ABSTRACT
Human adenovirus (HAdV) encodes a multifunctional DNA-binding protein pVII, which is involved in virus DNA packaging and extracellular immune signaling regulation. Although the pVII is an essential viral protein, its exact role in the virus life cycle and interplay with cellular proteins have remained to a large extent unclear. We have recently identified the cellular zinc finger protein 622 (ZNF622) as a potential pVII-interacting protein. In this study, we describe the functional consequences of the ZNF622-pVII interplay and the role of ZNF622 in the HAdV life cycle. ZNF622 protein expression increased, and it accumulated similarly to the pVII protein in the nuclei of virus-infected cells. The lack of the ZNF622 protein specifically increased pVII binding to viral DNA in the infected cells and elevated the pVII protein levels in the purified virions. In addition, ZNF622 knockout cells showed an increased cell lysis and enhanced accumulation of the infectious virus particles. Protein interaction studies revealed that ZNF622 forms a trimeric complex with the pVII protein and the cellular histone chaperon protein nucleophosmin 1 (NPM1). The integrity of this complex is important since ZNF622 mutations and NPM1 deficiency changed pVII ability to bind viral DNA. Collectively, our results implicate that ZNF622 may act as a cellular antiviral protein hindering lytic HAdV growth and limiting pVII protein binding to viral DNA.
IMPORTANCE Human adenoviruses (HAdVs) are common human pathogens causing a wide range of acute infections. To counteract viral pathogenicity, cells encode a variety of antiviral proteins and noncoding RNAs to block virus growth. In this study, we show that the cellular zinc finger protein 622 (ZNF622) interacts with an essential HAdV protein known as pVII. This mutual interaction limits pVII binding to viral DNA. Further, ZNF622 has a role in HAdV life cycle since the lack of ZNF622 correlates with increased lysis of the infected cells and accumulation of the infectious virions. Together, our study reveals a novel cellular antiviral protein ZNF622, which may impede lytic HAdV growth.
INTRODUCTION
Human adenoviruses (HAdVs) are widespread pathogens, and infections with HAdVs are associated with respiratory, gastrointestinal, and ocular diseases (1, 2). Efficient HAdV replication relies on coordinated action between early and late virus gene products. The early gene products ensure an optimal environment for HAdV DNA replication, whereas the late proteins are involved in virus genome encapsidation and virion assembly (3, 4). HAdV type 5 (HAdV-C5) protein VII (referred to here as pVII) is an abundant nuclear protein accumulating during the late phase of virus infection (5–7). The pVII is expressed as a precursor protein, and it is processed into a shorter mature form by the adenovirus protease during the final stage of virus particle maturation (8–10). The pVII is a DNA-binding protein (11, 12) that is able to condense virus DNA into core structure within the capsid (13–19). HAdV-C5 virus particles can assemble also in the absence of pVII, which suggests that pVII is not required to condense viral DNA within the capsid (20). Nevertheless, the mature pVII protein covers incoming virus DNA, remains associated with the virus genome even after nuclear entry, and acts as an essential viral protein during the early stages of infection (7, 13, 20–24). The mature pVII protein is believed to protect the virus genome from host cell DNA damage response (25, 26) and to control virus gene expression (21, 23, 27–29). The HAdV genome undergoes extensive remodeling during the virus life cycle (30). This involves partial removal of the mature pVII protein and incorporation of nucleosomal histones into the virus genome during the early phase of infection (21, 23, 31, 32). The de novo-synthesized precursor pVII protein accumulates and mediates virus genome packaging into the capsid during the late phase of infection (30, 33, 34). The cellular histones are removed from viral DNA by an unknown mechanism during the late phase of infection. This is supported by studies showing that histones are not found in the purified virus particles (5, 35). Cellular histone chaperon proteins, such as TAF-I (also known as SET) and TAF-III (also known as B23 or NPM1), interact with pVII and modulate its association with DNA (28, 36–38). However, the exact mechanisms controlling pVII assembly/disassembly on viral DNA throughout the virus life cycle have remained to a large extent enigmatic (30). The pVII protein can also associate with host cell DNA, where it blocks the release of the immune danger protein HMGB1 and inhibits DNA damage response signaling (39, 40).
Zinc finger protein 622 (ZNF622), also known as the zinc finger-like protein 9 (ZPR9), was originally identified as a substrate of murine protein serine/threonine kinase 38 (MPK38) (41). Human ZNF622 is a C2H2 zinc finger motif-containing phosphoprotein located both in the cytoplasm and in the cell nucleus (41, 42). Protein-protein interaction studies have revealed that ZNF622 interacts with the MPK38, Myb-related protein B (B-MYB), and apoptosis signal-regulating kinase 1 (ASK1) proteins (41, 43–45). MPK38 and ASK1 interaction with ZNF622 involves a disulfide linkage between cysteines, indicating a redox-dependent interaction between these proteins (43, 45). Phosphorylated ZNF622 positively regulates redox-sensitive ASK1, transforming growth factor β (TGF-β), and p53 signaling pathways, which in turn can cause enhanced apoptotic cell death (45). ZNF622 also copurifies with proteins involved in the ribosomal biogenesis; hence, it is possible that ZNF622 is involved in the 60S ribosomal subunit maturation process in mammalian cells (42).
High-throughput protein-protein interaction experiments have revealed multiple cellular proteins interacting with the pVII protein in noninfected cells (40, 46). However, only a few of these protein-protein interactions have been functionally characterized in HAdV-infected cells. Since we have previously identified ZNF622 as a potential pVII interacting protein in a yeast two-hybrid screen (46), we aimed to disclose the functional impact of this interaction in HAdV-C5 infection.
RESULTS
ZNF622 interacts with the pVII protein in HAdV-C5-infected cells.To validate ZNF622-pVII interaction in mammalian cells, a coimmunoprecipitation experiment was performed in U2OS cells infected with replication competent HAdV-pVII-Flag virus. This HAdV-C5-based virus encodes the pVII protein fused to a Flag epitope tag at the C terminus of pVII (46). The ZNF622 protein interacted with pVII-Flag, and this interaction was dependent on accumulation of the pVII-Flag protein in HAdV-C5-infected cells (Fig. 1A). The interaction was partially nucleic acid mediated since Benzonase endonuclease treatment of the immunoprecipitated protein complexes diminished the ZNF622-pVII interaction (Fig. 1A, lanes 3 and 4). In order to study which of the ZNF622 protein regions mediate binding to pVII, plasmids expressing three different ZNF622 N- and C-terminal deletion proteins were constructed (Fig. 1B). Two ZNF622 C-terminal mutant proteins, 1-108 and 1-359, did not show a detectable binding to the pVII-Flag protein (Fig. 1C, lanes 8 and 10). In contrast, the ZNF622 N-terminal mutant, 109-477, was still able to interact with the pVII-Flag protein (Fig. 1C, lane 9). These data indicate that ZNF622 C-terminal part mediates binding to the pVII protein.
HAdV-C5 pVII interacts with the ZNF622 protein. (A) HAdV-pVII-Flag (1 FFU/cell)-infected U2OS cells were harvested 48 hpi, and the pVII-Flag protein was immunoprecipitated (IP) with an anti-Flag antibody. Antibody-purified proteins were treated (lanes 2 and 4) with Benzonase endonuclease, which digests both RNA and DNA. Western blot (WB) membranes were probed with the anti-Flag and anti-ZNF622 antibodies. The ZNF622 protein binding to pVII-Flag was quantified after normalization of the ZNF622 IP values to the ZNF622 input values. The untreated sample (lane 3) was considered as 1, and data are shown as a relative ZNF622 binding to pVII-Flag in the Benzonase-treated cells (lane 4) from the presented WB experiment. The molecular weights of the ZNF622 and pVII proteins partially overlap the antibody heavy (IgH) and ligh (IgL)t chains; hence, some background staining of IgH and IgL is detectable on WB. (B) Illustration of the ZNF622 mutant proteins. The ZNF622(1-477) protein is referred to as the ZNF622(wt) protein throughout the text. Labeling of putative zinc finger domains (ZNF) is based on the Conserved Domain Database classification (72). (C) H1299 cell lysates transiently expressing the ZNF622(1-477)-HA, ZNF622(1-108)-HA, ZNF622(109-477)-HA, ZNF622(1-359)-HA, and pVII-Flag proteins were immunoprecipitated with an anti-Flag antibody. Proteins were detected with anti-HA and anti-Flag antibodies. (D) HEK293, A549, HeLa, and U2OS cells were infected with HAdV-pVII-Flag (5 FFU/cell) and harvested at 6, 12, 18, 24, and 30 hpi. WB membranes were probed with the anti-ZNF622, anti-Flag, anti-actin, and anti-HAdV capsid antibodies. The relative ZNF622 protein levels are shown below the images. The ZNF622 protein levels were normalized to the actin protein to correct for potential cell lysis and gel loading mistakes. (E) U2OS cells were infected with HAdV-pVII-Flag (5 FFU/cell) and harvested at 6, 12, 18, 24, and 30 hpi. ZNF622 mRNA expression was measured by qRT-PCR. Bars represent relative ZNF622 expression (means ± the standard deviations [SD]) after normalization to the internal control gene HPRT1 expression.
