Proteomic analysis of tyrosine phosphorylation induced by exogenous expression of oncogenic kinase fusions identified in lung adenocarcinoma

Kinase fusions are considered oncogenic drivers in numerous types of cancer. In lung adenocarcinoma 5–10% of patients harbor kinase fusions. The most frequently detected kinase fusion involves the Anaplastic Lymphoma Kinase (ALK) and Echinoderm Microtubule‐associated protein‐Like 4 (EML4). In addition, oncogenic kinase fusions involving the tyrosine kinases RET and ROS1 are found in smaller subsets of patients. In this study, we employed quantitative mass spectrometry‐based phosphoproteomics to define the cellular tyrosine phosphorylation patterns induced by different oncogenic kinase fusions identified in patients with lung adenocarcinoma. We show that exogenous expression of the kinase fusions in HEK 293T cells leads to widespread tyrosine phosphorylation. Direct comparison of different kinase fusions demonstrates that the kinase part and not the fusion partner primarily defines the phosphorylation pattern. The tyrosine phosphorylation patterns differed between ALK, ROS1, and RET fusions, suggesting that oncogenic signaling induced by these kinases involves the modulation of different cellular processes.


INTRODUCTION
In the last decade, advances in high-throughput sequencing technologies have enabled in-depth genetic characterization of tumors such as lung cancer and led to the discovery of numerous oncogenic drivers [1,2]. Gene fusions involving tyrosine kinases are found in diverse hematological and solid malignancies including acute lymphoblastic leukemia, chronic myeloid leukemia, lung cancer, and thyroid cancer [3][4][5][6][7]. Tyrosine kinase fusions result from genomic rearrangements such as chromosomal inversions or translocations and lead to expression of kinases that are constitutively activated through loss C-terminally to the Echinoderm Microtubule-associated protein-Like 4 (EML4). A number of different breakpoints have been reported but, in all cases, the EML4 part retains its coil-coiled domain that mediates dimerization of the EML4-ALK protein and thereby leads to constitutive activation of the kinase [10]. In rare cases, fusions of ALK with another microtubule-associated protein Kinesin-1 heavy chain (KIF5B) have been reported. As in case of EML4-ALK, the coiled-coil domain of KIF5B mediates dimerization and leads to constitutive activation of the KIF5B-ALK fusion protein [11].
In addition to ALK fusions, fusions of the tyrosine kinases Protooncogene tyrosine protein kinase receptor Ret and Proto-oncogene tyrosine kinase ROS1 have been described in LADC [12,13]. KIF5B-RET fusions have been identified in 1-2% of LADC patients. Functional investigations suggest that the kinase domain of RET is constitutively activated by artificial dimerization as in case of EML4-and KIF5B-ALK.
More recently, fusion proteins involving ROS1 have been described in LADC. Kinase fusions involving ROS1 can be identified in 1-2% of patients LADC. In most cases ROS1 is C-terminally fused to the transmembrane protein CD74. However, a number of other fusion partners including SDC4, EZR, CCDC6, and SLC34A2 have been identified [14].
The activation mechanism of ROS1 kinase fusions is not yet completely understood, but it is likely that the kinase activation occurs through artificial dimerization. Targeted therapies against constitutively active tyrosine kinase fusions involving ALK, ROS1, and RET have proven effective for the treatment of LADC [15][16][17][18][19].
Mass spectrometry (MS)-based proteomics is a powerful approach to study phosphorylation-dependent cellular signaling in physiology and disease [20]. Large scale identification of tyrosine phosphorylation sites can be achieved by enrichment of tyrosine phosphorylated peptides using specific antibodies followed by their identification using high performance liquid chromatography tandem mass spectrometry (HPLC-MS/MS) [21][22][23]. Previous MS-based proteomics studies have delivered important insights into tyrosine phosphorylation-dependent signaling after growth factor stimulation. In addition, phosphoproteomics combined with kinase inhibition revealed protein substrates of kinases functioning in diverse cellular processes [24][25][26][27][28][29] [30]. However, to date a systematic, quantitative and comparative proteomic analysis of the phosphorylation signaling induced by different oncogenic tyrosine kinase fusions in a controlled cellular model system has not been performed. In this study, we employed quantitative phosphoproteomics based on stable isotope labelling with amino acids in cell culture (SILAC) to characterize tyrosine phosphorylation-dependent signaling that is induced downstream of oncogenic kinase fusions. To this end, kinase fusions involving ALK, ROS1, and RET frequently found in LADC were investigated.
We found that exogenous expression of ALK, ROS1, and RET oncogenic L-lysine (13C6 15N2) (Cambridge Isotope Laboratories) as described previously [31]. All cells were cultured at 37 • C in a humidified incubator containing 5% CO 2 .

