The Epstein-Barr virus alkaline exonuclease BGLF5 serves pleiotropic functions in virus replication

Abstract

The Epstein-Barr virus (EBV) alkaline exonuclease BGLF5 has previously been recognized to contribute to immune evasion by downregulating production of HLA molecules during virus replication. We have constructed a BGLF5-null virus mutant to determine BGLF5's functions during EBV viral replication. Quantification of virus production in permissive 293 cells carrying a DeltaBGLF5 genome identified a 17- to 21-fold reduction relative to complemented or wild-type controls. Detailed monitoring of DeltaBGLF5 replication evidenced an impaired virus nucleocapsid maturation, a reduced primary egress and a 1.4-fold reduction in total viral DNA synthesis. DeltaBGLF5 single-unit-length viral genomes were not only less abundant but also migrated faster than expected in gel electrophoresis. We concluded that BGLF5 pertained both to the generation and to the processing of viral linear genomes. DeltaBGLF5 phenotypic traits were reminiscent of those previously identified in a mutant devoid of UL12, BGLF5's homolog in herpes simplex virus type 1, and indeed UL12 was found to partially complement the DeltaBGLF5 phenotype. However, BGLF5-specific functions could also be identified; the nuclear membrane of replicating cells displayed images of reduplication and complex folding that could be completely corrected by BGLF5 but not UL12. Similar nuclear abnormalities were previously observed in cells transfected with BFLF2 and BFRF1, two viral proteins crucial for EBV nuclear egress. Interestingly, DeltaBGLF5 cells produced more BFLF2 than wild-type or complemented counterparts. The present study provides an overview of BGLF5's functions that will guide future molecular studies. We anticipate that the 293/DeltaBGLF5 cell line will be instrumental in such developments.

FIG. 1.

Construction of a ΔBGLF5 recombinant virus. (A) Schematic map of the EBV genome segment that encompasses the BGLF5 gene in EBV-wt and after homologous recombination with the targeting vector carrying the kanamycin resistance gene. The overlapping BGLF4 and BBLF1 coding regions are not affected by the recombination. The cleavage sites for EcoRI (RI) and the expected fragment sizes for cleavage of EBV-wt and ΔBGLF5 genomes are given. pA, polyadenylation site, kana, kanamycin. (B) EcoRI restriction fragment analysis of EBV-wt (lane 1) and ΔBGLF5 mutant genomes (lane 2) after construction in E. coli or after rescue from stably transfected 293 cells (293/ΔBGLF5) (lane 3). The result is consistent with the predicted restriction pattern (see panel A). (C) Immunofluorescence analysis of lytically induced 293/ΔBGLF5, 293/EBV-wt and transcomplemented 293/ΔBGLF5 cells (293/ΔBGLF5-C) with a BGLF5-specific antibody. No BGLF5 protein was detected in 293/ΔBGLF5 mutant cells 2 days after induction. (D) Western blot analysis on extracts from induced 293/EBV-wt, 293/ΔBGLF5, and 293/ΔBGLF5-C cells, 48 h after induction. The transcomplemented 293/ΔBGLF5-C and 293/EBV-wt producer lines express BGLF5 at comparable levels, 293/ΔBGLF5 cells were BGLF5 negative. Staining of the same blot with an actin-specific antibody served as a loading control.

FIG. 2.

BGLF5 is required for virus production but not B-cell immortalization. (A) qPCR analysis with supernatants from lytically induced 293/ΔBGLF5, 293/ΔBGLF5-C and 293/EBV-wt cells. The concentration of viral geq is reduced in ΔBGLF5 supernatants compared to EBV-wt and ΔBGLF5-C supernatants. Presented are the mean values from eight independent experiments. (B) Electron micrographs of pelleted ΔBGLF5 and ΔBGLF5-C virus supernatants. ΔBGLF5 supernatants contained a lower number of mature virions in comparison with ΔBGLF5-C supernatants. Arrows indicate mature virions. Bar, 0.2 μm. (C) Raji B-cell infection assay. The concentration of infectious particles in ΔBGLF5 supernatants (as measured by green Raji cells) is reduced compared to EBV-wt and ΔBGLF5-C supernatants (mean value from eight independent experiments). (D) Multiple infections were carried out with 293 cells as target cells. Mean values of the percentages of infected cells are given. (E) Immortalization experiments with primary B cells from different donors. The results are given as mean values of the percentage of wells with cell outgrowth.

