Phosphorylation of human tristetraprolin in response to its interaction with the Cbl interacting protein CIN85

Abstract

Background: Tristetraprolin (TTP) is the prototype member of a family of CCCH tandem zinc finger proteins and is considered to be an anti-inflammatory protein in mammals. TTP plays a critical role in the decay of tumor necrosis factor alpha (TNF) mRNA, among others, by binding AU-rich RNA elements in the 3'-untranslated regions of this transcript and promoting its deadenylation and degradation.

Methodology/principal findings: We used yeast two-hybrid analysis to identify potential protein binding partners for human TTP (hTTP). Various regions of hTTP recovered 31 proteins that fell into 12 categories based on sequence similarities. Among these, the interactions between hTTP and CIN85, cytoplasmic poly (A) binding protein (PABP), nucleolin and heat shock protein 70 were confirmed by co-immunoprecipitation experiments. CIN85 and hTTP co-localized in the cytoplasm of cells as determined by confocal microscopy. CIN85 contains three SH3 domains that specifically bind a unique proline-arginine motif (PXXXPR) found in several CIN85 effectors. We found that the SH3 domains of CIN85 bound to a PXXXPR motif located near the C-terminus of hTTP. Co-expression of CIN85 with hTTP resulted in the increased phosphorylation of hTTP at serine residues in positions 66 and 93, possibly due in part to the demonstrated association of mitogen-activated protein kinase kinase kinase 4 (MEKK4) to both proteins. The presence of CIN85 did not appear to alter hTTP's binding to RNA probes or its stimulated breakdown of TNF mRNA.

Conclusions/significance: These studies describe interactions between hTTP and nucleolin, cytoplasmic PABP, heat shock protein 70 and CIN85; these interactions were initially discovered by two-hybrid analysis, and confirmed by co-immunoprecipitation. We found that CIN85 binding to a C-terminal motif within hTTP led to the increased phosphorylation of hTTP, possibly through enhanced association with MEKK4. The functional consequences to each of the members of this putative complex remain to be determined.

Figure 1. Co-immunoprecipitation of hTTP with potential interacting partners.

In this and subsequent figures, extracts were prepared in RIPA buffer from HEK 293 cells transfected with DNA encoding the HA- and Flag-tagged expression vectors indicated at the top of each gel lane by the “+” sign. Total DNA transfected was 5.0 µg per 100 mm petri dish. For each western blot shown, the immunoprecipitating antibody (IP) and the subsequent immunoblotting antibody (IB) are indicated to the left of each panel, as are the positions of protein molecular weight standards. The immunoreactive protein species are indicated by the labeled arrows to the right of each blot. Each immunoprecipitation used 1 mg of cellular lysate protein as the starting material. In some cases, the blots are of whole cell lysates (WCL) (50 µg of total protein per lane) instead of from immunoprecipitations to confirm expression of the respective protein in the lysates prior to immunoprecipitation. In addition to the epitope-tag antibodies indicated, western blotting in this case also used antibodies to endogenous nucleolin (NCL), CIN85, and HSP70. See the Results section for additional details.

Figure 2. Interaction of TTP and with PABP and CIN85.

Abbreviations and other details are as described in the legend to Fig. 1.

Figure 3. Association of CIN85 with hTTP family members.

Abbreviations and other details are as described in the legend to Fig. 1, with the exception of Panel 3B1, which shows an immunoblot of the whole cell lystate probed with anti-Flag antibody.

Figure 4. Sites of interaction in hTTP and CIN85.

In A, Flag-CIN85 was co-expressed with various C-terminal truncated forms of HA-hTTP, as indicated at the top of the figure. Abbreviations and other details are as described in the legend to Fig. 1. In B is shown a ClustalW alignment of the putative PXXXPR CIN85 binding motif near the C-terminus of hTTP aligned with the C-termini of TTP from various vertebrate species, as well as with the TTP family members ZFP36L1 and ZFP36L2 from human. The sequences shown are derived from the following GenBank accession numbers: human TTP (NP_003398.1), mouse (NP_035886.1), rat (NP_579824.2), chimpanzee (XP_001136016), rhesus (XP_001086084.1), horse (CD536523.1), sheep (NP_001009765.1), cow (NP_776918.1), dog (XP_541624.2), pig (DY419026), Xenopus tropicalis (Xtrop) (NP_001106542.1) and Xenopus laevis (Xlaev) (NP_001081884.1) The putative CIN85 binding PXXXPR motif in human and other mammalian TTPs (but not mouse) is boxed. This motif is also not present in the orthologues from the two frog species. The typical PXXXPR motif is also not present in the in the C-termini of the other human TTP family members ZFP36L1 (NP_004917) and ZFP36L2 (NP_008818). In C are shown schematic representations of full-length CIN85 and its truncations. The names of various truncations of CIN85 and their amino acid positions are indicated on the left. Positions of the Src homology domains 3 (SH3) A, B and C) are indicated in light grey boxes; the proline rich region (P- rich) is shown as a darker grey box; and the C-terminal coiled-coil region is shown as a black box. The ability of each construct to bind hTTP is indicated on the right. The data supporting this diagram are shown in D and F, with the transfected plasmids shown at the top of the blot, and with other aspects of the western blots as described in the legend to Fig. 1. Panel 3 documents the expression of full-length hTTP in each WCL (50 µg/lane). In E are shown data from mutations in the PXXXPR motif in hTTP, and the corresponding sequence in mouse TTP. Abbreviations and other details are as in the legend to Fig. 1.

