Monocyte Chemotactic Protein-induced Protein 1 and 4 Form a Complex but Act Independently in Regulation of Interleukin-6 mRNA Degradation

Results: chemotactic form they in of

Immunofluorescence staining showed that MCPIP4 was co-localized with MCPIP1 in GW-body, which is featured by GW182 and Argonaute 2. Further studies showed that MCPIP1 and MCPIP4 act independently in regulation of IL-6 mRNA degradation. These results suggest that MCPIP1 and MCPIP4 may additively contribute to control IL-6 production in vivo. MCPIP1 (also known as ZC3H12A and Regnase-1) is a prototype member of a novel protein family which includes MCPIP1, MCPIP2, MCPIP3 and MCPIP4 (1,2). The unique feature of MCPIP1 protein family is characterized by a CCCH-zinc finger motif located in the middle region of these proteins (1,2). We have identified the whole CCCH-zinc finger containing protein family in both human and mouse genomes (3). There are approximately 60 CCCH-zinc finger proteins in both human and mouse genomes. Most CCCH-type zinc finger proteins are involved in RNA metabolism pathways such as splicing, polyadenylation, and mRNA decay (3). For example, tristetraprolin, Roquin and ZAP are well-studied cytokines such as IL-6 and IL-12 (2,10). Through this central mechanism, MCPIP1 serves as an essential regulator in inflammatory cell activation and immune homeostasis (2). MCPIP1 knockout mice developed spontaneous inflammatory diseases accompanied by splenomegaly, lymphadenopathy and multi-organ inflammation especially in the lungs (2,11,12). T cell-specific deletion of MCPIP1 produces pathogenic T cells with hyperactivated phenotypes as well as autoimmune diseases (13).
MCPIP1 is a multi-domain containing protein that includes an ubiquitin association domain (UBA) at the N-terminus, a putative NYN-RNase domain, followed by a CCCH-zinc finger domain (ZF), and a prolinerich domain (PRD) at the C-terminus (11). A recent study compared the crystal structure of the putative NYN-RNase domain with other reported RNase proteins and suggested that MCPIP1 is a functional RNase (14). The mRNA targets of MCPIP1 nuclease are now expanding to c-Rel, IL-2, ICOS, Ox40, TNFR2, GATA3 and MCPIP1 self mRNA (13,15,16). MCPIP1 promotes their degradation by targeting their 3' untranslated region (3'UTR) (10,13). MCPIP1 specifically recognized a stem-loop structure on the 3' UTR of its substrate mRNAs (10).
MCPIP4 (also known as ZC3H12D, TFL and p34) is originally reported as a putative tumor suppressor that is deregulated in transformed follicular lymphoma (17)(18)(19). Similar to MCPIP1, MCPIP4 is also remarkably induced by Toll-like receptor activation in macrophages and overexpression of MCPIP4 also represses inflammatory activation of macrophages (20). The role of MCPIP4 in vivo seems overlapped but less important than MCPIP1. For example, MCPIP4 null mice showed pretty normal phenotypes under normal condition, but exhibited more activated lymphocytes upon stimulation (21).
In this study, we first found that MCPIP1 interacts with MCPIP4 to form a protein complex but they act independently in regulate IL-6 mRNA degradation, suggesting that MCPIP1 and MCPIP4 may additively contribute to control IL-6 production in vivo.

MATERIALS AND METHODS
Cells-HEK293, COS-7, HeLa and RAW264.7 cells were obtained from the American Type Culture Collection. These cells were grown as a monolayer in DMEM (Invitrogen) containing 10% FBS, 2 mM Lglutamine, with 100 U/mL penicillin and 100 µg/mL streptomycin in 5.0% CO2. HEK293-MCPIP1 stable cell line was established by lentiviral transduction of HEK293 with a GFP-MCPIP1 expressing construct and maintained in complete medium with 200 µg/mL of G418 and 0.25 µg/mL of puromycin.