Since HAdV-C5 infection modulates host cell protein expression and accumulation (47), we investigated whether ZNF622 protein expression is altered in virus-infected cells. For this purpose, HEK293, A549, HeLa, and U2OS cell lines were infected with HAdV-pVII-Flag, and ZNF622 protein accumulation during different time points postinfection was analyzed by Western blotting. Enhanced accumulation of the ZNF622 protein was observed in all four tested cell lines during a 24- to 30-h infection period (Fig. 1D). Quantitative PCR (qPCR) analysis revealed that ZNF622 mRNA was slightly elevated during the early phase (6 to 12 h postinfection [hpi]) but not during the late phase (18 to 30 hpi) of infection in U2OS cells (Fig. 1E).
Taken together, our data indicate that ZNF622 interacts with pVII in mammalian cells and that the ZNF622 protein accumulation increased in HAdV-C5-infected cells.
Lack of the ZNF622 protein increases accumulation of the extracellular pVII-Flag protein.Since ZNF622 protein accumulation increased in the HAdV-pVII-Flag-infected cells (Fig. 1D), we hypothesized that ZNF622 may influence virus growth by altering virus protein accumulation and virus DNA replication. To test this hypothesis, we generated a U2OS cell line lacking ZNF622 protein expression (Fig. 2A, WB:ZNF622) by introducing deletion within the first exon of the ZNF622 gene using the CRISPR/Cas9 genome editing approach (48). The U2OS cell line was chosen because the genome editing can be carried out with high efficiency in this particular cell line (49). Parental, wild-type U2OS [here referred to as U2OS(wt)] and ZNF622 knockout U2OS [here referred to as U2OS(KO)] cell lines were infected with HAdV-pVII-Flag, and virus protein accumulation was monitored. Virus late proteins (pVII-Flag, pV, and capsids) did not show enhanced accumulation in U2OS(KO) cells compared to U2OS(wt) cells up to 48 hpi (Fig. 2A). Analysis of the latter time points (72 hpi) revealed an increase in the late viral protein level in U2OS(KO) cells. Since virus gene expression does not always correlate with viral DNA amount (50), we tested whether HAdV-pVII-Flag genome amplification was altered in U2OS(wt) and U2OS(KO) cells. Quantitative PCR analysis revealed that relative virus DNA accumulation was similar in the U2OS(wt) and U2OS(KO) cells (Fig. 2B), indicating that ZNF622 does not alter HAdV-pVII-Flag replication.
Lack of the ZNF622 protein increases accumulation of the extracellular pVII-Flag protein. (A) Wild-type [U2OS(wt)] and ZNF622 knockout [U2OS(KO)] cells were infected with HAdV-pVII-Flag (1 FFU/cell). Total cell lysates were analyzed by WB using antibodies against the ZNF622, Flag, HAdV capsid, pV, and actin proteins. M, mock, noninfected cells. The relative accumulation of the pVII-Flag protein is shown below the images. The pVII-Flag protein in U2OS(wt) at 48 hpi was considered as 1 and the pVII-Flag protein was normalized to the actin protein. (B) Relative virus DNA accumulation (“Virus DNA”) was established in HAdV-pVII-Flag-infected (1 FFU/cell) U2OS(wt) and U2OS(KO) cells by quantification of the virus E1B-55K gene amplicons with qPCR. Bars denote the means ± the standard errors of the mean (SEM) from two independent experiments. (C) Accumulation of the extracellular pVII-Flag and cellular histone H3 proteins. Cell growth medium collected from samples shown in panel (A), separated by SDS-PAGE, and analyzed by WB with the anti-Flag and anti-H3 antibodies. Control (Ctrl) denotes U2OS(wt) total cell lysates infected with HAdV-pVII-Flag (72 hpi). An asterisk (*) indicates migration of the precursor pVII, and an arrowhead indicates migration of the processed mature pVII protein.
The small increase in viral proteins (Fig. 2A, 72 hpi) indicated that the lack of the ZNF622 protein might have a positive impact on virus life cycle. HAdV-C5 virions are released from infected cells by cell lysis (51) and can be detected by different cell biology methods. Hence, an increase in extracellular virus particles should correlate with enhanced detection of the capsid proteins in the cell growth medium (i.e., extracellular space). To test this, we monitored the pVII-Flag protein in the cell growth medium from HAdV-pVII-Flag-infected U2OS(wt) and U2OS(KO) cells. The pVII-Flag protein was detected at 72 and 96 hpi in U2OS(KO), but not in U2OS(wt), cell growth medium under our experimental conditions (Fig. 2C). Further analysis of the same experiment indicated that the detected protein corresponds to the mature pVII protein, which is the protein form found in correctly processed infectious virus particles (5). Extracellular histone H3 protein, which can be released from the virus-infected cells (52), was comparable in U2OS(wt) and U2OS(KO) cells.
Together, these data suggest that the lack of ZNF622 does not alter virus DNA amplification but increases accumulation of the extracellular pVII-Flag protein.
Lack of the ZNF622 protein enhances infected cell lysis and infectious virus accumulation.When analyzing HAdV-pVII-Flag infection kinetics (Fig. 2), we noticed a clear visual difference between infected U2OS(wt) and U2OS(KO) cells. In the case of U2OS(KO), the majority of cells were round, formed grape-like clusters, and detached from the cell culture plate at 72 hpi, whereas U2OS(wt) cells did not show a similar behavior (Fig. 3A). This observation suggested enhanced virus production in U2OS(KO) cells, since cell rounding, grape-like clusters, and detachment are typical marks of cytopathic infections (53, 54). These observations, together with detection of the pVII-Flag protein in cell growth medium (Fig. 2C), raised a possibility that infected U2OS(KO) cells are lysed faster than U2OS(wt) cells. The cell lysis was analyzed with the trypan blue exclusion assay to determine cell viability within the infected cell population. As shown in Fig. 3B, viability of infected U2OS(KO) cells decreased significantly at 72 and 96 hpi compared to U2OS(wt) cells.
Lack of the ZNF622 protein enhances infected cell lysis and infectious virion accumulation. (A) U2OS(wt) and U2OS(KO) cells were infected with HAdV-pVII-Flag (1 FFU/cell) and phase-contrast cell images were taken at 24, 48, 72, and 96 hpi. Scale bar, 200 μm. (B) The viability of the HAdV-pVII-Flag-infected cells (1 FFU/cell) was assayed with trypan blue exclusion assay. Bars denote the means ± the SD from two independent experiments. Unpaired t test was used to calculate the statistical significance (*, P < 0.05). (C) Purification of HAdV-pVII-Flag from U2OS(wt) and U2OS(KO) cells with single CsCl step gradient centrifugation. L, light-density particles; H, heavy-density particles; H/L, ratio of H to L band intensities established with ImageJ software (73). (D) Equal protein amount of the heavy- and light-density particles were loaded onto SDS-PAGE and analyzed by WB. Proteins were visualized with the anti-Flag and anti-capsid (detecting the hexon protein) antibodies. Relative pVII-Flag protein levels are shown after normalization to the hexon protein and considering the pVII-Flag/hexon ratio as 1 in the U2OS(wt) samples. Asterisk (*), precursor pVII; arrowhead mature pVII. (E) Titers of heavy-density particles from panel C. Physical titer (VP/ml) was established with qPCR, whereas the infectious titer (FFU/ml) with an anti-capsid antibody staining in infected 911 cells. The numbers show the ratio (fold change) between respective virus titers in the U2OS(KO) and U2OS(wt) cells. (F) Attached HAdV-pVII-Flag-infected (1 FFU/cell) cells were collected, lysed, and used to infect 911 cells to establish intracellular infectious virus titer (FFU/ml). Bars denote the mean virus titer ± the SD from two experiments. (G) Cell culture medium from the experiments on panel F was used to establish extracellular infectious virus titer (FFU/ml). The numbers show the virus titer ratio (fold change) between the U2OS(KO) and U2OS(wt) cells. An unpaired t test indicated a significant reduction in the formation of infectious virus particles in U2OS(wt) cells compared to U2OS(KO) cells (*, P < 0.05; **, P < 0.01; ***, P < 0.001).
