Proteomic analysis of SP600125‐controlled TrkA‐dependent targets in SK‐N‐MC neuroblastoma cells: Inhibition of TrkA activity by SP600125

The c‐Jun N‐terminal kinase (JNK) is well known to play an important role in cell death signaling of the p75 neurotrophin receptor. However, little has been studied about a role of JNK in the signaling pathways of the tropomyosin‐related kinase A (TrkA) neurotrophin receptor. In this study, we investigated JNK inhibitor SP600125‐controlled TrkA‐dependent targets by proteomic analysis to better understand an involvement of JNK in TrkA‐mediated signaling pathways. PDQuest image analysis and protein identification results showed that hnRNP C1/C2, α‐tubulin, β‐tubulin homolog, actin homolog, and eIF‐5A‐1 protein spots were upregulated by ectopic expression of TrkA, whereas α‐enolase, peroxiredoxin‐6, PROS‐27, HSP70, PP1‐gamma, and PDH E1‐alpha were downregulated by TrkA, and these TrkA‐dependent upregulation and downregulation were significantly suppressed by SP600125. Notably, TrkA largely affected certain PTM(s) but not total protein amounts of the SP600125‐controlled TrkA‐dependent targets. Moreover, SP600125 strongly suppressed TrkA‐mediated tyrosine phosphorylation signaling pathways as well as JNK signaling, indicating that SP600125 could function as a TrkA inhibitor. Taken together, our results suggest that TrkA could play an important role in the cytoskeleton, cell death, cellular processing, and glucose metabolism through activation or inactivation of the SP600125‐controlled TrkA‐dependent targets.


Introduction
Neurotrophins (NTs) such as nerve growth factor (NGF), brain-derived neurotrophic factor, NT-3, and NT-4 are a family of closely related proteins that regulate cell survival, death, maintenance, and development in both the central and peripheral nervous systems [1,2]. The effects of NTs are exhibited through activation of two distinct receptor types: the tropomyosin-related kinase (Trk) subfamily (TrkA, TrkB, and TrkC) and the p75 neurotrophin receptor (p75NTR) that belongs to the tumor necrosis factor receptor superfamily [3]. NGF activates various intracellular signaling pathways through activation of the TrkA and p75NTR receptors, involved in the regulation of both cell survival and death in neuronal and non-neuronal cell types.
The transmembrane TrkA receptor possesses an intrinsic tyrosine kinase activity in the intracellular domain. NGF binding to TrkA leads to receptor dimerization and enhances TrkA enzyme activity [4]. Thus, the activated TrkA phosphorylates Src homology 2-containing protein (Shc), and this adaptor protein induces an intermediate protein connection with Grb-2, Sos, and Raf, activating the extracellular signal-regulated protein kinase (ERK) pathway, which is one of the mitogen-activated protein kinase signaling pathways [4]. This NGF-mediated TrkA-dependent ERK signaling enhances transcription of proteins involved in cell proliferation and survival [4]. The activated TrkA also phosphorylates phosphatidylinositol 3-kinase and stimulates phosphatidylinositol 3-kinase-Akt signaling pathway, involved in the regulation of cell proliferation and survival [4,5]. In contrast, other mitogen-activated protein kinase members such as c-Jun N-terminal kinase (JNK) and p38 have been known to be activated by TrkA-dependent phospholipase C-␥ pathway, suggesting a role of TrkA in the cell death signaling [4]. At present, little has been studied about a role of JNK in the TrkA-dependent signaling pathways, whereas JNK has been well known to play a critical role in the p75NTR-dependent cell death signaling. In addition, p75NTR can function as a cell survival factor through activation of the nuclear factor-kappaB transcription factor [6].
To reveal TrkA-dependent cellular effects, ectopic expression of TrkA was widely used in various cell types. The effects of ectopic TrkA seemed to be different according to the cell type and protein expression level. It has been known that ectopic expression of TrkA plays an important role in neuronal responses in an NGF-dependent or NGFindependent manner [7]. TrkA overexpression caused activation of its tyrosine kinase activity and promoted cell growth, migration, and invasion in breast cancer cells [8]. In contrast, ectopic expression of TrkA inhibited angiogenesis and tumor growth in neuroblastoma cells [9,10] and induced apoptosis in a p53-dependent mechanism [11]. NGF-mediated TrkA-dependent apoptosis was also associated with a Ras/Raf signaling pathway in medulloblastoma cells [12]. Moreover, TrkA-dependent apoptotic cell death was strongly related to the production of ␥-H2AX, a phosphorylated form of histone H2AX at serine-139, in the absence of cellular stimuli such as DNA damage-inducing reagents in U2OS osteosarcoma cells, and this phenomenon was significantly suppressed by JNK inhibitor SP600125, suggesting a critical role of JNK in the TrkA-dependent cell death signaling [13,14]. The TrkAdependent cell death was promoted by ectopic expression of H2AX via upregulation of TrkA tyrosine-490 phosphorylation [15], whereas it was inhibited by ectopic expression of caveolin-1 via downregulation of tyrosine-490 phosphorylation [16], indicating that tyrosine-490 phosphorylation of TrkA is critically required for its cell death function. Recently, it was demonstrated that ectopic TrkA and TrkC, but not TrkB, caused death of mouse embryonic stem cells [17], further supporting a role of TrkA in cell death signaling.
Studies on the TrkA-dependent cellular effects have been somewhat accomplished using TrkA inhibitor GW441756 [18]. We have previously reported that the induction of TrkA tyrosine-490 phosphorylation by TrkA overexpression was significantly inhibited by GW441756, resulting in the downregulation of its cellular effects such as ␥-H2AX production [13], Bak cleavage [16], and inhibition of colony formation by cancer cells [19]. In addition, we recently identified TrkA-dependent targets associated with tyrosine phosphorylation signaling pathways using the effects of GW441756 in SK-N-MC neuroblastoma cells [20]. Moreover, we demonstrated that TrkA overexpression caused tyrosine phosphorylation of JNK as well as ERK, indicating an involvement of JNK in the TrkA-dependent signaling pathways [20]. To better understand this, here we identified SP600125-controlled TrkAdependent targets by proteomic analysis in TrkA-inducible SK-N-MC cells. This research could highly contribute on a new signal transduction and mechanism in various TrkAdependent signaling pathways through activation or inactivation of the SP600125-controlled TrkA-dependent targets.

