Cyclic AMP-Dependent Protein Kinase Exhibits Antagonistic Effects on the Replication Efficiency of Different Human Papillomavirus Types

ABSTRACT

Several types of widespread human papillomaviruses (HPVs) may induce the transformation of infected cells, provoking the development of neoplasms. Two main genera of HPVs are classified as mucosatropic alphapapillomaviruses and cutaneotropic betapapillomaviruses (α- and β-HPVs, respectively), and they both include high-risk cancer-associated species. The absence of antiviral drugs has driven investigations into the details of the molecular mechanisms of the HPV life cycle. HPV replication depends on the viral helicase E1 and the transcription factor E2. Their biological activities are controlled by numerous cellular proteins, including protein kinases. Here, we report that ubiquitously expressed cyclic AMP-dependent protein kinase A (PKA) differentially regulates the replication of α-HPV11, α-HPV18, and β-HPV5. PKA stimulates the replication of both α-HPVs studied but has a more profound effect on the replication of high-risk α-HPV18. However, the replication of β-HPV5 is inhibited by activated PKA in human primary keratinocytes and U2OS cells. We show that the activation of PKA signaling by different pharmacological agents induces the rapid proteasomal degradation of the HPV5 E2 protein, which in turn leads to the downregulation of E2-dependent transcription. In contrast, PKA-stimulated induction of HPV18 replication is the result of the downregulation of the E8^E2 transcript encoding a potent viral transcriptional inhibitor together with the rapid upregulation of E1 and E2 protein levels.

IMPORTANCE Several types of human papillomaviruses (HPVs) are causative agents of various types of epithelial cancers. Here, we report that ubiquitously expressed cyclic AMP-dependent protein kinase A (PKA) differentially regulates the replication of various types of HPVs during the initial amplification and maintenance phases of the viral life cycle. The replication of the skin cancer-related pathogen HPV5 is suppressed, whereas the replication of the cervical cancer-associated pathogen HPV18 is activated, in response to elevated PKA activity. To inhibit HPV5 replication, PKA targets the viral transcriptional activator E2, inducing its rapid proteasomal degradation. PKA-dependent stimulation of HPV18 replication relies on the downregulation of another E2 gene product, E8^E2, which encodes a potent transcriptional repressor. Our findings highlight, for the first time, protein kinase-related mechanistic differences in the regulation of the replication of mucosal and cutaneous HPV types.

INTRODUCTION

Papillomaviruses (PVs) are small epitheliotropic DNA viruses. Two main genera of human PVs (HPVs) include the mucosatropic Alphapapillomavirus genus and the cutaneotropic Betapapillomavirus genus (α- and β-HPVs, respectively). Persistent infection by a number of these types causes various types of malignancies, most notably cervical and oropharyngeal cancers (1). The cycle of HPV infection can be divided into three distinct stages. First, following the initial infection of basal cells in the stratified epithelium, the viral genome is amplified in cells, reaching a copy number of approximately 200 to 2,000 copies per cell. Second, during latent infection, the viral genome is maintained as a stable extrachromosomal multicopy episome. This phase of the HPV life cycle can last for years and occasionally leads to malignancies. The final stage of the viral life cycle proceeds only in differentiated keratinocytes. It involves a second amplification of the viral genome, the expression of viral capsid proteins, the assembly of virions, and the release of new viral particles (2).

Mechanisms of HPV replication are best understood during the initial amplification of the viral genome. Three viral elements are necessary and sufficient to drive the replication of the HPV genome at this stage. These elements include E1 and E2 gene products and a sequence element in the noncoding part of the viral genome termed the origin of replication (ori) (2). All other elements necessary for replicating the viral genome are provided by the host cell. The E1 protein binds to ori and acts as an initiator protein and helicase (3). The E2 protein has more diverse functions. Full-length E2 is the primary positive regulator of viral transcription. A spliced variant of E2, termed E8^E2, lacking the activation domain acts as a potent transcriptional repressor. Additionally, E2 participates in the partitioning of HPV genomes and ensures the correct binding of E1 to ori (4). Both E1 and E2 are phosphorylated proteins, and several cellular protein kinases can regulate their biological activities. However, only a subset of these kinases has been shown to modulate the replication of HPV genomes in assays involving either pharmacological or genetic regulation of kinase activity. These kinases include CK2α, PYK2, MK2, and p38 mitogen-activated protein kinase (MAPK). CK2α has been shown to be necessary for the initial, maintenance, and vegetative amplification phases of replication of several HPV types (5). MK2 and p38 MAPK are necessary for the second amplification of the HPV16, HPV18, and HPV31 genomes in differentiated keratinocytes (6), and PYK2 regulates the maintenance replication of HPV31 (7).

One of the kinases suggested to be involved in the regulation of HPV replication is a cAMP-dependent protein kinase, PKA. PKA is a ubiquitously expressed tetrameric kinase. The catalytically inactive PKA holoenzyme contains two catalytic and two regulatory subunits. Following the binding of cAMP to the PKA regulatory subunits, the catalytic subunits are released from the complex and act as active protein kinases (8). Interestingly, in the stratified epithelia of the skin and cervix, the highest level of expression of PKA was detected in basal cells, which are susceptible to HPV infection (https://www.proteinatlas.org/). The activity of PKA is regulated by the level of cAMP in the cell. PKA may be potentiated by cell-permeable pharmacological agents inducing an increase in cAMP levels, such as an analog of cAMP, dibutyryl-cAMP (dbcAMP); an inhibitor of phosphodiesterases, 3-isobutyl-1-methyl-xanthene (IBMX); or a stimulator of adenylate cyclases, forskolin. Alternatively, PKA catalytic activity may be suppressed with the inhibitors n-(2-p-bromocinnamylamino-ethyl)-5-isoquinolinesulfonamide (H89) or KT 5720, which blocks PKA signaling through the competitive inhibition of ATP (9). Additionally, PKA activity is regulated by complex phosphorylation events, such as autophosphorylation at threonine residue 197 (10).

It has been proposed that PKA phosphorylates the HPV8 E2 protein at S253, which is conserved only in β-PVs. This phosphorylation is required for the stabilization of the E2 protein, the binding of the protein to mitotic chromosomes, and the correct partitioning of viral genomes (11). Therefore, it can be assumed that PKA activity may primarily regulate the maintenance phase of viral genome replication.

