Human cytomegalovirus RNA2.7 regulates host cell cycle and facilitates viral DNA replication by inhibiting RNA polymerase II phosphorylation

Human cytomegalovirus (HCMV) is a ubiquitous pathogen belongs to the beta herpesvirus family. RNA2.7 is a viral long non-coding RNA accounting for more than 20% of total viral transcripts at early time of infection. By construction of RNA2.7 deleted mutant and genome transcriptomic analysis, RNA2.7 is demonstrated to repress host cellular RNA polymerase II (Pol II)-dependent transcription through inhibiting the phosphorylation of RNA polymerase II (Pol II). Co-immunoprecipitation, RNA immunoprecipitation and RNA electrophoretic mobility shift assay are followed to investigate its mechnism. A 145nt-in-length fragment in RNA2.7 is identified to bind to Pol II and block the interaction between Pol II and phosphorylated cyckin-dependent kinase 9 (phospho-CDK9). By inhibiting Pol II phosphorylation, RNA2.7 decreases the transcription and expression levels of chromatin licensing and DNA replication factor 1 (Cdt1) and cell division cycle gene 6 (Cdc6). Through above way, RNA2.7 prevents the entry of cells into S phase and facilitates viral DNA replication. Our results discover the functions of HCMV RNA2.7 in regulation of Pol II phosphorylation and cell cycle control during infection. Author summary Human cytomegalovirus (HCMV) RNA2.7 is a viral lncRNA that is most abundant during infection. Here we show that a 145nt-in-length fragment in RNA2.7 binds to RNA polymerase II (Pol II) and blocks the interaction between Pol II and phosphorylated cyckin-dependent kinase 9 (phospho-CDK9). By inhibiting Pol II phosphorylation, RNA2.7 decreases the transcription and expression levels of chromatin licensing and DNA replication factor 1 (Cdt1) and cell division cycle gene 6 (Cdc6), and blocks host cells entering into S phase. RNA2.7 is confirmed to facilitate viral DNA replication through decreasing Cdt1 and Cdc6. Therefore, our results discover the functions of HCMV RNA2.7 in regulation of Pol II phosphorylation and cell cycle control during infection.


Introduction
Human cytomegalovirus (HCMV) is a ubiquitous human pathogen belongs to the beta herpesvirus family [1]. Following primary infection, the virus establishes lifelong latent infection with episodes of reactivation, mainly in the immune compromised host.
Long non-coding RNAs (lncRNAs) are roughly defined as RNA molecules of more than 200 bases in length without protein-coding capacity. More and more evidences indicate that lncRNAs may play an important role in a variety of biological processes and be involved in many human diseases including tumors and infections. Kaposi's sarcoma-associated herpesvirus (KSHV) produces a highly abundant lncRNA known as PAN RNA. It has been demonstrated that PAN RNA controls KSHV gene expression by mediating the modification of chromatin through targeting KSHV repressed genome [2]. Recent high-resolution transcriptome mapping showed that most viral transcriptions during HCMV infection are concentrated in viral lncRNAs [3,4], including RNA1.2, RNA2.7, RNA4.9 and RNA5.0. HCMV RNA2.7 is a viral lncRNA of 2.7-kb in length, which was previously termed beta2.7 [5,6]. RNA2.7 was observed to be abundant at early time of infection, accounting for more than 20% of total viral transcripts [7,8]. It has been found to prevent cell apoptosis by interacts with mitochondrial complex I during HCMV infection, and to be essential to maintain high levels of energy 5 production in infected cells [9,10]. However, most functions of HCMV RNA2.7 involved in HCMV infection still remain unclear.
In this study, an activation of host cellular RNA polymerase II (Pol II)dependent transcription was observed in cells infected with an RNA2.7 deleted HCMV strain. It was then identified that HCMV RNA2.7 could inhibit Pol II Serine2 (S2) phosphorylation by blocking phosphorylated cyckin-dependent kinase 9 (phospho-CDK9) binding to Pol II. The inhibition of Pol II S2 phosphorylation by HCMV RNA2.7 could decrease the transcription and expression levels of chromatin licensing and DNA replication factor 1 (Cdt1) and cell division cycle gene 6 (Cdc6) to regulate host cell cycle and facilitate viral DNA replication. 6

