Pelota interacts with HAX1, EIF3G and SRPX and the resulting protein complexes are associated with the actin cytoskeleton

Abstract

Background: Pelota (PELO) is an evolutionary conserved protein, which has been reported to be involved in the regulation of cell proliferation and stem cell self-renewal. Recent studies revealed the essential role of PELO in the No-Go mRNA decay, by which mRNA with translational stall are endonucleotically cleaved and degraded. Further, PELO-deficient mice die early during gastrulation due to defects in cell proliferation and/or differentiation.

Results: We show here that PELO is associated with actin microfilaments of mammalian cells. Overexpression of human PELO in Hep2G cells had prominent effect on cell growth, cytoskeleton organization and cell spreading. To find proteins interacting with PELO, full-length human PELO cDNA was used as a bait in a yeast two-hybrid screening assay. Partial sequences of HAX1, EIF3G and SRPX protein were identified as PELO-interacting partners from the screening. The interactions between PELO and HAX1, EIF3G and SRPX were confirmed in vitro by GST pull-down assays and in vivo by co-immunoprecipitation. Furthermore, the PELO interaction domain was mapped to residues 268-385 containing the c-terminal and acidic tail domain. By bimolecular fluorescence complementation assay (BiFC), we found that protein complexes resulting from the interactions between PELO and either HAX1, EIF3G or SRPX were mainly localized to cytoskeletal filaments.

Conclusion: We could show that PELO is subcellularly localized at the actin cytoskeleton, interacts with HAX1, EIF3G and SRPX proteins and that this interaction occurs at the cytoskeleton. Binding of PELO to cytoskeleton-associated proteins may facilitate PELO to detect and degrade aberrant mRNAs, at which the ribosome is stalled during translation.

Figure 1

Subcellular localization of PELO at the cytoskeleton. A, Competition assay to determine the specificity of PELO antibodies. Blots with testis and lung protein extracts were incubated with anti-PELO antibodies in presence (+) or absence (-) of the GST-PELO fusion protein. B, Immunoblot with cellular extracts from different tissues was probed with anti-PELO antibodies. C, PELO is distributed in cytoskeletal and membrane fractions. Immunoblots with cellular extracts from different tissues were probed with anti-PELO antibodies. Blots containing samples of cytoskeletal (CS), cytoplasmic (C), nuclear (N) and membrane (M) fractions from HeLa cells were subsequently probed with polyclonal anti-PELO, monoclonal anti-α-actinin, anti-phospho H3 histone and anti-α-tubulin. D, Co-localization of the PELO antigen and actin stress fibers. Fibroblasts were immunolabeled with PELO antibodies (i, iv) or FITC-phalloidin (ii, v). PELO immunostaining was localized at the cytoskeleton (white arrows), at some regions of plasma membrane (yellow arrowhead), and near nuclear membranes (white arrowhead). The merged images (iii, vi) show that PELO largely co-localized with actin-stress fibers. Scale bars = 20 μm.

Figure 2

Overexpression of PELO affects actin stress fibers, cell growth and spreading in stable PELO-overexpressing cell lines. A, Hep2G cells were transfected with Myc-PELO fusion constructs or empty vector. Protein extracts from stable control (LC) and Myc-PELO expressing cell lines (L1, L3, L6 and L7) were analyzed by immunoblotting with monoclonal anti-Myc tag antibody. The membrane was subsequently reprobed with α-tubulin antibody. B-C, Cells of Myc-PELO-overexpressing L3 (B) and L6 (C) cell lines were stained with FITC-phalloidin. Cell showed diffusely cytoplasmic actin are labelled with arrows and that containing stress actin with arrowheads. D, In vitro growth properties of control and PELO-overexpressing cells. Growth curves of mock-transfected Hep2G (LC), and PELO-transfected clone L3 and L6. E, Growth of the same cell lines in soft agar. 0.2 × 10⁴ cells were plated per dish. 100% (LC) = 4.4 × 10² colonies. The data points represent the means ± SD of three independent experiments. F, Control cells (LC) as well as L3 and L6 that expressed PELO at low and high levels, respectively, were seeded onto fibronectin-coated slides for 30 and 120 min. Cells were then counted and scored for spreading. White bars represent L3, black bars L6 and grey bars LC cells. Results are presented as percentage of cells that adapted spread morphology at 30 and 120 min after seeding. 100 cells per well were counted in three separate experiments. Data represent means ± SD.