Cell culture
TrkA-inducible cells by the Tet-On system were previously established in SK-N-MC cells [19]. The cells were maintained with medium A (DMEM, Tet-screened 10% FBS, 1% penicillin/streptomycin) containing 1.25 g/mL blasticidine and 25 g/mL zeocin in a humidified 5% CO 2 incubator at 37ЊC. Ectopic expression of TrkA was induced by adding of 2 g/mL Tet for 12 h in the medium A and then treated with DMSO as a control solvent, SP600125, GW441756, or AS601245 with the indicated amounts and times.

Sample preparation, 2DE and Coomassie Blue stain
Samples were prepared for proteomic analysis, separated by 2DE and stained with Coomassie Blue staining solution as described previously [20]. Briefly, whole cells were washed with PBS, lysed in 2DE lysis buffer, treated with DNase I, and then precipitated with trichloroacetic acid. After washing with acetone, the pellet was dissolved in rehydration solution followed by protein quantification using Bradford reagent. IPG strips (pH 4-7, 17 cm) were rehydrated for 14 h at 20ЊC in the IEF cell with the protein samples (500 g each) containing 0.5% IPG buffer (pH 4-7) and performed by 1D IEF. The strips were reduced for 15 min in equilibration buffer (50 mM Tris (pH 8.8), 6 M urea, 30% glycerol, 2% SDS, and trace amount of bromophenol blue) containing 10 mg/mL DTT, and then alkylated for 15 min in the same buffer containing 25 mg/mL iodoacetamide. 2DE of the strips was performed on 8.5-14% sucrose gradient polyacrylamide gels, which were prepared using the light 8.5% SDS-PAGE solution and the heavy 14% SDS-PAGE solution containing 15% sucrose. The gels were fixed in fixing solution and stained with Coomassie Blue staining solution. After destaining with H 2 O, gel images were obtained using a UMAX scanner (PowerLook 2100XL).

