Kaposi's sarcoma-associated herpesvirus latency-associated nuclear antigen and angiogenin interact with common host proteins, including annexin A2, which is essential for survival of latently infected cells

Abstract

Kaposi's sarcoma-associated herpesvirus (KSHV) infection and latency-associated nuclear antigen (LANA-1) upregulate the multifunctional protein angiogenin (ANG). Our studies demonstrate that silencing ANG or inhibiting its nuclear translocation downregulates KSHV LANA-1 expression and ANG is necessary for KSHV latency, anti-apoptosis and angiogenesis (Sadagopan et al., J. Virol. 83:3342-3364, 2009; Sadagopan et al., J Virol. 85:2666-2685, 2011). Here we show that LANA-1 interacts with ANG and colocalizes in latently infected endothelial telomerase-immortalized human umbilical vein endothelial (TIVE-LTC) cells. Mass spectrometric analyses of TIVE-LTC proteins immunoprecipitated by anti-LANA-1 and ANG antibodies identified 28 common cellular proteins such as ribosomal proteins, structural proteins, tRNA synthetases, metabolic pathway enzymes, chaperons, transcription factors, antioxidants, and ubiquitin proteosome proteins. LANA-1 and ANG interaction with one of the proteins, annexin A2, was validated. Annexin A2 has been shown to play roles in cell proliferation, apoptosis, plasmin generation, exocytosis, endocytosis, and cytoskeleton reorganization. It is also known to associate with glycolytic enzyme 3-phosphoglyceratekinase in the primer recognition protein (PRP) complex that interacts with DNA polymerase α in the lagging strand of DNA during replication. A higher level of annexin A2 is expressed in KSHV+ but not in Epstein-Barr virus (EBV)+ B-lymphoma cell lines. Annexin A2 colocalized with several LANA-1 punctate spots in KSHV+ body cavity B-cell lymphoma (BCBL-1) cells. In triple-staining analyses, we observed annexin A2-ANG-LANA-1, annexin A2-ANG, and ANG-LANA-1 colocalizations. Annexin A2 appeared as punctate nuclear dots in LANA-1-positive TIVE-LTC cells. In LANA-1-negative TIVE-LTC cells, annexin A2 was detected predominately in the cytoplasm, with some nuclear spots, and colocalization with ANG was observed mostly in the cytoplasm. Annexin A2 coimmunoprecipitated with LANA-1 and ANG in TIVE-LTC and BCBL-1 cells and with ANG in 293T cells independent of LANA-1. This suggested that annexin A2 forms a complex with LANA-1 and ANG as well as a separate complex with ANG. Silencing annexin A2 in BCBL-1 cells resulted in significant cell death, downregulation of cell cycle-associated Cdk6 and of cyclin D, E, and A proteins, and downregulation of LANA-1 and ANG expression. No effect was seen in KSHV⁻ lymphoma (BJAB and Ramos) and 293T cells. These studies suggest that LANA-1 association with annexin A2/ANG could be more important than ANG association with annexin A2, and KSHV probably uses annexin A2 to maintain the viability and cell cycle regulation of latently infected cells. Since the identified LANA-1- and ANG-interacting common cellular proteins are hitherto unknown to KSHV and ANG biology, this offers a starting point for further analysis of their roles in KSHV biology, which may lead to identification of potential therapeutic targets to control KSHV latency and associated malignancies.