Considering the above-mentioned results, we analyzed the role of PKA in HPV replication in more detail using two independent cellular models: primary human epithelial keratinocytes (HPEKs) and human osteosarcoma U2OS cells. In this article, we show that PKA activity has a strong negative effect on the replication of cutaneotropic high-risk β-HPV5 during both the initial amplification and maintenance phases of the viral life cycle. In contrast, the elevation of PKA activity has a positive effect on the replication of mucosatropic low-risk α-HPV11 and high-risk α-HPV18. We further show that the primary target of this kinase in the regulation of HPV5 genome replication is the E2 protein, which is rapidly degraded via a proteasome-dependent pathway in response to PKA stimulation.

RESULTS

PKA has kinase activity-dependent antagonistic effects on the initial amplification of the α- and β-HPV genomes.

To investigate the role of PKA kinase activity in the life cycle of different HPV types, we first analyzed the expression of the PKA catalytic subunits PRKACA, PRKACB, and PRKACG in HPEKs, CIN612E cells, and U2OS cells, which are permissive for HPV replication, using reverse transcription-PCR (RT-PCR) and two independent pairs of oligonucleotides for each gene. Our analysis revealed that only PRKACA and PRKACB mRNAs were expressed in these cells (Fig. 1A). PKACα and PKACβ proteins encoded by PRKACA and PRKACB, respectively, share 92% identity, suggesting at least partially redundant functions of these proteins. Next, we generated constructs encoding N-terminally Flag-tagged wild-type (wt) PKA catalytic subunit α (referred to here as PKA) and its catalytically deficient mutants PKA(T197A) and PKA(K72R). To verify their predicted enzymatic activity, the overexpressed and immunopurified catalytic subunits were analyzed for their ability to phosphorylate a known PKA target, the Nth1 protein, in vitro (Fig. 1B). Only wt PKA had detectable kinase activity in this experiment.

FIG 1.

PKA regulates the replication of different HPV types in a kinase activity-dependent manner. (A) The expression of the PKA catalytic subunits PRKACA, PRKACB, and PRKACG was analyzed in HPEKs, CIN612E cells, and U2OS cells using RT-PCR. GAPDH expression was used as the positive control. (B) Proteins encoding wt and mutant PKAs were overexpressed in U2OS cells, immunopurified using anti-Flag-M2 resin, and subjected to an in vitro kinase assay in the presence of Nth1 protein. An empty vector was used as the negative control. (C) An empty vector or expression constructs encoding wt or mutant PKA were cotransfected with HPV11, HPV18, or HPV5 genomes into U2OS cells. Total DNA was isolated after 2, 3, and 4 days of incubation; treated with DpnI and restriction enzymes linearizing the respective HPV genomes; and subjected to SB. (D) SB signals corresponding to the replicated linearized HPV11, HPV18, and HPV5 genomes were quantified, and the signals in the control samples cotransfected with an empty vector and incubated for 2 days were set as 100%. The signals of the other samples were calculated relative to those of the control. The data are presented as the average means ± standard deviations (SD) (n ≥ 3) (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

To analyze the effect of the catalytic activity of PKA on the initial amplification of the viral genomes, we utilized U2OS cells permissive for replication of α- and β-HPVs. The expression vectors encoding the Flag-tagged PKA proteins were cotransfected with HPV11 (the low-risk α-HPV type), HPV18 (the high-risk α-HPV type), or HPV5 (the high-risk β-HPV type) into U2OS cells. Total DNA was extracted 2, 3, and 4 days after transfection, and the replication of the respective viral genomes was assessed by Southern blotting (SB).

Overexpressed wt PKA robustly inhibited HPV5 replication, while it stimulated the replication of both α-HPV types, HPV11 and HPV18 (Fig. 1C). The strongest effect of wt PKA on the HPV copy number increase was observed on the 2nd and 3rd days after transfection, which may be explained by the highest level of overexpressed PKA proteins in the transfected cells. Quantification of the results showed that the copy numbers of the HPV11 and HPV18 genomes increased approximately 2- and 10-fold, respectively, in response to catalytically active PKA (Fig. 1D). The mutant PKA proteins also influenced the efficiency of HPV replication but to a much lesser extent than wt PKA. Generally, the mutant PKA(T197A) had a stronger effect on the HPV11 and HPV18 copy numbers than PKA(K72R). These results are consistent with previously reported data demonstrating that the K72R mutation in the ATP binding domain almost completely destroys the catalytic activity of PKA. However, mutation T197A in the activation loop of PKA severely reduces the activity of the kinase but does not abolish it (12, 13). Both mutants can still form a holoenzyme complex with endogenous PKA subunits and either titer away endogenous PKA inhibitors (PKIs) or increase the number of cAMP-activatable holoenzyme complexes in a cell, where one of the catalytic subunits is endogenous PKA.

Next, we analyzed whether modulation of endogenous PKA activity had similar effects on the initial amplification of the HPV5 and HPV18 genomes in U2OS cells. HPV11 was excluded from further analyses because it was least affected by overexpressed PKA. We used three known pharmacological activators of PKA: IBMX, forskolin, and dbcAMP. All three chemicals induced rapid phosphorylation of the PKA target protein CREB1, indicating that there is active endogenous PKA in U2OS cells (Fig. 2A). Similar to our results obtained with overexpressed PKA, the activation of endogenous PKA by IBMX, forskolin, or dbcAMP stimulated the replication of HPV18 (Fig. 2B).

FIG 2.

