Intracellular β-Tubulin/Chaperonin Containing TCP1-β Complex Serves as a Novel Chemotherapeutic Target against Drug-Resistant Tumors Free

Iodoacetamide-based derivatives are known to interact with the low pKa cysteine residues of intracellular proteins, such as tubulin (1), myosin (2), actin (3), HMG-CoA reductase (4), sarcoplasmic reticulum Ca²⁺-ATPase (5), and protein tyrosine phosphatase 1B (6), to yield S-carboxymethyl cysteine derivatives. Recently, 3-(iodoacetamido)-benzoyl ethyl ester (3-IAABE) was found to inhibit tubulin polymerization; cause a double blockade in the cell cycle at the G₁-S transition and M phase through the increased kinase activities of cyclin E, cyclin-dependent kinase 2, cyclin B, and cdc2; and induce activation of caspase-9, caspase-3, and caspase-6, as well as the caspase-3 downstream effector, poly(ADP-ribose) polymerase (PARP), and accompanied cellular DNA fragmentation in CEM cells (7). In addition, 3-IAABE also inhibits the growth of human hepatocarcinoma in nude mice by a 72% decrease in tumor volume. This inhibition is more robust than treatment with vincristine (7), a tubulin-binding and antimitotic agent indicating the feasibility of using iodoacetamide-based derivatives in clinical cancer therapies.

Tubulin is a subunit of intracellular microtubules and is composed of an α and β heterodimer (8). Unlike α-tubulin, which binds GTP irreversibly, β-tubulin binds GTP reversibly and hydrolyzes it to GDP (9). Changes in the concentration of intracellular GDP-bound β-tubulin is known to affect intracellular microtubule dynamics (10), which play critical role in regulating the separation of chromosomes during meiosis and mitosis (11, 12). Several anticancer agents, such as paclitaxel (13), vinblastine (14), and colchicine (15), were designed to disrupt the α/β tubulin heterodimer so as to inhibit microtubule dynamics (16). However, occurrence of drug-resistant tumor cells limits the use of these antimitotic agents in current clinical cancer therapies. Although the molecular mechanisms promoting the induction of drug-resistant tumor cells remain unclear, the overexpression (17–20) or mutation (21, 22) of β-tubulin isotypes is thought to be crucial for this event.

Chaperonin containing TCP1 (CCT) is involved in the folding of cytoskeleton proteins, actin and tubulin (23), and other intracellular proteins, such as cyclin E1 (24), histone deacetylase (25), and protein phosphatase PP2A regulatory subunit B (26). CCT is divided into eight different subunits (α, β, γ, δ, ε, ζ, η, and 𝛉), each with an approximate molecular mass of 60 kDa. Although the functions of CCT containing complexes are still not well understood, the interaction of CCT with tubulin has been illustrated using cryoelectron microscopic (CryoEM) analysis (27). The docking model reveals that CCT subunits interact with different regions of α/β tubulin heterodimers to stabilize the microtubular structure (27). In the case of β-tubulin, for example, CCT-β interacts with the fragment V353-P357 within its COOH terminal region (27).

Although eukaryotic CCT plays an important function in maintaining cellular homoeostasis by assisting the folding of many proteins, the nature of those proteins and the involved cellular pathways have not yet been established. The present study offers a detailed illustration of β-tubulin/CCT-β complex interruption by a synthetic small molecule, N-iodoacetyl-tryptophan (I-Trp), via its incorporation at the Cys³⁵⁴ residue of β-tubulin. Furthermore, this interaction elicited caspase-dependent signaling and consequent cellular apoptosis. Our results also show that interrupting the constitutive β-tubulin/CCT-β complex causes more severe cell apoptosis in the multidrug-resistant MES-SA/Dx5 uterine cancer cells when compared with wild-type MES-SA cells. These observations indicate that targeting the β-tubulin/CCT-β complex may serve as a novel chemotherapeutic strategy for treating clinical multidrug-resistant tumors.

