Abstract

   The spider peptide toxins HNTX-III and JZTX-I are a specific inhibitor
   and activator of TTX-S VGSCs, respectively. They play important roles
   in regulating MAT-LyLu cell metastasis in prostate cancer. In order to
   identify key biomarkers involved in the regulation of MAT-LyLu cell
   metastasis, iTRAQ-based quantitative phosphoproteomics analysis was
   performed on cells treated with HNTX-III, JZTX-I and blank. A total of
   554 unique phosphorylated proteins and 1779 distinct phosphorylated
   proteins were identified, while 55 and 36 phosphorylated proteins were
   identified as differentially expressed proteins in HNTX-III and JZTX-I
   treated groups compared with control groups. Multiple bioinformatics
   analysis based on quantitative phosphoproteomics data suggested that
   the differentially expressed phosphorylated proteins and peptides were
   significantly associated with the migration and invasion of prostate
   tumors. Specifically, the toxins HNTX-III and JZTX-I have opposite
   effects on tumor formation and metastasis by regulating the expression
   and phosphorylation level of causal proteins. Herein, we highlighted
   three key proteins EEF2, U2AF2 and FLNC which were down-regulated in
   HNTX-III treated cells and up-regulated in JZTX-I treated cells. They
   played significant roles in cancer related physiological and
   pathological processes. The differentially expressed phosphorylated
   proteins identified in this study may serve as potential biomarkers for
   precision medicine for prostate cancer in the near future.

   Keywords: HNTX-III, JZTX-I, prostate cancer, quantitative
   phosphoproteomics, bioinformatics

1. Introduction

   Prostate cancer is one of the most common non-cutaneous genital
   malignancies in men, with an estimated 1.6 million cases and 366,000
   deaths worldwide every year [[40]1]. Metastatic disease is the leading
   cause of prostate-related death [[41]2]. Metastatic prostate cancer
   resists hormone therapy and other conventional treatments
   [[42]3,[43]4]. However, high metastasis is one of the inevitable
   characteristics of most cancers, therefore, it is necessary and urgent
   to develop efficient targeted drugs to inhibit the high metastasis of
   prostate cancer. In the past few years, research on the role of ion
   channels in cancer have attached importance to the involvement of
   various channel types in cancer cell metabolism and the tumor
   microenvironment [[44]5].

   Voltage-gated sodium channels (VGSCs) are a kind of transmembrane
   glycoprotein which consist of pore-forming α-subunits (260 kDa) and
   auxiliary β subunits (33–36 kDa) [[45]6]. It is known that nine
   different genes (SCN1A-SCN11A) encode nine distinct sodium channels
   (Nav1.1–Nav1.9) [[46]7,[47]8]. In pharmacology, according to the
   sensitivity of sodium channel subtypes to tetrodotoxin (TTX), sodium
   channels can be divided into TTX-sensitive (TTX-S) and TTX-insensitive
   (TTX-R) [[48]8]. In recent years, studies in rat prostate cancer models
   have shown that VGSCs are related to invasive metastasis in vitro
   [[49]9], and TTX-S VGSCs can reduce the invasiveness of prostate cancer
   cells [[50]10,[51]11,[52]12], which indicates that VGSCs may play an
   important role in tumor migration. Hainan Toxin-III (HNTX-III) is a 33
   amino acid peptide toxin isolated from the venom of the Chinese bird
   spider (Ornithoctonus hainana that can reduce the activity of TTX-S
   VGSCs [[53]13]. Jingzhao Toxin-I (JZTX-I) is a 33 amino acid and three
   disulfide bonds peptide toxin from the venom of the spider Chilobrachys
   jingzhao [[54]14]. It can delay the rapid inactivation mechanics of
   TTX-S VGSCs [[55]13]. The effect of the spider peptide toxins HNTX-III
   and JZTX-I on MAT-LyLu in prostate cancer cells is a very interesting
   and innovative study. In our previous work, the spider peptide toxins
   HNTX-III and JZTX-I have been used in the treatment of prostate cancer,
   and the preliminary cellular study suggested that both of them could
   regulate the metastasis of prostate cancer [[56]15]. To better
   understand the underlying molecular mechanisms of the regulation, we
   have now performed comparative phosphoproteomics in two groups:
   HNTX-III treatment (114 isobaric tags for relative and absolute
   quantification (iTRAQ) tags) with blank group (116 iTRAQ tags) are
   named as experimental group I; the JZTX-I treatment (115 iTRAQ tags)
   with the blank group (116 iTRAQ tags) are named as experimental group
   II.

