Identification of calgranulin B interacting proteins and network analysis in gastrointestinal cancer cells

Calgranulin B is known to be involved in tumor development, but the underlying molecular mechanism is not clear. To gain insight into possible roles of calgranulin B, we screened for calgranulin B-interacting molecules in the SNU-484 gastric cancer and the SNU-81 colon cancer cells. Calgranulin B-interacting partners were identified by yeast two-hybrid and functional information was obtained by computational analysis. Most of the calgranulin B-interacting partners were involved in metabolic and cellular processes, and found to have molecular function of binding and catalytic activities. Interestingly, 46 molecules in the network of the calgranulin B-interacting proteins are known to be associated with cancer and FKBP2 was found to interact with calgranulin B in both SNU-484 and SNU-81 cells. Polyubiquitin-C encoded by UBC, which exhibited an interaction with calgranulin B, has been associated with various molecules of the extracellular space and plasma membrane identified in our screening, including Na-K-Cl cotransporter 1 and dystonin in SNU-484 cells, and ATPase subunit beta-1 in SNU-81 cells. Our data provide novel insight into the roles of calgranulin B of gastrointestinal cancer cells, and offer new clues suggesting calgranulin B acts as an effector molecule through which the cell can communicate with the tumor microenvironment via polyubiquitin-C.


Introduction
Calprotectin is heterotetrameric calgranulin A and B complex that were noncovalently bonded without a peptide bridge between two subunits. The amounts of calprotectin in blood or extracellular body fluids are reportedly increased under many pathological conditions, such as rheumatoid arthritis, inflammatory bowel diseases, viral infection, microbial infection, tumors, and many inflammatory conditions [1]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Various functions of calprotectin have been reported, such as stimulation of fibroblast growth and beta 2-integrin-mediated neutrophil adhesion, neutrophil chemoattraction, and macrophage deactivation [2][3][4][5]. Calprotectin is also believed to function in altering the cytoskeleton and cell shape, transducing signals, and modulating intracellular calcium.
One of the two units of calprotectin, calgranulin B, is a small calcium-binding protein that is mainly found in granulocytes, monocytes, and activated keratinocytes [6][7][8][9][10]. It has also emerged as a marker for non-inflammatory pathological conditions, such as tumor development. Calgranulin B is reportedly overexpressed in various tumor types, including ovarian cancer, head and neck tumors, pulmonary carcinoma, and prostate cancer [11]. In addition, it is secreted by intestinal monocytes and epithelial cells, and elevated levels of calgranulin B have been detected in stool samples from colorectal cancer patients. We previously reported that calgranulin B is a candidate fecal marker for the diagnosis of colorectal cancer [12], and more recently showed that combining the fecal occult blood test (the established means of colorectal cancer screening) with calgranulin B screening can increase the sensitivity of colorectal cancer detection [13]. However, the intracellular molecular mechanism underlying the involvement of calgranulin B in tumor development is unknown.
Here, we set out to investigate the role of calgranulin B in gastrointestinal cancer by identifying calgranulin B-interacting partners in cancer cell lines.

Materials and methods
Human cell lines SNU-81 colorectal carcinoma cells, SNU-484 gastric carcinoma cells, and HEK293 human embryonic kidney cells were obtained from the Korean Cell Line Bank (KCLB, Seoul, Korea).

Yeast two hybrid (Y2H)
The full-length cDNA of human calgranulin B was PCR amplified and cloned into the pGBKT7 vector (containing the GAL4 DNA-binding domain). The pGBKT7-calgranulin B construct did not show any autonomous transcriptional activation or cytotoxicity following transformation into the yeast strain, Y2H Gold.
SNU-484, SNU-81, and HEK293 cells were used to construct cDNA libraries in the pGAD-T7-Rec vector (containing a GAL4 activation domain) using Matchmaker Library Construction and Screening kits (Clontech, Santa Clara, CA, USA). Each library was then transformed into the Y187 yeast strain, and Y2H screening was performed using the Matchmaker Two-Hybrid system (Clontech). Positive clones were selected based on their ability to grow on synthetic dropout (SD) medium/-LTH/X-α-Gal (TDO). Their cDNA inserts were PCR amplified, sequenced, and subjected to BLAST alignment. Interaction between the bait and identified prey clones was verified by co-transforming the purified prey plasmid plus the bait pGBKT7calgranulin B construct into Y2H Gold cells, and then selecting clones on SD/-LTHA/ X-α-Gal medium. Co-transformation of pGBKT7-p53 plus pGADT7-SV40 was used as a positive control, while co-transformation of pGBKT7-p53 with empty pGADT7 vector was used as a negative control.