Materials. All the chemicals for organic synthesis were purchased from Aldrich-Sigma. N-iodoacetylated amino acids (1), dansylated I-Trp (I-Trp-Dan) (2), and biotinylated I-Trp (I-Trp-B) (3) were synthesized according to Supplementary Scheme S1. The human cDNA and pET32Xa/Lic vector were purchased from Invitrogen. Tubulin polymerization kit was purchased from Cytoskeleton. Protein A-agarose and antibodies against β-tubulin and CCT-β were purchased from Santa Cruz Biotechnology. Antibodies against caspase-7, caspase-9, and PARP were purchased from Cell Signaling. HEK-293 cells and MES-SA and MES-SA/Dx-5 uterine cancer cells were obtained from the Bioresource Collection and Research Center in Taiwan and cultivated in accordance with manufacturer's guideline.

Plasmid construction and protein expression. The full-length β-tubulin cDNA was subcloned into pET32 Xa/LIC vector according to manufacturer's protocol (Invitrogen). The constructed pET32 Xa/LIC/β-tubulin plasmid was used as a DNA template for site-directed mutagenesis of the Cys³⁵⁴ residue to Ala. Site-directed mutagenesis was carried out by using QuikChange II Site-Directed Mutagenesis kit (Strategene). Recombinant wild-type β-tubulin and C354A mutant β-tubulin proteins were overexpressed in the Escherichia coli strain BL-21(DE3) as COOH terminally hexa-His-tagged proteins.

Protein purification. Cell lysates obtained from a 2-L culture of E. coli strain BL-21(DE3) were resuspended in 80 mL buffer A [12 mmol/L Tris-HCl (pH 7.5), 120 mmol/L NaCl] and loaded onto a 20-mL Ni-NTA resin that was equilibrated with buffer A. The column was washed with excess buffer A containing 10 mmol/L imidazole. His-tagged β-tubulin was eluted with buffer A containing 300 mmol/L imidazole. His-tagged β-tubulin protein was then digested with thrombin protease to remove the peptide tag.

For identifying I-Trp-B–incorporated protein(s), cell lysate (20 mg) of HEK-293 cells was prepared in 1 mL buffer [25 mmol/L Tris-HCl (pH 7.4), 150 mmol/L NaCl] and incubated with I-Trp-B compound (200 μmol/L) overnight at 4°C. Proteins linked with I-Trp-B were purified by affinity chromatography using a HiTrap Streptavidin affinity column according to manufacturer's guideline (GE Healthcare).

Mass spectrometric analysis. HEK-293 cells or recombinant β-tubulin (100 μg) were treated with I-Trp-Dan (1 μmol/L) for 24 h at 37°C or 2 h at room temperature in the dark, respectively. I-Trp-Dan–treated cell lysates (100 μg) or recombinant β-tubulin (30 μg) were subjected to electrophoresis on a 10% SDS-PAGE gel. The fluorescence-labeled protein band, visualized under a UV lamp, was excised from the gel and subjected to in-gel tryptic digestion and liquid chromatography-electrospray ionization-tandem mass spectrometric analysis. The raw data were converted to the Mascot (Matrix Science) generic format and searched against the NCBInr human protein database. For protein fingerprinting analysis, the resultant m/z values of each peptide fragment were compared with predicted molecular weight of peptide fragment derived from β-tubulin IVb subtype to assess possible I-Trp-Dan–mediated alkylation of cysteines.

Cell transfection. The gene encoding the β-tubulin T351-S364 fragment was synthesized by Mission Biotech and subcloned into an EGFP-IRES2 vector with an XhoI/EcoRI restriction site. The transfection of plasmid DNA or CCT-β small interfering RNA (siRNA) was performed according to Lipofectamine transfection guidelines (Invitrogen).

SDS-PAGE and Western blotting. Cell lysate (100 μg) or recombinant β-tubulin (20 μg) was subjected to SDS-PAGE. In the case of I-Trp-Dan compound treatment, proteins were detected by exposure to UV light (280 nm) before Coomassie blue staining. For Western blotting, proteins were transferred onto a polyvinylidene difluoride membrane (Millipore) and incubated with an antibody against β-tubulin, CCT-β, caspase-7, caspase-9, or PARP. After incubation with horseradish peroxidase–conjugated secondary antibody, immunoreactive protein bands were visualized using the enhanced chemiluminescence system (Amersham Bioscience).