Image analysis and statistical significance
Quantitative analysis of Coomassie Blue stained images was carried out using PDQuest software (Bio-Rad) according to the manufacturer's instructions. Quantity of total about 1500 protein spots was normalized by valid spot intensity. SP600125-controlled TrkA-dependent targets were determined by the inhibitory effects of SP600125 on more than twofold upregulated or downregulated TrkA-dependent protein spots. In addition, certain protein spots with a significant SP600125 effect were also determined as SP600125controlled TrkA-dependent targets, although they were less than twofold regulated by TrkA. The statistical significance of image analysis was determined by the Student's t-test (statistical level of *p < 0.05 is significant).

MALDI-TOF MS and database searching
SP600125-controlled TrkA-dependent protein spots were excised from Coomassie Blue stained 2DE gels and performed with in-gel digestion using modified porcine trypsin as described [20]. The tryptic digested peptides were mixed with saturated matrix solution (15 mg CHCA, 15 mg NC membrane, 75% acetone, 25% 2-propanol) and calibrants (bradykinin and neurotensin). And then, the mixed tryptic peptides were immediately spotted onto MALDI-TOF sample target plate. Mass measurement of tryptic peptides was performed with MALDI-TOF MS (Voyager-DE-STR; Applied Biosystems), and mass spectra were acquired for the mass range of 800-3500 Da and calibrated with bradikinin m/z (904.4681) and neurotensin m/z (1672.9175) as standard peaks. The proteins were identified by PMF on the basis of the Swiss-Prot and NCBInr database using the search program MASCOT (http://www.matrixscience.com), allowing peptide mass tolerance of 1.2 Da and one missed cleavage. Significance of the searched data was judged from the scores more than 56 (p < 0.05) and at least seven matched peptide masses.
Search for protein identity was carried out using the NCBInr database by ProteinPilot v.3.0 software (with MASCOT as the database search engine). The parameters for searching were peptide and fragment ion mass tolerance of 50 ppm, one missed trypsin cleavage, carbamidomethylation of cysteine, oxidation of methionine and monoisotopic. Significance of the identified proteins was based on the number of matching peptide masses and comparison of experimental and theoretical properties of the proteins, in addition to database searched protein scores; greater than 84% was considered as a statistically significant (p < 0.05).

Confirmation of the identified proteins
To confirm the results identified by MALDI-TOF MS and MS/MS analysis, SP600125-controlled TrkA-dependent protein spots were excised from Coomassie Blue stained 2DE gels followed by 10% SDS-PAGE. In addition, protein samples (5 g each) prepared for 2DE were separated by 10% SDS-PAGE. After 1DE, the proteins were transferred to NC membrane and analyzed by Western blot using the super signal ECL detection system [20].

2DE/Western blot analysis
IPG strips (pH 4-7, 7 cm) were rehydrated at 20ЊC for 14 h in the IEF cell with protein samples (10 g each) containing 0.5% IPG buffer (pH 4-7) followed by 1D IEF in a maximum current of 50 A/IPG strip; at 250 V for 15 min, at 4000 V for 2 h, and then at 4000 V for 30 000 V-h. The strips were reduced for 15 min in equilibration buffer containing 10 mg/mL DTT, and then alkylated for 15 min in the same buffer containing 25 mg/mL iodoacetamide. 2DE of the strips was performed on 10% SDS polyacrylamide gels, and then separated protein spots were transferred to NC membrane and analyzed by Western blot [20].

2DE and image analysis of SP600125 effects on TrkA-dependent targets
SP600125, which is one of the most extensively studied ATPcompetitive JNK inhibitors, has been known as a useful tool for studying JNK-mediated signal transduction [22][23][24]. To reveal an involvement of JNK in the TrkA-dependent signaling pathways, the JNK inhibitor SP600125-controlled TrkAdependent targets were investigated by 2DE and Coomassie Blue stain in TrkA-inducible SK-N-MC cells. The representative 2DE images were selected from five independent experiments ( Fig. 1), and PDQuest image analysis was performed with two sets of well-separated 2DE images. Thus, 16 protein spots were determined as major SP600125-controlled TrkA-dependent targets; ten protein spots were upregulated by TrkA, whereas six protein spots were downregulated by TrkA, and these TrkA-dependent upregulation and downregulation were significantly suppressed by SP600125 treatment.