Fig 1

KSHV LANA-1 interacts with angiogenin (ANG). (A) TIVE-LTC cells were stained for angiogenin (green) and LANA-1 (red) after permeabilization. DAPI (blue) was used as a nuclear stain and merged with LANA-1 and ANG. Punctate ANG and LANA-1 colocalization can be readily seen. (B) TIVE-LTC cells were stained for ANG (green). The yellow arrows in the enlarged insert indicate diffused cytoplasmic and nuclear staining of ANG in uninfected, LANA-1-negative cells. (C) TIVE-LTC cells were grown to confluence, and nuclear extract was prepared. Equal protein quantities of each lysate were immunoprecipitated with equal concentrations of either rabbit anti-angiogenin IgG, rabbit LANA-1 IgG antibody, or control rabbit IgG antibody and Western blotted for LANA-1 and ANG. Input control for each IP is shown to demonstrate that an equal amount of LANA-1 and ANG were present in both the samples. (D) Schematics of the experimental design for proteomic analysis. Confluent TIVE-LTC cells were harvested with RIPA buffer, and 500 μg of cell lysates were immunoprecipitated using protein G-Sepharose beads and rabbit anti-angiogenin IgG, rabbit anti-LANA-1 IgG, or rabbit anti-IgG antibodies. The slurry was incubated at 4°C in a rotating chamber for 2 h. The beads were then washed three times with RIPA buffer, resuspended in 2× sample solubilizing buffer, boiled at 95°C for 5 min, and run on a 4% to 20% gradient SDS-PAGE gel. Protein bands unique to LANA-1- and angiogenin-immunoprecipitated lanes were excised and subjected to mass spectrometric analysis. (E) Representative samples of Coomassie stained gel used for mass spectrometric analysis. TIVE-LTC lysates were immunoprecipitated with rabbit IgG (lane 1), rabbit anti-ANG IgG (lane 2), and rabbit anti-LANA-1 IgG (lane 3) and analyzed as describe above. Molecular weight markers are shown on the left.

Fig 2

STRING analysis of all the common proteins immunoprecipitated by angiogenin and LANA-1. The STRING program generates functional protein association networks. (A) Action view; uses different-colored lines to depict the types of interaction between proteins. (B) Evidence view; uses different-colored lines to depict the type of evidence that supports each interaction.

Fig 3

Pie chart showing the percentages of functionally classified proteins found to commonly interact with both angiogenin and LANA-1.

Fig 4

Detection of annexin A2 in various B-cell lines. (A and B) RNA from the indicated B cells was extracted, and cDNA was prepared and used in real-time RT-PCR to determine annexin A2 gene expression using the SYBR green detection protocol. Results were normalized to 18S rRNA expression levels, and the annexin A2 level in the cell line with the lowest value was assigned a value of 1 for comparisons. (C and D) Equal concentrations of protein lysates from the indicated B-cell lines were separated by SDS-PAGE and Western blotted for annexin A2. β-Actin was blotted to show equal protein loading. (E) Nuclear and cytoplasmic extracts from BCBL-1 cells were prepared and Western blotted for annexin A2. Lamin and β-actin were blotted to check the purity of the extracts.

Fig 5

Colocalization of annexin A2-ANG and LANA-1 in KSHV-infected and noninfected cells. (A and B) Immunofluorescence colocalization of ANG and LANA-1 with annexin A2 in BCBL-1 cells. BCBL-1 cells were stained for ANG (red) and annexin A2 (green) in panel A and for LANA-1 (red) and annexin A2 (green) in panel B after permeabilization. DAPI (blue) was used as a nuclear stain and merged with ANG and annexin A2 or LANA-1 and annexin A2. Long white arrows indicate ANG-annexin A2 colocalization. LANA-1-annexin A2 colocalization is indicated by thick white arrows. (C and D) TIVE-LTC cells were stained for LANA-1 (red) and annexin A2 (green) in panel C and for ANG (red) and annexin A2 (green) in panel D after permeabilization. ANG-annexin A2 (long white arrows) and LANA-1-annexin A2 (thick white arrows) colocalizations are indicated. Long yellow arrows indicate the LANA-1-negative TIVE-LTC cells. Boxed area is enlarged. (E) TIVE-LTC cells were stained for ANG (blue), LANA-1 (red), and annexin A2 (green) and merged in different combinations. White arrowheads indicate ANG-LANA-1-annexin A2 triple colocalizations. Yellow short arrows indicate ANG-LANA-1 colocalizations. ANG-annexin A2 (long white arrows) and LANA-1-annexin A2 (thick white arrows) colocalizations are also indicated.