Typically, HAdV preparation consists of a mixture of infectious mature (i.e., heavy particles) and noninfectious defective (i.e., light particles) virions, which can be separated from each other by virtue of their different densities in cesium chloride (CsCl) gradient (55, 56). Both, heavy- and light-density bands were detected when HAdV-pVII-Flag was purified from the U2OS(wt) and U2OS(KO) cell lysates in CsCl step gradient (Fig. 3C). Based on visual observation, the heavy-density band was more intense and sharper in the U2OS(KO) sample compared to the U2OS(wt) sample. Further, the light-density band was more diffused in the U2OS(wt) compared to the U2OS(KO) sample. To better elucidate the heavy- to light-density virion ratio (H/L), the intensities of the heavy- and light-density bands were quantified using ImageJ software (Fig. 3C). The H/L values were 0.8 in the U2OS(wt) sample and 1.4 in the U2OS(KO) sample. This quantification reinforces that U2OS(wt) cells contain more light-density virions compared to heavy-density virions, while U2OS(KO) cells contain more heavy-density virions compared to light-density virions.
The purified virus preparations were further analyzed for the presence of the pVII-Flag and viral hexon proteins. The mature pVII-Flag protein showed a very weak detection in the purified mature virions from U2OS(wt) cells, whereas it was clearly enriched in the purified mature virions isolated from U2OS(KO) cells (Fig. 3D, lanes 1 and 2). Normalization to the viral hexon protein showed 6.3-fold enrichment of the pVII-Flag protein in the U2OS(KO) sample compared to the U2OS(wt) sample (Fig. 3D, lanes 1 and 2). The light-density particles are considered virus assembly intermediates containing unprocessed and processed viral precursor proteins (56). This was also true in our virus preparations, since the pVII was detected as a slow-migrating precursor protein in the light-density band (Fig. 3D, lanes 3 and 4).
In accordance with the heavy band intensity (Fig. 3C), the physical and infectious virus titers were higher for HAdV-pVII-Flag isolated from the U2OS(KO) cells (Fig. 3E). The established physical virus titers showed that 1.9-fold more virus was recovered from the U2OS(KO) cells compared to U2OS(wt) cells (Fig. 3E). Interestingly, the infectious virus titers showed that heavy-density particles isolated from U2OS(KO) cells were 7.1-fold more infectious compared to the U2OS(wt) sample. Next, we calculated the ratio between infectious and physical titers for the virus purified from the same [e.g., U2OS(wt) or U2OS(KO)] cell line. This calculation revealed that HAdV-pVII-Flag purified from U2OS(wt) and U2OS(KO) cells contained approximately 3 and 13% infectious virus particles, respectively. Together, our data indicate that the lack of ZNF622 increases formation of the infectious heavy-density virus particles, and this correlates with the mature pVII protein enrichment in the purified virus particles.
The enhanced cell lysis (Fig. 3B) might indicate that HAdV-pVII-Flag is released more efficiently from U2OS(KO) compared to U2OS(wt) cells. If the virus was released more efficiently from U2OS(KO) cells, it should correlate with a higher virus accumulation in the extracellular space, while intracellular production of HAdV-pVII-Flag should remain similar in the U2OS(wt) and U2OS(KO) cells. To test this hypothesis, we collected both infected cells attached to the cell culture plastic (i.e., intracellular virus) and cell growth medium (i.e., extracellular virus) and established the infectious virus titers in 911 cells. As shown in Fig. 3F, intracellular virus significantly increased in U2OS(KO) cells compared to U2OS(wt) cells during three different time points. Similar to the intracellular virus, the extracellular virus was significantly higher in the U2OS(KO) cell growth medium (Fig. 3G). These data suggest that the higher level of extracellular HAdV-pVII-Flag corresponds to increased virus accumulation in U2OS(KO) cells.
ZNF622 knockout enhances pVII protein binding to viral DNA.Since ZNF622 interacts with pVII (Fig. 1A and C) (46), we were interested to test whether ZNF622 knockout alters pVII function as the DNA-binding protein. Therefore, the pVII protein binding to viral DNA (Fig. 4A) was analyzed with a chromatin immunoprecipitation (ChIP) assay in HAdV-pVII-Flag-infected cells at 48 hpi. This particular time point was chosen because the pVII-Flag protein accumulation pattern was very similar in the U2OS(wt) and U2OS(KO) cells (Fig. 2A). Binding of the pVII-Flag protein to viral DNA clearly increased in infected U2OS(KO) cells compared to U2OS(wt) cells (Fig. 4B). Analysis of five different HAdV-C5 genome regions revealed that pVII-Flag interaction was divergent, with the highest level of interaction detected at the E1A gene promoter and the lowest at the end of major late transcription unit L2 (L2 unit). In addition, our qPCR data indicated that the cellular histone H3 binding was also enhanced on viral DNA in infected U2OS(KO) cells, although not to the same extent as it was observed for the pVII-Flag protein (Fig. 4B). Analysis of the histone H3 and pVII-Flag binding at the cellular GAPDH (glyceraldehyde-3-phosphate dehydrogenase) promoter confirmed the ChIP assay specificity. As shown in Fig. 4B (GAPDH promoter), the pVII-Flag protein binding was undetectable at the cellular GAPDH promoter, whereas histone H3 was similarly enriched in both tested cell lines.
Elimination of the ZNF622 protein expression enhances pVII binding to viral DNA. (A) Simplified schematic overview of the HAdV-C5 genome with indicated ChIP assay qPCR amplicons (horizontal red lines). (B) Binding of the pVII-Flag and histone H3 proteins to the virus genome in U2OS(wt) and U2OS(KO) cells at 48 hpi analyzed by a ChIP assay. Protein binding was analyzed at the virus early E1A gene promoter (E1A promoter), the end of the E1A gene (E1A gene), the major late transcription unit promoter (MLTU promoter), the end of the L2 transcription unit (L2 unit), and the end of the L4 transcription unit (L4 unit). The cellular GAPDH promoter (GAPDH promoter) was used as the specificity control. Ctrl, no-antibody-added samples. Bars denote the mean % of input ± the SEM from three independent experiments. An unpaired t test was used to calculate the statistical significance (***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant).
Collectively, our results imply that lack of the ZNF622 protein enhances the overall pVII-Flag binding to viral DNA.
Reintroduction of ZNF622 affects virus growth in U2OS(KO) cells.One of the concerns when using the CRISPR/Cas9 genome editing approach is the risk for off-target genome modification by the Cas9 nuclease. Although we present the data using one particular ZNF622(KO) cell clone, several other isolated single cell ZNF622(KO) clones showed very similar phenotypic behavior after being infected with HAdV-pVII-Flag (data not shown). To further confirm that lack of ZNF622 per se enhanced HAdV-pVII-Flag lytic growth, we reintroduced exogenous wild-type ZNF622-HA(wt) cDNA into U2OS(KO) cells. Since the ZNF622 C terminus is involved in binding to pVII (Fig. 1C), we generated a stable cell line expressing ZNF622 C-terminal deletion mutant [KO+ZNF622-HA(1-359)]. Similarly, we generated a cell line expressing ZNF622 N-terminal deletion mutant [KO+ZNF622-HA(109-477)], which is still able to interact with pVII (Fig. 1C) but lacks the defined zinc finger motifs (Fig. 1B). As a control, a cell line stably transfected with an empty pcDNA3.1 plasmid (KO+pcDNA) was used. All three ZNF622 proteins (wt, 1-359, and 109-477) were expressed in HAdV-pVII-Flag-infected cells, and no obvious alterations in the pVII-Flag protein accumulation were observed at 48 hpi (Fig. 5A, lanes 5, 8, and 11). Similar to the data shown in noninfected cells (Fig. 1C), the ZNF622(wt)-HA and ZNF622(109-477)-HA proteins interacted with pVII-Flag, whereas the ZNF622(1-359)-HA was deficient in this function in HAdV-pVII-Flag-infected cells (Fig. 5B, lanes 4 to 6). Further, visual observations revealed that KO+ZNF622-HA(wt) cells were better attached to the cell culture plate and did not show extensive cell detachment compared to KO+pcDNA, KO+ZNF622-HA(1-359), and KO+ZNF622-HA(109-477) cells (Fig. 5C).