   Protein phosphorylation has a profound effect on the dynamic processes
   of the cell because the role of kinases and phosphatases is the basis
   of many major biological functions [[57]16]. LC-MS-based
   phosphoproteomics is an efficient and powerful method in recognizing
   large amounts of phosphorylated proteins and studying cell signaling
   pathways. Phosphoproteomics can reveal the key role of phosphorylated
   proteins in the signaling pathways. It has been widely applied in
   exploring the roles of multiple toxins such as Clostridioides
   difficile, Microcystis aeruginosa and snake venom in human diseases
   [[58]17,[59]18,[60]19]. Multiple bioinformatics analysis based on
   quantitative phosphoproteomics data help us to better understand the
   biological function and topological associations of the significantly
   differentially expressed phosphorylated proteins (DEPs) between case
   and control groups. It may largely help us to uncover the underlying
   molecular mechanisms of prostate cancer regulation by the peptide
   toxins HNTX-III and JZTX-I.

2. Results

2.1. Identification of Phosphorylated Proteins and Phosphopeptides in the
MAT-LyLu Cells

   To identify significant biomarkers involved in toxin regulation of
   prostate cancer, we conducted a quantitative phosphoproteomic analysis
   by iTRAQ integrated with LC-MS/MS to identify differentially expressed
   phosphorylated proteins and phosphopeptides in high metastatic MAT-LyLu
   cells. To obtain statistically reliable data, triplicate runs were
   performed for protein and phosphopeptide identification and
   quantification. We successfully identified 436, 436, and 408
   phosphorylated proteins in replicate 1, replicate 2 and replicate 3,
   respectively. At the same time, 1144, 1182 and 1126 phosphorylated
   peptides were successfully identified in replicate 1, replicate 2 and
   replicate 3, respectively ([61]Figure 1A,B, [62]Supplementary
   Materials, Tables S1 and S2). In total, 554 non-redundant
   phosphorylated proteins and 1779 distinct phosphopeptides were
   identified. Among the 554 identified proteins, 308 common proteins were
   identified in three replicates; more than 75.45% (418) were identified
   in at least two replicates. Among the 1779 identified phosphopetides,
   638 common peptides with same phosphorylation sites were identified in
   three replicates, more than 71.16% (1266) were identified at least two
   replicates.

Figure 1.

   [63]Figure 1
   [64]Open in a new tab

   Statistical information of phosphoproteomics in the Mat-lyLu cells
   based on iTRAQ method. (A,B) showed the phosphoproteins and
   phosphopeptides identified in the triplicate analyses respectively
   based on iTRAQ method. (C) The proportions of different phosphorylation
   sites (serine (p-Ser), threonine (p-Thr) and tyrosine (p-Tyr)); (D)
   Distribution of phosphopeptides depending on their number of p-sites;
   (E) Distribution of phosphopeptides based on their length; (F)
   Distribution of phosphorylation proteins based on their number of
   p-sites.

   Furthermore, we statistically analyzed the phosphorylation sites and
   the length of peptides. A total of 3452 phosphorylated peptides and
   3784 p-sites were identified. Among all the p-sites, 3407 serine
   p-sites, 361 threonine p-sites and 16 tyrosine p-sites were identified,
   respectively. Thus, the percentage distribution of p-sites was
   Ser/Thr/Tyr = 90.04%/9.54%/0.42%, which was in accordance with the
   value reported previously (Ser:Thr:Tyr = 90%:9.9%:0.1%) ([65]Figure 1C)
   [[66]20]. Among the 3452 phosphorylation peptides, 90.53% (3125/3452)
   singly phosphorylated phosphopeptides, 9.3% (321/3452) doubly
   phosphorylated phosphopeptides, and 0.17% (6/3452) phosphopeptides
   carrying three p-sites ([67]Figure 1D) were identified. The length of
   most phosphopeptides (2663/3452 or approximately 77.14%) ranged from 8
   to 20 amino acids, ([68]Figure 1E). Moreover, we detected the number of
   p-sites in a phosphoprotein, the results showed that the proteins
   contained one, two and three p-sites accounted for 45.71% (272/595),
   22.69% (135/595) and 14.79% (88/595), respectively. 16.81% (100/595) of
   the proteins contained more than three p-sites ([69]Figure 1F).

