Human cytomegalovirus (HCMV) remains an important cause of mortality and morbidity in immuno-compromised patients such as transplant recipients on immunosuppressive therapy. Furthermore, HCMV is the most common cause of intrauterine infection which can lead to sensorineural hearing loss and mental retardation. Although progress in the diagnosis of HCMV infections has led to improved therapeutic strategies, treatment of HCMV induced disease suffers from the toxicity of the currently available antiviral drugs. In addition, in patients where longer therapy is necessary the emergence of drug-resistant viral strains is frequent. Therefore, the development of novel antiviral drugs is urgently required in order to improve the therapy of HCMV infections. Thus, one important aim of our laboratory is the characterization of molecular events that can be used as targets for new antiviral strategies.
Microbial infections are not only controlled by innate and adaptive immune mechanisms but also by cellular restriction factors which give cells the capacity to resist pathogens. Unlike the innate and adaptive part of the immune system, that require pathogen-induced signaling cascades in order to be switched on, these so-called intrinsic immune mechanisms are mediated by cellular proteins that are constitutively expressed and active before a pathogen enters the cell, thus serving as a front-line defense. Our laboratory investigates whether a subnuclear structure, termed PML nuclear bodies (PML-NBs), contributes to the resistance of cells against herpesvirus infections. PML-NBs are dot-like structures of the cell nucleus, that are defined by the distinct accumulation of specific cellular proteins like PML, hDaxx, Sp100 and ATRX (Figure 1). During the last ten years our laboratory could show, that these proteins act as cellular restriction factors by inducing a silencing of viral gene expression. Moreover, our research revealed that specific viral proteins (pp71 and IE1 of HCMV) are able to antagonize this cellular silencing mechanism (Figure 1). This establishes a delicate balance between cellular defense and viral antagonism which determines whether the herpesvirus enters the productive cycle or viral gene expression is silenced.
During the last two years we could delineate the molecular mechanism how IE1 antagonizes PML-NBs. As evident from the crystal structure of IE1 that we recently solved in collaboration with Prof. Y. Muller and Prof. H. Sticht, this protein directly interacts with the PML coiled-coil domain via its globular core region and disrupts NB foci by inducing a loss of PML SUMOylation. We were able to demonstrate that IE1 acts via abrogating the de novo SUMOylation of PML. In order to overcome reversible SUMOylation dynamics, we made use of a cell-based assay that combines inducible IE1 expression with a SUMO mutant resistant to SUMO proteases. Interestingly, we observed that IE1 expression did not affect preSUMOylated PML, however, it clearly prevented de novo SUMO conjugation. Consistent results were obtained by in vitro SUMOylation assays demonstrating that IE1 alone is sufficient for this effect. Furthermore, IE1 acts in a selective manner since K160 was identified as the main target lysine. This is strengthened by the fact that IE1 also prevents As2O3-mediated hyperSUMOylation of K160 thereby blocking PML degradation. Since IE1 did not interfere with coiled-coil mediated PML dimerization we propose that IE1 either affects PML autoSUMOylation by directly abrogating PML E3 ligase function or by preventing the access to SUMO sites. Thus, our data suggest a novel mechanism how a viral protein counteracts a cellular restriction factor by selectively preventing the de novo SUMOylation at specific lysine residues without affecting global protein SUMOylation.
Studies performed during the last years provided evidence for an extended cross-talk between intrinsic and innate immune mechanisms in relation to PML-NBs. For instance, specific PML-NB factors such as PML and Sp100 are known to be induced by interferons (IFN) and the overall size and number of PML-NBs also increases upon IFN treatment of cells. This indicates that the restrictive function of PML-NBs towards viral infections can be enhanced via IFNs. However, recent evidence now suggests that PML itself acts as an important co-regulatory factor during the induction of interferon stimulated genes (ISGs) (Figure 1). For instance, we observed in collaboration with PD. Dr. P. Hemmerich (Jena) that the PML protein is able to promote interferon-γ (IFN-γ) induced MHC class II gene expression. However, the activities of PML as a coactivator of the IFN response are not confined to type II IFNs but also extend to type I IFN-regulated genes. This was observed in our recent study that detected a significantly reduced induction of specific ISGs after type I IFN treatment of cells exhibiting a depletion of the endogenous PML protein. The molecular mechanism by which PML stimulates the induction of type I IFN-regulated genes is not completely clear so far, but may involve the PML-NB-mediated stabilization of transcriptional complexes that control ISG expression. Consequently, as we were able to demonstrate for the IE1 protein of HCMV, viral proteins that modulate PML-NBs may not only antagonize intrinsic antiviral defense but also innate immune responses.