Reintroduction of exogenous ZNF622 into U2OS(KO) cells. (A) U2OS(KO+pcDNA), U2OS(KO+ZNF622-HA(wt)), U2OS(KO+ZNF622-HA(1-359)), and U2OS(KO+ZNF622-HA(109-477)) cell lines were infected with HAdV-pVII-Flag (1 FFU/cell), and total cell lysates were prepared at 48 and 96 hpi. WB analyses were performed with anti-HA, anti-Flag, and anti-actin antibodies. Migration of the ZNF622-HA proteins is indicated by arrowheads. (B) The aforementioned stable U2OS cell lines expressing different ZNF622 proteins were infected with HAdV-pVII-Flag (5 FFU/cell), followed by immunoprecipitation of the pVII-Flag protein with an anti-Flag antibody at 48 hpi. (C) The indicated cell lines were infected with HAdV-pVII-Flag (1 FFU/cell), and phase-contrast cell images were taken at 48, 72, and 96 hpi. Scale bar, 200 μm.
Previous reports have shown that the ZNF622 protein is located both in the cytoplasm and in the cell nucleus (41, 42). Therefore, we tested whether subcellular localization of the ZNF622 protein is altered in the HAdV-pVII-Flag-infected U2OS cells. The ZNF622(wt)-HA protein showed mainly cytoplasmic localization with specific staining at the nuclear rim in noninfected cells (Fig. 6A). This staining pattern changed in the infected cells, where the protein had partially lost nuclear rim staining and showed diffused localization both in the cytoplasm and in the nucleus (Fig. 6B). The ZNF622(1-359)-HA protein showed a staining pattern similar to the ZNF622(wt)-HA protein (Fig. 6C and D). Of note is the ZNF622(109-477)-HA protein, which showed diffused nuclear and cytoplasmic localization in both noninfected and infected U2OS cells (Fig. 6E and F).
Subcellular localization of the ZNF622 proteins in infected cells. U2OS cell lines constitutively expressing the ZNF622(wt)-HA, ZNF622(1-359)-HA, or ZNF622(109-477)-HA proteins were infected with HAdV-pVII-Flag (1 FFU/cell), or left noninfected, for 48 h. Indirect immunofluorescence was performed with the anti-HA and anti-Flag antibodies to detect the ZNF622-HA (green) and pVII-Flag (red) proteins. Nuclei were counterstained with DAPI (blue). Scale bar, 50 μm.
Reintroduced full-length ZNF622 reduces pVII binding to virus genome.Since the lack of ZNF622 protein (Fig. 4) enhanced pVII-Flag binding to viral DNA in U2OS(KO) cells, we hypothesized that reintroduction of ZNF622 into this cell line will decrease pVII-Flag binding to viral DNA. For that purpose, the ChIP assay was carried out with the aforementioned U2OS(KO) cell lines expressing different ZNF622 mutants (Fig. 5A). Expression of ZNF622-HA(wt) attenuated pVII-Flag binding to viral DNA irrespectively of the analyzed virus DNA region (Fig. 7A). These results are in line with our previous observations (Fig. 4B) and confirm that lack of ZNF622 enhances pVII-Flag binding to viral DNA. In contrast, pVII-Flag DNA binding was not affected in cells expressing the ZNF622 mutant [ZNF622-HA(1-359) and ZNF622-HA(109-477)] proteins, and it was comparable to what was observed in cells lacking ZNF622 expression (i.e., KO+pcDNA) (Fig. 7A). Similarly to the experiments performed in U2OS(wt) and U2OS(KO) cells (Fig. 4B), the most drastic effect on pVII-Flag DNA binding was detected at the E1A promoter, whereas pVII-Flag binding at the L2 unit was less affected (Fig. 7A). The ZNF622 protein contains C2H2 zinc finger motifs (Fig. 1B), which are putative DNA-binding modules (57). This prompted us to analyze whether the reintroduced ZNF622(wt)-HA could interact with viral DNA. However, we were unable to detect specific ZNF622(wt)-HA binding to viral DNA under our experimental conditions (Fig. 7B). Since the accumulation of infectious virions was enhanced in U2OS(KO) cells (Fig. 3), we hypothesized that reintroduction of the ZNF622(wt)-HA protein expression in U2OS(KO) cells might reduce infectious virus progeny in this particular cell line. Indeed, infectious intracellular virus was significantly reduced in the KO+ZNF622-HA(wt) cells compared to KO+pcDNA cells (Fig. 7C) at 72 and 96 hpi. In contrast, the formation of infectious virus progeny was similar in KO+ZNF622-HA(1-359), KO+ZNF622-HA(109-477), and KO+pcDNA cells under the same experimental conditions (Fig. 7C).
Reintroduced ZNF622(wt) reduces pVII binding to virus genome. (A) Binding of pVII-Flag to viral DNA in the ZNF622-HA-expressing cells. The U2OS(KO) cells constitutively expressing different versions of the ZNF622 proteins were infected with HAdV-pVII-Flag, and the ChIP assay was performed with an anti-Flag antibody at 48 hpi. Bars denote the mean % of input ± the SEM from two experiments. Ctrl denotes no antibody added sample. (B) Binding of the ZNF622-HA(wt) to viral DNA (E1A promoter and L4 unit) in U2OS(KO+pcDNA) and U2OS(KO+wt-HA) cells infected with HAdV-pVII-Flag (1 FFU/cell) at 48 hpi. A ChIP assay was performed with an anti-HA antibody. The data are from a single experiment, analyzed in duplicate. (C) Total cell lysates from respective cell lines infected with HAdV-pVII-Flag (1 FFU/cell, 72 and 96 hpi) were used to infect 911 cells to establish HAdV-pVII-Flag virus titer (FFU/ml). Bars denote the mean virus titer ± the SD from two experiments. Unpaired t test indicated a significant virus titer reduction in U2OS(KO+wt-HA) cells compared to U2OS(KO+pcDNA) cells (*, P < 0.05; ***, P < 0.001).
Together, our results confirm that the full-length ZNF622 protein has a negative effect on pVII binding to viral DNA, which correlates with reduced accumulation of infectious virus progeny.
ZNF622 forms a complex with the NPM1 protein in HAdV-pVII-Flag-infected cells.Although the ZNF622(109-477) mutant protein localized to the nucleus (Fig. 6F) and was able to bind to the pVII-Flag protein (Fig. 1C and 5B), it did not affect pVII binding to viral DNA (Fig. 7A). These puzzling observations suggested us that the N terminus of the ZNF622 protein might provide an interaction surface for other protein(s), which might influence pVII-Flag loading to viral DNA. A previous study has shown that a known cellular histone chaperon nucleophosmin 1 (NPM1) (58) can inhibit pVII loading to virus DNA in the infected HeLa cells (37). Therefore, we hypothesized that ZNF622 might interact with the NPM1 protein and thereby influence pVII-Flag DNA binding. Based on our coimmunoprecipitation experiment (Fig. 8A), the endogenous ZNF622 protein specifically interacts with NPM1 in U2OS(wt) cells. Further, the ZNF622(109-477)-HA mutant protein showed reduced interaction with NPM1 compared to the ZNF622(wt)-HA protein in U2OS(wt) cells (Fig. 8B, lanes 2 and 4). NPM1 is an abundant nucleolar protein (59). To study whether the ZNF622 and NPM1 proteins colocalize in the nucleoli, U2OS stable cell lines expressing the ZNF622-HA proteins were transiently transfected with plasmid expressing the Myc-tagged NPM1 protein. As shown in Fig. 8C, both ZNF622(wt)-HA and Myc-NPM1 were detected in the cell nucleus, with overlapping staining of nucleoli. In contrast, the ZNF622(109-477)-HA protein did not show colocalization with Myc-NPM1 in the nucleolus. Interestingly, the transfected Myc-NPM1 did not show a typical nucleolar staining pattern in the majority of the KO+ZNF622(109-477) and KO+pcDNA cells. It is important to remember that stable cell lines expressing the ZNF622-HA proteins were generated using U2OS(KO) cells. Hence, it is possible that lack of ZNF622 (e.g., KO+pcDNA) or presence of NPM1-binding deficient ZNF622 [e.g., KO+ZNF622(109-477)-HA] alter the typical nucleolar localization pattern of the Myc-NPM1 protein. Note that nucleolar localization of the Myc-NPM1 protein was not changed in the cells expressing NPM1-binding ZNF622 proteins (e.g., wt and 1-359).