Identification of SP600125-controlled TrkA-dependent protein spots
We

Confirmation of SP600125-controlled TrkA-dependent protein spots
To prove the identification results analyzed by MALDI-TOF MS and MALDI-TOF/TOF MS/MS, SP600125-controlled TrkA-dependent protein spots and ␤-actin spot as a negative control were excised from Coomassie Blue stained 2DE gels and analyzed by 1DE/Western blot. Except for SSP 8101 spot, which was very weakly recognized by peroxiredoxin-6 antibody for some reason and thus visualized by long-time exposure, most of the identified protein spots were strongly recognized by the indicated antibodies as expected: SSP 3209, 3212, 3350, and 3352 spots by hnRNP C1/C2 antibody; SSP 4415, 4506, and 4535 spots by ␣-tubulin antibody; SSP 7211 spot by PP1-gamma antibody; SSP 7420 spot by ␣-enolase antibody (Fig. 4). Next, the samples prepared for 2DE were analyzed by 1DE/Western blot to prove whether the identified proteins are regulated by TrkA and SP600125. As shown in Fig. 5A, ectopic expression of TrkA induced its tyrosine-490 phosphorylation in TrkA-inducible SK-N-MC cells, and both TrkA protein amount and tyrosine-490 phosphorylation level seemed to have no effect by SP600125 (first and second panels). In addition, ectopic expression of TrkA did not influence on acetylation of ␣-tubulin and protein amounts of ␣-tubulin, hnRNP C1/C2, PP1-gamma, and full length HSP70 (Fig. 5A, third to seventh panels). However, a certain protein of lower molecular weight than full length HSP70 was detected by HSP70 antibody in a control condition containing 0.1% DMSO, and it was disappeared by ectopic expression of TrkA (Fig. 5A, seventh panel; indicated by an arrow and asterisk). Moreover, the lower protein that could be a cleaved HSP70 or a certain HSP70 homolog was detected as a couple of protein spots by 2DE/Western blot analysis using 17 cm IPG strip (pH 4-7) in a normally proliferating TrkA-inducible SK-N-MC cells (Fig. 5B, indicated by an arrow and asterisk). These results suggest that TrkA could be involved in downregulation of full length HSP70 cleavage or a certain HSP70 homolog.

Investigation of GW441756 effects on SP600125-controlled TrkA-dependent protein spots
Recently, hnRNP C1/C2, ␣-tubulin, ␣-enolase, peroxiredoxin-6, and PROS-27 were identified as major GW441756controlled TrkA-dependent targets in SK-N-MC neuroblastoma cells by us [20]. Thus, here we investigated the effects of GW441756 on the SP600125-controlled novel TrkAdependent protein spots. The results revealed that unknown TrkA-dependent ␤-tubulin homolog, actin homolog, eIF-5A-1, PP1-gamma, and PDH E1-alpha protein spots were also controlled by GW441756, but not as much as the major GW441756-controlled TrkA-dependent protein spots (data not shown). However, SSP 3350, one of the SP600125controlled hnRNP C1/C2 protein spots, was not controlled by GW441756 on the TrkA-dependent upregulation, whereas , separated by 10% SDS-PAGE, and transferred to NC membrane. Panels a and b were analyzed by Western blot analysis using hnRNP C1/C2 antibody, and the blots were reprobed with ␤-actin and hnRNP C1/C2 antibodies (panels a`and b`). Panels c and d were detected with ␣-enolase antibody, and the blots were reprobed with ␤-actin and ␣-enolase antibodies (panels c`and d`).