MS sample preparation
Proteins were precipitated in fourfold excess of ice-cold acetone and subsequently redissolved in denaturation buffer (

MS analysis
Peptide fractions were analyzed on a quadrupole Orbitrap mass spectrometer (Q Exactive Plus, Thermo Scientific) equipped with a UHPLC system (EASY-nLC 1000, Thermo Scientific) as described [33]. Peptide samples were loaded onto C18 reversed phase columns (15 cm length, 75 μm inner diameter, 1.9 μm bead size) and eluted with a linear gradient from 8 to 40% acetonitrile containing 0.1% formic acid in 2 h. The mass spectrometer was operated in data dependent mode, automatically switching between MS and MS2 acquisition. Survey fullscan MS spectra (m/z 300-1650) were acquired in the Orbitrap. The ten most intense ions were sequentially isolated and fragmented by higher energy C-trap dissociation (HCD) [34]. Peptides with unassigned charge states, as well as with charge states less than +2 were excluded from fragmentation. Fragment spectra were acquired in the Orbitrap mass analyzer.

Peptide identification
Raw data files were analyzed using MaxQuant (development version 1.5.2.8) [35]. Parent ion and MS2 spectra were searched against a database containing 88,473 human protein sequences obtained from the UniProtKB released in December 2016 using Andromeda search engine [36]. Spectra were searched with a mass tolerance of 6 ppm in MS mode, 20 ppm in HCD MS2 mode, strict trypsin specificity and allowing up to three miscleavages. Cysteine carbamidomethylation was searched as a fixed modification, whereas protein N-terminal acetylation, methionine oxidation, and phosphorylation of serine, threonine, and tyrosine were searched as variable modifications. Site localization probabilities were determined by MaxQuant using the PTM scoring algorithm as described previously [35,37]. The dataset was filtered based on posterior error probability to arrive at a false discovery rate (FDR) below 1% estimated using a target-decoy approach [38].

Data analysis
Statistical analysis was performed using the R software environment.
To identify significantly regulated phosphorylation sites a moderated T test (limma) was employed [39]. Only sites with a FDR-adjusted p value ≤ 0.01 were considered regulated. Kinase activities were estimated using the KSEA algorithm [40] and the R implementation of the KSEA App [41]. Kinase-Substrate annotations were obtained from PhosphoSitePlus (PSP) [42] and from the NetworKIN database [43].
The analysis was performed with a minimum NetworKIN score of 5 for upregulated and downregulated phosphorylation sites with a p value ≤ 0.05. Phosphosite-specific signature analysis was performed using PTM-SEA [44]. As input data p values generated during statistical analysis were transformed and multiplied by the sign of the averaged log 2transformed fold changes. As identifier for the phosphorylation sites the flanking sequence (+/-7 amino acids) was used. PTM-SEA analysis was performed in R. 'PTM site-specific phosphorylation signatures of kinases, perturbations and signaling pathways' (PTMsigDB) was used as reference database. Minimal overlap between reference data and input data was set to 3.

Cloning and expression of ALK, RET, and ROS1 kinase fusions
Oncogenic kinase fusions involving ALK, RET, and ROS1 are frequently found in patients with LADC. We employed the COSMIC database [45] to identify the most frequently observed fusion partners and breakpoints of the tyrosine kinases ALK, RET, and ROS1 ( Figure 1A and Table 1 Figure 1C).
It has been previously reported that the kinase fusions investigated in this study localize to different cellular compartments [9,11,46,47]. To verify that the generated constructs retain these properties in

Oncogenic kinase fusions induce widespread tyrosine phosphorylation
To investigate the effect of the oncogenic kinase fusions on the cellular phosphorylation patterns, we employed quantitative mass spectrome-

Cellular phosphorylation patterns are primarily determined by the kinase in the fusion protein
To investigate the differences/similarities between the phosphorylation patterns induced by the kinase fusions, we performed multidi-

ALK, RET, and ROS1 fusions activate different cellular signaling pathways
We subsequently performed annotation enrichment analysis to identify biological processes that are regulated by phosphorylation after overexpression of the investigated fusion kinases as well as to define the affected cellular compartments. To this end, annotations from the GO Ontology and the Reactome pathway database were employed.
The analysis revealed that phosphorylation induced by all investi-

Kinase fusions lead to phosphorylation of fusion partners and interacting proteins
The annotation enrichment analysis suggested that kinase fusions comprising KIF5B induce tyrosine phosphorylation of KIF5B itself and

ALK, RET, and ROS1 kinase fusions activate STAT signaling
The STAT signaling pathway has been previously identified as one of the signaling pathways mediating the oncogenic effect of ALK [48], RET [49], and ROS1 [50], fusion proteins. To gain further insights into the activation of the STAT transcription factors in cells expressing different kinase fusions, we investigated the phosphorylation of STAT Microscopy core facilities at IMB is gratefully acknowledged.
Open access funding enabled and organized by Projekt DEAL.

CONFLICT OF INTEREST
The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT
The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium (http://proteomecentral. proteomexchange.org) via the PRIDE partner repository [53] with the dataset identifier PXD021904.