Results
An RNA2.7 deleted mutant is constructed based on HCMV bacterial artificial chromosome HAN HCMV RNA2.7 is a transcript from the antisense strand in HCMV genome neighbored to RL1, RL6, RL8A and RL9A (Fig 1A). To study the functions of HCMV RNA2.7, an RNA2.7 deleted mutant (HANΔRNA2.7) was constructed based on a previously constructed HCMV bacterial artificial chromosome (BAC) HAN, which is the first characterized HCMV clinical strain in China [11]. By reversetranscription PCR and sequencing, it was confirmed that the RNA2.7 sequence was deleted successfully in HANΔRNA2.7 genome (Fig 1B).
To verify whether the deletion of RNA2.7 disturb the transcriptions of its flanking genes or not, human embryonic lung fibroblast (HELF) cells were infected with reconstituted viruses HAN and HANΔRNA2.7 at a multiplicity of infection (MOI) of 0.5, respectively. Total RNA was extracted at 72 hours post infection (hpi). The transcriptions of RNA2.7 and its flanking genes were measured by quantitative PCR respectively. Transcript of RNA2.7 was not detected in HANΔRNA2.7 infected cells (Fig 1C). The transcription levels of RL1, RL6, RL8A and RL9A in HANΔRNA2.7 infected cells were similar to those in HAN infected cells. Based on these results, HANΔRNA2.7 reconstituted virus could be used as an RNA2.7 knocked out HCMV stock in our further study. 7

Host gene transcriptions are altered after infections with different HCMV constructs
To address the effects of RNA2.7 on host gene transcription, HELF cells were infected with HAN or HANΔRNA2.7 at an MOI of 0.5. HELF cells treated with PBS were used as a control. Total RNA was harvested at 72 hpi and subjected to whole genome transcriptomic analysis.
In total, transcriptions of 2,520 cellular genes were substantially altered by infection with HAN compared with uninfected cells, while transcriptions of 4,286 cellular genes were changed by HANΔRNA2.7 infection (Fig 2A). After sorting according to relative transcription levels, 4 patterns were identified: transcriptions of 509 genes were only affected by HAN infection and were designated as HAN specific transcriptional genes. Transcriptions of 54 cellular genes were inversely regulated between HAN and HANΔRNA2.7 infections, 1,957 genes were similarly regulated by both strains, and a total of 2,275 genes were HANΔRNA2.7 specific. In general, more cellular genes were transcribed in response to HANΔRNA2.7 infection compared with HAN infection, suggesting either an activation of gene transcription or a lack of suppression at transcriptional level due to deletion of RNA2.7.