Gene ontology and top disease information analysis
Computational analysis was applied to all of the molecules identified as interacting with calgranulin B in the three cell lines. Gene ontology (GO) analysis of the relevant biological processes, cellular components, and molecular functions was performed using the Protein Analysis Through Evolutionary Relationships program (PANTHER, www.pantherdb.org), which refers to a curated database of protein families, functions and pathways [14,15]. GO terms assigned into identified molecules were classified according to their function.
The most highly represented diseases and disorders information was obtained from Ingenuity Pathway Analysis (IPA, www.quiagen.com/ingenuity), which determines functions of identified molecules from literature-based information.

Protein class and pathway analysis
Furthermore, the protein class levels and pathways of all identified molecules were assessed using the PANTHER pathway program, which refers to a database containing: 165 expertly curated metabolic and signaling pathways; 20,851 proteins directly associated with these pathways, along with an evidence code; and 3569 distinct literature references (http://pid.nci.nih. gov/2009/090414/full/pid.2009.1.shtml).

Molecular interaction and network analysis
IPA, which determines the interactions and functions of identified gene products based on the information available in the literature, was used to map the identified molecules with respect to their interconnections. IPA scans input molecules to identify networks using the Ingenuity Pathways Knowledge Base (IPKB) for interaction between identified molecules. Molecular interactions are shown by grey lines, the identified calgranulin B-interacting molecules are highlighted in purple, and molecules that have been associated with cancer are indicated by dotted blue lines.
STRING analysis (https://www.string-db.org) was used to find direct and indirect interactions between all identified molecules interacting with calgranulin B in the extracellular space and plasma membrane. STRING is a database which provides direct and indirect associations from computational prediction, and interactions from other databases for interaction analysis [16,17]. STRING score is calculated from the combination of all predictions and network analysis was set at a medium confidence. Eight different colored lines were used to represent the types of evidence for associations, as follows: green, neighborhood evidence; red, gene fusion; blue, co-occurrence; black, co-expression; purple, experimental; light blue, database; yellow, text mining; sky blue, protein homology.

Results
Calgranulin B-interacting molecules were successfully identified and various computational analyses draw functional information, which could provide the biological role of calgranulin B in the context of a complex molecular interaction network. The whole scheme of the experimental procedures was represented in Fig 1. Identification of calgranulin B-interacting partners using a yeast twohybrid system cDNA libraries were derived from the SNU-484 human gastric cancer cell line, the SNU-81 human colon cancer cell line, and the HEK293 normal human kidney cell line, and calgranulin B-interacting partners were screened using the bait plasmid, pGBKT7-calgranulin B. Interactions of bait and prey proteins were examined by assessing growth on two selective media with different levels of restrictiveness: TDO (-His/X-a-Gal/AbA/) and QDO (-His/-Ade/X-a-Gal/ AbA) (Fig 2). Screening of 1.55x10 7 and 1.87x10 7 primary transformants obtained from SNU-484 and SNU-81 cells, respectively, yielded 25 and 27 positive clones, respectively (Fig 2A and  2B, upper). These clones were isolated, sequenced and aligned using the NCBI BLAST alignment search tool. We found that 14 and 13 of the clones, respectively, were in-frame; these were selected as potential candidate calgranulin B-interacting partners in gastric and colon cancers, respectively (Fig 2A and 2B, lower). We also performed two screenings in HEK293 cells, and obtained 1.76x10 7 and 1.41x10 7 primary transformants that yielded 10 and 34 positive clones, respectively. Four and 15 of the positive clones were found to be in-frame (Fig 2C), and were subjected to further analysis to reveal the biological meaning and relationship with diseases in the context of an interaction with calgranulin B. All of the selected candidate calgranulin B-interacting genes identified in the three cell lines by Y2H assay system are listed in Table 1.

