Multiple functions of protein phosphatases in receptor tyrosine kinase signaling revealed by interactome analysis

ABSTRACT

To obtain a global picture of how protein phosphatases are involved in receptor tyrosine kinase (RTK) signaling, we mapped the RTK-phosphatase interactome. Analyses of selected interactions revealed detailed mechanisms of their actions. This study provides new knowledge to better understand cancer development and to identify novel therapeutic targets.

Receptor tyrosine kinases (RTKs) receive multiple extracellular signals, and process and transmit them into intracellular space. They operate through the activation of their tyrosine kinase domains upon ligand engagement, and the following tyrosine phosphorylation. These phosphorylated tyrosines recruit signaling molecules containing SRC homology 2 (SH2) or phosphotyrosine-binding (PTB) domains and relay the signal to downstream pathways.¹ Malfunction of RTKs is involved in several disorders. Self-sufficient activation derived from overexpression or mutation of RTKs or related regulated molecules is one of the driving causes of various types of cancer. In addition, mutations in RTKs also contribute to resistance to anti-cancer drugs.

Protein tyrosine phosphatases (PTPs) are a group of enzymes that remove phosphate from tyrosine residues and thereby are the major means of terminating RTK signal and maintaining control of RTK signaling. An oversimplified model describes that PTPs function as signal erasers. However, accumulating data have demonstrated that PTPs behave in a complex manner; besides terminating signals, some PTPs actually potentiate RTK signaling, and some play controversial roles depending on different cellular and RTK context. In addition to tyrosine phosphorylation, serine/theronine phosphorylation also plays certain roles in fine-tuning RTK signaling, although most of these roles are unknown. Therefore, theoretically all protein phosphatases, including protein Ser/Thr phosphatases (PSPs), are involved in RTK signaling.

To fully understand RTK function, it is necessary to comprehensively examine the RTK-phosphatase relationship. We aimed to tackle this problem from the perspective of protein-protein interaction (PPI).^2,3 However, the fact that RTKs are integral transmembrane proteins makes it extremely difficult to study RTK-phosphatase interactions in a systems fashion using traditional PPI methods, due to the special physical and chemical features of the membrane environment. To overcome this problem, we developed a unique Membrane Yeast Two-Hybrid (MYTH) system to study membrane protein PPIs (Fig. 1A), which utilizes split ubiquitin to sense PPIs.⁴ It should be noted that the MYTH system detects PPIs in a yeast environment instead of in mammalian cells for mammalian PPIs, but MYTH is still a valuable system since it maximally eliminates the interference of endogenous mammalian interactors and regulators. Recently, we also developed a mammalian version of MYTH, the Mammalian Membrane Two-Hybrid (MaMTH) system, to complement the MYTH assay.⁵

Figure 1.

Mapping the receptor tyrosine kinase (RTK)-phosphatase interactome with MYTH and MaMTH screens. (A) Membrane Yeast Two-Hybrid (MYTH) and Mammalian Membrane Two-Hybrid (MaMTH) assays. In both systems, the C-terminal half of split ubiquitin (Cub) is fused to bait protein and the N-terminal half (Nub) is fused to prey protein. Interaction of bait and prey creates a fully functional ubiquitin, which recruits deubiquitinating enzymes (DUBs) to cleave the transcription factor (TF) in the bait construct. TF enters the nucleus and induces expression of different reporter genes depending on MYTH or MaMTH system. (B) PTPRH and PTPRB inhibit epidermal growth factor receptor (EGFR) signaling through dephosphorylating key tyrosine residues on EGFR responsible for signal relay. (C) PTPRA potentiates EGFR signaling by increasing local concentration of effective SRC molecules through dephosphorylating EGFR-bound SRC.

We performed a RTK-phosphatase interactome screen using MYTH and MaMTH assays.⁶ The MYTH assay included 48 RTK baits and 108 phosphatase preys, and the MaMTH assay included 4 ErbB family RTKs and 18 receptor type PTPs (RPTPs). More than 300 PPIs were identified in the screen, most of which have not been reported in previous studies. Interestingly, almost all phosphatase families showed some degree of interactions with RTKs, suggesting the potential roles of various phosphatases. Further characterization of RTK-PTP interactions in MYTH using inactive PTP mutants interestingly revealed three modes of interaction: 1) trapping mutant type interactions, suggesting a phosphatase-substrate relationship, 2) constant interactions, and 3) interactions indicating a reciprocal inhibition relationship.

The MaMTH screen identified a panel of RPTP-epidermal growth factor receptor (EGFR) interaction hits. Our drill-down analyses of several interactions provided molecular details of how these RPTPs regulate EGFR signaling. Among them, PTPRH downregulates EGFR signaling through direct interaction and dephosphorylation of several key EGFR tyrosine residues responsible for signal relay (Fig. 1B). PTPRH knockout by CRISPR Cas9 massively enhanced EGFR signaling in OV-90 cells, suggesting that PTPRH is the major phosphatase to inhibit EGFR signaling in this cell line. PTPRB functions in a similar manner. Thus, PTPRH and PTPRB are potential tumor suppressor genes against EGFR-driven cancer development. Indeed, genomic changes of PTPRH and PTPRB in some cancer patients are recorded in cancer genome databases. Whether disruption of PTPRH and PTPRB are really involved in the development of EGFR or other RTK-driven cancers needs further clarification.

Both PTPRH and PTPRB belong to the R3 subgroup of RPTPs.⁷ Interestingly, PTPRJ, an R3 member similar to PTPRH and PTPRB, has also been shown to downregulate EGFR.⁸ However, the EGFR/PTPRJ interaction was not captured in our screen, a fact probably reflecting different characteristics of interaction. It is not clear whether these R3 RPTPs are functionally redundant. In addition, whether two other R3 RPTPs, PTPRO and PTPRQ, are also involved in EGFR signaling merits further investigation.

Our analysis of PTPRA demonstrated both the diversity and the specificity of phosphatase action in RTK signaling. Contrary to PTPRH and PTPRB, PTPRA enhances EGFR signaling. We found that this function operates through dephosphorylating the inhibitory phosphotyrosine of SRC (Fig. 1C).⁹ By comparing with a highly similar SRC phosphatase PTPRE,¹⁰ we concluded that the specialized effect of PTPRA on EGFR signaling is a result of its capability of interaction with EGFR, whereby PTPRA increases the local active SRC concentration through activating EGFR-bound SRC molecules. This function suggests a therapeutic value of inhibiting PTPRA for treating EGFR-driven cancer development.

Our analyses of selected RTK-phosphatase interactions presented examples in which the RTK-phosphatase interactome was used as a start point for in-depth study of phosphatase function in RTK signaling. Many open questions still remain, such as the functions of many interacting PSPs. We believe the presented RTK-phosphatase interactome provides a wealth of information for the research field of cancer signaling. It will help us not only understand the molecular mechanism of cancer development but also identify novel targets for cancer therapy.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgment

We thank Dr. Laura Riley for editing this manuscript.

Funding

Work in the Stagljar laboratory was supported by grants from Ontario Research Fund (GL2–01–018), Canadian Cancer Society Research Institute (CCSRI #702109 and #703889), Canadian Cystic Fibrosis Foundation (CFC #2847), Genome Canada/Ontario Genomics (#9427 and #9428), and Pancreatic Cancer Canada.