Immunoprecipitation. Cell lysates (200 μg) were diluted in 1 mL of cell lysis buffer [10 mmol/L Tris-HCl (pH 7.4), 140 mmol/L NaCl, 3 mmol/L MgCl₂, 2 mmol/L EDTA, 5 mmol/L EGTA, 0.5% Triton X-100, 4% protease inhibitor cocktail; Merck Biosciences] and incubated with 2 μg of anti–β-tubulin antibody for 2 h at 4°C, followed by precipitation with 20 μL of protein A-agarose beads for 1 h at 4°C. The immunoprecipitates were analyzed by SDS-PAGE/Western blotting using a CCT-β antibody. In addition, the blots were stripped using a commercial kit (Millipore) and reprobed with an anti–β-tubulin antibody.

Flow cytometric analysis. Cells (6 × 10⁵) were fixed in 70% ice-cold ethanol for 30 min at room temperature. Cells were washed with PBS once and then incubated with propidium iodide (PI) staining solution (PBS containing 0.1% bovine serum albumin, 0.1% RNase A, and 20 ng/mL of PI) for 30 min at room temperature in the dark. After the incubation, the cells were analyzed by flow cytometry to assess their DNA content.

Determination of caspase activities. The caspase activity assay was performed according to the manufacturer's guidelines (BioVision). Fold increases in caspase activity was determined by comparing these results with the level of the untreated control.

Computer modeling. Molecular docking of I-Trp compounds into the β-tubulin protein structure was evaluated using GEMDOCK software (28). Intermolecular interactions were predicted using Swiss-PdbViewer 3.7.

Synthesis of N-iodoacetylated amino acids and evaluation of their cytotoxic effects. A series of N-iodoacetylated amino acids were synthesized in accordance with the procedure described in Supplementary Scheme S1. The identities of all synthesized compounds were confirmed by mass spectrometric and nuclear magnetic resonance analyses (data not shown). The treatment of HEK-293 cells with 1 μmol/L I-Asp, I-Glu, I-Ser, and I-Trp (chemical structures shown in Fig. 1A), but not I-Gly, I-Lys, or iodoacetamide, induced the accumulation of sub-G₀ cell population compared with untreated cells (Supplementary Fig. S1). Effects of these compounds on sub-G₀ cell accumulation of HEK-293 cells were dose dependent at concentrations ranging from 0.001 to 1 μmol/L (Fig. 1B,, left) and also caused intracellular DNA fragmentation (Fig. 1B , right), indicating intracellular apoptotic signaling. Based on flow cytometric analysis, I-Trp exhibited the most potent cytotoxic effect on HEK-293 cells. Thereafter, we used the I-Trp compound in subsequent experiments to elucidate the mechanism by which it induced cell apoptosis.

Because caspase activation is thought to be critical for intracellular apoptotic signaling (29), the activities of intracellular caspases were determined in I-Trp–treated HEK-293 cells. Except caspase-1, the increased activities of caspase-2 through caspase-10 were detected after HEK-293 cells were treated for 24 hours with I-Trp (1 μmol/L), but not iodoacetamide (Fig. 1C). Especially, I-Trp induced an approximate of 2.5-fold elevation in caspase-3/caspase-7 activities in HEK-293 cells compared with control cells (Fig. 1C). Furthermore, the inclusion of caspase-3/caspase-7 peptide inhibitor (DEVD) at concentrations ranging from 0.3 to 3 μmol/L significantly (P < 0.05) suppressed the I-Trp–induced sub-G₀ cell accumulation of HEK-293 cells in a dose-dependent manner (Fig. 1D).

Identification of intracellular protein target(s) of I-Trp in HEK-293 cells. To identify the intracellular protein target(s) of I-Trp, a dansylated I-Trp compound (I-Trp-Dan; Fig. 2A,, insert) was synthesized and given to HEK-293 cell cultures. As shown in Fig. 2A, I-Trp-Dan induced generation of apoptotic cells in a dose-dependent manner at concentrations ranging from 0.05 to 2 μmol/L. SDS-PAGE analysis of cell lysates obtained from I-Trp-Dan (1 μmol/L)–treated cells revealed a fluorescent protein band with a molecular weight of ∼55 kDa after exposure of the gel to UV light (Fig. 2B). The fluorescent protein band was excised from gel and subjected to mass spectrometric analysis. As shown in Supplementary Table S1, CCT-β subunit, tubulin α/β subunits, and other proteins were identified in the mass spectrometric analysis. These proteins all have similar molecular weights and run on SDS-PAGE at similar positions. To identify the real target, an I-Trp-B compound (Fig. 2C,, insert) was synthesized and incubated with HEK-293 cell lysates. Subsequently, proteins covalently linked to this I-Trp-B compound were purified by affinity chromatography using a streptavidin-immobilized column. Large amounts of unbound proteins were washed from the column, and a small quantity of streptavidin-bound protein was eluted using a urea solution (Fig. 2D). Western blot analysis revealed that β-tubulin, not α-tubulin or CCT-β, was present in the eluted solution (Fig. 2D), suggesting that β-tubulin acts as an intracellular protein target of I-Trp in HEK-293 cells.