Gene ontology and top diseases information
PANTHER was used to identify the GO terms associated with the molecular function, biological activity, and cellular component(s) associated with each candidate calgranulin B-interacting protein. The results of this analysis (Figs 3-5) revealed that calgranulin B appears to be associated with slightly different molecular functions, biological processes and cellular components in SNU-484, SNU-81, and HEK293 cells. However, in all three lines, binding (GO:0005488) and catalytic activity (GO:0003824) were the most highly represented molecular functions, with representations of 31.3% and 33.3% in SNU-484 and HEK293 cells, respectively and 40% for both cases in SNU-81 cells. The next most common functions were structural molecule activity (GO:0005198) in SNU-484 cells, enzyme regulation (GO:0030234) in HEK293 cells,     and structural molecule activity (GO:0005198) and transportation (GO:00052158) in SNU-81 cells (Fig 3). The calgranulin B-interacting molecules were also classified according to their related biological activities/processes (i.e., recognized series of events or molecular functions; www. geneontology.org). In all three cell lines, the most highly represented biological process was metabolic process (GO:0008152), followed by cellular processes (GO:0009987) in SNU-484 and HEK293 cells and immune processes (GO:0002376) in SNU-81 cells (Fig 4).
The cellular component portion of the GO analysis was used to classify the candidate proteins by their known locations at the levels of subcellular structure and macromolecular complexes (www.geneontology.org). Our results revealed that the calgranulin B-interacting molecules were largely localized in the cell part (GO:0044464) and organelle (GO:0043226) components of SNU-484 and SNU-81 cells, while they were distributed to the cell part (GO: 0044464), macromolecular complex (GO: 0032991) and organelle (GO: 0043226) components of HEK293 cells (Fig 5).
Ingenuity Pathway Analysis (IPA) was used to identify the most highly represented diseases and disorders among the calgranulin B-interacting proteins of each cell line. Cancer was the most highly represented disease/disorder in SNU-81 and HEK293 cells, whereas cardiovascular disease was the most highly represented disease/disorder in SNU-484 cells. The full results of this analysis are presented in Table 2.

Protein classes and pathways related to the identified calgranulin Binteracting partners
The protein class distribution in each cell line was evaluated by PANTHER analysis (Fig 6). Our results revealed that nucleic acid binding (PC00171) was the most highly represented protein class in all three cell lines, followed by: transporter (PC00227), oxidoreductase (PC00176), and transfer/carrier (PC00219) functions in SNU-484 cells ( Fig 6A); chaperone (PC00072) and isomerase (PC00135) functions in SNU-81 cells (Fig 6B); and hydrolase (PC00121) and signaling (PC00207) functions in HEK293 cells (Fig 6C).

Molecular network analysis and cancer-related calgranulin B-interacting proteins
To obtain a more comprehensive view of the molecular network related to the calgranulin Binteracting molecules identified herein, we used IPA to analyze their direct and indirect molecular interactions. The results are shown in Fig 8, where molecular interactions are indicated with grey lines, the identified calgranulin B-interacting molecules are highlighted in purple, the relevant cell lines are indicated in red circles, and the 46 calgranulin B-interacting molecules that have been associated with cancer are indicated with dotted blue lines. All candidate molecules interacting with calgranulin B were unique in three cell lines except FKBP2 that was found in two cancer cell lines, SNU-484 and SNU-81. However molecular network analysis revealed they had direct and/or indirect interaction across 3 cell lines and most of them were involved in cancer disease development. Calgranulin B-interacting proteins associated with the extracellular space and plasma membrane Among the identified calgranulin B-interacting proteins, those known to be localized to the extracellular space and plasma membrane were further characterized by network analysis to reveal interactive relationship with calgranulin B. As shown in Fig 9, calgranulin B (S100A9) was found to directly interact with polyubiquitin-C (encoded by UBC), which is known to directly interact with various other calgranulin B-interacting partners, including dystonin (encoded by DST) in SNU-484 cells, ATPase subunit beta-1 (ATP1B1, encoded by ATP1B1) in SNU-81 cells, and nodal modulator 1 (encoded by NOMO1) in HEK293 cells. Direct interactions with calgranulin B (S100A9) were predicted using STRING analysis, and are indicated by different colored lines: green, red, blue, black, purple, light blue, yellow, sky blue based on the types of evidence for associations. doi:10.1371/journal.pone.0171232.g009