2.2. Identification of DEPs

   To ensure the reliability of the quantitative data, only the
   phosphorylated proteins (417) identified in two or more replicates were
   used for subsequent analysis. p-value < 0.05 and |log2(foldchange)| >
   0.263 were set as the statistical thresholds for DEPs identification.
   In experimental group I (HNTX-III treatment group (114 isobaric tags
   for relative and absolute quantification (iTRAQ) tags) with blank group
   (116 iTRAQ tags)) and II (JZTX-I treatment group (115 iTRAQ tags) with
   the blank group (116 iTRAQ tags)), 55 (15 up-regulated and 40
   down-regulated proteins) and 36 (20 up-regulated and 16 down-regulated
   proteins) phosphorylated DEPs were identified, respectively ([70]Figure
   2A,B). Among them, EEF2, U2AF2 and FLNC were significantly
   down-regulated in experimental group I and significantly up-regulated
   in experimental group II. The three pivotal DEPs specific statistical
   indicators are presented in the [71]Table 1. The clustering of
   experimental group I and experimental group II in the three replicates
   is shown in [72]Figure 2C,D.

Figure 2.

   [73]Figure 2
   [74]Open in a new tab

   Identification of DEPs. (A,B) showed the DEPs identified in
   experimental group I and II, respectively. They were both compared with
   the control group. And each group of experiments were repeated three
   times. The red dots are upregulated DEPs with log2(foldchange) > 0.263.
   The green dots are down DEPs with log2(foldchange) < 0.263. Among them,
   the red triangles and the green triangles in (A,B) are the three
   specific DEPs that reversely regulated in experimental group I and II.
   (C,D) Cluster analysis based on 55 and 36 DEPs in the experimental
   group I and II, respectively. Red and blue show the higher and smaller
   fold-change of up-regulated and down-regulated DEPs, respectively.
   REP1: Repeat the experiment for the first time; REP2: Repeat the
   experiment for the second time; REP3: Repeat the experiment for the
   third time.

Table 1.

   Three specific DEPs reversely regulated after HNTX-III and JZTX-I
   treatment.
   Official Symbol Official Full Name p-Value ^1 Fold Change ^2
   114:116 ^3 115:116 ^4 114:116 115:116
   U2AF2 U2 snRNP auxiliary factor large subunit 1.18 × 10^−2 1.72 × 10^−2
   0.43 1.61
   EEF2 Elongation factor 2 2.12 × 10^−2 9.79 × 10^−3 0.63 4.13
   FLNC Filamin-C 1.75 × 10^−2 9.34 × 10^−3 0.64 2.17
   [75]Open in a new tab

   ^1p-value < 0.05 is defined as a significant difference between the
   experimental group and the control group. The values here are presented
   in scientific notation. ^2 Fold change value refers to the ratio of the
   expression level of the experimental group and the control group. Here
   it is defined as >1.20 is up-regulation, and <0.83 is down-regulation.
   ^3 The HNTX-III treatment (114 iTRAQ tags) with blank group (116 iTRAQ
   tags) named as experimental group I. ^4 The JZTX-I treatment (115 iTRAQ
   tags) with the blank group (116 iTRAQ tags) named as experimental group
   II.

2.3. Identification of Overrepresented Kinase Targeting Peptide Motifs

   In total, 12 serine motifs and one threonine motif were significantly
   enriched ([76]Figure 3A). Each motif score and their corresponding
   kinases are presented in [77]Figure 3B. Among the 12 serine motifs, the
   top four known motifs were psDXE, RSXps, psPXR, and RRXps according to
   motif score, there were potential substrates of casein kinase II (CK2),
   AKT-like kinase, cyclin-dependent kinases (CDKs) and protein kinase A
   (PKA), respectively. However, ptP is the most common threonine motif
   which located in the nucleus, cytosol, and secreted proteins could be
   targeted for mitogen-activated protein kinase (MAPK) [[78]21,[79]22]. A
   column diagram showed that the number of phosphopeptides that contain
   each of the overrepresented phosphorylation motif, we observed a lot of
   phosphopeptides containing the psP motif ([80]Figure 3C), which is a
   known target for proline-directed kinase including cyclin-dependent
   kinases and mitogen-activated kinases. Remarkably, our result shows
   that the phosphorylation level of cyclin dependent kinase 1 (CDK1) was
   significantly changed ([81]Figure 3D).