ZNF622 interacts with the pVII and NPM1 proteins. (A) The ZNF622 protein was immunoprecipitated from U2OS(wt) and U2OS(KO) cell lysates with an anti-ZNF622 antibody. WB with anti-NPM1 and anti-ZNF622 antibodies. (B) Total cell lysates from U2OS(KO) stable cell lines expressing different versions of the ZNF622-HA proteins were immunoprecipitated with an anti-HA antibody. WB analysis was performed with anti-HA and anti-NPM1 antibodies. (C) U2OS stable cell lines [KO+ZNF622(wt)-HA, KO+ZNF622(1-359)-HA, KO+ZNF622(109-477)-HA, and KO+pcDNA] were transiently transfected with the plasmid expressing the Myc-NPM1 protein for 48 h. Indirect immunofluorescence was performed with the anti-HA and anti-Myc antibodies to detect the ZNF622-HA (green) and Myc-NPM1 (red) proteins. Nuclei were counterstained with DAPI (blue). Scale bar, 50 μm. (D) U2OS(wt) and U2OS(KO) cells were infected with HAdV-pVII-Flag (1 FFU/cell), and the pVII-Flag protein was immunoprecipitated at 48 hpi. The NPM1 protein binding to pVII-Flag was quantified after normalization of the NPM1 IP values to NPM1 input values. U2OS(KO) sample (lane 3) was considered as 1, and data are shown as the relative NPM1 binding to pVII-Flag in U2OS(wt) cells (lane 4) from the presented WB experiment. (E) U2OS(wt) and U2OS(KO) were transfected with scrambled siRNA (siScr) or NPM1 siRNA (siNPM1) for 24 h, followed by HAdV-pVII-Flag infection (48 hpi, 1 FFU/cell). Total cell lysates were analyzed by WB with the indicated antibodies. Note that there is a very weak background staining of the ZNF622 protein in U2OS(KO) cells, probably due to contamination with wild-type cells in this particular experiment. (F) Binding of pVII-Flag to viral DNA in siNPM1-treated cells. ChIP assay with an anti-Flag antibody on cell lysates treated as described in panel D. The data are shown as the fold change of pVII-Flag binding to viral DNA considering siScr-treated U2OS(wt) samples as 1 (lanes 1 and 5). Since the Ctrl samples (i.e., beads only) showed very low signal, they were excluded from the graph. Bars denote means ± the SEM from two experiments. Unpaired t test indicated a significant increase of pVII-Flag DNA binding in siNPM1-treated cells compared to siScr-treated cells (****, P < 0.0001; **, P < 0.01). (G) U2OS(wt) cells were treated exactly as described in panel D. The pVII-Flag protein was immunoprecipitated from the total cell lysates, and the interacting proteins were revealed by WB.
Since NPM1 has been shown to interact with pVII (38), we tested whether the NPM1 protein copurifies together with the pVII and ZNF622 proteins in virus-infected cells. Both NPM1 and ZNF622 were detected when the pVII-Flag protein was immunoprecipitated from virus-infected U2OS cells (Fig. 8D, lanes 3 and 4). Interestingly, binding between the NPM1 and pVII proteins was stronger in the U2OS(wt) cells compared to the U2OS(KO) cells (Fig. 8D, see quantification).
To better understand the interplay between ZNF622, pVII and NPM1, we analyzed the impact of NPM1 on pVII DNA binding. For this purpose, NPM1 expression was knocked down using the RNAi approach (Fig. 8E) and pVII-Flag binding to viral DNA was analyzed with a ChIP assay. As shown in Fig. 8F, pVII-Flag DNA binding significantly increased at two analyzed virus genome regions (MLTU and E1A promoters) in siNPM1-treated U2OS(wt) cells (lanes 1 and 2 and lanes 5 and 6). In contrast, pVII-Flag binding to virus genome was not enhanced, but rather nonsignificantly decreased in U2OS(KO) cells (Fig. 8F). The latter might be due to reduced pVII-Flag protein expression in this siNPM-treated cell line (Fig. 8E, lanes 3 and 6). The pVII protein binds to both, the ZNF622 and NPM1 proteins (Fig. 8D). Therefore, we tested whether the NPM1 protein mediates pVII interaction with ZNF622. siNPM1 treatment eliminated NPM1 from the pVII-Flag protein complex, but it did not affect ZNF622 binding to the pVII-Flag protein (Fig. 8G). This indicates that the pVII-ZNF622 interaction is not mediated by the NPM1 protein.
Together, our results suggest that ZNF622 interacts with NPM1 and that they form a functional trimeric protein complex with pVII in HAdV-C5-infected U2OS cells.
The ZNF622 protein associates with chromatin in HAdV-C5-infected cells.In order to better understand how ZNF622 limits HAdV-C5 growth, we decided to study the subcellular localization of the protein in virus-infected cells. Biochemical fractionation of the U2OS(wt) cells revealed that the ZNF622 protein localized mainly in the cytoplasm and also to some extent in the cell nucleus (Fig. 9A, lanes 1 and 2). The same fractionation approach in virus-infected cells showed elevated accumulation of the ZNF622 protein in the nuclear fraction and confirmed the presence of the pVII-Flag protein in the same subcellular fraction (Fig. 9A, lanes 3 and 4). This biochemical fractionation experiment confirms our previous immunofluorescence data (Fig. 6A and B) about the subcellular localization of the ZNF622 protein. Nuclear proteins often associate with the chromatin, which alters their solubility (60, 61). Therefore, we performed successive salt fractionation of infected U2OS(wt) cell nuclei to elucidate the intranuclear localization of ZNF622. This biochemical approach allows isolation of different soluble chromatin fractions (e.g., highly accessible chromatin, condensed chromatin) after micrococcal nuclease (MNase) treatment in the presence of increasing Na+ concentrations (60, 61). Proteins weakly interacting with DNA become soluble at a low NaCl concentration (e.g., 75 mM), whereas proteins tightly bound to DNA become soluble at a high NaCl concentration (e.g., 600 mM) (61). The nuclear ZNF622 protein localized mainly in chromatin fractions solubilized with 150 to 600 mM NaCl, which overlaps with the pVII-Flag localization pattern (Fig. 9B, lanes 8 to 10). Slight increase of the NPM1 protein was detected in the high salt fractions of the infected nuclei compared to noninfected nuclei. This overlaps with the ZNF622 protein localization pattern. The cellular HMGB1 and HMGB2 proteins showed a drastic alterations in HAdV-C5-infected nuclei (Fig. 9B, WB:HMGB1 and HMGB2), as reported previously (39, 40). Hence, these observations imply that ZNF622 chromatin association is an HAdV-C5 infection-specific event.
The ZNF622 protein associates with chromatin in virus-infected cells. (A) U2OS cells were infected with HAdV-pVII-Flag (5 FFU/cell), and biochemical cell fractionation was performed at 36 hpi. Nuclear (N) and cytoplasmic (C) cell fractions were analyzed with WB using the indicated antibodies. Detection of the GAPDH and histone H3 proteins confirms the purity of the cytoplasmic and nuclear fractions, respectively. (B) U2OS cells were infected with HAdV-pVII-Flag (5 FFU/cell), and successive salt fractionation (75, 150, 300, and 600 mM NaCl) on MNase-treated cell nuclei was done at 36 hpi. WB analysis was performed with the anti-ZNF622, anti-Flag, anti-HMGB1, anti-HMGB2, anti-NPM1, and anti-histone H3 antibodies. T, total cell lysate. The arrowhead indicates migration of the ZNF622 protein. (C) U2OS(wt) and U2OS(KO) cells were infected with HAdV-pVII-Flag (5 FFU/cell), and salt fractionation and WB analysis were done as in panel B. (D) H1299 cells were transiently transfected with plasmid expressing the precursor pVII-Flag protein, and successive salt fractionation was done 36 h posttransfection. WB analyses were performed with anti-ZNF622, anti-Flag, and anti-histone H3 antibodies.