FIG. 3.

Electron micrographs of induced 293/ΔBGLF5 (A, C, and D) and 293/ΔBGLF5-C cells (B). (A) Overview of a replicating 293/ΔBGLF5 cell at low (×8,000) magnification. The nucleus contains numerous immature B capsids with scaffolding structures but only rare mature C-capsids carrying electron-dense DNA (inset I). No viral structures could be detected in the cytoplasm. The nuclear membrane was highly irregular in shape and displayed electron-dense membrane foldings and/or duplications within the perinuclear space resulting in the formation of nuclear pockets, some of which were deeply located in the cytoplasm (inset II). (B) Complementation of 293/ΔBGLF5 cells corrected these abnormalities; all three types of nucleocapsids (inset III) and mature capsids were visible in the cytoplasm, and the nuclear membrane displayed normal morphological features (inset IV). (C and D) Several micrographs provide a more detailed view of the nuclear membrane from induced 293/ΔBGLF5 cells at lower and high magnifications. The nuclear membrane appears thickened and by places connected to heterochromatin (panel C, left and middle images). The cellular material accumulated in the nuclear membrane has a similar density as heterochromatin. The remaining pictures (panel C, right image, and panel D) exemplify what are suspected to be earlier events; the nuclear membrane appears duplicated with four stacked membrane layers visible (arrow). The electron-dense material appears to progressively fill in the space between nuclear membranes to form multiple projections of nuclear envelope that produce an appearance of concentric nuclear pockets on these two-dimensional sections. Note that some nuclear pockets enclose a profile of cytoplasm (panel B, middle image). INM, inner nuclear membrane; ONM, outer nuclear membrane; nuc, nucleus; cyt, cytoplasm; HC, heterochromatin; A, B, and C, A-, B-, and C-type capsids. Bar, 0.2 μm.

FIG. 4.

BGLF5's functions in viral lytic DNA replication. (A) Viral DNA replication in uninduced and induced (ind) cells was quantitated by qPCR. Mean values and standard deviations from three independent experiments are presented. (B) Southern blot analysis of BamHI-cleaved DNA fragments from induced cells hybridized with a TR-specific probe. The 10-kb fragment results from restriction of complete BamHI Nhet fragments that are present only in circular genomes or genome concatemers. In contrast, the smaller fragments are generated by restriction of single unit-length linear genomes (see the adjacent schematic). The “1” and “2” represent circular and concatemeric DNA; the “3” corresponds to linear DNA segments. (C) Gardella gel electrophoresis coupled to Southern blot analysis using a nonrepetitive EBV-specific probe. The linear B95.8 genomes are 172 kb, whereas the ΔBGLF5 and EBV-wt genomes are 183 kb.

FIG. 5.

Expression of BFLF2/BFRF1 nuclear proteins in ΔBGLF5 cells. (A) Western blot analysis of protein extracts from lytically induced 293/EBV-wt (lane 1), 293/ΔBGLF5 (lane 2), and 293/ΔBGLF5-C (lane 3) cells. The same blot was sequentially immunostained with antibodies specific to BFLF2, BFRF1, BGLF5, or actin as a loading control. (B) Induced cells were immunostained with antibodies specific to BFLF2 or BFRF1, and nuclei were counterstained with Hoechst dye (blue). Immunofluorescence was recorded by using a confocal microscope using identical exposure times to analyze both the distribution and the intensity of the fluorescent signals. Localization of both proteins at the nuclear rim is not altered in the absence of BGLF5.

FIG. 6.

Complementation assays with UL12. UL12 was transfected into induced 293/ΔBGLF5 cells. Genome DNA equivalents (A) and green Raji titers (B) were assessed and compared to those obtained with noncomplemented 293/ΔBGLF5 cells, BGLF5-complemented 293/ΔBGLF5 cells, and induced 293/EBV-wt cells (i.e., the same data as presented in Fig. 2). (C) The same experiments were conducted with 293 cells as target cells. (D) Two electron micrographs of induced 293/ΔBGLF5 cells complemented with UL12. Note the irregular nuclear envelope and the presence of C-type nucleocapsids and of extracellular mature virions (arrows). INM, inner nuclear membrane; ONM, outer nuclear membrane; nuc, nucleus; cyt, cytoplasm; HC, heterochromatin; B and C, B- and C-type capsids. Bar, 0.2 μm.