Identification of MCPIP1-interacting proteins by Co-IP and Mass-Spec analysis-HEK293 cells were
transfected with empty pCMV-Flag vector or pCMV-Flag-MCPIP1 and incubated for 48 h. Transfected cells were lysed in a buffer containing 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton X-100, and protease inhibitor cocktail (Roche). Cell lysates were precleared with mouse immunoglobulin G (IgG) agarose and incubated with 1 μl of mouse anti-Flag M2 agarose beads (Sigma) on ice for 2 h. After extensive wash with the wash buffer containing 50 mM Tris-HCl (pH 7.4) and 150 mM NaCl, proteins bound to the beads were then eluted into 1× sodium dodecyl sulfate (SDS) running buffer by heating at 95 °C for 5 min. The proteins were separated on a 10% SDS-PAGE and stained by Sypro Ruby. Stained bands were excised out and proteins were identified by LTQ-orbitrap-velos mass spectrometer.
Confocal microscopic analysis-COS-7 cells seeded on glass coverslips were transfected with pEGFP-C1, pEGFP-MCPIP1 or pEGFP-MCPIP4 in a six-well plate using Lipofectamine 2000. After transfection for 24 h, the cells were washed twice with cold PBS, fixed with 4% paraformaldehyde (pH 7.4) in PBS for 20 min, permeabilized with 0.2% Triton X-100 in PBS for 15 min. and then stained with DAPI (Life technologies) to visualize the nuclear DNA. The cell images were recorded with a LEICA laser-scanning confocal microscope.

Co-immunoprecipitation-HEK293
Cells were transfected with Flag-MCPIP4 and HA-MCPIP1 by the calcium precipitation. After 24 h, the cells were lysed with cold CelLytic M lysis buffer (sigma) with protease inhibitors including 1 mM PMSF, 1 µg/ml aprotinin and 1 µg/ml leupeptin. The lysate supernatant was pre-cleared by incubating the cell lysates with protein A agarose beads (Invitrogen) for 60 minutes at 4°C with gentle agitation, and then incubated with 25 µl the ANTI-FLAG M2 or anti-HA agarose beads at 4 °C for 4 h with gentle mixing. The samples were extensively washed two times using wash buffer (50 mM Tris-HCl pH7.4, 150 mM NaCl and 0.05% Triton-X100) with protease inhibitors. The precipitates were treated with or without 100 unit/ml of RNase A (Sigma) at room temperature for 30 min. The agarose beads were washed three times with the wash buffer. The immunoprecipitates were eluted from the beads with 100 µl loading buffer, resolved by 12% SDS-PAGE and analyzed by immunoblotting with Flag or HA antibodies. Membranes were developed using an enhanced chemiluminescence (ECL) detection system (GE Healthcare).
Immunofluorescence-HEK293 and HeLa cells were transfected with the expression plasmids as indicated. After 24 h, the transfected cells were fixed with 4% paraformaldehyde for 20 min, permeabilized with 0.2% Triton X-100 in PBS for 20 min, and then incubated in BlockAid™ Blocking Solution (Molecuar probes) for 1 h at room temperature. Then the cells were incubated with primary antibody or control IgG diluted in PBS at 4 ℃ overnight. The cells were briefly rinsed with PBS, incubated with Alexa Fluor® 488 or Alexa Fluor® 594 fluorescent labeled secondary antibody (Life Technologies) for 1 h at room temperature. After intensive wash with PBS, the slides were mounted by cold Vectashield Hard Set mounting medium with DAPI, and visualized by a Nikon C1 plus confocal.
Mammalian two-hybrid assay-To analyze the interaction between MCPIP1 and MCPIP4 in mammalian cells, HEK293 cells were co-transfected with various pBIND fusion expression plasmids, pACT expression plasmids and the luciferase reporter pG5luc (Promega) in a 24-well plate using Lipofectamine 2000. 24 h after transfection, the transfected cells were harvested, washed twice with PBS and lysed with 1×Passive Lysis Buffer (Promega). Luciferase assay was performed using the Dual-Luciferase Reporter Assay system (Promega) according to the manufacturer's instructions. All the experiments were carried out in triplicate and luciferase value was determined using a GloMax® 96 Microplate Luminometer (Promega).
Transfection-Transient transfection into RAW264.7 cells was performed by electroporation following the manufacturer's instruction (Amaxa). Briefly, RAW264.7 cells were grown to confluence in DMEM medium supplemented with 10% FBS. Cells were collected and washed once with DMEM medium and re-suspended with the electroporation buffer (Amaxa). After electroporation, the cells were plated on 6-well plates and the transfection efficiency was monitored by fluorescent microscopy. Quantitative real-time PCR (QPCR)-After removing the genomic DNA using DNase I (Ambion), 2 μg of total RNA was reverse-transcribed to cDNA using a high-capacity cDNA reverse transcription kit (Life Technologies). QPCR was performed with StepOne Plus real-time PCR system (ABI) using SYBR Green master mix (ABI). Forty cycles were conducted as follows: 95 °C for 30 s, 60 °C for 30 s, preceded by 1 min at 95 °C for polymerase activation. Primer sequences for all genes we measured in this report are available upon request. Quantification was performed by the delta cycle time method, with β-actin used for normalization.