Since both pVII-Flag and ZNF622 showed overlapping chromatin localization patterns (Fig. 9B), we tested whether ZNF622 influences pVII-Flag chromatin localization. Successive salt fractionation of U2OS(KO) nuclei revealed that lack of ZNF622 did not detectably change subnuclear localization of pVII-Flag (Fig. 9C). Similarly, nuclear localization of the control proteins, HMGB1 and HMGB2, was not altered in U2OS(KO) cells. The NPM1 protein showed reduced association with chromatin in U2OS(KO) cells (Fig. 9C, WB:NPM1), suggesting that accumulation and/or stabilization of the NPM1 protein in HAdV-C5-infected cells is partially regulated by the ZNF622 protein. Since ZNF622 interacts with pVII-Flag (Fig. 1), we tested whether this interaction per se is enough for ZNF622 chromatin localization. The successive salt fractionation of H1299 cell nuclei transiently expressing the pVII-Flag protein revealed that ZNF622 does not accumulate in chromatin fractions in these cells (Fig. 9D).
Collectively, our results suggest that the ZNF622 protein accumulates in the chromatin in HAdV-C5-infected cells.
DISCUSSION
Although pVII association with viral DNA was proposed more than four decades ago (14, 15), the exact details about how this histone-like protein is assembled and maintained on viral DNA have remained to a large extent enigmatic (30). The present study reveals a cellular protein, ZNF622, which influences pVII interaction with viral DNA.
HAdV-C5 infection increased the ZNF622 protein level and enhanced its nuclear accumulation. The ZNF622 mRNA showed reduction after 18 hpi, whereas the protein started to accumulate from the same time point in the U2OS cells (Fig. 1D and E). These observations suggest that the ZNF622 protein accumulation is probably not due to enhanced gene transcription. The exact mechanism(s) of how the ZNF622 protein accumulates in the infected cell nucleus remains unclear. Since transient expression of pVII alone did not determine ZNF622 nuclear localization (Fig. 9D), it is possible that additional viral proteins are involved in this process. Potential candidates are other known pVII-interacting viral proteins, such as pV and L1 52/55 kDa (38, 62), which together with pVII might lead to nuclear localization of ZNF622. An alternative possibility is that site-specific phosphorylation may increase ZNF622 nuclear accumulation in virus-infected cells. This is supported by the study showing that transient overexpression of MPK38 causes ZNF622 accumulation in the nucleus (41).
We envision two possible scenarios how ZNF622 influences pVII interaction with viral DNA in the infected cell nucleus (Fig. 10). Notably, these scenarios are not mutually exclusive. One possible scenario is that ZNF622 forms an inhibitory protein-protein complex with pVII, thereby limiting pVII binding to viral DNA. This is supported by our ChIP assays, where elimination of the ZNF622 protein significantly enhanced pVII binding to viral DNA (Fig. 4B). Further, the mutant ZNF622(1-359) protein, which is defective in pVII binding, did not reduce pVII-DNA interaction (Fig. 5B and 7A). Hence, the ZNF622 protein may function as a sponge by physically keeping the pVII protein away from binding to viral DNA. The ZNF622 protein is a potential DNA-binding protein, by virtue of the C2H2 motifs (Fig. 1B). We considered an option that ZNF622 directly binds to viral DNA and thereby obstructs pVII DNA binding. However, we did not detect ZNF622 binding to viral DNA (Fig. 7B). Thus, it is likely that ZNF622 does not compete with pVII for the same binding sites on viral DNA.
Proposed model of ZNF622, NPM1 and pVII interplay in HAdV-C5-infected U2OS cells. ZNF622 (ZNF622WT) may form a protein complex by binding to the NPM1 and pVII proteins. This inhibitory protein complex will physically keep pVII away from DNA and remove pVII from viral DNA with the help of the NPM1 protein. Alternatively, ZNF622 sequesters the NPM1 and pVII proteins to cellular DNA to impede pVII binding to viral DNA. Lack of ZNF622 (ZNF622KO) correlates with enhanced pVII binding to viral DNA and enhanced formation of the infectious virions.
The negative effect of ZNF622 on pVII DNA binding can be only partially explained by a direct physical interaction with the pVII protein. The results showing that ZNF622(109-477) was still able to interact with pVII (Fig. 1C and 5B), but did not block pVII binding to viral DNA (Fig. 7A), pointed toward a more complex mechanism. In this regard the observation that ZNF622 interacts with NPM1 can explain the versatile inhibitory function of ZNF622. NPM1 is a multifunctional protein, and it can act as the histone chaperone protein (58). The NPM1 protein associates with core and linker histones, mediates histone assembly into chromatin and controls spatial organization of nucleolus-associated heterochromatin (59, 63). Previous studies have shown that knockdown of NPM1 reduces infectious virus particle production and enhances pVII binding to viral DNA (37, 64). Therefore, it has been proposed that NPM1 regulates assembly/disassembly of pVII with the virus genome in HAdV-C5-infected HeLa cells (37). Together with our experiments (Fig. 8), it is likely that NPM1 is involved in pVII removal from virus genome (see below).
Elimination of either NPM1 (37) or ZNF622 (this study) enhances pVII binding to viral DNA. However, it is important to mention that the consequences of pVII enrichment on viral DNA are different. In our study the lack of the ZNF622 protein significantly increased formation of infectious virus progeny (Fig. 3 and 7C), whereas in the aforementioned studies (37, 64) the NPM1 deficiency correlated with reduced formation of the infectious virus particles. We provide two interpretations to clarify these opposing results.
First, pVII-Flag DNA binding in NPM1 knockdown U2OS(wt) cells did not reach the same level as was observed in the U2OS(KO) cells (Fig. 8F, lanes 2 and 3 and lanes 6 and 7). In addition, NPM1 knockdown in U2OS(KO) cells did not show any additive effect on pVII DNA binding, indicating that ZNF622 has a more prominent role in pVII-Flag DNA binding compared to NPM1. A likely scenario here is that ZNF622 functions as a scaffold to keep the NPM1 and pVII proteins in a trimeric protein complex (Fig. 10). The ZNF622 as a scaffold protein can both physically seclude the pVII protein and keep NPM1 active as the histone chaperon protein to remove the pVII protein from the viral DNA. The importance of the ZNF622-NPM1 connection is further strengthened by our immunofluorescence data showing that presence of the ZNF622(wt) protein can influence the nucleolar localization of NPM1 (Fig. 8C).
Second, we show a significant loss of cell viability due to cell lysis in ZNF622(KO) cells. This might indicate that the ZNF622 protein has additional, cell lysis-associated functions in the infected cells. Notably, we have not detected enhanced lysis of the HAdV-infected NPM1 knockdown cells, and it has not been reported in other studies (37, 64). Surprisingly, little is known about how HAdV-infected cells are lysed. It has been proposed that HAdV-mediated cell lysis is achieved via the mechanisms that resemble necrosis-like programmed cell death and autophagic cell death (65, 66). Interestingly, the ZNF622 protein has been shown to enhance p53- and TGF-β-mediated apoptosis (45). However, our data show that ZNF622 deficiency, and not its presence, correlates with increased cell lysis. Hence, the observed infected cell lysis (Fig. 3) cannot be assigned to the reported ZNF622 apoptotic functions (45).
We also provide an alternative scenario how ZNF622 limits pVII binding to viral DNA. This is based on the observation that the ZNF622 protein associates with particular chromatin fractions (Fig. 9). Hence, it is possible that ZNF622 sequesters pVII and NPM1 to cellular DNA as part of a potential antiviral mechanism (Fig. 10). This is supported by our experiments showing that the ZNF622, pVII, and NPM1 proteins form a complex and that they have an overlapping chromatin localization pattern (Fig. 8 and 9). Also, the observation that the ZNF622 and pVII protein-protein interaction was partially mediated by nucleic acids (Fig. 1A) reinforces the proposed model (Fig. 10). A possible protein sequestration scenario is also supported by a recent study showing that pVII association with the cellular HMGB1 protein on host chromatin blocks HMGB1 antiviral function (40).