Endogenous PKA catalytic activity alters the replication of HPV18 and HPV5 in different ways. (A) U2OS cells were treated with the PKA activators IBMX, forskolin (FORSK), and dbcAMP for 10 min, 30 min, 1 h, and 2 h. DMSO was used as a negative control. The levels of total and phosphorylated CREB1 protein in whole-cell extracts were analyzed using immunoblotting. In the top panel, the upper and lower bands correspond to phosphorylated CREB1 and ATF1, respectively. (B and C) U2OS cells were transfected with either the HPV18 or HPV5 genome. The next day, the cells were treated with DMSO, forskolin, IBMX, dbcAMP, H89, or KT 5720 and cultured for an additional 1, 2, or 3 days. Total DNA was isolated 2, 3, and 4 days after transfection; treated with DpnI and either BglI or SacI for linearizing HPV18 and HPV5, respectively; and subjected to SB. (D) SB signals obtained in the experiments described above for panels B and C were quantified. The intensity of the signals obtained in the control samples treated with DMSO for 1 day was set as 100%. The data are presented as average percentages ± SD compared to the control (n ≥ 3) (*, P < 0.05; **, P < 0.01; ***, P < 0.001). (E) U2OS cells were transfected with the HPV18 genome; treated the next day with forskolin, IBMX, or H89; and incubated for 2 additional days. The viability of the cells was assayed using MTT. The cell cycle profile was analyzed using propidium iodide by flow cytometry.

Quantification of the SB signals showed that in the case of HPV18, the strongest effect was achieved using forskolin (Fig. 2D, left). The HPV18 copy number increased approximately 2- and 3-fold in the presence of IBMX or dbcAMP and forskolin, respectively, during the early phase of initial replication 2 and 3 days after transfection. However, when the copy number of HPV18 approached the level of that in the maintenance phase, the positive effect of the PKA activators diminished.

In contrast to that of HPV18, the replication of the HPV5 genome was strongly inhibited by IBMX, but forskolin mediated a milder inhibitory effect (Fig. 2C). Therefore, these data were corroborated using a PKA activator, dbcAMP, and inhibitors, KT 5720 and H89. The HPV5 copy number increased with time in the presence of KT 5720 or H89 and decreased in response to dbcAMP (Fig. 2C and D, right). However, inhibition of the PKA catalytic activity generally provided an insignificant increase in the HPV5 genome copy numbers. The mild effect of PKA inhibition on HPV5 replication may be accounted for by the overall low basal activity of PKA in unstimulated U2OS cells, which is consistent with the very low levels of phospho-CREB1 protein in these cells (Fig. 2A). Taken together, these data confirm the PKA-mediated inhibition of HPV5 replication.

To rule out the possibility of indirect effects of the used chemicals on HPV replication, we performed cell cycle analysis and an MTT [3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H-tetrazolium bromide] assay for U2OS cells treated with forskolin, IBMX, and H89 for 3 days (Fig. 2E). None of the chemicals induced significant changes in the U2OS cell cycle or viability.

Elevated activity of PKA suppresses transient replication of HPV5 in HPEKs.

To corroborate our findings in more physiological settings, we performed similar experiments with HPEKs. HPEKs are able to differentiate in response to elevated calcium levels, and the HPV copy number increases in differentiated cells, mimicking to some extent the events taking place during productive infection.

To measure the replication efficiency of the HPV5 genome in HPEKs, we utilized a modified HPV5 genome, HPV5-Nluc. We have previously shown that nanoluciferase (Nluc) activity is directly proportional to the viral genome copy number and therefore can be used to measure the replication efficiency of a transfected genome (5, 14).

First, we cotransfected the HPV5-Nluc genome together with wt PKA- or PKA(T197A) mutant-expressing constructs into HPEKs and measured the replication efficiency 3 days after transfection. Similar to the results obtained using U2OS cells (Fig. 1B), wt PKA repressed the replication of the HPV5-Nluc genome by more than 60% (Fig. 3A, left). Mutant PKA induced a statistically insignificant mild repressive effect: the HPV5-Nluc copy number was decreased by approximately 20%. The relative luminescence units (RLU) of the vector-transfected cell lysates varied between 0.07 and 3% compared to the HPV-Nluc-transfected HPEK lysates. A similar experiment was performed using the HPV18-E1HA-Nluc-E2Flag genome. In contrast to the mutant PKA, wt PKA was able to induce HPV18-E1HA-Nluc-E2Flag genome copy numbers approximately 3.8-fold, as measured using quantitative PCR (qPCR) (Fig. 3A, right).

FIG 3.

Modulation of endogenous PKA catalytic activity affects the copy number of the HPV5-Nluc and HPV18-E1HA-Nluc-E2Flag genomes in HPEKs. (A) HPEKs were cotransfected with the HPV5-Nluc genome and either an empty vector or expression constructs encoding wt or kinase activity-deficient PKA. Alternatively, the HPV18-E1HA-Nluc-E2Flag genome was used for cotransfection. The HPV5-Nluc copy number was measured using a luciferase assay 3 or 4 days after transfection. Nluc activity was normalized to the total protein concentrations, and the normalized Nluc activity was set to 100% for the control cells transfected with a vector. The HPV18-E1HA-Nluc-E2Flag genome copy numbers were measured by qPCR using DpnI-treated total DNA as a template and normalized to DNA sequences from the m10q locus. Levels of HPV18-E1HA-Nluc-E2Flag in control cells transfected with an empty vector were set to 100%. RLU, relative luminescence units; RCN, relative genome copy number. (B) HPEKs were treated with 1.5 mM CaCl₂ for 2 or 3 days. The levels of the housekeeping protein GAPDH, the keratinocyte differentiation markers involucrin and cytokeratin 10 (K10), CREB1, and S133 phospho-CREB1 were analyzed by immunoblotting. (C) HPEKs were transfected with either the HPV5-Nluc or HPV18-E1HA-Nluc-E2Flag genome, incubated for 1 day, and treated with DMSO, IBMX, forskolin, or H89 in the presence of 1.5 mM CaCl₂, as indicated, for an additional 2 or 3 days. Nluc activity was measured, normalized to the total protein concentrations, and set to 100% for the control cells treated with DMSO in the absence of CaCl₂. For panels A and C, the data are shown as the average percentages ± SD compared to the control (n = 3) (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

To verify that calcium-mediated differentiation of HPEKs takes place in our experiments, we cultivated HPEKs in differentiation medium for 2 or 3 days and analyzed the expression of the differentiation markers involucrin and keratin 10 by immunoblotting. Both proteins were induced after 2 days of treatment (Fig. 3B). We also observed that phosphorylated CREB1, a target of both activated PKA and calcium influx, was induced during differentiation (Fig. 3B).