The ability of human cytomegalovirus (HCMV) to establish lifelong infection in humans and to reactivate from latency in immunocompromised individuals underscores the importance of understanding the mechanisms which regulate latency and reactivation. As primary targets of HCMV, monocytes are considered as a site of latency, and differentiation of these cells to a mature phenotype is linked to reactivation. Silencing of viral IE gene expression during latency is achieved by a repressive chromatin structure around the major IE promoter. Since results from chromatin immunoprecipitation experiments performed in our laboratory indicated that PML-NB proteins mediate the silencing of IE gene expression via epigenetic mechanisms we hypothesized that PML-NB factors might be of importance for the establishment of HCMV latency. In order to further clarify this, we utilized the monocytic cell line THP-1 as an in vitro latency model for human cytomegalovirus infection (HCMV). Characterization of THP-1 cells by immunofluorescence and Western blot analysis confirmed the expression of all major PML-NB components. THP-1 cells with a stable, individual knockdown of either PML, hDaxx or Sp100 were generated. Importantly, depletion of the major PML-NB proteins did not prevent the terminal cellular differentiation of THP-1 monocytes. After construction of a recombinant, endotheliotropic human cytomegalovirus expressing IE2-EYFP, we investigated whether the depletion of PML-NB proteins affects the onset of viral IE gene expression. No effect was observed in non-differentiated monocytes, indicating that PML-NBs do not serve as key determinants for the establishment of HCMV latency. In contrast, an increase in the number of IE-expressing cells was readily detected both after infection of differentiated, THP-1-derived macrophages and during differentiation induced latency. We conclude that PML, hDaxx and Sp100 not only act as cellular restriction factors during lytic HCMV replication but also act as a cellular barrier during the dynamic process of reactivation from latency.
The open reading frame UL69 of human cytomegalovirus encodes the multi-functional, regulatory protein pUL69 which acts as a viral mRNA export factor via its nucleocytoplasmic shuttling activity and via recruitment of the cellular mRNA export machinery by interacting with the cellular DExD/H-box RNA helicase UAP56. We reported previously, that subcellular localization and mRNA export activity of pUL69 are modulated via phosphorylation by the viral kinase pUL97. In recent studies were able to obtain evidence that pUL69 is subject to additional posttranslational modifications like arginine methylation. First, we demonstrated a specific immunoprecipitation of full-length pUL69 as well as pUL69aa1-146 by a mono/dimethylarginine-specific antibody. Second, we observed a specific electrophoretic mobility shift upon overexpression of the catalytically active protein arginine methyltransferase 6 (PRMT6). Third, a direct interaction of pUL69 and PRMT6 was confirmed by yeast two-hybrid and coimmunoprecipitation analyses. We mapped the PRMT6 interaction motif to the pUL69 N terminus and identified critical amino acids within the arginine-rich R1 box of pUL69 that were crucial for PRMT6 and/or UAP56 recruitment. In order to test the impact of putative methylation substrates on the functions of pUL69, we constructed various pUL69 derivatives harboring arginine-to-alanine substitutions and tested them for RNA export activity. Thus, we were able to discriminate between arginines within the R1 box of pUL69 that were crucial for UAP56/PRMT6-interaction and/or mRNA export activity. Remarkably, nuclear magnetic resonance (NMR) analyses revealed the same α-helical structures for pUL69 sequences encoding either the wild type R1/R2 boxes or a UAP56/PRMT6 binding-deficient derivative, thereby excluding the possibility that R/A amino acid substitutions within R1 affected the secondary structure of pUL69. We therefore conclude that the pUL69 N terminus is methylated by PRMT6 and that this critically affects the functions of pUL69 for efficient mRNA export