Figure 5. Localization of hTTP and MEKK4 with CIN85 in cultured cells.

In A, HEK 293 cells were transfected with plasmids for the expression of HA-hTTP (red) and Flag-CIN85 (Green), or (B) HA-MEKK4 (red) and Flag-CIN85 (green). The cells were stained with primary antibodies, either an anti-HA polyclonal or anti-Flag monoclonal, followed by the secondary antibodies Alexa 594 anti-rabbit (red) or Alexa 488 anti-mouse (green), respectively. Nuclei (blue) were stained with DAPI. The cells were visualized and images obtained by confocal microscopy. The merged images of two protein signals, indicating areas of apparent co-localization, are shown in yellow.

Figure 6. Interaction between hTTP and both MEKK4 and CIN85.

In A, tagged expression constructs of MEKK4, CIN85 and hTTP were co-transfected in pairs (lanes 1 to 7) or together (lane 8) with or without empty vectors. Abbreviations and other details are as described in the legend to Fig. 1. In B are shown whole cell lysates, demonstrating the protein expression from the various expression plasmids or empty vectors, as well as the apparent shift in the M_r of hTTP when co-expressed with CIN85.

Figure 7. Phosphorylation sites in hTTP co-expressed with CIN85.

Panels A, B, and C are extracted ion chromatograms (EICs) for residues 66–82 (A), 83–103 (B), and 183–194 (C) generated from nanoLC-ESI-MS runs derived from tandem MS data of ions m/z 915.4, m/z 1083.5, and m/z 658.5, corresponding to the phosphorylated peptides 66–82, 83–103, and 183–194, respectively (Data/Fig. not shown). The normalized responses demonstrated in the EICs are an estimation of the abundance of the ion of interest. The currents from the non-phosphorylated peptides are represented as solid lines, and the ion currents attributed to the phosphorylated forms of the same peptides are shown as dashed lines. Duplicate technical replicates yielded similar results.

Figure 8. Effect of co-expressed CIN85 on hTTP binding to an RNA probe.

Cytosolic extracts of HEK293 cells transfected with vector alone (BS), or vectors expressing HA-hTTP alone and Flag-CIN85 alone, were used in RNA gel shift analysis, using a 5′ biotin-labeled TNF-ARE based RNA probe. In A, protein extracts of containing decreasing amounts of HA-hTTP incubated in the presence or absence of a constant concentration of FLAG-CIN85 were incubated with 0.6 ng of the RNA probe. The migration positions of the hTTP-ARE complexes, the non-specific complexes seen in the HEK 293 cell extract alone, and the RNA probe alone, are all indicated with arrows to the right of Fig. 8A. In B, 2 µg of cellular protein from FLAG-CIN85 expressing cell extracts was incubated with or without HA-hTTP or FLAG-hTTP (1.6 µg) in the presence or absence of the respective epitope tag antibodies. Arrows to the right of the panel are the same as in A, except for the addition of an arrow pointing to the hTTP supershifts. In C, immunoblots were performed using 10 µg of cellular protein from the same extracts, demonstrating expression of the epitope-tagged proteins.

Figure 9. Effects of co-expression of CIN85 on hTTP-promoted destabilization of a TNF mRNA.

In A, a CMV-driven mouse TNF-encoding plasmid was co-transfected into HEK 293 cells (lanes 2–6, 8–12) with either vector alone or with the indicated amounts of an hTTP expression construct in the presence or absence of 2 µg of a CIN85 expression construct. Total cellular RNA was harvested 24 h later, and used for northern blotting. Each lane was loaded with 10 µg of total RNA. Lanes 1 and 7 were from mock-transfected HEK 293 cells. Lanes 2 and 8 were from cells transfected with vector alone (BS+; 5 µg/plate). Lanes 3–6, were from cells co-transfected with CMV.mTNF (1 µg) and CMV.hTTP.tag (0.005, 0.05, 0.1, and 0.5 µg/plate, respectively). Lanes 9–12 were from cells co-transfected with CMV.mTNF (1 µg) and CMV.CIN85.tag (2 µg) and CMV.hTTP.tag (0.005, 0.05, 0.1 and 0.5 µg/plate, respectively). Vector was also added as needed to make the total amount of co-transfected plasmids 5 µg/plate in each case. As indicated, the northern blots were probed with either a ³²P-labeled mTNF cDNA probe (panels 1 and 2, duplicate experiments), an hTTP probe (panel 3), a CIN85 probe (panel 4) or a GAPDH probe (panel 5). Film exposure was 4 h and 7 h, respectively, for panels, A1 and A2 for filters hybridized with an mTNF probe. All other filters were exposed to films for 7 h. The two parallel lines labeled TNF indicate the two species of TNF mRNA discussed in the text. The positions of the 18S rRNA are indicated. In B are shown the phosphorimager values for both species of TNF mRNA as a function of various TTP plasmid amounts, transfected with or without the CIN85 vector. The graph in B is from a single experiment, but is representative of three similar experiments.