, Alain Trouvé2 and Luba Tchertanov1,2*

1 Laboratoire de Biologie et Pharmacologie Appliquée, Ecole Normale Supérieure de Cachan, 61 avenue du Président Wilson, 94235 Cachan, France

2 Centre de Mathématiques et de Leurs Applications, Ecole Normale Supérieure de Cachan, 61 avenue du Président Wilson, 94235 Cachan, France

Abstract

Signal Transducer and Activator of Transcription STAT5 is a key mediator of cell proliferation, differentiation and survival. While STAT5 activity is tightly regulated in normal cells, its constitutive activation directly contributes to oncogenesis and is associated with a broad range of hematological and solid tumor cancers. Therefore the development of compounds able to modulate pathogenic activation of this protein is a very challenging endeavor. A crucial step of drug design is the understanding of the protein conformational features and the definition of putative binding site(s) for such modulators. Currently, there is no structural data available for human STAT5, and our study is the first footprint towards the description of structure and dynamics of this protein. We investigated the structural and dynamical features of the two STAT5 isoforms, STAT5a and STAT5b, taken into account their phosphorylation status. The study was based on the exploration of molecular dynamics simulations by different analytical methods. Despite the overall folding similarity of STAT5 proteins, the MD conformations display specific structural and dynamical features for each protein, indicating first, sequence-encoded structural properties and second, phosphorylation-induced effects which contribute to local and long-distance structural rearrangements interpreted as allosteric event. Further examination of the dynamical coupling between distant sites provides evidence for alternative profiles of the communication pathways inside and between the STAT5 domains. These results add a new insight to the understanding of the crucial role of intrinsic molecular dynamics in mediating intramolecular signaling in STAT5. Two pockets, localized in close proximity to the phosphotyrosine-binding site and adjacent to the channel for communication pathways across STAT5, may constitute valid targets to develop inhibitors able to modulate the function-related communication properties of this signaling protein.

PMID: 26717567

Supplementary

The Signal Transducer and Activator of Transcription (STAT) proteins are transcriptional factors mediating a cellular signal transfer from the cytoplasm to the DNA, and regulating the transcription of major genes contributing to cell growth and survival. Two proteins from this family, STAT5a and STAT5b, are involved in the physiological control of apoptosis, cell cycle progression and reactive oxygen species (ROS) production. Either the deregulation of STAT5 activity initiated by phosphorylation events or its overexpression contributes to the progression of hematological malignancies such as chronic myeloid leukemia (CML) [1,2] and mediates resistance to tyrosine kinase inhibitors (TKIs) [3].

STAT proteins share a common structural organization comprising the N-terminal domain, the Coiled-Coil domain (CCD), driving the nuclear import-export, the DNA Binding domain (DBD), recognizing the specific DNA sequences, the Linker domain (LD), the SRC homology 2 domain (SH2), controlling the STAT dimerization), the phosphotyrosyl Tail (p-Tail),  bearing a specific phosphor-tyrosine residue, and the Trans-Activation Domain (TAD), recruiting proteins to form transcription complexes, at the C-terminus (Fig 1 A).

 

The STAT5 activation is initiated by phosphorylation of the p-Tail tyrosine that mediates reciprocal interactions between the monomers through the phosphotyrosyl residue and the SH2 domain [4]. The STAT5 monomer, reported as the major cytoplasmic form of the protein [5], dimerizes upon activation and translocates into the cell nucleus, where it binds to a specific double-stranded DNA and activates the transcription through recruitment of transcriptional protein partners [4]. The phosphorylation event of STAT5 is tightly regulated and its deregulation by mutated upstream proteins promotes the development of different tumors, depending on the signaling pathway (Bcl-xL in CML [6], KIT in mastocytosis [7]). The finding of STAT5 mutants in the large granular lymphocytic (LGL) leukemia patients [8] emphasizes the role of STAT5 in cancers.