To further confirm that β-tubulin was an authentic target of I-Trp, recombinant β-tubulin was produced as described in Materials and Methods. As shown in Fig. 3A, the treatment of recombinant β-tubulin, not CCT-β (Supplementary Fig. S2), with the I-Trp-Dan compound generated a fluorescent protein band on SDS-PAGE after UV light exposure. Furthermore, the incubation of recombinant β-tubulin with I-Asp, I-Glu, I-Ser, I-Trp, or I-Trp-B compounds before the I-Trp-Dan treatment abrogated generation of the fluorescent protein band (Fig. 3B), indicating that those synthesized compounds likely interacted and alkylated the same cysteine residue of β-tubulin.

Furthermore, protein fingerprinting analyses indicated that the Cys³⁵⁴ residue within the β-tubulin fragment TAVCDIPPR was alkylated with the I-Trp-Dan compound. This observation is based on the resultant mass spectrometer peak at m/z = 1532.751, approximately equal to the predicted mass values of the TAVCDIPPR peptide fragment incorporated with an N-acetyl-Trp-Dan group (Fig. 3B). In contrast to the wild-type β-tubulin, a Cys³⁵⁴Ala mutant failed to produce a fluorescently labeled band on SDS-PAGE (Fig. 3C), consistent with the conclusion that Cys³⁵⁴ reacted with I-Trp. Interestingly, the Cys³⁵⁴ residue is located within the peptide fragment VCDIP of β-tubulin (the magenta region in Supplementary Fig. S3), which serves as the CCT-β binding site (27). We further found that I-Trp caused the disruption of the constitutively formed β-tubulin/CCT-β complex, as judged by the reduced immunoprecipitates of β-tubulin/CCT-β complex in cell lysates of I-Trp–treated HEK-293 cells (Fig. 3D).

Intermolecular interactions between the I-Trp compound and β-tubulin. Because the I-Trp compound disrupts constitutively associated β-tubulin/CCT-β complex by alkylating the Cys³⁵⁴ residue of β-tubulin, computer modeling was used to simulate the intermolecular interactions between I-Trp and β-tubulin. Docking experiments display that I-Trp binds to β-tubulin at the region near the Cys³⁵⁴ residue (Fig. 4A). The intermolecular distance between the iodo group of I-Trp and the SH group of Cys³⁵⁴ was calculated to be 3.24 Å (enlarged in Fig. 4A). Possible hydrogen bond interactions with the Asp³⁵⁵ residue and hydrophobic interactions with Leu⁴², Pro²⁴³, Gly²⁴⁴, and Gln²⁴⁵ residues of β-tubulin may also be formed (Supplementary Fig. S4B). These interactions ensure the specific incorporation of I-Trp into this region of β-tubulin. Similar binding modes of I-Trp-Dan and I-Trp-B with β-tubulin were also predicted in the same region (Supplementary Fig. S4A and B).

Effect of the I-Trp compound on tubulin polymerization. The binding site of I-Trp in β-tubulin (Fig. 4B, peptide fragment VCDIP in magenta and Cys³⁵⁴ in yellow) is different from those of other β-tubulin–binding agents, such as Taxol (Fig. 4B,, TXL), colchicine, and vinblastine (Supplementary Fig. S3). As shown in the insert in Fig. 4C, the pretreatment of recombinant β-tubulin with paclitaxel, as well as colchicine and vinblastine (data not shown), did not affect the incorporation of I-Trp-Dan into β-tubulin. Unlike paclitaxel, the treatment with I-Trp failed to induce in vitro tubulin polymerization (Fig. 4C).