and replication of human cytomegalovirus. Furthermore, we set out to identify phosphorylation sites within the functionally important N-terminus of pUL69 and characterized the in vivo importance for HCMV-replication. In silico analyses predicted the existence of 13 serines and 3 threonines as putative phosphorylation sites within the functionally important N-terminal 140 amino acids of pUL69. MS-based phosphosite mapping of immunopurified FLAG-pUL69 identified S18, S51 and T52 as highly probable phosphorylation sites in transfected HEK293T cells. CoIP analyses revealed that combinatorial exchange of S13+15+16+18 to alanine abolished pUL69´s capacity to interact with UAP56. We therefore hypothesize that cellular kinases phosphorylate this serine stretch upstream of alpha-helix1 and thereby determine the recruitment of UAP56-recruitment by pUL69. By comparing the expression profiles of pUL69-mutants that carry individual or combinatorial substitutions of putative phosphosites via Phos-tag-SDS-PAGE, we provide strong evidence that several serines were phosphorylated when pUL97 wildtype but not when its catalytically inactive derivative pUL97-K355M were coexpressed. Since pS46 and pS49 within alpha-helix2 are both preceding a proline, we speculated that they might be substrate for Pin1-mediated peptidyl-prolyl cis-trans-isomerization and confirmed a complex formation of pUL69 and Pin1 by CoIP experiments. NMR-experiments to define the impact of phosphorylation on the structure of pUL69 are in progress. In accordance with our biochemical in vitro studies, preliminary multistep growth-curve analyses revealed a severe growth defect of recombinant HB15-derived viruses that carry S/T to A-substitutions within the pUL69 N-terminus. We therefore identified phosphosites of pUL69 which are required for UAP56- or Pin1-recruitement, respectively, and confirmed their in vivo importance for HCMV replication.
G protein-coupled receptors (GPCRs) are key regulators of numerous cellular processes. Thus, virally encoded GPCRs illustrate an effective means to bypass the immune system, modulate cellular functions and redirect cellular signaling networks. Human cytomegalovirus (HCMV) encodes four GPCR homologues, termed pUS27, pUS28, pUL33, and pUL78. Whereas pUS28 and pUL33 constitutively activate multiple signaling cascades, the functions of pUS27 and pUL78 are not fully understood, yet. To determine whether pUS27 or pUL78 display any signaling properties, we performed luciferase reporter assays in 293T cells transduced with CREB-, NFAT and NF-κB-specific reporter constructs. Our experiments demonstrate for the first time that pUS27 activates NF-κB dependent gene expression. Intriguingly, it turned out that NF-κB activation differed significantly depending on whether N- or C-terminally tagged versions of pUS27 were applied: while transfection of N-terminally tagged pUS27 did not activate NF-κB, the expression of C-terminally tagged versions strongly induced NF-κB signaling. Disruption of a putative PDZ domain binding motif by adding one serine to the C-terminus of pUS27 induced high NF-κB signaling suggesting that pUS27 may be negatively regulated via a PDZ domain protein. Bioinformatic analysis revealed the existence of four putative TRAF binding motifs within the C-terminus of pUS27. Interestingly, mutation of a predicted TRAF6 binding motif led to a complete loss of NF-κB signaling. To answer the question whether the C-terminal region of pUS27 alone is essential and sufficient for NF-κB activation, we generated chimeric proteins with either CD8 or GFP. These chimeras strongly activated NF-κB signaling independent of their localization at either the cell surface or endosomes, respectively. This indicates that the subcellular localization of the pUS27 cytoplasmic domain is not critical for NF-κB activation. Moreover, we could show that pUS27 specifically induced the canonical NF-κB pathway through TRAF6 by using either ACHP, an inhibitor of IKKβ phosphorylation, or a dominant negative IκBα. Taken together, our data reveal a novel and highly complex signaling capability of the HCMV GPCR pUS27.