Hepatoma-derived growth factor and nucleolin exist in the same ribonucleoprotein complex

Abstract

Background: Hepatoma-derived growth factor (HDGF) is a protein which is highly expressed in a variety of tumours. HDGF has mitogenic, angiogenic, neurotrophic and antiapoptotic activity but the molecular mechanisms by which it exerts these activities are largely unknown nor has its biological function in tumours been elucidated. Mass spectrometry was performed to analyse the HDGFStrep-tag interactome. By Pull-down-experiments using different protein and nucleic acid constructs the interaction of HDGF and nucleolin was investigated further.

Results: A number of HDGFStrep-tag copurifying proteins were identified which interact with RNA or are involved in the cellular DNA repair machinery. The most abundant protein, however, copurifying with HDGF in this approach was nucleolin. Therefore we focus on the characterization of the interaction of HDGF and nucleolin in this study. We show that expression of a cytosolic variant of HDGF causes a redistribution of nucleolin into the cytoplasm. Furthermore, formation of HDGF/nucleolin complexes depends on bcl-2 mRNA. Overexpression of full length bcl-2 mRNA increases the number of HDGF/nucleolin complexes whereas expression of only the bcl-2 coding sequence abolishes interaction completely. Further examination reveals that the coding sequence of bcl-2 mRNA together with either the 5' or 3' UTR is sufficient for formation of HDGF/nucleolin complexes. When bcl-2 coding sequence within the full length cDNA is replaced by a sequence coding for secretory alkaline phosphatase complex formation is not enhanced.

Conclusion: The results provide evidence for the existence of HDGF and nucleolin containing nucleoprotein complexes which formation depends on the presence of specific mRNAs. The nature of these RNAs and other components of the complexes should be investigated in future.

Figure 1

Mass spectrometry analysis of purified HDGF protein complexes. HDGF interaction partners were isolated using Strep-Tactin®-MacroPrep matrices as affinity resin to purify HDGFStrep-tag proteins. Trapped proteins were digested by trypsin and peptides analyzed by ESI-MS/MS. (A) Flow chart of the experiment. (B) Venn diagramm showing the overlap between HDGFStrep-tag and GFPStrep-tag interactome. (C) Pie chart of molecular functions of HDGFStrep-tag interaction partners (see also Additional file 4).

Figure 2

Verification of nucleolin copurification with HDGF. (A) Analysis of most abundant HDGFStrep-tag copurifying proteins by Coomassie stained SDS-PAGE. Proteins purified via Strep-Tactin®-MacroPrep matrices after transfection of cells with HDGFStrep-tag or untagged HDGF (Mock) were separated using SDS-PAGE and visualized by Coomassie staining. Visible proteins specific for HDGFStrep-tag sample correspond to Ku86 (86 kDa), nucleolin (110 kDa), PARP-1 (113 kDa) and the catalytic subunit of DNA-PK (460 kDa). (B) To confirm mass spectrometric results, lysates of HEK293 cells expressing HDGFStrep-tag fusion proteins and their eluate fractions from HDGFStrep-tag purification were examined by Western blot with a specific nucleolin antibody. Nucleolin is only present in eluates from cells expressing HDGFStrep-tag fusion proteins. (C) To investigate whether endogenous HDGF also interacts with nucleolin a cellular extract of untransfected HeLa cells was loaded onto a human HDGF antibody column. Non-specific proteins were washed out with 300 mM NaCl. Specifically binding proteins were eluted with 50 mM citric acid pH 3. Western blot analysis with a monoclonal anti nucleolin antibody confirmed copurification of nucleolin with endogenous HDGF.

Figure 3

HDGF changes intracellular distribution of nucleolin. (A) To examine HDGF/nucleolin complex formation in HeLa cells, cells were transfected with either wild type HDGF or a HDGF-NLS2 mutant expression vector and fixed after 48 hours with methanol. Cells transfected with wild type HDGF (HDGFwt) show predominant nuclear localization of HDGF (red, Figure 3B panel a) as well as nucleolin (green, Figure 3B panel c). In contrast, induction of cytoplasmic expression of HDGF-NLS2 leads to a significant increase of cytoplasmic HDGF (red, Figure 3B panel b) as well as nucleolin (green, Figure 3B panel d). Cells were counterstained with DAPI (blue). (B) HDGF positive cells were counted (n = 100) and localisation of nucleolin was analysed. 67% of HDGF-NLS2 transfected cells show cytoplasmic localisation of nucleolin in comparison to only 7% of wild type transfected cells. *p < 0.001. (C) Protein lysates from HDGFStrep-tag expressing HeLa cells were incubated with or without RNAse A (25 μg/ml) over night at 4°C followed by affinity purification via Strep-Tactin®-MacroPrep spin columns. Proteins from eluate fractions were subjected to Western blot analysis using a specific nucleolin antibody (upper panel) or with peroxidase coupled Strep-Tactin® to detect HDGFStrep-tag (lower panel). In samples treated with RNAse A a complete loss of interaction can be observed.

Figure 4

HDGF and nucleolin complex formation depends on bcl-2 full length mRNA. (A) To investigate a possible involvement of bcl-2 mRNA in HDGF-nucleolin complex formation cells were treated with Paclitaxel and protein lysates were examined for nucleolin copurification as described above. Incubation with the bcl-2 mRNA destabilizing agent Paclitaxel led to a significant decrease of nucleolin copurification. (B + C) Different deletion constructs of bcl-2 mRNA were generated and cotransfected together with HDGFStrep-tag to investigate their influence on the copurification of nucleolin. The full-length bcl-2 mRNA (FL Bcl-2) had a positive whereas the coding sequence of bcl-2 (cds) had a negative effect on the amount of coprecipitated nucleolin. Interestingly, addition of 400 bp of the 3^′UTR (ARE1) containing a known nucleolin binding motive was already sufficient to rescue the negative effect on the amount of precipitated nucleolin. Furthermore, beside the 3′UTR also the 5^′UTR had a rescuing effect on the interaction.

Figure 5

Cds of bcl-2 mRNA is necessary for HDGF and nucleolin complex formation. (A) To investigate a possible role of the coding sequence of bcl-2 (cds) it was replaced by the cds of secreted alkaline phosphatase (SEAP). (B) Both constructs led to a significant increase in bcl-2 UTRs in the cells as detected by RT-PCR. (C) Replacement of the cds of bcl-2 against the cds of SEAP was not able to increase the amount of coprecipitated nucleolin.