Figure 3

The acidic tail (AT, residues 371-385), and c-terminal (Pelo-1, residues 268-371) region of PELO is required for binding to HAX1, and EIF3G and SPRX, respectively. A, To determine the regions of PELO that are required for its association with HAX1, EIF3G and SRPX, four GST-PELO fusion proteins were generated (GST-1, GST-2, GST-3 and GST-4) and tested for their ability to bind in GST pull-down assays. GST alone (GST-0) was used as control in the assay. B, Recombinant GST-PELO fusion proteins were subjected to SDS-PAGE and stained with Coomassie blue. C, Equivalent amounts of GST-PELO and GST-0 proteins bound to glutathione-matrices were subsequently incubated with extracts of HeLa cells, which were transfected with either Myc-tagged HAX1, EIF3G or SRPX constructs. Western blots containing the GST pull-down assays were incubated with monoclonal anti-Myc antibodies. Immunoblot analysis revealed that GST-1 and GST-2, but not GST-3, GST-4 and GST-0 bound to Myc-HAX1 protein. Except GST-4 and GST-0, all other GST-PELO fusion proteins exhibit specific association to Myc-EIF3G and Myc-SRPX.

Figure 4

PELO interacts with HAX1, EIF3G and SRPX in vivo. A, Total protein from HeLa cells, which are co-transfected with HA-tagged PELO and either Myc-tagged HAX1, EIF3G or SRPX constructs, were precipitated with monoclonal anti-Myc antibodies. The precipitates of co-transfected cells with PELO-HA:Myc-HAX1 (IP1), PELO-HA:Myc-EIF3G (IP2) and PELO-HA:Myc-SRPX (IP3) were probed in immunoblot analysis with polyclonal anti-HA antibodies. IPC, control co-immunoprecipitation (coIP) assay using anti-mouse IgG serum. B-C, In reciprocal coIP experiments, cells lysates from co-transfected HeLa cells were precipitated with either polyclonal anti-HA (B) or anti-PELO antibodies (C). Precipitated proteins (IP) were detected with anti-Myc antibodies. IPC, control coIP experiments using anti-rabbit IgG serum.

Figure 5

Cellular localization of PELO-interacting proteins. A, HeLa cells were transiently co-transfected with PELO-HA and Myc-HAX1 and immunostained with anti-HA (panel i) and anti-Myc (panel ii) antibodies. A co-localization of both proteins is mainly diffused in cytoplasm (panel iii). B-N, Bimolecular fluorescence complementation analysis (BiFC) revealed that protein complexes between PELO and HAX1, EIF3G or SRPX are localized at the cytoskeleton. B-E, Cells transfected with GFPN (B) and GFPC (C), PELO-GFPC (D) and HAX1-GFPN (E) alone did not show any fluorescence. F-N, HeLa cells were transiently co-transfected with PELO-GFPC and either HAX1-GFPN (F), EIF3G-GFPN (G), SRPX-GFPN (M) or MOS-GFPN (N) as control. Complementation of both non-fluorescent GFP halves resulted in functional fluorophores, which is diagnostic for stable interaction of the fusion proteins. The reconstituted GFP-fluorescence (BiFC) is shown in green and mainly localized at the cytoskeleton. Absence of GFP-fluorescence (N) indicates lack of BiFC. DNA staining is shown in blue. Scale bars = 20 μm.