Fig 6

Coimmunoprecipitation of annexin A2 with ANG and LANA-1. (A) TIVE and TIVE-LTC cells were grown to confluence, harvested in RIPA buffer, and immunoprecipitated using rabbit anti-angiogenin antibody and Western blotted for annexin A2. Annexin A2 and ANG input control and β-actin loading control are shown for each IP. (B) Same conditions as for panel A but with BJAB and BCBL-1 cells. (C) Nuclear extracts of TIVE and TIVE-LTC cells were prepared and immunoprecipitated using rabbit anti-LANA-1 antibody and Western blotted for annexin A2. Annexin A2 and LANA-1 input control and β-actin loading control are shown for each IP. (D) Same conditions as for panel C but with BJAB and BCBL cells. (E) 293T cells were transfected with an empty vector plasmid or full-length GFP-tagged angiogenin plasmid and harvested with RIPA buffer after 48 h. The lysates were immunoprecipitated with rabbit anti-angiogenin antibody and Western blotted for annexin A2. GFP input controls for pCDNA-GFP (lower band) and ANG-GFP (upper band) are shown.

Fig 7

Silencing annexin A2 affects viability and LANA-1 and ANG expression in BCBL-1 cells without affecting Ramos and BJAB cells. (A) 293T cells were infected with lentiviruses carrying four short hairpin RNAs targeting annexin A2 and one targeting GFP as a control. Annexin A2 knockdown was checked by Western blotting at 72 h post-lentivirus infection. (B) BCBL-1, BJAB, and Ramos cells were infected with si-GFP control or si-annexin A2-4 lentivirus particles and blotted for annexin A2 and β-actin at 72 h post-lentivirus infection. BCBL-1 cells were also blotted for β-integrin. (C, D, and E) Ramos (C), BJAB (D), and BCBL-1 (E) cells infected as described above were stained with YO-PRO-1/PI and analyzed by FACS at 72 h post-lentivirus infection. Cells were gated for propidium iodide (PI) and GFP. (F, G, and H) Quantitative analysis of YO-PRO staining. One of the representative samples from three independent experiments is shown here. The percentages of live and dead cells are represented for each cell type with si-GFP and si-annexin A2. The values were normalized to account for the differences in transduction efficiency. (I and J) RNA was extracted from BCBL-1 cells infected with si-GFP control or si-annexin A2 lentivirus particles as described above, and mRNA levels of LANA-1 and ANG were measured using real-time RT-PCR at 72 h post-lentivirus infection. (K and L) Protein levels of LANA-1 and ANG were measured by Western blotting BCBL-1 cells infected with si-GFP control or si-annexin A2 lentivirus particles as described above.

Fig 8

Silencing annexin A2 affects cell cycle progression in BCBL-1 cells but not in Ramos and BJAB cells. (A, B, and C) Ramos, BJAB, and BCBL-1 cells infected with si-GFP and si-annexin A2 lentivirus particles were stained with propidium iodide (PI) and analyzed by FACS at 72 h post-lentivirus infection for cell cycle distribution. One representative figure is shown. (D, E, and F) Quantitative analysis of cell cycle distribution of Ramos (D), BJAB (E), and BCBL-1 (F) cells treated with si-GFP or si-annexin A2 lentivirus from three independent experiments. The percentages of cells in G₀+G1 and G₂+S phases are graphed. (G, H, I, and J) 293T, Ramos, BJAB, and BCBL-1 cells treated as described above were harvested with RIPA buffer after 72 h. The lysates were Western blotted for cyclin D, Cdk6, cyclin E, cyclin A, and β-actin. The percent reduction of proteins in BCBL-1 cells is indicated.