HCMV RNA2.7 represses cellular Pol II-dependent transcription
Cellular transcripts altered by HAN or HANΔRNA2.7 infection were analyzed for their comparisons on diseases and biofunctions. Results with 8 different z-scores between HAN and HANΔRNA2.7 infected cells are listed ( Fig   2B). In HAN infected cells, the apoptosis of endothelial cells was indicated to be repressed with z-score of -2.40. On the other hand, DNA endogenous promoter was indicated significantly activated in HANΔRNA2.7 infected cells (z-score=3.74). Genes involved in activation of DNA endogenous promoter were analyzed and their information is listed in S1  10 Phosphorylation of Pol II S2 promotes Pol II to overcome transcriptional blocks and increases the efficiency of 3'-end processing during elongation.
Phospho-CDK9 is an important molecule mediating Pol II S2 phosphorylation.
Therefore, protein levels of CDK9 and phospho-CDK9 were also measured and compared between cells infected with HAN or HANΔRNA2.7. Both CDK9 and phospho-CDK9 were increased and accumulated along with infection, while no difference was found between cells infected with HAN and HANΔRNA2.7 (Fig 4A).
It was confirmed that inhibition of Pol II S2 phosphorylation by RNA2.7 was not directly caused by alteration of CDK9 or phospho-CDK9 proteins.
Since the inhibition of Pol II S2 phosphorylation by RNA2.7 was not due to the quantity changes of CDK9 or phospho-CDK9, the effects of RNA2.7 on the interaction between phospho-CDK9 and Pol II was then studied. Pol II binding proteins were immunoprecipitated using anti-Pol II antibody in HELF cells infected with different strains. Phospho-CDK9 was measured in captured proteins by western blot. Pol II was used as a reference for calculation. When RNA2.7 was deleted during HCMV infection, phospho-CDK9 binding to Pol II was increased by more than 47.6% ( Fig 4B). It was illustrated that RNA2.7 could block the binding of phospho-CDK9 to Pol II.
To explore the interaction between HCMV RNA2.7 and Pol II or CDK9, RNA immunoprecipitation (RIP) was performed to immunoprecipitate Pol II or CDK9 binding RNAs in HAN infected HELF cells. To exclude the disturbance of the binding between the undergoing transcriptional RNA2.7 and Pol II, alfaamanitin, which can block all undergoing Pol II-dependent transcription 11 processes, was added into cells 2 hours before immunoprecipitation. Captured RNA was reverse transcribed and amplified using RNA2.7 specific primers. RNA2.7 was found to bind to both Pol II and CDK9 proteins physically (Fig 4C). was identified to functionally inhibit Pol II S2 phosphorylation ( Fig 4D).
Compared to reference value, the level of endogenous Pol II S2 protein was decreased by 54.9% in cells transfected with vector transcribing RNA2.7C2c. The block effect of RNA2.7C2c was then validated using transfection and coimmunoprecipitation with anti-Pol II antibody. Compared to cells transfected with pcDNA3.1, the amount of phospho-CDK9 binding to Pol II was decreased by about 42.6% in cells transfected with vector transcribing RNA2.7C2c (Fig 4E).
To address the mechnism of RNA2.7C2c blocking the interaction between phospho-CDK9 and Pol II, RNA electrophoretic mobility shift assay (RNA EMSA) was carried out using biotin-labeled RNA2.7C2c RNA probes. A strip appeared after nucleoprotein was incubated with RNA2.7C2c probes ( Fig 4F). The strip 12 was weakened after competitive RNA or anti-Pol II antibody was added into the reaction systems. Although no classic super-shift strip was obtained after incubation with Pol II antibody, the strip was more weakened when more anti-Pol II antibody was added. However, no change of the strip was observed after incubation with anti-CDK9 antibody.
In addition, Pol II proteins were purified from nucleoproteins using magnetic beads and antibodies. Purified proteins with gradient increasing concentrations were incubated with biotin-labeled RNA2.7C2c RNA probes.
When the concentration of purified Pol II protein was increased, the quantity of detected RNA probes was decreased correspondingly (Fig 4F). The results confirmed that RNA2.7C2c is a functional fragment of RNA2.7 to inhibit Pol II S2 phosphorylation by a physical binding to Pol II protein directly. Serine2-phosphorylated Pol II [13,14]. 17 HCMV processes gene transcription through cellular Pol II and therefore it might also develop mechanisms to regulate Pol II activities contributing viral growth. It has been reported that HCMV UL79 interacts with Pol II to benefit accumulation of viral transcription during late stages of infection [15]. In our results, RNA2.7 inhibits the phosphorylation of Pol II S2 site without altering CDK9 and phospho-CDK9 levels. By a series of experiments including Coimmunoprecipitation and RNA immunoprecipitation, a 145nt-in-length fragment (RNA2.7C2c) in RNA2.7 was identified to inhibit Pol II S2 phosphorylation by blocking the interaction between phospho-CDK9 and Pol II.
The interaction between RNA2.7C2c and Pol II or CDK9 was detected using RNA EMSA. Different to adding anti-Pol II antibody into the reaction system, there was no change observed in the strips after adding anti-CDK9 antibody. No super-shift strip was obtained by adding anti-Pol II antibody into reaction system. It might be due to the large volume of RNA-protein complex that prevents them entering gel in electrophoresis. However, the detected quantity of biotin-labeled RNA2.7C2c reduced gradually along with the increasing input of purified Pol II correspondingly, which confirmed that the inhibition of Pol II S2 phosphorylation by RNA2.7C2c is mediated by a physical interaction between RNA2.7C2c and Pol II protein directly. The binding of RNA2.7C2c to Pol II might induce a competitive inhibition that reduces the binding of phospho-CDK9 to Pol II. 18 Deregulation of gene expression programs in cancer cells depends on continuous active transcription [16]. For example, the development of triplenegative breast cancer is mainly due to the uninterrupted transcription of oncogenes [17]. Disturbing transcription has been aimed for cancer therapy and chemical drugs triggers degradation of phosphorylated Pol II have been evaluated among cancer patients in recent years [18]. Transfection of small RNA molecule might cause less cytotoxicity than intake of chemical drugs. It has been reported that a small domain of RNA2.7 termed p137 can prevent dopaminergic cell death by protecting mitochondrial complex I activity and is expected to be used in therapy for Parkinson's disease [10,19]. It might be similarly expected for RNA2.7C2c fragment to be used in clinical cancer therapy, considered to its effect of inhibiting Pol II S2 phosphorylation.
Cell cycle progression is regulated by a wide variety of factors. It is established that HCMV infection can lead to cell cycle arrest at some points [20,21]. The assembly of pre-replication complex at DNA replication origins during G1 phase of the cell cycle initiates cellular G1-S-phase transition. Cdt1 and Cdc6 are key components of cellular pre-replication complex [22][23][24][25]. In our study, both mRNA and protein levels of Cdt1 and Cdc6 were decreased along with the inhibition of POLR II S2 phosphorylation by RNA2.7. Without RNA2.7 during infection, more cells drove into S phase for cellular DNA replication.
Cellular DNA replication may compete against viral DNA replication, since more molecules are used for cellular DNA replication and less is available for viral DNA replication. Arresting host cells at G0/G1 phase is essential to the 19 initiation of HCMV gene expression at the time of infection [26]. It has been reported that depletion of pre-replication factors in HCMV infected cell could promote viral replication [27]. Our data showed that HCMV DNA levels HCMV has both lytic and latent phases in its life cycle like other human herpesviruses. Although deletion of RNA2.7 did not result in growth defects of HCMV in fibroblast cells [28], the study on the functions of RNA2.7 should not be only limited in lytic phase of infection and permissive cells for lytic infection.
RNA4.9 was proposed to play a role in transcriptional repression of viral IE gene expression for viral latency infection [4]. Similar to RNA4.9, RNA2.7 was detected in latent CD14+ and CD34+ cells at relatively high levels [4]. Based on our results, it is prospected that HCMV RNA2.7 might play a role in intracellular modulation in some aspects for the progress of latency or reactivation, such as cellular transcription and cell cycle control. More work on RNA2.7 functions for HCMV latency and reactivation is needed in future study. 20  RNase free water and then treated by TURBO DNA-free TM Kit (ThermoFisher).