The reported inhibitors targeting STAT5 or upstream activators [9,10] show very limited potency and low selectivity within the STAT family. It is therefore a great challenge to develop highly selective molecules capable of controlling STAT5 activity. To apply structure-based methodologies, widely and successfully used for the development of therapeutic agents, a STAT5 structural characterization is a prerequisite step. We build the 3D structures of two STAT5 isomorphs, STAT5a and STAT5b, in a phosphorylated (pSTAT5) and unphosphorylated (STAT5) states (Fig 1B). The generated models were studied by computational methods ‒ molecular dynamics simulations, normal mode analysis, and modular network analysis ‒ with the aim to retrieve the biologically-relevant information that can guide the rational development of STAT5 inhibitors.

STAT5a and STAT5b, having a high sequence homology, showed similar 3D structures accurately preserved over the simulations. Nevertheless, detailed analysis of the secondary structures indicates to a slight but noticeable difference between these proteins, reflecting a sequence-dependent structural arrangement and conformational dynamics. The most important structural divergence is observed in SH2 domain, the less conserved region in STATs.

Similarly to other STATs, STAT5 exhibits a large range of internal motions, ranged from individual atomic displacement to collective large-scale movements. The global dynamics of STAT5s characterizing the functionally-related collective movements are comparable. The distal region of CCD, composed of two extended α-helices, oscillates in different directions, while the proximal part of CCD (the four-helix bundle) displays sparse motions (Fig 1C). This dynamics has not been described in the other STATs in which it is probably not present. Since no residue of this region has been described as crucial for STAT5 functions, the CCD motions may reveal dynamics whose function needs to be explored. We evidenced highly coupled motions between largely distant sites of STAT5s, in particular, between the distal region of CCD and the SH2 domain separated by 80-100 Å, suggesting a long-distance allosteric regulation of dynamically-dependent processes operating in these proteins.

The structural and dynamical properties of STAT5 are sequence-dependent and are sensitive to the phosphorylation events, as evidenced by the phosphorylation-induced adaptation of secondary structures in the CCD and DBD in both STAT5s, and by a difference of the concerted motions of largely distant regions, the distal CCD and the SH2 domain between STAT5a and STAT5b. The advanced description of these effects was achieved thought a study of the internal dynamics of the proteins by Principal Feature Analysis (PFD) and Modular Network Analysis (MONETA) [11].

Representing the groups of residues showing the most striking local dynamics as the Independent Dynamics Segments (IDSs) [12], we identified the internal dynamics landscape of each studied protein. This description revealed on the one side, typical dynamics for all STAT5s and on the other side, subtle but meaningful differences of IDSs dynamics between the two STAT5 isoforms and between the p-STAT5 and STAT5s. The structure-related connections between IDSs were described as Communications Pathways (CPs), characterizing the routes of possible transmission of information between the distant regions of the protein through interacting residues.

We defined the ‘shortest’ intramolecular pathway connecting the SH2 and CCD domains as the succession of CPs that involve the minimal number of residues (Fig 1D). The shortest pathway traced in all STAT5s, were dissimilar in the number of residues contributing to this path, which was considerably diminished in phosphorylated species.

Further, we searched for new putative small molecules binding sites in STAT5s, viewed as possible targets for the STAT5 inhibition. Despite of a commonly-accepted outlook on using the dimerization interface as the target, its high conservation among STATs limits its application for development of inhibitors highly selective to STAT5. Exploring STAT5 conformations, we detected two novel putative sites in the interface of the SH2-LD domains showing less conserved content among the STATs. A plausible use of these sites for development of inhibitors is enforced by their proximity to the CP connecting the SH2 domain and the distal region of CCD (Fig 1E).  It suggests a development of inhibitors able to modulate communication properties of this signaling protein. Such communication-inspired and communication-targeted modulation may block several post-transduction processes, such as dimerization, DNA binding or upstream STAT5 activators recognition.

Acknowledgements

This work was supported by La Ligue Nationale contre le Cancer, Société Française d’Hématologie, Fondation de France and Institut Farman de l’ENS de Cachan. We are grateful to Dr. Isaure Chauvot de Beauchêne for the useful discussions. We thanks BULL for attribution of the computer time on their cluster NOVA, and the Centre Informatique National de l’Enseignement Supérieur (CINES) supported by Grand Equipement National de Calcul Intensif (GENCI) for attribution of the computer time on Supercomputer JADE (allocation 2014-c2013077107).