Next, we challenged HPV5-Nluc-transfected normal and differentiating HPEKs with IBMX or H89. Consistent with the results obtained using U2OS cells (Fig. 2), we observed that elevated endogenous PKA activity had repressive effects on the replication efficiency of the HPV5-Nluc genome, whereas the PKA inhibitor H89 stimulated HPV5-Nluc replication (Fig. 3C, left). These effects were also observed in differentiating HPEKs, where the HPV5-Nluc copy number increased in response to calcium. This increase was reduced in the presence of IBMX and aggravated in the presence of H89 (Fig. 3C, left). However, forskolin and IBMX augmented HPV18-E1HA-Nluc-E2Flag copy numbers 2.5- and 1.5-fold, respectively (Fig. 3C, right). This increase was further potentiated by elevated calcium in the differentiating HPEKs.

PKA represses the replication of the HPV5 genome but stimulates HPV18 replication during the maintenance phase of viral infection.

We have developed and reported stable cell lines generated with U2OS cells where HPV genomes replicate synchronously with cellular DNA as multicopy episomes, mostly in the form of viral genome oligomers of various sizes (14, 15). These cell lines can be used as models to study maintenance replication of HPV genomes, similar to that of persistent infection. In addition, it has been demonstrated that HPV18-positive (HPV18⁺) cells initiate a second round of amplification of the viral genome when cultured at confluence for 3 to 6 days, mimicking to some extent the events taking place during the last stage of productive infection (14, 15). We challenged HPV5⁺ and HPV18-E1HA⁺ cells with IBMX and forskolin to determine whether modulation of PKA activity has an effect on the maintenance replication of these viral genomes. SB of the low-molecular-weight (LMW) DNA showed that the copy number of the HPV18-E1HA genome increased with time in response to the PKA activators (Fig. 4A and B, top). In contrast, both PKA activators had a negative effect on HPV5 maintenance replication, with IBMX being more potent, similar to the observations from transient-transfection studies. In stable cells, IBMX and forskolin induced approximately 70% and 60% decreases in the HPV5 copy number, respectively. A negative effect was observed in both subconfluent (day 2) and confluent (days 4 and 6) cultures. Notably, in contrast to that of the HPV18 genome, the copy number of the HPV5 genome remained stable in the confluent cells, as has been reported previously (16) (Fig. 4A and B).

FIG 4.

Replication of the HPV18-E1HA and HPV5 genomes in stable cell lines is affected by PKA catalytic activity. (A) HPV18-E1HA⁺ and HPV5⁺ cells were treated with DMSO, IBMX, or forskolin (FORSK) for 2, 3, or 6 days. LMW DNA was isolated; treated with the BglI and SacI restriction endonucleases to linearize the HPV18-E1HA and HPV5 genomes, respectively; and subjected to SB. (B) The signals corresponding to the linearized HPV18-E1HA or HPV5 genomes in the experiments described above for panel A were quantified and set as 100% for control cells treated with DMSO and incubated for 2 days. The data are presented as the average percentages ± SD compared to the control (n = 3) (*, P < 0.05; **, P < 0.01; ***, P < 0.001). (C) LMW DNA was isolated from HPV5⁺ cells treated with DMSO, IBMX, or forskolin for 3 or 5 days; treated with the HPV5-noncutting restriction enzyme NheI; and subjected to SB. *, dominant replicon of HPV5.

SB of the uncut HPV5 DNA revealed a pattern of different oligomeric forms that was similar to that observed for HPV18 DNA (Fig. 4C). Quantification of SB signals revealed that the most prevalent replicon, accounting for approximately 30% of all HPV5 DNA, was detected in the LMW DNA samples isolated from the control cells (this oligomer is indicated with an asterisk in Fig. 4C). Treatment of HPV5⁺ cells with IBMX and forskolin resulted in the inhibition of all HPV5 DNA replicons except the dominant oligomeric form. A similar effect has been observed in HPV18⁺ cells in response to depletion of the E1 protein (14). These results suggest that the PKA-mediated inhibition of HPV5 replication involves the E1 and/or E2 protein.

PKA directs HPV5 E2 protein to proteasomal degradation.

To analyze the effect of PKA activation on the HPV replication proteins E1 and E2, we used a modified HPV5 genome named HPV5-E1HA-Nluc-E2Flag. This genome facilitates the detection of the endogenous E1 and E2 proteins using tag-specific antibodies.

The introduced modifications did not hinder the replication of the HPV5-E1HA-Nluc-E2Flag genome in U2OS cells, as verified by SB (Fig. 5A). Next, we tested whether the overexpression of the wt or mutant PKA catalytic subunits or modulation of endogenous PKA activity affected the copy number of the HPV5-E1HA-Nluc-E2Flag genome in a manner similar to that observed for the HPV5 and HPV5-Nluc genomes (Fig. 1 to 3). A luciferase assay showed that Nluc activity was reduced in the presence of overexpressed PKA or PKA activators, whereas inhibitors of PKA, H89 and KT 5720, potentiated Nluc activity by approximately 2-fold (Fig. 5B).

FIG 5.