Protein isolation and Western blot-Protein isolation
and Western blot were essentially performed as described previously (1).
Statistics-Data were expressed as mean±SD. For comparison between two groups, the unpaired Student's test was used. For multiple comparisons, analysis of variance followed by unpaired Student's test was used. A value of p<0.05 was considered significant.

Identification of MCPIP4 as a MCPIP1-interacting
protein-Previously, we and others have demonstrated that MCPIP1 is crucial to control inflammation and immune homeostasis (10,11,(24)(25)(26). However, the molecular mechanisms need to be further elucidated. To identify the interacting proteins that may involve in MCPIP1-mediated repression of inflammation and immunity, we performed immunoprecipitation (IP) followed with Mass-Spec (MS) analysis. Flag-tagged MCPIP1 was transiently expressed in HEK293 cells, and MCPIP1-bound proteins were immunoprecipitated with anti-Flag M2 agarose beads. After extensive wash, eluted proteins were separated on a 10% SDS-PAGE and stained by Sypro Ruby. Stained bands were excised out and proteins were identified by LTQorbitrap-velos mass spectrometer. One protein band with an apparent molecular mass of 58 kDa was repeatedly identified in the IP assay compared to the lysate from control HEK293 cells transfected with pCMV-Flag vector ( Figure 1A). MS analysis identified this protein as MCPIP4 (ZC3H12D) ( Figure  1B). Other proteins identified in the assay will be described elsewhere.
As reported previously, MCPIP1 is a protein containing multiple domains (2,11). As shown in Figure 1C, the NYN-RNase domain (133-300) and CCCH-zinc finger (305-325) are located at the middle and both are critical for its RNase activity (2). A proline-rich domain (PRD) is located at its C-terminal without known function yet. MCPIP4 also has a similar RNase, CCCH-zinc finger and proline-rich domains ( Figure 1C).
To confirm the interaction of MCPIP1 with MCPIP4, HA-tagged MCPIP1 and Flag-tagged MCPIP4 were co-transfected into HEK293 cells, and Co-IP of cell lysates with either anti-Flag or anti-HA antibodies were performed. Both cell lysate input and immunoprecipitates were then analyzed by Western blotting using anti-Flag and anti-HA antibodies. As shown in Figure 1D, immunoprecipitation with antibodies targeting either MCPIP1 or MCPIP4 results in pull-down of both proteins, confirming their interaction. To exclude the possibility that the interaction of MCPIP1 with MCPIP4 is mediated by RNA, the immunoprecipitates were treated with or without RNase A, and then detected by Western blot with anti-Flag or anti-HA. As shown in Figure 1E, the interaction of MCPIP1 with MCPIP4 was not affected by RNase A treatment.
To further confirm the interaction of MCPIP1 with MCPIP4, we performed mammalian two-hybrid assay. First, we inserted the gene fragments encoding MCPIP1 and MCPIP4 into the vector pACT containing the herpes simplex virus VP16 activation domain or the pBIND vector containing the yeast Gal4 DNA-binding domain to generate fusion proteins. Then, we transiently co-transfected these vectors into HEK293 cells with a luciferase reporter pG5luc. As positive control, pBIND-ID and pACT-MyoD were co-transfected with the reporter into HEK293 cells, the reporter activity were increased by ~225 folds than the negative control. As shown in Figure 1F, either pBIND-MCPIP1 co-transfected with pACT-MCPIP4 or pBIND-MCPIP4 co-transfected with pACT-MCPIP1 resulted in luciferase activity increase by more than 20 folds. Taken together, these data confirm the MS result that MCPIP1 interacts with MCPIP4 in mammalian cells.