Inhibition of TrkA activity by SP600125
Intriguingly, we showed here that most of the SP600125controlled TrkA-dependent targets were similarly regulated by TrkA inhibitor GW441756 although there was somewhat difference in 2DE image (Fig. 8). To further investigate this, TrkA was induced with Tet for 12 h in SK-N-MC-TrkA cells and then treated with GW441756, SP600125, and other JNK inhibitor AS601245 for 8 h. After removing media thoroughly, the cells were immediately extracted with SDS sample buffer to exclude a potential dephosphorylation and other modifications followed by 10% SDS-PAGE and Western blot analysis. Surprisingly, TrkA-mediated tyrosine phosphorylation processes including ERK phosphorylation were gradually suppressed in a dose dependent manner by both GW441756 and SP600125, indicating that SP600125 could function as a TrkA inhibitor similarly to GW441756 (Fig. 9A, second and third panels). As expected, activation of TrkA-dependent JNK signaling including c-Jun phosphorylation was significantly suppressed by SP600125 but not by GW441756 at least at these cellular circumstances (Fig. 9A, fifth and seventh panels). In addition, TrkA-dependent JNK activation was inhibited by AS601245, but TrkA-mediated tyrosine phosphorylation processes were not suppressed by AS601245 (Fig. 9B, second and third panels). Again, our results demonstrated here that TrkA did not largely affect protein amount of full length HSP70 and ␣-tubulin, however, a cleaved form of HSP70 or a certain HSP70 homolog indicated by an arrow and asterisk was significantly inhibited by TrkA, in consistent with the results of Fig. 5A (Fig. 9A, eighth and ninth panels).
Interestingly, we found that a cleaved form of ␣-tubulin or a certain ␣-tubulin homolog was significantly upregulated by TrkA, and it was blocked by SP600125 (Fig. 10A, indicated by an arrow and asterisk). Moreover, ectopic expression of TrkA altered acetylation levels of cellular proteins as shown in Fig. 10B (especially in areas indicated by arrows). Taken together, our results strongly suggest that TrkA could be involved in various PTM signaling pathways such as acetylation and cleavage of cellular proteins as well as tyrosine phosphorylation.