Discussion
Yeast two-hybrid screening analysis is a molecular biology technique that enables researchers to uncover protein-protein and protein-DNA interactions by testing for physical interactions between relevant fragments. Y2H screening has been steadily used to identify new interacting partners for a protein of interest, or they may confirm binary protein-protein interactions, which play crucial roles in almost all biological processes [18]. Here, we used Y2H analysis to identify candidate calgranulin B-interacting molecules (Fig 2), and subjected these proteins to PANTHER analysis of: their predicted activities, which emphasized binding and catalytic activity (Figs 3 and 6); their cellular involvements, which emphasized metabolic and cellular processes (Fig 4); and their cellular localizations (Fig 5). Molecular function, biological activity and cellular components of identified molecules showed slight differences among three cell lines but most terms were common in all three cell lines tested. To determine further biological characteristics of calgranulin B-interacting proteins, previously reported calgranulin B-interacting molecules on the BioGRID database (http://thebiogrid.org) were analyzed by PAN-THER and found to have correlation with the present data. For example, the most highly represented molecular function; cellular component, biological process, and protein class of the previously reported 92 molecules are same with our present study as binding, cell part, metabolic process, and nucleic acid binding, respectively (data not shown). However, signaling pathways involved calgranulin B interacting partners were totally different in each cell line (Fig 7) providing an involvement of unique signaling cascades. Ubiquitin proteasome pathway represented in SNU-81 colon cell is very interesting considering internalization pathway of calgranulin B in colon cancer cells. Ubiquitination, one of the post-translational modifications of proteins can lead to targeting for degradation by the proteasome machinery as well as can serve as a sorting signal for regulation of internalization such as activation of receptor endocytosis [19,20]. In addition, ubiquitination of integral plasma membrane proteins triggers their rapid internalization into endocytic pathway [21]. Recently we found that calgranulin B released from immune cells can be internalized specifically into colon cancer cells and suppressed cell proliferation [22] but the molecular mechanism of internalization pathway was not clear. Ubiquitin proteasome pathway proposed in this study could provide a possible mechanism of calgranulin B internalization in colon cancer cells.
Among the calgranulin B-interacting molecules known to be involved in cancer (Fig 8, purple), FKBP2 (FK506-binding protein 2) was identified from both SNU-484 and SNU-81 cells. FKBP2, which is a member of the immunophilin protein family, contributes to immunoregulation and is thought to function as an ER chaperone and a component of membrane cytoskeletal scaffolding [23]. The FKBP2 gene has been associated with pathological osteoporosis [24], and its protein product was suggested to be associated with the development of type 2 diabetes [25]. Northern blotting revealed that FKBP2 is expressed in tissues that are predisposed to hyperplasia in multiple endocrine neoplasia type 1 (MEN1) patients; however, mutation analysis of MEN1 kindreds and sporadic tumors excluded FKBP2 as a candidate gene for MEN1 [26]. No previous report has suggested any molecular cooperation between FKBP2 and calgranulin B. Here, we show for the first time that calgranulin B may have FKBP2-mediated function in gastrointestinal cancer cells.
NKCC1 is a membrane-transport protein that governs the active bi-directional transport of sodium, potassium, and chloride [40]. NKCC1 expression has been shown to predict poor prognosis in lung adenocarcinoma [41] and affect the G2/M checkpoint in esophageal squamous cell carcinoma [42]. In glioma cells, inhibition of NKCC1 was found to reduce invasion [43,44] and augment temozolomide-induced apoptosis [45]. Blockade of NKCC1 function also diminished the proliferation of poorly differentiated human gastric cancer cells by affecting G0/G1 phase [42]. It is not yet known whether the interaction between calgranulin B and NKCC1 blocks or augments the function of NKCC1, but studies showing that calgranulin B and calprotectin exert pro-apoptotic effects [46][47][48] may suggest that calgranulin B inhibits the activity of NKCC1 in gastric cancer cells.
Dystonin, which is a neural isoform of bullous pemphigoid antigen 1, has an N-terminal actin-binding domain and is essential for maintaining the cytoskeletal integrity of neurons [49]. No previous study has linked dystonin to gastric cancer. Further work is needed to clarify the potential for calgranulin B to communicate with the extracellular environment via dystonin.
Among the identified proteins, ATPase subunit beta-1 is unique in that it may interact with calgranulin B to facilitate colon cancer progression (Fig 8) and/or communicate directly with polyubiquitin-C (Fig 9). ATPase is a well-known membrane protein that is responsible for maintaining the electrochemical gradients of Na and K ions across the plasma membrane. It can also be a target of TGF-beta1 during TGF-beta1-induced epithelial-to-mesenchymal transition (EMT) [50], and participate in the drug sensitization of cancer cells [51]. In contrast to the interaction between calgranulin B and NKCC1 we do not expect to see a calgranulin Bmediated change in ATPase activity. However, our overall data demonstrate that ATPase subunit beta-1 is a most promising candidate for future efforts to study the potential function of calgranulin B in colon cancer cells.
In sum, we herein screened for calgranulin B-interacting partners and found that they included both cytosolic and plasma membrane proteins. The molecular linkages identified herein suggest that calgranulin B may have multiple anti-tumor functions acting against the progression of gastrointestinal cancer. Our novel findings suggest that further functional studies are warranted to investigate the signaling downstream of the encounters between calgranulin B and its interacting partners.