PKA activation directs the HPV5 E2 protein to rapid proteasomal degradation and induces the downregulation of E2-dependent transcription. (A) Total DNA from U2OS cells transfected with the HPV5-E1HA-Nluc-E2Flag genome was isolated 2, 3, or 4 days after transfection; treated with the DpnI and SacI restriction endonucleases; and subjected to SB. (B) U2OS cells were cotransfected with the HPV5-E1HA-Nluc-E2Flag genome and either an empty vector or constructs encoding wt or catalytically deficient PKA. Alternatively, U2OS cells were transfected with the HPV5-E1HA-Nluc-E2Flag genome and treated the next day with DMSO, the PKA activator IBMX or forskolin (FORSK), or the PKA inhibitor H89 or KT 5720. All cells were incubated for 3 days and subjected to luciferase assays. Nluc activity was normalized to alkaline phosphatase activity and set as 100% for the control samples, which were either cotransfected with an empty vector (left) or treated with DMSO (right). The data obtained for the other samples were calculated relative to the control. The data are presented as the average means ± SD (n = 3) (***, P < 0.001). (C) U2OS cells were transfected with either HPV5 or HPV5-E1HA-Nluc-E2Flag and incubated for 4 days. Cells transfected with the HPV5-E1HA-Nluc-E2Flag genome were treated with the PKA activator IBMX, dbcAMP, or forskolin for 6 h. HA-tagged E1 was detected in the whole-cell extract samples using anti-HA-HRP antibody. The Flag-tagged E2 protein was immunopurified using anti-Flag-M2 antibody and detected using anti-Flag-HRP antibody. The housekeeping protein GAPDH was used as the loading control. (D) U2OS cells were transfected with the HPV5-E1HA-Nluc-E2Flag genome, incubated for 4 days, and treated with either DMSO or IBMX in the presence of the proteasome inhibitor MG132, as indicated, for 1, 2, 4, or 6 h. The E1 and E2 proteins were detected as described above for panel C. (E) WB signals corresponding to the HPV5 E2 protein were quantified. The intensity of the signals in the control samples treated with DMSO was set to 100%, and the signals from the other samples were calculated relative to the control. The data are shown as the average means ± SD (n = 3) (*, P < 0.05; **, P < 0.01; ***, P < 0.001). (F) U2OS cells were transfected with either an empty vector or the expression construct encoding the HPV5 Flag-tagged E2 protein. The cells were incubated for 1 day and treated with DMSO, IBMX, or forskolin, as indicated. The overexpressed E2 protein was analyzed using immunoblotting and anti-Flag-HRP antibody (*, nonspecific band). (G and H) U2OS cells were transfected with replication-deficient HPV5E1⁻ or HPV5E2⁻ genomes, incubated for 3 days, and treated with DMSO, IBMX, or dbcAMP for 8 h. E1, E2, E6, E7, E1^E4, and E8^E2 mRNA expression levels were analyzed using qPCR in triplicate and normalized to GAPDH expression levels. The data were calculated relative to the normalized expression level of each particular gene in control cells treated with DMSO (set to 1). The data from the IBMX- or dbcAMP-treated samples were calculated relative to the respective controls. The data are expressed as the average means ± SD. *, P < 0.05; ***, P < 0.001.

To analyze the status of the endogenous E1 and E2 proteins, U2OS cells were transfected with the HPV5-E1HA-Nluc-E2Flag genome, incubated for 4 days, and challenged with IBMX, forskolin, or dbcAMP for 6 h. The immunoblot analysis revealed that HPV5 E1-hemagglutinin (HA) and E2-Flag proteins were approximately 90 and 65 kDa, respectively (Fig. 5C). Elevated PKA catalytic activity resulted in a reduction in HPV5 E2 levels but not E1 levels. The strongest and lowest reductions in the E2 protein level were observed in IBMX- and forskolin-treated cells, respectively, which is consistent with the effects of these PKA activators on HPV5 replication (Fig. 2B and C and Fig. 5B).

The reduction in HPV5 E2 protein levels can be due to either a dramatic blockade of viral transcription or the degradation of the E2 protein itself. To distinguish between these possibilities, we pharmacologically inhibited the proteasome-mediated protein degradation pathway in HPV5-E1HA-Nluc-E2Flag-transfected cells, which we challenged with IBMX for 1, 2, 4, or 6 h. A reduction in HPV5 E2 protein levels was detected after 2 h of IBMX challenge, and it was reversed by the addition of a challenge cocktail containing the proteasome inhibitor MG132 to the cultured cells. No major changes in the HPV5 E1 protein levels were observed (Fig. 5D). Quantification of the immunoblot analysis signals showed that the endogenous E2 protein levels decreased by approximately 90%, 80%, and 25% in response to IBMX, dbcAMP, and forskolin, respectively (Fig. 5E). Similar results were obtained by analyzing the Flag-tagged E2 protein overexpressed in U2OS cells (Fig. 5F). Challenging these cells with IBMX for 3, 6, or 24 h resulted in the rapid degradation of overexpressed E2, and a similar effect was observed in the presence of forskolin (Fig. 5F, top and bottom, IBMX and forskolin, respectively). Quantification of the Western blot (WB) signals showed that the levels of the overexpressed HPV5 E2 protein were reduced by approximately 85% and 75% in response to IBMX and forskolin, respectively.

To rule out PKA-mediated E2-independent effects on HPV5 transcription, we utilized replication-deficient HPV5 E1-deficient (HPV5E1⁻) or HPV5E2⁻ genomes. These genomes were transfected into U2OS cells, followed by challenge with either IBMX or dbcAMP for 8 h and analysis of different HPV5 transcripts using qPCR. Analysis of HPV5E1⁻ genome transcription revealed that IBMX and dbcAMP induced a significant reduction in the levels of the transcripts encoding E8^E2, E1, and E2 but not those encoding E1^E4, E6, and E7, which are expressed from different promoters (Fig. 5G) (16). In contrast, there was no significant difference in the levels of any of the analyzed transcripts expressed by the E2-deficient HPV5 genome in the presence of PKA activity-elevating agents (Fig. 5H). These data indicate that E2-independent viral transcription was not perturbed by elevated PKA activity. However, the levels of all the analyzed viral genes expressed in the HPV5E2⁻ genome were approximately 10-fold lower than those expressed in the E1-deficient HPV5 genome.

HPV18 E8^E2 transcripts and E1/E2 proteins are targets of PKA activity.

To analyze the mechanism of the PKA-mediated activation of HPV18 replication, we cotransfected U2OS cells with the replication-deficient HPV18E1⁻ genome combined with either an empty vector or constructs encoding wt or mutant PKA. The HPV18E1⁻ genome does not replicate in transfected cells. The cells were incubated for 3 days, and the expression levels of the viral genes encoding the main regulatory proteins, the transcriptional activator E2, the transcriptional repressor E8^E2, and the helicase E1, were analyzed by qPCR. Our analysis showed that the expression levels of E1 and E2 mRNAs were similar in all the samples analyzed (Fig. 6A). However, the expression of E8^E2 mRNA was downregulated by approximately 95% in cells transfected with catalytically active PKA.

FIG 6.