MCPIP1 and MCPIP4 are co-localized in GW-
body-To examine if MCPIP1 and MCPIP4 are colocalized in cells, GFP-MCPIP1 or GFP-MCPIP4 was transiently transfected into COS-7 cells and the protein localization was visualized by a fluorescence microscopy. The results showed both MCPIP1 and MCPIP4 formed granule-like structure in cytoplasm ( Figure 2A). To further examine if endogenous MCPIP1 and MCPIP4 also form granule-like structure, Raw264.7 cells were stimulated with PamCSK4 for 6 h to induce the expression of MCPIP1 and MCPIP4. The cells were then stained with the primary anti-MCPIP1 (Genetex) or MCPIP4 (kindly provided by Dr. Matsui) or control IgG, and then stained with fluorescence-labeled anti-IgG. As shown in Figure 2B, both endogenous MCPIP1 and MCPIP4 also formed granule-like structure in cytoplasm. To determine the specificity of anti-MCPIP1, Raw264.7 cells were transfected with small interference RNA for control (si-Control) or MCPIP1 (si-MCPIP1). 24 h later, the transfected cells were stimulated with PamCSK4 for 6 h. MCPIP1 expression in the cells was determined by Western blot with anti-MCPIP1 (Genetex). As shown in Figure 2C, MCPIP1 protein was detected as a single band around 72 kd (migrated behind theoretical protein size). The specificity of anti-MCPIP4 monoclonal antibody was determined previously (20,21). To determine if MCPIP1 and MCPIP4 are colocalized in cells, HeLa cells were co-transfected with Flag-MCPIP1 and EGFP-MCPIP4 and visualized by staining with anti-Flag and Alexa Flour596-labeled second antibody. As shown in Figure 2D, the MCPIP1-granules and MCPIP4-granules were nicely overlapped in cytoplasm. To characterize the identity of MCPIP1-granules, we have previously performed immunostaining with the antibodies against different organelles and granule markers. MCPIP1-granules were not overlapped with mitochondria, Golgi network, lysosome, endosome and stress granule etc. (27). However, as shown in Figure 2E, MCPIP1granules were partly overlapped with GW182 and Argonaute 2 (Ago2), which are the markers of GW body (28). Interestingly, MCPIP4 was also overlapped with GW182 ( Figure 2E). Next, we determined if MCPIP1 can interact to GW182 or Ago2 using mammalian two-hybrid assay. As shown in Figure 2F, when pBIND-MCPIP1 and pACT-GW182-C were cotransfected with the reporter pG5luc, the luciferase activity was increased by 55 folds compared with the control group, suggesting that MCPIP1 is associated with the C-terminal of GW182. Same strategy was used to determine if MCPIP1 is also associated with Ago2 and showed that MCPIP1 did not associate with Ago2 (data not shown).

Mapping the interacting domains of MCPIP1 and MCPIP4-After establishing that full-length MCPIP1
and full-length MCPIP4 associate in vivo, we asked, which domains could be responsible for the observed interactions. For that reason, we performed mammalian two-hybrid assay. First, we inserted the gene fragments encoding the serial deletions of MCPIP1 and MCPIP4 into the vector pACT or the pBIND ( Figure 3A). Then, we transiently cotransfected these vectors into HEK293 cells with the reporter pG5luc. As shown in Figure 3B, 1-356 and 259-527 of MCPIP4 have similar binding ability compared with full-length MCPIP4, However, 1-258 and 357-527 of MCPIP4 lost the binding ability with MCPIP1. These results suggest that the region 259-356 of MCPIP4 is required for interaction with MCPIP1. On the other hand, 1-457 of MCPIP1 has similar binding ability with MCPIP4 compared with full-length MCPIP1, whereas 1-300 of MCPIP1 lost the binding ability with MCPIP4. These results suggest that the region 301-457 of MCPIP1 is crucial for the interaction with MCPIP4. To further examine if 259-356 of MCPIP4 or 301-457 of MCPIP1 is sufficient for their interaction, we performed similar transfection experiments in HEK293 cells. As shown in Figure 3C, both 259-356 of MCPIP4 and 301-457 of MCPIP1 were sufficient for their interaction. Moreover, further removing their CCCH-zinc finger did not affect their interaction ( Figure 3C).

The interaction domain of MCPIP1 and MCPIP4 is required, but not sufficient for their granule-like structure formation-To determine if the interaction
domain of MCPIP1 and MCPIP4 is required for their granule-like structure formation, we mapped the domains responsible for the granule-like structure. First, we inserted the gene fragments encoding serial deletions of MCPIP4 and MCPIP1 into pEGFP-C1 vector. These vectors were transfected into HEK293 cells and the correct expression of the gene fragments were determined by Western blot analysis ( Figure 4A). Then we transfected the vectors into HeLa cells and the protein localization was visualized by fluorescence microscopy. As shown in Figure 4B, the region 300-536 of MCPIP1 and the region 259-457 of MCPIP4 are responsible for granule-like structure formation. These regions include their CCCH-zinc finger motif, interaction domain and proline-rich domain, which may contribute to their granule-like structure formation.