Figure 3.

   [82]Figure 3
   [83]Open in a new tab

   Motif analysis of identified phosphopeptides. (A) 12 serine motifs and
   one threonine motif were significantly enriched according to Motif-X
   online evaluation. (B) Score and the type of overrepresented kinase of
   each identified motif. (C) Number of phosphopeptides that contain a
   specific overrepresented phosphorylation motif. (D) Phosphorylation
   level of CDK1 after HNTX-III and JZTX-I treatment.

2.4. Gene Functional Annotation

   In biological processes (BP), the DEPs identified in experimental group
   I were mainly significantly enriched in “cell-cell adhesion”,
   “ATP-dependent chromatin remodeling”, “negative regulation of mRNA
   splicing, via spliceosome”, “nucleosome positioning”, “RNA splicing”,
   “positive regulation of RNA splicing”, “mRNA processing” and
   “NLS-bearing protein import into nucleus”. “Cell adhesion” and
   “spliceosome” were closely related to cell proliferation. The molecular
   function (MF) is mainly involved in “cell-cell adhesion” and “cadherin
   binding involved in cell-cell adhesion”, which easily leads to the
   occurrence of cancer ([84]Figure 4A). The DEPs identified in
   experimental group II accounted for multiple entries, including “cell
   adhesion”, “actin filaments”, “phosphorylation protein kinase
   regulation”, they are significantly related to cancer [[85]23]. As for
   molecular functions, they mainly involved in “poly(A) RNA binding”,
   “nucleotide binding”, “cadherin binding involved in cell-cell
   adhesion”, “protein kinase binding”, “C2H2 zinc finger domain binding”,
   “ankyrin binding” and “enzyme binding” ([86]Figure 4B). These items
   play key roles in regulating the changes of cell protein
   phosphorylation level, participating in cell signal transduction and
   cell migration. Subcellular localization analysis showed that DEPs are
   mainly located in exogenous components of cytoplasmic membrane, the
   expected results supported that toxin stimulation mediated VGSCs leads
   to changes in membrane protein expression, resulting in cell adhesion.
   Among them, candidate proteins EEF2, U2AF2 and FLNC were significantly
   annotated in “cell adhesion” (p-value: 6.02 × 10^−3). “Cell adhesion”,
   as the most prominently annotated cell process, is closely related to
   the occurrence of cancer. Once the proteins involved in this process
   are overexpressed, it is very easy to cause tumors
   [[87]24,[88]25,[89]26].

Figure 4.

   [90]Figure 4
   [91]Open in a new tab

   Functional annotation and enrichment analysis based on DEPs. (A,B)
   showed significant terms that the DEPs enriched in for experimental
   group I and II. Most of DEPs including EEF2, U2AF2 and FLNC from both
   groups are enriched in cell activity related terms such as “cell-cell
   adhesion”, “negative regulation of mRNA splicing, via spliceosome” and
   “RNA splicing”. BP: biological process; CC: cell component; MF:
   molecular function.

2.5. DEPs Based Pathway Analysis

   The DEPs identified in experimental group I and II were significantly
   enriched in the “spliceosome”, “regulation of actin cytoskeleton”,
   “proteoglycans in cancer”, “focal adhesion”, “citrate cycle” (TCA
   cycle) and “FC gamma R-mediated phagocytosis” pathways ([92]Figure 5).
   Among them, candidate proteins EEF2, U2AF2 and FLNC are significantly
   enriched in the three pathways of “spliceosome” (p-value: 1.01 ×
   10^−3), “proteoglycans in cancer” (p-value: 8.58 × 10^−2) and “focal
   adhesion” (p-value: 1.48 × 10^−3). “Spliceosome” and “Fc gamma
   R-mediated phagocytosis” pathways were closely associated with the
   proliferation and metastasis of prostate cancer cells. “Focal adhesion”
   may lead to uncontrolled proliferation of cells, which is one of the
   main factors of tumor formation. “TCA cycle” is the central pathway for
   oxidative phosphorylation of cells and meets their biological energy,
   biosynthesis and redox balance requirements [[93]27]. With the
   formation and development of tumors, changes in cell and enzyme
   activity may alter the composition and structure of proteoglycans, and
   thus alter their function. Many cancers, including prostate cancer, use
   these proteoglycan changes to promote their survival, growth, and
   spread [[94]28].

Figure 5.