PKA regulates HPV18 transcription and E1/E2 proteins. (A) U2OS cells were cotransfected with the replication-deficient HPV18E1⁻ genome and either an empty vector or plasmids encoding wt or catalytically deficient PKA. The cells were incubated for 2 or 3 days. The expression levels of the HPV18 transcripts E1, E2, and E8^E2 were analyzed using qPCR and normalized to GAPDH mRNA expression levels. The normalized mRNA expression level of each particular gene was set as 1 for the control cells cotransfected with an empty vector, and the data from other samples were calculated relative to the respective controls. The data are presented as average means ± SD (n = 3) (***, P < 0.001). (B, top) U2OS cells were cotransfected with either HPV18 or HPV18E8⁻ genomes and an empty vector or a plasmid encoding wt PKA. Cells were incubated for 2 days. (Bottom) U2OS cells were transfected with the HPV18E8⁻ genome and challenged with forskolin (FORSK) for 2, 3, or 4 days. Total DNA was extracted, treated with DpnI and BglI, and subjected to SB. (C) U2OS cells were transfected with the HPV18-E1HA-Nluc-E2Flag genome, incubated for 4 days, and treated with DMSO or forskolin for 2, 4, or 8 h. GAPDH, E1, and immunopurified E2 proteins were analyzed using immunoblotting.

To analyze whether the increased HPV18 replication is the result of only E8^E2 downregulation, we utilized the HPV18E8⁻ genome, which is deficient for E8^E2 expression due to a point mutation in the E8^E2 ATG. The replication efficiency of the HPV18E8⁻ genome in the U2OS cells in the presence of overexpressed PKA was compared with that of the HPV18 genome (Fig. 6B, top). The loss of E8^E2 led to a robust increase in HPV replication. However, overexpressed PKA was able to further enhance the copy number of the HPV18E8⁻ genome although to a lesser extent than that of HPV18. Quantification of the SB signals showed that the copy number of the HPV18E8⁻ genome increased approximately 3.5- ± 1.7-fold (P < 0.001; n = 3). Also, forskolin induced the copy number of the HPV18E8⁻ genome approximately 2- ± 0.7-fold (P < 0.01; n = 3) (Fig. 6B, bottom). These data indicate that the absence of the E8^E2 repressor does not limit the PKA-mediated positive regulation of HPV18 replication.

Next, we transfected the HPV18-E1HA-Nluc-E2Flag genome into U2OS cells. At 4 days posttransfection, we challenged the cells with forskolin for 2 or 4 h and analyzed the levels of the E1 and E2 proteins using immunoblotting (Fig. 6C). Our analysis revealed approximately 1.8- ± 0.35-fold increases in the E1 protein level in response to the forskolin challenge (P < 0.01; n = 4). E2 protein levels increased 1.5- ± 0.17-fold (P < 0.001; n = 4).

DISCUSSION

The basic determinants of HPV replication are the availability and biological activity of E1 and E2 proteins and the ability of the virus genome-bearing cell to enter the S phase. Both E1 and E2 are multiphosphorylated proteins. It was previously shown that they can be targets of a variety of cellular protein kinases. The consequences of these posttranslational modifications are less understood and are mainly studied in the context of biochemical assays pinpointing changes in E1 and/or E2 stability, subcellular localization, binding to other proteins, or binding to DNA. Very few studies have addressed the physiological role of any protein kinase in the regulation of HPV genome replication.

In the present article, we have demonstrated the role of PKA in the regulation of HPV replication. PKA was chosen as the protein kinase of interest for several reasons. First, it is expressed in keratinocytes, the natural host cells of HPV infection. Second, the availability of cAMP, the effector of PKA, regulates the proliferation and differentiation of keratinocytes in the skin (17–20). Third, it has been suggested that PKA influences the HPV life cycle by targeting at least two viral proteins, the E2 transcription factor in the case of β-HPV8 and the E6 oncoprotein in the case of α-HPV18 (11, 21, 22). Finally, bioinformatics analysis indicated multiple PKA consensus sites, RRXS/T, in the E1 and E2 proteins of different HPV types.

Our results show that the effects of PKA activity on the efficiency of HPV replication depend on the virus type. We used HPV5 as a prototypic β genus member and HPV11 and HPV18 as members of the α genus. Elevation of PKA activity stimulated the efficiency of the initial amplification of the HPV11 and HPV18 genomes while strongly inhibiting that of HPV5. We further showed PKA-mediated inhibition of HPV5 replication in HPEKs, the natural host cells of HPV5, and in a stable cell line bearing the HPV5 genome in the form of oligomeric multicopy episomes. Some deviations in the efficiency of the effects obtained using different pharmacological PKA activators (e.g., IBMX and forskolin) may be explained by their possible off-target effects. For instance, IBMX, a nonspecific phosphodiesterase inhibitor, may also activate cGMP-dependent kinase signaling, and forskolin is a specific activator of membrane-associated adenylate cyclase but not soluble adenylate cyclase (23).

The observation that PKA induced effects antagonistic to the replication of different HPV types led us to the idea that PKA differentially modulates the activity of the viral replication-related proteins E1 and/or E2. Indeed, we found that elevated PKA activity directs the HPV5 E2 protein to proteasomal degradation, whereas it has no negative effect on HPV5 E1. It was previously shown that E2 proteins of different α-HPV types are subjected to ubiquitin-mediated proteasomal degradation through their amino-terminal activation domain (24–26).

Two possible mechanisms may explain how PKA-dependent phosphorylation induces the degradation of the HPV5 E2 protein. First, a specific ubiquitin ligase may recognize the PKA-phosphorylated part of E2. Second, phosphorylation may induce conformational changes in the E2 protein that in turn allow easier access of ubiquitin ligases to E2. Both mechanisms have been characterized for mammalian transcription factors (27–30). Currently, we are unable to distinguish between these possibilities. It is noteworthy that CK2-dependent phosphorylation induces conformational changes in the deltapapillomavirus BPV1 E2 protein that triggers the degradation of the viral protein (31).

In contrast to its effect on HPV5, PKA catalytic activity stimulates HPV18 replication via the downregulation of the E8^E2 transcript, which is initiated from its specific promoter p1193 (16). The downregulation of E8^E2 expression may occur via either the interference of p1193 or general alterations to the HPV18 splicing pattern. Although the exact mechanism of the PKA-mediated activation of HPV18 replication remains to be elucidated, our results clearly show that PKA may be a positive regulator of the HPV18 life cycle and also a positive regulator of E1 and E2 proteins.