Co-expression of MCPIP1 and MCPIP4 enhances the repression on the reporter of IL-6 3'UTR-It was
reported that MCPIP1 is an RNase that destabilizes a set of mRNAs, including IL-6 and IL-12, through cleavage of their 3' UTRs (2, 10). The role of MCPIP4 remains unknown. The results described above suggest that MCPIP4 may functionally associate with MCPIP1. To test this idea, we first transfected the expression plasmids encoding MCPIP1/2/3/4 with the reporter of IL-6 3'UTR into HEK293 cells respectively. The results showed that MCPIP1 is the most potent member to repress IL-6 3'UTR. MCPIP4 has moderate effect on the reporter, whereas MCPIP2 and MCPIP3 have subtle effect on this reporter ( Figure  5A). Next, we co-transfected MCPIP1 or/and MCPIP4 with the reporter of IL-6 3'UTR into HEK293 cells. As shown in Figure 5B, overexpression of MCPIP1 or/and MCPIP4 had no effect on the pGL3-control reporter without IL-6 3'UTR. However, both MCPIP1 and MCPIP4 repressed the reporter activity of IL-6 3'UTR. Co-expression of MCPIP1 and MCPIP4 enhanced the repression on the reporter of IL-6 3'UTR ( Figure 5B). In another experiment, MCPIP1-inducible HEK293 stable cells were co-transfected with MCPIP4 and the reporter of IL-6 3'UTR or control. 24 hours post-transfection, the cells were induced with 10 ng/ml of doxycycline for 24 hours to induce MCPIP1 expression. As shown in Figure 5D, overexpression of MCPIP1 or MCPIP4 alone significantly repressed the reporter activity of IL-6 3'UTR, but did not affect the reporter activity of the pGL3-control. Co-expression of MCPIP1 and MCPIP4 had enhanced effect on the IL-6 3'UTR. The inducible expression of MCPIP1 in the cell line upon Dox treatment was determined by western blot with anti-GFP antibody ( Figure 5C).

MCPIP1 and MCPIP4 additively contribute to control the IL-6 mRNA levels in activated
macrophages-To further compare the functional importance of MCPIP1 and MCPIP4 in inflammatory cytokine production, we purchased the plasmids encoding short hairpin RNA (shRNA) targeting to different positions of mouse MCPIP1 and MCPIP4 mRNA from Sigma. To test their efficacy to knockdown the expression of MCPIP1 and MCPIP4 in activated macrophages, we transfected these plasmids into RAW264.7 cells, a murine macrophage cell line, by electroporation (Amaxa). After 48 h, the transfected cells were treated with 20 ng/ml of Pam3CSK4 (an agonist of Toll-like receptor 2) for 6 h to induce the expression of both MCPIP1 and MCPIP4 (20,29). The mRNA levels of MCPIP1 and MCPIP4 were examined by QPCR. As shown in Figure 6A, the shRNA#1 for MCPIP1 was more efficiently to knockdown MCPIP1 expression and the shRNA#1 for MCPIP4 was more efficiently to knockdown MCPIP4 expression in activated macrophages. The efficiency of sh-MCPIP1#1 and sh-MCPIP4#1 was further examined by Western blot. As shown in Figure 6B, the protein levels of MCPIP1 and MCPIP4 were efficiently decreased by sh-MCPIP1#1 and sh-MCPIP4#1 respectively. Next, we transfected sh-Control, sh-MCPIP1#1, sh-MCPIP4#1 or sh-MCPIP1#1 plus sh-MCPIP4#1 into RAW264.7 cells and the transfected cells were then treated with Pam3CSK4 (20 ng/ml) for 6 h. The mRNA level of IL-6 was examined by QPCR. As shown in Figure 6C, knocking down of either MCPIP1 or MCPIP4 alone increased the level of IL-6 mRNA, whereas knocking down both MCPIP1 and MCPIP4 showed more enhanced effect on the IL-6 mRNA in activated macrophages. To exclude any offtarget effects, we also transfected another pairs of shRNAs into RAW264.7 cells and showed similar effects ( Figure 6C, right). The knocking-down efficiency of these shRNAs in the experiments described above was also confirmed by QPCR (data not shown). In contrast, overexpression of either MCPIP1 or MCPIP4 alone decreased IL-6 mRNA level. Co-expression of MCPIP1 and MCPIP4 enhanced the repression on IL-6 mRNA level ( Figure  6D). The mutants of MCPIP1(D141N) and MCPIP4(D94N) failed to repress IL-6 mRNA expression.
Neither overexpression of MCPIP1(D141N) with wild-type of MCPIP4 nor overexpression MCPIP4(D94N) with wild-type of MCPIP1 further enhanced their effects on IL-6 mRNA ( Figure D). These results suggest that the interaction of MCPIP1 and MCPIP4 is not required for their action in the regulation of IL-6 mRNA degradation. MCPIP1 and MCPIP4 may act independently in the regulation of IL-6 mRNA level.