The ZNF622 protein deficiency increases accumulation of the infectious virus progeny. This increase was not due to boosted virus DNA replication or massive capsid protein production (Fig. 2). However, it correlated with enhanced pVII binding to viral DNA and with elevated accumulation of the mature pVII protein in the heavy-density virions (Fig. 3 and 4). It is possible that the pVII protein enrichment may shield virus genome from various DNA damage response pathways in the HAdV-C5-infected cells. This is supported by studies showing that the pVII protein can indeed protect viral DNA against an activated cellular DNA damage response pathways (25, 39). An alternative explanation is that elevated pVII DNA binding can be beneficial for the packaging of the viral DNA into infectious virions. Hypothetically, accumulation and increased DNA binding of the mature pVII protein might result in a denser compression of viral DNA, which in turn can facilitate the formation of fully infectious virus particles (i.e., heavy-density particles). However, the essential role of pVII on HAdV DNA packaging has been questioned by a recent study (20).
In conclusion, the present study has revealed a potential cellular antiviral protein ZNF622, which limits binding of the pVII protein to viral DNA and hinders HAdV-C5 lytic growth.
MATERIALS AND METHODS
Plasmids.Human full-length ZNF622 (NM_033414.2) cDNA (GenScript) was cloned into pcDNA3.1-HA plasmid (46) to express the C-terminal HA-tagged ZNF622 protein. ZNF622 mutants (1-108, 109-477, and 1-359) were generated with the PCR-mediated DNA deletion approach and cloned into the pcDNA3.1-HA background vector. Myc-tagged NPM1(1-294) expressing plasmid was kindly provided by Mikael Lindström (59). The plasmid expressing the pVII-Flag protein has been described previously (67).
Cell lines.The H1299, U2OS, HEK293, A549, and 911 cell lines were obtained from American Type Culture Collection. The cell lines were grown in Dulbecco modified Eagle medium (DMEM; Invitrogen) supplemented with 10% fetal calf serum (FCS; PAA) and penicillin-streptomycin solution (PEST; Gibco) at 37°C in a 5% CO2 incubator. Stable U2OS cell lines constitutively expressing the ZNF622 proteins were established by transfection of the pcDNA-ZNF622-HA (1-477, 1-359, and 109-477) and pcDNA3.1 plasmids into U2OS(KO) cells. Single cell clones were selected in the presence of G418 (Sigma; 700 μg/ml). Transfections were performed with the Turbofect (Thermo Scientific) transfection reagent according to the manufacturer’s protocol.
CRISPR/Cas9 genome editing.U2OS cells were transiently transfected with the plasmid pMB1442 (48), which is a modified version of the original plasmid pX330-U6-Chimeric_BB-CBh-hSpCas9 (Addgene, catalog no. 42230). The gRNA was designed to target the first exon of ZNF622 gene (5′-GCAGCGGGCCCACTATAAGA-3′) using Benchling software and cloned into the pMB1442 plasmid. Single cell clones were isolated using the limiting dilution method and verified by Western blotting and DNA sequencing.
HAdV-C5 infections.All HAdV-C5 infections were done using a replication-competent HAdV-pVII-Flag virus. This virus is based on HAdV-C5 and expresses the pVII-Flag fusion protein, since the Flag epitope tag was incorporated into the C terminus of the pVII protein (46). In most of the experiments cells were infected at a multiplicity of infection of 1, defined as fluorescence-forming units (FFU)/cell), in DMEM without any supplements. After 1 h of incubation at 37°C in the presence of 5% CO2, virus-containing growth medium was replaced with growth medium containing FCS and PEST.
siRNA transfection and virus infection.U2OS cells (1.2 × 105 cells per well) grown on a 6-well plate were transfected with scrambled small interfering RNA [siRNA; siScr, 5′-CUUACGCUGAGUACUUCGA(dTdT)-3′] or NPM1 siRNA [siNPM1, 5′-GUAGAAGACAUUAAAGCAA(dTdT)-3′] synthesized at Eurofins Genomics. Transfections were performed with INTERFERin (Polyplus transfection) reagent according to the manufacturer’s recommendations at a final concentration of 5 nM. Transfected cells were infected with HAdV-pVII-Flag (1 FFU/cell) 24 h after siRNA transfection, and the cells were harvested 48 hpi.
Cell lysates and cell fractionation.Total cells lysate were done using RIPA buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS]) (68) on ice for 30 min, followed by sonication using Bioruptor (Diagenode). Successive salt fractionation of cell nuclei was performed as described previously (60). Briefly, approximately 2 × 107 U2OS cells were suspended in hypotonic buffer (10 mM HEPES [pH 7.9], 1.5 mM MgCl2, 10 mM KCl, 0.5 mM dithiothreitol, and Complete protease inhibitor [Roche]). Cells were disrupted with a Dounce homogenizer and pelleted 1,500 × g for 5 min at 4°C. The supernatant was collected as a cytoplasmic fraction, while the pellet at (nuclear fraction) was digested with MNase for 1 h at 37°C. Thereafter, the MNase-treated nuclear fraction was sequentially fractionated by a set of buffers containing increasing amounts of NaCl (75, 150, 300, and 600 mM) (60).
IP.Approximately 8 × 106 cells were lysed in immunoprecipitation (IP) lysis buffer (50 mM Tris-HCl [pH 7.5], 150 mM NaCl, 0.1% Triton X-100, 0.2 mM ZnCl2, 1 mg/ml bovine serum albumin [BSA]) for 10 min on ice and then sonicated for 30 min with a Bioruptor (Diagenode) (Fig. 5B, 8A, 8B, 8C, and 8F). Alternatively, cells were lysed in RIPA buffer for 30 min on ice. Lysed cells were centrifuged at 21,130 × g, 10 min at 4°C. The supernatant was incubated with an anti-Flag M2 affinity gel (Sigma) for 3 h at 4°C. In the case of anti-hemagglutinin (anti-HA) immunoprecipitation, cell lysates were incubated with an anti-rabbit HA antibody overnight and thereafter collected using protein G-Sepharose (GE Healthcare). The beads were washed three times, using 1 ml each time, in the respective lysis buffer. In the case of Benzonase treatment, immunoprecipitated protein complexes were treated with 500 U of Benzonase (Sigma) for 1 h at 37°C. Proteins were eluted with 2× SDS loading buffer and analyzed by Western blotting.
Antibodies.The following primary antibodies were used for Western blot and immunoprecipitation experiments: anti-rabbit HMGB2 (Abcam, ab67282), anti-rabbit HMGB1 (Abcam, ab18256), anti-mouse NPM1 (Abcam, ab10530), anti-rabbit histone H3 (Abcam, ab1791), anti-goat actin (Santa Cruz, sc-1616), anti-mouse pV (kindly provided by D. Matthews [69]), anti-rabbit HAdV-5 capsid (Abcam, ab6982), anti-rabbit ZNF622 (Bethyl Laboratories, A304-076A), anti-mouse Flag (Sigma, M2, F1804), anti-rabbit Flag (Sigma, F7425), anti-rabbit HA (Santa Cruz, sc-805), anti-mouse HA (BioLegend, 901501), and anti-mouse GAPDH (Santa Cruz, sc-365062).
Western blotting.Equal amounts of total protein lysate (30 μg) were loaded on AnyKD (Bio-Rad) or a homemade SDS-containing polyacrylamide gel. Western blotting was performed using wet gel transfer to the nitrocellulose or polyvinylidene difluoride membranes, followed by membrane incubation with primary antibodies in blocking buffer (LI-COR) at 4°C. Proteins were detected with fluorescence-labeled secondary antibodies (LI-COR) and visualized using an Odyssey CLX imaging system (LI-COR). Proteins were quantified using Image Studio 2.1 software (LI-COR) according to the manufacturer’s instructions.