   [95]Figure 5
   [96]Open in a new tab

   Pathway analysis based on DEPs. The results of pathway enrichment
   analysis based on DEPs identified in experimental group I and II. As
   expected, the candidate proteins EEF2, U2AF2 and FLNC play multiple and
   crucial roles in cancer-related pathways such as “focal adhesion”,
   “spliceosome” and “TCA cycle”.

2.6. The Construction of PPI Network

   The protein-protein interaction network diagram constructed by the DEPs
   identified in experimental group I ([97]Figure 6A) and experimental
   group II ([98]Figure 6B) shows that candidate proteins EEF2, U2AF2 and
   FLNC are key proteins in the PPI network. They can coordinate with
   other DEPs or their upstream and downstream key factors to regulate the
   proliferation and metastasis of prostate cancer cells.

Figure 6.

   [99]Figure 6
   [100]Open in a new tab

   Protein-protein interaction analysis and network reconstruction based
   on differentially expressed genes. (A,B) showed the protein-protein
   interactions of DEPs from experimental group I and II. The results
   indicated that the candidate proteins EEF2, U2AF2 and FLNC could
   regulate the proliferation and metastasis of prostate cancer cells in a
   coordinated way with other DEPs or their upstream and downstream key
   factors.

2.7. Analysis of Gene-Disease Associations

   In order to focus on understanding the diseases associated with the
   DEPs identified in experimental group I and experimental group II, the
   gene-disease association was analyzed. Among them, we mainly focused on
   the disease types associated with three key factors: EEF2, U2AF2 and
   FLNC, which were significantly down-regulated in experimental group I
   and significantly up-regulated in experimental group II. According to
   the network diagram ([101]Figure 7), it can be seen that these three
   types are involved in the formation of various diseases, among which,
   they are all related to malignant neoplasms, neoplasms, primary
   malignant neoplasm, carcinogenesis and neoplasia. Both EEF2 and FLNC
   were associated with the formation of prostate cancer.

Figure 7.

   [102]Figure 7
   [103]Open in a new tab

   Gene-disease association study of the 3 candidate factors. The results
   indicated that the candidate proteins EEF2, U2AF2 and FLNC mainly
   involved in the diseases of “malignant neoplasms”, “carcinogenesis” and
   “neoplasia”.

3. Discussion

   In this study, we aimed to elucidate the biological activity of
   HNTX-III and JZTX-I toxins on migration, invasion and proliferation of
   MAT-LyLu cell lines in prostate cancer. Tumor metastasis and invasion
   are multistep and extremely complex biological processes including the
   detachment, angiogenesis, colonization and proliferation which could be
   regulated by various signaling pathways [[104]29]. To intuitively
   understand the effects of these two toxins, we used the iTRAQ method to
   conduct quantitative phosphorylated proteomics analysis to compare the
   changes of protein expression level between case (HNTX-III and JZTX-I
   treated MAT-LyLu cells) and control samples. Quantitative
   phosphoproteomics integrated with multiple bioinformatics analysis
   revealed a series of phosphopeptides and candidate proteins such as
   EEF2, U2AF2 and FLNC involved in tumor metastasis and invasion. It can
   be clearly concluded that the three differentially phosphorylated
   proteins, EEF2, U2AF2 and FLNC, were down-regulated in experimental
   group I and up-regulated in experimental group II. They play
   significant roles in cancer-related pathways including “cell adhesion”,
   “spliceosome” and “cadherin binding involved in cell-cell adhesion” by
   forming synergistically functional regulatory networks. The results may
   largely help us to uncover the potential molecular mechanisms of
   HNTX-III and JZTX-I in regulating tumor metastasis and invasion.

   In most tumors, the signaling pathways driven by protein kinases are
   significantly altered. By motif analysis of peptide motifs, the first
   four motifs in the 12 serine motifs were psDXE, RSXps, psPXR, and RRXps
   according to their motif scores, and their corresponding kinases were
   CK2, AKT kinase, CDKS, and PKA [[105]21]. Among them, P21 activated
   kinase, a serine/threonine kinase family, is involved in a variety of
   tumor-related signaling pathways as a downstream node, including
   regulating cytoskeleton remodeling and cell movement, affecting cell
   proliferation and regulating apoptosis [[106]30]. The basis of the
   structural homology of PAK subtype can be divided into two groups: the
   first group PAKs including PAK1–3 and the second group PAKs including
   PAK4–6 [[107]31]. A large number of literature references have shown