Mapping the functional domains of MCPIP1 and MCPIP4 in regulation of IL-6 3'UTR-To further
understand how MCPIP1 and MCPIP4 regulate IL-6 3'UTR, we co-transfected the vectors containing serial deletions of MCPIP1 or MCPIP4 with the reporter of IL-6 3'UTR into HEK293 cells. As shown in Figure  7A, the region 81-457 of MCPIP1, containing RNase domain and MCPIP4-interaction domain, is critical in the regulation of IL-6 3'UTR. In addition, the region of 1-356 of MCPIP4, containing both RNase domain and MCPIP1-interaction domain, is required for suppressing the reporter of IL-6 3'UTR. To further confirm that their RNase domain is critical in repression of IL-6 3'UTR, we transfected the point mutations of both MCPIP1(D141N) and MCPIP4(D94N), which was previously demonstrated to be the active site for its RNase activity (2). The results showed that these point mutations also diminished their repressing activity on IL-6 3'UTR. Taken together, these results suggest that the RNase activity of MCPIP1 and MCPIP4 is required for the regulation of mRNA destabilization.

DISCUSSION
MCPIP1 is a newly identified CCCH-zinc finger containing protein, which belongs to a subfamily including MCPIP1, 2, 3 and 4. Though it is recently demonstrated that MCPIP1 is a critical factor that controls inflammatory cell activation, immune homeostasis and viral infection (30-34), the function of other three members remains largely unknown. In addition, emerging evidence suggest that MCPIP1 is an endonuclease and selectively destabilizes mRNAs that encode certain inflammatory cytokines and immune molecules (24). However, the detailed molecular mechanisms by which MCPIP1 promotes mRNA degradation are obscure. To understand the molecular mechanisms, we searched MCPIP1interacting proteins by a Co-IP combined with proteomic analysis. Interestingly, MCPIP4 was copurified with MCPIP1 in this approach. Further studies demonstrated that MCPIP4 is a MCPIP1-interacting protein, which may work together with MCPIP1 to control inflammatory cytokine production.
Both MCPIP1 and MCPIP4 contain a single CCCH-type zinc finger domain at the middle region, a highly conserved NYN-type RNase domain at the Nterminus and a proline-rich domain at the C-terminus ( Figure 1C). Many studies have shown that MCPIP1 exerts the most potent RNase activity on certain mRNAs and pre-miRNAs among the MCPIP family (22,33). In the gene knockout mouse models, MCPIP1 deficient mice developed severe and complex inflammatory phenotypes including splenomegaly, lymphadenopathy, multi-organ inflammatory cell infiltration and premature death (2,11). However, MCPIP4-deficient mice did not show gross developmental defects (21). On the other hand, emerging evidence suggests that MCPIP1 and MCPIP4 have some similar effects on the regulation of IL-6 mRNA degradation and endothelial inflammation (21,35,36). Taken together, these studies suggest that both MCPIP1 and MCPIP4 may involve in the regulation of inflammatory cytokine production. In our present study, we have evidence to suggest that MCPIP1 interacts with MCPIP4 to form a protein complex, but they may act independently in the regulation of IL-6 mRNA degradation. Based current evidence, MCPIP1 and MCPIP4 may additively contribute to control IL-6 production in vivo.
We and others have previously observed that overexpressed exogenous or endogenous MCPIP1 protein formed granule-like structure in the cytoplasm (22,27). In our previous work, we have observed that MCPIP1-granules were not overlapped with mitochondria, Golgi network, lysosome, endosome and stress granules etc. (27). The exact identity of MCPIP1-granules is still obscure. In the present study, we have found that both MCPIP1 and MCPIP4 formed similar granule-like structures in cytoplasm. Moreover, both MCPIP1 and MCPIP4 granules are partly overlapped with GW182 and Ago2, which are typical markers for GW-body (28) and the major components of miRNA-induced silencing complex (miRISC) (37). Further studies suggest that MCPIP1 is associated with the C-terminal of GW182, but not associated with Ago2. As it was known that Ago2 interacts with the Nterminal of GW182 (38), we propose that GW182 may act as an adaptor protein to recruit both Ago2 and MCPIP1 into GW-body through its N-terminal and Cterminal respectively. Nevertheless, the relationship between MCPIP1 and miRISC need to be further investigated both in vitro and in vivo.
In summary, we here reported that MCPIP4 is a MCPIP1-interacting protein and both MCPIP1 and MCPIP4 may be important in controlling IL-6 production. However, the interaction of MCPIP1 and MCPIP4 is not required for their regulation of IL-6 mRNA degradation. MCPIP1 and MCPIP4 act independently and additively contribute to controlling IL-6 production in activated macrophages.