Effects of interrupting the β-tubulin/CCT-β interaction in multidrug-resistant cancer cells. Cys³⁵⁴ residue of β-tubulin is located within the CCT-β–binding sequence VCDIP, which is conserved in all β-tubulin isotypes of mammalian cells (Supplementary Fig. S5). Therefore, MES-SA/Dx5 cell line, a multidrug-resistant derivative from uterine MES-SA sarcoma cell line (30), was used to examine whether interrupting β-tubulin/CCT-β association contributes a new strategy to treat tumors with drug resistance to β-tubulin–binding agents, such as paclitaxel, vinblastine, and colchicine. MES-SA and MES-SA/Dx5 cells were cultivated and transfected with the gene encoding the peptide fragment TAVCDIPPR to block the formation of the intracellular β-tubulin/CCT-β complex. As shown in Fig. 5A, overexpression of this peptide fragment caused a more predominant accumulation of sub-G₀ cells in MES-SA/Dx5 cells compared with the wild-type MES-SA cells. Particularly, in the cells transfected with 2 μg of plasmid containing this truncated β-tubulin gene, an approximate of 10-fold larger population of sub-G₀ cells was observed in MES-SA/Dx5 cells relative to MES-SA cells (Fig. 5A,, left) and intracellular DNA fragmentation (right) was detected. Similar observations were made using 1 μmol/L I-Trp or 10 to 100 μmol/L myristoylated synthetic peptide (Myr-TAVCDIPPRG; Supplementary Fig. S6). Furthermore, immunoprecipition revealed that overexpression of the truncated β-tubulin reduced the formation of intracellular β-tubulin/CCT-β complex in MES-SA/Dx5 cells (Fig. 5B). Surprisingly, Western blot analysis revealed that an increased CCT-β, not β-tubulin, protein expression was detected in MES-SA/Dx5 cells compared with the wild-type MES-SA cells (Fig. 5C). This could explain the predominant activity of overexpressed β-tubulin fragment in causing sub-G₀ cell accumulation in MES-SA/Dx5 cells. The knockdown of CCT-β expression using CCT-β–specific siRNA markedly suppressed endogenous CCT-β protein levels (Fig. 5D,, top) and significantly (P < 0.05) inhibited I-Trp compound–induced sub-G₀ cell accumulation in MES-SA/Dx5 cells compared with untreated and control experiments (Fig. 5D).

Caspase activation in MES-SA/Dx5 cells transfected with truncated β-tubulin. Further experiments were performed to examine the activation of caspase proteins during apoptosis of MES-SA/Dx5 cells induced by the overexpression of the β-tubulin peptide fragment T351-S364. As shown in Fig. 6A, except caspase-1, all the other caspases were activated in the MES-SA/Dx5 cells transfected with the truncated β-tubulin. The cleaved forms of caspase-3, caspase-7, and caspase-9 (Fig. 6B) and the fragmentation of PARP, an intracellular downstream target of caspase-3/caspase-7 (Fig. 6C), were observed in the MES-SA/Dx5 cells transfected (T) with truncated β-tubulin gene compared with the untreated (U) control. Interestingly, the profile of caspase activation in the truncated β-tubulin–expressing MES-SA/Dx5 cells was similar to that in the I-Trp–treated HEK-293 cells, indicating that a unique intracellular caspase cascade is involved in mediating apoptotic signaling induced by interrupting constitutive β-tubulin/CCT-β complexes (see Discussion).

CryoEM analysis reveals that CCT-β interacts with amino acid residues 353 to 357 (VCDIP) of β-tubulin (27), a region highly conserved in β-tubulin isotypes (see Supplementary Fig. S5). By mass spectrometric analysis, the conserved Cys³⁵⁴ residue of β-tubulin was identified to serve as the incorporation site of I-Trp compound. Furthermore, treatment with I-Trp or overexpression of truncated β-tubulin (T351-S364) indeed reduced the formation of intracellular β-tubulin/CCT-β complexes and caused cell apoptosis. Because I-Trp did not affect in vitro tubulin polymerization (Fig. 4C), the disruption of intracellular β-tubulin/CCT-β complexes by I-Trp may induce cell apoptosis by causing microtubule disassembly in accordance with the results presented in a previous study (31).