The fact that PKA has a strong negative effect on β genus HPV5 replication suggests that the virus should target cells in the epidermis with no or low PKA activity to achieve successful infection. In contrast, α-HPV types can infect cells of the mucosal epithelium with active PKA. In reality, it is unclear which cells are the targets of HPV infection in both cutaneous and mucosal epithelia, as stem cells for these tissues are poorly characterized (32). However, some indirect evidence suggests that β-HPV types might indeed infect cells with low PKA activity in the skin. Namely, it has been shown that one of the stem cell populations of the skin lies in the bulge of the hair follicle (33). This particular cell population in mice is dependent on the mitogen Shh and its downstream effector Gli transcription factors (34). Effective SHH signaling can occur only in cells with low PKA activity, as PKA-mediated phosphorylation of GLI proteins directs these transcription factors to proteasomal degradation (35, 36). It has been suggested that these hair follicle stem cells are also targets of β-HPV infections in healthy individuals (37).

In summary, we have identified PKA as an important regulator of HPV5 and HPV18 replication. Our results pinpoint the importance of the activities of cellular protein kinases in the regulation of HPV replication during different stages of the viral life cycle.

MATERIALS AND METHODS

Plasmids and chemicals.

HPV5, HPV5-Nluc, HPV5E1⁻, HPV5E2⁻, HPV11, HPV18, HPV18-E1HA-Nluc-E2Flag, HPV18E8⁻, and HPV18E1⁻ genomes and their parental plasmids were described previously (5, 38, 39). The HPV5-E1HA-Nluc-E2Flag genome was generated on the basis of the HPV5-Nluc genome upon the insertion of a sequence encoding the HA tag after the 15th nucleotide of the E1 open reading frame (ORF). The sequences encoding the codon-optimized nanoluciferase (Nluc) and 2A region of foot-and-mouth disease virus were inserted after the 60th nucleotide in the E2 ORF, immediately after the E1 stop codon. The full-length wt E2 ORF containing the Flag tag-encoding sequence after the first ATG begins after the 2A sequence.

The sequence encoding catalytic subunit α of PKA was amplified by RT-PCR using a Phusion high-fidelity DNA polymerase kit (Thermo Fisher Scientific) and cDNA from U2OS cells. The sequence was cloned into a pFlag-CMV-4 vector (Sigma-Aldrich) between the HindIII and NotI sites. The constructs encoding PKA mutants T197A and K72R were generated on the basis of a construct encoding Flag-tagged wt PKA using PCR-based mutagenesis. The sequence encoding Flag-tagged E2 of HPV5 was amplified using a Phusion high-fidelity DNA polymerase kit and the HPV5 genome as a template and cloned into a pFlag-CMV-4 vector between the HindIII and BamHI sites. The oligonucleotides are listed in Table 1.

TABLE 1.

Oligonucleotides used in the present study

Primer	Oligonucleotide	Sequence (5′–3′)
1	PRKACA Fw1	CTTATACATGGTCATGGAGTAC
2	PRKACA Rv1	CTGTAGATGAGATCCAGCGAG
3	PRKACA Fw2	CTACCCGCCCTTCTTCGCAGAC
4	PRKACA Rv2	CAAAGCGCTTGGTGAGATCTAC
5	PRKACB Fw1	GGGTGAAATGTTTTCACATC
6	PRKACB Rv1	GATCTCTGTAGATGAGGTCTAG
7	PRKACB Fw2	GGCAGAACTTGGACATTATGTG
8	PRKACB Rv2	GGTTGGTCTGCAAAGAATGG
9	PRKACG Fw1	CTGCAGGTGGACCTCACCAAG
10	PRKACG Rv1	CGTCAAAGTTACTGGCATCCC
11	PRKACG Fw2	CTACAGCGCGTCGGAAGGTTTAG
12	PRKACG Rv2	GAAGTCCGTCACCTGCAGGTAG
13	PRKACA Fw HindIII	AAGCTTGGCAACGCCGCCGCCGCCAAGA
14	PRKACA Rv BamHI	GGATCCCTAAAACTCAGAAAACTCCTTG
15	PKA A1 K72R Fw	GAACCACTATGCCATGAGGATCCTCGACAAACAG
16	PKA A1 K72R Rv	CTGTTTGTCGAGGATCCTCATGGCATAGTGGTTC
17	PKA A1 T197A Fw	GAAGGGCCGCACTTGGGCCTTGTGCGGCACCCCT
18	PKA A1 T197A Rv	AGGGGTGCCGCACAAGGCCCAAGTGCGGCCCTTC
19	HPV5 E2 HindIII Fw	AAGCTTGAGAATCTCAGCGAGCGTTTC
20	HPV5 E2 BamHI Rv	GGATCCTTAAAGACTGTCCAGGTTGCC
21	HPV5 E8^E2 Fw1	TGAAGCTGAAGATGTTACTC
22	HPV5 E8^E2 and E1^E4 Rv1	GATGTGGTGAGCTGTCTGGAC
23	HPV5 E8^E2 Fw2	GATGTTACTCCTGAGGTGGAG
24	HPV5 E8^E2 and E1^E4 Rv2	CTCTGGTTTCGGTTTGTTGTG
25	HPV5 E1^E4 Fw1	CTGCAAACATGACGGATCC
26	HPV5 E1^E4 Fw2	GACGGATCCTAATTCTAAAG
27	HPV5 E1 Fw1	CAGTATGCAAGGCTTGCTCC
28	HPV5 E1 Rv1	CTGACCAGTGCCCTTCCCCTTC
29	HPV5 E1 Fw2	GCTGCAACCCCTTTCAGAGTG
30	HPV5 E1 Rv2	CCCCATGCACATTAATGTTAG
31	HPV5 E2 Fw1	GCTTGAGTCACTACAGACATC
32	HPV5 E2 Rv1	GGCATTATCTGGATCATTGTC
33	HPV5 E2 Fw2	GTATACAATGTGGACCTATGTG
34	HPV5 E2 Rv2	CTTCCCATTCTCCAGTTGTAC
35	HPV5 E6 Fw	GTATGCTGTGGCGCCACTG
36	HPV5 E6 Rv	GAAGGCCTCTGCCACAGCAATC
37	HPV5 E7 Fw	CTGGAGCTCAGTGAGGTGCAG
38	HPV5 E7 Rv	CTCACAGTTCCTGCAACCGCAC
39	HPV18 Fw1	GAAACACACCACAATACTATG
40	HPV18 Rv1	GCAGTGAAGTGTTCAGTTCC
41	HPV18 Fw2	CCTGTCAAAAGGATGCTGCAC
42	HPV18 Rv2	GACGCAATCCAGCCTGAAC
43	m10q F	TAGACCCAGGAGGGAGTTATTTAAGAG
44	m10q R	TTGGGAATGCAATGCAGTGTGTAC