CONFLICT OF INTEREST
The authors declare that they have no conflicts of interest with the contents of this article.

AUTHOR CONTRIBUTIONS
MF conceived and coordinated the study and wrote the paper. SH designed, performed and analyzed the experiments shown in Figure 1 Ruby; B) The stained bands were excised out and analyzed by LTQ-orbitrap-velos mass-spectrometer. Two fragments targeted on MCPIP4 amino acid sequence were indicated. C) Scheme of MCPIP1 and MCPIP4 protein domains. D) HA-tagged MCPIP1 and Flag-tagged MCPIP4 expression vectors were co-transfected into HEK293 cells. Cell lysates were prepared and incubated with either anti-HA or anti-Flag beads to precipitate MCPIP1 and MCPIP4, respectively. The cell lysates and immunoprecipitates were subjected to Western blot analysis with anti-HA or anti-Flag antibodies. E) HA-tagged MCPIP1 and Flag-tagged MCPIP4 expression vectors were cotransfected into HEK293 cells. Cell lysates were incubated with anti-Flag beads to precipitate MCPIP4. The immunoprecipitates were treated with RNase A (100 unit/ml) for 30 min and then subjected to Western blot analysis with anti-HA or anti-Flag antibodies. F) Mammalian two-hybrid assay for the interaction of MCPIP1 and MCPIP4. Different combinations of pBIND-and pACT-derived expression vectors were co-transfected with a reporter containing five Gal4 binding sites upstream of a minimum promoter-drive luciferase gene into HEK293 cells. The luciferase activity was measured using a dual-luciferase assay system. Data were presented as mean±SD, n=4.    A) The expression plasmids of MCPIP1/2/3/4 or an empty vector were co-transfected with the luciferase reporter of IL-6 3'UTR into HEK293 cells. A control reporter vector pRL-TK was also transfected to normalize the values. After 24 h, the cell lysates were prepared and the luciferase activity were measured by dual luciferase assay system. Data were presented as mean±SD, n=4, *P<0.05, **P<0.01 vs control group. B) The expression plasmids of MCPIP1 or/and MCPIP4 were co-transfected with the reporter of IL-6 3'UTR or the pGL3-control reporter into HEK293 cells. After 24 h, the cell lysates were prepared and the luciferase activity were measured by dual luciferase assay system. Data were presented as mean±SD, n=4, *P<0.05, **P<0.01 vs vector group. C) Inducible MCPIP1-stable expressed HEK293 cell line was treated with 0, 5 or 10 ng/ml of doxycycline for 24 h. The inducible expression of MCPIP1 in the cells was determined by western blot analysis with anti-GFP antibody. Actin was probed as a loading control. D) The inducible MCPIP1-stable cells were transiently cotransfected with the expression plasmid of MCPIP4 or empty vector and the reporter of IL-6 3'UTR or the pGL3control reporter and followed by treatment with or without 10 ng/ml of doxycycline for 24 h. The cell lysates were prepared and the luciferase activity were measured by dual luciferase assay system. Data were presented as mean±SD, n=4, *P<0.05, **P<0.01 vs vector group.