Recently, tubulin-binding agent–resistant tumors were considered to be correlated with the occurrence of β-tubulin mutations (22) or overexpression of individual β-tubulin isotypes (17, 19, 20) such as class III β-tubulin. Our data indicate that CCT-β is highly expressed in multidrug-resistant MES-SA/Dx-5 uterine cancer cells, and the destruction of intracellular β-tubulin/CCT-β complex by the small molecule I-Trp, synthetic peptide Myr-TAVCDIPPRG (T351-G360), or overexpression of the β-tubulin fragment T351-S364 causes a predominant cell apoptosis response in MES-SA/Dx-5 cells when compared with wild-type MES-SA cells. Conversely, the knockdown of CCT-β protein expression rendered MES-SA/Dx5 cells less sensitive to apoptotic signals elicited through interrupting constitutive β-tubulin/CCT-β complex by I-Trp compound. Because increased CCT-β expression is also detectable in several malignant tumors (32) and might occur commonly in drug-resistant tumors, the constitutive β-tubulin/CCT-β complex may serve as a useful chemotherapeutic target for treating clinical multidrug-resistant or CCT-β-overexpressing tumors.

The interaction of microtubules with mitochondria was identified in myocytes (33) and brain cells (34). Recently, microtubular depolymerization was found to induce mitochondria-dependent apoptotic signaling accompanied with activation of caspase-9 and caspase-3 (35). On the other hand, the treatment with tubulin-binding agent Taxol transactivates the intracellular domain of the Notch receptor and modulates the Notch signaling pathway–dependent survival of several cancer cells (36, 37). Besides, the association of tubulin with other receptors, including the muscarinic receptor (38), γ-aminobutyric acid receptor (39), purinergic receptor (40), and interleukin-2 receptor (41), was also previously documented. This information supports our data indicating that the disruption of the constitutive β-tubulin/CCT-β complex by treatment with I-Trp or transfection with the β-tubulin fragment (T351-S364) activates both intrinsic (caspase-3, caspase-6, caspase-7, and caspase-9) and extrinsic (caspase-2, caspase-8, and caspase-10) caspase-dependent apoptotic signaling.

Although caspase-1 clusters phylogenetically with caspase-4 and caspase-5 and is a member of the group of the so-called inflammatory caspases, caspase-4 and caspase-5–dependent apoptotic signaling is not mediated by caspase-1 (42). Furthermore, the in vitro cleavage of pro-caspase-3 by caspase-4 and caspase-5 suggests that caspase-4 and caspase-5 might be involved in the intracellular apoptotic signaling through activated caspase-3 (42–44). In the present study, either I-Trp treatment or the truncated β-tubulin transfection induced the activation of intracellular caspase-4 and caspase-5, but not caspase-1. A reasonable interpretation is that caspase-1 might be specialized to mediate extracellular inflammatory signals (45, 46), because caspase-1 expression was detectable in HEK-293 cells (47, 48). On the other hand, recent studies show that caspase-4 is localized to the surface of the endoplasmic reticulum (ER) and is involved in ER stress–induced apoptosis. Besides, it was well known that intracellular ER movement is dependent on microtubule mobilization (49, 50). Thus, caspase-4 activation may reflect the involvement of ER-dependent apoptotic signaling during β-tubulin/CCT-β complex–dependent apoptosis.

In conclusion, the interaction of iodoacetamide derivatives with the sulfhydryl group of cysteine residues provides a useful approach for us to discover a novel intracellular target for cancer therapy. By using the I-Trp-Dan compound, which was specifically incorporated into β-tubulin in HEK-293 cells, we define a novel β-tubulin/CCT-β–dependent apoptotic event mediated by intrinsic, extrinsic, and ER-dependent caspase signaling cascades. Our data also show that disrupting constitutive β-tubulin/CCT-β complexes via overexpression of truncated β-tubulin causes a more severe cell apoptosis in multidrug-resistant MES-SA/Dx5 cells compared with wild-type MES-SA cells. Because the binding site of CCT-β is conserved among β-tubulin isotypes and is located away from the vinblastine-, colchicine-, and taxol-binding sites, the modulation of β-tubulin/CCT-β complex dynamics may serve as a useful chemotherapeutic strategy for treating tubulin-binding agent–resistant tumors.

No potential conflicts of interest were disclosed.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank the Core Facilities for Proteomics at Academia Sinica in Taiwan for the mass technical service and Shih-Hsun Chen for his assistance in computer modeling.