ChIP.The ChIP assays were performed as described previously (70) with some modifications. Briefly, infected cells (1 FFU/cell, 48 hpi) were cross-linked with 1% formaldehyde in 1× PBS. Cells were lysed in SDS lysis buffer (50 mM Tris-HCl [pH 8.0], 10 mM EDTA, 1% SDS) and sonicated using a Bioruptor Pico (Diagenode). Immunoprecipitations were performed from 2.5 × 106 cells using 3 μg of anti-Flag (Sigma, M2, F1804), anti-histone H3 (Abcam, ab1791), and anti-HA (an equal mixture of anti-mouse [BioLegend, 901501] and anti-rabbit [Santa Cruz, sc-805] antibodies). The control (Ctrl) reaction was carried out with magnetic beads only. Pierce protein A/G magnetic beads (Thermo Scientific) were used to collect antibody-protein complexes. The beads were washed, and DNA was eluted as described previously (70). Quantitative PCR on ChIP material was set up using FIREPol EvaGreen qPCR Supermix (Solis Biodyne), and the reactions were carried out using a QuantStudio 6 Flex real-time PCR system (Thermo Scientific). The following primers were used to amplify HAdV-C5 and cellular DNA: E1Apromoter_Fw (5′-TCCGCGTTCCGGGTCAAAGT-3′), E1Apromoter_Rev (5′-GTCGGAGCGGCTCGGAG-3′), E1Agene_Fw (5′-TCCTAAAATGGCGCCTGCTA-3′), E1Agene_Rev (5′-GGACCGGAGTCACAGCTATCC-3′), MLTUpromoter_Fw (5′-TGACTTCTGCGCTAAGATTGTCA-3′), MLTUpromoter_Rev (5′-ATGCGGCCACCCTCAAA-3′), L2unit_Fw (5′-CAAGTTGCATGTGGAAAAATCAAA-3′), L2unit_Rev (5′-CGCAAAGTTGATGTCTTCCATTC-3′), L4unit_Fw (5′-GTCTGGCGCCCAACGA-3′), L4unit_Rev (5′-GGCCCCTGCTCTGTTGAAA-3′), GAPDHpromoter_Fw (5′-TTCGCCCCAGGCTGGATGG-3′), and GAPDHpromoter_Rev (5′-AGGCGGAGGACAGGATGGC-3′).
RNA isolation and qRT-PCR.Total RNA was extracted according to the TRI Reagent (Sigma) protocol. Isolated RNA was DNase treated using a RapidOut DNA removal kit (Thermo Scientific). SuperScript II reverse transcriptase (Invitrogen) was used to synthesize cDNA from RNA primed with random primers. Quantitative reverse transcription-PCR (qRT-PCR) reactions were set up using FIREPol EvaGreen qPCR Supermix (Solis Biodyne) and run in three technical replicates using the QuantStudio 6 Flex real-time PCR system (Thermo Scientific). The primers ZNF622_Fw (5′-TGATTCTGCCTTCTGGTGCC-3′) and ZNF622_Rev (5′-CACGGCCTTCCGATTTTTGG-3′) were used. The data analysis and the HPRT1 oligonucleotides have been described previously (46, 68).
Genomic DNA isolation and qPCR.Genomic DNA was extracted according to the NucleoSpin DNA RapidLyse kit (Macherey-Nagel) protocol. qPCR was performed using FIREPol EvaGreen qPCR Supermix (Solis Biodyne) and analyzed with the QuantStudio 6 Flex real-time PCR system (Thermo Scientific). The following primers were used: E1B-55K_Fw (5′-CACGTAGCCAGCCACTCTC-3′), E1B-55K_Rev (5′-CAAATGCAAGGAACAGCGGG-3′), RNaseP_Fw (5′-AGATTTGGACCTGCGAGC-3′), and RNaseP_Rev (5′-GAGCGGCTGTCTCCACAAGT-3′). The E1B-55K values were normalized to cellular housekeeping gene RNaseP values. Data (Fig. 2B) are shown as the fold increases of HAdV-C5 genomes relative to virus genomes in U2OS(wt) cells at 24 hpi. The virus DNA amount in U2OS(wt) at 24 hpi was considered to be 1.
Indirect immunofluorescence assay.Indirect immunofluorescence assay was carried out as described previously (68). Briefly, stable U2OS cell lines [KO+ZNF622(wt)-HA, KO+ZNF622(1-359), and KO+ZNF622(109-477)] were infected with HAdV-pVII-Flag (1 FFU/cell) or mock infected (infection medium without virus) for 48 h. Cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 15 min and permeabilized with 0.1% Triton X-100 in PBS-T (PBS plus 0.01% Tween 20) for 15 min at room temperature. After blocking the cells with 3% BSA/PBS-T solution, the coverslips were subjected to immunofluorescence analysis using anti-Flag (Sigma, rabbit, 1:1,000) and anti-HA (BioLegend, mouse, 1:1,000) antibodies. Proteins were visualized using anti-mouse IgG(H+L) highly cross-adsorbed secondary antibody, Alexa Fluor 488 (Thermo Fisher Scientific, A-11029, 1:500) and anti-rabbit IgG(H+L) cross-adsorbed secondary antibody, and Alexa Fluor 594 (Thermo Fisher Scientific, A-11012, 1:500) secondary antibodies. Nuclei were counterstained with DAPI (1 μg/ml), and the coverslips were mounted using Fluoromount-G (Southern Biotech) mounting solution. Cells were visualized by a fluorescence microscope (Nikon Eclipse 90i), and the images were processed with NIS-elements (AR 3.10; Nikon) software. Immunofluorescence assays with U2OS cells expressing transfected Myc-NPM1 were performed as described above. Anti-Myc (Sigma, C3956, rabbit, 1:1,000) was used to detect the Myc-NPM1 protein.
Cell viability analysis.Cell viability was analyzed by a trypan blue exclusion test. HAdV-pVII-Flag (1 FFU/cell) infected U2OS(wt) cells were trypsinized, resuspended in PBS, and mixed with 0.4% trypan blue solution (Life Technologies, catalog no. 15250-061). Stained (i.e., lysed) and unstained (i.e., viable) cells were counted with a hemacytometer.
Virus purification and titration.HAdV-pVII-Flag virus was routinely propagated in 911 cells. U2OS(wt) and U2OS(KO) cells were seeded on twelve 15-cm cell culture plates to amplify the virus. At 72 hpi, cells were collected into growth medium, pelleted at 1,500 × g, 5 min at 4°C. Cell pellets were resuspended in 10 mM Tris-HCl (pH 8.0) and lysed by three freeze-thaw rounds to release the virus. The lysates were cleared at 3,162 × g for 30 min at 4°C. The cleared lysates were applied to a 1.25- to 1.40-g/ml CsCl step gradient and spun 175,117 × g in an SW28 rotor (Beckman) for 2 h at 15°C. Virus bands were extracted from the gradients, dialyzed against buffer (10 mM Tris-HCl [pH 7.4], 1 mM MgCl2, 150 mM NaCl, 10% glycerol), and stored at –80°C. The physical virus titer (viral particles [VP]/ml) was established by incubation of the purified virus particles with proteinase K (50 μg/μl) for 1 h at 55°C. Absolute quantification of virus DNA was carried out with qPCR detecting E1B-55K gene amplicons with the E1B-55K_Fw and E1B-55K_Rev primers. The standard curve was generated using the pCMVE1B-55K plasmid (71). The infectious virus titer (FFU/ml) was established by infecting 911 cells with serially diluted virus (102 to 109). Infected 911 cells were fixed with fixing solution (95% methanol in PBS) at 42 hpi, followed by incubation with mouse anti-adenovirus monoclonal antibody (EMD Millipore, MAB8052, 1:50). The cells were washed three times with 1 ml of PBS, followed by incubation with fluorescent dye-conjugated secondary antibodies as described earlier (68). After the final washing steps, the expression of the adenovirus capsid proteins was analyzed using a fluorescence microscope. Equivalent virus titers were estimated according to the capsid protein expression data and dilution factors.
ACKNOWLEDGMENTS
This study was supported by the Marcus Borgströms Foundation, the Åke Wibergs Foundation (M14-0155), the Swedish Cancer Society (CAN 2013/350), and the Swedish Research Council through a grant to the Uppsala RNA Research Centre (2006-5038-36531-16) to T.P. K.M. is supported by the Erasmus Mundus LOTUS+ program.
We thank Göran Akusjärvi, Anna Rostedt Punga, Raviteja Inturi, Rodrigo Villasenor, Philip Knuckles, and Mikael S. Lindström for fruitful discussions; Marc Bühler, David Matthews, Mikael S. Lindström, and Göran Akusjärvi for reagents; and Marta Lewandowska for proofreading the manuscript.
FOOTNOTES
- Received 17 September 2018.
- Accepted 2 November 2018.
- Accepted manuscript posted online 14 November 2018.
- Copyright © 2019 American Society for Microbiology.