Cells
RNA preparations were estimated by electrophoresis on 1% agarose gel and quantified by ND-1000 spectrophotometer (Nanodrop Technologies). Total RNAs were reverse-transcribed using the SuperScript III First-strand synthesis system (ThermoFisher). 22 Primers RNA2.7rt forward and reverse were designed and used to amplify RNA2.7 from the cDNAs. Transcript of HCMV UL83 was amplified as a control. The primer sequences used in reverse-transcription PCR are shown in Table 1. Products were analyzed by electrophoresis on a 1.5% agarose gel containing ethidium bromide and visualized under ultraviolet light.  25 The microarray data are available in the NCBI's Gene Expression Omnibus database (accession number GSE73954). Differentially transcribed genes were functionally categorized using ToppGene (https://toppgene.cchmc.org) and Ingenuity Pathway Analysis software (Ingenuity Systems).

Western blot analyses of Pol II and CDK9
HELF cells growing in 6-well plates were infected with HAN or  Table 1. PCR products were inserted into pcDNA3.1 (-) vector (ThermoFisher) and transformed into TOP10 (TianGen BioTech). All constructs were verified by sequencing.
HELF cells were prepared in 6-well plates. 2μg vectors per well were transfected into the cells using Attractene Transfection reagent (Qiagen) after cell preparation. Cells without transfection or transfected with pcDNA3.1 (-) vector were used as a negative control. 48 hours after transfection, proteins were extracted for western blot analysis or Co-IP.

Co-immunoprecipitation
HELF cells growing in 100mm plates were infected with different HCMV strains at an MOI of 0.5 or transfected with vectors transcribing RNA2.7C2c. Cells were harvested and re-suspended using M-PER (ThermoFisher) with protease and phosphatase inhibitor cocktail (Abcam). PureProteome TM ProteinA/G Mix Magnetic Beads (Millipore) were coated with anti-Pol II antibody and incubated 27 with lysates at 4℃overnight. The captured protein complex was eluted with 60μl SDS-PAGE sample loading buffer (Beyotime) and then heated at 70℃ for 10 minutes. After the beads were removed, the supernatants were loaded on 8% SDS-PAGE gel. Blotted PVDF membranes were incubated with antibodies against Pol II and phospho-CDK9, followed by peroxidase-conjugated goat anti-mouse or rabbit IgG (ZSGB-BIO) with ECL western blot reagent (ThermoFisher). Protein relative densities were quantified using ImageJ software 1.44p (NIH). Pol II was used as a quantitative reference to caculate the relative amount of phospho-CDK9 binding to Pol II.

Preparations of nucleoprotein and Pol II protein
Nucleoprotein was extracted from 1×10 7 HELF cells using NE-PER

Statistical analysis
Statistical analyses were performed using Excel and GraphPad Prism 5.0. Statistically significant differences were calculated using unpaired 2-tailed Student's t. In all cases, P value less than 0.05 were considered significant.  Data are presented as mean±SEM.     Data are presented as mean±SEM. Supporting Information S1 Table. Genes involved inactivation of DNA endogenous promoter