MG132, IBMX, and forskolin were purchased from Sigma-Aldrich. H89, KT 5720, and dbcAMP were purchased from Santa Cruz Biotechnologies. The chemicals were dissolved in dimethyl sulfoxide (DMSO) (Sigma-Aldrich).

Cell culture.

U2OS cells (ATCC HTB-96) and their derivative HPV5⁺ and HPV18-E1HA⁺ cells were grown in Iscove’s Dulbecco’s medium (Biowest) with 10% fetal calf serum and 1% penicillin-streptomycin (Sigma-Aldrich). HPV5⁺ (previously referred to as clone 15) and HPV18-E1HA⁺ cells were previously described (14, 15). The cells were transfected by electroporation as described previously (14). The following amounts of plasmids were used for the transfection of approximately 10⁶ cells: 0.5 μg of HPV11; 1 μg of HPV5, HPV5-Nluc, HPV5-E1HA-Nluc-E2Flag, and HPV18; 0.25 μg of wt PKA; 1 μg of PKA(T197A); and 1.5 μg of PKA(K72R). The following final concentrations of the chemicals were used in normal growth medium: 10 μM MG132, 0.5 mM IBMX, 20 μM forskolin, 1 mM dbcAMP, 1 μM H89, and 1 μM KT 5720. The medium was replenished daily.

HPEKs were propagated through passage 5 in keratinocyte serum-free medium (Gibco) supplemented with penicillin-streptomycin. The HPEKs were transfected using Lipofectamine LTX (Invitrogen). The following amounts of plasmids were added to each well of a 96-well plate: 70 ng of HPV18-E1HA-Nluc-E2Flag, 50 ng of HPV5-Nluc, and 10 ng or 50 ng of the wt or mutant PKA-encoding constructs, respectively. Differentiation of the HPEKs was induced with 1.5 mM CaCl₂ in keratinocyte-SFM medium containing 3-fold less supplements than used to culture undifferentiated HPEKs. The cells were treated with 4 μM forskolin, 100 μM IBMX, or 1 μM H89. Induction medium was replenished daily.

Cell cycle and MTT analyses of U2OS cells transfected with the HPV18 genome and treated with forskolin, IBMX, or H89 for 2 days were performed as previously described (5).

Southern blot analysis.

Total and low-molecular-weight (LMW) DNA was isolated as described previously (14). To analyze the linearized HPV genomes, 5 μg of total DNA was digested with DpnI and the following restriction endonucleases: SacI for HPV5 and its derivatives, HindIII for HPV11, and BglI for HPV18, HPV18E8⁻, and HPV18-Nluc. To analyze the uncut HPV5 genome, 20 μg of LMW DNA was treated with the NheI restriction endonuclease. All restriction enzymes were purchased from Thermo Fischer Scientific. Further DNA manipulations and Southern blot (SB) procedures were performed as described previously (14). Total DNA was isolated from 96-well plates using DNAzol (Invitrogen) according to the manufacturer’s instructions for the analysis of the HPV18-E1HA-Nluc-E2Flag genome copy numbers. Total DNA was treated with DpnI and used for subsequent quantitative PCR (qPCR) analysis.

Immunoprecipitation (IP) and Western blot analysis.

Flag-tagged PKA catalytic subunits and E2 proteins were immunoprecipitated using anti-Flag-M2 affinity gel (Sigma-Aldrich) according to the manufacturer's instructions. The following antibodies were used: HA-horseradish peroxidase (HRP) clone 3F10 (Sigma-Aldrich) (1:2,000), Flag-HRP clone M2 (Sigma-Aldrich) (1:3,000), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (Sigma-Aldrich) (1:5,000), involucrin clone SY5 (Sigma-Aldrich) (1:3,000), cytokeratin 10 (catalog no. ab76318; Abcam) (1:1,000), CREB1 (catalog no. 06-863; Merck Millipore) (1:1,000), phospho-CREB (Ser133) clone 87G3 (Cell Signaling Technology) (1:5,000), and goat anti-mouse or anti-rabbit IgG conjugated with HRP (Invitrogen) (1:15,000). The phospho-CREB protein was analyzed as described previously (40).

RNA isolation, RT-PCR, and qPCR.

Total RNA isolation, cDNA synthesis, and qPCR were performed as described previously (14). The HPV5 transcripts were analyzed using two independent primer pairs for E1, E2, E1^E4, and E8^E2 and one primer pair for E6 and E7 (Table 1). The primers for the HPV18 transcripts and GAPDH were described previously (14). The HPV18-E1HA-Nluc-E2Flag genome copy numbers in HPEKs were analyzed using primers 39 to 42 and normalized to DNA sequences from the human m10q locus (Table 1).

Luciferase assay.

HPEKs were seeded onto 96-well plates 48 h prior to transfection. For each treatment, six replicates were used. Prior to lysis, the cells were washed with phosphate-buffered saline (PBS). The HPEKs were lysed in 30 μl of radioimmunoprecipitation assay (RIPA) buffer per well, frozen for 30 min, and incubated for 15 min at room temperature (RT). Nluc activity was measured using a Nano-Glo luciferase assay system (Promega) and normalized based on the total protein concentration as measured using a bicinchoninic acid (BCA) protein assay kit (Pierce).

Statistical analysis.

The P values for the two-tailed t test were based on assumed unequal variances and calculated using Excel software.