Discussion
To better understand an involvement of JNK in TrkAmediated signaling pathways, we identified JNK inhibitor  (Tables 1 and 2). However, we could not confirm TrkAdependent regulation and SP600125 effect on the target proteins by 1DE/Western blot analysis (Fig. 5A). Thus, we investigated TrkA-dependent regulation on the target protein spots by 2DE/Western blot analysis. Interestingly, ectopic expression of TrkA resulted in upregulation or downregulation of certain modification(s) occurring normally on the targets, and it was suppressed by SP600125 (Figs. 6  and 7). Moreover, most of the SP600125-controlled TrkAdependent protein spots were also regulated by GW441756, indicating that these proteins are involved in the TrkAdependent tyrosine phosphorylation signaling pathways (data not shown). However, hnRNP C1/C2 protein spot SSP 3350 and HSP70 protein spot SSP 2620 had no significant effect by GW441756 (Fig. 8). To figure out these results, we compared the effects of GW441756 and SP600125 on TrkA-mediated signaling processes by Western blot analysis after immediately preparing whole cell extracts with SDS sample buffer. The results demonstrated that TrkA activity was significantly suppressed by both GW441756 and SP600125 but not AS601245, supporting similar effects of GW441756 and SP600125 in the inhibition of TrkA-mediated signaling pathways (Fig. 9). Therefore, the SP600125-controlled forms can be classified to GW441756controlled and -uncontrolled TrkA-dependent targets through analysis of GW441756 effects (Fig. 11). GW441756-controlled TrkA dependent targets consisted of ␣-tubulin, ␤-tubulin homolog, actin homolog, peroxiredoxin-6, hnRNP C1/C2, eIF-5A-1, PROS-27, PP1-gamma, ␣-enolase and PDH E-alpha, whereas GW441756-uncontrolled TrkA dependent targets included hnRNP C1/C2 and HSP70 (Fig. 11). Since activation of TrkA-dependent JNK signaling was also significantly suppressed by SP600125 (Fig. 9A), our results strongly suggest an involvement of JNK in various TrkA-mediated signaling pathways.
Neuronal microtubule cytoskeleton was regulated by tubulin PTM, and many neurodegenerative disorders were related with altered microtubule-based transport [25]. Moreover, JNK was activated by moscatilin-induced tubulin depolymerization in human colorectal cancer cells [26]. JNK activation and tubulin depolymerization were involved in anticancer drug-induced cell cycle arrest and apoptosis in human breast cancer cells [27] and prostate cancer cells [28]. In addition, JNK activation induced actin cytoskeleton changes [29] and played an important role in the regulation of actin stability and migration in neurons and cancer cells [30]. Both ␣tubulin and ␤-tubulin existed in numerous isotypic forms encoded by different genes and underwent various PTMs such as acetylation, phosphorylation, detyrosylation, polyglutamylation, and polyglycylation [31]. Thus, our results suggest that TrkA could play an important role in the cytoskeleton through regulation of ␣-tubulin, ␤-tubulin homolog, and actin homolog.
Peroxiredoxins and HSP70 have been well known to play an important role as a cell survival factor. Peroxiredoxins be- long to a family of multifunctional antioxidant thioredoxindependent peroxidases and play an important role in cellular protection against oxidative stress [32]. Six different peroxiredoxin isozymes showed distinct distribution profiles in different brain regions and different cell types, and peroxiredoxin-6 was expressed in glial cells but not in neurons [33]. Peroxiredoxin-6 was predominantly localized in the cytoplasm, and the intracellular location was changed to the mitochondria after ischemia-reperfusion in mice, protecting cells against mitochondrial dysfunction [34]. Moreover, peroxiredoxins suppressed JNK activation stimulated by tumor necrosis factor-␣ [35] and radiation [36], indicating a cell survival role of peroxiredoxins in various signaling pathways. HSP70 is a major member of the HSP family and protects cells against cellular stresses. HSP70 overexpression inhibited aminoglycoside-induced hair cell death [37], whereas reduced expression of HSP70 was associated with triptolidemediated cell death of neuroblastoma [38]. The chaperone function of HSP70 was required for protecting cells against stress-induced apoptosis [39]. HSP70 prevented activation of JNK [40] and c-Jun [41], leading to inhibition of apoptotic cell death. Thus, our results suggest that TrkA could play an important role in the regulation of cell death through an interaction with peroxiredoxin-6 and HSP70.
hnRNP C1/C2 belongs to the family of nucleic acid binding proteins that are involved in the regulation of pre-mRNA processing and mature mRNA exporting out of the nucleus [42,43]. It has been known that hnRNP C1/C2 was rapidly phosphorylated in response to hydrogen peroxide, indicating a PTM of hnRNP C1/C2 in cell death signaling [44,45]. eIF-5A-1 is one of the two human eIF-5A family members, which were originally proposed to be a translation initiation factor that stimulates the initiation phase of protein synthesis [46]. In addition, eIF-5A has been involved in apoptotic signaling via intrinsic mitochondrial pathway [46] and p53-dependent mechanism [47]. Ubiquitination-proteasome system plays an important role in the regulation of protein stability, and JNK has been involved in this signaling pathway [48,49]. PROS-27 belongs to ubiquitin-proteasome system-related proteins, which have been also involved in the control of cardiomyocyte differentiation of embryonic stem cells [50]. PP1-gamma is one of the three catalytic subunits of protein phosphatase type 1 (PP1), and invasive ductal carcinoma of the breast was associated with enhanced expression of PP1-gamma [51]. Moreover, PP1 was inactivated by JNK, influencing on phosphorylation state of eIF2␣ [52]. In addition, decrease of nuclear PP1 activity was associated with mitotic arrest and apoptosis induced by an extracellular polysaccharide in myeloid leukemia U937 cells [53]. Thus, our results suggest that TrkA could play an important role in the regulation of various cellular processes such as post-transcription, post-translation, ubiquitination, and dephosphorylation through control of hn-RNP C1/C2, eIF-5A-1, PROS-27, and PP1-gamma.
␣-enolase is one of the three enolase isoforms and controls pyruvate synthesis in glycolytic pathway [54]. In addition to glucose metabolism, abnormal regulation of enolase enzymes has been related with Alzheimer's disease [55]. PTMs of ␣-enolase such as phosphorylation, acetylation, and methylation have been identified in pancreatic ductal adenocarcinoma cells [56,57]. PDH E1-alpha is one of the enzymes consisting of pyruvate dehydrogenase (PDH) complex and plays an important role in regulating the flux of glycolytic metabolites into the tricarboxylic acid cycle [58]. Deficiency of the PDH E1-alpha was associated with brain damage and neurological symptoms [59], and PDH E1-alpha activity was regulated by phosphorylation and dephosphorylation in human skeletal muscle [60]. Therefore, our results suggest that TrkA could play an important role in the regulation of glucose metabolism through control of ␣-enolase and PDH E1-alpha, which are key glycolytic enzymes.